Documentation/ABI/testing/debugfs-aufs | 55 + Documentation/ABI/testing/sysfs-aufs | 31 + Documentation/admin-guide/kernel-parameters.txt | 20 + Documentation/admin-guide/sysctl/kernel.rst | 34 + Documentation/admin-guide/sysctl/vm.rst | 41 +- Documentation/filesystems/aufs/README | 396 ++ Documentation/filesystems/aufs/design/01intro.txt | 171 + Documentation/filesystems/aufs/design/02struct.txt | 258 ++ Documentation/filesystems/aufs/design/03atomic_open.txt | 85 + Documentation/filesystems/aufs/design/03lookup.txt | 113 + Documentation/filesystems/aufs/design/04branch.txt | 74 + Documentation/filesystems/aufs/design/05wbr_policy.txt | 64 + Documentation/filesystems/aufs/design/06dirren.dot | 31 + Documentation/filesystems/aufs/design/06dirren.txt | 102 + Documentation/filesystems/aufs/design/06fhsm.txt | 120 + Documentation/filesystems/aufs/design/06mmap.txt | 72 + Documentation/filesystems/aufs/design/06xattr.txt | 96 + Documentation/filesystems/aufs/design/07export.txt | 58 + Documentation/filesystems/aufs/design/08shwh.txt | 52 + Documentation/filesystems/aufs/design/10dynop.txt | 47 + Documentation/locking/futex2.rst | 198 + Documentation/locking/index.rst | 1 + Documentation/scheduler/sched-BFS.txt | 351 ++ Documentation/scheduler/sched-MuQSS.txt | 373 ++ Documentation/vm/index.rst | 1 + Documentation/vm/multigen_lru.rst | 143 + Documentation/vm/uksm.txt | 61 + MAINTAINERS | 15 +- Makefile | 7 +- arch/Kconfig | 9 + arch/alpha/Kconfig | 2 + arch/arm/Kconfig | 2 + arch/arm/tools/syscall.tbl | 4 + arch/arm64/Kconfig | 2 + arch/arm64/include/asm/unistd.h | 2 +- arch/arm64/include/asm/unistd32.h | 4 + arch/powerpc/Kconfig | 2 + arch/powerpc/platforms/cell/spufs/sched.c | 2 + arch/s390/include/asm/preempt.h | 8 +- arch/x86/Kconfig | 19 + arch/x86/Kconfig.cpu | 332 +- arch/x86/Makefile | 47 +- arch/x86/entry/syscalls/syscall_32.tbl | 4 + arch/x86/entry/syscalls/syscall_64.tbl | 4 + arch/x86/include/asm/pci.h | 6 + arch/x86/include/asm/pgtable.h | 2 +- arch/x86/include/asm/vermagic.h | 66 + arch/x86/mm/pgtable.c | 5 +- arch/x86/pci/common.c | 7 +- block/Kconfig.iosched | 9 + block/bfq-iosched.c | 3 + block/blk-merge.c | 16 - block/elevator.c | 11 + block/mq-deadline.c | 11 + drivers/Makefile | 15 +- drivers/accessibility/speakup/speakup_acntpc.c | 4 +- drivers/accessibility/speakup/speakup_apollo.c | 2 +- drivers/accessibility/speakup/speakup_decext.c | 2 +- drivers/accessibility/speakup/speakup_decpc.c | 2 +- drivers/accessibility/speakup/speakup_dectlk.c | 2 +- drivers/accessibility/speakup/speakup_dtlk.c | 4 +- drivers/accessibility/speakup/speakup_keypc.c | 4 +- drivers/accessibility/speakup/synth.c | 8 +- drivers/ata/ahci.c | 23 +- drivers/block/loop.c | 18 + drivers/block/swim.c | 6 +- drivers/bluetooth/btusb.c | 22 +- drivers/char/ipmi/ipmi_msghandler.c | 2 +- drivers/char/ipmi/ipmi_ssif.c | 2 +- drivers/comedi/drivers/ni_mio_common.c | 2 +- drivers/cpufreq/Kconfig | 1 + drivers/cpufreq/cpufreq_ondemand.c | 14 +- drivers/cpufreq/intel_pstate.c | 2 + drivers/gpu/drm/vmwgfx/vmwgfx_cmd.c | 2 +- drivers/gpu/drm/vmwgfx/vmwgfx_irq.c | 2 +- drivers/hwmon/fam15h_power.c | 2 +- drivers/i2c/busses/Kconfig | 9 + drivers/i2c/busses/Makefile | 1 + drivers/i2c/busses/i2c-nct6775.c | 647 +++ drivers/i2c/busses/i2c-piix4.c | 4 +- drivers/iio/light/tsl2563.c | 2 +- drivers/input/touchscreen/atmel_mxt_ts.c | 2 +- drivers/media/i2c/msp3400-driver.c | 4 +- drivers/media/pci/cx18/cx18-gpio.c | 4 +- drivers/media/pci/ivtv/ivtv-gpio.c | 6 +- drivers/media/pci/ivtv/ivtv-ioctl.c | 2 +- drivers/media/pci/ivtv/ivtv-streams.c | 2 +- drivers/media/radio/radio-mr800.c | 2 +- drivers/media/radio/radio-tea5777.c | 2 +- drivers/media/radio/tea575x.c | 2 +- drivers/mfd/ucb1x00-core.c | 2 +- drivers/misc/sgi-xp/xpc_channel.c | 2 +- drivers/net/caif/caif_hsi.c | 2 +- drivers/net/can/usb/peak_usb/pcan_usb.c | 2 +- drivers/net/usb/lan78xx.c | 2 +- drivers/net/usb/usbnet.c | 2 +- drivers/net/wireless/intel/ipw2x00/ipw2100.c | 4 +- drivers/parport/ieee1284.c | 2 +- drivers/parport/ieee1284_ops.c | 2 +- drivers/pci/controller/Makefile | 6 + drivers/pci/controller/intel-nvme-remap.c | 462 +++ drivers/pci/quirks.c | 101 + drivers/platform/x86/intel_ips.c | 8 +- drivers/rtc/rtc-wm8350.c | 6 +- drivers/scsi/Kconfig | 2 + drivers/scsi/Makefile | 1 + drivers/scsi/fnic/fnic_scsi.c | 29 +- drivers/scsi/lpfc/lpfc_scsi.c | 2 +- drivers/scsi/snic/snic_scsi.c | 2 +- drivers/scsi/vhba/Kconfig | 9 + drivers/scsi/vhba/Makefile | 4 + drivers/scsi/vhba/vhba.c | 1105 ++++++ drivers/staging/rts5208/rtsx.c | 2 +- drivers/staging/unisys/visornic/visornic_main.c | 6 +- drivers/usb/gadget/function/u_serial.c | 5 + drivers/video/fbdev/omap/hwa742.c | 2 +- drivers/video/fbdev/pxafb.c | 2 +- drivers/video/logo/logo_linux_clut224.ppm | 16006 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------- drivers/video/logo/logo_linux_mono.pbm | 273 +- drivers/video/logo/logo_linux_vga16.ppm | 16006 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++-------- fs/Kconfig | 1 + fs/Makefile | 1 + fs/aufs/Kconfig | 199 + fs/aufs/Makefile | 46 + fs/aufs/aufs.h | 62 + fs/aufs/branch.c | 1427 +++++++ fs/aufs/branch.h | 375 ++ fs/aufs/conf.mk | 40 + fs/aufs/cpup.c | 1457 +++++++ fs/aufs/cpup.h | 100 + fs/aufs/dbgaufs.c | 526 +++ fs/aufs/dbgaufs.h | 53 + fs/aufs/dcsub.c | 225 ++ fs/aufs/dcsub.h | 137 + fs/aufs/debug.c | 441 +++ fs/aufs/debug.h | 226 ++ fs/aufs/dentry.c | 1160 ++++++ fs/aufs/dentry.h | 269 ++ fs/aufs/dinfo.c | 554 +++ fs/aufs/dir.c | 765 ++++ fs/aufs/dir.h | 134 + fs/aufs/dirren.c | 1316 +++++++ fs/aufs/dirren.h | 140 + fs/aufs/dynop.c | 368 ++ fs/aufs/dynop.h | 77 + fs/aufs/export.c | 837 ++++ fs/aufs/f_op.c | 771 ++++ fs/aufs/fhsm.c | 427 ++ fs/aufs/file.c | 863 ++++ fs/aufs/file.h | 342 ++ fs/aufs/finfo.c | 149 + fs/aufs/fstype.h | 401 ++ fs/aufs/hbl.h | 65 + fs/aufs/hfsnotify.c | 288 ++ fs/aufs/hfsplus.c | 60 + fs/aufs/hnotify.c | 715 ++++ fs/aufs/i_op.c | 1513 +++++++ fs/aufs/i_op_add.c | 941 +++++ fs/aufs/i_op_del.c | 515 +++ fs/aufs/i_op_ren.c | 1251 ++++++ fs/aufs/iinfo.c | 286 ++ fs/aufs/inode.c | 531 +++ fs/aufs/inode.h | 705 ++++ fs/aufs/ioctl.c | 220 ++ fs/aufs/lcnt.h | 186 + fs/aufs/loop.c | 148 + fs/aufs/loop.h | 55 + fs/aufs/magic.mk | 31 + fs/aufs/module.c | 273 ++ fs/aufs/module.h | 166 + fs/aufs/mvdown.c | 706 ++++ fs/aufs/opts.c | 1880 +++++++++ fs/aufs/opts.h | 225 ++ fs/aufs/plink.c | 516 +++ fs/aufs/poll.c | 51 + fs/aufs/posix_acl.c | 106 + fs/aufs/procfs.c | 170 + fs/aufs/rdu.c | 384 ++ fs/aufs/rwsem.h | 85 + fs/aufs/sbinfo.c | 314 ++ fs/aufs/super.c | 1050 +++++ fs/aufs/super.h | 587 +++ fs/aufs/sysaufs.c | 93 + fs/aufs/sysaufs.h | 102 + fs/aufs/sysfs.c | 374 ++ fs/aufs/sysrq.c | 149 + fs/aufs/vdir.c | 896 +++++ fs/aufs/vfsub.c | 916 +++++ fs/aufs/vfsub.h | 358 ++ fs/aufs/wbr_policy.c | 830 ++++ fs/aufs/whout.c | 1070 +++++ fs/aufs/whout.h | 87 + fs/aufs/wkq.c | 372 ++ fs/aufs/wkq.h | 89 + fs/aufs/xattr.c | 368 ++ fs/aufs/xino.c | 1925 +++++++++ fs/dcache.c | 4 +- fs/exec.c | 4 + fs/fcntl.c | 5 +- fs/file_table.c | 2 + fs/fuse/dev.c | 3 +- fs/inode.c | 4 +- fs/namespace.c | 9 + fs/notify/group.c | 1 + fs/open.c | 1 + fs/proc/base.c | 4 +- fs/proc/meminfo.c | 4 + fs/proc/nommu.c | 5 +- fs/proc/task_mmu.c | 7 +- fs/proc/task_nommu.c | 5 +- fs/read_write.c | 2 + fs/splice.c | 12 +- fs/sync.c | 3 +- fs/xattr.c | 1 + include/linux/blkdev.h | 16 + include/linux/cgroup.h | 15 +- include/linux/compat.h | 23 + include/linux/elevator.h | 2 +- include/linux/freezer.h | 3 + include/linux/fs.h | 5 + include/linux/init_task.h | 4 + include/linux/ioprio.h | 2 + include/linux/ksm.h | 43 +- include/linux/lockdep.h | 2 + include/linux/memcontrol.h | 7 +- include/linux/mm.h | 27 + include/linux/mm_inline.h | 234 ++ include/linux/mm_types.h | 112 + include/linux/mmzone.h | 120 + include/linux/mnt_namespace.h | 3 + include/linux/nodemask.h | 1 + include/linux/page-flags-layout.h | 19 +- include/linux/page-flags.h | 4 +- include/linux/pgtable.h | 21 +- include/linux/sched.h | 89 +- include/linux/sched/deadline.h | 9 + include/linux/sched/nohz.h | 2 +- include/linux/sched/prio.h | 6 + include/linux/sched/rt.h | 2 + include/linux/sched/task.h | 2 +- include/linux/skip_list.h | 33 + include/linux/splice.h | 6 + include/linux/sradix-tree.h | 77 + include/linux/swap.h | 4 +- include/linux/syscalls.h | 18 + include/linux/tcp.h | 3 +- include/linux/uksm.h | 149 + include/linux/user_namespace.h | 4 + include/net/inet_connection_sock.h | 5 +- include/net/tcp.h | 49 +- include/uapi/asm-generic/unistd.h | 14 +- include/uapi/linux/Kbuild | 1 + include/uapi/linux/aufs_type.h | 452 +++ include/uapi/linux/futex.h | 51 + include/uapi/linux/inet_diag.h | 33 + include/uapi/linux/sched.h | 9 +- init/Kconfig | 81 +- init/init_task.c | 10 + init/main.c | 2 + kernel/Kconfig.MuQSS | 106 + kernel/Makefile | 4 +- kernel/bounds.c | 6 + kernel/delayacct.c | 2 +- kernel/events/uprobes.c | 2 +- kernel/exit.c | 5 +- kernel/fork.c | 30 +- kernel/futex.c | 351 +- kernel/futex2.c | 1297 ++++++ kernel/irq/Kconfig | 17 + kernel/irq/manage.c | 11 + kernel/kthread.c | 31 +- kernel/locking/lockdep.c | 4 +- kernel/sched/Makefile | 10 +- kernel/sched/MuQSS.c | 8262 +++++++++++++++++++++++++++++++++++++++ kernel/sched/MuQSS.h | 1084 ++++++ kernel/sched/core.c | 6 + kernel/sched/cpufreq_schedutil.c | 4 + kernel/sched/cpupri.h | 2 + kernel/sched/cputime.c | 22 +- kernel/sched/fair.c | 25 + kernel/sched/idle.c | 4 + kernel/sched/sched.h | 37 + kernel/sched/topology.c | 8 + kernel/skip_list.c | 148 + kernel/sys_ni.c | 6 + kernel/sysctl.c | 117 +- kernel/task_work.c | 1 + kernel/time/Kconfig | 2 +- kernel/time/clockevents.c | 6 + kernel/time/hrtimer.c | 112 + kernel/time/posix-cpu-timers.c | 4 +- kernel/time/timer.c | 47 +- kernel/trace/trace_selftest.c | 5 + kernel/user_namespace.c | 7 + lib/Makefile | 2 +- lib/sradix-tree.c | 476 +++ mm/Kconfig | 119 + mm/Makefile | 5 +- mm/filemap.c | 2 +- mm/huge_memory.c | 9 +- mm/internal.h | 1 + mm/khugepaged.c | 2 +- mm/ksm.c | 11 - mm/memcontrol.c | 28 + mm/memory.c | 43 +- mm/migrate.c | 2 +- mm/mm_init.c | 6 +- mm/mmap.c | 70 +- mm/mmzone.c | 2 + mm/nommu.c | 10 +- mm/page_alloc.c | 30 +- mm/prfile.c | 86 + mm/rmap.c | 6 + mm/swap.c | 22 +- mm/swapfile.c | 6 +- mm/uksm.c | 5614 ++++++++++++++++++++++++++ mm/userfaultfd.c | 2 +- mm/vmscan.c | 2961 +++++++++++++- mm/vmstat.c | 3 + mm/workingset.c | 167 +- net/core/pktgen.c | 2 +- net/ipv4/Kconfig | 22 + net/ipv4/Makefile | 1 + net/ipv4/bpf_tcp_ca.c | 2 +- net/ipv4/tcp.c | 1 + net/ipv4/tcp_bbr.c | 38 +- net/ipv4/tcp_bbr2.c | 2671 +++++++++++++ net/ipv4/tcp_cong.c | 1 + net/ipv4/tcp_input.c | 38 +- net/ipv4/tcp_output.c | 25 +- net/ipv4/tcp_rate.c | 36 +- net/ipv4/tcp_timer.c | 1 + scripts/headers_install.sh | 1 + scripts/mkcompile_h | 4 +- security/security.c | 8 + sound/pci/maestro3.c | 4 +- sound/soc/codecs/rt5631.c | 4 +- sound/soc/codecs/wm8350.c | 12 +- sound/soc/codecs/wm8900.c | 2 +- sound/soc/codecs/wm9713.c | 4 +- sound/soc/soc-dapm.c | 2 +- sound/usb/card.h | 9 +- sound/usb/endpoint.c | 1 + sound/usb/line6/pcm.c | 2 +- sound/usb/pcm.c | 214 +- sound/usb/pcm.h | 3 - tools/include/uapi/asm-generic/unistd.h | 11 +- tools/perf/arch/x86/entry/syscalls/syscall_64.tbl | 3 + tools/testing/selftests/futex/functional/.gitignore | 3 + tools/testing/selftests/futex/functional/Makefile | 8 +- tools/testing/selftests/futex/functional/futex2_requeue.c | 164 + tools/testing/selftests/futex/functional/futex2_wait.c | 209 + tools/testing/selftests/futex/functional/futex2_waitv.c | 157 + tools/testing/selftests/futex/functional/futex_wait_timeout.c | 58 +- tools/testing/selftests/futex/functional/futex_wait_wouldblock.c | 33 +- tools/testing/selftests/futex/functional/run.sh | 6 + tools/testing/selftests/futex/include/futex2test.h | 121 + 357 files changed, 97168 insertions(+), 4112 deletions(-) diff --git a/Documentation/ABI/testing/debugfs-aufs b/Documentation/ABI/testing/debugfs-aufs new file mode 100644 index 00000000000..45b739879d7 --- /dev/null +++ b/Documentation/ABI/testing/debugfs-aufs @@ -0,0 +1,55 @@ +What: /debug/aufs/si_/ +Date: March 2009 +Contact: J. R. Okajima +Description: + Under /debug/aufs, a directory named si_ is created + per aufs mount, where is a unique id generated + internally. + +What: /debug/aufs/si_/plink +Date: Apr 2013 +Contact: J. R. Okajima +Description: + It has three lines and shows the information about the + pseudo-link. The first line is a single number + representing a number of buckets. The second line is a + number of pseudo-links per buckets (separated by a + blank). The last line is a single number representing a + total number of psedo-links. + When the aufs mount option 'noplink' is specified, it + will show "1\n0\n0\n". + +What: /debug/aufs/si_/xib +Date: March 2009 +Contact: J. R. Okajima +Description: + It shows the consumed blocks by xib (External Inode Number + Bitmap), its block size and file size. + When the aufs mount option 'noxino' is specified, it + will be empty. About XINO files, see the aufs manual. + +What: /debug/aufs/si_/xi +Date: March 2009 +Contact: J. R. Okajima +Description: + It shows the consumed blocks by xino (External Inode Number + Translation Table), its link count, block size and file + size. + Due to the file size limit, there may exist multiple + xino files per branch. In this case, "-N" is added to + the filename and it corresponds to the index of the + internal xino array. "-0" is omitted. + When the aufs mount option 'noxino' is specified, Those + entries won't exist. About XINO files, see the aufs + manual. + +What: /debug/aufs/si_/xigen +Date: March 2009 +Contact: J. R. Okajima +Description: + It shows the consumed blocks by xigen (External Inode + Generation Table), its block size and file size. + If CONFIG_AUFS_EXPORT is disabled, this entry will not + be created. + When the aufs mount option 'noxino' is specified, it + will be empty. About XINO files, see the aufs manual. diff --git a/Documentation/ABI/testing/sysfs-aufs b/Documentation/ABI/testing/sysfs-aufs new file mode 100644 index 00000000000..48500c0569e --- /dev/null +++ b/Documentation/ABI/testing/sysfs-aufs @@ -0,0 +1,31 @@ +What: /sys/fs/aufs/si_/ +Date: March 2009 +Contact: J. R. Okajima +Description: + Under /sys/fs/aufs, a directory named si_ is created + per aufs mount, where is a unique id generated + internally. + +What: /sys/fs/aufs/si_/br +Date: March 2009 +Contact: J. R. Okajima +Description: + It shows the abolute path of a member directory (which + is called branch) in aufs, and its permission. + +What: /sys/fs/aufs/si_/brid +Date: July 2013 +Contact: J. R. Okajima +Description: + It shows the id of a member directory (which is called + branch) in aufs. + +What: /sys/fs/aufs/si_/xi_path +Date: March 2009 +Contact: J. R. Okajima +Description: + It shows the abolute path of XINO (External Inode Number + Bitmap, Translation Table and Generation Table) file + even if it is the default path. + When the aufs mount option 'noxino' is specified, it + will be empty. About XINO files, see the aufs manual. diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt index 995deccc28b..ca39a2a6697 100644 --- a/Documentation/admin-guide/kernel-parameters.txt +++ b/Documentation/admin-guide/kernel-parameters.txt @@ -1928,6 +1928,9 @@ disable Do not enable intel_pstate as the default scaling driver for the supported processors + enable + Enable intel_pstate in-case "disable" was passed + previously in the kernel boot parameters passive Use intel_pstate as a scaling driver, but configure it to work with generic cpufreq governors (instead of @@ -3798,6 +3801,15 @@ nomsi [MSI] If the PCI_MSI kernel config parameter is enabled, this kernel boot option can be used to disable the use of MSI interrupts system-wide. + pcie_acs_override = + [PCIE] Override missing PCIe ACS support for: + downstream + All downstream ports - full ACS capabilities + multfunction + All multifunction devices - multifunction ACS subset + id:nnnn:nnnn + Specfic device - full ACS capabilities + Specified as vid:did (vendor/device ID) in hex noioapicquirk [APIC] Disable all boot interrupt quirks. Safety option to keep boot IRQs enabled. This should never be necessary. @@ -4868,6 +4880,14 @@ Memory area to be used by remote processor image, managed by CMA. + rqshare= [X86] Select the MuQSS scheduler runqueue sharing type. + Format: + smt -- Share SMT (hyperthread) sibling runqueues + mc -- Share MC (multicore) sibling runqueues + smp -- Share SMP runqueues + none -- So not share any runqueues + Default value is mc + rw [KNL] Mount root device read-write on boot S [KNL] Run init in single mode diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst index 68b21395a74..2881a2c4110 100644 --- a/Documentation/admin-guide/sysctl/kernel.rst +++ b/Documentation/admin-guide/sysctl/kernel.rst @@ -436,6 +436,16 @@ this allows system administrators to override the ``IA64_THREAD_UAC_NOPRINT`` ``prctl`` and avoid logs being flooded. +iso_cpu: (MuQSS CPU scheduler only) +=================================== + +This sets the percentage cpu that the unprivileged SCHED_ISO tasks can +run effectively at realtime priority, averaged over a rolling five +seconds over the -whole- system, meaning all cpus. + +Set to 70 (percent) by default. + + kexec_load_disabled =================== @@ -1078,6 +1088,20 @@ ROM/Flash boot loader. Maybe to tell it what to do after rebooting. ??? +rr_interval: (MuQSS CPU scheduler only) +======================================= + +This is the smallest duration that any cpu process scheduling unit +will run for. Increasing this value can increase throughput of cpu +bound tasks substantially but at the expense of increased latencies +overall. Conversely decreasing it will decrease average and maximum +latencies but at the expense of throughput. This value is in +milliseconds and the default value chosen depends on the number of +cpus available at scheduler initialisation with a minimum of 6. + +Valid values are from 1-1000. + + sched_energy_aware ================== @@ -1527,3 +1551,13 @@ is 10 seconds. The softlockup threshold is (``2 * watchdog_thresh``). Setting this tunable to zero will disable lockup detection altogether. + + +yield_type: (MuQSS CPU scheduler only) +====================================== + +This determines what type of yield calls to sched_yield will perform. + + 0: No yield. + 1: Yield only to better priority/deadline tasks. (default) + 2: Expire timeslice and recalculate deadline. diff --git a/Documentation/admin-guide/sysctl/vm.rst b/Documentation/admin-guide/sysctl/vm.rst index 586cd4b8642..4372a0d671a 100644 --- a/Documentation/admin-guide/sysctl/vm.rst +++ b/Documentation/admin-guide/sysctl/vm.rst @@ -26,6 +26,8 @@ Currently, these files are in /proc/sys/vm: - admin_reserve_kbytes - block_dump +- clean_low_kbytes +- clean_min_kbytes - compact_memory - compaction_proactiveness - compact_unevictable_allowed @@ -106,6 +108,41 @@ On x86_64 this is about 128MB. Changing this takes effect whenever an application requests memory. +clean_low_kbytes +================ + +This knob provides *best-effort* protection of clean file pages. The clean file +pages on the current node won't be reclaimed under memory pressure when their +amount is below vm.clean_low_kbytes *unless* we threaten to OOM or have no +free swap space or vm.swappiness=0. + +Protection of clean file pages may be used to prevent thrashing and +reducing I/O under low-memory conditions. + +Setting it to a high value may result in a early eviction of anonymous pages +into the swap space by attempting to hold the protected amount of clean file +pages in memory. + +The default value is defined by CONFIG_CLEAN_LOW_KBYTES. + + +clean_min_kbytes +================ + +This knob provides *hard* protection of clean file pages. The clean file pages +on the current node won't be reclaimed under memory pressure when their amount +is below vm.clean_min_kbytes. + +Hard protection of clean file pages may be used to avoid high latency and +prevent livelock in near-OOM conditions. + +Setting it to a high value may result in a early out-of-memory condition due to +the inability to reclaim the protected amount of clean file pages when other +types of pages cannot be reclaimed. + +The default value is defined by CONFIG_CLEAN_MIN_KBYTES. + + block_dump ========== @@ -800,6 +837,7 @@ allocated in any single per_cpu_pagelist. This entry only changes the value of hot per cpu pagelists. User can specify a number like 100 to allocate 1/100th of each zone to each per cpu page list. + The batch value of each per cpu pagelist is also updated as a result. It is set to pcp->high/4. The upper limit of batch is (PAGE_SHIFT * 8) @@ -808,6 +846,7 @@ the high water marks for each per cpu page list. If the user writes '0' to this sysctl, it will revert to this default behavior. + stat_interval ============= @@ -999,4 +1038,4 @@ of other processes running on other nodes will not be affected. Allowing regular swap effectively restricts allocations to the local node unless explicitly overridden by memory policies or cpuset -configurations. +configurations. \ No newline at end of file diff --git a/Documentation/filesystems/aufs/README b/Documentation/filesystems/aufs/README new file mode 100644 index 00000000000..d08a56444e2 --- /dev/null +++ b/Documentation/filesystems/aufs/README @@ -0,0 +1,396 @@ + +Aufs5 -- advanced multi layered unification filesystem version 5.x +http://aufs.sf.net +Junjiro R. Okajima + + +0. Introduction +---------------------------------------- +In the early days, aufs was entirely re-designed and re-implemented +Unionfs Version 1.x series. Adding many original ideas, approaches, +improvements and implementations, it became totally different from +Unionfs while keeping the basic features. +Later, Unionfs Version 2.x series began taking some of the same +approaches to aufs1's. +Unionfs was being developed by Professor Erez Zadok at Stony Brook +University and his team. + +Aufs5 supports linux-v5.0 and later, If you want older kernel version +support, +- for linux-v4.x series, try aufs4-linux.git or aufs4-standalone.git +- for linux-v3.x series, try aufs3-linux.git or aufs3-standalone.git +- for linux-v2.6.16 and later, try aufs2-2.6.git, aufs2-standalone.git + or aufs1 from CVS on SourceForge. + +Note: it becomes clear that "Aufs was rejected. Let's give it up." + According to Christoph Hellwig, linux rejects all union-type + filesystems but UnionMount. + + +PS. Al Viro seems have a plan to merge aufs as well as overlayfs and + UnionMount, and he pointed out an issue around a directory mutex + lock and aufs addressed it. But it is still unsure whether aufs will + be merged (or any other union solution). + + + +1. Features +---------------------------------------- +- unite several directories into a single virtual filesystem. The member + directory is called as a branch. +- you can specify the permission flags to the branch, which are 'readonly', + 'readwrite' and 'whiteout-able.' +- by upper writable branch, internal copyup and whiteout, files/dirs on + readonly branch are modifiable logically. +- dynamic branch manipulation, add, del. +- etc... + +Also there are many enhancements in aufs, such as: +- test only the highest one for the directory permission (dirperm1) +- copyup on open (coo=) +- 'move' policy for copy-up between two writable branches, after + checking free space. +- xattr, acl +- readdir(3) in userspace. +- keep inode number by external inode number table +- keep the timestamps of file/dir in internal copyup operation +- seekable directory, supporting NFS readdir. +- whiteout is hardlinked in order to reduce the consumption of inodes + on branch +- do not copyup, nor create a whiteout when it is unnecessary +- revert a single systemcall when an error occurs in aufs +- remount interface instead of ioctl +- maintain /etc/mtab by an external command, /sbin/mount.aufs. +- loopback mounted filesystem as a branch +- kernel thread for removing the dir who has a plenty of whiteouts +- support copyup sparse file (a file which has a 'hole' in it) +- default permission flags for branches +- selectable permission flags for ro branch, whether whiteout can + exist or not +- export via NFS. +- support /fs/aufs and /aufs. +- support multiple writable branches, some policies to select one + among multiple writable branches. +- a new semantics for link(2) and rename(2) to support multiple + writable branches. +- no glibc changes are required. +- pseudo hardlink (hardlink over branches) +- allow a direct access manually to a file on branch, e.g. bypassing aufs. + including NFS or remote filesystem branch. +- userspace wrapper for pathconf(3)/fpathconf(3) with _PC_LINK_MAX. +- and more... + +Currently these features are dropped temporary from aufs5. +See design/08plan.txt in detail. +- nested mount, i.e. aufs as readonly no-whiteout branch of another aufs + (robr) +- statistics of aufs thread (/sys/fs/aufs/stat) + +Features or just an idea in the future (see also design/*.txt), +- reorder the branch index without del/re-add. +- permanent xino files for NFSD +- an option for refreshing the opened files after add/del branches +- light version, without branch manipulation. (unnecessary?) +- copyup in userspace +- inotify in userspace +- readv/writev + + +2. Download +---------------------------------------- +There are three GIT trees for aufs5, aufs5-linux.git, +aufs5-standalone.git, and aufs-util.git. Note that there is no "5" in +"aufs-util.git." +While the aufs-util is always necessary, you need either of aufs5-linux +or aufs5-standalone. + +The aufs5-linux tree includes the whole linux mainline GIT tree, +git://git.kernel.org/.../torvalds/linux.git. +And you cannot select CONFIG_AUFS_FS=m for this version, eg. you cannot +build aufs5 as an external kernel module. +Several extra patches are not included in this tree. Only +aufs5-standalone tree contains them. They are described in the later +section "Configuration and Compilation." + +On the other hand, the aufs5-standalone tree has only aufs source files +and necessary patches, and you can select CONFIG_AUFS_FS=m. +But you need to apply all aufs patches manually. + +You will find GIT branches whose name is in form of "aufs5.x" where "x" +represents the linux kernel version, "linux-5.x". For instance, +"aufs5.0" is for linux-5.0. For latest "linux-5.x-rcN", use +"aufs5.x-rcN" branch. + +o aufs5-linux tree +$ git clone --reference /your/linux/git/tree \ + git://github.com/sfjro/aufs5-linux.git aufs5-linux.git +- if you don't have linux GIT tree, then remove "--reference ..." +$ cd aufs5-linux.git +$ git checkout origin/aufs5.0 + +Or You may want to directly git-pull aufs into your linux GIT tree, and +leave the patch-work to GIT. +$ cd /your/linux/git/tree +$ git remote add aufs5 git://github.com/sfjro/aufs5-linux.git +$ git fetch aufs5 +$ git checkout -b my5.0 v5.0 +$ (add your local change...) +$ git pull aufs5 aufs5.0 +- now you have v5.0 + your_changes + aufs5.0 in you my5.0 branch. +- you may need to solve some conflicts between your_changes and + aufs5.0. in this case, git-rerere is recommended so that you can + solve the similar conflicts automatically when you upgrade to 5.1 or + later in the future. + +o aufs5-standalone tree +$ git clone git://github.com/sfjro/aufs5-standalone.git aufs5-standalone.git +$ cd aufs5-standalone.git +$ git checkout origin/aufs5.0 + +o aufs-util tree +$ git clone git://git.code.sf.net/p/aufs/aufs-util aufs-util.git +- note that the public aufs-util.git is on SourceForge instead of + GitHUB. +$ cd aufs-util.git +$ git checkout origin/aufs5.0 + +Note: The 5.x-rcN branch is to be used with `rc' kernel versions ONLY. +The minor version number, 'x' in '5.x', of aufs may not always +follow the minor version number of the kernel. +Because changes in the kernel that cause the use of a new +minor version number do not always require changes to aufs-util. + +Since aufs-util has its own minor version number, you may not be +able to find a GIT branch in aufs-util for your kernel's +exact minor version number. +In this case, you should git-checkout the branch for the +nearest lower number. + +For (an unreleased) example: +If you are using "linux-5.10" and the "aufs5.10" branch +does not exist in aufs-util repository, then "aufs5.9", "aufs5.8" +or something numerically smaller is the branch for your kernel. + +Also you can view all branches by + $ git branch -a + + +3. Configuration and Compilation +---------------------------------------- +Make sure you have git-checkout'ed the correct branch. + +For aufs5-linux tree, +- enable CONFIG_AUFS_FS. +- set other aufs configurations if necessary. +- for aufs5.13 and later + Because aufs is not only an ordinary filesystem (callee of VFS), but + also a caller of VFS functions for branch filesystems, subclassing of + the internal locks for LOCKDEP is necessary. LOCKDEP is a debugging + feature of linux kernel. If you enable CONFIG_LOCKDEP, then you will + need to customize some LOCKDEP numbers. Here are what I use on my + test environment. + CONFIG_LOCKDEP_BITS=21 + CONFIG_LOCKDEP_CHAINS_BITS=21 + CONFIG_LOCKDEP_STACK_TRACE_BITS=24 + +For aufs5-standalone tree, +There are several ways to build. + +1. +- apply ./aufs5-kbuild.patch to your kernel source files. +- apply ./aufs5-base.patch too. +- apply ./aufs5-mmap.patch too. +- apply ./aufs5-standalone.patch too, if you have a plan to set + CONFIG_AUFS_FS=m. otherwise you don't need ./aufs5-standalone.patch. +- copy ./{Documentation,fs,include/uapi/linux/aufs_type.h} files to your + kernel source tree. Never copy $PWD/include/uapi/linux/Kbuild. +- enable CONFIG_AUFS_FS, you can select either + =m or =y. +- and build your kernel as usual. +- install the built kernel. +- install the header files too by "make headers_install" to the + directory where you specify. By default, it is $PWD/usr. + "make help" shows a brief note for headers_install. +- and reboot your system. + +2. +- module only (CONFIG_AUFS_FS=m). +- apply ./aufs5-base.patch to your kernel source files. +- apply ./aufs5-mmap.patch too. +- apply ./aufs5-standalone.patch too. +- build your kernel, don't forget "make headers_install", and reboot. +- edit ./config.mk and set other aufs configurations if necessary. + Note: You should read $PWD/fs/aufs/Kconfig carefully which describes + every aufs configurations. +- build the module by simple "make". +- you can specify ${KDIR} make variable which points to your kernel + source tree. +- install the files + + run "make install" to install the aufs module, or copy the built + $PWD/aufs.ko to /lib/modules/... and run depmod -a (or reboot simply). + + run "make install_headers" (instead of headers_install) to install + the modified aufs header file (you can specify DESTDIR which is + available in aufs standalone version's Makefile only), or copy + $PWD/usr/include/linux/aufs_type.h to /usr/include/linux or wherever + you like manually. By default, the target directory is $PWD/usr. +- no need to apply aufs5-kbuild.patch, nor copying source files to your + kernel source tree. + +Note: The header file aufs_type.h is necessary to build aufs-util + as well as "make headers_install" in the kernel source tree. + headers_install is subject to be forgotten, but it is essentially + necessary, not only for building aufs-util. + You may not meet problems without headers_install in some older + version though. + +And then, +- read README in aufs-util, build and install it +- note that your distribution may contain an obsoleted version of + aufs_type.h in /usr/include/linux or something. When you build aufs + utilities, make sure that your compiler refers the correct aufs header + file which is built by "make headers_install." +- if you want to use readdir(3) in userspace or pathconf(3) wrapper, + then run "make install_ulib" too. And refer to the aufs manual in + detail. + +There several other patches in aufs5-standalone.git. They are all +optional. When you meet some problems, they will help you. +- aufs5-loopback.patch + Supports a nested loopback mount in a branch-fs. This patch is + unnecessary until aufs produces a message like "you may want to try + another patch for loopback file". +- vfs-ino.patch + Modifies a system global kernel internal function get_next_ino() in + order to stop assigning 0 for an inode-number. Not directly related to + aufs, but recommended generally. +- tmpfs-idr.patch + Keeps the tmpfs inode number as the lowest value. Effective to reduce + the size of aufs XINO files for tmpfs branch. Also it prevents the + duplication of inode number, which is important for backup tools and + other utilities. When you find aufs XINO files for tmpfs branch + growing too much, try this patch. + + +4. Usage +---------------------------------------- +At first, make sure aufs-util are installed, and please read the aufs +manual, aufs.5 in aufs-util.git tree. +$ man -l aufs.5 + +And then, +$ mkdir /tmp/rw /tmp/aufs +# mount -t aufs -o br=/tmp/rw:${HOME} none /tmp/aufs + +Here is another example. The result is equivalent. +# mount -t aufs -o br=/tmp/rw=rw:${HOME}=ro none /tmp/aufs + Or +# mount -t aufs -o br:/tmp/rw none /tmp/aufs +# mount -o remount,append:${HOME} /tmp/aufs + +Then, you can see whole tree of your home dir through /tmp/aufs. If +you modify a file under /tmp/aufs, the one on your home directory is +not affected, instead the same named file will be newly created under +/tmp/rw. And all of your modification to a file will be applied to +the one under /tmp/rw. This is called the file based Copy on Write +(COW) method. +Aufs mount options are described in aufs.5. +If you run chroot or something and make your aufs as a root directory, +then you need to customize the shutdown script. See the aufs manual in +detail. + +Additionally, there are some sample usages of aufs which are a +diskless system with network booting, and LiveCD over NFS. +See sample dir in CVS tree on SourceForge. + + +5. Contact +---------------------------------------- +When you have any problems or strange behaviour in aufs, please let me +know with: +- /proc/mounts (instead of the output of mount(8)) +- /sys/module/aufs/* +- /sys/fs/aufs/* (if you have them) +- /debug/aufs/* (if you have them) +- linux kernel version + if your kernel is not plain, for example modified by distributor, + the url where i can download its source is necessary too. +- aufs version which was printed at loading the module or booting the + system, instead of the date you downloaded. +- configuration (define/undefine CONFIG_AUFS_xxx) +- kernel configuration or /proc/config.gz (if you have it) +- LSM (linux security module, if you are using) +- behaviour which you think to be incorrect +- actual operation, reproducible one is better +- mailto: aufs-users at lists.sourceforge.net + +Usually, I don't watch the Public Areas(Bugs, Support Requests, Patches, +and Feature Requests) on SourceForge. Please join and write to +aufs-users ML. + + +6. Acknowledgements +---------------------------------------- +Thanks to everyone who have tried and are using aufs, whoever +have reported a bug or any feedback. + +Especially donators: +Tomas Matejicek(slax.org) made a donation (much more than once). + Since Apr 2010, Tomas M (the author of Slax and Linux Live + scripts) is making "doubling" donations. + Unfortunately I cannot list all of the donators, but I really + appreciate. + It ends Aug 2010, but the ordinary donation URL is still available. + +Dai Itasaka made a donation (2007/8). +Chuck Smith made a donation (2008/4, 10 and 12). +Henk Schoneveld made a donation (2008/9). +Chih-Wei Huang, ASUS, CTC donated Eee PC 4G (2008/10). +Francois Dupoux made a donation (2008/11). +Bruno Cesar Ribas and Luis Carlos Erpen de Bona, C3SL serves public + aufs2 GIT tree (2009/2). +William Grant made a donation (2009/3). +Patrick Lane made a donation (2009/4). +The Mail Archive (mail-archive.com) made donations (2009/5). +Nippy Networks (Ed Wildgoose) made a donation (2009/7). +New Dream Network, LLC (www.dreamhost.com) made a donation (2009/11). +Pavel Pronskiy made a donation (2011/2). +Iridium and Inmarsat satellite phone retailer (www.mailasail.com), Nippy + Networks (Ed Wildgoose) made a donation for hardware (2011/3). +Max Lekomcev (DOM-TV project) made a donation (2011/7, 12, 2012/3, 6 and +11). +Sam Liddicott made a donation (2011/9). +Era Scarecrow made a donation (2013/4). +Bor Ratajc made a donation (2013/4). +Alessandro Gorreta made a donation (2013/4). +POIRETTE Marc made a donation (2013/4). +Alessandro Gorreta made a donation (2013/4). +lauri kasvandik made a donation (2013/5). +"pemasu from Finland" made a donation (2013/7). +The Parted Magic Project made a donation (2013/9 and 11). +Pavel Barta made a donation (2013/10). +Nikolay Pertsev made a donation (2014/5). +James B made a donation (2014/7 and 2015/7). +Stefano Di Biase made a donation (2014/8). +Daniel Epellei made a donation (2015/1). +OmegaPhil made a donation (2016/1, 2018/4). +Tomasz Szewczyk made a donation (2016/4). +James Burry made a donation (2016/12). +Carsten Rose made a donation (2018/9). +Porteus Kiosk made a donation (2018/10). + +Thank you very much. +Donations are always, including future donations, very important and +helpful for me to keep on developing aufs. + + +7. +---------------------------------------- +If you are an experienced user, no explanation is needed. Aufs is +just a linux filesystem. + + +Enjoy! + +# Local variables: ; +# mode: text; +# End: ; diff --git a/Documentation/filesystems/aufs/design/01intro.txt b/Documentation/filesystems/aufs/design/01intro.txt new file mode 100644 index 00000000000..47e0a01a8af --- /dev/null +++ b/Documentation/filesystems/aufs/design/01intro.txt @@ -0,0 +1,171 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Introduction +---------------------------------------- + +aufs [ei ju: ef es] | /ey-yoo-ef-es/ | [a u f s] +1. abbrev. for "advanced multi-layered unification filesystem". +2. abbrev. for "another unionfs". +3. abbrev. for "auf das" in German which means "on the" in English. + Ex. "Butter aufs Brot"(G) means "butter onto bread"(E). + But "Filesystem aufs Filesystem" is hard to understand. +4. abbrev. for "African Urban Fashion Show". + +AUFS is a filesystem with features: +- multi layered stackable unification filesystem, the member directory + is called as a branch. +- branch permission and attribute, 'readonly', 'real-readonly', + 'readwrite', 'whiteout-able', 'link-able whiteout', etc. and their + combination. +- internal "file copy-on-write". +- logical deletion, whiteout. +- dynamic branch manipulation, adding, deleting and changing permission. +- allow bypassing aufs, user's direct branch access. +- external inode number translation table and bitmap which maintains the + persistent aufs inode number. +- seekable directory, including NFS readdir. +- file mapping, mmap and sharing pages. +- pseudo-link, hardlink over branches. +- loopback mounted filesystem as a branch. +- several policies to select one among multiple writable branches. +- revert a single systemcall when an error occurs in aufs. +- and more... + + +Multi Layered Stackable Unification Filesystem +---------------------------------------------------------------------- +Most people already knows what it is. +It is a filesystem which unifies several directories and provides a +merged single directory. When users access a file, the access will be +passed/re-directed/converted (sorry, I am not sure which English word is +correct) to the real file on the member filesystem. The member +filesystem is called 'lower filesystem' or 'branch' and has a mode +'readonly' and 'readwrite.' And the deletion for a file on the lower +readonly branch is handled by creating 'whiteout' on the upper writable +branch. + +On LKML, there have been discussions about UnionMount (Jan Blunck, +Bharata B Rao and Valerie Aurora) and Unionfs (Erez Zadok). They took +different approaches to implement the merged-view. +The former tries putting it into VFS, and the latter implements as a +separate filesystem. +(If I misunderstand about these implementations, please let me know and +I shall correct it. Because it is a long time ago when I read their +source files last time). + +UnionMount's approach will be able to small, but may be hard to share +branches between several UnionMount since the whiteout in it is +implemented in the inode on branch filesystem and always +shared. According to Bharata's post, readdir does not seems to be +finished yet. +There are several missing features known in this implementations such as +- for users, the inode number may change silently. eg. copy-up. +- link(2) may break by copy-up. +- read(2) may get an obsoleted filedata (fstat(2) too). +- fcntl(F_SETLK) may be broken by copy-up. +- unnecessary copy-up may happen, for example mmap(MAP_PRIVATE) after + open(O_RDWR). + +In linux-3.18, "overlay" filesystem (formerly known as "overlayfs") was +merged into mainline. This is another implementation of UnionMount as a +separated filesystem. All the limitations and known problems which +UnionMount are equally inherited to "overlay" filesystem. + +Unionfs has a longer history. When I started implementing a stackable +filesystem (Aug 2005), it already existed. It has virtual super_block, +inode, dentry and file objects and they have an array pointing lower +same kind objects. After contributing many patches for Unionfs, I +re-started my project AUFS (Jun 2006). + +In AUFS, the structure of filesystem resembles to Unionfs, but I +implemented my own ideas, approaches and enhancements and it became +totally different one. + +Comparing DM snapshot and fs based implementation +- the number of bytes to be copied between devices is much smaller. +- the type of filesystem must be one and only. +- the fs must be writable, no readonly fs, even for the lower original + device. so the compression fs will not be usable. but if we use + loopback mount, we may address this issue. + for instance, + mount /cdrom/squashfs.img /sq + losetup /sq/ext2.img + losetup /somewhere/cow + dmsetup "snapshot /dev/loop0 /dev/loop1 ..." +- it will be difficult (or needs more operations) to extract the + difference between the original device and COW. +- DM snapshot-merge may help a lot when users try merging. in the + fs-layer union, users will use rsync(1). + +You may want to read my old paper "Filesystems in LiveCD" +(http://aufs.sourceforge.net/aufs2/report/sq/sq.pdf). + + +Several characters/aspects/persona of aufs +---------------------------------------------------------------------- + +Aufs has several characters, aspects or persona. +1. a filesystem, callee of VFS helper +2. sub-VFS, caller of VFS helper for branches +3. a virtual filesystem which maintains persistent inode number +4. reader/writer of files on branches such like an application + +1. Callee of VFS Helper +As an ordinary linux filesystem, aufs is a callee of VFS. For instance, +unlink(2) from an application reaches sys_unlink() kernel function and +then vfs_unlink() is called. vfs_unlink() is one of VFS helper and it +calls filesystem specific unlink operation. Actually aufs implements the +unlink operation but it behaves like a redirector. + +2. Caller of VFS Helper for Branches +aufs_unlink() passes the unlink request to the branch filesystem as if +it were called from VFS. So the called unlink operation of the branch +filesystem acts as usual. As a caller of VFS helper, aufs should handle +every necessary pre/post operation for the branch filesystem. +- acquire the lock for the parent dir on a branch +- lookup in a branch +- revalidate dentry on a branch +- mnt_want_write() for a branch +- vfs_unlink() for a branch +- mnt_drop_write() for a branch +- release the lock on a branch + +3. Persistent Inode Number +One of the most important issue for a filesystem is to maintain inode +numbers. This is particularly important to support exporting a +filesystem via NFS. Aufs is a virtual filesystem which doesn't have a +backend block device for its own. But some storage is necessary to +keep and maintain the inode numbers. It may be a large space and may not +suit to keep in memory. Aufs rents some space from its first writable +branch filesystem (by default) and creates file(s) on it. These files +are created by aufs internally and removed soon (currently) keeping +opened. +Note: Because these files are removed, they are totally gone after + unmounting aufs. It means the inode numbers are not persistent + across unmount or reboot. I have a plan to make them really + persistent which will be important for aufs on NFS server. + +4. Read/Write Files Internally (copy-on-write) +Because a branch can be readonly, when you write a file on it, aufs will +"copy-up" it to the upper writable branch internally. And then write the +originally requested thing to the file. Generally kernel doesn't +open/read/write file actively. In aufs, even a single write may cause a +internal "file copy". This behaviour is very similar to cp(1) command. + +Some people may think it is better to pass such work to user space +helper, instead of doing in kernel space. Actually I am still thinking +about it. But currently I have implemented it in kernel space. diff --git a/Documentation/filesystems/aufs/design/02struct.txt b/Documentation/filesystems/aufs/design/02struct.txt new file mode 100644 index 00000000000..0092b3dcfbe --- /dev/null +++ b/Documentation/filesystems/aufs/design/02struct.txt @@ -0,0 +1,258 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Basic Aufs Internal Structure + +Superblock/Inode/Dentry/File Objects +---------------------------------------------------------------------- +As like an ordinary filesystem, aufs has its own +superblock/inode/dentry/file objects. All these objects have a +dynamically allocated array and store the same kind of pointers to the +lower filesystem, branch. +For example, when you build a union with one readwrite branch and one +readonly, mounted /au, /rw and /ro respectively. +- /au = /rw + /ro +- /ro/fileA exists but /rw/fileA + +Aufs lookup operation finds /ro/fileA and gets dentry for that. These +pointers are stored in a aufs dentry. The array in aufs dentry will be, +- [0] = NULL (because /rw/fileA doesn't exist) +- [1] = /ro/fileA + +This style of an array is essentially same to the aufs +superblock/inode/dentry/file objects. + +Because aufs supports manipulating branches, ie. add/delete/change +branches dynamically, these objects has its own generation. When +branches are changed, the generation in aufs superblock is +incremented. And a generation in other object are compared when it is +accessed. When a generation in other objects are obsoleted, aufs +refreshes the internal array. + + +Superblock +---------------------------------------------------------------------- +Additionally aufs superblock has some data for policies to select one +among multiple writable branches, XIB files, pseudo-links and kobject. +See below in detail. +About the policies which supports copy-down a directory, see +wbr_policy.txt too. + + +Branch and XINO(External Inode Number Translation Table) +---------------------------------------------------------------------- +Every branch has its own xino (external inode number translation table) +file. The xino file is created and unlinked by aufs internally. When two +members of a union exist on the same filesystem, they share the single +xino file. +The struct of a xino file is simple, just a sequence of aufs inode +numbers which is indexed by the lower inode number. +In the above sample, assume the inode number of /ro/fileA is i111 and +aufs assigns the inode number i999 for fileA. Then aufs writes 999 as +4(8) bytes at 111 * 4(8) bytes offset in the xino file. + +When the inode numbers are not contiguous, the xino file will be sparse +which has a hole in it and doesn't consume as much disk space as it +might appear. If your branch filesystem consumes disk space for such +holes, then you should specify 'xino=' option at mounting aufs. + +Aufs has a mount option to free the disk blocks for such holes in XINO +files on tmpfs or ramdisk. But it is not so effective actually. If you +meet a problem of disk shortage due to XINO files, then you should try +"tmpfs-ino.patch" (and "vfs-ino.patch" too) in aufs4-standalone.git. +The patch localizes the assignment inumbers per tmpfs-mount and avoid +the holes in XINO files. + +Also a writable branch has three kinds of "whiteout bases". All these +are existed when the branch is joined to aufs, and their names are +whiteout-ed doubly, so that users will never see their names in aufs +hierarchy. +1. a regular file which will be hardlinked to all whiteouts. +2. a directory to store a pseudo-link. +3. a directory to store an "orphan"-ed file temporary. + +1. Whiteout Base + When you remove a file on a readonly branch, aufs handles it as a + logical deletion and creates a whiteout on the upper writable branch + as a hardlink of this file in order not to consume inode on the + writable branch. +2. Pseudo-link Dir + See below, Pseudo-link. +3. Step-Parent Dir + When "fileC" exists on the lower readonly branch only and it is + opened and removed with its parent dir, and then user writes + something into it, then aufs copies-up fileC to this + directory. Because there is no other dir to store fileC. After + creating a file under this dir, the file is unlinked. + +Because aufs supports manipulating branches, ie. add/delete/change +dynamically, a branch has its own id. When the branch order changes, +aufs finds the new index by searching the branch id. + + +Pseudo-link +---------------------------------------------------------------------- +Assume "fileA" exists on the lower readonly branch only and it is +hardlinked to "fileB" on the branch. When you write something to fileA, +aufs copies-up it to the upper writable branch. Additionally aufs +creates a hardlink under the Pseudo-link Directory of the writable +branch. The inode of a pseudo-link is kept in aufs super_block as a +simple list. If fileB is read after unlinking fileA, aufs returns +filedata from the pseudo-link instead of the lower readonly +branch. Because the pseudo-link is based upon the inode, to keep the +inode number by xino (see above) is essentially necessary. + +All the hardlinks under the Pseudo-link Directory of the writable branch +should be restored in a proper location later. Aufs provides a utility +to do this. The userspace helpers executed at remounting and unmounting +aufs by default. +During this utility is running, it puts aufs into the pseudo-link +maintenance mode. In this mode, only the process which began the +maintenance mode (and its child processes) is allowed to operate in +aufs. Some other processes which are not related to the pseudo-link will +be allowed to run too, but the rest have to return an error or wait +until the maintenance mode ends. If a process already acquires an inode +mutex (in VFS), it has to return an error. + + +XIB(external inode number bitmap) +---------------------------------------------------------------------- +Addition to the xino file per a branch, aufs has an external inode number +bitmap in a superblock object. It is also an internal file such like a +xino file. +It is a simple bitmap to mark whether the aufs inode number is in-use or +not. +To reduce the file I/O, aufs prepares a single memory page to cache xib. + +As well as XINO files, aufs has a feature to truncate/refresh XIB to +reduce the number of consumed disk blocks for these files. + + +Virtual or Vertical Dir, and Readdir in Userspace +---------------------------------------------------------------------- +In order to support multiple layers (branches), aufs readdir operation +constructs a virtual dir block on memory. For readdir, aufs calls +vfs_readdir() internally for each dir on branches, merges their entries +with eliminating the whiteout-ed ones, and sets it to file (dir) +object. So the file object has its entry list until it is closed. The +entry list will be updated when the file position is zero and becomes +obsoleted. This decision is made in aufs automatically. + +The dynamically allocated memory block for the name of entries has a +unit of 512 bytes (by default) and stores the names contiguously (no +padding). Another block for each entry is handled by kmem_cache too. +During building dir blocks, aufs creates hash list and judging whether +the entry is whiteouted by its upper branch or already listed. +The merged result is cached in the corresponding inode object and +maintained by a customizable life-time option. + +Some people may call it can be a security hole or invite DoS attack +since the opened and once readdir-ed dir (file object) holds its entry +list and becomes a pressure for system memory. But I'd say it is similar +to files under /proc or /sys. The virtual files in them also holds a +memory page (generally) while they are opened. When an idea to reduce +memory for them is introduced, it will be applied to aufs too. +For those who really hate this situation, I've developed readdir(3) +library which operates this merging in userspace. You just need to set +LD_PRELOAD environment variable, and aufs will not consume no memory in +kernel space for readdir(3). + + +Workqueue +---------------------------------------------------------------------- +Aufs sometimes requires privilege access to a branch. For instance, +in copy-up/down operation. When a user process is going to make changes +to a file which exists in the lower readonly branch only, and the mode +of one of ancestor directories may not be writable by a user +process. Here aufs copy-up the file with its ancestors and they may +require privilege to set its owner/group/mode/etc. +This is a typical case of a application character of aufs (see +Introduction). + +Aufs uses workqueue synchronously for this case. It creates its own +workqueue. The workqueue is a kernel thread and has privilege. Aufs +passes the request to call mkdir or write (for example), and wait for +its completion. This approach solves a problem of a signal handler +simply. +If aufs didn't adopt the workqueue and changed the privilege of the +process, then the process may receive the unexpected SIGXFSZ or other +signals. + +Also aufs uses the system global workqueue ("events" kernel thread) too +for asynchronous tasks, such like handling inotify/fsnotify, re-creating a +whiteout base and etc. This is unrelated to a privilege. +Most of aufs operation tries acquiring a rw_semaphore for aufs +superblock at the beginning, at the same time waits for the completion +of all queued asynchronous tasks. + + +Whiteout +---------------------------------------------------------------------- +The whiteout in aufs is very similar to Unionfs's. That is represented +by its filename. UnionMount takes an approach of a file mode, but I am +afraid several utilities (find(1) or something) will have to support it. + +Basically the whiteout represents "logical deletion" which stops aufs to +lookup further, but also it represents "dir is opaque" which also stop +further lookup. + +In aufs, rmdir(2) and rename(2) for dir uses whiteout alternatively. +In order to make several functions in a single systemcall to be +revertible, aufs adopts an approach to rename a directory to a temporary +unique whiteouted name. +For example, in rename(2) dir where the target dir already existed, aufs +renames the target dir to a temporary unique whiteouted name before the +actual rename on a branch, and then handles other actions (make it opaque, +update the attributes, etc). If an error happens in these actions, aufs +simply renames the whiteouted name back and returns an error. If all are +succeeded, aufs registers a function to remove the whiteouted unique +temporary name completely and asynchronously to the system global +workqueue. + + +Copy-up +---------------------------------------------------------------------- +It is a well-known feature or concept. +When user modifies a file on a readonly branch, aufs operate "copy-up" +internally and makes change to the new file on the upper writable branch. +When the trigger systemcall does not update the timestamps of the parent +dir, aufs reverts it after copy-up. + + +Move-down (aufs3.9 and later) +---------------------------------------------------------------------- +"Copy-up" is one of the essential feature in aufs. It copies a file from +the lower readonly branch to the upper writable branch when a user +changes something about the file. +"Move-down" is an opposite action of copy-up. Basically this action is +ran manually instead of automatically and internally. +For desgin and implementation, aufs has to consider these issues. +- whiteout for the file may exist on the lower branch. +- ancestor directories may not exist on the lower branch. +- diropq for the ancestor directories may exist on the upper branch. +- free space on the lower branch will reduce. +- another access to the file may happen during moving-down, including + UDBA (see "Revalidate Dentry and UDBA"). +- the file should not be hard-linked nor pseudo-linked. they should be + handled by auplink utility later. + +Sometimes users want to move-down a file from the upper writable branch +to the lower readonly or writable branch. For instance, +- the free space of the upper writable branch is going to run out. +- create a new intermediate branch between the upper and lower branch. +- etc. + +For this purpose, use "aumvdown" command in aufs-util.git. diff --git a/Documentation/filesystems/aufs/design/03atomic_open.txt b/Documentation/filesystems/aufs/design/03atomic_open.txt new file mode 100644 index 00000000000..fb8cf0bc8c7 --- /dev/null +++ b/Documentation/filesystems/aufs/design/03atomic_open.txt @@ -0,0 +1,85 @@ + +# Copyright (C) 2015-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Support for a branch who has its ->atomic_open() +---------------------------------------------------------------------- +The filesystems who implement its ->atomic_open() are not majority. For +example NFSv4 does, and aufs should call NFSv4 ->atomic_open, +particularly for open(O_CREAT|O_EXCL, 0400) case. Other than +->atomic_open(), NFSv4 returns an error for this open(2). While I am not +sure whether all filesystems who have ->atomic_open() behave like this, +but NFSv4 surely returns the error. + +In order to support ->atomic_open() for aufs, there are a few +approaches. + +A. Introduce aufs_atomic_open() + - calls one of VFS:do_last(), lookup_open() or atomic_open() for + branch fs. +B. Introduce aufs_atomic_open() calling create, open and chmod. this is + an aufs user Pip Cet's approach + - calls aufs_create(), VFS finish_open() and notify_change(). + - pass fake-mode to finish_open(), and then correct the mode by + notify_change(). +C. Extend aufs_open() to call branch fs's ->atomic_open() + - no aufs_atomic_open(). + - aufs_lookup() registers the TID to an aufs internal object. + - aufs_create() does nothing when the matching TID is registered, but + registers the mode. + - aufs_open() calls branch fs's ->atomic_open() when the matching + TID is registered. +D. Extend aufs_open() to re-try branch fs's ->open() with superuser's + credential + - no aufs_atomic_open(). + - aufs_create() registers the TID to an internal object. this info + represents "this process created this file just now." + - when aufs gets EACCES from branch fs's ->open(), then confirm the + registered TID and re-try open() with superuser's credential. + +Pros and cons for each approach. + +A. + - straightforward but highly depends upon VFS internal. + - the atomic behavaiour is kept. + - some of parameters such as nameidata are hard to reproduce for + branch fs. + - large overhead. +B. + - easy to implement. + - the atomic behavaiour is lost. +C. + - the atomic behavaiour is kept. + - dirty and tricky. + - VFS checks whether the file is created correctly after calling + ->create(), which means this approach doesn't work. +D. + - easy to implement. + - the atomic behavaiour is lost. + - to open a file with superuser's credential and give it to a user + process is a bad idea, since the file object keeps the credential + in it. It may affect LSM or something. This approach doesn't work + either. + +The approach A is ideal, but it hard to implement. So here is a +variation of A, which is to be implemented. + +A-1. Introduce aufs_atomic_open() + - calls branch fs ->atomic_open() if exists. otherwise calls + vfs_create() and finish_open(). + - the demerit is that the several checks after branch fs + ->atomic_open() are lost. in the ordinary case, the checks are + done by VFS:do_last(), lookup_open() and atomic_open(). some can + be implemented in aufs, but not all I am afraid. diff --git a/Documentation/filesystems/aufs/design/03lookup.txt b/Documentation/filesystems/aufs/design/03lookup.txt new file mode 100644 index 00000000000..5c3c97f11c5 --- /dev/null +++ b/Documentation/filesystems/aufs/design/03lookup.txt @@ -0,0 +1,113 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Lookup in a Branch +---------------------------------------------------------------------- +Since aufs has a character of sub-VFS (see Introduction), it operates +lookup for branches as VFS does. It may be a heavy work. But almost all +lookup operation in aufs is the simplest case, ie. lookup only an entry +directly connected to its parent. Digging down the directory hierarchy +is unnecessary. VFS has a function lookup_one_len() for that use, and +aufs calls it. + +When a branch is a remote filesystem, aufs basically relies upon its +->d_revalidate(), also aufs forces the hardest revalidate tests for +them. +For d_revalidate, aufs implements three levels of revalidate tests. See +"Revalidate Dentry and UDBA" in detail. + + +Test Only the Highest One for the Directory Permission (dirperm1 option) +---------------------------------------------------------------------- +Let's try case study. +- aufs has two branches, upper readwrite and lower readonly. + /au = /rw + /ro +- "dirA" exists under /ro, but /rw. and its mode is 0700. +- user invoked "chmod a+rx /au/dirA" +- the internal copy-up is activated and "/rw/dirA" is created and its + permission bits are set to world readable. +- then "/au/dirA" becomes world readable? + +In this case, /ro/dirA is still 0700 since it exists in readonly branch, +or it may be a natively readonly filesystem. If aufs respects the lower +branch, it should not respond readdir request from other users. But user +allowed it by chmod. Should really aufs rejects showing the entries +under /ro/dirA? + +To be honest, I don't have a good solution for this case. So aufs +implements 'dirperm1' and 'nodirperm1' mount options, and leave it to +users. +When dirperm1 is specified, aufs checks only the highest one for the +directory permission, and shows the entries. Otherwise, as usual, checks +every dir existing on all branches and rejects the request. + +As a side effect, dirperm1 option improves the performance of aufs +because the number of permission check is reduced when the number of +branch is many. + + +Revalidate Dentry and UDBA (User's Direct Branch Access) +---------------------------------------------------------------------- +Generally VFS helpers re-validate a dentry as a part of lookup. +0. digging down the directory hierarchy. +1. lock the parent dir by its i_mutex. +2. lookup the final (child) entry. +3. revalidate it. +4. call the actual operation (create, unlink, etc.) +5. unlock the parent dir + +If the filesystem implements its ->d_revalidate() (step 3), then it is +called. Actually aufs implements it and checks the dentry on a branch is +still valid. +But it is not enough. Because aufs has to release the lock for the +parent dir on a branch at the end of ->lookup() (step 2) and +->d_revalidate() (step 3) while the i_mutex of the aufs dir is still +held by VFS. +If the file on a branch is changed directly, eg. bypassing aufs, after +aufs released the lock, then the subsequent operation may cause +something unpleasant result. + +This situation is a result of VFS architecture, ->lookup() and +->d_revalidate() is separated. But I never say it is wrong. It is a good +design from VFS's point of view. It is just not suitable for sub-VFS +character in aufs. + +Aufs supports such case by three level of revalidation which is +selectable by user. +1. Simple Revalidate + Addition to the native flow in VFS's, confirm the child-parent + relationship on the branch just after locking the parent dir on the + branch in the "actual operation" (step 4). When this validation + fails, aufs returns EBUSY. ->d_revalidate() (step 3) in aufs still + checks the validation of the dentry on branches. +2. Monitor Changes Internally by Inotify/Fsnotify + Addition to above, in the "actual operation" (step 4) aufs re-lookup + the dentry on the branch, and returns EBUSY if it finds different + dentry. + Additionally, aufs sets the inotify/fsnotify watch for every dir on branches + during it is in cache. When the event is notified, aufs registers a + function to kernel 'events' thread by schedule_work(). And the + function sets some special status to the cached aufs dentry and inode + private data. If they are not cached, then aufs has nothing to + do. When the same file is accessed through aufs (step 0-3) later, + aufs will detect the status and refresh all necessary data. + In this mode, aufs has to ignore the event which is fired by aufs + itself. +3. No Extra Validation + This is the simplest test and doesn't add any additional revalidation + test, and skip the revalidation in step 4. It is useful and improves + aufs performance when system surely hide the aufs branches from user, + by over-mounting something (or another method). diff --git a/Documentation/filesystems/aufs/design/04branch.txt b/Documentation/filesystems/aufs/design/04branch.txt new file mode 100644 index 00000000000..da5200be41a --- /dev/null +++ b/Documentation/filesystems/aufs/design/04branch.txt @@ -0,0 +1,74 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Branch Manipulation + +Since aufs supports dynamic branch manipulation, ie. add/remove a branch +and changing its permission/attribute, there are a lot of works to do. + + +Add a Branch +---------------------------------------------------------------------- +o Confirm the adding dir exists outside of aufs, including loopback + mount, and its various attributes. +o Initialize the xino file and whiteout bases if necessary. + See struct.txt. + +o Check the owner/group/mode of the directory + When the owner/group/mode of the adding directory differs from the + existing branch, aufs issues a warning because it may impose a + security risk. + For example, when a upper writable branch has a world writable empty + top directory, a malicious user can create any files on the writable + branch directly, like copy-up and modify manually. If something like + /etc/{passwd,shadow} exists on the lower readonly branch but the upper + writable branch, and the writable branch is world-writable, then a + malicious guy may create /etc/passwd on the writable branch directly + and the infected file will be valid in aufs. + I am afraid it can be a security issue, but aufs can do nothing except + producing a warning. + + +Delete a Branch +---------------------------------------------------------------------- +o Confirm the deleting branch is not busy + To be general, there is one merit to adopt "remount" interface to + manipulate branches. It is to discard caches. At deleting a branch, + aufs checks the still cached (and connected) dentries and inodes. If + there are any, then they are all in-use. An inode without its + corresponding dentry can be alive alone (for example, inotify/fsnotify case). + + For the cached one, aufs checks whether the same named entry exists on + other branches. + If the cached one is a directory, because aufs provides a merged view + to users, as long as one dir is left on any branch aufs can show the + dir to users. In this case, the branch can be removed from aufs. + Otherwise aufs rejects deleting the branch. + + If any file on the deleting branch is opened by aufs, then aufs + rejects deleting. + + +Modify the Permission of a Branch +---------------------------------------------------------------------- +o Re-initialize or remove the xino file and whiteout bases if necessary. + See struct.txt. + +o rw --> ro: Confirm the modifying branch is not busy + Aufs rejects the request if any of these conditions are true. + - a file on the branch is mmap-ed. + - a regular file on the branch is opened for write and there is no + same named entry on the upper branch. diff --git a/Documentation/filesystems/aufs/design/05wbr_policy.txt b/Documentation/filesystems/aufs/design/05wbr_policy.txt new file mode 100644 index 00000000000..0262084bf63 --- /dev/null +++ b/Documentation/filesystems/aufs/design/05wbr_policy.txt @@ -0,0 +1,64 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Policies to Select One among Multiple Writable Branches +---------------------------------------------------------------------- +When the number of writable branch is more than one, aufs has to decide +the target branch for file creation or copy-up. By default, the highest +writable branch which has the parent (or ancestor) dir of the target +file is chosen (top-down-parent policy). +By user's request, aufs implements some other policies to select the +writable branch, for file creation several policies, round-robin, +most-free-space, and other policies. For copy-up, top-down-parent, +bottom-up-parent, bottom-up and others. + +As expected, the round-robin policy selects the branch in circular. When +you have two writable branches and creates 10 new files, 5 files will be +created for each branch. mkdir(2) systemcall is an exception. When you +create 10 new directories, all will be created on the same branch. +And the most-free-space policy selects the one which has most free +space among the writable branches. The amount of free space will be +checked by aufs internally, and users can specify its time interval. + +The policies for copy-up is more simple, +top-down-parent is equivalent to the same named on in create policy, +bottom-up-parent selects the writable branch where the parent dir +exists and the nearest upper one from the copyup-source, +bottom-up selects the nearest upper writable branch from the +copyup-source, regardless the existence of the parent dir. + +There are some rules or exceptions to apply these policies. +- If there is a readonly branch above the policy-selected branch and + the parent dir is marked as opaque (a variation of whiteout), or the + target (creating) file is whiteout-ed on the upper readonly branch, + then the result of the policy is ignored and the target file will be + created on the nearest upper writable branch than the readonly branch. +- If there is a writable branch above the policy-selected branch and + the parent dir is marked as opaque or the target file is whiteouted + on the branch, then the result of the policy is ignored and the target + file will be created on the highest one among the upper writable + branches who has diropq or whiteout. In case of whiteout, aufs removes + it as usual. +- link(2) and rename(2) systemcalls are exceptions in every policy. + They try selecting the branch where the source exists as possible + since copyup a large file will take long time. If it can't be, + ie. the branch where the source exists is readonly, then they will + follow the copyup policy. +- There is an exception for rename(2) when the target exists. + If the rename target exists, aufs compares the index of the branches + where the source and the target exists and selects the higher + one. If the selected branch is readonly, then aufs follows the + copyup policy. diff --git a/Documentation/filesystems/aufs/design/06dirren.dot b/Documentation/filesystems/aufs/design/06dirren.dot new file mode 100644 index 00000000000..2d62bb6dd55 --- /dev/null +++ b/Documentation/filesystems/aufs/design/06dirren.dot @@ -0,0 +1,31 @@ + +// to view this graph, run dot(1) command in GRAPHVIZ. + +digraph G { +node [shape=box]; +whinfo [label="detailed info file\n(lower_brid_root-hinum, h_inum, namelen, old name)"]; + +node [shape=oval]; + +aufs_rename -> whinfo [label="store/remove"]; + +node [shape=oval]; +inode_list [label="h_inum list in branch\ncache"]; + +node [shape=box]; +whinode [label="h_inum list file"]; + +node [shape=oval]; +brmgmt [label="br_add/del/mod/umount"]; + +brmgmt -> inode_list [label="create/remove"]; +brmgmt -> whinode [label="load/store"]; + +inode_list -> whinode [style=dashed,dir=both]; + +aufs_rename -> inode_list [label="add/del"]; + +aufs_lookup -> inode_list [label="search"]; + +aufs_lookup -> whinfo [label="load/remove"]; +} diff --git a/Documentation/filesystems/aufs/design/06dirren.txt b/Documentation/filesystems/aufs/design/06dirren.txt new file mode 100644 index 00000000000..38ae77b2c84 --- /dev/null +++ b/Documentation/filesystems/aufs/design/06dirren.txt @@ -0,0 +1,102 @@ + +# Copyright (C) 2017-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Special handling for renaming a directory (DIRREN) +---------------------------------------------------------------------- +First, let's assume we have a simple usecase. + +- /u = /rw + /ro +- /rw/dirA exists +- /ro/dirA and /ro/dirA/file exist too +- there is no dirB on both branches +- a user issues rename("dirA", "dirB") + +Now, what should aufs behave against this rename(2)? +There are a few possible cases. + +A. returns EROFS. + since dirA exists on a readonly branch which cannot be renamed. +B. returns EXDEV. + it is possible to copy-up dirA (only the dir itself), but the child + entries ("file" in this case) should not be. it must be a bad + approach to copy-up recursively. +C. returns a success. + even the branch /ro is readonly, aufs tries renaming it. Obviously it + is a violation of aufs' policy. +D. construct an extra information which indicates that /ro/dirA should + be handled as the name of dirB. + overlayfs has a similar feature called REDIRECT. + +Until now, aufs implements the case B only which returns EXDEV, and +expects the userspace application behaves like mv(1) which tries +issueing rename(2) recursively. + +A new aufs feature called DIRREN is introduced which implements the case +D. There are several "extra information" added. + +1. detailed info per renamed directory + path: /rw/dirB/$AUFS_WH_DR_INFO_PFX. +2. the inode-number list of directories on a branch + path: /rw/dirB/$AUFS_WH_DR_BRHINO + +The filename of "detailed info per directory" represents the lower +branch, and its format is +- a type of the branch id + one of these. + + uuid (not implemented yet) + + fsid + + dev +- the inode-number of the branch root dir + +And it contains these info in a single regular file. +- magic number +- branch's inode-number of the logically renamed dir +- the name of the before-renamed dir + +The "detailed info per directory" file is created in aufs rename(2), and +loaded in any lookup. +The info is considered in lookup for the matching case only. Here +"matching" means that the root of branch (in the info filename) is same +to the current looking-up branch. After looking-up the before-renamed +name, the inode-number is compared. And the matched dentry is used. + +The "inode-number list of directories" is a regular file which contains +simply the inode-numbers on the branch. The file is created or updated +in removing the branch, and loaded in adding the branch. Its lifetime is +equal to the branch. +The list is refered in lookup, and when the current target inode is +found in the list, the aufs tries loading the "detailed info per +directory" and get the changed and valid name of the dir. + +Theoretically these "extra informaiton" may be able to be put into XATTR +in the dir inode. But aufs doesn't choose this way because +1. XATTR may not be supported by the branch (or its configuration) +2. XATTR may have its size limit. +3. XATTR may be less easy to convert than a regular file, when the + format of the info is changed in the future. +At the same time, I agree that the regular file approach is much slower +than XATTR approach. So, in the future, aufs may take the XATTR or other +better approach. + +This DIRREN feature is enabled by aufs configuration, and is activated +by a new mount option. + +For the more complicated case, there is a work with UDBA option, which +is to dected the direct access to the branches (by-passing aufs) and to +maintain the cashes in aufs. Since a single cached aufs dentry may +contains two names, before- and after-rename, the name comparision in +UDBA handler may not work correctly. In this case, the behaviour will be +equivalen to udba=reval case. diff --git a/Documentation/filesystems/aufs/design/06fhsm.txt b/Documentation/filesystems/aufs/design/06fhsm.txt new file mode 100644 index 00000000000..df985662bef --- /dev/null +++ b/Documentation/filesystems/aufs/design/06fhsm.txt @@ -0,0 +1,120 @@ + +# Copyright (C) 2011-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + + +File-based Hierarchical Storage Management (FHSM) +---------------------------------------------------------------------- +Hierarchical Storage Management (or HSM) is a well-known feature in the +storage world. Aufs provides this feature as file-based with multiple +writable branches, based upon the principle of "Colder, the Lower". +Here the word "colder" means that the less used files, and "lower" means +that the position in the order of the stacked branches vertically. +These multiple writable branches are prioritized, ie. the topmost one +should be the fastest drive and be used heavily. + +o Characters in aufs FHSM story +- aufs itself and a new branch attribute. +- a new ioctl interface to move-down and to establish a connection with + the daemon ("move-down" is a converse of "copy-up"). +- userspace tool and daemon. + +The userspace daemon establishes a connection with aufs and waits for +the notification. The notified information is very similar to struct +statfs containing the number of consumed blocks and inodes. +When the consumed blocks/inodes of a branch exceeds the user-specified +upper watermark, the daemon activates its move-down process until the +consumed blocks/inodes reaches the user-specified lower watermark. + +The actual move-down is done by aufs based upon the request from +user-space since we need to maintain the inode number and the internal +pointer arrays in aufs. + +Currently aufs FHSM handles the regular files only. Additionally they +must not be hard-linked nor pseudo-linked. + + +o Cowork of aufs and the user-space daemon + During the userspace daemon established the connection, aufs sends a + small notification to it whenever aufs writes something into the + writable branch. But it may cost high since aufs issues statfs(2) + internally. So user can specify a new option to cache the + info. Actually the notification is controlled by these factors. + + the specified cache time. + + classified as "force" by aufs internally. + Until the specified time expires, aufs doesn't send the info + except the forced cases. When aufs decide forcing, the info is always + notified to userspace. + For example, the number of free inodes is generally large enough and + the shortage of it happens rarely. So aufs doesn't force the + notification when creating a new file, directory and others. This is + the typical case which aufs doesn't force. + When aufs writes the actual filedata and the files consumes any of new + blocks, the aufs forces notifying. + + +o Interfaces in aufs +- New branch attribute. + + fhsm + Specifies that the branch is managed by FHSM feature. In other word, + participant in the FHSM. + When nofhsm is set to the branch, it will not be the source/target + branch of the move-down operation. This attribute is set + independently from coo and moo attributes, and if you want full + FHSM, you should specify them as well. +- New mount option. + + fhsm_sec + Specifies a second to suppress many less important info to be + notified. +- New ioctl. + + AUFS_CTL_FHSM_FD + create a new file descriptor which userspace can read the notification + (a subset of struct statfs) from aufs. +- Module parameter 'brs' + It has to be set to 1. Otherwise the new mount option 'fhsm' will not + be set. +- mount helpers /sbin/mount.aufs and /sbin/umount.aufs + When there are two or more branches with fhsm attributes, + /sbin/mount.aufs invokes the user-space daemon and /sbin/umount.aufs + terminates it. As a result of remounting and branch-manipulation, the + number of branches with fhsm attribute can be one. In this case, + /sbin/mount.aufs will terminate the user-space daemon. + + +Finally the operation is done as these steps in kernel-space. +- make sure that, + + no one else is using the file. + + the file is not hard-linked. + + the file is not pseudo-linked. + + the file is a regular file. + + the parent dir is not opaqued. +- find the target writable branch. +- make sure the file is not whiteout-ed by the upper (than the target) + branch. +- make the parent dir on the target branch. +- mutex lock the inode on the branch. +- unlink the whiteout on the target branch (if exists). +- lookup and create the whiteout-ed temporary name on the target branch. +- copy the file as the whiteout-ed temporary name on the target branch. +- rename the whiteout-ed temporary name to the original name. +- unlink the file on the source branch. +- maintain the internal pointer array and the external inode number + table (XINO). +- maintain the timestamps and other attributes of the parent dir and the + file. + +And of course, in every step, an error may happen. So the operation +should restore the original file state after an error happens. diff --git a/Documentation/filesystems/aufs/design/06mmap.txt b/Documentation/filesystems/aufs/design/06mmap.txt new file mode 100644 index 00000000000..9184f671043 --- /dev/null +++ b/Documentation/filesystems/aufs/design/06mmap.txt @@ -0,0 +1,72 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +mmap(2) -- File Memory Mapping +---------------------------------------------------------------------- +In aufs, the file-mapped pages are handled by a branch fs directly, no +interaction with aufs. It means aufs_mmap() calls the branch fs's +->mmap(). +This approach is simple and good, but there is one problem. +Under /proc, several entries show the mmapped files by its path (with +device and inode number), and the printed path will be the path on the +branch fs's instead of virtual aufs's. +This is not a problem in most cases, but some utilities lsof(1) (and its +user) may expect the path on aufs. + +To address this issue, aufs adds a new member called vm_prfile in struct +vm_area_struct (and struct vm_region). The original vm_file points to +the file on the branch fs in order to handle everything correctly as +usual. The new vm_prfile points to a virtual file in aufs, and the +show-functions in procfs refers to vm_prfile if it is set. +Also we need to maintain several other places where touching vm_file +such like +- fork()/clone() copies vma and the reference count of vm_file is + incremented. +- merging vma maintains the ref count too. + +This is not a good approach. It just fakes the printed path. But it +leaves all behaviour around f_mapping unchanged. This is surely an +advantage. +Actually aufs had adopted another complicated approach which calls +generic_file_mmap() and handles struct vm_operations_struct. In this +approach, aufs met a hard problem and I could not solve it without +switching the approach. + +There may be one more another approach which is +- bind-mount the branch-root onto the aufs-root internally +- grab the new vfsmount (ie. struct mount) +- lazy-umount the branch-root internally +- in open(2) the aufs-file, open the branch-file with the hidden + vfsmount (instead of the original branch's vfsmount) +- ideally this "bind-mount and lazy-umount" should be done atomically, + but it may be possible from userspace by the mount helper. + +Adding the internal hidden vfsmount and using it in opening a file, the +file path under /proc will be printed correctly. This approach looks +smarter, but is not possible I am afraid. +- aufs-root may be bind-mount later. when it happens, another hidden + vfsmount will be required. +- it is hard to get the chance to bind-mount and lazy-umount + + in kernel-space, FS can have vfsmount in open(2) via + file->f_path, and aufs can know its vfsmount. But several locks are + already acquired, and if aufs tries to bind-mount and lazy-umount + here, then it may cause a deadlock. + + in user-space, bind-mount doesn't invoke the mount helper. +- since /proc shows dev and ino, aufs has to give vma these info. it + means a new member vm_prinode will be necessary. this is essentially + equivalent to vm_prfile described above. + +I have to give up this "looks-smater" approach. diff --git a/Documentation/filesystems/aufs/design/06xattr.txt b/Documentation/filesystems/aufs/design/06xattr.txt new file mode 100644 index 00000000000..d0f6aedfe2d --- /dev/null +++ b/Documentation/filesystems/aufs/design/06xattr.txt @@ -0,0 +1,96 @@ + +# Copyright (C) 2014-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write to the Free Software +# Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + + +Listing XATTR/EA and getting the value +---------------------------------------------------------------------- +For the inode standard attributes (owner, group, timestamps, etc.), aufs +shows the values from the topmost existing file. This behaviour is good +for the non-dir entries since the bahaviour exactly matches the shown +information. But for the directories, aufs considers all the same named +entries on the lower branches. Which means, if one of the lower entry +rejects readdir call, then aufs returns an error even if the topmost +entry allows it. This behaviour is necessary to respect the branch fs's +security, but can make users confused since the user-visible standard +attributes don't match the behaviour. +To address this issue, aufs has a mount option called dirperm1 which +checks the permission for the topmost entry only, and ignores the lower +entry's permission. + +A similar issue can happen around XATTR. +getxattr(2) and listxattr(2) families behave as if dirperm1 option is +always set. Otherwise these very unpleasant situation would happen. +- listxattr(2) may return the duplicated entries. +- users may not be able to remove or reset the XATTR forever, + + +XATTR/EA support in the internal (copy,move)-(up,down) +---------------------------------------------------------------------- +Generally the extended attributes of inode are categorized as these. +- "security" for LSM and capability. +- "system" for posix ACL, 'acl' mount option is required for the branch + fs generally. +- "trusted" for userspace, CAP_SYS_ADMIN is required. +- "user" for userspace, 'user_xattr' mount option is required for the + branch fs generally. + +Moreover there are some other categories. Aufs handles these rather +unpopular categories as the ordinary ones, ie. there is no special +condition nor exception. + +In copy-up, the support for XATTR on the dst branch may differ from the +src branch. In this case, the copy-up operation will get an error and +the original user operation which triggered the copy-up will fail. It +can happen that even all copy-up will fail. +When both of src and dst branches support XATTR and if an error occurs +during copying XATTR, then the copy-up should fail obviously. That is a +good reason and aufs should return an error to userspace. But when only +the src branch support that XATTR, aufs should not return an error. +For example, the src branch supports ACL but the dst branch doesn't +because the dst branch may natively un-support it or temporary +un-support it due to "noacl" mount option. Of course, the dst branch fs +may NOT return an error even if the XATTR is not supported. It is +totally up to the branch fs. + +Anyway when the aufs internal copy-up gets an error from the dst branch +fs, then aufs tries removing the just copied entry and returns the error +to the userspace. The worst case of this situation will be all copy-up +will fail. + +For the copy-up operation, there two basic approaches. +- copy the specified XATTR only (by category above), and return the + error unconditionally if it happens. +- copy all XATTR, and ignore the error on the specified category only. + +In order to support XATTR and to implement the correct behaviour, aufs +chooses the latter approach and introduces some new branch attributes, +"icexsec", "icexsys", "icextr", "icexusr", and "icexoth". +They correspond to the XATTR namespaces (see above). Additionally, to be +convenient, "icex" is also provided which means all "icex*" attributes +are set (here the word "icex" stands for "ignore copy-error on XATTR"). + +The meaning of these attributes is to ignore the error from setting +XATTR on that branch. +Note that aufs tries copying all XATTR unconditionally, and ignores the +error from the dst branch according to the specified attributes. + +Some XATTR may have its default value. The default value may come from +the parent dir or the environment. If the default value is set at the +file creating-time, it will be overwritten by copy-up. +Some contradiction may happen I am afraid. +Do we need another attribute to stop copying XATTR? I am unsure. For +now, aufs implements the branch attributes to ignore the error. diff --git a/Documentation/filesystems/aufs/design/07export.txt b/Documentation/filesystems/aufs/design/07export.txt new file mode 100644 index 00000000000..6fcb00d7dbd --- /dev/null +++ b/Documentation/filesystems/aufs/design/07export.txt @@ -0,0 +1,58 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Export Aufs via NFS +---------------------------------------------------------------------- +Here is an approach. +- like xino/xib, add a new file 'xigen' which stores aufs inode + generation. +- iget_locked(): initialize aufs inode generation for a new inode, and + store it in xigen file. +- destroy_inode(): increment aufs inode generation and store it in xigen + file. it is necessary even if it is not unlinked, because any data of + inode may be changed by UDBA. +- encode_fh(): for a root dir, simply return FILEID_ROOT. otherwise + build file handle by + + branch id (4 bytes) + + superblock generation (4 bytes) + + inode number (4 or 8 bytes) + + parent dir inode number (4 or 8 bytes) + + inode generation (4 bytes)) + + return value of exportfs_encode_fh() for the parent on a branch (4 + bytes) + + file handle for a branch (by exportfs_encode_fh()) +- fh_to_dentry(): + + find the index of a branch from its id in handle, and check it is + still exist in aufs. + + 1st level: get the inode number from handle and search it in cache. + + 2nd level: if not found in cache, get the parent inode number from + the handle and search it in cache. and then open the found parent + dir, find the matching inode number by vfs_readdir() and get its + name, and call lookup_one_len() for the target dentry. + + 3rd level: if the parent dir is not cached, call + exportfs_decode_fh() for a branch and get the parent on a branch, + build a pathname of it, convert it a pathname in aufs, call + path_lookup(). now aufs gets a parent dir dentry, then handle it as + the 2nd level. + + to open the dir, aufs needs struct vfsmount. aufs keeps vfsmount + for every branch, but not itself. to get this, (currently) aufs + searches in current->nsproxy->mnt_ns list. it may not be a good + idea, but I didn't get other approach. + + test the generation of the gotten inode. +- every inode operation: they may get EBUSY due to UDBA. in this case, + convert it into ESTALE for NFSD. +- readdir(): call lockdep_on/off() because filldir in NFSD calls + lookup_one_len(), vfs_getattr(), encode_fh() and others. diff --git a/Documentation/filesystems/aufs/design/08shwh.txt b/Documentation/filesystems/aufs/design/08shwh.txt new file mode 100644 index 00000000000..d7e58319086 --- /dev/null +++ b/Documentation/filesystems/aufs/design/08shwh.txt @@ -0,0 +1,52 @@ + +# Copyright (C) 2005-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Show Whiteout Mode (shwh) +---------------------------------------------------------------------- +Generally aufs hides the name of whiteouts. But in some cases, to show +them is very useful for users. For instance, creating a new middle layer +(branch) by merging existing layers. + +(borrowing aufs1 HOW-TO from a user, Michael Towers) +When you have three branches, +- Bottom: 'system', squashfs (underlying base system), read-only +- Middle: 'mods', squashfs, read-only +- Top: 'overlay', ram (tmpfs), read-write + +The top layer is loaded at boot time and saved at shutdown, to preserve +the changes made to the system during the session. +When larger changes have been made, or smaller changes have accumulated, +the size of the saved top layer data grows. At this point, it would be +nice to be able to merge the two overlay branches ('mods' and 'overlay') +and rewrite the 'mods' squashfs, clearing the top layer and thus +restoring save and load speed. + +This merging is simplified by the use of another aufs mount, of just the +two overlay branches using the 'shwh' option. +# mount -t aufs -o ro,shwh,br:/livesys/overlay=ro+wh:/livesys/mods=rr+wh \ + aufs /livesys/merge_union + +A merged view of these two branches is then available at +/livesys/merge_union, and the new feature is that the whiteouts are +visible! +Note that in 'shwh' mode the aufs mount must be 'ro', which will disable +writing to all branches. Also the default mode for all branches is 'ro'. +It is now possible to save the combined contents of the two overlay +branches to a new squashfs, e.g.: +# mksquashfs /livesys/merge_union /path/to/newmods.squash + +This new squashfs archive can be stored on the boot device and the +initramfs will use it to replace the old one at the next boot. diff --git a/Documentation/filesystems/aufs/design/10dynop.txt b/Documentation/filesystems/aufs/design/10dynop.txt new file mode 100644 index 00000000000..d55cae285df --- /dev/null +++ b/Documentation/filesystems/aufs/design/10dynop.txt @@ -0,0 +1,47 @@ + +# Copyright (C) 2010-2020 Junjiro R. Okajima +# +# This program is free software; you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation; either version 2 of the License, or +# (at your option) any later version. +# +# This program is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program. If not, see . + +Dynamically customizable FS operations +---------------------------------------------------------------------- +Generally FS operations (struct inode_operations, struct +address_space_operations, struct file_operations, etc.) are defined as +"static const", but it never means that FS have only one set of +operation. Some FS have multiple sets of them. For instance, ext2 has +three sets, one for XIP, for NOBH, and for normal. +Since aufs overrides and redirects these operations, sometimes aufs has +to change its behaviour according to the branch FS type. More importantly +VFS acts differently if a function (member in the struct) is set or +not. It means aufs should have several sets of operations and select one +among them according to the branch FS definition. + +In order to solve this problem and not to affect the behaviour of VFS, +aufs defines these operations dynamically. For instance, aufs defines +dummy direct_IO function for struct address_space_operations, but it may +not be set to the address_space_operations actually. When the branch FS +doesn't have it, aufs doesn't set it to its address_space_operations +while the function definition itself is still alive. So the behaviour +itself will not change, and it will return an error when direct_IO is +not set. + +The lifetime of these dynamically generated operation object is +maintained by aufs branch object. When the branch is removed from aufs, +the reference counter of the object is decremented. When it reaches +zero, the dynamically generated operation object will be freed. + +This approach is designed to support AIO (io_submit), Direct I/O and +XIP (DAX) mainly. +Currently this approach is applied to address_space_operations for +regular files only. diff --git a/Documentation/locking/futex2.rst b/Documentation/locking/futex2.rst new file mode 100644 index 00000000000..13a7699bd6f --- /dev/null +++ b/Documentation/locking/futex2.rst @@ -0,0 +1,198 @@ +.. SPDX-License-Identifier: GPL-2.0 + +====== +futex2 +====== + +:Author: André Almeida + +futex, or fast user mutex, is a set of syscalls to allow userspace to create +performant synchronization mechanisms, such as mutexes, semaphores and +conditional variables in userspace. C standard libraries, like glibc, uses it +as a means to implement more high level interfaces like pthreads. + +The interface +============= + +uAPI functions +-------------- + +.. kernel-doc:: kernel/futex2.c + :identifiers: sys_futex_wait sys_futex_wake sys_futex_waitv sys_futex_requeue + +uAPI structures +--------------- + +.. kernel-doc:: include/uapi/linux/futex.h + +The ``flag`` argument +--------------------- + +The flag is used to specify the size of the futex word +(FUTEX_[8, 16, 32]). It's mandatory to define one, since there's no +default size. + +By default, the timeout uses a monotonic clock, but can be used as a realtime +one by using the FUTEX_REALTIME_CLOCK flag. + +By default, futexes are of the private type, that means that this user address +will be accessed by threads that share the same memory region. This allows for +some internal optimizations, so they are faster. However, if the address needs +to be shared with different processes (like using ``mmap()`` or ``shm()``), they +need to be defined as shared and the flag FUTEX_SHARED_FLAG is used to set that. + +By default, the operation has no NUMA-awareness, meaning that the user can't +choose the memory node where the kernel side futex data will be stored. The +user can choose the node where it wants to operate by setting the +FUTEX_NUMA_FLAG and using the following structure (where X can be 8, 16, or +32):: + + struct futexX_numa { + __uX value; + __sX hint; + }; + +This structure should be passed at the ``void *uaddr`` of futex functions. The +address of the structure will be used to be waited on/waken on, and the +``value`` will be compared to ``val`` as usual. The ``hint`` member is used to +define which node the futex will use. When waiting, the futex will be +registered on a kernel-side table stored on that node; when waking, the futex +will be searched for on that given table. That means that there's no redundancy +between tables, and the wrong ``hint`` value will lead to undesired behavior. +Userspace is responsible for dealing with node migrations issues that may +occur. ``hint`` can range from [0, MAX_NUMA_NODES), for specifying a node, or +-1, to use the same node the current process is using. + +When not using FUTEX_NUMA_FLAG on a NUMA system, the futex will be stored on a +global table on allocated on the first node. + +The ``timo`` argument +--------------------- + +As per the Y2038 work done in the kernel, new interfaces shouldn't add timeout +options known to be buggy. Given that, ``timo`` should be a 64-bit timeout at +all platforms, using an absolute timeout value. + +Implementation +============== + +The internal implementation follows a similar design to the original futex. +Given that we want to replicate the same external behavior of current futex, +this should be somewhat expected. + +Waiting +------- + +For the wait operations, they are all treated as if you want to wait on N +futexes, so the path for futex_wait and futex_waitv is the basically the same. +For both syscalls, the first step is to prepare an internal list for the list +of futexes to wait for (using struct futexv_head). For futex_wait() calls, this +list will have a single object. + +We have a hash table, where waiters register themselves before sleeping. Then +the wake function checks this table looking for waiters at uaddr. The hash +bucket to be used is determined by a struct futex_key, that stores information +to uniquely identify an address from a given process. Given the huge address +space, there'll be hash collisions, so we store information to be later used on +collision treatment. + +First, for every futex we want to wait on, we check if (``*uaddr == val``). +This check is done holding the bucket lock, so we are correctly serialized with +any futex_wake() calls. If any waiter fails the check above, we dequeue all +futexes. The check (``*uaddr == val``) can fail for two reasons: + +- The values are different, and we return -EAGAIN. However, if while + dequeueing we found that some futexes were awakened, we prioritize this + and return success. + +- When trying to access the user address, we do so with page faults + disabled because we are holding a bucket's spin lock (and can't sleep + while holding a spin lock). If there's an error, it might be a page + fault, or an invalid address. We release the lock, dequeue everyone + (because it's illegal to sleep while there are futexes enqueued, we + could lose wakeups) and try again with page fault enabled. If we + succeed, this means that the address is valid, but we need to do + all the work again. For serialization reasons, we need to have the + spin lock when getting the user value. Additionally, for shared + futexes, we also need to recalculate the hash, since the underlying + mapping mechanisms could have changed when dealing with page fault. + If, even with page fault enabled, we can't access the address, it + means it's an invalid user address, and we return -EFAULT. For this + case, we prioritize the error, even if some futexes were awaken. + +If the check is OK, they are enqueued on a linked list in our bucket, and +proceed to the next one. If all waiters succeed, we put the thread to sleep +until a futex_wake() call, timeout expires or we get a signal. After waking up, +we dequeue everyone, and check if some futex was awakened. This dequeue is done +by iteratively walking at each element of struct futex_head list. + +All enqueuing/dequeuing operations requires to hold the bucket lock, to avoid +racing while modifying the list. + +Waking +------ + +We get the bucket that's storing the waiters at uaddr, and wake the required +number of waiters, checking for hash collision. + +There's an optimization that makes futex_wake() not take the bucket lock if +there's no one to be woken on that bucket. It checks an atomic counter that each +bucket has, if it says 0, then the syscall exits. In order for this to work, the +waiter thread increases it before taking the lock, so the wake thread will +correctly see that there's someone waiting and will continue the path to take +the bucket lock. To get the correct serialization, the waiter issues a memory +barrier after increasing the bucket counter and the waker issues a memory +barrier before checking it. + +Requeuing +--------- + +The requeue path first checks for each struct futex_requeue and their flags. +Then, it will compare the expected value with the one at uaddr1::uaddr. +Following the same serialization explained at Waking_, we increase the atomic +counter for the bucket of uaddr2 before taking the lock. We need to have both +buckets locks at same time so we don't race with other futex operation. To +ensure the locks are taken in the same order for all threads (and thus avoiding +deadlocks), every requeue operation takes the "smaller" bucket first, when +comparing both addresses. + +If the compare with user value succeeds, we proceed by waking ``nr_wake`` +futexes, and then requeuing ``nr_requeue`` from bucket of uaddr1 to the uaddr2. +This consists in a simple list deletion/addition and replacing the old futex key +with the new one. + +Futex keys +---------- + +There are two types of futexes: private and shared ones. The private are futexes +meant to be used by threads that share the same memory space, are easier to be +uniquely identified and thus can have some performance optimization. The +elements for identifying one are: the start address of the page where the +address is, the address offset within the page and the current->mm pointer. + +Now, for uniquely identifying a shared futex: + +- If the page containing the user address is an anonymous page, we can + just use the same data used for private futexes (the start address of + the page, the address offset within the page and the current->mm + pointer); that will be enough for uniquely identifying such futex. We + also set one bit at the key to differentiate if a private futex is + used on the same address (mixing shared and private calls does not + work). + +- If the page is file-backed, current->mm maybe isn't the same one for + every user of this futex, so we need to use other data: the + page->index, a UUID for the struct inode and the offset within the + page. + +Note that members of futex_key don't have any particular meaning after they +are part of the struct - they are just bytes to identify a futex. Given that, +we don't need to use a particular name or type that matches the original data, +we only need to care about the bitsize of each component and make both private +and shared fit in the same memory space. + +Source code documentation +========================= + +.. kernel-doc:: kernel/futex2.c + :no-identifiers: sys_futex_wait sys_futex_wake sys_futex_waitv sys_futex_requeue diff --git a/Documentation/locking/index.rst b/Documentation/locking/index.rst index 7003bd5aeff..9bf03c7fa1e 100644 --- a/Documentation/locking/index.rst +++ b/Documentation/locking/index.rst @@ -24,6 +24,7 @@ locking percpu-rw-semaphore robust-futexes robust-futex-ABI + futex2 .. only:: subproject and html diff --git a/Documentation/scheduler/sched-BFS.txt b/Documentation/scheduler/sched-BFS.txt new file mode 100644 index 00000000000..c0282002a07 --- /dev/null +++ b/Documentation/scheduler/sched-BFS.txt @@ -0,0 +1,351 @@ +BFS - The Brain Fuck Scheduler by Con Kolivas. + +Goals. + +The goal of the Brain Fuck Scheduler, referred to as BFS from here on, is to +completely do away with the complex designs of the past for the cpu process +scheduler and instead implement one that is very simple in basic design. +The main focus of BFS is to achieve excellent desktop interactivity and +responsiveness without heuristics and tuning knobs that are difficult to +understand, impossible to model and predict the effect of, and when tuned to +one workload cause massive detriment to another. + + +Design summary. + +BFS is best described as a single runqueue, O(n) lookup, earliest effective +virtual deadline first design, loosely based on EEVDF (earliest eligible virtual +deadline first) and my previous Staircase Deadline scheduler. Each component +shall be described in order to understand the significance of, and reasoning for +it. The codebase when the first stable version was released was approximately +9000 lines less code than the existing mainline linux kernel scheduler (in +2.6.31). This does not even take into account the removal of documentation and +the cgroups code that is not used. + +Design reasoning. + +The single runqueue refers to the queued but not running processes for the +entire system, regardless of the number of CPUs. The reason for going back to +a single runqueue design is that once multiple runqueues are introduced, +per-CPU or otherwise, there will be complex interactions as each runqueue will +be responsible for the scheduling latency and fairness of the tasks only on its +own runqueue, and to achieve fairness and low latency across multiple CPUs, any +advantage in throughput of having CPU local tasks causes other disadvantages. +This is due to requiring a very complex balancing system to at best achieve some +semblance of fairness across CPUs and can only maintain relatively low latency +for tasks bound to the same CPUs, not across them. To increase said fairness +and latency across CPUs, the advantage of local runqueue locking, which makes +for better scalability, is lost due to having to grab multiple locks. + +A significant feature of BFS is that all accounting is done purely based on CPU +used and nowhere is sleep time used in any way to determine entitlement or +interactivity. Interactivity "estimators" that use some kind of sleep/run +algorithm are doomed to fail to detect all interactive tasks, and to falsely tag +tasks that aren't interactive as being so. The reason for this is that it is +close to impossible to determine that when a task is sleeping, whether it is +doing it voluntarily, as in a userspace application waiting for input in the +form of a mouse click or otherwise, or involuntarily, because it is waiting for +another thread, process, I/O, kernel activity or whatever. Thus, such an +estimator will introduce corner cases, and more heuristics will be required to +cope with those corner cases, introducing more corner cases and failed +interactivity detection and so on. Interactivity in BFS is built into the design +by virtue of the fact that tasks that are waking up have not used up their quota +of CPU time, and have earlier effective deadlines, thereby making it very likely +they will preempt any CPU bound task of equivalent nice level. See below for +more information on the virtual deadline mechanism. Even if they do not preempt +a running task, because the rr interval is guaranteed to have a bound upper +limit on how long a task will wait for, it will be scheduled within a timeframe +that will not cause visible interface jitter. + + +Design details. + +Task insertion. + +BFS inserts tasks into each relevant queue as an O(1) insertion into a double +linked list. On insertion, *every* running queue is checked to see if the newly +queued task can run on any idle queue, or preempt the lowest running task on the +system. This is how the cross-CPU scheduling of BFS achieves significantly lower +latency per extra CPU the system has. In this case the lookup is, in the worst +case scenario, O(n) where n is the number of CPUs on the system. + +Data protection. + +BFS has one single lock protecting the process local data of every task in the +global queue. Thus every insertion, removal and modification of task data in the +global runqueue needs to grab the global lock. However, once a task is taken by +a CPU, the CPU has its own local data copy of the running process' accounting +information which only that CPU accesses and modifies (such as during a +timer tick) thus allowing the accounting data to be updated lockless. Once a +CPU has taken a task to run, it removes it from the global queue. Thus the +global queue only ever has, at most, + + (number of tasks requesting cpu time) - (number of logical CPUs) + 1 + +tasks in the global queue. This value is relevant for the time taken to look up +tasks during scheduling. This will increase if many tasks with CPU affinity set +in their policy to limit which CPUs they're allowed to run on if they outnumber +the number of CPUs. The +1 is because when rescheduling a task, the CPU's +currently running task is put back on the queue. Lookup will be described after +the virtual deadline mechanism is explained. + +Virtual deadline. + +The key to achieving low latency, scheduling fairness, and "nice level" +distribution in BFS is entirely in the virtual deadline mechanism. The one +tunable in BFS is the rr_interval, or "round robin interval". This is the +maximum time two SCHED_OTHER (or SCHED_NORMAL, the common scheduling policy) +tasks of the same nice level will be running for, or looking at it the other +way around, the longest duration two tasks of the same nice level will be +delayed for. When a task requests cpu time, it is given a quota (time_slice) +equal to the rr_interval and a virtual deadline. The virtual deadline is +offset from the current time in jiffies by this equation: + + jiffies + (prio_ratio * rr_interval) + +The prio_ratio is determined as a ratio compared to the baseline of nice -20 +and increases by 10% per nice level. The deadline is a virtual one only in that +no guarantee is placed that a task will actually be scheduled by this time, but +it is used to compare which task should go next. There are three components to +how a task is next chosen. First is time_slice expiration. If a task runs out +of its time_slice, it is descheduled, the time_slice is refilled, and the +deadline reset to that formula above. Second is sleep, where a task no longer +is requesting CPU for whatever reason. The time_slice and deadline are _not_ +adjusted in this case and are just carried over for when the task is next +scheduled. Third is preemption, and that is when a newly waking task is deemed +higher priority than a currently running task on any cpu by virtue of the fact +that it has an earlier virtual deadline than the currently running task. The +earlier deadline is the key to which task is next chosen for the first and +second cases. Once a task is descheduled, it is put back on the queue, and an +O(n) lookup of all queued-but-not-running tasks is done to determine which has +the earliest deadline and that task is chosen to receive CPU next. + +The CPU proportion of different nice tasks works out to be approximately the + + (prio_ratio difference)^2 + +The reason it is squared is that a task's deadline does not change while it is +running unless it runs out of time_slice. Thus, even if the time actually +passes the deadline of another task that is queued, it will not get CPU time +unless the current running task deschedules, and the time "base" (jiffies) is +constantly moving. + +Task lookup. + +BFS has 103 priority queues. 100 of these are dedicated to the static priority +of realtime tasks, and the remaining 3 are, in order of best to worst priority, +SCHED_ISO (isochronous), SCHED_NORMAL, and SCHED_IDLEPRIO (idle priority +scheduling). When a task of these priorities is queued, a bitmap of running +priorities is set showing which of these priorities has tasks waiting for CPU +time. When a CPU is made to reschedule, the lookup for the next task to get +CPU time is performed in the following way: + +First the bitmap is checked to see what static priority tasks are queued. If +any realtime priorities are found, the corresponding queue is checked and the +first task listed there is taken (provided CPU affinity is suitable) and lookup +is complete. If the priority corresponds to a SCHED_ISO task, they are also +taken in FIFO order (as they behave like SCHED_RR). If the priority corresponds +to either SCHED_NORMAL or SCHED_IDLEPRIO, then the lookup becomes O(n). At this +stage, every task in the runlist that corresponds to that priority is checked +to see which has the earliest set deadline, and (provided it has suitable CPU +affinity) it is taken off the runqueue and given the CPU. If a task has an +expired deadline, it is taken and the rest of the lookup aborted (as they are +chosen in FIFO order). + +Thus, the lookup is O(n) in the worst case only, where n is as described +earlier, as tasks may be chosen before the whole task list is looked over. + + +Scalability. + +The major limitations of BFS will be that of scalability, as the separate +runqueue designs will have less lock contention as the number of CPUs rises. +However they do not scale linearly even with separate runqueues as multiple +runqueues will need to be locked concurrently on such designs to be able to +achieve fair CPU balancing, to try and achieve some sort of nice-level fairness +across CPUs, and to achieve low enough latency for tasks on a busy CPU when +other CPUs would be more suited. BFS has the advantage that it requires no +balancing algorithm whatsoever, as balancing occurs by proxy simply because +all CPUs draw off the global runqueue, in priority and deadline order. Despite +the fact that scalability is _not_ the prime concern of BFS, it both shows very +good scalability to smaller numbers of CPUs and is likely a more scalable design +at these numbers of CPUs. + +It also has some very low overhead scalability features built into the design +when it has been deemed their overhead is so marginal that they're worth adding. +The first is the local copy of the running process' data to the CPU it's running +on to allow that data to be updated lockless where possible. Then there is +deference paid to the last CPU a task was running on, by trying that CPU first +when looking for an idle CPU to use the next time it's scheduled. Finally there +is the notion of cache locality beyond the last running CPU. The sched_domains +information is used to determine the relative virtual "cache distance" that +other CPUs have from the last CPU a task was running on. CPUs with shared +caches, such as SMT siblings, or multicore CPUs with shared caches, are treated +as cache local. CPUs without shared caches are treated as not cache local, and +CPUs on different NUMA nodes are treated as very distant. This "relative cache +distance" is used by modifying the virtual deadline value when doing lookups. +Effectively, the deadline is unaltered between "cache local" CPUs, doubled for +"cache distant" CPUs, and quadrupled for "very distant" CPUs. The reasoning +behind the doubling of deadlines is as follows. The real cost of migrating a +task from one CPU to another is entirely dependant on the cache footprint of +the task, how cache intensive the task is, how long it's been running on that +CPU to take up the bulk of its cache, how big the CPU cache is, how fast and +how layered the CPU cache is, how fast a context switch is... and so on. In +other words, it's close to random in the real world where we do more than just +one sole workload. The only thing we can be sure of is that it's not free. So +BFS uses the principle that an idle CPU is a wasted CPU and utilising idle CPUs +is more important than cache locality, and cache locality only plays a part +after that. Doubling the effective deadline is based on the premise that the +"cache local" CPUs will tend to work on the same tasks up to double the number +of cache local CPUs, and once the workload is beyond that amount, it is likely +that none of the tasks are cache warm anywhere anyway. The quadrupling for NUMA +is a value I pulled out of my arse. + +When choosing an idle CPU for a waking task, the cache locality is determined +according to where the task last ran and then idle CPUs are ranked from best +to worst to choose the most suitable idle CPU based on cache locality, NUMA +node locality and hyperthread sibling business. They are chosen in the +following preference (if idle): + +* Same core, idle or busy cache, idle threads +* Other core, same cache, idle or busy cache, idle threads. +* Same node, other CPU, idle cache, idle threads. +* Same node, other CPU, busy cache, idle threads. +* Same core, busy threads. +* Other core, same cache, busy threads. +* Same node, other CPU, busy threads. +* Other node, other CPU, idle cache, idle threads. +* Other node, other CPU, busy cache, idle threads. +* Other node, other CPU, busy threads. + +This shows the SMT or "hyperthread" awareness in the design as well which will +choose a real idle core first before a logical SMT sibling which already has +tasks on the physical CPU. + +Early benchmarking of BFS suggested scalability dropped off at the 16 CPU mark. +However this benchmarking was performed on an earlier design that was far less +scalable than the current one so it's hard to know how scalable it is in terms +of both CPUs (due to the global runqueue) and heavily loaded machines (due to +O(n) lookup) at this stage. Note that in terms of scalability, the number of +_logical_ CPUs matters, not the number of _physical_ CPUs. Thus, a dual (2x) +quad core (4X) hyperthreaded (2X) machine is effectively a 16X. Newer benchmark +results are very promising indeed, without needing to tweak any knobs, features +or options. Benchmark contributions are most welcome. + + +Features + +As the initial prime target audience for BFS was the average desktop user, it +was designed to not need tweaking, tuning or have features set to obtain benefit +from it. Thus the number of knobs and features has been kept to an absolute +minimum and should not require extra user input for the vast majority of cases. +There are precisely 2 tunables, and 2 extra scheduling policies. The rr_interval +and iso_cpu tunables, and the SCHED_ISO and SCHED_IDLEPRIO policies. In addition +to this, BFS also uses sub-tick accounting. What BFS does _not_ now feature is +support for CGROUPS. The average user should neither need to know what these +are, nor should they need to be using them to have good desktop behaviour. + +rr_interval + +There is only one "scheduler" tunable, the round robin interval. This can be +accessed in + + /proc/sys/kernel/rr_interval + +The value is in milliseconds, and the default value is set to 6 on a +uniprocessor machine, and automatically set to a progressively higher value on +multiprocessor machines. The reasoning behind increasing the value on more CPUs +is that the effective latency is decreased by virtue of there being more CPUs on +BFS (for reasons explained above), and increasing the value allows for less +cache contention and more throughput. Valid values are from 1 to 1000 +Decreasing the value will decrease latencies at the cost of decreasing +throughput, while increasing it will improve throughput, but at the cost of +worsening latencies. The accuracy of the rr interval is limited by HZ resolution +of the kernel configuration. Thus, the worst case latencies are usually slightly +higher than this actual value. The default value of 6 is not an arbitrary one. +It is based on the fact that humans can detect jitter at approximately 7ms, so +aiming for much lower latencies is pointless under most circumstances. It is +worth noting this fact when comparing the latency performance of BFS to other +schedulers. Worst case latencies being higher than 7ms are far worse than +average latencies not being in the microsecond range. + +Isochronous scheduling. + +Isochronous scheduling is a unique scheduling policy designed to provide +near-real-time performance to unprivileged (ie non-root) users without the +ability to starve the machine indefinitely. Isochronous tasks (which means +"same time") are set using, for example, the schedtool application like so: + + schedtool -I -e amarok + +This will start the audio application "amarok" as SCHED_ISO. How SCHED_ISO works +is that it has a priority level between true realtime tasks and SCHED_NORMAL +which would allow them to preempt all normal tasks, in a SCHED_RR fashion (ie, +if multiple SCHED_ISO tasks are running, they purely round robin at rr_interval +rate). However if ISO tasks run for more than a tunable finite amount of time, +they are then demoted back to SCHED_NORMAL scheduling. This finite amount of +time is the percentage of _total CPU_ available across the machine, configurable +as a percentage in the following "resource handling" tunable (as opposed to a +scheduler tunable): + + /proc/sys/kernel/iso_cpu + +and is set to 70% by default. It is calculated over a rolling 5 second average +Because it is the total CPU available, it means that on a multi CPU machine, it +is possible to have an ISO task running as realtime scheduling indefinitely on +just one CPU, as the other CPUs will be available. Setting this to 100 is the +equivalent of giving all users SCHED_RR access and setting it to 0 removes the +ability to run any pseudo-realtime tasks. + +A feature of BFS is that it detects when an application tries to obtain a +realtime policy (SCHED_RR or SCHED_FIFO) and the caller does not have the +appropriate privileges to use those policies. When it detects this, it will +give the task SCHED_ISO policy instead. Thus it is transparent to the user. +Because some applications constantly set their policy as well as their nice +level, there is potential for them to undo the override specified by the user +on the command line of setting the policy to SCHED_ISO. To counter this, once +a task has been set to SCHED_ISO policy, it needs superuser privileges to set +it back to SCHED_NORMAL. This will ensure the task remains ISO and all child +processes and threads will also inherit the ISO policy. + +Idleprio scheduling. + +Idleprio scheduling is a scheduling policy designed to give out CPU to a task +_only_ when the CPU would be otherwise idle. The idea behind this is to allow +ultra low priority tasks to be run in the background that have virtually no +effect on the foreground tasks. This is ideally suited to distributed computing +clients (like setiathome, folding, mprime etc) but can also be used to start +a video encode or so on without any slowdown of other tasks. To avoid this +policy from grabbing shared resources and holding them indefinitely, if it +detects a state where the task is waiting on I/O, the machine is about to +suspend to ram and so on, it will transiently schedule them as SCHED_NORMAL. As +per the Isochronous task management, once a task has been scheduled as IDLEPRIO, +it cannot be put back to SCHED_NORMAL without superuser privileges. Tasks can +be set to start as SCHED_IDLEPRIO with the schedtool command like so: + + schedtool -D -e ./mprime + +Subtick accounting. + +It is surprisingly difficult to get accurate CPU accounting, and in many cases, +the accounting is done by simply determining what is happening at the precise +moment a timer tick fires off. This becomes increasingly inaccurate as the +timer tick frequency (HZ) is lowered. It is possible to create an application +which uses almost 100% CPU, yet by being descheduled at the right time, records +zero CPU usage. While the main problem with this is that there are possible +security implications, it is also difficult to determine how much CPU a task +really does use. BFS tries to use the sub-tick accounting from the TSC clock, +where possible, to determine real CPU usage. This is not entirely reliable, but +is far more likely to produce accurate CPU usage data than the existing designs +and will not show tasks as consuming no CPU usage when they actually are. Thus, +the amount of CPU reported as being used by BFS will more accurately represent +how much CPU the task itself is using (as is shown for example by the 'time' +application), so the reported values may be quite different to other schedulers. +Values reported as the 'load' are more prone to problems with this design, but +per process values are closer to real usage. When comparing throughput of BFS +to other designs, it is important to compare the actual completed work in terms +of total wall clock time taken and total work done, rather than the reported +"cpu usage". + + +Con Kolivas Fri Aug 27 2010 diff --git a/Documentation/scheduler/sched-MuQSS.txt b/Documentation/scheduler/sched-MuQSS.txt new file mode 100644 index 00000000000..ae28b85c999 --- /dev/null +++ b/Documentation/scheduler/sched-MuQSS.txt @@ -0,0 +1,373 @@ +MuQSS - The Multiple Queue Skiplist Scheduler by Con Kolivas. + +MuQSS is a per-cpu runqueue variant of the original BFS scheduler with +one 8 level skiplist per runqueue, and fine grained locking for much more +scalability. + + +Goals. + +The goal of the Multiple Queue Skiplist Scheduler, referred to as MuQSS from +here on (pronounced mux) is to completely do away with the complex designs of +the past for the cpu process scheduler and instead implement one that is very +simple in basic design. The main focus of MuQSS is to achieve excellent desktop +interactivity and responsiveness without heuristics and tuning knobs that are +difficult to understand, impossible to model and predict the effect of, and when +tuned to one workload cause massive detriment to another, while still being +scalable to many CPUs and processes. + + +Design summary. + +MuQSS is best described as per-cpu multiple runqueue, O(log n) insertion, O(1) +lookup, earliest effective virtual deadline first tickless design, loosely based +on EEVDF (earliest eligible virtual deadline first) and my previous Staircase +Deadline scheduler, and evolved from the single runqueue O(n) BFS scheduler. +Each component shall be described in order to understand the significance of, +and reasoning for it. + + +Design reasoning. + +In BFS, the use of a single runqueue across all CPUs meant that each CPU would +need to scan the entire runqueue looking for the process with the earliest +deadline and schedule that next, regardless of which CPU it originally came +from. This made BFS deterministic with respect to latency and provided +guaranteed latencies dependent on number of processes and CPUs. The single +runqueue, however, meant that all CPUs would compete for the single lock +protecting it, which would lead to increasing lock contention as the number of +CPUs rose and appeared to limit scalability of common workloads beyond 16 +logical CPUs. Additionally, the O(n) lookup of the runqueue list obviously +increased overhead proportionate to the number of queued proecesses and led to +cache thrashing while iterating over the linked list. + +MuQSS is an evolution of BFS, designed to maintain the same scheduling +decision mechanism and be virtually deterministic without relying on the +constrained design of the single runqueue by splitting out the single runqueue +to be per-CPU and use skiplists instead of linked lists. + +The original reason for going back to a single runqueue design for BFS was that +once multiple runqueues are introduced, per-CPU or otherwise, there will be +complex interactions as each runqueue will be responsible for the scheduling +latency and fairness of the tasks only on its own runqueue, and to achieve +fairness and low latency across multiple CPUs, any advantage in throughput of +having CPU local tasks causes other disadvantages. This is due to requiring a +very complex balancing system to at best achieve some semblance of fairness +across CPUs and can only maintain relatively low latency for tasks bound to the +same CPUs, not across them. To increase said fairness and latency across CPUs, +the advantage of local runqueue locking, which makes for better scalability, is +lost due to having to grab multiple locks. + +MuQSS works around the problems inherent in multiple runqueue designs by +making its skip lists priority ordered and through novel use of lockless +examination of each other runqueue it can decide if it should take the earliest +deadline task from another runqueue for latency reasons, or for CPU balancing +reasons. It still does not have a balancing system, choosing to allow the +next task scheduling decision and task wakeup CPU choice to allow balancing to +happen by virtue of its choices. + +As a further evolution of the design, MuQSS normally configures sharing of +runqueues in a logical fashion for when CPU resources are shared for improved +latency and throughput. By default it shares runqueues and locks between +multicore siblings. Optionally it can be configured to run with sharing of +SMT siblings only, all SMP packages or no sharing at all. Additionally it can +be selected at boot time. + + +Design details. + +Custom skip list implementation: + +To avoid the overhead of building up and tearing down skip list structures, +the variant used by MuQSS has a number of optimisations making it specific for +its use case in the scheduler. It uses static arrays of 8 'levels' instead of +building up and tearing down structures dynamically. This makes each runqueue +only scale O(log N) up to 64k tasks. However as there is one runqueue per CPU +it means that it scales O(log N) up to 64k x number of logical CPUs which is +far beyond the realistic task limits each CPU could handle. By being 8 levels +it also makes the array exactly one cacheline in size. Additionally, each +skip list node is bidirectional making insertion and removal amortised O(1), +being O(k) where k is 1-8. Uniquely, we are only ever interested in the very +first entry in each list at all times with MuQSS, so there is never a need to +do a search and thus look up is always O(1). In interactive mode, the queues +will be searched beyond their first entry if the first task is not suitable +for affinity or SMT nice reasons. + +Task insertion: + +MuQSS inserts tasks into a per CPU runqueue as an O(log N) insertion into +a custom skip list as described above (based on the original design by William +Pugh). Insertion is ordered in such a way that there is never a need to do a +search by ordering tasks according to static priority primarily, and then +virtual deadline at the time of insertion. + +Niffies: + +Niffies are a monotonic forward moving timer not unlike the "jiffies" but are +of nanosecond resolution. Niffies are calculated per-runqueue from the high +resolution TSC timers, and in order to maintain fairness are synchronised +between CPUs whenever both runqueues are locked concurrently. + +Virtual deadline: + +The key to achieving low latency, scheduling fairness, and "nice level" +distribution in MuQSS is entirely in the virtual deadline mechanism. The one +tunable in MuQSS is the rr_interval, or "round robin interval". This is the +maximum time two SCHED_OTHER (or SCHED_NORMAL, the common scheduling policy) +tasks of the same nice level will be running for, or looking at it the other +way around, the longest duration two tasks of the same nice level will be +delayed for. When a task requests cpu time, it is given a quota (time_slice) +equal to the rr_interval and a virtual deadline. The virtual deadline is +offset from the current time in niffies by this equation: + + niffies + (prio_ratio * rr_interval) + +The prio_ratio is determined as a ratio compared to the baseline of nice -20 +and increases by 10% per nice level. The deadline is a virtual one only in that +no guarantee is placed that a task will actually be scheduled by this time, but +it is used to compare which task should go next. There are three components to +how a task is next chosen. First is time_slice expiration. If a task runs out +of its time_slice, it is descheduled, the time_slice is refilled, and the +deadline reset to that formula above. Second is sleep, where a task no longer +is requesting CPU for whatever reason. The time_slice and deadline are _not_ +adjusted in this case and are just carried over for when the task is next +scheduled. Third is preemption, and that is when a newly waking task is deemed +higher priority than a currently running task on any cpu by virtue of the fact +that it has an earlier virtual deadline than the currently running task. The +earlier deadline is the key to which task is next chosen for the first and +second cases. + +The CPU proportion of different nice tasks works out to be approximately the + + (prio_ratio difference)^2 + +The reason it is squared is that a task's deadline does not change while it is +running unless it runs out of time_slice. Thus, even if the time actually +passes the deadline of another task that is queued, it will not get CPU time +unless the current running task deschedules, and the time "base" (niffies) is +constantly moving. + +Task lookup: + +As tasks are already pre-ordered according to anticipated scheduling order in +the skip lists, lookup for the next suitable task per-runqueue is always a +matter of simply selecting the first task in the 0th level skip list entry. +In order to maintain optimal latency and fairness across CPUs, MuQSS does a +novel examination of every other runqueue in cache locality order, choosing the +best task across all runqueues. This provides near-determinism of how long any +task across the entire system may wait before receiving CPU time. The other +runqueues are first examine lockless and then trylocked to minimise the +potential lock contention if they are likely to have a suitable better task. +Each other runqueue lock is only held for as long as it takes to examine the +entry for suitability. In "interactive" mode, the default setting, MuQSS will +look for the best deadline task across all CPUs, while in !interactive mode, +it will only select a better deadline task from another CPU if it is more +heavily laden than the current one. + +Lookup is therefore O(k) where k is number of CPUs. + + +Latency. + +Through the use of virtual deadlines to govern the scheduling order of normal +tasks, queue-to-activation latency per runqueue is guaranteed to be bound by +the rr_interval tunable which is set to 6ms by default. This means that the +longest a CPU bound task will wait for more CPU is proportional to the number +of running tasks and in the common case of 0-2 running tasks per CPU, will be +under the 7ms threshold for human perception of jitter. Additionally, as newly +woken tasks will have an early deadline from their previous runtime, the very +tasks that are usually latency sensitive will have the shortest interval for +activation, usually preempting any existing CPU bound tasks. + +Tickless expiry: + +A feature of MuQSS is that it is not tied to the resolution of the chosen tick +rate in Hz, instead depending entirely on the high resolution timers where +possible for sub-millisecond accuracy on timeouts regarless of the underlying +tick rate. This allows MuQSS to be run with the low overhead of low Hz rates +such as 100 by default, benefiting from the improved throughput and lower +power usage it provides. Another advantage of this approach is that in +combination with the Full No HZ option, which disables ticks on running task +CPUs instead of just idle CPUs, the tick can be disabled at all times +regardless of how many tasks are running instead of being limited to just one +running task. Note that this option is NOT recommended for regular desktop +users. + + +Scalability and balancing. + +Unlike traditional approaches where balancing is a combination of CPU selection +at task wakeup and intermittent balancing based on a vast array of rules set +according to architecture, busyness calculations and special case management, +MuQSS indirectly balances on the fly at task wakeup and next task selection. +During initialisation, MuQSS creates a cache coherency ordered list of CPUs for +each logical CPU and uses this to aid task/CPU selection when CPUs are busy. +Additionally it selects any idle CPUs, if they are available, at any time over +busy CPUs according to the following preference: + + * Same thread, idle or busy cache, idle or busy threads + * Other core, same cache, idle or busy cache, idle threads. + * Same node, other CPU, idle cache, idle threads. + * Same node, other CPU, busy cache, idle threads. + * Other core, same cache, busy threads. + * Same node, other CPU, busy threads. + * Other node, other CPU, idle cache, idle threads. + * Other node, other CPU, busy cache, idle threads. + * Other node, other CPU, busy threads. + +Mux is therefore SMT, MC and Numa aware without the need for extra +intermittent balancing to maintain CPUs busy and make the most of cache +coherency. + + +Features + +As the initial prime target audience for MuQSS was the average desktop user, it +was designed to not need tweaking, tuning or have features set to obtain benefit +from it. Thus the number of knobs and features has been kept to an absolute +minimum and should not require extra user input for the vast majority of cases. +There are 3 optional tunables, and 2 extra scheduling policies. The rr_interval, +interactive, and iso_cpu tunables, and the SCHED_ISO and SCHED_IDLEPRIO +policies. In addition to this, MuQSS also uses sub-tick accounting. What MuQSS +does _not_ now feature is support for CGROUPS. The average user should neither +need to know what these are, nor should they need to be using them to have good +desktop behaviour. However since some applications refuse to work without +cgroups, one can enable them with MuQSS as a stub and the filesystem will be +created which will allow the applications to work. + +rr_interval: + + /proc/sys/kernel/rr_interval + +The value is in milliseconds, and the default value is set to 6. Valid values +are from 1 to 1000 Decreasing the value will decrease latencies at the cost of +decreasing throughput, while increasing it will improve throughput, but at the +cost of worsening latencies. It is based on the fact that humans can detect +jitter at approximately 7ms, so aiming for much lower latencies is pointless +under most circumstances. It is worth noting this fact when comparing the +latency performance of MuQSS to other schedulers. Worst case latencies being +higher than 7ms are far worse than average latencies not being in the +microsecond range. + +interactive: + + /proc/sys/kernel/interactive + +The value is a simple boolean of 1 for on and 0 for off and is set to on by +default. Disabling this will disable the near-determinism of MuQSS when +selecting the next task by not examining all CPUs for the earliest deadline +task, or which CPU to wake to, instead prioritising CPU balancing for improved +throughput. Latency will still be bound by rr_interval, but on a per-CPU basis +instead of across the whole system. + +Runqueue sharing. + +By default MuQSS chooses to share runqueue resources (specifically the skip +list and locking) between multicore siblings. It is configurable at build time +to select between None, SMT, MC and SMP, corresponding to no sharing, sharing +only between simultaneous mulithreading siblings, multicore siblings, or +symmetric multiprocessing physical packages. Additionally it can be se at +bootime with the use of the rqshare parameter. The reason for configurability +is that some architectures have CPUs with many multicore siblings (>= 16) +where it may be detrimental to throughput to share runqueues and another +sharing option may be desirable. Additionally, more sharing than usual can +improve latency on a system-wide level at the expense of throughput if desired. + +The options are: +none, smt, mc, smp + +eg: + rqshare=mc + +Isochronous scheduling: + +Isochronous scheduling is a unique scheduling policy designed to provide +near-real-time performance to unprivileged (ie non-root) users without the +ability to starve the machine indefinitely. Isochronous tasks (which means +"same time") are set using, for example, the schedtool application like so: + + schedtool -I -e amarok + +This will start the audio application "amarok" as SCHED_ISO. How SCHED_ISO works +is that it has a priority level between true realtime tasks and SCHED_NORMAL +which would allow them to preempt all normal tasks, in a SCHED_RR fashion (ie, +if multiple SCHED_ISO tasks are running, they purely round robin at rr_interval +rate). However if ISO tasks run for more than a tunable finite amount of time, +they are then demoted back to SCHED_NORMAL scheduling. This finite amount of +time is the percentage of CPU available per CPU, configurable as a percentage in +the following "resource handling" tunable (as opposed to a scheduler tunable): + +iso_cpu: + + /proc/sys/kernel/iso_cpu + +and is set to 70% by default. It is calculated over a rolling 5 second average +Because it is the total CPU available, it means that on a multi CPU machine, it +is possible to have an ISO task running as realtime scheduling indefinitely on +just one CPU, as the other CPUs will be available. Setting this to 100 is the +equivalent of giving all users SCHED_RR access and setting it to 0 removes the +ability to run any pseudo-realtime tasks. + +A feature of MuQSS is that it detects when an application tries to obtain a +realtime policy (SCHED_RR or SCHED_FIFO) and the caller does not have the +appropriate privileges to use those policies. When it detects this, it will +give the task SCHED_ISO policy instead. Thus it is transparent to the user. + + +Idleprio scheduling: + +Idleprio scheduling is a scheduling policy designed to give out CPU to a task +_only_ when the CPU would be otherwise idle. The idea behind this is to allow +ultra low priority tasks to be run in the background that have virtually no +effect on the foreground tasks. This is ideally suited to distributed computing +clients (like setiathome, folding, mprime etc) but can also be used to start a +video encode or so on without any slowdown of other tasks. To avoid this policy +from grabbing shared resources and holding them indefinitely, if it detects a +state where the task is waiting on I/O, the machine is about to suspend to ram +and so on, it will transiently schedule them as SCHED_NORMAL. Once a task has +been scheduled as IDLEPRIO, it cannot be put back to SCHED_NORMAL without +superuser privileges since it is effectively a lower scheduling policy. Tasks +can be set to start as SCHED_IDLEPRIO with the schedtool command like so: + +schedtool -D -e ./mprime + +Subtick accounting: + +It is surprisingly difficult to get accurate CPU accounting, and in many cases, +the accounting is done by simply determining what is happening at the precise +moment a timer tick fires off. This becomes increasingly inaccurate as the timer +tick frequency (HZ) is lowered. It is possible to create an application which +uses almost 100% CPU, yet by being descheduled at the right time, records zero +CPU usage. While the main problem with this is that there are possible security +implications, it is also difficult to determine how much CPU a task really does +use. Mux uses sub-tick accounting from the TSC clock to determine real CPU +usage. Thus, the amount of CPU reported as being used by MuQSS will more +accurately represent how much CPU the task itself is using (as is shown for +example by the 'time' application), so the reported values may be quite +different to other schedulers. When comparing throughput of MuQSS to other +designs, it is important to compare the actual completed work in terms of total +wall clock time taken and total work done, rather than the reported "cpu usage". + +Symmetric MultiThreading (SMT) aware nice: + +SMT, a.k.a. hyperthreading, is a very common feature on modern CPUs. While the +logical CPU count rises by adding thread units to each CPU core, allowing more +than one task to be run simultaneously on the same core, the disadvantage of it +is that the CPU power is shared between the tasks, not summating to the power +of two CPUs. The practical upshot of this is that two tasks running on +separate threads of the same core run significantly slower than if they had one +core each to run on. While smart CPU selection allows each task to have a core +to itself whenever available (as is done on MuQSS), it cannot offset the +slowdown that occurs when the cores are all loaded and only a thread is left. +Most of the time this is harmless as the CPU is effectively overloaded at this +point and the extra thread is of benefit. However when running a niced task in +the presence of an un-niced task (say nice 19 v nice 0), the nice task gets +precisely the same amount of CPU power as the unniced one. MuQSS has an +optional configuration feature known as SMT-NICE which selectively idles the +secondary niced thread for a period proportional to the nice difference, +allowing CPU distribution according to nice level to be maintained, at the +expense of a small amount of extra overhead. If this is configured in on a +machine without SMT threads, the overhead is minimal. + + +Con Kolivas Sat, 29th October 2016 diff --git a/Documentation/vm/index.rst b/Documentation/vm/index.rst index eff5fbd492d..c353b3f5592 100644 --- a/Documentation/vm/index.rst +++ b/Documentation/vm/index.rst @@ -17,6 +17,7 @@ various features of the Linux memory management swap_numa zswap + multigen_lru Kernel developers MM documentation ================================== diff --git a/Documentation/vm/multigen_lru.rst b/Documentation/vm/multigen_lru.rst new file mode 100644 index 00000000000..a18416ed7e9 --- /dev/null +++ b/Documentation/vm/multigen_lru.rst @@ -0,0 +1,143 @@ +.. SPDX-License-Identifier: GPL-2.0 + +===================== +Multigenerational LRU +===================== + +Quick Start +=========== +Build Options +------------- +:Required: Set ``CONFIG_LRU_GEN=y``. + +:Optional: Set ``CONFIG_LRU_GEN_ENABLED=y`` to turn the feature on by + default. + +:Optional: Change ``CONFIG_NR_LRU_GENS`` to a number ``X`` to support + a maximum of ``X`` generations. + +:Optional: Change ``CONFIG_TIERS_PER_GEN`` to a number ``Y`` to + support a maximum of ``Y`` tiers per generation. + +Runtime Options +--------------- +:Required: Write ``1`` to ``/sys/kernel/mm/lru_gen/enable`` if the + feature was not turned on by default. + +:Optional: Change ``/sys/kernel/mm/lru_gen/spread`` to a number ``N`` + to spread pages out across ``N+1`` generations. ``N`` should be less + than ``X``. Larger values make the background aging more aggressive. + +:Optional: Read ``/sys/kernel/debug/lru_gen`` to verify the feature. + This file has the following output: + +:: + + memcg memcg_id memcg_path + node node_id + min_gen birth_time anon_size file_size + ... + max_gen birth_time anon_size file_size + +Given a memcg and a node, ``min_gen`` is the oldest generation +(number) and ``max_gen`` is the youngest. Birth time is in +milliseconds. The sizes of anon and file types are in pages. + +Recipes +------- +:Android on ARMv8.1+: ``X=4``, ``Y=3`` and ``N=0``. + +:Android on pre-ARMv8.1 CPUs: Not recommended due to the lack of + ``ARM64_HW_AFDBM``. + +:Laptops and workstations running Chrome on x86_64: Use the default + values. + +:Working set estimation: Write ``+ memcg_id node_id gen [swappiness]`` + to ``/sys/kernel/debug/lru_gen`` to account referenced pages to + generation ``max_gen`` and create the next generation ``max_gen+1``. + ``gen`` should be equal to ``max_gen``. A swap file and a non-zero + ``swappiness`` are required to scan anon type. If swapping is not + desired, set ``vm.swappiness`` to ``0``. + +:Proactive reclaim: Write ``- memcg_id node_id gen [swappiness] + [nr_to_reclaim]`` to ``/sys/kernel/debug/lru_gen`` to evict + generations less than or equal to ``gen``. ``gen`` should be less + than ``max_gen-1`` as ``max_gen`` and ``max_gen-1`` are active + generations and therefore protected from the eviction. Use + ``nr_to_reclaim`` to limit the number of pages to evict. Multiple + command lines are supported, so does concatenation with delimiters + ``,`` and ``;``. + +Framework +========= +For each ``lruvec``, evictable pages are divided into multiple +generations. The youngest generation number is stored in ``max_seq`` +for both anon and file types as they are aged on an equal footing. The +oldest generation numbers are stored in ``min_seq[2]`` separately for +anon and file types as clean file pages can be evicted regardless of +swap and write-back constraints. These three variables are +monotonically increasing. Generation numbers are truncated into +``order_base_2(CONFIG_NR_LRU_GENS+1)`` bits in order to fit into +``page->flags``. The sliding window technique is used to prevent +truncated generation numbers from overlapping. Each truncated +generation number is an index to an array of per-type and per-zone +lists. Evictable pages are added to the per-zone lists indexed by +``max_seq`` or ``min_seq[2]`` (modulo ``CONFIG_NR_LRU_GENS``), +depending on their types. + +Each generation is then divided into multiple tiers. Tiers represent +levels of usage from file descriptors only. Pages accessed N times via +file descriptors belong to tier order_base_2(N). Each generation +contains at most CONFIG_TIERS_PER_GEN tiers, and they require +additional CONFIG_TIERS_PER_GEN-2 bits in page->flags. In contrast to +moving across generations which requires the lru lock for the list +operations, moving across tiers only involves an atomic operation on +``page->flags`` and therefore has a negligible cost. A feedback loop +modeled after the PID controller monitors the refault rates across all +tiers and decides when to activate pages from which tiers in the +reclaim path. + +The framework comprises two conceptually independent components: the +aging and the eviction, which can be invoked separately from user +space for the purpose of working set estimation and proactive reclaim. + +Aging +----- +The aging produces young generations. Given an ``lruvec``, the aging +scans page tables for referenced pages of this ``lruvec``. Upon +finding one, the aging updates its generation number to ``max_seq``. +After each round of scan, the aging increments ``max_seq``. + +The aging maintains either a system-wide ``mm_struct`` list or +per-memcg ``mm_struct`` lists, and it only scans page tables of +processes that have been scheduled since the last scan. + +The aging is due when both of ``min_seq[2]`` reaches ``max_seq-1``, +assuming both anon and file types are reclaimable. + +Eviction +-------- +The eviction consumes old generations. Given an ``lruvec``, the +eviction scans the pages on the per-zone lists indexed by either of +``min_seq[2]``. It first tries to select a type based on the values of +``min_seq[2]``. When anon and file types are both available from the +same generation, it selects the one that has a lower refault rate. + +During a scan, the eviction sorts pages according to their new +generation numbers, if the aging has found them referenced. It also +moves pages from the tiers that have higher refault rates than tier 0 +to the next generation. + +When it finds all the per-zone lists of a selected type are empty, the +eviction increments ``min_seq[2]`` indexed by this selected type. + +To-do List +========== +KVM Optimization +---------------- +Support shadow page table scanning. + +NUMA Optimization +----------------- +Optimize page table scan for NUMA. diff --git a/Documentation/vm/uksm.txt b/Documentation/vm/uksm.txt new file mode 100644 index 00000000000..be19a312700 --- /dev/null +++ b/Documentation/vm/uksm.txt @@ -0,0 +1,61 @@ +The Ultra Kernel Samepage Merging feature +---------------------------------------------- +/* + * Ultra KSM. Copyright (C) 2011-2012 Nai Xia + * + * This is an improvement upon KSM. Some basic data structures and routines + * are borrowed from ksm.c . + * + * Its new features: + * 1. Full system scan: + * It automatically scans all user processes' anonymous VMAs. Kernel-user + * interaction to submit a memory area to KSM is no longer needed. + * + * 2. Rich area detection: + * It automatically detects rich areas containing abundant duplicated + * pages based. Rich areas are given a full scan speed. Poor areas are + * sampled at a reasonable speed with very low CPU consumption. + * + * 3. Ultra Per-page scan speed improvement: + * A new hash algorithm is proposed. As a result, on a machine with + * Core(TM)2 Quad Q9300 CPU in 32-bit mode and 800MHZ DDR2 main memory, it + * can scan memory areas that does not contain duplicated pages at speed of + * 627MB/sec ~ 2445MB/sec and can merge duplicated areas at speed of + * 477MB/sec ~ 923MB/sec. + * + * 4. Thrashing area avoidance: + * Thrashing area(an VMA that has frequent Ksm page break-out) can be + * filtered out. My benchmark shows it's more efficient than KSM's per-page + * hash value based volatile page detection. + * + * + * 5. Misc changes upon KSM: + * * It has a fully x86-opitmized memcmp dedicated for 4-byte-aligned page + * comparison. It's much faster than default C version on x86. + * * rmap_item now has an struct *page member to loosely cache a + * address-->page mapping, which reduces too much time-costly + * follow_page(). + * * The VMA creation/exit procedures are hooked to let the Ultra KSM know. + * * try_to_merge_two_pages() now can revert a pte if it fails. No break_ + * ksm is needed for this case. + * + * 6. Full Zero Page consideration(contributed by Figo Zhang) + * Now uksmd consider full zero pages as special pages and merge them to an + * special unswappable uksm zero page. + */ + +ChangeLog: + +2012-05-05 The creation of this Doc +2012-05-08 UKSM 0.1.1.1 libc crash bug fix, api clean up, doc clean up. +2012-05-28 UKSM 0.1.1.2 bug fix release +2012-06-26 UKSM 0.1.2-beta1 first beta release for 0.1.2 +2012-07-2 UKSM 0.1.2-beta2 +2012-07-10 UKSM 0.1.2-beta3 +2012-07-26 UKSM 0.1.2 Fine grained speed control, more scan optimization. +2012-10-13 UKSM 0.1.2.1 Bug fixes. +2012-12-31 UKSM 0.1.2.2 Minor bug fixes. +2014-07-02 UKSM 0.1.2.3 Fix a " __this_cpu_read() in preemptible bug". +2015-04-22 UKSM 0.1.2.4 Fix a race condition that can sometimes trigger anonying warnings. +2016-09-10 UKSM 0.1.2.5 Fix a bug in dedup ratio calculation. +2017-02-26 UKSM 0.1.2.6 Fix a bug in hugetlbpage handling and a race bug with page migration. diff --git a/MAINTAINERS b/MAINTAINERS index 0cce91cd562..bbaa188a2ee 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -3040,6 +3040,19 @@ F: include/uapi/linux/audit.h F: kernel/audit* F: lib/*audit.c +AUFS (advanced multi layered unification filesystem) FILESYSTEM +M: "J. R. Okajima" +L: aufs-users@lists.sourceforge.net (members only) +L: linux-unionfs@vger.kernel.org +S: Supported +W: http://aufs.sourceforge.net +T: git://github.com/sfjro/aufs4-linux.git +F: Documentation/ABI/testing/debugfs-aufs +F: Documentation/ABI/testing/sysfs-aufs +F: Documentation/filesystems/aufs/ +F: fs/aufs/ +F: include/uapi/linux/aufs_type.h + AUXILIARY DISPLAY DRIVERS M: Miguel Ojeda S: Maintained @@ -7521,7 +7534,7 @@ F: Documentation/locking/*futex* F: include/asm-generic/futex.h F: include/linux/futex.h F: include/uapi/linux/futex.h -F: kernel/futex.c +F: kernel/futex* F: tools/perf/bench/futex* F: tools/testing/selftests/futex/ diff --git a/Makefile b/Makefile index a846e4a86c4..a227e810d26 100644 --- a/Makefile +++ b/Makefile @@ -2,8 +2,8 @@ VERSION = 5 PATCHLEVEL = 13 SUBLEVEL = 9 -EXTRAVERSION = -gnu -NAME = Opossums on Parade +EXTRAVERSION = -gnu-pck1 +NAME = The City on the Edge of Forever # *DOCUMENTATION* # To see a list of typical targets execute "make help" @@ -1003,8 +1003,7 @@ KBUILD_CFLAGS += -fno-strict-overflow # Make sure -fstack-check isn't enabled (like gentoo apparently did) KBUILD_CFLAGS += -fno-stack-check -# conserve stack if available -KBUILD_CFLAGS += $(call cc-option,-fconserve-stack) + # Prohibit date/time macros, which would make the build non-deterministic KBUILD_CFLAGS += -Werror=date-time diff --git a/arch/Kconfig b/arch/Kconfig index c45b770d357..e3812adc69f 100644 --- a/arch/Kconfig +++ b/arch/Kconfig @@ -826,6 +826,15 @@ config HAVE_ARCH_TRANSPARENT_HUGEPAGE config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD bool +config HAVE_ARCH_PARENT_PMD_YOUNG + bool + depends on PGTABLE_LEVELS > 2 + help + Architectures that select this are able to set the accessed bit on + non-leaf PMD entries in addition to leaf PTE entries where pages are + mapped. For them, page table walkers that clear the accessed bit may + stop at non-leaf PMD entries if they do not see the accessed bit. + config HAVE_ARCH_HUGE_VMAP bool diff --git a/arch/alpha/Kconfig b/arch/alpha/Kconfig index 5998106faa6..e1a96348428 100644 --- a/arch/alpha/Kconfig +++ b/arch/alpha/Kconfig @@ -674,6 +674,8 @@ config HZ default 1200 if HZ_1200 default 1024 +source "kernel/Kconfig.MuQSS" + config SRM_ENV tristate "SRM environment through procfs" depends on PROC_FS diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig index 24804f11302..2c31f181c49 100644 --- a/arch/arm/Kconfig +++ b/arch/arm/Kconfig @@ -1197,6 +1197,8 @@ config SCHED_SMT MultiThreading at a cost of slightly increased overhead in some places. If unsure say N here. +source "kernel/Kconfig.MuQSS" + config HAVE_ARM_SCU bool help diff --git a/arch/arm/tools/syscall.tbl b/arch/arm/tools/syscall.tbl index 28e03b5fec0..25f175ada12 100644 --- a/arch/arm/tools/syscall.tbl +++ b/arch/arm/tools/syscall.tbl @@ -460,3 +460,7 @@ 444 common landlock_create_ruleset sys_landlock_create_ruleset 445 common landlock_add_rule sys_landlock_add_rule 446 common landlock_restrict_self sys_landlock_restrict_self +447 common futex_wait sys_futex_wait +448 common futex_wake sys_futex_wake +449 common futex_waitv sys_futex_waitv +450 common futex_requeue sys_futex_requeue diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig index 9f1d8566bbf..120ef842092 100644 --- a/arch/arm64/Kconfig +++ b/arch/arm64/Kconfig @@ -1006,6 +1006,8 @@ config SCHED_SMT MultiThreading at a cost of slightly increased overhead in some places. If unsure say N here. +source "kernel/Kconfig.MuQSS" + config NR_CPUS int "Maximum number of CPUs (2-4096)" range 2 4096 diff --git a/arch/arm64/include/asm/unistd.h b/arch/arm64/include/asm/unistd.h index 727bfc3be99..4e65da3445c 100644 --- a/arch/arm64/include/asm/unistd.h +++ b/arch/arm64/include/asm/unistd.h @@ -38,7 +38,7 @@ #define __ARM_NR_compat_set_tls (__ARM_NR_COMPAT_BASE + 5) #define __ARM_NR_COMPAT_END (__ARM_NR_COMPAT_BASE + 0x800) -#define __NR_compat_syscalls 447 +#define __NR_compat_syscalls 451 #endif #define __ARCH_WANT_SYS_CLONE diff --git a/arch/arm64/include/asm/unistd32.h b/arch/arm64/include/asm/unistd32.h index 5dab69d2c22..1749cc10844 100644 --- a/arch/arm64/include/asm/unistd32.h +++ b/arch/arm64/include/asm/unistd32.h @@ -900,6 +900,10 @@ __SYSCALL(__NR_landlock_create_ruleset, sys_landlock_create_ruleset) __SYSCALL(__NR_landlock_add_rule, sys_landlock_add_rule) #define __NR_landlock_restrict_self 446 __SYSCALL(__NR_landlock_restrict_self, sys_landlock_restrict_self) +#define __NR_futex_wait 447 +__SYSCALL(__NR_futex_wait, sys_futex_wait) +#define __NR_futex_wake 448 +__SYSCALL(__NR_futex_wake, sys_futex_wake) /* * Please add new compat syscalls above this comment and update diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 088dd2afcfe..bf97c467a12 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -873,6 +873,8 @@ config SCHED_SMT when dealing with POWER5 cpus at a cost of slightly increased overhead in some places. If unsure say N here. +source "kernel/Kconfig.MuQSS" + config PPC_DENORMALISATION bool "PowerPC denormalisation exception handling" depends on PPC_BOOK3S_64 diff --git a/arch/powerpc/platforms/cell/spufs/sched.c b/arch/powerpc/platforms/cell/spufs/sched.c index 36920648989..53723e28240 100644 --- a/arch/powerpc/platforms/cell/spufs/sched.c +++ b/arch/powerpc/platforms/cell/spufs/sched.c @@ -54,7 +54,9 @@ static struct timer_list spuloadavg_timer; /* * Priority of a normal, non-rt, non-niced'd process (aka nice level 0). */ +#ifndef CONFIG_SCHED_MUQSS #define NORMAL_PRIO 120 +#endif /* * Frequency of the spu scheduler tick. By default we do one SPU scheduler diff --git a/arch/s390/include/asm/preempt.h b/arch/s390/include/asm/preempt.h index d9d5350cc3e..36df78112d5 100644 --- a/arch/s390/include/asm/preempt.h +++ b/arch/s390/include/asm/preempt.h @@ -29,6 +29,9 @@ static inline void preempt_count_set(int pc) old, new) != old); } + + + static inline void set_preempt_need_resched(void) { __atomic_and(~PREEMPT_NEED_RESCHED, &S390_lowcore.preempt_count); @@ -82,6 +85,9 @@ static inline void preempt_count_set(int pc) S390_lowcore.preempt_count = pc; } + + + static inline void set_preempt_need_resched(void) { } @@ -129,4 +135,4 @@ extern void preempt_schedule_notrace(void); #define __preempt_schedule_notrace() preempt_schedule_notrace() #endif /* CONFIG_PREEMPTION */ -#endif /* __ASM_PREEMPT_H */ +#endif /* __ASM_PREEMPT_H */ \ No newline at end of file diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig index 0045e1b4419..e42751ed469 100644 --- a/arch/x86/Kconfig +++ b/arch/x86/Kconfig @@ -170,6 +170,7 @@ config X86 select HAVE_ARCH_TRACEHOOK select HAVE_ARCH_TRANSPARENT_HUGEPAGE select HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD if X86_64 + select HAVE_ARCH_PARENT_PMD_YOUNG if X86_64 select HAVE_ARCH_USERFAULTFD_WP if X86_64 && USERFAULTFD select HAVE_ARCH_USERFAULTFD_MINOR if X86_64 && USERFAULTFD select HAVE_ARCH_VMAP_STACK if X86_64 @@ -1010,6 +1011,22 @@ config NR_CPUS config SCHED_SMT def_bool y if SMP +config SMT_NICE + bool "SMT (Hyperthreading) aware nice priority and policy support" + depends on SCHED_MUQSS && SCHED_SMT + default y + help + Enabling Hyperthreading on Intel CPUs decreases the effectiveness + of the use of 'nice' levels and different scheduling policies + (e.g. realtime) due to sharing of CPU power between hyperthreads. + SMT nice support makes each logical CPU aware of what is running on + its hyperthread siblings, maintaining appropriate distribution of + CPU according to nice levels and scheduling policies at the expense + of slightly increased overhead. + + If unsure say Y here. + + config SCHED_MC def_bool y prompt "Multi-core scheduler support" @@ -1040,6 +1057,8 @@ config SCHED_MC_PRIO If unsure say Y here. +source "kernel/Kconfig.MuQSS" + config UP_LATE_INIT def_bool y depends on !SMP && X86_LOCAL_APIC diff --git a/arch/x86/Kconfig.cpu b/arch/x86/Kconfig.cpu index 814fe0d349b..d53775ceb36 100644 --- a/arch/x86/Kconfig.cpu +++ b/arch/x86/Kconfig.cpu @@ -157,7 +157,7 @@ config MPENTIUM4 config MK6 - bool "K6/K6-II/K6-III" + bool "AMD K6/K6-II/K6-III" depends on X86_32 help Select this for an AMD K6-family processor. Enables use of @@ -165,7 +165,7 @@ config MK6 flags to GCC. config MK7 - bool "Athlon/Duron/K7" + bool "AMD Athlon/Duron/K7" depends on X86_32 help Select this for an AMD Athlon K7-family processor. Enables use of @@ -173,12 +173,98 @@ config MK7 flags to GCC. config MK8 - bool "Opteron/Athlon64/Hammer/K8" + bool "AMD Opteron/Athlon64/Hammer/K8" help Select this for an AMD Opteron or Athlon64 Hammer-family processor. Enables use of some extended instructions, and passes appropriate optimization flags to GCC. +config MK8SSE3 + bool "AMD Opteron/Athlon64/Hammer/K8 with SSE3" + help + Select this for improved AMD Opteron or Athlon64 Hammer-family processors. + Enables use of some extended instructions, and passes appropriate + optimization flags to GCC. + +config MK10 + bool "AMD 61xx/7x50/PhenomX3/X4/II/K10" + help + Select this for an AMD 61xx Eight-Core Magny-Cours, Athlon X2 7x50, + Phenom X3/X4/II, Athlon II X2/X3/X4, or Turion II-family processor. + Enables use of some extended instructions, and passes appropriate + optimization flags to GCC. + +config MBARCELONA + bool "AMD Barcelona" + help + Select this for AMD Family 10h Barcelona processors. + + Enables -march=barcelona + +config MBOBCAT + bool "AMD Bobcat" + help + Select this for AMD Family 14h Bobcat processors. + + Enables -march=btver1 + +config MJAGUAR + bool "AMD Jaguar" + help + Select this for AMD Family 16h Jaguar processors. + + Enables -march=btver2 + +config MBULLDOZER + bool "AMD Bulldozer" + help + Select this for AMD Family 15h Bulldozer processors. + + Enables -march=bdver1 + +config MPILEDRIVER + bool "AMD Piledriver" + help + Select this for AMD Family 15h Piledriver processors. + + Enables -march=bdver2 + +config MSTEAMROLLER + bool "AMD Steamroller" + help + Select this for AMD Family 15h Steamroller processors. + + Enables -march=bdver3 + +config MEXCAVATOR + bool "AMD Excavator" + help + Select this for AMD Family 15h Excavator processors. + + Enables -march=bdver4 + +config MPCK + bool "AMD Zen" + help + Select this for AMD Family 17h Zen processors. + + Enables -march=znver1 + +config MPCK2 + bool "AMD Zen 2" + help + Select this for AMD Family 17h Zen 2 processors. + + Enables -march=znver2 + +config MPCK3 + bool "AMD Zen 3" + depends on GCC_VERSION > 100300 + help + Select this for AMD Family 19h Zen 3 processors. + + Enables -march=znver3 + config MCRUSOE bool "Crusoe" depends on X86_32 @@ -270,7 +356,7 @@ config MPSC in /proc/cpuinfo. Family 15 is an older Xeon, Family 6 a newer one. config MCORE2 - bool "Core 2/newer Xeon" + bool "Intel Core 2" help Select this for Intel Core 2 and newer Core 2 Xeons (Xeon 51xx and @@ -278,6 +364,8 @@ config MCORE2 family in /proc/cpuinfo. Newer ones have 6 and older ones 15 (not a typo) + Enables -march=core2 + config MATOM bool "Intel Atom" help @@ -287,6 +375,182 @@ config MATOM accordingly optimized code. Use a recent GCC with specific Atom support in order to fully benefit from selecting this option. +config MNEHALEM + bool "Intel Nehalem" + select X86_P6_NOP + help + + Select this for 1st Gen Core processors in the Nehalem family. + + Enables -march=nehalem + +config MWESTMERE + bool "Intel Westmere" + select X86_P6_NOP + help + + Select this for the Intel Westmere formerly Nehalem-C family. + + Enables -march=westmere + +config MSILVERMONT + bool "Intel Silvermont" + select X86_P6_NOP + help + + Select this for the Intel Silvermont platform. + + Enables -march=silvermont + +config MGOLDMONT + bool "Intel Goldmont" + select X86_P6_NOP + help + + Select this for the Intel Goldmont platform including Apollo Lake and Denverton. + + Enables -march=goldmont + +config MGOLDMONTPLUS + bool "Intel Goldmont Plus" + select X86_P6_NOP + help + + Select this for the Intel Goldmont Plus platform including Gemini Lake. + + Enables -march=goldmont-plus + +config MSANDYBRIDGE + bool "Intel Sandy Bridge" + select X86_P6_NOP + help + + Select this for 2nd Gen Core processors in the Sandy Bridge family. + + Enables -march=sandybridge + +config MIVYBRIDGE + bool "Intel Ivy Bridge" + select X86_P6_NOP + help + + Select this for 3rd Gen Core processors in the Ivy Bridge family. + + Enables -march=ivybridge + +config MHASWELL + bool "Intel Haswell" + select X86_P6_NOP + help + + Select this for 4th Gen Core processors in the Haswell family. + + Enables -march=haswell + +config MBROADWELL + bool "Intel Broadwell" + select X86_P6_NOP + help + + Select this for 5th Gen Core processors in the Broadwell family. + + Enables -march=broadwell + +config MSKYLAKE + bool "Intel Skylake" + select X86_P6_NOP + help + + Select this for 6th Gen Core processors in the Skylake family. + + Enables -march=skylake + +config MSKYLAKEX + bool "Intel Skylake X" + select X86_P6_NOP + help + + Select this for 6th Gen Core processors in the Skylake X family. + + Enables -march=skylake-avx512 + +config MCANNONLAKE + bool "Intel Cannon Lake" + select X86_P6_NOP + help + + Select this for 8th Gen Core processors + + Enables -march=cannonlake + +config MICELAKE + bool "Intel Ice Lake" + select X86_P6_NOP + help + + Select this for 10th Gen Core processors in the Ice Lake family. + + Enables -march=icelake-client + +config MCASCADELAKE + bool "Intel Cascade Lake" + select X86_P6_NOP + help + + Select this for Xeon processors in the Cascade Lake family. + + Enables -march=cascadelake + +config MCOOPERLAKE + bool "Intel Cooper Lake" + depends on GCC_VERSION > 100100 + select X86_P6_NOP + help + + Select this for Xeon processors in the Cooper Lake family. + + Enables -march=cooperlake + +config MTIGERLAKE + bool "Intel Tiger Lake" + depends on GCC_VERSION > 100100 + select X86_P6_NOP + help + + Select this for third-generation 10 nm process processors in the Tiger Lake family. + + Enables -march=tigerlake + +config MSAPPHIRERAPIDS + bool "Intel Sapphire Rapids" + depends on GCC_VERSION > 110000 + select X86_P6_NOP + help + + Select this for third-generation 10 nm process processors in the Sapphire Rapids family. + + Enables -march=sapphirerapids + +config MROCKETLAKE + bool "Intel Rocket Lake" + depends on GCC_VERSION > 110000 + select X86_P6_NOP + help + + Select this for eleventh-generation processors in the Rocket Lake family. + + Enables -march=rocketlake + +config MALDERLAKE + bool "Intel Alder Lake" + depends on GCC_VERSION > 110000 + select X86_P6_NOP + help + + Select this for twelfth-generation processors in the Alder Lake family. + + Enables -march=alderlake + config GENERIC_CPU bool "Generic-x86-64" depends on X86_64 @@ -294,6 +558,50 @@ config GENERIC_CPU Generic x86-64 CPU. Run equally well on all x86-64 CPUs. +config GENERIC_CPU2 + bool "Generic-x86-64-v2" + depends on GCC_VERSION > 110000 + depends on X86_64 + help + Generic x86-64 CPU. + Run equally well on all x86-64 CPUs with min support of x86-64-v2. + +config GENERIC_CPU3 + bool "Generic-x86-64-v3" + depends on GCC_VERSION > 110000 + depends on X86_64 + help + Generic x86-64-v3 CPU with v3 instructions. + Run equally well on all x86-64 CPUs with min support of x86-64-v3. + +config GENERIC_CPU4 + bool "Generic-x86-64-v4" + depends on GCC_VERSION > 110000 + depends on X86_64 + help + Generic x86-64 CPU with v4 instructions. + Run equally well on all x86-64 CPUs with min support of x86-64-v4. + +config MNATIVE_INTEL + bool "Intel-Native optimizations autodetected by GCC" + help + + GCC 4.2 and above support -march=native, which automatically detects + the optimum settings to use based on your processor. Do NOT use this + for AMD CPUs. Intel Only! + + Enables -march=native + +config MNATIVE_AMD + bool "AMD-Native optimizations autodetected by GCC" + help + + GCC 4.2 and above support -march=native, which automatically detects + the optimum settings to use based on your processor. Do NOT use this + for Intel CPUs. AMD Only! + + Enables -march=native + endchoice config X86_GENERIC @@ -318,7 +626,7 @@ config X86_INTERNODE_CACHE_SHIFT config X86_L1_CACHE_SHIFT int default "7" if MPENTIUM4 || MPSC - default "6" if MK7 || MK8 || MPENTIUMM || MCORE2 || MATOM || MVIAC7 || X86_GENERIC || GENERIC_CPU + default "6" if MK7 || MK8 || MPENTIUMM || MCORE2 || MATOM || MVIAC7 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD || X86_GENERIC || GENERIC_CPU || GENERIC_CPU2 || GENERIC_CPU3 || GENERIC_CPU4 default "4" if MELAN || M486SX || M486 || MGEODEGX1 default "5" if MWINCHIP3D || MWINCHIPC6 || MCRUSOE || MEFFICEON || MCYRIXIII || MK6 || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || M586 || MVIAC3_2 || MGEODE_LX @@ -336,11 +644,11 @@ config X86_ALIGNMENT_16 config X86_INTEL_USERCOPY def_bool y - depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK7 || MEFFICEON || MCORE2 + depends on MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M586MMX || X86_GENERIC || MK8 || MK7 || MEFFICEON || MCORE2 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL config X86_USE_PPRO_CHECKSUM def_bool y - depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MATOM + depends on MWINCHIP3D || MWINCHIPC6 || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MK8 || MVIAC3_2 || MVIAC7 || MEFFICEON || MGEODE_LX || MCORE2 || MATOM || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD config X86_USE_3DNOW def_bool y @@ -360,26 +668,26 @@ config X86_USE_3DNOW config X86_P6_NOP def_bool y depends on X86_64 - depends on (MCORE2 || MPENTIUM4 || MPSC) + depends on (MCORE2 || MPENTIUM4 || MPSC || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL) config X86_TSC def_bool y - depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MATOM) || X86_64 + depends on (MWINCHIP3D || MCRUSOE || MEFFICEON || MCYRIXIII || MK7 || MK6 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586MMX || M586TSC || MK8 || MVIAC3_2 || MVIAC7 || MGEODEGX1 || MGEODE_LX || MCORE2 || MATOM || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD) || X86_64 config X86_CMPXCHG64 def_bool y - depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586TSC || M586MMX || MATOM || MGEODE_LX || MGEODEGX1 || MK6 || MK7 || MK8 + depends on X86_PAE || X86_64 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || M586TSC || M586MMX || MATOM || MGEODE_LX || MGEODEGX1 || MK6 || MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD # this should be set for all -march=.. options where the compiler # generates cmov. config X86_CMOV def_bool y - depends on (MK8 || MK7 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MATOM || MGEODE_LX) + depends on (MK8 || MK7 || MCORE2 || MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MCRUSOE || MEFFICEON || X86_64 || MATOM || MGEODE_LX || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD) config X86_MINIMUM_CPU_FAMILY int default "64" if X86_64 - default "6" if X86_32 && (MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MEFFICEON || MATOM || MCRUSOE || MCORE2 || MK7 || MK8) + default "6" if X86_32 && (MPENTIUM4 || MPENTIUMM || MPENTIUMIII || MPENTIUMII || M686 || MVIAC3_2 || MVIAC7 || MEFFICEON || MATOM || MCRUSOE || MCORE2 || MK7 || MK8 || MK8SSE3 || MK10 || MBARCELONA || MBOBCAT || MJAGUAR || MBULLDOZER || MPILEDRIVER || MSTEAMROLLER || MEXCAVATOR || MPCK || MZEN2 || MZEN3 || MNEHALEM || MWESTMERE || MSILVERMONT || MGOLDMONT || MGOLDMONTPLUS || MSANDYBRIDGE || MIVYBRIDGE || MHASWELL || MBROADWELL || MSKYLAKE || MSKYLAKEX || MCANNONLAKE || MICELAKE || MCASCADELAKE || MCOOPERLAKE || MTIGERLAKE || MSAPPHIRERAPIDS || MROCKETLAKE || MALDERLAKE || MNATIVE_INTEL || MNATIVE_AMD) default "5" if X86_32 && X86_CMPXCHG64 default "4" diff --git a/arch/x86/Makefile b/arch/x86/Makefile index cb5e8d39cac..5a9e712a8b0 100644 --- a/arch/x86/Makefile +++ b/arch/x86/Makefile @@ -122,11 +122,48 @@ else # FIXME - should be integrated in Makefile.cpu (Makefile_32.cpu) cflags-$(CONFIG_MK8) += $(call cc-option,-march=k8) cflags-$(CONFIG_MPSC) += $(call cc-option,-march=nocona) - - cflags-$(CONFIG_MCORE2) += \ - $(call cc-option,-march=core2,$(call cc-option,-mtune=generic)) - cflags-$(CONFIG_MATOM) += $(call cc-option,-march=atom) \ - $(call cc-option,-mtune=atom,$(call cc-option,-mtune=generic)) + cflags-$(CONFIG_MK8SSE3) += $(call cc-option,-march=k8-sse3) + cflags-$(CONFIG_MK10) += $(call cc-option,-march=amdfam10) + cflags-$(CONFIG_MBARCELONA) += $(call cc-option,-march=barcelona) + cflags-$(CONFIG_MBOBCAT) += $(call cc-option,-march=btver1) + cflags-$(CONFIG_MJAGUAR) += $(call cc-option,-march=btver2) + cflags-$(CONFIG_MBULLDOZER) += $(call cc-option,-march=bdver1) + cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-march=bdver2) + cflags-$(CONFIG_MPILEDRIVER) += $(call cc-option,-mno-tbm) + cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-march=bdver3) + cflags-$(CONFIG_MSTEAMROLLER) += $(call cc-option,-mno-tbm) + cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-march=bdver4) + cflags-$(CONFIG_MEXCAVATOR) += $(call cc-option,-mno-tbm) + cflags-$(CONFIG_MPCK) += $(call cc-option,-march=znver1) + cflags-$(CONFIG_MPCK2) += $(call cc-option,-march=znver2) + cflags-$(CONFIG_MPCK3) += $(call cc-option,-march=znver3) + + cflags-$(CONFIG_MNATIVE_INTEL) += $(call cc-option,-march=native) + cflags-$(CONFIG_MNATIVE_AMD) += $(call cc-option,-march=native) + cflags-$(CONFIG_MATOM) += $(call cc-option,-march=bonnell) + cflags-$(CONFIG_MCORE2) += $(call cc-option,-march=core2) + cflags-$(CONFIG_MNEHALEM) += $(call cc-option,-march=nehalem) + cflags-$(CONFIG_MWESTMERE) += $(call cc-option,-march=westmere) + cflags-$(CONFIG_MSILVERMONT) += $(call cc-option,-march=silvermont) + cflags-$(CONFIG_MGOLDMONT) += $(call cc-option,-march=goldmont) + cflags-$(CONFIG_MGOLDMONTPLUS) += $(call cc-option,-march=goldmont-plus) + cflags-$(CONFIG_MSANDYBRIDGE) += $(call cc-option,-march=sandybridge) + cflags-$(CONFIG_MIVYBRIDGE) += $(call cc-option,-march=ivybridge) + cflags-$(CONFIG_MHASWELL) += $(call cc-option,-march=haswell) + cflags-$(CONFIG_MBROADWELL) += $(call cc-option,-march=broadwell) + cflags-$(CONFIG_MSKYLAKE) += $(call cc-option,-march=skylake) + cflags-$(CONFIG_MSKYLAKEX) += $(call cc-option,-march=skylake-avx512) + cflags-$(CONFIG_MCANNONLAKE) += $(call cc-option,-march=cannonlake) + cflags-$(CONFIG_MICELAKE) += $(call cc-option,-march=icelake-client) + cflags-$(CONFIG_MCASCADELAKE) += $(call cc-option,-march=cascadelake) + cflags-$(CONFIG_MCOOPERLAKE) += $(call cc-option,-march=cooperlake) + cflags-$(CONFIG_MTIGERLAKE) += $(call cc-option,-march=tigerlake) + cflags-$(CONFIG_MSAPPHIRERAPIDS) += $(call cc-option,-march=sapphirerapids) + cflags-$(CONFIG_MROCKETLAKE) += $(call cc-option,-march=rocketlake) + cflags-$(CONFIG_MALDERLAKE) += $(call cc-option,-march=alderlake) + cflags-$(CONFIG_GENERIC_CPU2) += $(call cc-option,-march=x86-64-v2) + cflags-$(CONFIG_GENERIC_CPU3) += $(call cc-option,-march=x86-64-v3) + cflags-$(CONFIG_GENERIC_CPU4) += $(call cc-option,-march=x86-64-v4) cflags-$(CONFIG_GENERIC_CPU) += $(call cc-option,-mtune=generic) KBUILD_CFLAGS += $(cflags-y) diff --git a/arch/x86/entry/syscalls/syscall_32.tbl b/arch/x86/entry/syscalls/syscall_32.tbl index 4bbc267fb36..1c3ca8b5024 100644 --- a/arch/x86/entry/syscalls/syscall_32.tbl +++ b/arch/x86/entry/syscalls/syscall_32.tbl @@ -451,3 +451,7 @@ 444 i386 landlock_create_ruleset sys_landlock_create_ruleset 445 i386 landlock_add_rule sys_landlock_add_rule 446 i386 landlock_restrict_self sys_landlock_restrict_self +447 i386 futex_wait sys_futex_wait +448 i386 futex_wake sys_futex_wake +449 i386 futex_waitv sys_futex_waitv compat_sys_futex_waitv +450 i386 futex_requeue sys_futex_requeue compat_sys_futex_requeue diff --git a/arch/x86/entry/syscalls/syscall_64.tbl b/arch/x86/entry/syscalls/syscall_64.tbl index ce18119ea0d..a1a39ed156e 100644 --- a/arch/x86/entry/syscalls/syscall_64.tbl +++ b/arch/x86/entry/syscalls/syscall_64.tbl @@ -368,6 +368,10 @@ 444 common landlock_create_ruleset sys_landlock_create_ruleset 445 common landlock_add_rule sys_landlock_add_rule 446 common landlock_restrict_self sys_landlock_restrict_self +447 common futex_wait sys_futex_wait +448 common futex_wake sys_futex_wake +449 common futex_waitv sys_futex_waitv +450 common futex_requeue sys_futex_requeue # # Due to a historical design error, certain syscalls are numbered differently diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h index d2c76c8d8cf..47c3b333f2c 100644 --- a/arch/x86/include/asm/pci.h +++ b/arch/x86/include/asm/pci.h @@ -27,6 +27,7 @@ struct pci_sysdata { #if IS_ENABLED(CONFIG_VMD) struct pci_dev *vmd_dev; /* VMD Device if in Intel VMD domain */ #endif + struct pci_dev *nvme_remap_dev; /* AHCI Device if NVME remapped bus */ }; extern int pci_routeirq; @@ -70,6 +71,11 @@ static inline bool is_vmd(struct pci_bus *bus) #define is_vmd(bus) false #endif /* CONFIG_VMD */ +static inline bool is_nvme_remap(struct pci_bus *bus) +{ + return to_pci_sysdata(bus)->nvme_remap_dev != NULL; +} + /* Can be used to override the logic in pci_scan_bus for skipping already-configured bus numbers - to be used for buggy BIOSes or architectures with incomplete PCI setup by the loader */ diff --git a/arch/x86/include/asm/pgtable.h b/arch/x86/include/asm/pgtable.h index b1099f2d980..3a24d2af4e9 100644 --- a/arch/x86/include/asm/pgtable.h +++ b/arch/x86/include/asm/pgtable.h @@ -846,7 +846,7 @@ static inline unsigned long pmd_page_vaddr(pmd_t pmd) static inline int pmd_bad(pmd_t pmd) { - return (pmd_flags(pmd) & ~_PAGE_USER) != _KERNPG_TABLE; + return ((pmd_flags(pmd) | _PAGE_ACCESSED) & ~_PAGE_USER) != _KERNPG_TABLE; } static inline unsigned long pages_to_mb(unsigned long npg) diff --git a/arch/x86/include/asm/vermagic.h b/arch/x86/include/asm/vermagic.h index 75884d2cdec..176f9c298f3 100644 --- a/arch/x86/include/asm/vermagic.h +++ b/arch/x86/include/asm/vermagic.h @@ -17,6 +17,48 @@ #define MODULE_PROC_FAMILY "586MMX " #elif defined CONFIG_MCORE2 #define MODULE_PROC_FAMILY "CORE2 " +#elif defined CONFIG_MNATIVE_INTEL +#define MODULE_PROC_FAMILY "NATIVE_INTEL " +#elif defined CONFIG_MNATIVE_AMD +#define MODULE_PROC_FAMILY "NATIVE_AMD " +#elif defined CONFIG_MNEHALEM +#define MODULE_PROC_FAMILY "NEHALEM " +#elif defined CONFIG_MWESTMERE +#define MODULE_PROC_FAMILY "WESTMERE " +#elif defined CONFIG_MSILVERMONT +#define MODULE_PROC_FAMILY "SILVERMONT " +#elif defined CONFIG_MGOLDMONT +#define MODULE_PROC_FAMILY "GOLDMONT " +#elif defined CONFIG_MGOLDMONTPLUS +#define MODULE_PROC_FAMILY "GOLDMONTPLUS " +#elif defined CONFIG_MSANDYBRIDGE +#define MODULE_PROC_FAMILY "SANDYBRIDGE " +#elif defined CONFIG_MIVYBRIDGE +#define MODULE_PROC_FAMILY "IVYBRIDGE " +#elif defined CONFIG_MHASWELL +#define MODULE_PROC_FAMILY "HASWELL " +#elif defined CONFIG_MBROADWELL +#define MODULE_PROC_FAMILY "BROADWELL " +#elif defined CONFIG_MSKYLAKE +#define MODULE_PROC_FAMILY "SKYLAKE " +#elif defined CONFIG_MSKYLAKEX +#define MODULE_PROC_FAMILY "SKYLAKEX " +#elif defined CONFIG_MCANNONLAKE +#define MODULE_PROC_FAMILY "CANNONLAKE " +#elif defined CONFIG_MICELAKE +#define MODULE_PROC_FAMILY "ICELAKE " +#elif defined CONFIG_MCASCADELAKE +#define MODULE_PROC_FAMILY "CASCADELAKE " +#elif defined CONFIG_MCOOPERLAKE +#define MODULE_PROC_FAMILY "COOPERLAKE " +#elif defined CONFIG_MTIGERLAKE +#define MODULE_PROC_FAMILY "TIGERLAKE " +#elif defined CONFIG_MSAPPHIRERAPIDS +#define MODULE_PROC_FAMILY "SAPPHIRERAPIDS " +#elif defined CONFIG_ROCKETLAKE +#define MODULE_PROC_FAMILY "ROCKETLAKE " +#elif defined CONFIG_MALDERLAKE +#define MODULE_PROC_FAMILY "ALDERLAKE " #elif defined CONFIG_MATOM #define MODULE_PROC_FAMILY "ATOM " #elif defined CONFIG_M686 @@ -35,6 +77,30 @@ #define MODULE_PROC_FAMILY "K7 " #elif defined CONFIG_MK8 #define MODULE_PROC_FAMILY "K8 " +#elif defined CONFIG_MK8SSE3 +#define MODULE_PROC_FAMILY "K8SSE3 " +#elif defined CONFIG_MK10 +#define MODULE_PROC_FAMILY "K10 " +#elif defined CONFIG_MBARCELONA +#define MODULE_PROC_FAMILY "BARCELONA " +#elif defined CONFIG_MBOBCAT +#define MODULE_PROC_FAMILY "BOBCAT " +#elif defined CONFIG_MBULLDOZER +#define MODULE_PROC_FAMILY "BULLDOZER " +#elif defined CONFIG_MPILEDRIVER +#define MODULE_PROC_FAMILY "PILEDRIVER " +#elif defined CONFIG_MSTEAMROLLER +#define MODULE_PROC_FAMILY "STEAMROLLER " +#elif defined CONFIG_MJAGUAR +#define MODULE_PROC_FAMILY "JAGUAR " +#elif defined CONFIG_MEXCAVATOR +#define MODULE_PROC_FAMILY "EXCAVATOR " +#elif defined CONFIG_MPCK +#define MODULE_PROC_FAMILY "PCK " +#elif defined CONFIG_MPCK2 +#define MODULE_PROC_FAMILY "PCK2 " +#elif defined CONFIG_MPCK3 +#define MODULE_PROC_FAMILY "PCK3 " #elif defined CONFIG_MELAN #define MODULE_PROC_FAMILY "ELAN " #elif defined CONFIG_MCRUSOE diff --git a/arch/x86/mm/pgtable.c b/arch/x86/mm/pgtable.c index d27cf69e811..b968d6bd28b 100644 --- a/arch/x86/mm/pgtable.c +++ b/arch/x86/mm/pgtable.c @@ -550,7 +550,7 @@ int ptep_test_and_clear_young(struct vm_area_struct *vma, return ret; } -#ifdef CONFIG_TRANSPARENT_HUGEPAGE +#if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG) int pmdp_test_and_clear_young(struct vm_area_struct *vma, unsigned long addr, pmd_t *pmdp) { @@ -562,6 +562,9 @@ int pmdp_test_and_clear_young(struct vm_area_struct *vma, return ret; } +#endif + +#ifdef CONFIG_TRANSPARENT_HUGEPAGE int pudp_test_and_clear_young(struct vm_area_struct *vma, unsigned long addr, pud_t *pudp) { diff --git a/arch/x86/pci/common.c b/arch/x86/pci/common.c index 3507f456fcd..732b4cf516b 100644 --- a/arch/x86/pci/common.c +++ b/arch/x86/pci/common.c @@ -715,12 +715,15 @@ int pci_ext_cfg_avail(void) return 0; } -#if IS_ENABLED(CONFIG_VMD) struct pci_dev *pci_real_dma_dev(struct pci_dev *dev) { +#if IS_ENABLED(CONFIG_VMD) if (is_vmd(dev->bus)) return to_pci_sysdata(dev->bus)->vmd_dev; +#endif + + if (is_nvme_remap(dev->bus)) + return to_pci_sysdata(dev->bus)->nvme_remap_dev; return dev; } -#endif diff --git a/block/Kconfig.iosched b/block/Kconfig.iosched index 2f2158e05a9..92a47e48fb3 100644 --- a/block/Kconfig.iosched +++ b/block/Kconfig.iosched @@ -9,6 +9,15 @@ config MQ_IOSCHED_DEADLINE help MQ version of the deadline IO scheduler. +config MQ_IOSCHED_DEADLINE_NODEFAULT + bool "Rename mq-deadline scheduler to mq-deadline-nodefault" + depends on MQ_IOSCHED_DEADLINE + default n + help + This renames the mq-deadline scheduler to "mq-deadline-nodefault" and + also drops its alias of "deadline". This can prevent existing + userspace from forcing this scheduler over the kernel's choice. + config MQ_IOSCHED_KYBER tristate "Kyber I/O scheduler" default y diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c index eccbe2aed7c..4df33cc08ee 100644 --- a/block/bfq-iosched.c +++ b/block/bfq-iosched.c @@ -2333,6 +2333,9 @@ static int bfq_request_merge(struct request_queue *q, struct request **req, __rq = bfq_find_rq_fmerge(bfqd, bio, q); if (__rq && elv_bio_merge_ok(__rq, bio)) { *req = __rq; + + if (blk_discard_mergable(__rq)) + return ELEVATOR_DISCARD_MERGE; return ELEVATOR_FRONT_MERGE; } diff --git a/block/blk-merge.c b/block/blk-merge.c index bcdff1879c3..7abbea25ae0 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -705,22 +705,6 @@ static void blk_account_io_merge_request(struct request *req) } } -/* - * Two cases of handling DISCARD merge: - * If max_discard_segments > 1, the driver takes every bio - * as a range and send them to controller together. The ranges - * needn't to be contiguous. - * Otherwise, the bios/requests will be handled as same as - * others which should be contiguous. - */ -static inline bool blk_discard_mergable(struct request *req) -{ - if (req_op(req) == REQ_OP_DISCARD && - queue_max_discard_segments(req->q) > 1) - return true; - return false; -} - static enum elv_merge blk_try_req_merge(struct request *req, struct request *next) { diff --git a/block/elevator.c b/block/elevator.c index 440699c2811..b23a9fbd996 100644 --- a/block/elevator.c +++ b/block/elevator.c @@ -336,6 +336,9 @@ enum elv_merge elv_merge(struct request_queue *q, struct request **req, __rq = elv_rqhash_find(q, bio->bi_iter.bi_sector); if (__rq && elv_bio_merge_ok(__rq, bio)) { *req = __rq; + + if (blk_discard_mergable(__rq)) + return ELEVATOR_DISCARD_MERGE; return ELEVATOR_BACK_MERGE; } @@ -621,11 +624,19 @@ static inline bool elv_support_iosched(struct request_queue *q) */ static struct elevator_type *elevator_get_default(struct request_queue *q) { +#ifndef CONFIG_PCK_INTERACTIVE if (q->nr_hw_queues != 1 && !blk_mq_is_sbitmap_shared(q->tag_set->flags)) return NULL; +#endif +#if defined(CONFIG_PCK_INTERACTIVE) && defined(CONFIG_IOSCHED_BFQ) + return elevator_get(q, "bfq", false); +#elif defined(CONFIG_MQ_IOSCHED_DEADLINE_NODEFAULT) + return elevator_get(q, "mq-deadline-nodefault", false); +#else return elevator_get(q, "mq-deadline", false); +#endif } /* diff --git a/block/mq-deadline.c b/block/mq-deadline.c index 8eea2cbf2bf..7ca79892bb7 100644 --- a/block/mq-deadline.c +++ b/block/mq-deadline.c @@ -454,6 +454,8 @@ static int dd_request_merge(struct request_queue *q, struct request **rq, if (elv_bio_merge_ok(__rq, bio)) { *rq = __rq; + if (blk_discard_mergable(__rq)) + return ELEVATOR_DISCARD_MERGE; return ELEVATOR_FRONT_MERGE; } } @@ -790,12 +792,21 @@ static struct elevator_type mq_deadline = { .queue_debugfs_attrs = deadline_queue_debugfs_attrs, #endif .elevator_attrs = deadline_attrs, +#ifdef CONFIG_MQ_IOSCHED_DEADLINE_NODEFAULT + .elevator_name = "mq-deadline-nodefault", + .elevator_alias = "deadline-nodefault", +#else .elevator_name = "mq-deadline", .elevator_alias = "deadline", +#endif .elevator_features = ELEVATOR_F_ZBD_SEQ_WRITE, .elevator_owner = THIS_MODULE, }; +#ifdef CONFIG_MQ_IOSCHED_DEADLINE_NODEFAULT +MODULE_ALIAS("mq-deadline-nodefault-iosched"); +#else MODULE_ALIAS("mq-deadline-iosched"); +#endif static int __init deadline_init(void) { diff --git a/drivers/Makefile b/drivers/Makefile index 5a6d613e868..0e1a7a5d308 100644 --- a/drivers/Makefile +++ b/drivers/Makefile @@ -60,15 +60,8 @@ obj-y += char/ # iommu/ comes before gpu as gpu are using iommu controllers obj-y += iommu/ -# gpu/ comes after char for AGP vs DRM startup and after iommu -obj-y += gpu/ - obj-$(CONFIG_CONNECTOR) += connector/ -# i810fb and intelfb depend on char/agp/ -obj-$(CONFIG_FB_I810) += video/fbdev/i810/ -obj-$(CONFIG_FB_INTEL) += video/fbdev/intelfb/ - obj-$(CONFIG_PARPORT) += parport/ obj-$(CONFIG_NVM) += lightnvm/ obj-y += base/ block/ misc/ mfd/ nfc/ @@ -82,6 +75,14 @@ obj-$(CONFIG_IDE) += ide/ obj-y += scsi/ obj-y += nvme/ obj-$(CONFIG_ATA) += ata/ + +# gpu/ comes after char for AGP vs DRM startup and after iommu +obj-y += gpu/ + +# i810fb and intelfb depend on char/agp/ +obj-$(CONFIG_FB_I810) += video/fbdev/i810/ +obj-$(CONFIG_FB_INTEL) += video/fbdev/intelfb/ + obj-$(CONFIG_TARGET_CORE) += target/ obj-$(CONFIG_MTD) += mtd/ obj-$(CONFIG_SPI) += spi/ diff --git a/drivers/accessibility/speakup/speakup_acntpc.c b/drivers/accessibility/speakup/speakup_acntpc.c index c1ec087dca1..b2d0d4266f6 100644 --- a/drivers/accessibility/speakup/speakup_acntpc.c +++ b/drivers/accessibility/speakup/speakup_acntpc.c @@ -198,7 +198,7 @@ static void do_catch_up(struct spk_synth *synth) full_time_val = full_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); if (synth_full()) { - schedule_timeout(msecs_to_jiffies(full_time_val)); + schedule_msec_hrtimeout((full_time_val)); continue; } set_current_state(TASK_RUNNING); @@ -226,7 +226,7 @@ static void do_catch_up(struct spk_synth *synth) jiffy_delta_val = jiffy_delta->u.n.value; delay_time_val = delay_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout(delay_time_val); jiff_max = jiffies + jiffy_delta_val; } } diff --git a/drivers/accessibility/speakup/speakup_apollo.c b/drivers/accessibility/speakup/speakup_apollo.c index cd63581b2e9..d636157a284 100644 --- a/drivers/accessibility/speakup/speakup_apollo.c +++ b/drivers/accessibility/speakup/speakup_apollo.c @@ -165,7 +165,7 @@ static void do_catch_up(struct spk_synth *synth) if (!synth->io_ops->synth_out(synth, ch)) { synth->io_ops->tiocmset(synth, 0, UART_MCR_RTS); synth->io_ops->tiocmset(synth, UART_MCR_RTS, 0); - schedule_timeout(msecs_to_jiffies(full_time_val)); + schedule_msec_hrtimeout(full_time_val); continue; } if (time_after_eq(jiffies, jiff_max) && (ch == SPACE)) { diff --git a/drivers/accessibility/speakup/speakup_decext.c b/drivers/accessibility/speakup/speakup_decext.c index 092cfd08a9e..e7fc85f8ce5 100644 --- a/drivers/accessibility/speakup/speakup_decext.c +++ b/drivers/accessibility/speakup/speakup_decext.c @@ -180,7 +180,7 @@ static void do_catch_up(struct spk_synth *synth) if (ch == '\n') ch = 0x0D; if (synth_full() || !synth->io_ops->synth_out(synth, ch)) { - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout(delay_time_val); continue; } set_current_state(TASK_RUNNING); diff --git a/drivers/accessibility/speakup/speakup_decpc.c b/drivers/accessibility/speakup/speakup_decpc.c index dec314dee21..2a5deb5256b 100644 --- a/drivers/accessibility/speakup/speakup_decpc.c +++ b/drivers/accessibility/speakup/speakup_decpc.c @@ -398,7 +398,7 @@ static void do_catch_up(struct spk_synth *synth) if (ch == '\n') ch = 0x0D; if (dt_sendchar(ch)) { - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout((delay_time_val)); continue; } set_current_state(TASK_RUNNING); diff --git a/drivers/accessibility/speakup/speakup_dectlk.c b/drivers/accessibility/speakup/speakup_dectlk.c index 580ec796816..67c156b90dd 100644 --- a/drivers/accessibility/speakup/speakup_dectlk.c +++ b/drivers/accessibility/speakup/speakup_dectlk.c @@ -256,7 +256,7 @@ static void do_catch_up(struct spk_synth *synth) if (ch == '\n') ch = 0x0D; if (synth_full_val || !synth->io_ops->synth_out(synth, ch)) { - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout(delay_time_val); continue; } set_current_state(TASK_RUNNING); diff --git a/drivers/accessibility/speakup/speakup_dtlk.c b/drivers/accessibility/speakup/speakup_dtlk.c index 92838d3ae9e..b687cb4d326 100644 --- a/drivers/accessibility/speakup/speakup_dtlk.c +++ b/drivers/accessibility/speakup/speakup_dtlk.c @@ -211,7 +211,7 @@ static void do_catch_up(struct spk_synth *synth) delay_time_val = delay_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); if (synth_full()) { - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout((delay_time_val)); continue; } set_current_state(TASK_RUNNING); @@ -227,7 +227,7 @@ static void do_catch_up(struct spk_synth *synth) delay_time_val = delay_time->u.n.value; jiffy_delta_val = jiffy_delta->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout((delay_time_val)); jiff_max = jiffies + jiffy_delta_val; } } diff --git a/drivers/accessibility/speakup/speakup_keypc.c b/drivers/accessibility/speakup/speakup_keypc.c index 311f4aa0be2..99c523fdcc9 100644 --- a/drivers/accessibility/speakup/speakup_keypc.c +++ b/drivers/accessibility/speakup/speakup_keypc.c @@ -199,7 +199,7 @@ static void do_catch_up(struct spk_synth *synth) full_time_val = full_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); if (synth_full()) { - schedule_timeout(msecs_to_jiffies(full_time_val)); + schedule_msec_hrtimeout((full_time_val)); continue; } set_current_state(TASK_RUNNING); @@ -232,7 +232,7 @@ static void do_catch_up(struct spk_synth *synth) jiffy_delta_val = jiffy_delta->u.n.value; delay_time_val = delay_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); - schedule_timeout(msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout(delay_time_val); jiff_max = jiffies + jiffy_delta_val; } } diff --git a/drivers/accessibility/speakup/synth.c b/drivers/accessibility/speakup/synth.c index 2b8699673ba..1dd3710771b 100644 --- a/drivers/accessibility/speakup/synth.c +++ b/drivers/accessibility/speakup/synth.c @@ -98,7 +98,7 @@ static void _spk_do_catch_up(struct spk_synth *synth, int unicode) else ret = synth->io_ops->synth_out(synth, ch); if (!ret) { - schedule_timeout(msecs_to_jiffies(full_time_val)); + schedule_msec_hrtimeout(full_time_val); continue; } if (time_after_eq(jiffies, jiff_max) && (ch == SPACE)) { @@ -108,11 +108,9 @@ static void _spk_do_catch_up(struct spk_synth *synth, int unicode) full_time_val = full_time->u.n.value; spin_unlock_irqrestore(&speakup_info.spinlock, flags); if (synth->io_ops->synth_out(synth, synth->procspeech)) - schedule_timeout( - msecs_to_jiffies(delay_time_val)); + schedule_msec_hrtimeout(delay_time_val); else - schedule_timeout( - msecs_to_jiffies(full_time_val)); + schedule_msec_hrtimeout(full_time_val); jiff_max = jiffies + jiffy_delta_val; } set_current_state(TASK_RUNNING); diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c index 33192a8f687..66df8aabf42 100644 --- a/drivers/ata/ahci.c +++ b/drivers/ata/ahci.c @@ -1497,7 +1497,7 @@ static irqreturn_t ahci_thunderx_irq_handler(int irq, void *dev_instance) } #endif -static void ahci_remap_check(struct pci_dev *pdev, int bar, +static int ahci_remap_check(struct pci_dev *pdev, int bar, struct ahci_host_priv *hpriv) { int i; @@ -1510,7 +1510,7 @@ static void ahci_remap_check(struct pci_dev *pdev, int bar, pci_resource_len(pdev, bar) < SZ_512K || bar != AHCI_PCI_BAR_STANDARD || !(readl(hpriv->mmio + AHCI_VSCAP) & 1)) - return; + return 0; cap = readq(hpriv->mmio + AHCI_REMAP_CAP); for (i = 0; i < AHCI_MAX_REMAP; i++) { @@ -1525,18 +1525,11 @@ static void ahci_remap_check(struct pci_dev *pdev, int bar, } if (!hpriv->remapped_nvme) - return; - - dev_warn(&pdev->dev, "Found %u remapped NVMe devices.\n", - hpriv->remapped_nvme); - dev_warn(&pdev->dev, - "Switch your BIOS from RAID to AHCI mode to use them.\n"); + return 0; - /* - * Don't rely on the msi-x capability in the remap case, - * share the legacy interrupt across ahci and remapped devices. - */ - hpriv->flags |= AHCI_HFLAG_NO_MSI; + /* Abort probe, allowing intel-nvme-remap to step in when available */ + dev_info(&pdev->dev, "Device will be handled by intel-nvme-remap.\n"); + return -ENODEV; } static int ahci_get_irq_vector(struct ata_host *host, int port) @@ -1761,7 +1754,9 @@ static int ahci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) hpriv->mmio = pcim_iomap_table(pdev)[ahci_pci_bar]; /* detect remapped nvme devices */ - ahci_remap_check(pdev, ahci_pci_bar, hpriv); + rc = ahci_remap_check(pdev, ahci_pci_bar, hpriv); + if (rc) + return rc; sysfs_add_file_to_group(&pdev->dev.kobj, &dev_attr_remapped_nvme.attr, diff --git a/drivers/block/loop.c b/drivers/block/loop.c index e81298b9122..c46d5db279e 100644 --- a/drivers/block/loop.c +++ b/drivers/block/loop.c @@ -814,6 +814,24 @@ static int loop_change_fd(struct loop_device *lo, struct block_device *bdev, return error; } +/* + * for AUFS + * no get/put for file. + */ +struct file *loop_backing_file(struct super_block *sb) +{ + struct file *ret; + struct loop_device *l; + + ret = NULL; + if (MAJOR(sb->s_dev) == LOOP_MAJOR) { + l = sb->s_bdev->bd_disk->private_data; + ret = l->lo_backing_file; + } + return ret; +} +EXPORT_SYMBOL_GPL(loop_backing_file); + /* loop sysfs attributes */ static ssize_t loop_attr_show(struct device *dev, char *page, diff --git a/drivers/block/swim.c b/drivers/block/swim.c index 2917b21f48f..e8ab61f8df3 100644 --- a/drivers/block/swim.c +++ b/drivers/block/swim.c @@ -328,7 +328,7 @@ static inline void swim_motor(struct swim __iomem *base, if (swim_readbit(base, MOTOR_ON)) break; set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); } } else if (action == OFF) { swim_action(base, MOTOR_OFF); @@ -347,7 +347,7 @@ static inline void swim_eject(struct swim __iomem *base) if (!swim_readbit(base, DISK_IN)) break; set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); } swim_select(base, RELAX); } @@ -371,7 +371,7 @@ static inline int swim_step(struct swim __iomem *base) for (wait = 0; wait < HZ; wait++) { set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); swim_select(base, RELAX); if (!swim_readbit(base, STEP)) diff --git a/drivers/bluetooth/btusb.c b/drivers/bluetooth/btusb.c index ad5c60f616c..2c22ccc6ef8 100644 --- a/drivers/bluetooth/btusb.c +++ b/drivers/bluetooth/btusb.c @@ -516,6 +516,7 @@ static const struct dmi_system_id btusb_needs_reset_resume_table[] = { #define BTUSB_HW_RESET_ACTIVE 12 #define BTUSB_TX_WAIT_VND_EVT 13 #define BTUSB_WAKEUP_DISABLE 14 +#define BTUSB_USE_ALT3_FOR_WBS 15 struct btusb_data { struct hci_dev *hdev; @@ -1748,16 +1749,20 @@ static void btusb_work(struct work_struct *work) /* Bluetooth USB spec recommends alt 6 (63 bytes), but * many adapters do not support it. Alt 1 appears to * work for all adapters that do not have alt 6, and - * which work with WBS at all. + * which work with WBS at all. Some devices prefer + * alt 3 (HCI payload >= 60 Bytes let air packet + * data satisfy 60 bytes), requiring + * MTU >= 3 (packets) * 25 (size) - 3 (headers) = 72 + * see also Core spec 5, vol 4, B 2.1.1 & Table 2.1. */ - new_alts = btusb_find_altsetting(data, 6) ? 6 : 1; - /* Because mSBC frames do not need to be aligned to the - * SCO packet boundary. If support the Alt 3, use the - * Alt 3 for HCI payload >= 60 Bytes let air packet - * data satisfy 60 bytes. - */ - if (new_alts == 1 && btusb_find_altsetting(data, 3)) + if (btusb_find_altsetting(data, 6)) + new_alts = 6; + else if (test_bit(BTUSB_USE_ALT3_FOR_WBS, &data->flags) && + hdev->sco_mtu >= 72 && + btusb_find_altsetting(data, 3)) new_alts = 3; + else + new_alts = 1; } if (btusb_switch_alt_setting(hdev, new_alts) < 0) @@ -4732,6 +4737,7 @@ static int btusb_probe(struct usb_interface *intf, * (DEVICE_REMOTE_WAKEUP) */ set_bit(BTUSB_WAKEUP_DISABLE, &data->flags); + set_bit(BTUSB_USE_ALT3_FOR_WBS, &data->flags); } if (!reset) diff --git a/drivers/char/ipmi/ipmi_msghandler.c b/drivers/char/ipmi/ipmi_msghandler.c index 8a0e97b33ca..b5b22e487c0 100644 --- a/drivers/char/ipmi/ipmi_msghandler.c +++ b/drivers/char/ipmi/ipmi_msghandler.c @@ -3548,7 +3548,7 @@ static void cleanup_smi_msgs(struct ipmi_smi *intf) /* Current message first, to preserve order */ while (intf->curr_msg && !list_empty(&intf->waiting_rcv_msgs)) { /* Wait for the message to clear out. */ - schedule_timeout(1); + schedule_min_hrtimeout(); } /* No need for locks, the interface is down. */ diff --git a/drivers/char/ipmi/ipmi_ssif.c b/drivers/char/ipmi/ipmi_ssif.c index 20d5af92966..05b8d6a8171 100644 --- a/drivers/char/ipmi/ipmi_ssif.c +++ b/drivers/char/ipmi/ipmi_ssif.c @@ -1247,7 +1247,7 @@ static void shutdown_ssif(void *send_info) /* make sure the driver is not looking for flags any more. */ while (ssif_info->ssif_state != SSIF_NORMAL) - schedule_timeout(1); + schedule_min_hrtimeout(); ssif_info->stopping = true; del_timer_sync(&ssif_info->watch_timer); diff --git a/drivers/comedi/drivers/ni_mio_common.c b/drivers/comedi/drivers/ni_mio_common.c index 4f80a4991f9..c164c852490 100644 --- a/drivers/comedi/drivers/ni_mio_common.c +++ b/drivers/comedi/drivers/ni_mio_common.c @@ -4747,7 +4747,7 @@ static int cs5529_wait_for_idle(struct comedi_device *dev) if ((status & NI67XX_CAL_STATUS_BUSY) == 0) break; set_current_state(TASK_INTERRUPTIBLE); - if (schedule_timeout(1)) + if (schedule_min_hrtimeout()) return -EIO; } if (i == timeout) { diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig index c3038cdc686..da7ac1fdedb 100644 --- a/drivers/cpufreq/Kconfig +++ b/drivers/cpufreq/Kconfig @@ -40,6 +40,7 @@ choice default CPU_FREQ_DEFAULT_GOV_USERSPACE if ARM_SA1100_CPUFREQ || ARM_SA1110_CPUFREQ default CPU_FREQ_DEFAULT_GOV_SCHEDUTIL if ARM64 || ARM default CPU_FREQ_DEFAULT_GOV_SCHEDUTIL if X86_INTEL_PSTATE && SMP + default CPU_FREQ_DEFAULT_GOV_ONDEMAND if !X86_INTEL_PSTATE default CPU_FREQ_DEFAULT_GOV_PERFORMANCE help This option sets which CPUFreq governor shall be loaded at diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c index ac361a8b1d3..5376a2b5c84 100644 --- a/drivers/cpufreq/cpufreq_ondemand.c +++ b/drivers/cpufreq/cpufreq_ondemand.c @@ -18,10 +18,22 @@ #include "cpufreq_ondemand.h" /* On-demand governor macros */ +#if defined(CONFIG_PCK_INTERACTIVE) && defined(CONFIG_SCHED_MUQSS) +/* MuQSS is less likely to keep a single threaded task on the same core, so + have on-demand run at max speed from a much lower core utilization */ +#define DEF_FREQUENCY_UP_THRESHOLD (40) +#define MICRO_FREQUENCY_UP_THRESHOLD (45) +#define DEF_SAMPLING_DOWN_FACTOR (5) +#elif defined(CONFIG_PCK_INTERACTIVE) #define DEF_FREQUENCY_UP_THRESHOLD (80) +#define MICRO_FREQUENCY_UP_THRESHOLD (95) +#define DEF_SAMPLING_DOWN_FACTOR (5) +#else +#define DEF_FREQUENCY_UP_THRESHOLD (80) +#define MICRO_FREQUENCY_UP_THRESHOLD (95) #define DEF_SAMPLING_DOWN_FACTOR (1) +#endif #define MAX_SAMPLING_DOWN_FACTOR (100000) -#define MICRO_FREQUENCY_UP_THRESHOLD (95) #define MICRO_FREQUENCY_MIN_SAMPLE_RATE (10000) #define MIN_FREQUENCY_UP_THRESHOLD (1) #define MAX_FREQUENCY_UP_THRESHOLD (100) diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c index 0e69dffd5a7..2353ccbc48f 100644 --- a/drivers/cpufreq/intel_pstate.c +++ b/drivers/cpufreq/intel_pstate.c @@ -3157,6 +3157,8 @@ static int __init intel_pstate_setup(char *str) pr_info("HWP disabled\n"); no_hwp = 1; } + if (!strcmp(str, "enable")) + no_load = 0; if (!strcmp(str, "force")) force_load = 1; if (!strcmp(str, "hwp_only")) diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_cmd.c b/drivers/gpu/drm/vmwgfx/vmwgfx_cmd.c index 20246a7c97c..499a4542383 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_cmd.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_cmd.c @@ -227,7 +227,7 @@ static int vmw_fifo_wait_noirq(struct vmw_private *dev_priv, DRM_ERROR("SVGA device lockup.\n"); break; } - schedule_timeout(1); + schedule_min_hrtimeout(); if (interruptible && signal_pending(current)) { ret = -ERESTARTSYS; break; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c index 6c2a569f1fc..8d7feeb0d7a 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_irq.c @@ -201,7 +201,7 @@ int vmw_fallback_wait(struct vmw_private *dev_priv, break; } if (lazy) - schedule_timeout(1); + schedule_min_hrtimeout(); else if ((++count & 0x0F) == 0) { /** * FIXME: Use schedule_hr_timeout here for diff --git a/drivers/hwmon/fam15h_power.c b/drivers/hwmon/fam15h_power.c index 29f5fed28c2..974cb08c7aa 100644 --- a/drivers/hwmon/fam15h_power.c +++ b/drivers/hwmon/fam15h_power.c @@ -221,7 +221,7 @@ static ssize_t power1_average_show(struct device *dev, prev_ptsc[cu] = data->cpu_sw_pwr_ptsc[cu]; } - leftover = schedule_timeout_interruptible(msecs_to_jiffies(data->power_period)); + leftover = schedule_msec_hrtimeout_interruptible((data->power_period)); if (leftover) return 0; diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig index 10acece9d7b..3d43e83bad0 100644 --- a/drivers/i2c/busses/Kconfig +++ b/drivers/i2c/busses/Kconfig @@ -219,6 +219,15 @@ config I2C_CHT_WC combined with a FUSB302 Type-C port-controller as such it is advised to also select CONFIG_TYPEC_FUSB302=m. +config I2C_NCT6775 + tristate "Nuvoton NCT6775 and compatible SMBus controller" + help + If you say yes to this option, support will be included for the + Nuvoton NCT6775 and compatible SMBus controllers. + + This driver can also be built as a module. If so, the module + will be called i2c-nct6775. + config I2C_NFORCE2 tristate "Nvidia nForce2, nForce3 and nForce4" depends on PCI diff --git a/drivers/i2c/busses/Makefile b/drivers/i2c/busses/Makefile index 69e9963615f..443eeec4854 100644 --- a/drivers/i2c/busses/Makefile +++ b/drivers/i2c/busses/Makefile @@ -17,6 +17,7 @@ obj-$(CONFIG_I2C_CHT_WC) += i2c-cht-wc.o obj-$(CONFIG_I2C_I801) += i2c-i801.o obj-$(CONFIG_I2C_ISCH) += i2c-isch.o obj-$(CONFIG_I2C_ISMT) += i2c-ismt.o +obj-$(CONFIG_I2C_NCT6775) += i2c-nct6775.o obj-$(CONFIG_I2C_NFORCE2) += i2c-nforce2.o obj-$(CONFIG_I2C_NFORCE2_S4985) += i2c-nforce2-s4985.o obj-$(CONFIG_I2C_NVIDIA_GPU) += i2c-nvidia-gpu.o diff --git a/drivers/i2c/busses/i2c-nct6775.c b/drivers/i2c/busses/i2c-nct6775.c new file mode 100644 index 00000000000..0462f095204 --- /dev/null +++ b/drivers/i2c/busses/i2c-nct6775.c @@ -0,0 +1,647 @@ +/* + * i2c-nct6775 - Driver for the SMBus master functionality of + * Nuvoton NCT677x Super-I/O chips + * + * Copyright (C) 2019 Adam Honse + * + * Derived from nct6775 hwmon driver + * Copyright (C) 2012 Guenter Roeck + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. + * + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#define DRVNAME "i2c-nct6775" + +/* Nuvoton SMBus address offsets */ +#define SMBHSTDAT (0 + nuvoton_nct6793d_smba) +#define SMBBLKSZ (1 + nuvoton_nct6793d_smba) +#define SMBHSTCMD (2 + nuvoton_nct6793d_smba) +#define SMBHSTIDX (3 + nuvoton_nct6793d_smba) //Index field is the Command field on other controllers +#define SMBHSTCTL (4 + nuvoton_nct6793d_smba) +#define SMBHSTADD (5 + nuvoton_nct6793d_smba) +#define SMBHSTERR (9 + nuvoton_nct6793d_smba) +#define SMBHSTSTS (0xE + nuvoton_nct6793d_smba) + +/* Command register */ +#define NCT6793D_READ_BYTE 0 +#define NCT6793D_READ_WORD 1 +#define NCT6793D_READ_BLOCK 2 +#define NCT6793D_BLOCK_WRITE_READ_PROC_CALL 3 +#define NCT6793D_PROC_CALL 4 +#define NCT6793D_WRITE_BYTE 8 +#define NCT6793D_WRITE_WORD 9 +#define NCT6793D_WRITE_BLOCK 10 + +/* Control register */ +#define NCT6793D_MANUAL_START 128 +#define NCT6793D_SOFT_RESET 64 + +/* Error register */ +#define NCT6793D_NO_ACK 32 + +/* Status register */ +#define NCT6793D_FIFO_EMPTY 1 +#define NCT6793D_FIFO_FULL 2 +#define NCT6793D_MANUAL_ACTIVE 4 + +#define NCT6775_LD_SMBUS 0x0B + +/* Other settings */ +#define MAX_RETRIES 400 + +enum kinds { nct6106, nct6775, nct6776, nct6779, nct6791, nct6792, nct6793, + nct6795, nct6796, nct6798 }; + +struct nct6775_sio_data { + int sioreg; + enum kinds kind; +}; + +/* used to set data->name = nct6775_device_names[data->sio_kind] */ +static const char * const nct6775_device_names[] = { + "nct6106", + "nct6775", + "nct6776", + "nct6779", + "nct6791", + "nct6792", + "nct6793", + "nct6795", + "nct6796", + "nct6798", +}; + +static const char * const nct6775_sio_names[] __initconst = { + "NCT6106D", + "NCT6775F", + "NCT6776D/F", + "NCT6779D", + "NCT6791D", + "NCT6792D", + "NCT6793D", + "NCT6795D", + "NCT6796D", + "NCT6798D", +}; + +#define SIO_REG_LDSEL 0x07 /* Logical device select */ +#define SIO_REG_DEVID 0x20 /* Device ID (2 bytes) */ +#define SIO_REG_SMBA 0x62 /* SMBus base address register */ + +#define SIO_NCT6106_ID 0xc450 +#define SIO_NCT6775_ID 0xb470 +#define SIO_NCT6776_ID 0xc330 +#define SIO_NCT6779_ID 0xc560 +#define SIO_NCT6791_ID 0xc800 +#define SIO_NCT6792_ID 0xc910 +#define SIO_NCT6793_ID 0xd120 +#define SIO_NCT6795_ID 0xd350 +#define SIO_NCT6796_ID 0xd420 +#define SIO_NCT6798_ID 0xd428 +#define SIO_ID_MASK 0xFFF0 + +static inline void +superio_outb(int ioreg, int reg, int val) +{ + outb(reg, ioreg); + outb(val, ioreg + 1); +} + +static inline int +superio_inb(int ioreg, int reg) +{ + outb(reg, ioreg); + return inb(ioreg + 1); +} + +static inline void +superio_select(int ioreg, int ld) +{ + outb(SIO_REG_LDSEL, ioreg); + outb(ld, ioreg + 1); +} + +static inline int +superio_enter(int ioreg) +{ + /* + * Try to reserve and for exclusive access. + */ + if (!request_muxed_region(ioreg, 2, DRVNAME)) + return -EBUSY; + + outb(0x87, ioreg); + outb(0x87, ioreg); + + return 0; +} + +static inline void +superio_exit(int ioreg) +{ + outb(0xaa, ioreg); + outb(0x02, ioreg); + outb(0x02, ioreg + 1); + release_region(ioreg, 2); +} + +/* + * ISA constants + */ + +#define IOREGION_ALIGNMENT (~7) +#define IOREGION_LENGTH 2 +#define ADDR_REG_OFFSET 0 +#define DATA_REG_OFFSET 1 + +#define NCT6775_REG_BANK 0x4E +#define NCT6775_REG_CONFIG 0x40 + +static struct i2c_adapter *nct6775_adapter; + +struct i2c_nct6775_adapdata { + unsigned short smba; +}; + +/* Return negative errno on error. */ +static s32 nct6775_access(struct i2c_adapter * adap, u16 addr, + unsigned short flags, char read_write, + u8 command, int size, union i2c_smbus_data * data) +{ + struct i2c_nct6775_adapdata *adapdata = i2c_get_adapdata(adap); + unsigned short nuvoton_nct6793d_smba = adapdata->smba; + int i, len, cnt; + union i2c_smbus_data tmp_data; + int timeout = 0; + + tmp_data.word = 0; + cnt = 0; + len = 0; + + outb_p(NCT6793D_SOFT_RESET, SMBHSTCTL); + + switch (size) { + case I2C_SMBUS_QUICK: + outb_p((addr << 1) | read_write, + SMBHSTADD); + break; + case I2C_SMBUS_BYTE_DATA: + tmp_data.byte = data->byte; + case I2C_SMBUS_BYTE: + outb_p((addr << 1) | read_write, + SMBHSTADD); + outb_p(command, SMBHSTIDX); + if (read_write == I2C_SMBUS_WRITE) { + outb_p(tmp_data.byte, SMBHSTDAT); + outb_p(NCT6793D_WRITE_BYTE, SMBHSTCMD); + } + else { + outb_p(NCT6793D_READ_BYTE, SMBHSTCMD); + } + break; + case I2C_SMBUS_WORD_DATA: + outb_p((addr << 1) | read_write, + SMBHSTADD); + outb_p(command, SMBHSTIDX); + if (read_write == I2C_SMBUS_WRITE) { + outb_p(data->word & 0xff, SMBHSTDAT); + outb_p((data->word & 0xff00) >> 8, SMBHSTDAT); + outb_p(NCT6793D_WRITE_WORD, SMBHSTCMD); + } + else { + outb_p(NCT6793D_READ_WORD, SMBHSTCMD); + } + break; + case I2C_SMBUS_BLOCK_DATA: + outb_p((addr << 1) | read_write, + SMBHSTADD); + outb_p(command, SMBHSTIDX); + if (read_write == I2C_SMBUS_WRITE) { + len = data->block[0]; + if (len == 0 || len > I2C_SMBUS_BLOCK_MAX) + return -EINVAL; + outb_p(len, SMBBLKSZ); + + cnt = 1; + if (len >= 4) { + for (i = cnt; i <= 4; i++) { + outb_p(data->block[i], SMBHSTDAT); + } + + len -= 4; + cnt += 4; + } + else { + for (i = cnt; i <= len; i++ ) { + outb_p(data->block[i], SMBHSTDAT); + } + + len = 0; + } + + outb_p(NCT6793D_WRITE_BLOCK, SMBHSTCMD); + } + else { + return -ENOTSUPP; + } + break; + default: + dev_warn(&adap->dev, "Unsupported transaction %d\n", size); + return -EOPNOTSUPP; + } + + outb_p(NCT6793D_MANUAL_START, SMBHSTCTL); + + while ((size == I2C_SMBUS_BLOCK_DATA) && (len > 0)) { + if (read_write == I2C_SMBUS_WRITE) { + timeout = 0; + while ((inb_p(SMBHSTSTS) & NCT6793D_FIFO_EMPTY) == 0) + { + if(timeout > MAX_RETRIES) + { + return -ETIMEDOUT; + } + usleep_range(250, 500); + timeout++; + } + + //Load more bytes into FIFO + if (len >= 4) { + for (i = cnt; i <= (cnt + 4); i++) { + outb_p(data->block[i], SMBHSTDAT); + } + + len -= 4; + cnt += 4; + } + else { + for (i = cnt; i <= (cnt + len); i++) { + outb_p(data->block[i], SMBHSTDAT); + } + + len = 0; + } + } + else { + return -ENOTSUPP; + } + + } + + //wait for manual mode to complete + timeout = 0; + while ((inb_p(SMBHSTSTS) & NCT6793D_MANUAL_ACTIVE) != 0) + { + if(timeout > MAX_RETRIES) + { + return -ETIMEDOUT; + } + usleep_range(250, 500); + timeout++; + } + + if ((inb_p(SMBHSTERR) & NCT6793D_NO_ACK) != 0) { + return -ENXIO; + } + else if ((read_write == I2C_SMBUS_WRITE) || (size == I2C_SMBUS_QUICK)) { + return 0; + } + + switch (size) { + case I2C_SMBUS_QUICK: + case I2C_SMBUS_BYTE_DATA: + data->byte = inb_p(SMBHSTDAT); + break; + case I2C_SMBUS_WORD_DATA: + data->word = inb_p(SMBHSTDAT) + (inb_p(SMBHSTDAT) << 8); + break; + } + return 0; +} + +static u32 nct6775_func(struct i2c_adapter *adapter) +{ + return I2C_FUNC_SMBUS_QUICK | I2C_FUNC_SMBUS_BYTE | + I2C_FUNC_SMBUS_BYTE_DATA | I2C_FUNC_SMBUS_WORD_DATA | + I2C_FUNC_SMBUS_BLOCK_DATA; +} + +static const struct i2c_algorithm smbus_algorithm = { + .smbus_xfer = nct6775_access, + .functionality = nct6775_func, +}; + +static int nct6775_add_adapter(unsigned short smba, const char *name, struct i2c_adapter **padap) +{ + struct i2c_adapter *adap; + struct i2c_nct6775_adapdata *adapdata; + int retval; + + adap = kzalloc(sizeof(*adap), GFP_KERNEL); + if (adap == NULL) { + return -ENOMEM; + } + + adap->owner = THIS_MODULE; + adap->class = I2C_CLASS_HWMON | I2C_CLASS_SPD; + adap->algo = &smbus_algorithm; + + adapdata = kzalloc(sizeof(*adapdata), GFP_KERNEL); + if (adapdata == NULL) { + kfree(adap); + return -ENOMEM; + } + + adapdata->smba = smba; + + snprintf(adap->name, sizeof(adap->name), + "SMBus NCT67xx adapter%s at %04x", name, smba); + + i2c_set_adapdata(adap, adapdata); + + retval = i2c_add_adapter(adap); + if (retval) { + kfree(adapdata); + kfree(adap); + return retval; + } + + *padap = adap; + return 0; +} + +static void nct6775_remove_adapter(struct i2c_adapter *adap) +{ + struct i2c_nct6775_adapdata *adapdata = i2c_get_adapdata(adap); + + if (adapdata->smba) { + i2c_del_adapter(adap); + kfree(adapdata); + kfree(adap); + } +} + +//static SIMPLE_DEV_PM_OPS(nct6775_dev_pm_ops, nct6775_suspend, nct6775_resume); + +/* + * when Super-I/O functions move to a separate file, the Super-I/O + * bus will manage the lifetime of the device and this module will only keep + * track of the nct6775 driver. But since we use platform_device_alloc(), we + * must keep track of the device + */ +static struct platform_device *pdev[2]; + +static int nct6775_probe(struct platform_device *pdev) +{ + struct device *dev = &pdev->dev; + struct nct6775_sio_data *sio_data = dev_get_platdata(dev); + struct resource *res; + + res = platform_get_resource(pdev, IORESOURCE_IO, 0); + if (!devm_request_region(&pdev->dev, res->start, IOREGION_LENGTH, + DRVNAME)) + return -EBUSY; + + switch (sio_data->kind) { + case nct6791: + case nct6792: + case nct6793: + case nct6795: + case nct6796: + case nct6798: + nct6775_add_adapter(res->start, "", &nct6775_adapter); + break; + default: + return -ENODEV; + } + + return 0; +} +/* +static void nct6791_enable_io_mapping(int sioaddr) +{ + int val; + + val = superio_inb(sioaddr, NCT6791_REG_HM_IO_SPACE_LOCK_ENABLE); + if (val & 0x10) { + pr_info("Enabling hardware monitor logical device mappings.\n"); + superio_outb(sioaddr, NCT6791_REG_HM_IO_SPACE_LOCK_ENABLE, + val & ~0x10); + } +}*/ + +static struct platform_driver i2c_nct6775_driver = { + .driver = { + .name = DRVNAME, +// .pm = &nct6775_dev_pm_ops, + }, + .probe = nct6775_probe, +}; + +static void __exit i2c_nct6775_exit(void) +{ + int i; + + if(nct6775_adapter) + nct6775_remove_adapter(nct6775_adapter); + + for (i = 0; i < ARRAY_SIZE(pdev); i++) { + if (pdev[i]) + platform_device_unregister(pdev[i]); + } + platform_driver_unregister(&i2c_nct6775_driver); +} + +/* nct6775_find() looks for a '627 in the Super-I/O config space */ +static int __init nct6775_find(int sioaddr, struct nct6775_sio_data *sio_data) +{ + u16 val; + int err; + int addr; + + err = superio_enter(sioaddr); + if (err) + return err; + + val = (superio_inb(sioaddr, SIO_REG_DEVID) << 8) | + superio_inb(sioaddr, SIO_REG_DEVID + 1); + + switch (val & SIO_ID_MASK) { + case SIO_NCT6106_ID: + sio_data->kind = nct6106; + break; + case SIO_NCT6775_ID: + sio_data->kind = nct6775; + break; + case SIO_NCT6776_ID: + sio_data->kind = nct6776; + break; + case SIO_NCT6779_ID: + sio_data->kind = nct6779; + break; + case SIO_NCT6791_ID: + sio_data->kind = nct6791; + break; + case SIO_NCT6792_ID: + sio_data->kind = nct6792; + break; + case SIO_NCT6793_ID: + sio_data->kind = nct6793; + break; + case SIO_NCT6795_ID: + sio_data->kind = nct6795; + break; + case SIO_NCT6796_ID: + sio_data->kind = nct6796; + break; + case SIO_NCT6798_ID: + sio_data->kind = nct6798; + break; + default: + if (val != 0xffff) + pr_debug("unsupported chip ID: 0x%04x\n", val); + superio_exit(sioaddr); + return -ENODEV; + } + + /* We have a known chip, find the SMBus I/O address */ + superio_select(sioaddr, NCT6775_LD_SMBUS); + val = (superio_inb(sioaddr, SIO_REG_SMBA) << 8) + | superio_inb(sioaddr, SIO_REG_SMBA + 1); + addr = val & IOREGION_ALIGNMENT; + if (addr == 0) { + pr_err("Refusing to enable a Super-I/O device with a base I/O port 0\n"); + superio_exit(sioaddr); + return -ENODEV; + } + + //if (sio_data->kind == nct6791 || sio_data->kind == nct6792 || + // sio_data->kind == nct6793 || sio_data->kind == nct6795 || + // sio_data->kind == nct6796) + // nct6791_enable_io_mapping(sioaddr); + + superio_exit(sioaddr); + pr_info("Found %s or compatible chip at %#x:%#x\n", + nct6775_sio_names[sio_data->kind], sioaddr, addr); + sio_data->sioreg = sioaddr; + + return addr; +} + +static int __init i2c_nct6775_init(void) +{ + int i, err; + bool found = false; + int address; + struct resource res; + struct nct6775_sio_data sio_data; + int sioaddr[2] = { 0x2e, 0x4e }; + + err = platform_driver_register(&i2c_nct6775_driver); + if (err) + return err; + + /* + * initialize sio_data->kind and sio_data->sioreg. + * + * when Super-I/O functions move to a separate file, the Super-I/O + * driver will probe 0x2e and 0x4e and auto-detect the presence of a + * nct6775 hardware monitor, and call probe() + */ + for (i = 0; i < ARRAY_SIZE(pdev); i++) { + address = nct6775_find(sioaddr[i], &sio_data); + if (address <= 0) + continue; + + found = true; + + pdev[i] = platform_device_alloc(DRVNAME, address); + if (!pdev[i]) { + err = -ENOMEM; + goto exit_device_unregister; + } + + err = platform_device_add_data(pdev[i], &sio_data, + sizeof(struct nct6775_sio_data)); + if (err) + goto exit_device_put; + + memset(&res, 0, sizeof(res)); + res.name = DRVNAME; + res.start = address; + res.end = address + IOREGION_LENGTH - 1; + res.flags = IORESOURCE_IO; + + err = acpi_check_resource_conflict(&res); + if (err) { + platform_device_put(pdev[i]); + pdev[i] = NULL; + continue; + } + + err = platform_device_add_resources(pdev[i], &res, 1); + if (err) + goto exit_device_put; + + /* platform_device_add calls probe() */ + err = platform_device_add(pdev[i]); + if (err) + goto exit_device_put; + } + if (!found) { + err = -ENODEV; + goto exit_unregister; + } + + return 0; + +exit_device_put: + platform_device_put(pdev[i]); +exit_device_unregister: + while (--i >= 0) { + if (pdev[i]) + platform_device_unregister(pdev[i]); + } +exit_unregister: + platform_driver_unregister(&i2c_nct6775_driver); + return err; +} + +MODULE_AUTHOR("Adam Honse "); +MODULE_DESCRIPTION("SMBus driver for NCT6775F and compatible chips"); +MODULE_LICENSE("GPL"); + +module_init(i2c_nct6775_init); +module_exit(i2c_nct6775_exit); diff --git a/drivers/i2c/busses/i2c-piix4.c b/drivers/i2c/busses/i2c-piix4.c index 8c1b31ed0c4..0b1b36fa0b7 100644 --- a/drivers/i2c/busses/i2c-piix4.c +++ b/drivers/i2c/busses/i2c-piix4.c @@ -467,11 +467,11 @@ static int piix4_transaction(struct i2c_adapter *piix4_adapter) if (srvrworks_csb5_delay) /* Extra delay for SERVERWORKS_CSB5 */ usleep_range(2000, 2100); else - usleep_range(250, 500); + usleep_range(25, 50); while ((++timeout < MAX_TIMEOUT) && ((temp = inb_p(SMBHSTSTS)) & 0x01)) - usleep_range(250, 500); + usleep_range(25, 50); /* If the SMBus is still busy, we give up */ if (timeout == MAX_TIMEOUT) { diff --git a/drivers/iio/light/tsl2563.c b/drivers/iio/light/tsl2563.c index 5bf2bfbc537..921f86fcd4e 100644 --- a/drivers/iio/light/tsl2563.c +++ b/drivers/iio/light/tsl2563.c @@ -275,7 +275,7 @@ static void tsl2563_wait_adc(struct tsl2563_chip *chip) * TODO: Make sure that we wait at least required delay but why we * have to extend it one tick more? */ - schedule_timeout_interruptible(msecs_to_jiffies(delay) + 2); + schedule_msec_hrtimeout_interruptible(delay + 2); } static int tsl2563_adjust_gainlevel(struct tsl2563_chip *chip, u16 adc) diff --git a/drivers/input/touchscreen/atmel_mxt_ts.c b/drivers/input/touchscreen/atmel_mxt_ts.c index e1a453a8d26..9ff8ad2b828 100644 --- a/drivers/input/touchscreen/atmel_mxt_ts.c +++ b/drivers/input/touchscreen/atmel_mxt_ts.c @@ -2234,7 +2234,7 @@ static int mxt_initialize(struct mxt_data *data) if (error) return error; - error = reject_firmware_nowait(THIS_MODULE, true, MXT_CFG_NAME, + error = request_firmware_nowait(THIS_MODULE, true, MXT_CFG_NAME, &client->dev, GFP_KERNEL, data, mxt_config_cb); if (error) { diff --git a/drivers/media/i2c/msp3400-driver.c b/drivers/media/i2c/msp3400-driver.c index 39530d43590..a7caf2eb577 100644 --- a/drivers/media/i2c/msp3400-driver.c +++ b/drivers/media/i2c/msp3400-driver.c @@ -170,7 +170,7 @@ static int msp_read(struct i2c_client *client, int dev, int addr) break; dev_warn(&client->dev, "I/O error #%d (read 0x%02x/0x%02x)\n", err, dev, addr); - schedule_timeout_interruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_interruptible((10)); } if (err == 3) { dev_warn(&client->dev, "resetting chip, sound will go off.\n"); @@ -211,7 +211,7 @@ static int msp_write(struct i2c_client *client, int dev, int addr, int val) break; dev_warn(&client->dev, "I/O error #%d (write 0x%02x/0x%02x)\n", err, dev, addr); - schedule_timeout_interruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_interruptible((10)); } if (err == 3) { dev_warn(&client->dev, "resetting chip, sound will go off.\n"); diff --git a/drivers/media/pci/cx18/cx18-gpio.c b/drivers/media/pci/cx18/cx18-gpio.c index cf7cfda9410..f63e1748954 100644 --- a/drivers/media/pci/cx18/cx18-gpio.c +++ b/drivers/media/pci/cx18/cx18-gpio.c @@ -81,11 +81,11 @@ static void gpio_reset_seq(struct cx18 *cx, u32 active_lo, u32 active_hi, /* Assert */ gpio_update(cx, mask, ~active_lo); - schedule_timeout_uninterruptible(msecs_to_jiffies(assert_msecs)); + schedule_msec_hrtimeout_uninterruptible((assert_msecs)); /* Deassert */ gpio_update(cx, mask, ~active_hi); - schedule_timeout_uninterruptible(msecs_to_jiffies(recovery_msecs)); + schedule_msec_hrtimeout_uninterruptible((recovery_msecs)); } /* diff --git a/drivers/media/pci/ivtv/ivtv-gpio.c b/drivers/media/pci/ivtv/ivtv-gpio.c index 856e7ab7f33..766a2625133 100644 --- a/drivers/media/pci/ivtv/ivtv-gpio.c +++ b/drivers/media/pci/ivtv/ivtv-gpio.c @@ -105,7 +105,7 @@ void ivtv_reset_ir_gpio(struct ivtv *itv) curout = (curout & ~0xF) | 1; write_reg(curout, IVTV_REG_GPIO_OUT); /* We could use something else for smaller time */ - schedule_timeout_interruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_interruptible((1)); curout |= 2; write_reg(curout, IVTV_REG_GPIO_OUT); curdir &= ~0x80; @@ -125,11 +125,11 @@ int ivtv_reset_tuner_gpio(void *dev, int component, int cmd, int value) curout = read_reg(IVTV_REG_GPIO_OUT); curout &= ~(1 << itv->card->xceive_pin); write_reg(curout, IVTV_REG_GPIO_OUT); - schedule_timeout_interruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_interruptible((1)); curout |= 1 << itv->card->xceive_pin; write_reg(curout, IVTV_REG_GPIO_OUT); - schedule_timeout_interruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_interruptible((1)); return 0; } diff --git a/drivers/media/pci/ivtv/ivtv-ioctl.c b/drivers/media/pci/ivtv/ivtv-ioctl.c index 35dccb31174..8181cd65e87 100644 --- a/drivers/media/pci/ivtv/ivtv-ioctl.c +++ b/drivers/media/pci/ivtv/ivtv-ioctl.c @@ -1139,7 +1139,7 @@ void ivtv_s_std_dec(struct ivtv *itv, v4l2_std_id std) TASK_UNINTERRUPTIBLE); if ((read_reg(IVTV_REG_DEC_LINE_FIELD) >> 16) < 100) break; - schedule_timeout(msecs_to_jiffies(25)); + schedule_msec_hrtimeout((25)); } finish_wait(&itv->vsync_waitq, &wait); mutex_lock(&itv->serialize_lock); diff --git a/drivers/media/pci/ivtv/ivtv-streams.c b/drivers/media/pci/ivtv/ivtv-streams.c index f04ee84bab5..c4469b4b8f9 100644 --- a/drivers/media/pci/ivtv/ivtv-streams.c +++ b/drivers/media/pci/ivtv/ivtv-streams.c @@ -849,7 +849,7 @@ int ivtv_stop_v4l2_encode_stream(struct ivtv_stream *s, int gop_end) while (!test_bit(IVTV_F_I_EOS, &itv->i_flags) && time_before(jiffies, then + msecs_to_jiffies(2000))) { - schedule_timeout(msecs_to_jiffies(10)); + schedule_msec_hrtimeout((10)); } /* To convert jiffies to ms, we must multiply by 1000 diff --git a/drivers/media/radio/radio-mr800.c b/drivers/media/radio/radio-mr800.c index cb0437b4c33..163fffc0e1d 100644 --- a/drivers/media/radio/radio-mr800.c +++ b/drivers/media/radio/radio-mr800.c @@ -366,7 +366,7 @@ static int vidioc_s_hw_freq_seek(struct file *file, void *priv, retval = -ENODATA; break; } - if (schedule_timeout_interruptible(msecs_to_jiffies(10))) { + if (schedule_msec_hrtimeout_interruptible((10))) { retval = -ERESTARTSYS; break; } diff --git a/drivers/media/radio/radio-tea5777.c b/drivers/media/radio/radio-tea5777.c index fb9de7bbcd1..e53cf45e7f3 100644 --- a/drivers/media/radio/radio-tea5777.c +++ b/drivers/media/radio/radio-tea5777.c @@ -235,7 +235,7 @@ static int radio_tea5777_update_read_reg(struct radio_tea5777 *tea, int wait) } if (wait) { - if (schedule_timeout_interruptible(msecs_to_jiffies(wait))) + if (schedule_msec_hrtimeout_interruptible((wait))) return -ERESTARTSYS; } diff --git a/drivers/media/radio/tea575x.c b/drivers/media/radio/tea575x.c index c37315226c4..e73e6393403 100644 --- a/drivers/media/radio/tea575x.c +++ b/drivers/media/radio/tea575x.c @@ -401,7 +401,7 @@ int snd_tea575x_s_hw_freq_seek(struct file *file, struct snd_tea575x *tea, for (;;) { if (time_after(jiffies, timeout)) break; - if (schedule_timeout_interruptible(msecs_to_jiffies(10))) { + if (schedule_msec_hrtimeout_interruptible((10))) { /* some signal arrived, stop search */ tea->val &= ~TEA575X_BIT_SEARCH; snd_tea575x_set_freq(tea); diff --git a/drivers/mfd/ucb1x00-core.c b/drivers/mfd/ucb1x00-core.c index b690796d24d..448b13da62b 100644 --- a/drivers/mfd/ucb1x00-core.c +++ b/drivers/mfd/ucb1x00-core.c @@ -250,7 +250,7 @@ unsigned int ucb1x00_adc_read(struct ucb1x00 *ucb, int adc_channel, int sync) break; /* yield to other processes */ set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); } return UCB_ADC_DAT(val); diff --git a/drivers/misc/sgi-xp/xpc_channel.c b/drivers/misc/sgi-xp/xpc_channel.c index 8e6607fc8a6..b9ab770bbdb 100644 --- a/drivers/misc/sgi-xp/xpc_channel.c +++ b/drivers/misc/sgi-xp/xpc_channel.c @@ -834,7 +834,7 @@ xpc_allocate_msg_wait(struct xpc_channel *ch) atomic_inc(&ch->n_on_msg_allocate_wq); prepare_to_wait(&ch->msg_allocate_wq, &wait, TASK_INTERRUPTIBLE); - ret = schedule_timeout(1); + ret = schedule_min_hrtimeout(); finish_wait(&ch->msg_allocate_wq, &wait); atomic_dec(&ch->n_on_msg_allocate_wq); diff --git a/drivers/net/caif/caif_hsi.c b/drivers/net/caif/caif_hsi.c index 3d63b15bbaa..164071e9d45 100644 --- a/drivers/net/caif/caif_hsi.c +++ b/drivers/net/caif/caif_hsi.c @@ -924,7 +924,7 @@ static void cfhsi_wake_down(struct work_struct *work) break; set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); retry--; } diff --git a/drivers/net/can/usb/peak_usb/pcan_usb.c b/drivers/net/can/usb/peak_usb/pcan_usb.c index 899a3d21b77..3747195b6cf 100644 --- a/drivers/net/can/usb/peak_usb/pcan_usb.c +++ b/drivers/net/can/usb/peak_usb/pcan_usb.c @@ -300,7 +300,7 @@ static int pcan_usb_write_mode(struct peak_usb_device *dev, u8 onoff) } else { /* the PCAN-USB needs time to init */ set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(msecs_to_jiffies(PCAN_USB_STARTUP_TIMEOUT)); + schedule_msec_hrtimeout((PCAN_USB_STARTUP_TIMEOUT)); } return err; diff --git a/drivers/net/usb/lan78xx.c b/drivers/net/usb/lan78xx.c index 02bce40a67e..69758f825a4 100644 --- a/drivers/net/usb/lan78xx.c +++ b/drivers/net/usb/lan78xx.c @@ -2655,7 +2655,7 @@ static void lan78xx_terminate_urbs(struct lan78xx_net *dev) while (!skb_queue_empty(&dev->rxq) && !skb_queue_empty(&dev->txq) && !skb_queue_empty(&dev->done)) { - schedule_timeout(msecs_to_jiffies(UNLINK_TIMEOUT_MS)); + schedule_msec_hrtimeout((UNLINK_TIMEOUT_MS)); set_current_state(TASK_UNINTERRUPTIBLE); netif_dbg(dev, ifdown, dev->net, "waited for %d urb completions\n", temp); diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c index ecf62849f4c..ed6b424a597 100644 --- a/drivers/net/usb/usbnet.c +++ b/drivers/net/usb/usbnet.c @@ -764,7 +764,7 @@ static void wait_skb_queue_empty(struct sk_buff_head *q) spin_lock_irqsave(&q->lock, flags); while (!skb_queue_empty(q)) { spin_unlock_irqrestore(&q->lock, flags); - schedule_timeout(msecs_to_jiffies(UNLINK_TIMEOUT_MS)); + schedule_msec_hrtimeout((UNLINK_TIMEOUT_MS)); set_current_state(TASK_UNINTERRUPTIBLE); spin_lock_irqsave(&q->lock, flags); } diff --git a/drivers/net/wireless/intel/ipw2x00/ipw2100.c b/drivers/net/wireless/intel/ipw2x00/ipw2100.c index ef11e985aee..9f88f6b9d97 100644 --- a/drivers/net/wireless/intel/ipw2x00/ipw2100.c +++ b/drivers/net/wireless/intel/ipw2x00/ipw2100.c @@ -815,7 +815,7 @@ static int ipw2100_hw_send_command(struct ipw2100_priv *priv, * doesn't seem to have as many firmware restart cycles... * * As a test, we're sticking in a 1/100s delay here */ - schedule_timeout_uninterruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_uninterruptible((10)); return 0; @@ -1266,7 +1266,7 @@ static int ipw2100_start_adapter(struct ipw2100_priv *priv) IPW_DEBUG_FW("Waiting for f/w initialization to complete...\n"); i = 5000; do { - schedule_timeout_uninterruptible(msecs_to_jiffies(40)); + schedule_msec_hrtimeout_uninterruptible((40)); /* Todo... wait for sync command ... */ read_register(priv->net_dev, IPW_REG_INTA, &inta); diff --git a/drivers/parport/ieee1284.c b/drivers/parport/ieee1284.c index 4547ac44c8d..8fa1a7fdf12 100644 --- a/drivers/parport/ieee1284.c +++ b/drivers/parport/ieee1284.c @@ -202,7 +202,7 @@ int parport_wait_peripheral(struct parport *port, /* parport_wait_event didn't time out, but the * peripheral wasn't actually ready either. * Wait for another 10ms. */ - schedule_timeout_interruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_interruptible((10)); } } diff --git a/drivers/parport/ieee1284_ops.c b/drivers/parport/ieee1284_ops.c index 2c11bd3fe1f..8cb6b61c088 100644 --- a/drivers/parport/ieee1284_ops.c +++ b/drivers/parport/ieee1284_ops.c @@ -520,7 +520,7 @@ size_t parport_ieee1284_ecp_read_data (struct parport *port, /* Yield the port for a while. */ if (count && dev->port->irq != PARPORT_IRQ_NONE) { parport_release (dev); - schedule_timeout_interruptible(msecs_to_jiffies(40)); + schedule_msec_hrtimeout_interruptible((40)); parport_claim_or_block (dev); } else diff --git a/drivers/pci/controller/Makefile b/drivers/pci/controller/Makefile index 63e3880a3e8..64ab5f12160 100644 --- a/drivers/pci/controller/Makefile +++ b/drivers/pci/controller/Makefile @@ -1,4 +1,10 @@ # SPDX-License-Identifier: GPL-2.0 +ifdef CONFIG_X86_64 +ifdef CONFIG_SATA_AHCI +obj-y += intel-nvme-remap.o +endif +endif + obj-$(CONFIG_PCIE_CADENCE) += cadence/ obj-$(CONFIG_PCI_FTPCI100) += pci-ftpci100.o obj-$(CONFIG_PCI_HYPERV) += pci-hyperv.o diff --git a/drivers/pci/controller/intel-nvme-remap.c b/drivers/pci/controller/intel-nvme-remap.c new file mode 100644 index 00000000000..e105e6f5cc9 --- /dev/null +++ b/drivers/pci/controller/intel-nvme-remap.c @@ -0,0 +1,462 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Intel remapped NVMe device support. + * + * Copyright (c) 2019 Endless Mobile, Inc. + * Author: Daniel Drake + * + * Some products ship by default with the SATA controller in "RAID" or + * "Intel RST Premium With Intel Optane System Acceleration" mode. Under this + * mode, which we refer to as "remapped NVMe" mode, any installed NVMe + * devices disappear from the PCI bus, and instead their I/O memory becomes + * available within the AHCI device BARs. + * + * This scheme is understood to be a way of avoiding usage of the standard + * Windows NVMe driver under that OS, instead mandating usage of Intel's + * driver instead, which has better power management, and presumably offers + * some RAID/disk-caching solutions too. + * + * Here in this driver, we support the remapped NVMe mode by claiming the + * AHCI device and creating a fake PCIe root port. On the new bus, the + * original AHCI device is exposed with only minor tweaks. Then, fake PCI + * devices corresponding to the remapped NVMe devices are created. The usual + * ahci and nvme drivers are then expected to bind to these devices and + * operate as normal. + * + * The PCI configuration space for the NVMe devices is completely + * unavailable, so we fake a minimal one and hope for the best. + * + * Interrupts are shared between the AHCI and NVMe devices. For simplicity, + * we only support the legacy interrupt here, although MSI support + * could potentially be added later. + */ + +#define MODULE_NAME "intel-nvme-remap" + +#include +#include +#include +#include +#include + +#define AHCI_PCI_BAR_STANDARD 5 + +struct nvme_remap_dev { + struct pci_dev *dev; /* AHCI device */ + struct pci_bus *bus; /* our fake PCI bus */ + struct pci_sysdata sysdata; + int irq_base; /* our fake interrupts */ + + /* + * When we detect an all-ones write to a BAR register, this flag + * is set, so that we return the BAR size on the next read (a + * standard PCI behaviour). + * This includes the assumption that an all-ones BAR write is + * immediately followed by a read of the same register. + */ + bool bar_sizing; + + /* + * Resources copied from the AHCI device, to be regarded as + * resources on our fake bus. + */ + struct resource ahci_resources[PCI_NUM_RESOURCES]; + + /* Resources corresponding to the NVMe devices. */ + struct resource remapped_dev_mem[AHCI_MAX_REMAP]; + + /* Number of remapped NVMe devices found. */ + int num_remapped_devices; +}; + +static inline struct nvme_remap_dev *nrdev_from_bus(struct pci_bus *bus) +{ + return container_of(bus->sysdata, struct nvme_remap_dev, sysdata); +} + + +/******** PCI configuration space **********/ + +/* + * Helper macros for tweaking returned contents of PCI configuration space. + * + * value contains len bytes of data read from reg. + * If fixup_reg is included in that range, fix up the contents of that + * register to fixed_value. + */ +#define NR_FIX8(fixup_reg, fixed_value) do { \ + if (reg <= fixup_reg && fixup_reg < reg + len) \ + ((u8 *) value)[fixup_reg - reg] = (u8) (fixed_value); \ + } while (0) + +#define NR_FIX16(fixup_reg, fixed_value) do { \ + NR_FIX8(fixup_reg, fixed_value); \ + NR_FIX8(fixup_reg + 1, fixed_value >> 8); \ + } while (0) + +#define NR_FIX24(fixup_reg, fixed_value) do { \ + NR_FIX8(fixup_reg, fixed_value); \ + NR_FIX8(fixup_reg + 1, fixed_value >> 8); \ + NR_FIX8(fixup_reg + 2, fixed_value >> 16); \ + } while (0) + +#define NR_FIX32(fixup_reg, fixed_value) do { \ + NR_FIX16(fixup_reg, (u16) fixed_value); \ + NR_FIX16(fixup_reg + 2, fixed_value >> 16); \ + } while (0) + +/* + * Read PCI config space of the slot 0 (AHCI) device. + * We pass through the read request to the underlying device, but + * tweak the results in some cases. + */ +static int nvme_remap_pci_read_slot0(struct pci_bus *bus, int reg, + int len, u32 *value) +{ + struct nvme_remap_dev *nrdev = nrdev_from_bus(bus); + struct pci_bus *ahci_dev_bus = nrdev->dev->bus; + int ret; + + ret = ahci_dev_bus->ops->read(ahci_dev_bus, nrdev->dev->devfn, + reg, len, value); + if (ret) + return ret; + + /* + * Adjust the device class, to prevent this driver from attempting to + * additionally probe the device we're simulating here. + */ + NR_FIX24(PCI_CLASS_PROG, PCI_CLASS_STORAGE_SATA_AHCI); + + /* + * Unset interrupt pin, otherwise ACPI tries to find routing + * info for our virtual IRQ, fails, and complains. + */ + NR_FIX8(PCI_INTERRUPT_PIN, 0); + + /* + * Truncate the AHCI BAR to not include the region that covers the + * hidden devices. This will cause the ahci driver to successfully + * probe th new device (instead of handing it over to this driver). + */ + if (nrdev->bar_sizing) { + NR_FIX32(PCI_BASE_ADDRESS_5, ~(SZ_16K - 1)); + nrdev->bar_sizing = false; + } + + return PCIBIOS_SUCCESSFUL; +} + +/* + * Read PCI config space of a remapped device. + * Since the original PCI config space is inaccessible, we provide a minimal, + * fake config space instead. + */ +static int nvme_remap_pci_read_remapped(struct pci_bus *bus, unsigned int port, + int reg, int len, u32 *value) +{ + struct nvme_remap_dev *nrdev = nrdev_from_bus(bus); + struct resource *remapped_mem; + + if (port > nrdev->num_remapped_devices) + return PCIBIOS_DEVICE_NOT_FOUND; + + *value = 0; + remapped_mem = &nrdev->remapped_dev_mem[port - 1]; + + /* Set a Vendor ID, otherwise Linux assumes no device is present */ + NR_FIX16(PCI_VENDOR_ID, PCI_VENDOR_ID_INTEL); + + /* Always appear on & bus mastering */ + NR_FIX16(PCI_COMMAND, PCI_COMMAND_MEMORY | PCI_COMMAND_MASTER); + + /* Set class so that nvme driver probes us */ + NR_FIX24(PCI_CLASS_PROG, PCI_CLASS_STORAGE_EXPRESS); + + if (nrdev->bar_sizing) { + NR_FIX32(PCI_BASE_ADDRESS_0, + ~(resource_size(remapped_mem) - 1)); + nrdev->bar_sizing = false; + } else { + resource_size_t mem_start = remapped_mem->start; + + mem_start |= PCI_BASE_ADDRESS_MEM_TYPE_64; + NR_FIX32(PCI_BASE_ADDRESS_0, mem_start); + mem_start >>= 32; + NR_FIX32(PCI_BASE_ADDRESS_1, mem_start); + } + + return PCIBIOS_SUCCESSFUL; +} + +/* Read PCI configuration space. */ +static int nvme_remap_pci_read(struct pci_bus *bus, unsigned int devfn, + int reg, int len, u32 *value) +{ + if (PCI_SLOT(devfn) == 0) + return nvme_remap_pci_read_slot0(bus, reg, len, value); + else + return nvme_remap_pci_read_remapped(bus, PCI_SLOT(devfn), + reg, len, value); +} + +/* + * Write PCI config space of the slot 0 (AHCI) device. + * Apart from the special case of BAR sizing, we disable all writes. + * Otherwise, the ahci driver could make changes (e.g. unset PCI bus master) + * that would affect the operation of the NVMe devices. + */ +static int nvme_remap_pci_write_slot0(struct pci_bus *bus, int reg, + int len, u32 value) +{ + struct nvme_remap_dev *nrdev = nrdev_from_bus(bus); + struct pci_bus *ahci_dev_bus = nrdev->dev->bus; + + if (reg >= PCI_BASE_ADDRESS_0 && reg <= PCI_BASE_ADDRESS_5) { + /* + * Writing all-ones to a BAR means that the size of the + * memory region is being checked. Flag this so that we can + * reply with an appropriate size on the next read. + */ + if (value == ~0) + nrdev->bar_sizing = true; + + return ahci_dev_bus->ops->write(ahci_dev_bus, + nrdev->dev->devfn, + reg, len, value); + } + + return PCIBIOS_SET_FAILED; +} + +/* + * Write PCI config space of a remapped device. + * Since the original PCI config space is inaccessible, we reject all + * writes, except for the special case of BAR probing. + */ +static int nvme_remap_pci_write_remapped(struct pci_bus *bus, + unsigned int port, + int reg, int len, u32 value) +{ + struct nvme_remap_dev *nrdev = nrdev_from_bus(bus); + + if (port > nrdev->num_remapped_devices) + return PCIBIOS_DEVICE_NOT_FOUND; + + /* + * Writing all-ones to a BAR means that the size of the memory + * region is being checked. Flag this so that we can reply with + * an appropriate size on the next read. + */ + if (value == ~0 && reg >= PCI_BASE_ADDRESS_0 + && reg <= PCI_BASE_ADDRESS_5) { + nrdev->bar_sizing = true; + return PCIBIOS_SUCCESSFUL; + } + + return PCIBIOS_SET_FAILED; +} + +/* Write PCI configuration space. */ +static int nvme_remap_pci_write(struct pci_bus *bus, unsigned int devfn, + int reg, int len, u32 value) +{ + if (PCI_SLOT(devfn) == 0) + return nvme_remap_pci_write_slot0(bus, reg, len, value); + else + return nvme_remap_pci_write_remapped(bus, PCI_SLOT(devfn), + reg, len, value); +} + +static struct pci_ops nvme_remap_pci_ops = { + .read = nvme_remap_pci_read, + .write = nvme_remap_pci_write, +}; + + +/******** Initialization & exit **********/ + +/* + * Find a PCI domain ID to use for our fake bus. + * Start at 0x10000 to not clash with ACPI _SEG domains (16 bits). + */ +static int find_free_domain(void) +{ + int domain = 0xffff; + struct pci_bus *bus = NULL; + + while ((bus = pci_find_next_bus(bus)) != NULL) + domain = max_t(int, domain, pci_domain_nr(bus)); + + return domain + 1; +} + +static int find_remapped_devices(struct nvme_remap_dev *nrdev, + struct list_head *resources) +{ + void __iomem *mmio; + int i, count = 0; + u32 cap; + + mmio = pcim_iomap(nrdev->dev, AHCI_PCI_BAR_STANDARD, + pci_resource_len(nrdev->dev, + AHCI_PCI_BAR_STANDARD)); + if (!mmio) + return -ENODEV; + + /* Check if this device might have remapped nvme devices. */ + if (pci_resource_len(nrdev->dev, AHCI_PCI_BAR_STANDARD) < SZ_512K || + !(readl(mmio + AHCI_VSCAP) & 1)) + return -ENODEV; + + cap = readq(mmio + AHCI_REMAP_CAP); + for (i = AHCI_MAX_REMAP-1; i >= 0; i--) { + struct resource *remapped_mem; + + if ((cap & (1 << i)) == 0) + continue; + if (readl(mmio + ahci_remap_dcc(i)) + != PCI_CLASS_STORAGE_EXPRESS) + continue; + + /* We've found a remapped device */ + remapped_mem = &nrdev->remapped_dev_mem[count++]; + remapped_mem->start = + pci_resource_start(nrdev->dev, AHCI_PCI_BAR_STANDARD) + + ahci_remap_base(i); + remapped_mem->end = remapped_mem->start + + AHCI_REMAP_N_SIZE - 1; + remapped_mem->flags = IORESOURCE_MEM | IORESOURCE_PCI_FIXED; + pci_add_resource(resources, remapped_mem); + } + + pcim_iounmap(nrdev->dev, mmio); + + if (count == 0) + return -ENODEV; + + nrdev->num_remapped_devices = count; + dev_info(&nrdev->dev->dev, "Found %d remapped NVMe devices\n", + nrdev->num_remapped_devices); + return 0; +} + +static void nvme_remap_remove_root_bus(void *data) +{ + struct pci_bus *bus = data; + + pci_stop_root_bus(bus); + pci_remove_root_bus(bus); +} + +static int nvme_remap_probe(struct pci_dev *dev, + const struct pci_device_id *id) +{ + struct nvme_remap_dev *nrdev; + LIST_HEAD(resources); + int i; + int ret; + struct pci_dev *child; + + nrdev = devm_kzalloc(&dev->dev, sizeof(*nrdev), GFP_KERNEL); + nrdev->sysdata.domain = find_free_domain(); + nrdev->sysdata.nvme_remap_dev = dev; + nrdev->dev = dev; + pci_set_drvdata(dev, nrdev); + + ret = pcim_enable_device(dev); + if (ret < 0) + return ret; + + pci_set_master(dev); + + ret = find_remapped_devices(nrdev, &resources); + if (ret) + return ret; + + /* Add resources from the original AHCI device */ + for (i = 0; i < PCI_NUM_RESOURCES; i++) { + struct resource *res = &dev->resource[i]; + + if (res->start) { + struct resource *nr_res = &nrdev->ahci_resources[i]; + + nr_res->start = res->start; + nr_res->end = res->end; + nr_res->flags = res->flags; + pci_add_resource(&resources, nr_res); + } + } + + /* Create virtual interrupts */ + nrdev->irq_base = devm_irq_alloc_descs(&dev->dev, -1, 0, + nrdev->num_remapped_devices + 1, + 0); + if (nrdev->irq_base < 0) + return nrdev->irq_base; + + /* Create and populate PCI bus */ + nrdev->bus = pci_create_root_bus(&dev->dev, 0, &nvme_remap_pci_ops, + &nrdev->sysdata, &resources); + if (!nrdev->bus) + return -ENODEV; + + if (devm_add_action_or_reset(&dev->dev, nvme_remap_remove_root_bus, + nrdev->bus)) + return -ENOMEM; + + /* We don't support sharing MSI interrupts between these devices */ + nrdev->bus->bus_flags |= PCI_BUS_FLAGS_NO_MSI; + + pci_scan_child_bus(nrdev->bus); + + list_for_each_entry(child, &nrdev->bus->devices, bus_list) { + /* + * Prevent PCI core from trying to move memory BARs around. + * The hidden NVMe devices are at fixed locations. + */ + for (i = 0; i < PCI_NUM_RESOURCES; i++) { + struct resource *res = &child->resource[i]; + + if (res->flags & IORESOURCE_MEM) + res->flags |= IORESOURCE_PCI_FIXED; + } + + /* Share the legacy IRQ between all devices */ + child->irq = dev->irq; + } + + pci_assign_unassigned_bus_resources(nrdev->bus); + pci_bus_add_devices(nrdev->bus); + + return 0; +} + +static const struct pci_device_id nvme_remap_ids[] = { + /* + * Match all Intel RAID controllers. + * + * There's overlap here with the set of devices detected by the ahci + * driver, but ahci will only successfully probe when there + * *aren't* any remapped NVMe devices, and this driver will only + * successfully probe when there *are* remapped NVMe devices that + * need handling. + */ + { + PCI_VDEVICE(INTEL, PCI_ANY_ID), + .class = PCI_CLASS_STORAGE_RAID << 8, + .class_mask = 0xffffff00, + }, + {0,} +}; +MODULE_DEVICE_TABLE(pci, nvme_remap_ids); + +static struct pci_driver nvme_remap_drv = { + .name = MODULE_NAME, + .id_table = nvme_remap_ids, + .probe = nvme_remap_probe, +}; +module_pci_driver(nvme_remap_drv); + +MODULE_AUTHOR("Daniel Drake "); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/pci/quirks.c b/drivers/pci/quirks.c index 6d74386eadc..98887326a61 100644 --- a/drivers/pci/quirks.c +++ b/drivers/pci/quirks.c @@ -3547,6 +3547,106 @@ static void quirk_no_bus_reset(struct pci_dev *dev) dev->dev_flags |= PCI_DEV_FLAGS_NO_BUS_RESET; } +static bool acs_on_downstream; +static bool acs_on_multifunction; + +#define NUM_ACS_IDS 16 +struct acs_on_id { + unsigned short vendor; + unsigned short device; +}; +static struct acs_on_id acs_on_ids[NUM_ACS_IDS]; +static u8 max_acs_id; + +static __init int pcie_acs_override_setup(char *p) +{ + if (!p) + return -EINVAL; + + while (*p) { + if (!strncmp(p, "downstream", 10)) + acs_on_downstream = true; + if (!strncmp(p, "multifunction", 13)) + acs_on_multifunction = true; + if (!strncmp(p, "id:", 3)) { + char opt[5]; + int ret; + long val; + + if (max_acs_id >= NUM_ACS_IDS - 1) { + pr_warn("Out of PCIe ACS override slots (%d)\n", + NUM_ACS_IDS); + goto next; + } + + p += 3; + snprintf(opt, 5, "%s", p); + ret = kstrtol(opt, 16, &val); + if (ret) { + pr_warn("PCIe ACS ID parse error %d\n", ret); + goto next; + } + acs_on_ids[max_acs_id].vendor = val; + + p += strcspn(p, ":"); + if (*p != ':') { + pr_warn("PCIe ACS invalid ID\n"); + goto next; + } + + p++; + snprintf(opt, 5, "%s", p); + ret = kstrtol(opt, 16, &val); + if (ret) { + pr_warn("PCIe ACS ID parse error %d\n", ret); + goto next; + } + acs_on_ids[max_acs_id].device = val; + max_acs_id++; + } +next: + p += strcspn(p, ","); + if (*p == ',') + p++; + } + + if (acs_on_downstream || acs_on_multifunction || max_acs_id) + pr_warn("Warning: PCIe ACS overrides enabled; This may allow non-IOMMU protected peer-to-peer DMA\n"); + + return 0; +} +early_param("pcie_acs_override", pcie_acs_override_setup); + +static int pcie_acs_overrides(struct pci_dev *dev, u16 acs_flags) +{ + int i; + + /* Never override ACS for legacy devices or devices with ACS caps */ + if (!pci_is_pcie(dev) || + pci_find_ext_capability(dev, PCI_EXT_CAP_ID_ACS)) + return -ENOTTY; + + for (i = 0; i < max_acs_id; i++) + if (acs_on_ids[i].vendor == dev->vendor && + acs_on_ids[i].device == dev->device) + return 1; + + switch (pci_pcie_type(dev)) { + case PCI_EXP_TYPE_DOWNSTREAM: + case PCI_EXP_TYPE_ROOT_PORT: + if (acs_on_downstream) + return 1; + break; + case PCI_EXP_TYPE_ENDPOINT: + case PCI_EXP_TYPE_UPSTREAM: + case PCI_EXP_TYPE_LEG_END: + case PCI_EXP_TYPE_RC_END: + if (acs_on_multifunction && dev->multifunction) + return 1; + } + + return -ENOTTY; +} /* * Some NVIDIA GPU devices do not work with bus reset, SBR needs to be * prevented for those affected devices. @@ -4862,6 +4962,7 @@ static const struct pci_dev_acs_enabled { { PCI_VENDOR_ID_ZHAOXIN, 0x9083, pci_quirk_mf_endpoint_acs }, /* Zhaoxin Root/Downstream Ports */ { PCI_VENDOR_ID_ZHAOXIN, PCI_ANY_ID, pci_quirk_zhaoxin_pcie_ports_acs }, + { PCI_ANY_ID, PCI_ANY_ID, pcie_acs_overrides }, { 0 } }; diff --git a/drivers/platform/x86/intel_ips.c b/drivers/platform/x86/intel_ips.c index bffe548187e..c2918ee3e10 100644 --- a/drivers/platform/x86/intel_ips.c +++ b/drivers/platform/x86/intel_ips.c @@ -798,7 +798,7 @@ static int ips_adjust(void *data) ips_gpu_lower(ips); sleep: - schedule_timeout_interruptible(msecs_to_jiffies(IPS_ADJUST_PERIOD)); + schedule_msec_hrtimeout_interruptible((IPS_ADJUST_PERIOD)); } while (!kthread_should_stop()); dev_dbg(ips->dev, "ips-adjust thread stopped\n"); @@ -974,7 +974,7 @@ static int ips_monitor(void *data) seqno_timestamp = get_jiffies_64(); old_cpu_power = thm_readl(THM_CEC); - schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD)); + schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD)); /* Collect an initial average */ for (i = 0; i < IPS_SAMPLE_COUNT; i++) { @@ -1001,7 +1001,7 @@ static int ips_monitor(void *data) mchp_samples[i] = mchp; } - schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD)); + schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD)); if (kthread_should_stop()) break; } @@ -1028,7 +1028,7 @@ static int ips_monitor(void *data) * us to reduce the sample frequency if the CPU and GPU are idle. */ old_cpu_power = thm_readl(THM_CEC); - schedule_timeout_interruptible(msecs_to_jiffies(IPS_SAMPLE_PERIOD)); + schedule_msec_hrtimeout_interruptible((IPS_SAMPLE_PERIOD)); last_sample_period = IPS_SAMPLE_PERIOD; timer_setup(&ips->timer, monitor_timeout, TIMER_DEFERRABLE); diff --git a/drivers/rtc/rtc-wm8350.c b/drivers/rtc/rtc-wm8350.c index 2018614f258..fc19b312c34 100644 --- a/drivers/rtc/rtc-wm8350.c +++ b/drivers/rtc/rtc-wm8350.c @@ -114,7 +114,7 @@ static int wm8350_rtc_settime(struct device *dev, struct rtc_time *tm) /* Wait until confirmation of stopping */ do { rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL); - schedule_timeout_uninterruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_uninterruptible((1)); } while (--retries && !(rtc_ctrl & WM8350_RTC_STS)); if (!retries) { @@ -197,7 +197,7 @@ static int wm8350_rtc_stop_alarm(struct wm8350 *wm8350) /* Wait until confirmation of stopping */ do { rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL); - schedule_timeout_uninterruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_uninterruptible((1)); } while (retries-- && !(rtc_ctrl & WM8350_RTC_ALMSTS)); if (!(rtc_ctrl & WM8350_RTC_ALMSTS)) @@ -220,7 +220,7 @@ static int wm8350_rtc_start_alarm(struct wm8350 *wm8350) /* Wait until confirmation */ do { rtc_ctrl = wm8350_reg_read(wm8350, WM8350_RTC_TIME_CONTROL); - schedule_timeout_uninterruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_uninterruptible((1)); } while (retries-- && rtc_ctrl & WM8350_RTC_ALMSTS); if (rtc_ctrl & WM8350_RTC_ALMSTS) diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig index 3d114be5b66..10a710cb3ee 100644 --- a/drivers/scsi/Kconfig +++ b/drivers/scsi/Kconfig @@ -1516,4 +1516,6 @@ endif # SCSI_LOWLEVEL source "drivers/scsi/device_handler/Kconfig" +source "drivers/scsi/vhba/Kconfig" + endmenu diff --git a/drivers/scsi/Makefile b/drivers/scsi/Makefile index bc3882f5cc6..dcc9518e846 100644 --- a/drivers/scsi/Makefile +++ b/drivers/scsi/Makefile @@ -151,6 +151,7 @@ obj-$(CONFIG_CHR_DEV_SCH) += ch.o obj-$(CONFIG_SCSI_ENCLOSURE) += ses.o obj-$(CONFIG_SCSI_HISI_SAS) += hisi_sas/ +obj-$(CONFIG_VHBA) += vhba/ # This goes last, so that "real" scsi devices probe earlier obj-$(CONFIG_SCSI_DEBUG) += scsi_debug.o diff --git a/drivers/scsi/fnic/fnic_scsi.c b/drivers/scsi/fnic/fnic_scsi.c index 762cc8bd265..75c557c08ac 100644 --- a/drivers/scsi/fnic/fnic_scsi.c +++ b/drivers/scsi/fnic/fnic_scsi.c @@ -217,7 +217,7 @@ int fnic_fw_reset_handler(struct fnic *fnic) /* wait for io cmpl */ while (atomic_read(&fnic->in_flight)) - schedule_timeout(msecs_to_jiffies(1)); + schedule_msec_hrtimeout((1)); spin_lock_irqsave(&fnic->wq_copy_lock[0], flags); @@ -1367,13 +1367,16 @@ static bool fnic_cleanup_io_iter(struct scsi_cmnd *sc, void *data, struct fnic *fnic = data; struct fnic_io_req *io_req; unsigned long flags = 0; + spinlock_t *io_lock; unsigned long start_time = 0; struct fnic_stats *fnic_stats = &fnic->fnic_stats; io_lock = fnic_io_lock_tag(fnic, sc->request->tag); + spin_lock_irqsave(io_lock, flags); + io_req = (struct fnic_io_req *)CMD_SP(sc); if ((CMD_FLAGS(sc) & FNIC_DEVICE_RESET) && !(CMD_FLAGS(sc) & FNIC_DEV_RST_DONE)) { @@ -1415,6 +1418,7 @@ static bool fnic_cleanup_io_iter(struct scsi_cmnd *sc, void *data, "fnic_cleanup_io: tag:0x%x : sc:0x%p duration = %lu DID_TRANSPORT_DISRUPTED\n", sc->request->tag, sc, (jiffies - start_time)); + if (atomic64_read(&fnic->io_cmpl_skip)) atomic64_dec(&fnic->io_cmpl_skip); else @@ -1570,14 +1574,17 @@ static bool fnic_rport_abort_io_iter(struct scsi_cmnd *sc, void *data, struct fnic_io_req *io_req; spinlock_t *io_lock; unsigned long flags; + struct reset_stats *reset_stats = &fnic->fnic_stats.reset_stats; struct terminate_stats *term_stats = &fnic->fnic_stats.term_stats; struct scsi_lun fc_lun; enum fnic_ioreq_state old_ioreq_state; io_lock = fnic_io_lock_tag(fnic, abt_tag); + spin_lock_irqsave(io_lock, flags); + io_req = (struct fnic_io_req *)CMD_SP(sc); if (!io_req || io_req->port_id != iter_data->port_id) { @@ -1622,6 +1629,7 @@ static bool fnic_rport_abort_io_iter(struct scsi_cmnd *sc, void *data, if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) { atomic64_inc(&reset_stats->device_reset_terminates); abt_tag |= FNIC_TAG_DEV_RST; + } FNIC_SCSI_DBG(KERN_DEBUG, fnic->lport->host, "fnic_rport_exch_reset dev rst sc 0x%p\n", sc); @@ -1687,10 +1695,12 @@ static void fnic_rport_exch_reset(struct fnic *fnic, u32 port_id) void fnic_terminate_rport_io(struct fc_rport *rport) { + struct fc_rport_libfc_priv *rdata; struct fc_lport *lport; struct fnic *fnic; + if (!rport) { printk(KERN_ERR "fnic_terminate_rport_io: rport is NULL\n"); return; @@ -1718,6 +1728,7 @@ void fnic_terminate_rport_io(struct fc_rport *rport) return; fnic_rport_exch_reset(fnic, rport->port_id); + } /* @@ -2019,6 +2030,7 @@ struct fnic_pending_aborts_iter_data { int ret; }; + static bool fnic_pending_aborts_iter(struct scsi_cmnd *sc, void *data, bool reserved) { @@ -2029,7 +2041,9 @@ static bool fnic_pending_aborts_iter(struct scsi_cmnd *sc, struct fnic_io_req *io_req; spinlock_t *io_lock; unsigned long flags; + struct scsi_lun fc_lun; + DECLARE_COMPLETION_ONSTACK(tm_done); enum fnic_ioreq_state old_ioreq_state; @@ -2038,10 +2052,12 @@ static bool fnic_pending_aborts_iter(struct scsi_cmnd *sc, if (reserved) return true; + io_lock = fnic_io_lock_tag(fnic, abt_tag); spin_lock_irqsave(io_lock, flags); io_req = (struct fnic_io_req *)CMD_SP(sc); if (!io_req) { + spin_unlock_irqrestore(io_lock, flags); return true; } @@ -2083,6 +2099,7 @@ static bool fnic_pending_aborts_iter(struct scsi_cmnd *sc, BUG_ON(io_req->abts_done); + if (CMD_FLAGS(sc) & FNIC_DEVICE_RESET) { abt_tag |= FNIC_TAG_DEV_RST; FNIC_SCSI_DBG(KERN_INFO, fnic->lport->host, @@ -2117,6 +2134,7 @@ static bool fnic_pending_aborts_iter(struct scsi_cmnd *sc, CMD_FLAGS(sc) |= FNIC_IO_INTERNAL_TERM_ISSUED; wait_for_completion_timeout(&tm_done, msecs_to_jiffies + (fnic->config.ed_tov)); /* Recheck cmd state to check if it is now aborted */ @@ -2189,7 +2207,7 @@ static int fnic_clean_pending_aborts(struct fnic *fnic, ret = iter_data.ret; goto clean_pending_aborts_end; } - schedule_timeout(msecs_to_jiffies(2 * fnic->config.ed_tov)); + schedule_msec_hrtimeout((2 * fnic->config.ed_tov)); /* walk again to check, if IOs are still pending in fw */ if (fnic_is_abts_pending(fnic, lr_sc)) @@ -2689,6 +2707,7 @@ void fnic_exch_mgr_reset(struct fc_lport *lp, u32 sid, u32 did) static bool fnic_abts_pending_iter(struct scsi_cmnd *sc, void *data, bool reserved) + { struct fnic_pending_aborts_iter_data *iter_data = data; struct fnic *fnic = iter_data->fnic; @@ -2697,6 +2716,8 @@ static bool fnic_abts_pending_iter(struct scsi_cmnd *sc, void *data, spinlock_t *io_lock; unsigned long flags; + + /* * ignore this lun reset cmd or cmds that do not belong to * this lun @@ -2711,6 +2732,7 @@ static bool fnic_abts_pending_iter(struct scsi_cmnd *sc, void *data, io_req = (struct fnic_io_req *)CMD_SP(sc); if (!io_req) { + spin_unlock_irqrestore(io_lock, flags); return true; } @@ -2723,6 +2745,7 @@ static bool fnic_abts_pending_iter(struct scsi_cmnd *sc, void *data, "Found IO in %s on lun\n", fnic_ioreq_state_to_str(CMD_STATE(sc))); cmd_state = CMD_STATE(sc); + spin_unlock_irqrestore(io_lock, flags); if (cmd_state == FNIC_IOREQ_ABTS_PENDING) iter_data->ret = 1; @@ -2755,4 +2778,4 @@ int fnic_is_abts_pending(struct fnic *fnic, struct scsi_cmnd *lr_sc) fnic_abts_pending_iter, &iter_data); return iter_data.ret; -} +} \ No newline at end of file diff --git a/drivers/scsi/lpfc/lpfc_scsi.c b/drivers/scsi/lpfc/lpfc_scsi.c index eefbb9b2279..ab7d7869dd1 100644 --- a/drivers/scsi/lpfc/lpfc_scsi.c +++ b/drivers/scsi/lpfc/lpfc_scsi.c @@ -5829,7 +5829,7 @@ lpfc_reset_flush_io_context(struct lpfc_vport *vport, uint16_t tgt_id, tgt_id, lun_id, context); later = msecs_to_jiffies(2 * vport->cfg_devloss_tmo * 1000) + jiffies; while (time_after(later, jiffies) && cnt) { - schedule_timeout_uninterruptible(msecs_to_jiffies(20)); + schedule_msec_hrtimeout_uninterruptible((20)); cnt = lpfc_sli_sum_iocb(vport, tgt_id, lun_id, context); } if (cnt) { diff --git a/drivers/scsi/snic/snic_scsi.c b/drivers/scsi/snic/snic_scsi.c index 6dd0ff188bb..aedf0b78f62 100644 --- a/drivers/scsi/snic/snic_scsi.c +++ b/drivers/scsi/snic/snic_scsi.c @@ -2349,7 +2349,7 @@ snic_reset(struct Scsi_Host *shost, struct scsi_cmnd *sc) /* Wait for all the IOs that are entered in Qcmd */ while (atomic_read(&snic->ios_inflight)) - schedule_timeout(msecs_to_jiffies(1)); + schedule_msec_hrtimeout((1)); ret = snic_issue_hba_reset(snic, sc); if (ret) { diff --git a/drivers/scsi/vhba/Kconfig b/drivers/scsi/vhba/Kconfig new file mode 100644 index 00000000000..e70a381fe3d --- /dev/null +++ b/drivers/scsi/vhba/Kconfig @@ -0,0 +1,9 @@ +config VHBA + tristate "Virtual (SCSI) Host Bus Adapter" + depends on SCSI + help + This is the in-kernel part of CDEmu, a CD/DVD-ROM device + emulator. + + This driver can also be built as a module. If so, the module + will be called vhba. diff --git a/drivers/scsi/vhba/Makefile b/drivers/scsi/vhba/Makefile new file mode 100644 index 00000000000..b31e07b6891 --- /dev/null +++ b/drivers/scsi/vhba/Makefile @@ -0,0 +1,4 @@ +VHBA_VERSION := 20210418 + +obj-$(CONFIG_VHBA) += vhba.o +ccflags-y := -DVHBA_VERSION=\"$(VHBA_VERSION)\" -Werror diff --git a/drivers/scsi/vhba/vhba.c b/drivers/scsi/vhba/vhba.c new file mode 100644 index 00000000000..574ea08ce0d --- /dev/null +++ b/drivers/scsi/vhba/vhba.c @@ -0,0 +1,1105 @@ +/* + * vhba.c + * + * Copyright (C) 2007-2012 Chia-I Wu + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License along + * with this program; if not, write to the Free Software Foundation, Inc., + * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. + */ + +#define pr_fmt(fmt) "vhba: " fmt + +#include + +#include +#include +#include +#include +#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 11, 0) +#include +#else +#include +#endif +#include +#include +#include +#include +#include +#ifdef CONFIG_COMPAT +#include +#endif +#include +#include +#include +#include +#include +#include + + +MODULE_AUTHOR("Chia-I Wu"); +MODULE_VERSION(VHBA_VERSION); +MODULE_DESCRIPTION("Virtual SCSI HBA"); +MODULE_LICENSE("GPL"); + + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 15, 0) +#define sdev_dbg(sdev, fmt, a...) \ + dev_dbg(&(sdev)->sdev_gendev, fmt, ##a) +#define scmd_dbg(scmd, fmt, a...) \ + dev_dbg(&(scmd)->device->sdev_gendev, fmt, ##a) +#endif + +#define VHBA_MAX_SECTORS_PER_IO 256 +#define VHBA_MAX_BUS 16 +#define VHBA_MAX_ID 16 +#define VHBA_MAX_DEVICES (VHBA_MAX_BUS * (VHBA_MAX_ID-1)) +#define VHBA_KBUF_SIZE PAGE_SIZE + +#define DATA_TO_DEVICE(dir) ((dir) == DMA_TO_DEVICE || (dir) == DMA_BIDIRECTIONAL) +#define DATA_FROM_DEVICE(dir) ((dir) == DMA_FROM_DEVICE || (dir) == DMA_BIDIRECTIONAL) + + +static int vhba_can_queue = 32; +module_param_named(can_queue, vhba_can_queue, int, 0); + + +enum vhba_req_state { + VHBA_REQ_FREE, + VHBA_REQ_PENDING, + VHBA_REQ_READING, + VHBA_REQ_SENT, + VHBA_REQ_WRITING, +}; + +struct vhba_command { + struct scsi_cmnd *cmd; + /* metatags are per-host. not to be confused with + queue tags that are usually per-lun */ + unsigned long metatag; + int status; + struct list_head entry; +}; + +struct vhba_device { + unsigned int num; + spinlock_t cmd_lock; + struct list_head cmd_list; + wait_queue_head_t cmd_wq; + atomic_t refcnt; + + unsigned char *kbuf; + size_t kbuf_size; +}; + +struct vhba_host { + struct Scsi_Host *shost; + spinlock_t cmd_lock; + int cmd_next; + struct vhba_command *commands; + spinlock_t dev_lock; + struct vhba_device *devices[VHBA_MAX_DEVICES]; + int num_devices; + DECLARE_BITMAP(chgmap, VHBA_MAX_DEVICES); + int chgtype[VHBA_MAX_DEVICES]; + struct work_struct scan_devices; +}; + +#define MAX_COMMAND_SIZE 16 + +struct vhba_request { + __u32 metatag; + __u32 lun; + __u8 cdb[MAX_COMMAND_SIZE]; + __u8 cdb_len; + __u32 data_len; +}; + +struct vhba_response { + __u32 metatag; + __u32 status; + __u32 data_len; +}; + + + +struct vhba_command *vhba_alloc_command (void); +void vhba_free_command (struct vhba_command *vcmd); + +static struct platform_device vhba_platform_device; + + + +/* These functions define a symmetric 1:1 mapping between device numbers and + the bus and id. We have reserved the last id per bus for the host itself. */ +void devnum_to_bus_and_id(unsigned int devnum, unsigned int *bus, unsigned int *id) +{ + *bus = devnum / (VHBA_MAX_ID-1); + *id = devnum % (VHBA_MAX_ID-1); +} + +unsigned int bus_and_id_to_devnum(unsigned int bus, unsigned int id) +{ + return (bus * (VHBA_MAX_ID-1)) + id; +} + +struct vhba_device *vhba_device_alloc (void) +{ + struct vhba_device *vdev; + + vdev = kzalloc(sizeof(struct vhba_device), GFP_KERNEL); + if (!vdev) { + return NULL; + } + + spin_lock_init(&vdev->cmd_lock); + INIT_LIST_HEAD(&vdev->cmd_list); + init_waitqueue_head(&vdev->cmd_wq); + atomic_set(&vdev->refcnt, 1); + + vdev->kbuf = NULL; + vdev->kbuf_size = 0; + + return vdev; +} + +void vhba_device_put (struct vhba_device *vdev) +{ + if (atomic_dec_and_test(&vdev->refcnt)) { + kfree(vdev); + } +} + +struct vhba_device *vhba_device_get (struct vhba_device *vdev) +{ + atomic_inc(&vdev->refcnt); + + return vdev; +} + +int vhba_device_queue (struct vhba_device *vdev, struct scsi_cmnd *cmd) +{ + struct vhba_host *vhost; + struct vhba_command *vcmd; + unsigned long flags; + + vhost = platform_get_drvdata(&vhba_platform_device); + + vcmd = vhba_alloc_command(); + if (!vcmd) { + return SCSI_MLQUEUE_HOST_BUSY; + } + + vcmd->cmd = cmd; + + spin_lock_irqsave(&vdev->cmd_lock, flags); + vcmd->metatag = vcmd->cmd->request->tag; + list_add_tail(&vcmd->entry, &vdev->cmd_list); + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + wake_up_interruptible(&vdev->cmd_wq); + + return 0; +} + +int vhba_device_dequeue (struct vhba_device *vdev, struct scsi_cmnd *cmd) +{ + struct vhba_command *vcmd; + int retval; + unsigned long flags; + + spin_lock_irqsave(&vdev->cmd_lock, flags); + list_for_each_entry(vcmd, &vdev->cmd_list, entry) { + if (vcmd->cmd == cmd) { + list_del_init(&vcmd->entry); + break; + } + } + + /* command not found */ + if (&vcmd->entry == &vdev->cmd_list) { + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + return SUCCESS; + } + + while (vcmd->status == VHBA_REQ_READING || vcmd->status == VHBA_REQ_WRITING) { + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + scmd_dbg(cmd, "wait for I/O before aborting\n"); + schedule_timeout(1); + spin_lock_irqsave(&vdev->cmd_lock, flags); + } + + retval = (vcmd->status == VHBA_REQ_SENT) ? FAILED : SUCCESS; + + vhba_free_command(vcmd); + + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + return retval; +} + +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 19, 0) +int vhba_slave_alloc(struct scsi_device *sdev) +{ + struct Scsi_Host *shost = sdev->host; + + sdev_dbg(sdev, "enabling tagging (queue depth: %i).\n", sdev->queue_depth); +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 17, 0) + if (!shost_use_blk_mq(shost) && shost->bqt) { +#else + if (shost->bqt) { +#endif + blk_queue_init_tags(sdev->request_queue, sdev->queue_depth, shost->bqt); + } + scsi_adjust_queue_depth(sdev, 0, sdev->queue_depth); + + return 0; +} +#endif + +void vhba_scan_devices_add (struct vhba_host *vhost, int bus, int id) +{ + struct scsi_device *sdev; + + sdev = scsi_device_lookup(vhost->shost, bus, id, 0); + if (!sdev) { + scsi_add_device(vhost->shost, bus, id, 0); + } else { + dev_warn(&vhost->shost->shost_gendev, "tried to add an already-existing device %d:%d:0!\n", bus, id); + scsi_device_put(sdev); + } +} + +void vhba_scan_devices_remove (struct vhba_host *vhost, int bus, int id) +{ + struct scsi_device *sdev; + + sdev = scsi_device_lookup(vhost->shost, bus, id, 0); + if (sdev) { + scsi_remove_device(sdev); + scsi_device_put(sdev); + } else { + dev_warn(&vhost->shost->shost_gendev, "tried to remove non-existing device %d:%d:0!\n", bus, id); + } +} + +void vhba_scan_devices (struct work_struct *work) +{ + struct vhba_host *vhost = container_of(work, struct vhba_host, scan_devices); + unsigned long flags; + int change, exists; + unsigned int devnum; + unsigned int bus, id; + + for (;;) { + spin_lock_irqsave(&vhost->dev_lock, flags); + + devnum = find_first_bit(vhost->chgmap, VHBA_MAX_DEVICES); + if (devnum >= VHBA_MAX_DEVICES) { + spin_unlock_irqrestore(&vhost->dev_lock, flags); + break; + } + change = vhost->chgtype[devnum]; + exists = vhost->devices[devnum] != NULL; + + vhost->chgtype[devnum] = 0; + clear_bit(devnum, vhost->chgmap); + + spin_unlock_irqrestore(&vhost->dev_lock, flags); + + devnum_to_bus_and_id(devnum, &bus, &id); + + if (change < 0) { + dev_dbg(&vhost->shost->shost_gendev, "trying to remove target %d:%d:0\n", bus, id); + vhba_scan_devices_remove(vhost, bus, id); + } else if (change > 0) { + dev_dbg(&vhost->shost->shost_gendev, "trying to add target %d:%d:0\n", bus, id); + vhba_scan_devices_add(vhost, bus, id); + } else { + /* quick sequence of add/remove or remove/add; we determine + which one it was by checking if device structure exists */ + if (exists) { + /* remove followed by add: remove and (re)add */ + dev_dbg(&vhost->shost->shost_gendev, "trying to (re)add target %d:%d:0\n", bus, id); + vhba_scan_devices_remove(vhost, bus, id); + vhba_scan_devices_add(vhost, bus, id); + } else { + /* add followed by remove: no-op */ + dev_dbg(&vhost->shost->shost_gendev, "no-op for target %d:%d:0\n", bus, id); + } + } + } +} + +int vhba_add_device (struct vhba_device *vdev) +{ + struct vhba_host *vhost; + unsigned int devnum; + unsigned long flags; + + vhost = platform_get_drvdata(&vhba_platform_device); + + vhba_device_get(vdev); + + spin_lock_irqsave(&vhost->dev_lock, flags); + if (vhost->num_devices >= VHBA_MAX_DEVICES) { + spin_unlock_irqrestore(&vhost->dev_lock, flags); + vhba_device_put(vdev); + return -EBUSY; + } + + for (devnum = 0; devnum < VHBA_MAX_DEVICES; devnum++) { + if (vhost->devices[devnum] == NULL) { + vdev->num = devnum; + vhost->devices[devnum] = vdev; + vhost->num_devices++; + set_bit(devnum, vhost->chgmap); + vhost->chgtype[devnum]++; + break; + } + } + spin_unlock_irqrestore(&vhost->dev_lock, flags); + + schedule_work(&vhost->scan_devices); + + return 0; +} + +int vhba_remove_device (struct vhba_device *vdev) +{ + struct vhba_host *vhost; + unsigned long flags; + + vhost = platform_get_drvdata(&vhba_platform_device); + + spin_lock_irqsave(&vhost->dev_lock, flags); + set_bit(vdev->num, vhost->chgmap); + vhost->chgtype[vdev->num]--; + vhost->devices[vdev->num] = NULL; + vhost->num_devices--; + spin_unlock_irqrestore(&vhost->dev_lock, flags); + + vhba_device_put(vdev); + + schedule_work(&vhost->scan_devices); + + return 0; +} + +struct vhba_device *vhba_lookup_device (int devnum) +{ + struct vhba_host *vhost; + struct vhba_device *vdev = NULL; + unsigned long flags; + + vhost = platform_get_drvdata(&vhba_platform_device); + + if (likely(devnum < VHBA_MAX_DEVICES)) { + spin_lock_irqsave(&vhost->dev_lock, flags); + vdev = vhost->devices[devnum]; + if (vdev) { + vdev = vhba_device_get(vdev); + } + + spin_unlock_irqrestore(&vhost->dev_lock, flags); + } + + return vdev; +} + +struct vhba_command *vhba_alloc_command (void) +{ + struct vhba_host *vhost; + struct vhba_command *vcmd; + unsigned long flags; + int i; + + vhost = platform_get_drvdata(&vhba_platform_device); + + spin_lock_irqsave(&vhost->cmd_lock, flags); + + vcmd = vhost->commands + vhost->cmd_next++; + if (vcmd->status != VHBA_REQ_FREE) { + for (i = 0; i < vhba_can_queue; i++) { + vcmd = vhost->commands + i; + + if (vcmd->status == VHBA_REQ_FREE) { + vhost->cmd_next = i + 1; + break; + } + } + + if (i == vhba_can_queue) { + vcmd = NULL; + } + } + + if (vcmd) { + vcmd->status = VHBA_REQ_PENDING; + } + + vhost->cmd_next %= vhba_can_queue; + + spin_unlock_irqrestore(&vhost->cmd_lock, flags); + + return vcmd; +} + +void vhba_free_command (struct vhba_command *vcmd) +{ + struct vhba_host *vhost; + unsigned long flags; + + vhost = platform_get_drvdata(&vhba_platform_device); + + spin_lock_irqsave(&vhost->cmd_lock, flags); + vcmd->status = VHBA_REQ_FREE; + spin_unlock_irqrestore(&vhost->cmd_lock, flags); +} + +int vhba_queuecommand (struct Scsi_Host *shost, struct scsi_cmnd *cmd) +{ + struct vhba_device *vdev; + int retval; + unsigned int devnum; + + scmd_dbg(cmd, "queue %p tag %i\n", cmd, cmd->request->tag); + + devnum = bus_and_id_to_devnum(cmd->device->channel, cmd->device->id); + vdev = vhba_lookup_device(devnum); + if (!vdev) { + scmd_dbg(cmd, "no such device\n"); + + cmd->result = DID_NO_CONNECT << 16; + cmd->scsi_done(cmd); + + return 0; + } + + retval = vhba_device_queue(vdev, cmd); + + vhba_device_put(vdev); + + return retval; +} + +int vhba_abort (struct scsi_cmnd *cmd) +{ + struct vhba_device *vdev; + int retval = SUCCESS; + unsigned int devnum; + + scmd_dbg(cmd, "abort %p\n", cmd); + + devnum = bus_and_id_to_devnum(cmd->device->channel, cmd->device->id); + vdev = vhba_lookup_device(devnum); + if (vdev) { + retval = vhba_device_dequeue(vdev, cmd); + vhba_device_put(vdev); + } else { + cmd->result = DID_NO_CONNECT << 16; + } + + return retval; +} + +static struct scsi_host_template vhba_template = { + .module = THIS_MODULE, + .name = "vhba", + .proc_name = "vhba", + .queuecommand = vhba_queuecommand, + .eh_abort_handler = vhba_abort, + .this_id = -1, + .max_sectors = VHBA_MAX_SECTORS_PER_IO, + .sg_tablesize = 256, +#if LINUX_VERSION_CODE < KERNEL_VERSION(3, 19, 0) + .slave_alloc = vhba_slave_alloc, +#endif +#if LINUX_VERSION_CODE >= KERNEL_VERSION(4, 0, 0) + .tag_alloc_policy = BLK_TAG_ALLOC_RR, +#endif +#if LINUX_VERSION_CODE >= KERNEL_VERSION(3, 19, 0) && LINUX_VERSION_CODE < KERNEL_VERSION(4, 4, 0) + .use_blk_tags = 1, +#endif +#if LINUX_VERSION_CODE >= KERNEL_VERSION(5, 0, 0) + .max_segment_size = VHBA_KBUF_SIZE, +#endif +}; + +ssize_t do_request (struct vhba_device *vdev, unsigned long metatag, struct scsi_cmnd *cmd, char __user *buf, size_t buf_len) +{ + struct vhba_request vreq; + ssize_t ret; + + scmd_dbg(cmd, "request %lu (%p), cdb 0x%x, bufflen %d, sg count %d\n", + metatag, cmd, cmd->cmnd[0], scsi_bufflen(cmd), scsi_sg_count(cmd)); + + ret = sizeof(vreq); + if (DATA_TO_DEVICE(cmd->sc_data_direction)) { + ret += scsi_bufflen(cmd); + } + + if (ret > buf_len) { + scmd_dbg(cmd, "buffer too small (%zd < %zd) for a request\n", buf_len, ret); + return -EIO; + } + + vreq.metatag = metatag; + vreq.lun = cmd->device->lun; + memcpy(vreq.cdb, cmd->cmnd, MAX_COMMAND_SIZE); + vreq.cdb_len = cmd->cmd_len; + vreq.data_len = scsi_bufflen(cmd); + + if (copy_to_user(buf, &vreq, sizeof(vreq))) { + return -EFAULT; + } + + if (DATA_TO_DEVICE(cmd->sc_data_direction) && vreq.data_len) { + buf += sizeof(vreq); + + if (scsi_sg_count(cmd)) { + unsigned char *kaddr, *uaddr; + struct scatterlist *sglist = scsi_sglist(cmd); + struct scatterlist *sg; + int i; + + uaddr = (unsigned char *) buf; + + for_each_sg(sglist, sg, scsi_sg_count(cmd), i) { + size_t len = sg->length; + + if (len > vdev->kbuf_size) { + scmd_dbg(cmd, "segment size (%zu) exceeds kbuf size (%zu)!", len, vdev->kbuf_size); + len = vdev->kbuf_size; + } + + kaddr = kmap_atomic(sg_page(sg)); + memcpy(vdev->kbuf, kaddr + sg->offset, len); + kunmap_atomic(kaddr); + + if (copy_to_user(uaddr, vdev->kbuf, len)) { + return -EFAULT; + } + uaddr += len; + } + } else { + if (copy_to_user(buf, scsi_sglist(cmd), vreq.data_len)) { + return -EFAULT; + } + } + } + + return ret; +} + +ssize_t do_response (struct vhba_device *vdev, unsigned long metatag, struct scsi_cmnd *cmd, const char __user *buf, size_t buf_len, struct vhba_response *res) +{ + ssize_t ret = 0; + + scmd_dbg(cmd, "response %lu (%p), status %x, data len %d, sg count %d\n", + metatag, cmd, res->status, res->data_len, scsi_sg_count(cmd)); + + if (res->status) { + if (res->data_len > SCSI_SENSE_BUFFERSIZE) { + scmd_dbg(cmd, "truncate sense (%d < %d)", SCSI_SENSE_BUFFERSIZE, res->data_len); + res->data_len = SCSI_SENSE_BUFFERSIZE; + } + + if (copy_from_user(cmd->sense_buffer, buf, res->data_len)) { + return -EFAULT; + } + + cmd->result = res->status; + + ret += res->data_len; + } else if (DATA_FROM_DEVICE(cmd->sc_data_direction) && scsi_bufflen(cmd)) { + size_t to_read; + + if (res->data_len > scsi_bufflen(cmd)) { + scmd_dbg(cmd, "truncate data (%d < %d)\n", scsi_bufflen(cmd), res->data_len); + res->data_len = scsi_bufflen(cmd); + } + + to_read = res->data_len; + + if (scsi_sg_count(cmd)) { + unsigned char *kaddr, *uaddr; + struct scatterlist *sglist = scsi_sglist(cmd); + struct scatterlist *sg; + int i; + + uaddr = (unsigned char *)buf; + + for_each_sg(sglist, sg, scsi_sg_count(cmd), i) { + size_t len = (sg->length < to_read) ? sg->length : to_read; + + if (len > vdev->kbuf_size) { + scmd_dbg(cmd, "segment size (%zu) exceeds kbuf size (%zu)!", len, vdev->kbuf_size); + len = vdev->kbuf_size; + } + + if (copy_from_user(vdev->kbuf, uaddr, len)) { + return -EFAULT; + } + uaddr += len; + + kaddr = kmap_atomic(sg_page(sg)); + memcpy(kaddr + sg->offset, vdev->kbuf, len); + kunmap_atomic(kaddr); + + to_read -= len; + if (to_read == 0) { + break; + } + } + } else { + if (copy_from_user(scsi_sglist(cmd), buf, res->data_len)) { + return -EFAULT; + } + + to_read -= res->data_len; + } + + scsi_set_resid(cmd, to_read); + + ret += res->data_len - to_read; + } + + return ret; +} + +struct vhba_command *next_command (struct vhba_device *vdev) +{ + struct vhba_command *vcmd; + + list_for_each_entry(vcmd, &vdev->cmd_list, entry) { + if (vcmd->status == VHBA_REQ_PENDING) { + break; + } + } + + if (&vcmd->entry == &vdev->cmd_list) { + vcmd = NULL; + } + + return vcmd; +} + +struct vhba_command *match_command (struct vhba_device *vdev, __u32 metatag) +{ + struct vhba_command *vcmd; + + list_for_each_entry(vcmd, &vdev->cmd_list, entry) { + if (vcmd->metatag == metatag) { + break; + } + } + + if (&vcmd->entry == &vdev->cmd_list) { + vcmd = NULL; + } + + return vcmd; +} + +struct vhba_command *wait_command (struct vhba_device *vdev, unsigned long flags) +{ + struct vhba_command *vcmd; + DEFINE_WAIT(wait); + + while (!(vcmd = next_command(vdev))) { + if (signal_pending(current)) { + break; + } + + prepare_to_wait(&vdev->cmd_wq, &wait, TASK_INTERRUPTIBLE); + + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + schedule(); + + spin_lock_irqsave(&vdev->cmd_lock, flags); + } + + finish_wait(&vdev->cmd_wq, &wait); + if (vcmd) { + vcmd->status = VHBA_REQ_READING; + } + + return vcmd; +} + +ssize_t vhba_ctl_read (struct file *file, char __user *buf, size_t buf_len, loff_t *offset) +{ + struct vhba_device *vdev; + struct vhba_command *vcmd; + ssize_t ret; + unsigned long flags; + + vdev = file->private_data; + + /* Get next command */ + if (file->f_flags & O_NONBLOCK) { + /* Non-blocking variant */ + spin_lock_irqsave(&vdev->cmd_lock, flags); + vcmd = next_command(vdev); + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + if (!vcmd) { + return -EWOULDBLOCK; + } + } else { + /* Blocking variant */ + spin_lock_irqsave(&vdev->cmd_lock, flags); + vcmd = wait_command(vdev, flags); + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + if (!vcmd) { + return -ERESTARTSYS; + } + } + + ret = do_request(vdev, vcmd->metatag, vcmd->cmd, buf, buf_len); + + spin_lock_irqsave(&vdev->cmd_lock, flags); + if (ret >= 0) { + vcmd->status = VHBA_REQ_SENT; + *offset += ret; + } else { + vcmd->status = VHBA_REQ_PENDING; + } + + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + return ret; +} + +ssize_t vhba_ctl_write (struct file *file, const char __user *buf, size_t buf_len, loff_t *offset) +{ + struct vhba_device *vdev; + struct vhba_command *vcmd; + struct vhba_response res; + ssize_t ret; + unsigned long flags; + + if (buf_len < sizeof(res)) { + return -EIO; + } + + if (copy_from_user(&res, buf, sizeof(res))) { + return -EFAULT; + } + + vdev = file->private_data; + + spin_lock_irqsave(&vdev->cmd_lock, flags); + vcmd = match_command(vdev, res.metatag); + if (!vcmd || vcmd->status != VHBA_REQ_SENT) { + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + pr_debug("ctl dev #%u not expecting response\n", vdev->num); + return -EIO; + } + vcmd->status = VHBA_REQ_WRITING; + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + ret = do_response(vdev, vcmd->metatag, vcmd->cmd, buf + sizeof(res), buf_len - sizeof(res), &res); + + spin_lock_irqsave(&vdev->cmd_lock, flags); + if (ret >= 0) { + vcmd->cmd->scsi_done(vcmd->cmd); + ret += sizeof(res); + + /* don't compete with vhba_device_dequeue */ + if (!list_empty(&vcmd->entry)) { + list_del_init(&vcmd->entry); + vhba_free_command(vcmd); + } + } else { + vcmd->status = VHBA_REQ_SENT; + } + + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + return ret; +} + +long vhba_ctl_ioctl (struct file *file, unsigned int cmd, unsigned long arg) +{ + struct vhba_device *vdev = file->private_data; + struct vhba_host *vhost = platform_get_drvdata(&vhba_platform_device); + + switch (cmd) { + case 0xBEEF001: { + unsigned int ident[4]; /* host, channel, id, lun */ + + ident[0] = vhost->shost->host_no; + devnum_to_bus_and_id(vdev->num, &ident[1], &ident[2]); + ident[3] = 0; /* lun */ + + if (copy_to_user((void *) arg, ident, sizeof(ident))) { + return -EFAULT; + } + + return 0; + } + case 0xBEEF002: { + unsigned int devnum = vdev->num; + + if (copy_to_user((void *) arg, &devnum, sizeof(devnum))) { + return -EFAULT; + } + + return 0; + } + } + + return -ENOTTY; +} + +#ifdef CONFIG_COMPAT +long vhba_ctl_compat_ioctl (struct file *file, unsigned int cmd, unsigned long arg) +{ + unsigned long compat_arg = (unsigned long)compat_ptr(arg); + return vhba_ctl_ioctl(file, cmd, compat_arg); +} +#endif + +unsigned int vhba_ctl_poll (struct file *file, poll_table *wait) +{ + struct vhba_device *vdev = file->private_data; + unsigned int mask = 0; + unsigned long flags; + + poll_wait(file, &vdev->cmd_wq, wait); + + spin_lock_irqsave(&vdev->cmd_lock, flags); + if (next_command(vdev)) { + mask |= POLLIN | POLLRDNORM; + } + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + return mask; +} + +int vhba_ctl_open (struct inode *inode, struct file *file) +{ + struct vhba_device *vdev; + int retval; + + pr_debug("ctl dev open\n"); + + /* check if vhba is probed */ + if (!platform_get_drvdata(&vhba_platform_device)) { + return -ENODEV; + } + + vdev = vhba_device_alloc(); + if (!vdev) { + return -ENOMEM; + } + + vdev->kbuf_size = VHBA_KBUF_SIZE; + vdev->kbuf = kzalloc(vdev->kbuf_size, GFP_KERNEL); + if (!vdev->kbuf) { + return -ENOMEM; + } + + if (!(retval = vhba_add_device(vdev))) { + file->private_data = vdev; + } + + vhba_device_put(vdev); + + return retval; +} + +int vhba_ctl_release (struct inode *inode, struct file *file) +{ + struct vhba_device *vdev; + struct vhba_command *vcmd; + unsigned long flags; + + vdev = file->private_data; + + pr_debug("ctl dev release\n"); + + vhba_device_get(vdev); + vhba_remove_device(vdev); + + spin_lock_irqsave(&vdev->cmd_lock, flags); + list_for_each_entry(vcmd, &vdev->cmd_list, entry) { + WARN_ON(vcmd->status == VHBA_REQ_READING || vcmd->status == VHBA_REQ_WRITING); + + scmd_dbg(vcmd->cmd, "device released with command %lu (%p)\n", vcmd->metatag, vcmd->cmd); + vcmd->cmd->result = DID_NO_CONNECT << 16; + vcmd->cmd->scsi_done(vcmd->cmd); + + vhba_free_command(vcmd); + } + INIT_LIST_HEAD(&vdev->cmd_list); + spin_unlock_irqrestore(&vdev->cmd_lock, flags); + + kfree(vdev->kbuf); + vdev->kbuf = NULL; + + vhba_device_put(vdev); + + return 0; +} + +static struct file_operations vhba_ctl_fops = { + .owner = THIS_MODULE, + .open = vhba_ctl_open, + .release = vhba_ctl_release, + .read = vhba_ctl_read, + .write = vhba_ctl_write, + .poll = vhba_ctl_poll, + .unlocked_ioctl = vhba_ctl_ioctl, +#ifdef CONFIG_COMPAT + .compat_ioctl = vhba_ctl_compat_ioctl, +#endif +}; + +static struct miscdevice vhba_miscdev = { + .minor = MISC_DYNAMIC_MINOR, + .name = "vhba_ctl", + .fops = &vhba_ctl_fops, +}; + +int vhba_probe (struct platform_device *pdev) +{ + struct Scsi_Host *shost; + struct vhba_host *vhost; + int i; + + vhba_can_queue = clamp(vhba_can_queue, 1, 256); + + shost = scsi_host_alloc(&vhba_template, sizeof(struct vhba_host)); + if (!shost) { + return -ENOMEM; + } + + shost->max_channel = VHBA_MAX_BUS-1; + shost->max_id = VHBA_MAX_ID; + /* we don't support lun > 0 */ + shost->max_lun = 1; + shost->max_cmd_len = MAX_COMMAND_SIZE; + shost->can_queue = vhba_can_queue; + shost->cmd_per_lun = vhba_can_queue; + + vhost = (struct vhba_host *)shost->hostdata; + memset(vhost, 0, sizeof(struct vhba_host)); + + vhost->shost = shost; + vhost->num_devices = 0; + spin_lock_init(&vhost->dev_lock); + spin_lock_init(&vhost->cmd_lock); + INIT_WORK(&vhost->scan_devices, vhba_scan_devices); + vhost->cmd_next = 0; + vhost->commands = kzalloc(vhba_can_queue * sizeof(struct vhba_command), GFP_KERNEL); + if (!vhost->commands) { + return -ENOMEM; + } + + for (i = 0; i < vhba_can_queue; i++) { + vhost->commands[i].status = VHBA_REQ_FREE; + } + + platform_set_drvdata(pdev, vhost); + +#if LINUX_VERSION_CODE < KERNEL_VERSION(4, 4, 0) + i = scsi_init_shared_tag_map(shost, vhba_can_queue); + if (i) return i; +#endif + + if (scsi_add_host(shost, &pdev->dev)) { + scsi_host_put(shost); + return -ENOMEM; + } + + return 0; +} + +int vhba_remove (struct platform_device *pdev) +{ + struct vhba_host *vhost; + struct Scsi_Host *shost; + + vhost = platform_get_drvdata(pdev); + shost = vhost->shost; + + scsi_remove_host(shost); + scsi_host_put(shost); + + kfree(vhost->commands); + + return 0; +} + +void vhba_release (struct device * dev) +{ + return; +} + +static struct platform_device vhba_platform_device = { + .name = "vhba", + .id = -1, + .dev = { + .release = vhba_release, + }, +}; + +static struct platform_driver vhba_platform_driver = { + .driver = { + .owner = THIS_MODULE, + .name = "vhba", + }, + .probe = vhba_probe, + .remove = vhba_remove, +}; + +int __init vhba_init (void) +{ + int ret; + + ret = platform_device_register(&vhba_platform_device); + if (ret < 0) { + return ret; + } + + ret = platform_driver_register(&vhba_platform_driver); + if (ret < 0) { + platform_device_unregister(&vhba_platform_device); + return ret; + } + + ret = misc_register(&vhba_miscdev); + if (ret < 0) { + platform_driver_unregister(&vhba_platform_driver); + platform_device_unregister(&vhba_platform_device); + return ret; + } + + return 0; +} + +void __exit vhba_exit(void) +{ + misc_deregister(&vhba_miscdev); + platform_driver_unregister(&vhba_platform_driver); + platform_device_unregister(&vhba_platform_device); +} + +module_init(vhba_init); +module_exit(vhba_exit); + diff --git a/drivers/staging/rts5208/rtsx.c b/drivers/staging/rts5208/rtsx.c index 898add4d1fc..0aa9dd46734 100644 --- a/drivers/staging/rts5208/rtsx.c +++ b/drivers/staging/rts5208/rtsx.c @@ -477,7 +477,7 @@ static int rtsx_polling_thread(void *__dev) for (;;) { set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(msecs_to_jiffies(POLLING_INTERVAL)); + schedule_msec_hrtimeout((POLLING_INTERVAL)); /* lock the device pointers */ mutex_lock(&dev->dev_mutex); diff --git a/drivers/staging/unisys/visornic/visornic_main.c b/drivers/staging/unisys/visornic/visornic_main.c index a3bc568c660..e7cac6df020 100644 --- a/drivers/staging/unisys/visornic/visornic_main.c +++ b/drivers/staging/unisys/visornic/visornic_main.c @@ -547,7 +547,7 @@ static int visornic_disable_with_timeout(struct net_device *netdev, } set_current_state(TASK_INTERRUPTIBLE); spin_unlock_irqrestore(&devdata->priv_lock, flags); - wait += schedule_timeout(msecs_to_jiffies(10)); + wait += schedule_msec_hrtimeout((10)); spin_lock_irqsave(&devdata->priv_lock, flags); } @@ -558,7 +558,7 @@ static int visornic_disable_with_timeout(struct net_device *netdev, while (1) { set_current_state(TASK_INTERRUPTIBLE); spin_unlock_irqrestore(&devdata->priv_lock, flags); - schedule_timeout(msecs_to_jiffies(10)); + schedule_msec_hrtimeout((10)); spin_lock_irqsave(&devdata->priv_lock, flags); if (atomic_read(&devdata->usage)) break; @@ -712,7 +712,7 @@ static int visornic_enable_with_timeout(struct net_device *netdev, } set_current_state(TASK_INTERRUPTIBLE); spin_unlock_irqrestore(&devdata->priv_lock, flags); - wait += schedule_timeout(msecs_to_jiffies(10)); + wait += schedule_msec_hrtimeout((10)); spin_lock_irqsave(&devdata->priv_lock, flags); } diff --git a/drivers/usb/gadget/function/u_serial.c b/drivers/usb/gadget/function/u_serial.c index 1e59204ec7a..b21260520a9 100644 --- a/drivers/usb/gadget/function/u_serial.c +++ b/drivers/usb/gadget/function/u_serial.c @@ -1386,8 +1386,13 @@ void gserial_disconnect(struct gserial *gser) gser->ioport = NULL; if (port->port.count > 0) { wake_up_interruptible(&port->drain_wait); +#if 0 if (port->port.tty) tty_hangup(port->port.tty); +#else + if (port->port.tty) + stop_tty(port->port.tty); +#endif } port->suspended = false; spin_unlock_irqrestore(&port->port_lock, flags); diff --git a/drivers/video/fbdev/omap/hwa742.c b/drivers/video/fbdev/omap/hwa742.c index b191bef22d9..a8c491e1d84 100644 --- a/drivers/video/fbdev/omap/hwa742.c +++ b/drivers/video/fbdev/omap/hwa742.c @@ -927,7 +927,7 @@ static void hwa742_resume(void) if (hwa742_read_reg(HWA742_PLL_DIV_REG) & (1 << 7)) break; set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(msecs_to_jiffies(5)); + schedule_msec_hrtimeout((5)); } hwa742_set_update_mode(hwa742.update_mode_before_suspend); } diff --git a/drivers/video/fbdev/pxafb.c b/drivers/video/fbdev/pxafb.c index f1551e00eb1..f0f651e9250 100644 --- a/drivers/video/fbdev/pxafb.c +++ b/drivers/video/fbdev/pxafb.c @@ -1287,7 +1287,7 @@ static int pxafb_smart_thread(void *arg) mutex_unlock(&fbi->ctrlr_lock); set_current_state(TASK_INTERRUPTIBLE); - schedule_timeout(msecs_to_jiffies(30)); + schedule_msec_hrtimeout((30)); } pr_debug("%s(): task ending\n", __func__); diff --git a/drivers/video/logo/logo_linux_clut224.ppm b/drivers/video/logo/logo_linux_clut224.ppm index 3c14e43b82f..36bb8b1cb30 100644 --- a/drivers/video/logo/logo_linux_clut224.ppm +++ b/drivers/video/logo/logo_linux_clut224.ppm @@ -1,1604 +1,14404 @@ P3 -# Standard 224-color Linux logo -80 80 -255 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 6 6 6 10 10 10 10 10 10 - 10 10 10 6 6 6 6 6 6 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 10 10 10 14 14 14 - 22 22 22 26 26 26 30 30 30 34 34 34 - 30 30 30 30 30 30 26 26 26 18 18 18 - 14 14 14 10 10 10 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 14 14 14 26 26 26 42 42 42 - 54 54 54 66 66 66 78 78 78 78 78 78 - 78 78 78 74 74 74 66 66 66 54 54 54 - 42 42 42 26 26 26 18 18 18 10 10 10 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 22 22 22 42 42 42 66 66 66 86 86 86 - 66 66 66 38 38 38 38 38 38 22 22 22 - 26 26 26 34 34 34 54 54 54 66 66 66 - 86 86 86 70 70 70 46 46 46 26 26 26 - 14 14 14 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 10 10 10 26 26 26 - 50 50 50 82 82 82 58 58 58 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 6 6 6 54 54 54 86 86 86 66 66 66 - 38 38 38 18 18 18 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 22 22 22 50 50 50 - 78 78 78 34 34 34 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 6 6 6 70 70 70 - 78 78 78 46 46 46 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 18 18 18 42 42 42 82 82 82 - 26 26 26 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 14 14 14 - 46 46 46 34 34 34 6 6 6 2 2 6 - 42 42 42 78 78 78 42 42 42 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 0 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 10 10 10 30 30 30 66 66 66 58 58 58 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 26 26 26 - 86 86 86 101 101 101 46 46 46 10 10 10 - 2 2 6 58 58 58 70 70 70 34 34 34 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 14 14 14 42 42 42 86 86 86 10 10 10 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 30 30 30 - 94 94 94 94 94 94 58 58 58 26 26 26 - 2 2 6 6 6 6 78 78 78 54 54 54 - 22 22 22 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 22 22 22 62 62 62 62 62 62 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 26 26 26 - 54 54 54 38 38 38 18 18 18 10 10 10 - 2 2 6 2 2 6 34 34 34 82 82 82 - 38 38 38 14 14 14 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 30 30 30 78 78 78 30 30 30 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 10 10 10 - 10 10 10 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 78 78 78 - 50 50 50 18 18 18 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 38 38 38 86 86 86 14 14 14 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 54 54 54 - 66 66 66 26 26 26 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 42 42 42 82 82 82 2 2 6 2 2 6 - 2 2 6 6 6 6 10 10 10 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 6 6 6 - 14 14 14 10 10 10 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 18 18 18 - 82 82 82 34 34 34 10 10 10 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 46 46 46 86 86 86 2 2 6 2 2 6 - 6 6 6 6 6 6 22 22 22 34 34 34 - 6 6 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 18 18 18 34 34 34 - 10 10 10 50 50 50 22 22 22 2 2 6 - 2 2 6 2 2 6 2 2 6 10 10 10 - 86 86 86 42 42 42 14 14 14 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 46 46 46 86 86 86 2 2 6 2 2 6 - 38 38 38 116 116 116 94 94 94 22 22 22 - 22 22 22 2 2 6 2 2 6 2 2 6 - 14 14 14 86 86 86 138 138 138 162 162 162 -154 154 154 38 38 38 26 26 26 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 86 86 86 46 46 46 14 14 14 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 46 46 46 86 86 86 2 2 6 14 14 14 -134 134 134 198 198 198 195 195 195 116 116 116 - 10 10 10 2 2 6 2 2 6 6 6 6 -101 98 89 187 187 187 210 210 210 218 218 218 -214 214 214 134 134 134 14 14 14 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 86 86 86 50 50 50 18 18 18 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 1 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 46 46 46 86 86 86 2 2 6 54 54 54 -218 218 218 195 195 195 226 226 226 246 246 246 - 58 58 58 2 2 6 2 2 6 30 30 30 -210 210 210 253 253 253 174 174 174 123 123 123 -221 221 221 234 234 234 74 74 74 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 70 70 70 58 58 58 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 46 46 46 82 82 82 2 2 6 106 106 106 -170 170 170 26 26 26 86 86 86 226 226 226 -123 123 123 10 10 10 14 14 14 46 46 46 -231 231 231 190 190 190 6 6 6 70 70 70 - 90 90 90 238 238 238 158 158 158 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 70 70 70 58 58 58 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 1 0 0 0 - 0 0 1 0 0 1 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 42 42 42 86 86 86 6 6 6 116 116 116 -106 106 106 6 6 6 70 70 70 149 149 149 -128 128 128 18 18 18 38 38 38 54 54 54 -221 221 221 106 106 106 2 2 6 14 14 14 - 46 46 46 190 190 190 198 198 198 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 74 74 74 62 62 62 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 1 0 0 0 - 0 0 1 0 0 0 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 42 42 42 94 94 94 14 14 14 101 101 101 -128 128 128 2 2 6 18 18 18 116 116 116 -118 98 46 121 92 8 121 92 8 98 78 10 -162 162 162 106 106 106 2 2 6 2 2 6 - 2 2 6 195 195 195 195 195 195 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 74 74 74 62 62 62 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 1 0 0 1 - 0 0 1 0 0 0 0 0 1 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 38 38 38 90 90 90 14 14 14 58 58 58 -210 210 210 26 26 26 54 38 6 154 114 10 -226 170 11 236 186 11 225 175 15 184 144 12 -215 174 15 175 146 61 37 26 9 2 2 6 - 70 70 70 246 246 246 138 138 138 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 70 70 70 66 66 66 26 26 26 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 38 38 38 86 86 86 14 14 14 10 10 10 -195 195 195 188 164 115 192 133 9 225 175 15 -239 182 13 234 190 10 232 195 16 232 200 30 -245 207 45 241 208 19 232 195 16 184 144 12 -218 194 134 211 206 186 42 42 42 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 50 50 50 74 74 74 30 30 30 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 34 34 34 86 86 86 14 14 14 2 2 6 -121 87 25 192 133 9 219 162 10 239 182 13 -236 186 11 232 195 16 241 208 19 244 214 54 -246 218 60 246 218 38 246 215 20 241 208 19 -241 208 19 226 184 13 121 87 25 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 50 50 50 82 82 82 34 34 34 10 10 10 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 34 34 34 82 82 82 30 30 30 61 42 6 -180 123 7 206 145 10 230 174 11 239 182 13 -234 190 10 238 202 15 241 208 19 246 218 74 -246 218 38 246 215 20 246 215 20 246 215 20 -226 184 13 215 174 15 184 144 12 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 26 26 26 94 94 94 42 42 42 14 14 14 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 78 78 78 50 50 50 104 69 6 -192 133 9 216 158 10 236 178 12 236 186 11 -232 195 16 241 208 19 244 214 54 245 215 43 -246 215 20 246 215 20 241 208 19 198 155 10 -200 144 11 216 158 10 156 118 10 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 6 6 6 90 90 90 54 54 54 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 78 78 78 46 46 46 22 22 22 -137 92 6 210 162 10 239 182 13 238 190 10 -238 202 15 241 208 19 246 215 20 246 215 20 -241 208 19 203 166 17 185 133 11 210 150 10 -216 158 10 210 150 10 102 78 10 2 2 6 - 6 6 6 54 54 54 14 14 14 2 2 6 - 2 2 6 62 62 62 74 74 74 30 30 30 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 34 34 34 78 78 78 50 50 50 6 6 6 - 94 70 30 139 102 15 190 146 13 226 184 13 -232 200 30 232 195 16 215 174 15 190 146 13 -168 122 10 192 133 9 210 150 10 213 154 11 -202 150 34 182 157 106 101 98 89 2 2 6 - 2 2 6 78 78 78 116 116 116 58 58 58 - 2 2 6 22 22 22 90 90 90 46 46 46 - 18 18 18 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 38 38 38 86 86 86 50 50 50 6 6 6 -128 128 128 174 154 114 156 107 11 168 122 10 -198 155 10 184 144 12 197 138 11 200 144 11 -206 145 10 206 145 10 197 138 11 188 164 115 -195 195 195 198 198 198 174 174 174 14 14 14 - 2 2 6 22 22 22 116 116 116 116 116 116 - 22 22 22 2 2 6 74 74 74 70 70 70 - 30 30 30 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 50 50 50 101 101 101 26 26 26 10 10 10 -138 138 138 190 190 190 174 154 114 156 107 11 -197 138 11 200 144 11 197 138 11 192 133 9 -180 123 7 190 142 34 190 178 144 187 187 187 -202 202 202 221 221 221 214 214 214 66 66 66 - 2 2 6 2 2 6 50 50 50 62 62 62 - 6 6 6 2 2 6 10 10 10 90 90 90 - 50 50 50 18 18 18 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 10 10 10 34 34 34 - 74 74 74 74 74 74 2 2 6 6 6 6 -144 144 144 198 198 198 190 190 190 178 166 146 -154 121 60 156 107 11 156 107 11 168 124 44 -174 154 114 187 187 187 190 190 190 210 210 210 -246 246 246 253 253 253 253 253 253 182 182 182 - 6 6 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 62 62 62 - 74 74 74 34 34 34 14 14 14 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 10 10 10 22 22 22 54 54 54 - 94 94 94 18 18 18 2 2 6 46 46 46 -234 234 234 221 221 221 190 190 190 190 190 190 -190 190 190 187 187 187 187 187 187 190 190 190 -190 190 190 195 195 195 214 214 214 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 - 82 82 82 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 14 14 14 - 86 86 86 54 54 54 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 18 18 18 46 46 46 90 90 90 - 46 46 46 18 18 18 6 6 6 182 182 182 -253 253 253 246 246 246 206 206 206 190 190 190 -190 190 190 190 190 190 190 190 190 190 190 190 -206 206 206 231 231 231 250 250 250 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -202 202 202 14 14 14 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 42 42 42 86 86 86 42 42 42 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 14 14 14 38 38 38 74 74 74 66 66 66 - 2 2 6 6 6 6 90 90 90 250 250 250 -253 253 253 253 253 253 238 238 238 198 198 198 -190 190 190 190 190 190 195 195 195 221 221 221 -246 246 246 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 82 82 82 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 78 78 78 70 70 70 34 34 34 - 14 14 14 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 34 34 34 66 66 66 78 78 78 6 6 6 - 2 2 6 18 18 18 218 218 218 253 253 253 -253 253 253 253 253 253 253 253 253 246 246 246 -226 226 226 231 231 231 246 246 246 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 178 178 178 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 18 18 18 90 90 90 62 62 62 - 30 30 30 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 10 10 10 26 26 26 - 58 58 58 90 90 90 18 18 18 2 2 6 - 2 2 6 110 110 110 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -250 250 250 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 231 231 231 18 18 18 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 18 18 18 94 94 94 - 54 54 54 26 26 26 10 10 10 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 22 22 22 50 50 50 - 90 90 90 26 26 26 2 2 6 2 2 6 - 14 14 14 195 195 195 250 250 250 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -250 250 250 242 242 242 54 54 54 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 38 38 38 - 86 86 86 50 50 50 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 14 14 14 38 38 38 82 82 82 - 34 34 34 2 2 6 2 2 6 2 2 6 - 42 42 42 195 195 195 246 246 246 253 253 253 -253 253 253 253 253 253 253 253 253 250 250 250 -242 242 242 242 242 242 250 250 250 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 250 250 250 246 246 246 238 238 238 -226 226 226 231 231 231 101 101 101 6 6 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 38 38 38 82 82 82 42 42 42 14 14 14 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 10 10 10 26 26 26 62 62 62 66 66 66 - 2 2 6 2 2 6 2 2 6 6 6 6 - 70 70 70 170 170 170 206 206 206 234 234 234 -246 246 246 250 250 250 250 250 250 238 238 238 -226 226 226 231 231 231 238 238 238 250 250 250 -250 250 250 250 250 250 246 246 246 231 231 231 -214 214 214 206 206 206 202 202 202 202 202 202 -198 198 198 202 202 202 182 182 182 18 18 18 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 62 62 62 66 66 66 30 30 30 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 14 14 14 42 42 42 82 82 82 18 18 18 - 2 2 6 2 2 6 2 2 6 10 10 10 - 94 94 94 182 182 182 218 218 218 242 242 242 -250 250 250 253 253 253 253 253 253 250 250 250 -234 234 234 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 246 246 246 -238 238 238 226 226 226 210 210 210 202 202 202 -195 195 195 195 195 195 210 210 210 158 158 158 - 6 6 6 14 14 14 50 50 50 14 14 14 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 6 6 6 86 86 86 46 46 46 - 18 18 18 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 22 22 22 54 54 54 70 70 70 2 2 6 - 2 2 6 10 10 10 2 2 6 22 22 22 -166 166 166 231 231 231 250 250 250 253 253 253 -253 253 253 253 253 253 253 253 253 250 250 250 -242 242 242 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 246 246 246 -231 231 231 206 206 206 198 198 198 226 226 226 - 94 94 94 2 2 6 6 6 6 38 38 38 - 30 30 30 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 62 62 62 66 66 66 - 26 26 26 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 74 74 74 50 50 50 2 2 6 - 26 26 26 26 26 26 2 2 6 106 106 106 -238 238 238 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 246 246 246 218 218 218 202 202 202 -210 210 210 14 14 14 2 2 6 2 2 6 - 30 30 30 22 22 22 2 2 6 2 2 6 - 2 2 6 2 2 6 18 18 18 86 86 86 - 42 42 42 14 14 14 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 42 42 42 90 90 90 22 22 22 2 2 6 - 42 42 42 2 2 6 18 18 18 218 218 218 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 250 250 250 221 221 221 -218 218 218 101 101 101 2 2 6 14 14 14 - 18 18 18 38 38 38 10 10 10 2 2 6 - 2 2 6 2 2 6 2 2 6 78 78 78 - 58 58 58 22 22 22 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 54 54 54 82 82 82 2 2 6 26 26 26 - 22 22 22 2 2 6 123 123 123 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 250 250 250 -238 238 238 198 198 198 6 6 6 38 38 38 - 58 58 58 26 26 26 38 38 38 2 2 6 - 2 2 6 2 2 6 2 2 6 46 46 46 - 78 78 78 30 30 30 10 10 10 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 10 10 10 30 30 30 - 74 74 74 58 58 58 2 2 6 42 42 42 - 2 2 6 22 22 22 231 231 231 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 250 250 250 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 246 246 246 46 46 46 38 38 38 - 42 42 42 14 14 14 38 38 38 14 14 14 - 2 2 6 2 2 6 2 2 6 6 6 6 - 86 86 86 46 46 46 14 14 14 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 14 14 14 42 42 42 - 90 90 90 18 18 18 18 18 18 26 26 26 - 2 2 6 116 116 116 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 250 250 250 238 238 238 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 94 94 94 6 6 6 - 2 2 6 2 2 6 10 10 10 34 34 34 - 2 2 6 2 2 6 2 2 6 2 2 6 - 74 74 74 58 58 58 22 22 22 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 10 10 10 26 26 26 66 66 66 - 82 82 82 2 2 6 38 38 38 6 6 6 - 14 14 14 210 210 210 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 246 246 246 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 144 144 144 2 2 6 - 2 2 6 2 2 6 2 2 6 46 46 46 - 2 2 6 2 2 6 2 2 6 2 2 6 - 42 42 42 74 74 74 30 30 30 10 10 10 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 14 14 14 42 42 42 90 90 90 - 26 26 26 6 6 6 42 42 42 2 2 6 - 74 74 74 250 250 250 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 242 242 242 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 182 182 182 2 2 6 - 2 2 6 2 2 6 2 2 6 46 46 46 - 2 2 6 2 2 6 2 2 6 2 2 6 - 10 10 10 86 86 86 38 38 38 10 10 10 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 10 10 10 26 26 26 66 66 66 82 82 82 - 2 2 6 22 22 22 18 18 18 2 2 6 -149 149 149 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 234 234 234 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 206 206 206 2 2 6 - 2 2 6 2 2 6 2 2 6 38 38 38 - 2 2 6 2 2 6 2 2 6 2 2 6 - 6 6 6 86 86 86 46 46 46 14 14 14 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 18 18 18 46 46 46 86 86 86 18 18 18 - 2 2 6 34 34 34 10 10 10 6 6 6 -210 210 210 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 234 234 234 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 221 221 221 6 6 6 - 2 2 6 2 2 6 6 6 6 30 30 30 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 82 82 82 54 54 54 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 26 26 26 66 66 66 62 62 62 2 2 6 - 2 2 6 38 38 38 10 10 10 26 26 26 -238 238 238 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 238 238 238 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 6 6 6 - 2 2 6 2 2 6 10 10 10 30 30 30 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 66 66 66 58 58 58 22 22 22 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 38 38 38 78 78 78 6 6 6 2 2 6 - 2 2 6 46 46 46 14 14 14 42 42 42 -246 246 246 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 234 234 234 10 10 10 - 2 2 6 2 2 6 22 22 22 14 14 14 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 66 66 66 62 62 62 22 22 22 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 50 50 50 74 74 74 2 2 6 2 2 6 - 14 14 14 70 70 70 34 34 34 62 62 62 -250 250 250 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 246 246 246 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 234 234 234 14 14 14 - 2 2 6 2 2 6 30 30 30 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 66 66 66 62 62 62 22 22 22 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 54 54 54 62 62 62 2 2 6 2 2 6 - 2 2 6 30 30 30 46 46 46 70 70 70 -250 250 250 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 246 246 246 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 226 226 226 10 10 10 - 2 2 6 6 6 6 30 30 30 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 66 66 66 58 58 58 22 22 22 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 22 22 22 - 58 58 58 62 62 62 2 2 6 2 2 6 - 2 2 6 2 2 6 30 30 30 78 78 78 -250 250 250 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 246 246 246 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 206 206 206 2 2 6 - 22 22 22 34 34 34 18 14 6 22 22 22 - 26 26 26 18 18 18 6 6 6 2 2 6 - 2 2 6 82 82 82 54 54 54 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 26 26 26 - 62 62 62 106 106 106 74 54 14 185 133 11 -210 162 10 121 92 8 6 6 6 62 62 62 -238 238 238 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 246 246 246 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 158 158 158 18 18 18 - 14 14 14 2 2 6 2 2 6 2 2 6 - 6 6 6 18 18 18 66 66 66 38 38 38 - 6 6 6 94 94 94 50 50 50 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 10 10 10 10 10 10 18 18 18 38 38 38 - 78 78 78 142 134 106 216 158 10 242 186 14 -246 190 14 246 190 14 156 118 10 10 10 10 - 90 90 90 238 238 238 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 250 250 250 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 246 230 190 -238 204 91 238 204 91 181 142 44 37 26 9 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 38 38 38 46 46 46 - 26 26 26 106 106 106 54 54 54 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 14 14 14 22 22 22 - 30 30 30 38 38 38 50 50 50 70 70 70 -106 106 106 190 142 34 226 170 11 242 186 14 -246 190 14 246 190 14 246 190 14 154 114 10 - 6 6 6 74 74 74 226 226 226 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 231 231 231 250 250 250 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 228 184 62 -241 196 14 241 208 19 232 195 16 38 30 10 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 6 6 6 30 30 30 26 26 26 -203 166 17 154 142 90 66 66 66 26 26 26 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 18 18 18 38 38 38 58 58 58 - 78 78 78 86 86 86 101 101 101 123 123 123 -175 146 61 210 150 10 234 174 13 246 186 14 -246 190 14 246 190 14 246 190 14 238 190 10 -102 78 10 2 2 6 46 46 46 198 198 198 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 234 234 234 242 242 242 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 224 178 62 -242 186 14 241 196 14 210 166 10 22 18 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 6 6 6 121 92 8 -238 202 15 232 195 16 82 82 82 34 34 34 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 14 14 14 38 38 38 70 70 70 154 122 46 -190 142 34 200 144 11 197 138 11 197 138 11 -213 154 11 226 170 11 242 186 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -225 175 15 46 32 6 2 2 6 22 22 22 -158 158 158 250 250 250 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 250 250 250 242 242 242 224 178 62 -239 182 13 236 186 11 213 154 11 46 32 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 61 42 6 225 175 15 -238 190 10 236 186 11 112 100 78 42 42 42 - 14 14 14 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 22 22 22 54 54 54 154 122 46 213 154 11 -226 170 11 230 174 11 226 170 11 226 170 11 -236 178 12 242 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -241 196 14 184 144 12 10 10 10 2 2 6 - 6 6 6 116 116 116 242 242 242 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 231 231 231 198 198 198 214 170 54 -236 178 12 236 178 12 210 150 10 137 92 6 - 18 14 6 2 2 6 2 2 6 2 2 6 - 6 6 6 70 47 6 200 144 11 236 178 12 -239 182 13 239 182 13 124 112 88 58 58 58 - 22 22 22 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 70 70 70 180 133 36 226 170 11 -239 182 13 242 186 14 242 186 14 246 186 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 232 195 16 98 70 6 2 2 6 - 2 2 6 2 2 6 66 66 66 221 221 221 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 206 206 206 198 198 198 214 166 58 -230 174 11 230 174 11 216 158 10 192 133 9 -163 110 8 116 81 8 102 78 10 116 81 8 -167 114 7 197 138 11 226 170 11 239 182 13 -242 186 14 242 186 14 162 146 94 78 78 78 - 34 34 34 14 14 14 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 30 30 30 78 78 78 190 142 34 226 170 11 -239 182 13 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 241 196 14 203 166 17 22 18 6 - 2 2 6 2 2 6 2 2 6 38 38 38 -218 218 218 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -250 250 250 206 206 206 198 198 198 202 162 69 -226 170 11 236 178 12 224 166 10 210 150 10 -200 144 11 197 138 11 192 133 9 197 138 11 -210 150 10 226 170 11 242 186 14 246 190 14 -246 190 14 246 186 14 225 175 15 124 112 88 - 62 62 62 30 30 30 14 14 14 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 78 78 78 174 135 50 224 166 10 -239 182 13 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 241 196 14 139 102 15 - 2 2 6 2 2 6 2 2 6 2 2 6 - 78 78 78 250 250 250 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -250 250 250 214 214 214 198 198 198 190 150 46 -219 162 10 236 178 12 234 174 13 224 166 10 -216 158 10 213 154 11 213 154 11 216 158 10 -226 170 11 239 182 13 246 190 14 246 190 14 -246 190 14 246 190 14 242 186 14 206 162 42 -101 101 101 58 58 58 30 30 30 14 14 14 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 74 74 74 174 135 50 216 158 10 -236 178 12 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 241 196 14 226 184 13 - 61 42 6 2 2 6 2 2 6 2 2 6 - 22 22 22 238 238 238 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 226 226 226 187 187 187 180 133 36 -216 158 10 236 178 12 239 182 13 236 178 12 -230 174 11 226 170 11 226 170 11 230 174 11 -236 178 12 242 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 186 14 239 182 13 -206 162 42 106 106 106 66 66 66 34 34 34 - 14 14 14 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 26 26 26 70 70 70 163 133 67 213 154 11 -236 178 12 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 241 196 14 -190 146 13 18 14 6 2 2 6 2 2 6 - 46 46 46 246 246 246 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 221 221 221 86 86 86 156 107 11 -216 158 10 236 178 12 242 186 14 246 186 14 -242 186 14 239 182 13 239 182 13 242 186 14 -242 186 14 246 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -242 186 14 225 175 15 142 122 72 66 66 66 - 30 30 30 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 26 26 26 70 70 70 163 133 67 210 150 10 -236 178 12 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -232 195 16 121 92 8 34 34 34 106 106 106 -221 221 221 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -242 242 242 82 82 82 18 14 6 163 110 8 -216 158 10 236 178 12 242 186 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 242 186 14 163 133 67 - 46 46 46 18 18 18 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 10 10 10 - 30 30 30 78 78 78 163 133 67 210 150 10 -236 178 12 246 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -241 196 14 215 174 15 190 178 144 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 218 218 218 - 58 58 58 2 2 6 22 18 6 167 114 7 -216 158 10 236 178 12 246 186 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 186 14 242 186 14 190 150 46 - 54 54 54 22 22 22 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 38 38 38 86 86 86 180 133 36 213 154 11 -236 178 12 246 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 232 195 16 190 146 13 214 214 214 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 250 250 250 170 170 170 26 26 26 - 2 2 6 2 2 6 37 26 9 163 110 8 -219 162 10 239 182 13 246 186 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 186 14 236 178 12 224 166 10 142 122 72 - 46 46 46 18 18 18 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 50 50 50 109 106 95 192 133 9 224 166 10 -242 186 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -242 186 14 226 184 13 210 162 10 142 110 46 -226 226 226 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -253 253 253 253 253 253 253 253 253 253 253 253 -198 198 198 66 66 66 2 2 6 2 2 6 - 2 2 6 2 2 6 50 34 6 156 107 11 -219 162 10 239 182 13 246 186 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 242 186 14 -234 174 13 213 154 11 154 122 46 66 66 66 - 30 30 30 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 22 22 22 - 58 58 58 154 121 60 206 145 10 234 174 13 -242 186 14 246 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 186 14 236 178 12 210 162 10 163 110 8 - 61 42 6 138 138 138 218 218 218 250 250 250 -253 253 253 253 253 253 253 253 253 250 250 250 -242 242 242 210 210 210 144 144 144 66 66 66 - 6 6 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 61 42 6 163 110 8 -216 158 10 236 178 12 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 239 182 13 230 174 11 216 158 10 -190 142 34 124 112 88 70 70 70 38 38 38 - 18 18 18 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 22 22 22 - 62 62 62 168 124 44 206 145 10 224 166 10 -236 178 12 239 182 13 242 186 14 242 186 14 -246 186 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 236 178 12 216 158 10 175 118 6 - 80 54 7 2 2 6 6 6 6 30 30 30 - 54 54 54 62 62 62 50 50 50 38 38 38 - 14 14 14 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 6 6 6 80 54 7 167 114 7 -213 154 11 236 178 12 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 190 14 242 186 14 239 182 13 239 182 13 -230 174 11 210 150 10 174 135 50 124 112 88 - 82 82 82 54 54 54 34 34 34 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 18 18 18 - 50 50 50 158 118 36 192 133 9 200 144 11 -216 158 10 219 162 10 224 166 10 226 170 11 -230 174 11 236 178 12 239 182 13 239 182 13 -242 186 14 246 186 14 246 190 14 246 190 14 -246 190 14 246 190 14 246 190 14 246 190 14 -246 186 14 230 174 11 210 150 10 163 110 8 -104 69 6 10 10 10 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 6 6 6 91 60 6 167 114 7 -206 145 10 230 174 11 242 186 14 246 190 14 -246 190 14 246 190 14 246 186 14 242 186 14 -239 182 13 230 174 11 224 166 10 213 154 11 -180 133 36 124 112 88 86 86 86 58 58 58 - 38 38 38 22 22 22 10 10 10 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 14 14 14 - 34 34 34 70 70 70 138 110 50 158 118 36 -167 114 7 180 123 7 192 133 9 197 138 11 -200 144 11 206 145 10 213 154 11 219 162 10 -224 166 10 230 174 11 239 182 13 242 186 14 -246 186 14 246 186 14 246 186 14 246 186 14 -239 182 13 216 158 10 185 133 11 152 99 6 -104 69 6 18 14 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 2 2 6 2 2 6 2 2 6 - 2 2 6 6 6 6 80 54 7 152 99 6 -192 133 9 219 162 10 236 178 12 239 182 13 -246 186 14 242 186 14 239 182 13 236 178 12 -224 166 10 206 145 10 192 133 9 154 121 60 - 94 94 94 62 62 62 42 42 42 22 22 22 - 14 14 14 6 6 6 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 18 18 18 34 34 34 58 58 58 78 78 78 -101 98 89 124 112 88 142 110 46 156 107 11 -163 110 8 167 114 7 175 118 6 180 123 7 -185 133 11 197 138 11 210 150 10 219 162 10 -226 170 11 236 178 12 236 178 12 234 174 13 -219 162 10 197 138 11 163 110 8 130 83 6 - 91 60 6 10 10 10 2 2 6 2 2 6 - 18 18 18 38 38 38 38 38 38 38 38 38 - 38 38 38 38 38 38 38 38 38 38 38 38 - 38 38 38 38 38 38 26 26 26 2 2 6 - 2 2 6 6 6 6 70 47 6 137 92 6 -175 118 6 200 144 11 219 162 10 230 174 11 -234 174 13 230 174 11 219 162 10 210 150 10 -192 133 9 163 110 8 124 112 88 82 82 82 - 50 50 50 30 30 30 14 14 14 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 14 14 14 22 22 22 34 34 34 - 42 42 42 58 58 58 74 74 74 86 86 86 -101 98 89 122 102 70 130 98 46 121 87 25 -137 92 6 152 99 6 163 110 8 180 123 7 -185 133 11 197 138 11 206 145 10 200 144 11 -180 123 7 156 107 11 130 83 6 104 69 6 - 50 34 6 54 54 54 110 110 110 101 98 89 - 86 86 86 82 82 82 78 78 78 78 78 78 - 78 78 78 78 78 78 78 78 78 78 78 78 - 78 78 78 82 82 82 86 86 86 94 94 94 -106 106 106 101 101 101 86 66 34 124 80 6 -156 107 11 180 123 7 192 133 9 200 144 11 -206 145 10 200 144 11 192 133 9 175 118 6 -139 102 15 109 106 95 70 70 70 42 42 42 - 22 22 22 10 10 10 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 6 6 6 10 10 10 - 14 14 14 22 22 22 30 30 30 38 38 38 - 50 50 50 62 62 62 74 74 74 90 90 90 -101 98 89 112 100 78 121 87 25 124 80 6 -137 92 6 152 99 6 152 99 6 152 99 6 -138 86 6 124 80 6 98 70 6 86 66 30 -101 98 89 82 82 82 58 58 58 46 46 46 - 38 38 38 34 34 34 34 34 34 34 34 34 - 34 34 34 34 34 34 34 34 34 34 34 34 - 34 34 34 34 34 34 38 38 38 42 42 42 - 54 54 54 82 82 82 94 86 76 91 60 6 -134 86 6 156 107 11 167 114 7 175 118 6 -175 118 6 167 114 7 152 99 6 121 87 25 -101 98 89 62 62 62 34 34 34 18 18 18 - 6 6 6 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 6 6 6 10 10 10 - 18 18 18 22 22 22 30 30 30 42 42 42 - 50 50 50 66 66 66 86 86 86 101 98 89 -106 86 58 98 70 6 104 69 6 104 69 6 -104 69 6 91 60 6 82 62 34 90 90 90 - 62 62 62 38 38 38 22 22 22 14 14 14 - 10 10 10 10 10 10 10 10 10 10 10 10 - 10 10 10 10 10 10 6 6 6 10 10 10 - 10 10 10 10 10 10 10 10 10 14 14 14 - 22 22 22 42 42 42 70 70 70 89 81 66 - 80 54 7 104 69 6 124 80 6 137 92 6 -134 86 6 116 81 8 100 82 52 86 86 86 - 58 58 58 30 30 30 14 14 14 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 10 10 10 14 14 14 - 18 18 18 26 26 26 38 38 38 54 54 54 - 70 70 70 86 86 86 94 86 76 89 81 66 - 89 81 66 86 86 86 74 74 74 50 50 50 - 30 30 30 14 14 14 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 18 18 18 34 34 34 58 58 58 - 82 82 82 89 81 66 89 81 66 89 81 66 - 94 86 66 94 86 76 74 74 74 50 50 50 - 26 26 26 14 14 14 6 6 6 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 6 6 6 6 6 6 14 14 14 18 18 18 - 30 30 30 38 38 38 46 46 46 54 54 54 - 50 50 50 42 42 42 30 30 30 18 18 18 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 6 6 6 14 14 14 26 26 26 - 38 38 38 50 50 50 58 58 58 58 58 58 - 54 54 54 42 42 42 30 30 30 18 18 18 - 10 10 10 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 6 6 6 10 10 10 14 14 14 18 18 18 - 18 18 18 14 14 14 10 10 10 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 6 6 6 - 14 14 14 18 18 18 22 22 22 22 22 22 - 18 18 18 14 14 14 10 10 10 6 6 6 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 +# CREATOR: GIMP PNM Filter Version 1.1 +60 80 +255 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +19 +32 +37 +35 +52 +61 +24 +38 +43 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +41 +58 +67 +124 +195 +230 +132 +208 +249 +132 +208 +249 +56 +74 +84 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +35 +52 +61 +132 +208 +249 +139 +215 +255 +134 +211 +252 +134 +211 +252 +132 +203 +238 +71 +106 +125 +62 +97 +116 +24 +38 +43 +1 +7 +11 +46 +66 +79 +35 +52 +61 +14 +27 +32 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +54 +54 +42 +129 +126 +96 +36 +36 +29 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +22 +13 +19 +35 +20 +27 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +31 +45 +53 +76 +116 +141 +113 +182 +217 +132 +203 +238 +88 +129 +148 +110 +145 +159 +111 +150 +170 +107 +163 +193 +125 +202 +243 +132 +208 +249 +125 +202 +243 +80 +130 +153 +134 +211 +252 +139 +215 +255 +115 +170 +200 +19 +32 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +36 +36 +29 +111 +110 +85 +182 +183 +140 +129 +126 +96 +81 +80 +62 +63 +64 +50 +58 +36 +48 +192 +108 +156 +169 +92 +132 +94 +52 +74 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +41 +58 +67 +125 +202 +243 +125 +202 +243 +101 +131 +146 +107 +126 +136 +197 +217 +216 +213 +234 +233 +213 +234 +233 +149 +166 +166 +101 +131 +146 +134 +211 +252 +134 +211 +252 +119 +178 +208 +114 +164 +188 +114 +164 +188 +125 +202 +243 +76 +116 +141 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +36 +36 +29 +36 +36 +29 +91 +90 +71 +129 +126 +96 +138 +135 +104 +197 +195 +148 +95 +86 +75 +192 +108 +156 +169 +92 +132 +156 +86 +123 +143 +82 +116 +78 +43 +62 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +22 +40 +49 +125 +202 +243 +134 +211 +252 +88 +129 +148 +159 +179 +189 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +149 +166 +166 +94 +147 +175 +119 +178 +208 +156 +176 +175 +213 +234 +233 +206 +227 +226 +123 +149 +157 +67 +89 +102 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +182 +183 +140 +197 +195 +148 +152 +152 +119 +130 +138 +107 +152 +152 +119 +182 +183 +140 +113 +73 +88 +192 +108 +156 +129 +75 +105 +201 +109 +161 +201 +109 +161 +175 +97 +138 +22 +13 +19 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +94 +147 +175 +132 +203 +238 +119 +178 +208 +149 +166 +166 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +104 +128 +133 +111 +150 +170 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +156 +176 +175 +24 +31 +32 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +36 +36 +29 +91 +93 +68 +146 +144 +112 +146 +144 +112 +182 +183 +140 +129 +126 +96 +169 +92 +132 +201 +109 +161 +129 +75 +105 +192 +108 +156 +192 +108 +156 +201 +109 +161 +62 +34 +50 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +22 +40 +49 +132 +208 +249 +134 +211 +252 +104 +128 +133 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +149 +166 +166 +104 +128 +133 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +104 +116 +120 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +25 +25 +18 +63 +64 +50 +111 +110 +85 +168 +169 +129 +197 +195 +148 +207 +205 +158 +81 +80 +62 +192 +108 +156 +175 +97 +138 +140 +78 +112 +186 +101 +146 +201 +109 +161 +201 +109 +161 +94 +52 +74 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +56 +74 +84 +134 +211 +252 +132 +208 +249 +107 +126 +136 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +197 +217 +216 +130 +146 +146 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +179 +200 +199 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +54 +54 +42 +146 +144 +112 +182 +183 +140 +182 +183 +140 +182 +183 +140 +197 +195 +148 +63 +64 +50 +192 +108 +156 +156 +86 +123 +156 +86 +123 +201 +109 +161 +192 +108 +156 +192 +108 +156 +156 +86 +123 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +80 +130 +153 +134 +211 +252 +125 +202 +243 +130 +146 +146 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +131 +137 +133 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +69 +83 +88 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +17 +19 +16 +25 +25 +18 +36 +36 +29 +111 +110 +85 +152 +152 +119 +207 +205 +158 +91 +90 +71 +201 +109 +161 +140 +78 +112 +156 +86 +123 +192 +108 +156 +201 +109 +161 +201 +109 +161 +192 +108 +156 +22 +13 +19 +0 +1 +0 +0 +1 +0 +1 +7 +11 +94 +147 +175 +134 +211 +252 +132 +208 +249 +123 +149 +157 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +206 +227 +226 +124 +133 +134 +206 +227 +226 +197 +217 +216 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +89 +103 +106 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +91 +90 +71 +207 +205 +158 +207 +205 +158 +207 +205 +158 +207 +205 +158 +207 +205 +158 +91 +90 +71 +192 +108 +156 +156 +86 +123 +140 +78 +112 +201 +109 +161 +192 +108 +156 +192 +108 +156 +192 +108 +156 +22 +13 +19 +0 +1 +0 +0 +1 +0 +0 +1 +0 +94 +147 +175 +132 +203 +238 +134 +211 +252 +123 +149 +157 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +169 +188 +187 +175 +180 +182 +213 +234 +233 +197 +217 +216 +149 +166 +166 +149 +166 +166 +66 +73 +75 +213 +234 +233 +213 +234 +233 +213 +234 +233 +213 +234 +233 +69 +83 +88 +29 +31 +13 +18 +17 +7 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +36 +36 +29 +95 +86 +75 +129 +126 +96 +146 +144 +112 +157 +158 +118 +197 +195 +148 +111 +110 +85 +192 +108 +156 +186 +101 +146 +140 +78 +112 +192 +108 +156 +201 +109 +161 +201 +109 +161 +192 +108 +156 +22 +13 +19 +0 +1 +0 +0 +1 +0 +7 +10 +5 +46 +77 +90 +91 +139 +151 +91 +139 +151 +69 +99 +114 +156 +176 +175 +213 +234 +233 +213 +234 +233 +213 +234 +233 +197 +217 +216 +21 +25 +26 +80 +98 +92 +213 +234 +233 +130 +146 +146 +130 +146 +146 +102 +116 +108 +57 +62 +64 +213 +234 +233 +213 +234 +233 +213 +234 +233 +206 +227 +226 +112 +115 +71 +218 +217 +87 +198 +196 +82 +78 +78 +32 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +17 +19 +16 +111 +110 +85 +152 +152 +119 +197 +195 +148 +207 +205 +158 +207 +205 +158 +157 +158 +118 +175 +97 +138 +192 +108 +156 +156 +86 +123 +201 +109 +161 +192 +108 +156 +192 +108 +156 +175 +97 +138 +28 +14 +21 +0 +1 +0 +0 +1 +0 +29 +31 +13 +112 +115 +71 +171 +171 +73 +188 +186 +80 +188 +186 +80 +147 +154 +89 +134 +158 +150 +206 +227 +226 +213 +234 +233 +213 +234 +233 +102 +116 +108 +186 +191 +194 +206 +227 +226 +131 +150 +119 +95 +103 +65 +89 +103 +106 +123 +149 +157 +213 +234 +233 +213 +234 +233 +213 +234 +233 +104 +128 +133 +171 +171 +73 +151 +153 +62 +188 +186 +80 +213 +214 +91 +102 +102 +43 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +81 +80 +62 +146 +144 +112 +152 +152 +119 +152 +152 +119 +168 +169 +129 +197 +195 +148 +113 +73 +88 +192 +108 +156 +156 +86 +123 +192 +108 +156 +201 +109 +161 +201 +109 +161 +175 +97 +138 +17 +19 +16 +8 +0 +0 +8 +0 +0 +171 +171 +73 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +131 +150 +119 +169 +188 +187 +213 +234 +233 +213 +234 +233 +179 +200 +199 +102 +116 +108 +151 +153 +62 +181 +180 +75 +218 +217 +87 +160 +161 +68 +75 +83 +79 +156 +176 +175 +110 +131 +112 +198 +196 +82 +181 +180 +75 +126 +125 +53 +160 +161 +68 +198 +196 +82 +160 +161 +68 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +54 +54 +42 +168 +169 +129 +207 +205 +158 +207 +205 +158 +207 +205 +158 +207 +205 +158 +95 +86 +75 +192 +108 +156 +175 +97 +138 +175 +97 +138 +192 +108 +156 +192 +108 +156 +140 +78 +112 +8 +0 +0 +7 +10 +5 +51 +49 +21 +218 +217 +87 +213 +214 +91 +126 +125 +53 +71 +72 +31 +95 +92 +39 +143 +144 +61 +213 +214 +91 +213 +214 +91 +171 +171 +73 +134 +149 +107 +131 +150 +119 +147 +154 +89 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +142 +145 +68 +115 +116 +48 +218 +217 +87 +200 +202 +87 +95 +92 +39 +200 +202 +87 +218 +217 +87 +198 +196 +82 +51 +49 +21 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +63 +64 +50 +54 +54 +42 +63 +64 +50 +129 +126 +96 +138 +135 +104 +138 +135 +104 +111 +110 +85 +186 +101 +146 +192 +108 +156 +129 +75 +105 +201 +109 +161 +201 +109 +161 +94 +52 +74 +8 +0 +0 +0 +1 +0 +18 +17 +7 +198 +196 +82 +207 +208 +85 +213 +214 +91 +218 +217 +87 +160 +161 +68 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +198 +196 +82 +95 +92 +39 +137 +135 +53 +71 +72 +31 +198 +196 +82 +213 +214 +91 +181 +180 +75 +38 +38 +15 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +31 +30 +24 +138 +135 +104 +207 +205 +158 +207 +205 +158 +207 +205 +158 +207 +205 +158 +207 +205 +158 +98 +76 +74 +192 +108 +156 +186 +101 +146 +156 +86 +123 +192 +108 +156 +58 +36 +48 +0 +1 +0 +0 +1 +0 +8 +0 +0 +57 +60 +26 +200 +202 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +160 +161 +68 +143 +144 +61 +151 +153 +62 +160 +161 +68 +160 +161 +68 +143 +144 +61 +106 +108 +49 +194 +191 +78 +213 +214 +91 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +95 +92 +39 +188 +186 +80 +213 +214 +91 +213 +214 +91 +106 +108 +49 +24 +25 +12 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +31 +30 +24 +182 +183 +140 +138 +135 +104 +138 +135 +104 +152 +152 +119 +157 +158 +118 +207 +205 +158 +157 +158 +118 +129 +75 +105 +192 +108 +156 +192 +108 +156 +186 +101 +146 +22 +13 +19 +0 +1 +0 +0 +1 +0 +10 +7 +13 +8 +0 +0 +38 +38 +15 +143 +144 +61 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +194 +191 +78 +171 +171 +73 +171 +171 +73 +188 +186 +80 +198 +196 +82 +143 +144 +61 +188 +186 +80 +213 +214 +91 +213 +214 +91 +218 +217 +87 +143 +144 +61 +188 +186 +80 +213 +214 +91 +106 +108 +49 +46 +45 +20 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +18 +17 +7 +36 +36 +29 +36 +36 +29 +111 +110 +85 +111 +110 +85 +138 +135 +104 +197 +195 +148 +111 +110 +85 +175 +97 +138 +201 +109 +161 +124 +69 +101 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +10 +7 +13 +8 +0 +0 +38 +38 +15 +91 +93 +68 +142 +145 +68 +194 +191 +78 +213 +214 +91 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +160 +161 +68 +137 +135 +53 +171 +171 +73 +143 +144 +61 +143 +144 +61 +213 +214 +91 +82 +89 +46 +8 +0 +0 +8 +0 +0 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +36 +36 +29 +182 +183 +140 +168 +169 +129 +111 +110 +85 +138 +135 +104 +146 +144 +112 +63 +64 +50 +58 +36 +48 +192 +108 +156 +78 +43 +62 +8 +0 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +7 +10 +5 +8 +0 +0 +8 +0 +0 +8 +0 +0 +78 +114 +134 +63 +93 +95 +102 +102 +43 +194 +191 +78 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +194 +191 +78 +213 +214 +91 +213 +214 +91 +142 +145 +68 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +0 +1 +0 +7 +10 +5 +54 +54 +42 +8 +0 +0 +63 +64 +50 +138 +135 +104 +18 +17 +7 +18 +17 +7 +35 +20 +27 +201 +109 +161 +78 +43 +62 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +56 +74 +84 +139 +215 +255 +104 +128 +133 +198 +196 +82 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +188 +186 +80 +79 +105 +103 +22 +40 +49 +21 +25 +26 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +81 +80 +62 +31 +30 +24 +1 +7 +11 +0 +1 +0 +35 +20 +27 +201 +109 +161 +94 +52 +74 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +10 +7 +13 +46 +77 +90 +134 +211 +252 +139 +215 +255 +102 +116 +108 +200 +202 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +218 +217 +87 +188 +186 +80 +69 +83 +88 +132 +203 +238 +64 +82 +92 +182 +218 +240 +85 +103 +114 +13 +18 +20 +32 +35 +37 +41 +58 +67 +21 +25 +26 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +17 +19 +16 +8 +0 +0 +0 +1 +0 +0 +1 +0 +22 +13 +19 +192 +108 +156 +129 +75 +105 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +53 +83 +97 +139 +215 +255 +134 +211 +252 +132 +203 +238 +80 +98 +92 +147 +154 +89 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +159 +168 +82 +110 +141 +132 +132 +203 +238 +139 +215 +255 +62 +97 +116 +173 +210 +232 +195 +232 +254 +127 +145 +154 +107 +126 +136 +195 +232 +254 +154 +185 +201 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +22 +13 +19 +186 +101 +146 +175 +97 +138 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +10 +7 +13 +62 +97 +116 +134 +211 +252 +139 +215 +255 +139 +215 +255 +132 +203 +238 +119 +178 +208 +104 +128 +133 +110 +141 +132 +131 +150 +119 +130 +146 +146 +105 +172 +207 +139 +215 +255 +134 +211 +252 +139 +215 +255 +115 +170 +200 +97 +115 +126 +195 +232 +254 +195 +232 +254 +97 +115 +126 +182 +218 +240 +195 +232 +254 +144 +175 +191 +21 +25 +26 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +8 +0 +0 +186 +101 +146 +186 +101 +146 +22 +13 +19 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +105 +172 +207 +134 +211 +252 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +134 +211 +252 +134 +211 +252 +134 +211 +252 +139 +215 +255 +46 +66 +79 +195 +232 +254 +195 +232 +254 +182 +218 +240 +123 +149 +157 +195 +232 +254 +195 +232 +254 +107 +126 +136 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +7 +10 +5 +8 +0 +0 +58 +36 +48 +62 +69 +87 +80 +130 +153 +46 +77 +90 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +32 +35 +37 +139 +215 +255 +134 +211 +252 +132 +208 +249 +134 +211 +252 +132 +208 +249 +132 +203 +238 +139 +215 +255 +134 +211 +252 +132 +203 +238 +132 +203 +238 +139 +215 +255 +132 +203 +238 +139 +215 +255 +132 +208 +249 +132 +203 +238 +94 +147 +175 +110 +145 +159 +195 +232 +254 +195 +232 +254 +130 +162 +179 +195 +232 +254 +195 +232 +254 +195 +232 +254 +69 +83 +88 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +31 +45 +53 +115 +170 +200 +132 +203 +238 +134 +211 +252 +134 +211 +252 +80 +130 +153 +35 +52 +61 +1 +7 +11 +0 +1 +0 +1 +7 +11 +19 +32 +37 +35 +52 +61 +101 +155 +185 +134 +211 +252 +139 +215 +255 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +125 +202 +243 +111 +150 +170 +133 +151 +162 +130 +146 +146 +119 +135 +143 +119 +135 +143 +111 +150 +170 +132 +208 +249 +134 +211 +252 +64 +82 +92 +195 +232 +254 +195 +232 +254 +182 +218 +240 +195 +232 +254 +195 +232 +254 +195 +232 +254 +182 +218 +240 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +16 +23 +29 +62 +97 +116 +132 +208 +249 +139 +215 +255 +132 +203 +238 +139 +215 +255 +139 +215 +255 +132 +208 +249 +139 +215 +255 +76 +116 +141 +21 +25 +26 +72 +122 +146 +132 +208 +249 +132 +208 +249 +119 +178 +208 +125 +202 +243 +132 +208 +249 +134 +211 +252 +132 +208 +249 +132 +203 +238 +110 +145 +159 +166 +171 +173 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +230 +232 +229 +145 +153 +161 +114 +164 +188 +75 +93 +104 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +163 +195 +211 +31 +45 +53 +16 +23 +29 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +14 +27 +32 +106 +171 +199 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +134 +211 +252 +90 +138 +161 +139 +215 +255 +134 +211 +252 +132 +203 +238 +72 +122 +146 +132 +203 +238 +134 +211 +252 +139 +215 +255 +139 +215 +255 +132 +203 +238 +132 +203 +238 +134 +211 +252 +139 +215 +255 +106 +128 +145 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +200 +205 +208 +55 +66 +72 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +163 +195 +211 +67 +89 +102 +90 +138 +161 +124 +195 +230 +124 +195 +230 +72 +122 +146 +14 +27 +32 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +56 +74 +84 +134 +211 +252 +132 +208 +249 +132 +208 +249 +132 +208 +249 +132 +203 +238 +139 +215 +255 +94 +147 +175 +119 +178 +208 +134 +211 +252 +92 +124 +151 +105 +172 +207 +132 +208 +249 +132 +208 +249 +132 +208 +249 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +120 +152 +168 +239 +241 +238 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +130 +146 +146 +182 +218 +240 +195 +232 +254 +195 +232 +254 +86 +117 +131 +88 +129 +148 +132 +208 +249 +132 +203 +238 +139 +215 +255 +139 +215 +255 +132 +203 +238 +46 +77 +90 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +38 +68 +82 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +62 +97 +116 +80 +130 +153 +46 +77 +90 +132 +203 +238 +139 +215 +255 +132 +203 +238 +139 +215 +255 +134 +211 +252 +139 +215 +255 +132 +203 +238 +119 +178 +208 +209 +211 +208 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +230 +232 +229 +104 +128 +133 +195 +232 +254 +71 +106 +125 +119 +188 +224 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +105 +172 +207 +62 +97 +116 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +13 +18 +20 +119 +188 +224 +134 +211 +252 +134 +211 +252 +132 +208 +249 +134 +211 +252 +132 +208 +249 +132 +208 +249 +124 +195 +230 +31 +45 +53 +134 +211 +252 +139 +215 +255 +139 +215 +255 +134 +211 +252 +134 +211 +252 +139 +215 +255 +139 +215 +255 +134 +211 +252 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +139 +157 +167 +139 +157 +167 +114 +164 +188 +132 +208 +249 +134 +211 +252 +134 +211 +252 +134 +211 +252 +132 +208 +249 +139 +215 +255 +119 +178 +208 +35 +52 +61 +19 +32 +37 +31 +45 +53 +31 +45 +53 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +62 +97 +116 +134 +211 +252 +134 +211 +252 +139 +215 +255 +134 +211 +252 +139 +215 +255 +139 +215 +255 +134 +211 +252 +107 +163 +193 +119 +188 +224 +134 +211 +252 +134 +211 +252 +134 +211 +252 +134 +211 +252 +132 +208 +249 +132 +208 +249 +124 +195 +230 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +75 +93 +104 +132 +208 +249 +139 +215 +255 +134 +211 +252 +134 +211 +252 +134 +211 +252 +139 +215 +255 +132 +208 +249 +139 +215 +255 +105 +172 +207 +53 +83 +97 +134 +211 +252 +132 +208 +249 +78 +114 +134 +16 +23 +29 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +13 +18 +20 +53 +83 +97 +132 +203 +238 +139 +215 +255 +134 +211 +252 +132 +208 +249 +132 +208 +249 +134 +211 +252 +132 +208 +249 +80 +130 +153 +124 +195 +230 +134 +211 +252 +132 +203 +238 +119 +188 +224 +139 +215 +255 +139 +215 +255 +134 +211 +252 +114 +164 +188 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +111 +150 +170 +132 +208 +249 +132 +203 +238 +139 +215 +255 +132 +203 +238 +115 +170 +200 +101 +155 +185 +119 +188 +224 +132 +208 +249 +139 +215 +255 +107 +163 +193 +72 +122 +146 +139 +215 +255 +132 +203 +238 +78 +114 +134 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +16 +23 +29 +105 +172 +207 +88 +129 +148 +139 +215 +255 +132 +208 +249 +134 +211 +252 +113 +182 +217 +90 +138 +161 +106 +171 +199 +139 +215 +255 +62 +97 +116 +134 +211 +252 +139 +215 +255 +139 +215 +255 +90 +138 +161 +132 +208 +249 +132 +208 +249 +134 +211 +252 +120 +152 +168 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +220 +222 +219 +106 +171 +199 +124 +195 +230 +76 +116 +141 +56 +74 +84 +64 +82 +92 +106 +128 +145 +133 +151 +162 +110 +145 +159 +56 +74 +84 +105 +172 +207 +139 +215 +255 +78 +114 +134 +134 +211 +252 +139 +215 +255 +125 +202 +243 +19 +32 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +41 +58 +67 +134 +211 +252 +72 +122 +146 +134 +211 +252 +132 +203 +238 +75 +93 +104 +160 +93 +131 +186 +101 +146 +106 +110 +136 +119 +188 +224 +71 +106 +125 +134 +211 +252 +134 +211 +252 +139 +215 +255 +90 +138 +161 +139 +215 +255 +134 +211 +252 +139 +215 +255 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +175 +180 +182 +55 +66 +72 +163 +195 +211 +195 +232 +254 +195 +232 +254 +195 +232 +254 +182 +218 +240 +195 +232 +254 +127 +145 +154 +105 +172 +207 +125 +202 +243 +90 +138 +161 +134 +211 +252 +134 +211 +252 +139 +215 +255 +38 +68 +82 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +38 +68 +82 +134 +211 +252 +105 +172 +207 +107 +163 +193 +132 +208 +249 +88 +107 +129 +192 +108 +156 +201 +109 +161 +62 +69 +87 +62 +97 +116 +132 +203 +238 +132 +203 +238 +134 +211 +252 +132 +208 +249 +76 +116 +141 +139 +215 +255 +132 +203 +238 +134 +211 +252 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +209 +211 +208 +127 +145 +154 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +69 +83 +88 +173 +210 +232 +195 +232 +254 +86 +117 +131 +106 +171 +199 +139 +215 +255 +139 +215 +255 +132 +203 +238 +132 +208 +249 +62 +97 +116 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +41 +58 +67 +132 +203 +238 +139 +215 +255 +119 +178 +208 +134 +211 +252 +92 +124 +151 +175 +97 +138 +201 +109 +161 +77 +75 +95 +139 +215 +255 +139 +215 +255 +139 +215 +255 +132 +208 +249 +107 +163 +193 +76 +116 +141 +139 +215 +255 +139 +215 +255 +139 +215 +255 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +155 +160 +162 +163 +195 +211 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +123 +149 +157 +144 +175 +191 +195 +232 +254 +101 +131 +146 +124 +195 +230 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +38 +68 +82 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +22 +40 +49 +132 +208 +249 +134 +211 +252 +139 +215 +255 +139 +215 +255 +94 +147 +175 +175 +97 +138 +192 +108 +156 +92 +73 +98 +134 +211 +252 +101 +155 +185 +31 +45 +53 +21 +25 +26 +13 +18 +20 +71 +106 +125 +134 +211 +252 +139 +215 +255 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +116 +123 +125 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +110 +145 +159 +173 +210 +232 +195 +232 +254 +173 +210 +232 +80 +130 +153 +134 +211 +252 +132 +208 +249 +134 +211 +252 +119 +188 +224 +16 +23 +29 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +71 +106 +125 +134 +211 +252 +134 +211 +252 +139 +215 +255 +105 +172 +207 +156 +86 +123 +201 +109 +161 +105 +72 +102 +46 +66 +79 +13 +18 +20 +8 +0 +0 +8 +0 +0 +8 +0 +0 +76 +116 +141 +132 +208 +249 +139 +215 +255 +139 +215 +255 +130 +162 +179 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +230 +232 +229 +119 +135 +143 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +107 +126 +136 +195 +232 +254 +195 +232 +254 +195 +232 +254 +85 +103 +114 +134 +211 +252 +132 +203 +238 +124 +195 +230 +31 +45 +53 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +56 +74 +84 +115 +170 +200 +132 +208 +249 +124 +195 +230 +129 +75 +105 +201 +109 +161 +140 +78 +112 +1 +7 +11 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +78 +114 +134 +134 +211 +252 +134 +211 +252 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +152 +154 +151 +173 +210 +232 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +107 +126 +136 +195 +232 +254 +195 +232 +254 +195 +232 +254 +163 +195 +211 +88 +129 +148 +80 +130 +153 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +1 +7 +11 +1 +7 +11 +10 +7 +13 +32 +35 +37 +22 +40 +49 +113 +65 +94 +192 +108 +156 +175 +97 +138 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +92 +124 +151 +139 +215 +255 +134 +211 +252 +139 +215 +255 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +116 +123 +125 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +144 +175 +191 +130 +162 +179 +195 +232 +254 +195 +232 +254 +195 +232 +254 +64 +82 +92 +1 +7 +11 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +1 +7 +11 +1 +7 +11 +1 +7 +11 +1 +7 +11 +94 +52 +74 +192 +108 +156 +192 +108 +156 +28 +14 +21 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +107 +163 +193 +132 +208 +249 +134 +211 +252 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +200 +205 +208 +144 +175 +191 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +97 +115 +126 +195 +232 +254 +195 +232 +254 +195 +232 +254 +182 +218 +240 +31 +45 +53 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +0 +1 +0 +94 +52 +74 +201 +109 +161 +192 +108 +156 +45 +26 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +13 +18 +20 +132 +203 +238 +132 +208 +249 +134 +211 +252 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +140 +145 +147 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +104 +128 +133 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +123 +149 +157 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +78 +43 +62 +192 +108 +156 +201 +109 +161 +62 +34 +50 +0 +1 +0 +0 +1 +0 +0 +1 +0 +31 +45 +53 +139 +215 +255 +132 +203 +238 +134 +211 +252 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +123 +149 +157 +195 +232 +254 +195 +232 +254 +195 +232 +254 +163 +195 +211 +134 +160 +169 +120 +152 +168 +144 +175 +191 +182 +218 +240 +120 +152 +168 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +89 +103 +106 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +62 +34 +50 +192 +108 +156 +201 +109 +161 +78 +43 +62 +0 +1 +0 +0 +1 +0 +1 +7 +11 +71 +106 +125 +134 +211 +252 +139 +215 +255 +134 +211 +252 +139 +215 +255 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +133 +151 +162 +144 +175 +191 +101 +131 +146 +72 +122 +146 +106 +171 +199 +124 +195 +230 +124 +195 +230 +119 +178 +208 +78 +114 +134 +64 +82 +92 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +195 +232 +254 +85 +103 +114 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +45 +26 +37 +201 +109 +161 +192 +108 +156 +94 +52 +74 +0 +1 +0 +7 +10 +5 +16 +23 +29 +125 +202 +243 +134 +211 +252 +132 +208 +249 +139 +215 +255 +119 +188 +224 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +220 +222 +219 +75 +93 +104 +114 +164 +188 +139 +215 +255 +132 +203 +238 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +94 +147 +175 +144 +175 +191 +195 +232 +254 +195 +232 +254 +154 +185 +201 +75 +93 +104 +64 +82 +92 +75 +93 +104 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +35 +20 +27 +201 +109 +161 +192 +108 +156 +124 +69 +101 +8 +0 +0 +7 +10 +5 +71 +106 +125 +139 +215 +255 +139 +215 +255 +132 +208 +249 +139 +215 +255 +116 +162 +194 +239 +241 +238 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +104 +116 +120 +139 +215 +255 +132 +203 +238 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +56 +74 +84 +123 +149 +157 +35 +52 +61 +62 +97 +116 +124 +195 +230 +139 +215 +255 +139 +215 +255 +132 +203 +238 +53 +83 +97 +13 +18 +20 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +35 +20 +27 +192 +108 +156 +201 +109 +161 +160 +93 +131 +10 +7 +13 +35 +52 +61 +134 +211 +252 +134 +211 +252 +132 +208 +249 +139 +215 +255 +132 +208 +249 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +94 +99 +103 +88 +129 +148 +124 +195 +230 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +119 +188 +224 +106 +171 +199 +113 +182 +217 +106 +171 +199 +41 +58 +67 +132 +203 +238 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +101 +155 +185 +24 +38 +43 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +22 +13 +19 +192 +108 +156 +201 +109 +161 +186 +101 +146 +22 +13 +19 +101 +155 +185 +132 +203 +238 +139 +215 +255 +132 +208 +249 +132 +208 +249 +132 +208 +249 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +104 +116 +120 +132 +203 +238 +139 +215 +255 +182 +218 +240 +119 +178 +208 +107 +126 +136 +124 +133 +134 +116 +123 +125 +76 +83 +91 +145 +153 +161 +127 +145 +154 +75 +93 +104 +41 +58 +67 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +132 +203 +238 +119 +188 +224 +107 +163 +193 +24 +38 +43 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +22 +13 +19 +186 +101 +146 +201 +109 +161 +201 +109 +161 +62 +69 +87 +139 +215 +255 +139 +215 +255 +132 +203 +238 +139 +215 +255 +134 +211 +252 +125 +202 +243 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +97 +115 +126 +139 +215 +255 +114 +164 +188 +110 +145 +159 +195 +197 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +116 +123 +125 +182 +202 +214 +182 +202 +214 +75 +93 +104 +116 +162 +194 +139 +215 +255 +139 +215 +255 +139 +215 +255 +132 +203 +238 +88 +107 +129 +75 +93 +104 +75 +93 +104 +107 +126 +136 +106 +128 +145 +75 +93 +104 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +8 +0 +0 +175 +97 +138 +192 +108 +156 +201 +109 +161 +117 +117 +150 +139 +215 +255 +139 +215 +255 +134 +211 +252 +139 +215 +255 +132 +208 +249 +102 +125 +154 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +239 +241 +238 +85 +103 +114 +145 +153 +161 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +119 +135 +143 +144 +175 +191 +85 +103 +114 +124 +195 +230 +139 +215 +255 +139 +215 +255 +154 +185 +201 +56 +74 +84 +75 +93 +104 +159 +179 +189 +182 +202 +214 +182 +202 +214 +182 +202 +214 +182 +202 +214 +55 +66 +72 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +24 +38 +43 +90 +138 +161 +14 +27 +32 +0 +1 +0 +8 +0 +0 +7 +10 +5 +129 +75 +105 +201 +109 +161 +186 +101 +146 +117 +117 +150 +134 +211 +252 +139 +215 +255 +132 +203 +238 +139 +215 +255 +139 +215 +255 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +166 +171 +173 +68 +76 +83 +139 +215 +255 +139 +215 +255 +115 +170 +200 +56 +74 +84 +85 +103 +114 +169 +189 +201 +182 +202 +214 +182 +202 +214 +182 +202 +214 +182 +202 +214 +159 +179 +189 +55 +66 +72 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +10 +7 +13 +38 +68 +82 +139 +215 +255 +113 +182 +217 +35 +52 +61 +8 +0 +0 +7 +10 +5 +58 +36 +48 +201 +109 +161 +175 +97 +138 +116 +162 +194 +132 +208 +249 +139 +215 +255 +139 +215 +255 +139 +215 +255 +116 +162 +194 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +140 +145 +147 +139 +157 +167 +145 +153 +161 +49 +53 +55 +182 +202 +214 +182 +202 +214 +182 +202 +214 +182 +202 +214 +182 +202 +214 +169 +189 +201 +76 +83 +91 +21 +25 +26 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +56 +74 +84 +139 +215 +255 +139 +215 +255 +134 +211 +252 +107 +163 +193 +38 +68 +82 +56 +74 +84 +92 +124 +151 +116 +162 +194 +139 +215 +255 +139 +215 +255 +132 +208 +249 +132 +208 +249 +132 +208 +249 +140 +145 +147 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +166 +171 +173 +133 +151 +162 +182 +202 +214 +163 +195 +211 +97 +115 +126 +49 +53 +55 +22 +13 +19 +1 +7 +11 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +41 +58 +67 +134 +211 +252 +134 +211 +252 +132 +208 +249 +119 +188 +224 +53 +83 +97 +132 +203 +238 +139 +215 +255 +132 +203 +238 +134 +211 +252 +139 +215 +255 +132 +203 +238 +139 +215 +255 +116 +162 +194 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +49 +53 +55 +49 +53 +55 +13 +18 +20 +1 +7 +11 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +31 +45 +53 +134 +211 +252 +139 +215 +255 +132 +203 +238 +46 +66 +79 +119 +188 +224 +139 +215 +255 +134 +211 +252 +139 +215 +255 +139 +215 +255 +134 +211 +252 +139 +215 +255 +139 +215 +255 +106 +110 +136 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +94 +99 +103 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +16 +23 +29 +119 +188 +224 +139 +215 +255 +94 +147 +175 +72 +122 +146 +134 +211 +252 +132 +203 +238 +139 +215 +255 +132 +208 +249 +134 +211 +252 +134 +211 +252 +139 +215 +255 +119 +178 +208 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +116 +123 +125 +1 +7 +11 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +90 +138 +161 +134 +211 +252 +38 +68 +82 +125 +202 +243 +139 +215 +255 +139 +215 +255 +134 +211 +252 +139 +215 +255 +132 +203 +238 +134 +211 +252 +134 +211 +252 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +116 +123 +125 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +35 +52 +61 +113 +182 +217 +46 +77 +90 +134 +211 +252 +132 +208 +249 +134 +211 +252 +132 +203 +238 +139 +215 +255 +139 +215 +255 +134 +211 +252 +119 +188 +224 +175 +180 +182 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +140 +145 +147 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +14 +27 +32 +71 +106 +125 +94 +147 +175 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +139 +215 +255 +134 +211 +252 +134 +211 +252 +120 +152 +168 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +140 +145 +147 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +1 +7 +11 +10 +7 +13 +119 +188 +224 +134 +211 +252 +132 +208 +249 +134 +211 +252 +132 +208 +249 +132 +203 +238 +139 +215 +255 +134 +211 +252 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +124 +133 +134 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +7 +10 +5 +132 +203 +238 +139 +215 +255 +132 +203 +238 +134 +211 +252 +139 +215 +255 +139 +215 +255 +139 +215 +255 +119 +188 +224 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +94 +99 +103 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +10 +7 +13 +19 +32 +37 +134 +211 +252 +139 +215 +255 +139 +215 +255 +134 +211 +252 +134 +211 +252 +134 +211 +252 +139 +215 +255 +116 +162 +194 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +76 +83 +91 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +19 +32 +37 +132 +203 +238 +134 +211 +252 +134 +211 +252 +134 +211 +252 +134 +211 +252 +139 +215 +255 +139 +215 +255 +117 +117 +150 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +230 +232 +229 +32 +35 +37 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +16 +23 +29 +132 +208 +249 +132 +203 +238 +139 +215 +255 +134 +211 +252 +132 +208 +249 +139 +215 +255 +139 +215 +255 +130 +162 +179 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +140 +145 +147 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +10 +7 +13 +111 +150 +170 +139 +215 +255 +132 +203 +238 +139 +215 +255 +132 +203 +238 +134 +211 +252 +139 +215 +255 +133 +151 +162 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +68 +76 +83 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +46 +66 +79 +134 +211 +252 +132 +203 +238 +139 +215 +255 +139 +215 +255 +134 +211 +252 +139 +215 +255 +111 +150 +170 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +124 +133 +134 +10 +7 +13 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +13 +18 +20 +116 +162 +194 +139 +215 +255 +139 +215 +255 +134 +211 +252 +139 +215 +255 +134 +211 +252 +116 +162 +194 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +186 +191 +194 +49 +53 +55 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +1 +7 +11 +19 +32 +37 +119 +188 +224 +134 +211 +252 +132 +208 +249 +139 +215 +255 +139 +215 +255 +134 +211 +252 +186 +191 +194 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +200 +205 +208 +200 +205 +208 +239 +241 +238 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +186 +191 +194 +35 +20 +27 +51 +49 +21 +57 +60 +26 +78 +78 +32 +106 +108 +49 +115 +116 +48 +115 +116 +48 +95 +92 +39 +57 +60 +26 +29 +31 +13 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +24 +25 +12 +24 +25 +12 +29 +31 +13 +29 +31 +13 +17 +19 +16 +101 +131 +146 +139 +215 +255 +134 +211 +252 +134 +211 +252 +139 +215 +255 +145 +153 +161 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +200 +205 +208 +94 +99 +103 +45 +26 +37 +113 +73 +88 +116 +123 +125 +140 +145 +147 +183 +186 +182 +230 +232 +229 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +239 +241 +238 +131 +137 +133 +106 +108 +49 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +126 +125 +53 +38 +38 +15 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +38 +38 +15 +71 +72 +31 +115 +116 +48 +160 +161 +68 +198 +196 +82 +207 +208 +85 +213 +214 +91 +213 +214 +91 +213 +214 +91 +160 +161 +68 +91 +93 +68 +101 +131 +146 +125 +202 +243 +134 +211 +252 +119 +188 +224 +166 +171 +173 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +249 +251 +248 +200 +205 +208 +106 +110 +136 +98 +76 +74 +28 +14 +21 +8 +0 +0 +0 +1 +0 +8 +0 +0 +8 +0 +0 +28 +14 +21 +28 +14 +21 +46 +45 +20 +98 +76 +74 +113 +114 +94 +140 +145 +147 +140 +145 +147 +155 +160 +162 +140 +145 +147 +125 +128 +113 +134 +137 +80 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +200 +202 +87 +160 +161 +68 +181 +180 +75 +218 +217 +87 +151 +153 +62 +7 +10 +5 +0 +1 +0 +0 +1 +0 +38 +38 +15 +137 +135 +53 +213 +214 +91 +218 +217 +87 +218 +217 +87 +181 +180 +75 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +188 +186 +80 +102 +102 +43 +95 +103 +65 +109 +113 +104 +80 +98 +92 +113 +73 +88 +140 +145 +147 +116 +123 +125 +98 +76 +74 +95 +86 +75 +111 +110 +85 +51 +49 +21 +8 +0 +0 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +126 +125 +53 +218 +217 +87 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +160 +161 +68 +51 +49 +21 +8 +0 +0 +8 +0 +0 +0 +1 +0 +8 +0 +0 +102 +102 +43 +213 +214 +91 +213 +214 +91 +213 +214 +91 +115 +116 +48 +188 +186 +80 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +188 +186 +80 +8 +0 +0 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +137 +135 +53 +213 +214 +91 +213 +214 +91 +218 +217 +87 +198 +196 +82 +213 +214 +91 +198 +196 +82 +218 +217 +87 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +194 +191 +78 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +78 +78 +32 +7 +10 +5 +0 +1 +0 +7 +10 +5 +8 +0 +0 +18 +17 +7 +38 +38 +15 +36 +36 +29 +160 +161 +68 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +213 +214 +91 +188 +186 +80 +102 +102 +43 +46 +45 +20 +29 +31 +13 +0 +1 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +28 +14 +21 +46 +45 +20 +51 +49 +21 +95 +92 +39 +95 +92 +39 +95 +92 +39 +126 +125 +53 +143 +144 +61 +194 +191 +78 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +126 +125 +53 +106 +108 +49 +102 +102 +43 +115 +116 +48 +115 +116 +48 +126 +125 +53 +115 +116 +48 +71 +72 +31 +8 +0 +0 +8 +0 +0 +0 +1 +0 +8 +0 +0 +8 +0 +0 +8 +0 +0 +29 +31 +13 +218 +217 +87 +213 +214 +91 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +171 +171 +73 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +213 +214 +91 +218 +217 +87 +213 +214 +91 +126 +125 +53 +51 +49 +21 +31 +30 +24 +28 +14 +21 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +8 +0 +0 +8 +0 +0 +28 +14 +21 +28 +14 +21 +8 +0 +0 +28 +14 +21 +28 +14 +21 +28 +14 +21 +22 +13 +19 +46 +45 +20 +71 +72 +31 +126 +125 +53 +198 +196 +82 +218 +217 +87 +218 +217 +87 +218 +217 +87 +218 +217 +87 +78 +78 +32 +22 +13 +19 +8 +0 +0 +10 +7 +13 +8 +0 +0 +8 +0 +0 +7 +10 +5 +0 +1 +0 +0 +1 +0 +0 +1 +0 +1 +7 +11 +0 +1 +0 +0 +1 +0 +22 +13 +19 +95 +92 +39 +126 +125 +53 +95 +92 +39 +78 +78 +32 +71 +72 +31 +71 +72 +31 +198 +196 +82 +218 +217 +87 +218 +217 +87 +218 +217 +87 +213 +214 +91 +213 +214 +91 +126 +125 +53 +31 +30 +24 +22 +13 +19 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +8 +0 +0 +8 +0 +0 +22 +13 +19 +36 +36 +29 +71 +72 +31 +160 +161 +68 +213 +214 +91 +213 +214 +91 +200 +202 +87 +24 +25 +12 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +10 +7 +13 +8 +0 +0 +22 +13 +19 +8 +0 +0 +22 +13 +19 +10 +7 +13 +143 +144 +61 +218 +217 +87 +213 +214 +91 +218 +217 +87 +160 +161 +68 +36 +36 +29 +8 +0 +0 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +7 +10 +5 +8 +0 +0 +1 +7 +11 +8 +0 +0 +8 +0 +0 +10 +7 +13 +22 +13 +19 +22 +13 +19 +38 +38 +15 +61 +63 +26 +51 +49 +21 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +1 +7 +11 +8 +0 +0 +8 +0 +0 +18 +17 +7 +35 +20 +27 +51 +49 +21 +71 +72 +31 +46 +45 +20 +18 +17 +7 +8 +0 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +8 +0 +0 +1 +7 +11 +8 +0 +0 +10 +7 +13 +10 +7 +13 +8 +0 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 +0 +1 +0 diff --git a/drivers/video/logo/logo_linux_mono.pbm b/drivers/video/logo/logo_linux_mono.pbm index 2f14d9fedf6..664de625955 100644 --- a/drivers/video/logo/logo_linux_mono.pbm +++ b/drivers/video/logo/logo_linux_mono.pbm @@ -1,203 +1,72 @@ P1 -# Standard black and white Linux logo -80 80 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 0 0 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 0 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 -1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 0 0 1 1 1 1 1 1 0 0 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 0 0 1 1 1 1 0 0 0 0 0 0 1 1 -1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -0 1 1 0 0 1 1 1 0 0 1 1 0 0 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 1 1 1 0 1 1 1 1 0 0 1 -1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 1 1 1 0 1 1 1 0 1 1 1 1 0 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 -1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -0 1 1 0 0 0 0 0 0 0 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 1 0 0 1 0 0 0 0 0 1 1 1 -1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 1 -1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 0 0 0 0 0 1 1 0 0 0 0 1 1 -1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 1 1 1 1 1 1 1 0 0 0 0 1 1 0 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 0 1 0 0 0 0 0 0 0 0 1 1 0 0 0 1 -1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 -0 0 1 0 0 0 0 0 1 1 0 0 0 0 0 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 -1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 -0 0 0 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 0 1 1 1 1 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 1 1 1 -1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 0 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 0 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 1 1 1 1 1 1 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 -1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 1 1 1 1 0 1 1 1 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 1 0 1 1 1 1 1 1 0 1 1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 1 0 0 0 1 1 1 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 1 0 0 0 1 1 1 1 1 1 1 1 1 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 1 1 1 1 1 1 1 0 -0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 1 1 1 1 1 1 1 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 1 1 1 1 0 0 0 -0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 -1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 -1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 -0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 -0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 -1 1 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 -1 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 0 0 1 0 0 0 0 0 0 0 0 0 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 +# CREATOR: GIMP PNM Filter Version 1.1 +60 80 +1111111111111111111111111111111111111111111111111111111111111111111111 +1111111111111110001111111111111111111111111111111111111111111111111111 +1111000001111111111111111111111111111111111111111111111111111110011110 +0010011111111111111111111111111111011110011111111110011000110001101111 +1111111111111111111111111101001111111111001100000110100111111111111111 +1111111111100111010100011111110010000000110000111111111111111111111111 +1111101001000111111001000000001100000111111111111111111111111111001001 +0001111110010000000001000000111111111111111111111111000010110001111110 +0100000000010000001111111111111111111111111110101100001111100100000000 +0100000011111111111111111111111000001011000011111001000001100111000011 +1111111111111111111111111010010000111111111000011011110000100111111111 +1111111111111000100100001111110011000110111100011100111111111111111111 +1111110101000011110000001100011001111001101111111111111111111100000100 +0001111001111000111000001100100011111111111111111111111111001001111000 +0100000000000001110001111111111111111111111000001001011111000011111110 +0000010001111111111111111111111101111011000111111100000110010000100111 +1111111111111111111111111111010011111111111000000011111011111111111111 +1111111111111101111110111111111111100000000001111111111111111111111111 +1111111111101111111111101000000000011111111111111111111111111111111111 +1011111111111001000000001010111111111111111111111111111111111011111111 +1110001100001100100110111111111111111111111111111110011111111110000111 +1100000100100111111111111111111111111111100111111111100000000000000100 +0100111111111111111111111111111111111111110000000000000001100100011111 +1111111111111111111111000111111110000000111111001000000011111111111111 +1111111111000000011100000000110000001110000000111111111111111111111110 +0000100010000000010000000000100000110011111111111111111111000000100110 +0000001000000000001000110000011111111111111111110000000111000000000000 +0000000001010000001111111111111111111100000000100000000100000000000001 +1100000001111111111111111111100000001000000001000000000000001000000001 +1001111111111111111000000010000000100000000000000010000110001100111111 +1111111110000110100010001000000000000000101111111101000111111111111010 +0110101000100010000000000000001100000011010001111111111110010100110000 +1000100000000000000001000010011000011111111111100001001000011000100000 +0000000000100000110100000111111111111000010010111110001000000000000000 +1000001000100001111111111111000110111111100010000000000000001000001000 +1000111111111111111100101111111000100000000000000100000010000111111111 +1111111111111001111110001000000000000001000000110001111111111111111111 +1110011111100010000000000000110000000100001111111111111111111110011111 +0000100000000000001000000001000011111111111111111111100111110000100000 +0000000010000111010000011111111111111111111001111100001000000000000011 +1110001100000011111111111111111110011110000010000000000000110000000110 +0111111111111111111111100111100001000000000000010000000000111100001111 +1111111111111000110000010000000000000110000000101100000001111111111111 +1110001100000100000000000001000011111111000000001111111111111110001000 +0001000000000000010111000100110000111111111111111111100010000010000000 +0000000110000001010000110000011111111111111100100000100000000000000000 +0000011001110000001111111111001111000000010000000000000000000000001111 +0000001111111111110001111100000100000000000000000000000000011001111111 +1111111100001000000011000000000000000000000000000011111111111111111100 +0100000000100000000000000000000000000000111111111111111111001100000000 +1000000000000000000000000000001111111111111111111010000000010000000000 +0000000000000000000011111111111111111110100000000100000000000000000000 +0000000000111111111111111111111000000010000000000000000000000000000000 +1111111111111111111100000000100000000000000000000000000000001111111111 +1111111111000000001000000000000000000000000000000011111111111111111111 +0000000100000000000000000000000000000000111111111111111111110000000100 +0000000000000000000000000000001111111111111111111100000001000000000000 +0000000000000000000111111111111111111111100000010000000000000000000000 +0000000001111111111111111111111000000100000000000000000000000000000011 +1111111111111111111111000001000000000000000000000000000001111111111111 +1111111111110000001000000000000001000000000000111111111111111111111111 +1111000010000000000111111100000000011000000000011111111111000001110001 +0000001111111111111111111000000000011011111100010000000011111111111111 +1111110000000000000000000011111110001000000000000000000001111111110000 +0000000000000000001111111110000000000000000111111111111111111110000000 +0111111111111111000000100000001111111111111111111111111110000011111111 +1111111111110000001111111111111111111111111111111100011111111111111111 +1110001111111111111111111111111111111111111111111111111111111111111111 +1111111111111111111111111111111111111111 \ No newline at end of file diff --git a/drivers/video/logo/logo_linux_vga16.ppm b/drivers/video/logo/logo_linux_vga16.ppm index 1850c15e6fe..8615deb9444 100644 --- a/drivers/video/logo/logo_linux_vga16.ppm +++ b/drivers/video/logo/logo_linux_vga16.ppm @@ -1,1604 +1,14404 @@ P3 -# Standard 16-color Linux logo -80 80 -255 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 85 85 85 85 85 85 85 85 85 - 85 85 85 85 85 85 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 85 85 85 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 170 170 170 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 170 170 170 170 170 85 85 85 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 170 170 170 170 170 -170 170 170 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 85 85 85 170 170 170 170 170 170 170 170 170 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 170 170 170 255 255 255 255 255 255 -255 255 255 170 170 170 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 -170 170 170 170 170 170 255 255 255 255 255 255 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 170 170 255 255 255 170 170 170 170 170 170 -255 255 255 170 170 170 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 -170 170 170 0 0 0 0 0 0 255 255 255 - 85 85 85 0 0 0 0 0 0 0 0 0 -255 255 255 170 170 170 0 0 0 85 85 85 -170 170 170 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 - 85 85 85 0 0 0 0 0 0 170 170 170 - 85 85 85 0 0 0 0 0 0 0 0 0 -255 255 255 85 85 85 0 0 0 0 0 0 - 85 85 85 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 -170 170 170 0 0 0 0 0 0 170 170 170 - 85 85 85 85 85 85 85 85 85 85 85 85 -255 255 255 85 85 85 0 0 0 0 0 0 - 85 85 85 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 -255 255 255 0 0 0 0 0 0 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 0 0 0 0 0 0 - 85 85 85 255 255 255 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 -170 170 170 170 170 170 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 170 170 170 170 170 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 170 85 0 -170 85 0 170 85 0 85 85 85 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 - 85 85 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 85 85 85 0 0 0 - 0 0 0 85 85 85 170 170 170 85 85 85 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 - 85 85 85 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 170 170 170 170 170 170 170 170 0 0 0 - 0 0 0 0 0 0 170 170 170 170 170 170 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 85 85 85 170 170 170 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 170 170 170 170 170 -170 170 170 170 170 170 170 170 170 85 85 85 - 0 0 0 0 0 0 85 85 85 85 85 85 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 85 85 85 170 170 170 170 170 170 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 170 170 170 170 170 170 170 170 170 170 170 -255 255 255 255 255 255 255 255 255 170 170 170 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 85 85 85 -255 255 255 255 255 255 170 170 170 170 170 170 -170 170 170 170 170 170 170 170 170 170 170 170 -170 170 170 170 170 170 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 170 170 170 -255 255 255 255 255 255 170 170 170 170 170 170 -170 170 170 170 170 170 170 170 170 170 170 170 -170 170 170 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 85 85 85 255 255 255 -255 255 255 255 255 255 255 255 255 170 170 170 -170 170 170 170 170 170 170 170 170 170 170 170 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 85 85 85 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 85 85 85 170 170 170 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 170 170 170 170 170 170 170 170 170 -255 255 255 255 255 255 255 255 255 170 170 170 -170 170 170 170 170 170 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 170 170 170 170 170 170 170 170 170 -170 170 170 170 170 170 170 170 170 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 170 170 170 170 170 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 170 170 170 -170 170 170 170 170 170 170 170 170 85 85 85 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 170 170 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 170 170 170 170 170 170 - 0 0 0 0 0 0 0 0 0 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 85 85 85 0 0 0 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 170 170 170 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 170 170 170 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 - 0 0 0 0 0 0 85 85 85 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 0 0 0 85 85 85 - 85 85 85 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 - 0 0 0 85 85 85 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 0 0 0 85 85 85 - 85 85 85 0 0 0 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 85 85 85 - 0 0 0 170 170 170 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 85 85 85 0 0 0 - 0 0 0 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 85 85 85 0 0 0 - 85 85 85 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 85 85 85 0 0 0 0 0 0 -170 170 170 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 0 0 0 - 85 85 85 85 85 85 85 85 85 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 0 0 0 170 85 0 -255 255 85 170 85 0 0 0 0 0 0 0 - 85 85 85 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 85 85 85 85 85 85 0 0 0 - 0 0 0 85 85 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 0 0 0 - 0 0 0 85 85 85 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 85 170 85 0 255 255 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 - 0 0 0 0 0 0 85 85 85 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 85 -170 85 0 255 255 85 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 0 0 0 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 170 85 0 -255 255 85 170 85 0 255 255 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -255 255 85 170 85 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 0 0 0 0 0 0 0 0 0 - 85 85 85 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 85 -170 85 0 255 255 85 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 255 255 85 -170 85 0 255 255 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 0 0 0 0 0 0 - 0 0 0 85 85 85 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 0 0 0 - 0 0 0 0 0 0 85 85 85 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 170 170 170 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 170 170 170 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 170 170 170 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 170 170 170 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 0 0 0 0 0 0 0 0 0 - 85 85 85 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 170 170 170 85 85 85 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 85 85 85 85 85 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 85 85 85 85 85 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 170 170 170 - 85 85 85 0 0 0 0 0 0 170 85 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 85 85 85 - 0 0 0 0 0 0 0 0 0 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 170 170 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -170 170 170 85 85 85 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 170 85 0 -170 85 0 170 170 170 255 255 255 255 255 255 -255 255 255 255 255 255 255 255 255 255 255 255 -255 255 255 255 255 255 170 170 170 85 85 85 - 85 85 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 170 85 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 170 85 0 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 85 85 85 170 85 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 170 85 0 170 85 0 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 85 85 85 85 85 85 85 85 85 85 85 85 - 85 85 85 85 85 85 85 85 85 85 85 85 - 85 85 85 85 85 85 85 85 85 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -170 85 0 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 255 255 85 170 85 0 -170 85 0 170 85 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 255 255 85 170 85 0 -255 255 85 170 85 0 170 85 0 170 85 0 - 85 85 85 85 85 85 85 85 85 85 85 85 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 85 85 85 - 85 85 85 85 85 85 85 85 85 170 85 0 -170 85 0 170 85 0 170 85 0 255 255 85 -170 85 0 255 255 85 170 85 0 170 85 0 -170 85 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 170 85 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 -170 85 0 170 85 0 170 85 0 170 85 0 -170 85 0 170 85 0 170 85 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 - 0 0 0 0 0 0 0 0 0 0 0 0 +# CREATOR: GIMP PNM Filter Version 1.1 +60 80 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +170 +170 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +255 +255 +255 +85 +85 +85 +85 +85 +85 +85 +85 +85 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +170 +170 +0 +170 +170 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +170 +170 +170 +85 +85 +85 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +85 +85 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +170 +170 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +170 +170 +170 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +255 +255 +255 +255 +170 +170 +170 +85 +85 +85 +170 +170 +170 +170 +170 +170 +0 +0 +0 +255 +85 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +170 +170 +170 +85 +85 +85 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +170 +170 +170 +255 +255 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +255 +255 +255 +170 +170 +170 +170 +170 +170 +255 +255 +255 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +255 +255 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +255 +255 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +170 +170 +170 +0 +170 +0 +170 +0 +170 +170 +0 +170 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +170 +170 +0 +170 +170 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +255 +255 +255 +170 +170 +170 +170 +170 +170 +85 +85 +85 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +170 +170 +170 +170 +170 +170 +170 +170 +170 +0 +170 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +85 +0 +255 +255 +85 +255 +255 +85 +170 +170 +170 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +0 +170 +170 +85 +85 +85 +85 +85 +85 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +170 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +170 +170 +170 +85 +85 +85 +85 +85 +85 +170 +170 +170 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +170 +170 +170 +85 +85 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +170 +0 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +170 +85 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +85 +85 +85 +85 +85 +85 +170 +0 +170 +170 +0 +170 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +255 +255 +85 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +170 +255 +255 +255 +255 +255 +255 +170 +170 +170 +255 +255 +255 +170 +170 +170 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +170 +0 +85 +85 +85 +0 +170 +0 +170 +85 +0 +0 +170 +0 +170 +85 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +255 +85 +85 +85 +85 +85 +85 +170 +170 +170 +85 +85 +85 +255 +255 +255 +0 +170 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +170 +170 +170 +85 +85 +85 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +85 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +170 +0 +85 +85 +85 +170 +85 +0 +0 +170 +0 +170 +85 +0 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +255 +170 +170 +170 +85 +85 +85 +170 +170 +170 +170 +170 +170 +0 +0 +0 +0 +0 +0 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +85 +255 +255 +0 +0 +0 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +170 +85 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +170 +170 +170 +85 +85 +85 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +85 +85 +85 +0 +170 +170 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +170 +170 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +170 +170 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +170 +170 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +170 +170 +170 +170 +170 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +0 +0 +170 +0 +170 +170 +0 +170 +85 +85 +85 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +170 +170 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +0 +170 +170 +0 +170 +170 +85 +255 +255 +85 +85 +85 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +255 +255 +0 +170 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +170 +170 +170 +255 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +0 +170 +0 +170 +255 +85 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +0 +170 +170 +0 +170 +170 +0 +170 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +170 +170 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +255 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +85 +85 +85 +170 +170 +170 +255 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +170 +170 +170 +85 +85 +85 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +0 +0 +0 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +170 +170 +170 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +170 +0 +170 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +170 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +255 +85 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +85 +85 +85 +85 +85 +85 +0 +0 +0 +85 +85 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +85 +255 +255 +0 +170 +170 +85 +85 +85 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +255 +85 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +170 +170 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +170 +170 +0 +170 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +0 +170 +170 +0 +170 +170 +0 +170 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +170 +170 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +0 +170 +170 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +85 +255 +170 +0 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +170 +170 +85 +85 +85 +0 +0 +0 +0 +170 +170 +85 +255 +255 +0 +170 +170 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +85 +85 +85 +85 +85 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +170 +170 +85 +255 +255 +0 +170 +170 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +0 +170 +170 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +255 +255 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +0 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +85 +0 +0 +170 +0 +170 +85 +0 +85 +85 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +0 +0 +0 +0 +0 +0 +170 +170 +170 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +85 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +170 +85 +0 +0 +0 +0 +0 +170 +170 +85 +255 +255 +85 +255 +255 +85 +255 +255 +170 +170 +170 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +255 +85 +85 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +170 +85 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +85 +85 +255 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +170 +170 +170 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +85 +85 +85 +85 +85 +85 +0 +170 +0 +85 +85 +85 +170 +85 +0 +0 +170 +0 +255 +255 +85 +170 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +170 +170 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +170 +85 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +255 +255 +85 +85 +85 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +170 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +255 +255 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +85 +85 +85 +85 +85 +85 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 +0 diff --git a/fs/Kconfig b/fs/Kconfig index 141a856c50e..6741d695b1d 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -292,6 +292,7 @@ source "fs/sysv/Kconfig" source "fs/ufs/Kconfig" source "fs/erofs/Kconfig" source "fs/vboxsf/Kconfig" +source "fs/aufs/Kconfig" endif # MISC_FILESYSTEMS diff --git a/fs/Makefile b/fs/Makefile index 9c708e1fbe8..fde6be84a08 100644 --- a/fs/Makefile +++ b/fs/Makefile @@ -136,3 +136,4 @@ obj-$(CONFIG_EFIVAR_FS) += efivarfs/ obj-$(CONFIG_EROFS_FS) += erofs/ obj-$(CONFIG_VBOXSF_FS) += vboxsf/ obj-$(CONFIG_ZONEFS_FS) += zonefs/ +obj-$(CONFIG_AUFS_FS) += aufs/ diff --git a/fs/aufs/Kconfig b/fs/aufs/Kconfig new file mode 100644 index 00000000000..9f436425716 --- /dev/null +++ b/fs/aufs/Kconfig @@ -0,0 +1,199 @@ +# SPDX-License-Identifier: GPL-2.0 +config AUFS_FS + tristate "Aufs (Advanced multi layered unification filesystem) support" + help + Aufs is a stackable unification filesystem such as Unionfs, + which unifies several directories and provides a merged single + directory. + In the early days, aufs was entirely re-designed and + re-implemented Unionfs Version 1.x series. Introducing many + original ideas, approaches and improvements, it becomes totally + different from Unionfs while keeping the basic features. + +if AUFS_FS +choice + prompt "Maximum number of branches" + default AUFS_BRANCH_MAX_127 + help + Specifies the maximum number of branches (or member directories) + in a single aufs. The larger value consumes more system + resources and has a minor impact to performance. +config AUFS_BRANCH_MAX_127 + bool "127" + help + Specifies the maximum number of branches (or member directories) + in a single aufs. The larger value consumes more system + resources and has a minor impact to performance. +config AUFS_BRANCH_MAX_511 + bool "511" + help + Specifies the maximum number of branches (or member directories) + in a single aufs. The larger value consumes more system + resources and has a minor impact to performance. +config AUFS_BRANCH_MAX_1023 + bool "1023" + help + Specifies the maximum number of branches (or member directories) + in a single aufs. The larger value consumes more system + resources and has a minor impact to performance. +config AUFS_BRANCH_MAX_32767 + bool "32767" + help + Specifies the maximum number of branches (or member directories) + in a single aufs. The larger value consumes more system + resources and has a minor impact to performance. +endchoice + +config AUFS_SBILIST + bool + depends on AUFS_MAGIC_SYSRQ || PROC_FS + default y + help + Automatic configuration for internal use. + When aufs supports Magic SysRq or /proc, enabled automatically. + +config AUFS_HNOTIFY + bool "Detect direct branch access (bypassing aufs)" + help + If you want to modify files on branches directly, eg. bypassing aufs, + and want aufs to detect the changes of them fully, then enable this + option and use 'udba=notify' mount option. + Currently there is only one available configuration, "fsnotify". + It will have a negative impact to the performance. + See detail in aufs.5. + +choice + prompt "method" if AUFS_HNOTIFY + default AUFS_HFSNOTIFY +config AUFS_HFSNOTIFY + bool "fsnotify" + select FSNOTIFY +endchoice + +config AUFS_EXPORT + bool "NFS-exportable aufs" + depends on EXPORTFS + help + If you want to export your mounted aufs via NFS, then enable this + option. There are several requirements for this configuration. + See detail in aufs.5. + +config AUFS_INO_T_64 + bool + depends on AUFS_EXPORT + depends on 64BIT && !(ALPHA || S390) + default y + help + Automatic configuration for internal use. + /* typedef unsigned long/int __kernel_ino_t */ + /* alpha and s390x are int */ + +config AUFS_XATTR + bool "support for XATTR/EA (including Security Labels)" + help + If your branch fs supports XATTR/EA and you want to make them + available in aufs too, then enable this opsion and specify the + branch attributes for EA. + See detail in aufs.5. + +config AUFS_FHSM + bool "File-based Hierarchical Storage Management" + help + Hierarchical Storage Management (or HSM) is a well-known feature + in the storage world. Aufs provides this feature as file-based. + with multiple branches. + These multiple branches are prioritized, ie. the topmost one + should be the fastest drive and be used heavily. + +config AUFS_RDU + bool "Readdir in userspace" + help + Aufs has two methods to provide a merged view for a directory, + by a user-space library and by kernel-space natively. The latter + is always enabled but sometimes large and slow. + If you enable this option, install the library in aufs2-util + package, and set some environment variables for your readdir(3), + then the work will be handled in user-space which generally + shows better performance in most cases. + See detail in aufs.5. + +config AUFS_DIRREN + bool "Workaround for rename(2)-ing a directory" + help + By default, aufs returns EXDEV error in renameing a dir who has + his child on the lower branch, since it is a bad idea to issue + rename(2) internally for every lower branch. But user may not + accept this behaviour. So here is a workaround to allow such + rename(2) and store some extra infromation on the writable + branch. Obviously this costs high (and I don't like it). + To use this feature, you need to enable this configuration AND + to specify the mount option `dirren.' + See details in aufs.5 and the design documents. + +config AUFS_SHWH + bool "Show whiteouts" + help + If you want to make the whiteouts in aufs visible, then enable + this option and specify 'shwh' mount option. Although it may + sounds like philosophy or something, but in technically it + simply shows the name of whiteout with keeping its behaviour. + +config AUFS_BR_RAMFS + bool "Ramfs (initramfs/rootfs) as an aufs branch" + help + If you want to use ramfs as an aufs branch fs, then enable this + option. Generally tmpfs is recommended. + Aufs prohibited them to be a branch fs by default, because + initramfs becomes unusable after switch_root or something + generally. If you sets initramfs as an aufs branch and boot your + system by switch_root, you will meet a problem easily since the + files in initramfs may be inaccessible. + Unless you are going to use ramfs as an aufs branch fs without + switch_root or something, leave it N. + +config AUFS_BR_FUSE + bool "Fuse fs as an aufs branch" + depends on FUSE_FS + select AUFS_POLL + help + If you want to use fuse-based userspace filesystem as an aufs + branch fs, then enable this option. + It implements the internal poll(2) operation which is + implemented by fuse only (curretnly). + +config AUFS_POLL + bool + help + Automatic configuration for internal use. + +config AUFS_BR_HFSPLUS + bool "Hfsplus as an aufs branch" + depends on HFSPLUS_FS + default y + help + If you want to use hfsplus fs as an aufs branch fs, then enable + this option. This option introduces a small overhead at + copying-up a file on hfsplus. + +config AUFS_BDEV_LOOP + bool + depends on BLK_DEV_LOOP + default y + help + Automatic configuration for internal use. + Convert =[ym] into =y. + +config AUFS_DEBUG + bool "Debug aufs" + help + Enable this to compile aufs internal debug code. + It will have a negative impact to the performance. + +config AUFS_MAGIC_SYSRQ + bool + depends on AUFS_DEBUG && MAGIC_SYSRQ + default y + help + Automatic configuration for internal use. + When aufs supports Magic SysRq, enabled automatically. +endif diff --git a/fs/aufs/Makefile b/fs/aufs/Makefile new file mode 100644 index 00000000000..2c819a64935 --- /dev/null +++ b/fs/aufs/Makefile @@ -0,0 +1,46 @@ +# SPDX-License-Identifier: GPL-2.0 + +include ${src}/magic.mk +ifeq (${CONFIG_AUFS_FS},m) +include ${src}/conf.mk +endif +-include ${src}/priv_def.mk + +# cf. include/linux/kernel.h +# enable pr_debug +ccflags-y += -DDEBUG +# sparse requires the full pathname +ifdef M +ccflags-y += -include ${M}/../../include/uapi/linux/aufs_type.h +else +ccflags-y += -include ${srctree}/include/uapi/linux/aufs_type.h +endif + +obj-$(CONFIG_AUFS_FS) += aufs.o +aufs-y := module.o sbinfo.o super.o branch.o xino.o sysaufs.o opts.o \ + wkq.o vfsub.o dcsub.o \ + cpup.o whout.o wbr_policy.o \ + dinfo.o dentry.o \ + dynop.o \ + finfo.o file.o f_op.o \ + dir.o vdir.o \ + iinfo.o inode.o i_op.o i_op_add.o i_op_del.o i_op_ren.o \ + mvdown.o ioctl.o + +# all are boolean +aufs-$(CONFIG_PROC_FS) += procfs.o plink.o +aufs-$(CONFIG_SYSFS) += sysfs.o +aufs-$(CONFIG_DEBUG_FS) += dbgaufs.o +aufs-$(CONFIG_AUFS_BDEV_LOOP) += loop.o +aufs-$(CONFIG_AUFS_HNOTIFY) += hnotify.o +aufs-$(CONFIG_AUFS_HFSNOTIFY) += hfsnotify.o +aufs-$(CONFIG_AUFS_EXPORT) += export.o +aufs-$(CONFIG_AUFS_XATTR) += xattr.o +aufs-$(CONFIG_FS_POSIX_ACL) += posix_acl.o +aufs-$(CONFIG_AUFS_DIRREN) += dirren.o +aufs-$(CONFIG_AUFS_FHSM) += fhsm.o +aufs-$(CONFIG_AUFS_POLL) += poll.o +aufs-$(CONFIG_AUFS_RDU) += rdu.o +aufs-$(CONFIG_AUFS_BR_HFSPLUS) += hfsplus.o +aufs-$(CONFIG_AUFS_DEBUG) += debug.o +aufs-$(CONFIG_AUFS_MAGIC_SYSRQ) += sysrq.o diff --git a/fs/aufs/aufs.h b/fs/aufs/aufs.h new file mode 100644 index 00000000000..9a573445876 --- /dev/null +++ b/fs/aufs/aufs.h @@ -0,0 +1,62 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * all header files + */ + +#ifndef __AUFS_H__ +#define __AUFS_H__ + +#ifdef __KERNEL__ + +#define AuStub(type, name, body, ...) \ + static inline type name(__VA_ARGS__) { body; } + +#define AuStubVoid(name, ...) \ + AuStub(void, name, , __VA_ARGS__) +#define AuStubInt0(name, ...) \ + AuStub(int, name, return 0, __VA_ARGS__) + +#include "debug.h" + +#include "branch.h" +#include "cpup.h" +#include "dcsub.h" +#include "dbgaufs.h" +#include "dentry.h" +#include "dir.h" +#include "dirren.h" +#include "dynop.h" +#include "file.h" +#include "fstype.h" +#include "hbl.h" +#include "inode.h" +#include "lcnt.h" +#include "loop.h" +#include "module.h" +#include "opts.h" +#include "rwsem.h" +#include "super.h" +#include "sysaufs.h" +#include "vfsub.h" +#include "whout.h" +#include "wkq.h" + +#endif /* __KERNEL__ */ +#endif /* __AUFS_H__ */ diff --git a/fs/aufs/branch.c b/fs/aufs/branch.c new file mode 100644 index 00000000000..26d41e1bc0c --- /dev/null +++ b/fs/aufs/branch.c @@ -0,0 +1,1427 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * branch management + */ + +#include +#include +#include "aufs.h" + +/* + * free a single branch + */ +static void au_br_do_free(struct au_branch *br) +{ + int i; + struct au_wbr *wbr; + struct au_dykey **key; + + au_hnotify_fin_br(br); + /* always, regardless the mount option */ + au_dr_hino_free(&br->br_dirren); + au_xino_put(br); + + AuLCntZero(au_lcnt_read(&br->br_nfiles, /*do_rev*/0)); + au_lcnt_fin(&br->br_nfiles, /*do_sync*/0); + AuLCntZero(au_lcnt_read(&br->br_count, /*do_rev*/0)); + au_lcnt_fin(&br->br_count, /*do_sync*/0); + + wbr = br->br_wbr; + if (wbr) { + for (i = 0; i < AuBrWh_Last; i++) + dput(wbr->wbr_wh[i]); + AuDebugOn(atomic_read(&wbr->wbr_wh_running)); + AuRwDestroy(&wbr->wbr_wh_rwsem); + } + + if (br->br_fhsm) { + au_br_fhsm_fin(br->br_fhsm); + au_kfree_try_rcu(br->br_fhsm); + } + + key = br->br_dykey; + for (i = 0; i < AuBrDynOp; i++, key++) + if (*key) + au_dy_put(*key); + else + break; + + /* recursive lock, s_umount of branch's */ + /* synchronize_rcu(); */ /* why? */ + lockdep_off(); + path_put(&br->br_path); + lockdep_on(); + au_kfree_rcu(wbr); + au_lcnt_wait_for_fin(&br->br_nfiles); + au_lcnt_wait_for_fin(&br->br_count); + /* I don't know why, but percpu_refcount requires this */ + /* synchronize_rcu(); */ + au_kfree_rcu(br); +} + +/* + * frees all branches + */ +void au_br_free(struct au_sbinfo *sbinfo) +{ + aufs_bindex_t bmax; + struct au_branch **br; + + AuRwMustWriteLock(&sbinfo->si_rwsem); + + bmax = sbinfo->si_bbot + 1; + br = sbinfo->si_branch; + while (bmax--) + au_br_do_free(*br++); +} + +/* + * find the index of a branch which is specified by @br_id. + */ +int au_br_index(struct super_block *sb, aufs_bindex_t br_id) +{ + aufs_bindex_t bindex, bbot; + + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) + if (au_sbr_id(sb, bindex) == br_id) + return bindex; + return -1; +} + +/* ---------------------------------------------------------------------- */ + +/* + * add a branch + */ + +static int test_overlap(struct super_block *sb, struct dentry *h_adding, + struct dentry *h_root) +{ + if (unlikely(h_adding == h_root + || au_test_loopback_overlap(sb, h_adding))) + return 1; + if (h_adding->d_sb != h_root->d_sb) + return 0; + return au_test_subdir(h_adding, h_root) + || au_test_subdir(h_root, h_adding); +} + +/* + * returns a newly allocated branch. @new_nbranch is a number of branches + * after adding a branch. + */ +static struct au_branch *au_br_alloc(struct super_block *sb, int new_nbranch, + int perm) +{ + struct au_branch *add_branch; + struct dentry *root; + struct inode *inode; + int err; + + err = -ENOMEM; + add_branch = kzalloc(sizeof(*add_branch), GFP_NOFS); + if (unlikely(!add_branch)) + goto out; + add_branch->br_xino = au_xino_alloc(/*nfile*/1); + if (unlikely(!add_branch->br_xino)) + goto out_br; + err = au_hnotify_init_br(add_branch, perm); + if (unlikely(err)) + goto out_xino; + + if (au_br_writable(perm)) { + /* may be freed separately at changing the branch permission */ + add_branch->br_wbr = kzalloc(sizeof(*add_branch->br_wbr), + GFP_NOFS); + if (unlikely(!add_branch->br_wbr)) + goto out_hnotify; + } + + if (au_br_fhsm(perm)) { + err = au_fhsm_br_alloc(add_branch); + if (unlikely(err)) + goto out_wbr; + } + + root = sb->s_root; + err = au_sbr_realloc(au_sbi(sb), new_nbranch, /*may_shrink*/0); + if (!err) + err = au_di_realloc(au_di(root), new_nbranch, /*may_shrink*/0); + if (!err) { + inode = d_inode(root); + err = au_hinode_realloc(au_ii(inode), new_nbranch, + /*may_shrink*/0); + } + if (!err) + return add_branch; /* success */ + +out_wbr: + au_kfree_rcu(add_branch->br_wbr); +out_hnotify: + au_hnotify_fin_br(add_branch); +out_xino: + au_xino_put(add_branch); +out_br: + au_kfree_rcu(add_branch); +out: + return ERR_PTR(err); +} + +/* + * test if the branch permission is legal or not. + */ +static int test_br(struct inode *inode, int brperm, char *path) +{ + int err; + + err = (au_br_writable(brperm) && IS_RDONLY(inode)); + if (!err) + goto out; + + err = -EINVAL; + pr_err("write permission for readonly mount or inode, %s\n", path); + +out: + return err; +} + +/* + * returns: + * 0: success, the caller will add it + * plus: success, it is already unified, the caller should ignore it + * minus: error + */ +static int test_add(struct super_block *sb, struct au_opt_add *add, int remount) +{ + int err; + aufs_bindex_t bbot, bindex; + struct dentry *root, *h_dentry; + struct inode *inode, *h_inode; + + root = sb->s_root; + bbot = au_sbbot(sb); + if (unlikely(bbot >= 0 + && au_find_dbindex(root, add->path.dentry) >= 0)) { + err = 1; + if (!remount) { + err = -EINVAL; + pr_err("%s duplicated\n", add->pathname); + } + goto out; + } + + err = -ENOSPC; /* -E2BIG; */ + if (unlikely(AUFS_BRANCH_MAX <= add->bindex + || AUFS_BRANCH_MAX - 1 <= bbot)) { + pr_err("number of branches exceeded %s\n", add->pathname); + goto out; + } + + err = -EDOM; + if (unlikely(add->bindex < 0 || bbot + 1 < add->bindex)) { + pr_err("bad index %d\n", add->bindex); + goto out; + } + + inode = d_inode(add->path.dentry); + err = -ENOENT; + if (unlikely(!inode->i_nlink)) { + pr_err("no existence %s\n", add->pathname); + goto out; + } + + err = -EINVAL; + if (unlikely(inode->i_sb == sb)) { + pr_err("%s must be outside\n", add->pathname); + goto out; + } + + if (unlikely(au_test_fs_unsuppoted(inode->i_sb))) { + pr_err("unsupported filesystem, %s (%s)\n", + add->pathname, au_sbtype(inode->i_sb)); + goto out; + } + + if (unlikely(inode->i_sb->s_stack_depth)) { + pr_err("already stacked, %s (%s)\n", + add->pathname, au_sbtype(inode->i_sb)); + goto out; + } + + err = test_br(d_inode(add->path.dentry), add->perm, add->pathname); + if (unlikely(err)) + goto out; + + if (bbot < 0) + return 0; /* success */ + + err = -EINVAL; + for (bindex = 0; bindex <= bbot; bindex++) + if (unlikely(test_overlap(sb, add->path.dentry, + au_h_dptr(root, bindex)))) { + pr_err("%s is overlapped\n", add->pathname); + goto out; + } + + err = 0; + if (au_opt_test(au_mntflags(sb), WARN_PERM)) { + h_dentry = au_h_dptr(root, 0); + h_inode = d_inode(h_dentry); + if ((h_inode->i_mode & S_IALLUGO) != (inode->i_mode & S_IALLUGO) + || !uid_eq(h_inode->i_uid, inode->i_uid) + || !gid_eq(h_inode->i_gid, inode->i_gid)) + pr_warn("uid/gid/perm %s %u/%u/0%o, %u/%u/0%o\n", + add->pathname, + i_uid_read(inode), i_gid_read(inode), + (inode->i_mode & S_IALLUGO), + i_uid_read(h_inode), i_gid_read(h_inode), + (h_inode->i_mode & S_IALLUGO)); + } + +out: + return err; +} + +/* + * initialize or clean the whiteouts for an adding branch + */ +static int au_br_init_wh(struct super_block *sb, struct au_branch *br, + int new_perm) +{ + int err, old_perm; + aufs_bindex_t bindex; + struct inode *h_inode; + struct au_wbr *wbr; + struct au_hinode *hdir; + struct dentry *h_dentry; + + err = vfsub_mnt_want_write(au_br_mnt(br)); + if (unlikely(err)) + goto out; + + wbr = br->br_wbr; + old_perm = br->br_perm; + br->br_perm = new_perm; + hdir = NULL; + h_inode = NULL; + bindex = au_br_index(sb, br->br_id); + if (0 <= bindex) { + hdir = au_hi(d_inode(sb->s_root), bindex); + au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT); + } else { + h_dentry = au_br_dentry(br); + h_inode = d_inode(h_dentry); + inode_lock_nested(h_inode, AuLsc_I_PARENT); + } + if (!wbr) + err = au_wh_init(br, sb); + else { + wbr_wh_write_lock(wbr); + err = au_wh_init(br, sb); + wbr_wh_write_unlock(wbr); + } + if (hdir) + au_hn_inode_unlock(hdir); + else + inode_unlock(h_inode); + vfsub_mnt_drop_write(au_br_mnt(br)); + br->br_perm = old_perm; + + if (!err && wbr && !au_br_writable(new_perm)) { + au_kfree_rcu(wbr); + br->br_wbr = NULL; + } + +out: + return err; +} + +static int au_wbr_init(struct au_branch *br, struct super_block *sb, + int perm) +{ + int err; + struct kstatfs kst; + struct au_wbr *wbr; + + wbr = br->br_wbr; + au_rw_init(&wbr->wbr_wh_rwsem); + atomic_set(&wbr->wbr_wh_running, 0); + + /* + * a limit for rmdir/rename a dir + * cf. AUFS_MAX_NAMELEN in include/uapi/linux/aufs_type.h + */ + err = vfs_statfs(&br->br_path, &kst); + if (unlikely(err)) + goto out; + err = -EINVAL; + if (kst.f_namelen >= NAME_MAX) + err = au_br_init_wh(sb, br, perm); + else + pr_err("%pd(%s), unsupported namelen %ld\n", + au_br_dentry(br), + au_sbtype(au_br_dentry(br)->d_sb), kst.f_namelen); + +out: + return err; +} + +/* initialize a new branch */ +static int au_br_init(struct au_branch *br, struct super_block *sb, + struct au_opt_add *add) +{ + int err; + struct au_branch *brbase; + struct file *xf; + struct inode *h_inode; + + err = 0; + br->br_perm = add->perm; + br->br_path = add->path; /* set first, path_get() later */ + spin_lock_init(&br->br_dykey_lock); + au_lcnt_init(&br->br_nfiles, /*release*/NULL); + au_lcnt_init(&br->br_count, /*release*/NULL); + br->br_id = au_new_br_id(sb); + AuDebugOn(br->br_id < 0); + + /* always, regardless the given option */ + err = au_dr_br_init(sb, br, &add->path); + if (unlikely(err)) + goto out_err; + + if (au_br_writable(add->perm)) { + err = au_wbr_init(br, sb, add->perm); + if (unlikely(err)) + goto out_err; + } + + if (au_opt_test(au_mntflags(sb), XINO)) { + brbase = au_sbr(sb, 0); + xf = au_xino_file(brbase->br_xino, /*idx*/-1); + AuDebugOn(!xf); + h_inode = d_inode(add->path.dentry); + err = au_xino_init_br(sb, br, h_inode->i_ino, &xf->f_path); + if (unlikely(err)) { + AuDebugOn(au_xino_file(br->br_xino, /*idx*/-1)); + goto out_err; + } + } + + sysaufs_br_init(br); + path_get(&br->br_path); + goto out; /* success */ + +out_err: + memset(&br->br_path, 0, sizeof(br->br_path)); +out: + return err; +} + +static void au_br_do_add_brp(struct au_sbinfo *sbinfo, aufs_bindex_t bindex, + struct au_branch *br, aufs_bindex_t bbot, + aufs_bindex_t amount) +{ + struct au_branch **brp; + + AuRwMustWriteLock(&sbinfo->si_rwsem); + + brp = sbinfo->si_branch + bindex; + memmove(brp + 1, brp, sizeof(*brp) * amount); + *brp = br; + sbinfo->si_bbot++; + if (unlikely(bbot < 0)) + sbinfo->si_bbot = 0; +} + +static void au_br_do_add_hdp(struct au_dinfo *dinfo, aufs_bindex_t bindex, + aufs_bindex_t bbot, aufs_bindex_t amount) +{ + struct au_hdentry *hdp; + + AuRwMustWriteLock(&dinfo->di_rwsem); + + hdp = au_hdentry(dinfo, bindex); + memmove(hdp + 1, hdp, sizeof(*hdp) * amount); + au_h_dentry_init(hdp); + dinfo->di_bbot++; + if (unlikely(bbot < 0)) + dinfo->di_btop = 0; +} + +static void au_br_do_add_hip(struct au_iinfo *iinfo, aufs_bindex_t bindex, + aufs_bindex_t bbot, aufs_bindex_t amount) +{ + struct au_hinode *hip; + + AuRwMustWriteLock(&iinfo->ii_rwsem); + + hip = au_hinode(iinfo, bindex); + memmove(hip + 1, hip, sizeof(*hip) * amount); + au_hinode_init(hip); + iinfo->ii_bbot++; + if (unlikely(bbot < 0)) + iinfo->ii_btop = 0; +} + +static void au_br_do_add(struct super_block *sb, struct au_branch *br, + aufs_bindex_t bindex) +{ + struct dentry *root, *h_dentry; + struct inode *root_inode, *h_inode; + aufs_bindex_t bbot, amount; + + root = sb->s_root; + root_inode = d_inode(root); + bbot = au_sbbot(sb); + amount = bbot + 1 - bindex; + h_dentry = au_br_dentry(br); + au_sbilist_lock(); + au_br_do_add_brp(au_sbi(sb), bindex, br, bbot, amount); + au_br_do_add_hdp(au_di(root), bindex, bbot, amount); + au_br_do_add_hip(au_ii(root_inode), bindex, bbot, amount); + au_set_h_dptr(root, bindex, dget(h_dentry)); + h_inode = d_inode(h_dentry); + au_set_h_iptr(root_inode, bindex, au_igrab(h_inode), /*flags*/0); + au_sbilist_unlock(); +} + +int au_br_add(struct super_block *sb, struct au_opt_add *add, int remount) +{ + int err; + aufs_bindex_t bbot, add_bindex; + struct dentry *root, *h_dentry; + struct inode *root_inode; + struct au_branch *add_branch; + + root = sb->s_root; + root_inode = d_inode(root); + IMustLock(root_inode); + IiMustWriteLock(root_inode); + err = test_add(sb, add, remount); + if (unlikely(err < 0)) + goto out; + if (err) { + err = 0; + goto out; /* success */ + } + + bbot = au_sbbot(sb); + add_branch = au_br_alloc(sb, bbot + 2, add->perm); + err = PTR_ERR(add_branch); + if (IS_ERR(add_branch)) + goto out; + + err = au_br_init(add_branch, sb, add); + if (unlikely(err)) { + au_br_do_free(add_branch); + goto out; + } + + add_bindex = add->bindex; + sysaufs_brs_del(sb, add_bindex); /* remove successors */ + au_br_do_add(sb, add_branch, add_bindex); + sysaufs_brs_add(sb, add_bindex); /* append successors */ + dbgaufs_brs_add(sb, add_bindex, /*topdown*/0); /* rename successors */ + + h_dentry = add->path.dentry; + if (!add_bindex) { + au_cpup_attr_all(root_inode, /*force*/1); + sb->s_maxbytes = h_dentry->d_sb->s_maxbytes; + } else + au_add_nlink(root_inode, d_inode(h_dentry)); + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +static unsigned long long au_farray_cb(struct super_block *sb, void *a, + unsigned long long max __maybe_unused, + void *arg) +{ + unsigned long long n; + struct file **p, *f; + struct hlist_bl_head *files; + struct hlist_bl_node *pos; + struct au_finfo *finfo; + + n = 0; + p = a; + files = &au_sbi(sb)->si_files; + hlist_bl_lock(files); + hlist_bl_for_each_entry(finfo, pos, files, fi_hlist) { + f = finfo->fi_file; + if (file_count(f) + && !special_file(file_inode(f)->i_mode)) { + get_file(f); + *p++ = f; + n++; + AuDebugOn(n > max); + } + } + hlist_bl_unlock(files); + + return n; +} + +static struct file **au_farray_alloc(struct super_block *sb, + unsigned long long *max) +{ + struct au_sbinfo *sbi; + + sbi = au_sbi(sb); + *max = au_lcnt_read(&sbi->si_nfiles, /*do_rev*/1); + return au_array_alloc(max, au_farray_cb, sb, /*arg*/NULL); +} + +static void au_farray_free(struct file **a, unsigned long long max) +{ + unsigned long long ull; + + for (ull = 0; ull < max; ull++) + if (a[ull]) + fput(a[ull]); + kvfree(a); +} + +/* ---------------------------------------------------------------------- */ + +/* + * delete a branch + */ + +/* to show the line number, do not make it inlined function */ +#define AuVerbose(do_info, fmt, ...) do { \ + if (do_info) \ + pr_info(fmt, ##__VA_ARGS__); \ +} while (0) + +static int au_test_ibusy(struct inode *inode, aufs_bindex_t btop, + aufs_bindex_t bbot) +{ + return (inode && !S_ISDIR(inode->i_mode)) || btop == bbot; +} + +static int au_test_dbusy(struct dentry *dentry, aufs_bindex_t btop, + aufs_bindex_t bbot) +{ + return au_test_ibusy(d_inode(dentry), btop, bbot); +} + +/* + * test if the branch is deletable or not. + */ +static int test_dentry_busy(struct dentry *root, aufs_bindex_t bindex, + unsigned int sigen, const unsigned int verbose) +{ + int err, i, j, ndentry; + aufs_bindex_t btop, bbot; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry *d; + + err = au_dpages_init(&dpages, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_dcsub_pages(&dpages, root, NULL, NULL); + if (unlikely(err)) + goto out_dpages; + + for (i = 0; !err && i < dpages.ndpage; i++) { + dpage = dpages.dpages + i; + ndentry = dpage->ndentry; + for (j = 0; !err && j < ndentry; j++) { + d = dpage->dentries[j]; + AuDebugOn(au_dcount(d) <= 0); + if (!au_digen_test(d, sigen)) { + di_read_lock_child(d, AuLock_IR); + if (unlikely(au_dbrange_test(d))) { + di_read_unlock(d, AuLock_IR); + continue; + } + } else { + di_write_lock_child(d); + if (unlikely(au_dbrange_test(d))) { + di_write_unlock(d); + continue; + } + err = au_reval_dpath(d, sigen); + if (!err) + di_downgrade_lock(d, AuLock_IR); + else { + di_write_unlock(d); + break; + } + } + + /* AuDbgDentry(d); */ + btop = au_dbtop(d); + bbot = au_dbbot(d); + if (btop <= bindex + && bindex <= bbot + && au_h_dptr(d, bindex) + && au_test_dbusy(d, btop, bbot)) { + err = -EBUSY; + AuVerbose(verbose, "busy %pd\n", d); + AuDbgDentry(d); + } + di_read_unlock(d, AuLock_IR); + } + } + +out_dpages: + au_dpages_free(&dpages); +out: + return err; +} + +static int test_inode_busy(struct super_block *sb, aufs_bindex_t bindex, + unsigned int sigen, const unsigned int verbose) +{ + int err; + unsigned long long max, ull; + struct inode *i, **array; + aufs_bindex_t btop, bbot; + + array = au_iarray_alloc(sb, &max); + err = PTR_ERR(array); + if (IS_ERR(array)) + goto out; + + err = 0; + AuDbg("b%d\n", bindex); + for (ull = 0; !err && ull < max; ull++) { + i = array[ull]; + if (unlikely(!i)) + break; + if (i->i_ino == AUFS_ROOT_INO) + continue; + + /* AuDbgInode(i); */ + if (au_iigen(i, NULL) == sigen) + ii_read_lock_child(i); + else { + ii_write_lock_child(i); + err = au_refresh_hinode_self(i); + au_iigen_dec(i); + if (!err) + ii_downgrade_lock(i); + else { + ii_write_unlock(i); + break; + } + } + + btop = au_ibtop(i); + bbot = au_ibbot(i); + if (btop <= bindex + && bindex <= bbot + && au_h_iptr(i, bindex) + && au_test_ibusy(i, btop, bbot)) { + err = -EBUSY; + AuVerbose(verbose, "busy i%lu\n", i->i_ino); + AuDbgInode(i); + } + ii_read_unlock(i); + } + au_iarray_free(array, max); + +out: + return err; +} + +static int test_children_busy(struct dentry *root, aufs_bindex_t bindex, + const unsigned int verbose) +{ + int err; + unsigned int sigen; + + sigen = au_sigen(root->d_sb); + DiMustNoWaiters(root); + IiMustNoWaiters(d_inode(root)); + di_write_unlock(root); + err = test_dentry_busy(root, bindex, sigen, verbose); + if (!err) + err = test_inode_busy(root->d_sb, bindex, sigen, verbose); + di_write_lock_child(root); /* aufs_write_lock() calls ..._child() */ + + return err; +} + +static int test_dir_busy(struct file *file, aufs_bindex_t br_id, + struct file **to_free, int *idx) +{ + int err; + unsigned char matched, root; + aufs_bindex_t bindex, bbot; + struct au_fidir *fidir; + struct au_hfile *hfile; + + err = 0; + root = IS_ROOT(file->f_path.dentry); + if (root) { + get_file(file); + to_free[*idx] = file; + (*idx)++; + goto out; + } + + matched = 0; + fidir = au_fi(file)->fi_hdir; + AuDebugOn(!fidir); + bbot = au_fbbot_dir(file); + for (bindex = au_fbtop(file); bindex <= bbot; bindex++) { + hfile = fidir->fd_hfile + bindex; + if (!hfile->hf_file) + continue; + + if (hfile->hf_br->br_id == br_id) { + matched = 1; + break; + } + } + if (matched) + err = -EBUSY; + +out: + return err; +} + +static int test_file_busy(struct super_block *sb, aufs_bindex_t br_id, + struct file **to_free, int opened) +{ + int err, idx; + unsigned long long ull, max; + aufs_bindex_t btop; + struct file *file, **array; + struct dentry *root; + struct au_hfile *hfile; + + array = au_farray_alloc(sb, &max); + err = PTR_ERR(array); + if (IS_ERR(array)) + goto out; + + err = 0; + idx = 0; + root = sb->s_root; + di_write_unlock(root); + for (ull = 0; ull < max; ull++) { + file = array[ull]; + if (unlikely(!file)) + break; + + /* AuDbg("%pD\n", file); */ + fi_read_lock(file); + btop = au_fbtop(file); + if (!d_is_dir(file->f_path.dentry)) { + hfile = &au_fi(file)->fi_htop; + if (hfile->hf_br->br_id == br_id) + err = -EBUSY; + } else + err = test_dir_busy(file, br_id, to_free, &idx); + fi_read_unlock(file); + if (unlikely(err)) + break; + } + di_write_lock_child(root); + au_farray_free(array, max); + AuDebugOn(idx > opened); + +out: + return err; +} + +static void br_del_file(struct file **to_free, unsigned long long opened, + aufs_bindex_t br_id) +{ + unsigned long long ull; + aufs_bindex_t bindex, btop, bbot, bfound; + struct file *file; + struct au_fidir *fidir; + struct au_hfile *hfile; + + for (ull = 0; ull < opened; ull++) { + file = to_free[ull]; + if (unlikely(!file)) + break; + + /* AuDbg("%pD\n", file); */ + AuDebugOn(!d_is_dir(file->f_path.dentry)); + bfound = -1; + fidir = au_fi(file)->fi_hdir; + AuDebugOn(!fidir); + fi_write_lock(file); + btop = au_fbtop(file); + bbot = au_fbbot_dir(file); + for (bindex = btop; bindex <= bbot; bindex++) { + hfile = fidir->fd_hfile + bindex; + if (!hfile->hf_file) + continue; + + if (hfile->hf_br->br_id == br_id) { + bfound = bindex; + break; + } + } + AuDebugOn(bfound < 0); + au_set_h_fptr(file, bfound, NULL); + if (bfound == btop) { + for (btop++; btop <= bbot; btop++) + if (au_hf_dir(file, btop)) { + au_set_fbtop(file, btop); + break; + } + } + fi_write_unlock(file); + } +} + +static void au_br_do_del_brp(struct au_sbinfo *sbinfo, + const aufs_bindex_t bindex, + const aufs_bindex_t bbot) +{ + struct au_branch **brp, **p; + + AuRwMustWriteLock(&sbinfo->si_rwsem); + + brp = sbinfo->si_branch + bindex; + if (bindex < bbot) + memmove(brp, brp + 1, sizeof(*brp) * (bbot - bindex)); + sbinfo->si_branch[0 + bbot] = NULL; + sbinfo->si_bbot--; + + p = au_krealloc(sbinfo->si_branch, sizeof(*p) * bbot, AuGFP_SBILIST, + /*may_shrink*/1); + if (p) + sbinfo->si_branch = p; + /* harmless error */ +} + +static void au_br_do_del_hdp(struct au_dinfo *dinfo, const aufs_bindex_t bindex, + const aufs_bindex_t bbot) +{ + struct au_hdentry *hdp, *p; + + AuRwMustWriteLock(&dinfo->di_rwsem); + + hdp = au_hdentry(dinfo, bindex); + if (bindex < bbot) + memmove(hdp, hdp + 1, sizeof(*hdp) * (bbot - bindex)); + /* au_h_dentry_init(au_hdentry(dinfo, bbot); */ + dinfo->di_bbot--; + + p = au_krealloc(dinfo->di_hdentry, sizeof(*p) * bbot, AuGFP_SBILIST, + /*may_shrink*/1); + if (p) + dinfo->di_hdentry = p; + /* harmless error */ +} + +static void au_br_do_del_hip(struct au_iinfo *iinfo, const aufs_bindex_t bindex, + const aufs_bindex_t bbot) +{ + struct au_hinode *hip, *p; + + AuRwMustWriteLock(&iinfo->ii_rwsem); + + hip = au_hinode(iinfo, bindex); + if (bindex < bbot) + memmove(hip, hip + 1, sizeof(*hip) * (bbot - bindex)); + /* au_hinode_init(au_hinode(iinfo, bbot)); */ + iinfo->ii_bbot--; + + p = au_krealloc(iinfo->ii_hinode, sizeof(*p) * bbot, AuGFP_SBILIST, + /*may_shrink*/1); + if (p) + iinfo->ii_hinode = p; + /* harmless error */ +} + +static void au_br_do_del(struct super_block *sb, aufs_bindex_t bindex, + struct au_branch *br) +{ + aufs_bindex_t bbot; + struct au_sbinfo *sbinfo; + struct dentry *root, *h_root; + struct inode *inode, *h_inode; + struct au_hinode *hinode; + + SiMustWriteLock(sb); + + root = sb->s_root; + inode = d_inode(root); + sbinfo = au_sbi(sb); + bbot = sbinfo->si_bbot; + + h_root = au_h_dptr(root, bindex); + hinode = au_hi(inode, bindex); + h_inode = au_igrab(hinode->hi_inode); + au_hiput(hinode); + + au_sbilist_lock(); + au_br_do_del_brp(sbinfo, bindex, bbot); + au_br_do_del_hdp(au_di(root), bindex, bbot); + au_br_do_del_hip(au_ii(inode), bindex, bbot); + au_sbilist_unlock(); + + /* ignore an error */ + au_dr_br_fin(sb, br); /* always, regardless the mount option */ + + dput(h_root); + iput(h_inode); + au_br_do_free(br); +} + +static unsigned long long empty_cb(struct super_block *sb, void *array, + unsigned long long max, void *arg) +{ + return max; +} + +int au_br_del(struct super_block *sb, struct au_opt_del *del, int remount) +{ + int err, rerr, i; + unsigned long long opened; + unsigned int mnt_flags; + aufs_bindex_t bindex, bbot, br_id; + unsigned char do_wh, verbose; + struct au_branch *br; + struct au_wbr *wbr; + struct dentry *root; + struct file **to_free; + + err = 0; + opened = 0; + to_free = NULL; + root = sb->s_root; + bindex = au_find_dbindex(root, del->h_path.dentry); + if (bindex < 0) { + if (remount) + goto out; /* success */ + err = -ENOENT; + pr_err("%s no such branch\n", del->pathname); + goto out; + } + AuDbg("bindex b%d\n", bindex); + + err = -EBUSY; + mnt_flags = au_mntflags(sb); + verbose = !!au_opt_test(mnt_flags, VERBOSE); + bbot = au_sbbot(sb); + if (unlikely(!bbot)) { + AuVerbose(verbose, "no more branches left\n"); + goto out; + } + + br = au_sbr(sb, bindex); + AuDebugOn(!path_equal(&br->br_path, &del->h_path)); + if (unlikely(au_lcnt_read(&br->br_count, /*do_rev*/1))) { + AuVerbose(verbose, "br %pd2 is busy now\n", del->h_path.dentry); + goto out; + } + + br_id = br->br_id; + opened = au_lcnt_read(&br->br_nfiles, /*do_rev*/1); + if (unlikely(opened)) { + to_free = au_array_alloc(&opened, empty_cb, sb, NULL); + err = PTR_ERR(to_free); + if (IS_ERR(to_free)) + goto out; + + err = test_file_busy(sb, br_id, to_free, opened); + if (unlikely(err)) { + AuVerbose(verbose, "%llu file(s) opened\n", opened); + goto out; + } + } + + wbr = br->br_wbr; + do_wh = wbr && (wbr->wbr_whbase || wbr->wbr_plink || wbr->wbr_orph); + if (do_wh) { + /* instead of WbrWhMustWriteLock(wbr) */ + SiMustWriteLock(sb); + for (i = 0; i < AuBrWh_Last; i++) { + dput(wbr->wbr_wh[i]); + wbr->wbr_wh[i] = NULL; + } + } + + err = test_children_busy(root, bindex, verbose); + if (unlikely(err)) { + if (do_wh) + goto out_wh; + goto out; + } + + err = 0; + if (to_free) { + /* + * now we confirmed the branch is deletable. + * let's free the remaining opened dirs on the branch. + */ + di_write_unlock(root); + br_del_file(to_free, opened, br_id); + di_write_lock_child(root); + } + + sysaufs_brs_del(sb, bindex); /* remove successors */ + dbgaufs_xino_del(br); /* remove one */ + au_br_do_del(sb, bindex, br); + sysaufs_brs_add(sb, bindex); /* append successors */ + dbgaufs_brs_add(sb, bindex, /*topdown*/1); /* rename successors */ + + if (!bindex) { + au_cpup_attr_all(d_inode(root), /*force*/1); + sb->s_maxbytes = au_sbr_sb(sb, 0)->s_maxbytes; + } else + au_sub_nlink(d_inode(root), d_inode(del->h_path.dentry)); + if (au_opt_test(mnt_flags, PLINK)) + au_plink_half_refresh(sb, br_id); + + goto out; /* success */ + +out_wh: + /* revert */ + rerr = au_br_init_wh(sb, br, br->br_perm); + if (rerr) + pr_warn("failed re-creating base whiteout, %s. (%d)\n", + del->pathname, rerr); +out: + if (to_free) + au_farray_free(to_free, opened); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_ibusy(struct super_block *sb, struct aufs_ibusy __user *arg) +{ + int err; + aufs_bindex_t btop, bbot; + struct aufs_ibusy ibusy; + struct inode *inode, *h_inode; + + err = -EPERM; + if (unlikely(!capable(CAP_SYS_ADMIN))) + goto out; + + err = copy_from_user(&ibusy, arg, sizeof(ibusy)); + if (!err) + /* VERIFY_WRITE */ + err = !access_ok(&arg->h_ino, sizeof(arg->h_ino)); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + goto out; + } + + err = -EINVAL; + si_read_lock(sb, AuLock_FLUSH); + if (unlikely(ibusy.bindex < 0 || ibusy.bindex > au_sbbot(sb))) + goto out_unlock; + + err = 0; + ibusy.h_ino = 0; /* invalid */ + inode = ilookup(sb, ibusy.ino); + if (!inode + || inode->i_ino == AUFS_ROOT_INO + || au_is_bad_inode(inode)) + goto out_unlock; + + ii_read_lock_child(inode); + btop = au_ibtop(inode); + bbot = au_ibbot(inode); + if (btop <= ibusy.bindex && ibusy.bindex <= bbot) { + h_inode = au_h_iptr(inode, ibusy.bindex); + if (h_inode && au_test_ibusy(inode, btop, bbot)) + ibusy.h_ino = h_inode->i_ino; + } + ii_read_unlock(inode); + iput(inode); + +out_unlock: + si_read_unlock(sb); + if (!err) { + err = __put_user(ibusy.h_ino, &arg->h_ino); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + } + } +out: + return err; +} + +long au_ibusy_ioctl(struct file *file, unsigned long arg) +{ + return au_ibusy(file->f_path.dentry->d_sb, (void __user *)arg); +} + +#ifdef CONFIG_COMPAT +long au_ibusy_compat_ioctl(struct file *file, unsigned long arg) +{ + return au_ibusy(file->f_path.dentry->d_sb, compat_ptr(arg)); +} +#endif + +/* ---------------------------------------------------------------------- */ + +/* + * change a branch permission + */ + +static void au_warn_ima(void) +{ +#ifdef CONFIG_IMA + /* since it doesn't support mark_files_ro() */ + AuWarn1("RW -> RO makes IMA to produce wrong message\n"); +#endif +} + +static int do_need_sigen_inc(int a, int b) +{ + return au_br_whable(a) && !au_br_whable(b); +} + +static int need_sigen_inc(int old, int new) +{ + return do_need_sigen_inc(old, new) + || do_need_sigen_inc(new, old); +} + +static int au_br_mod_files_ro(struct super_block *sb, aufs_bindex_t bindex) +{ + int err, do_warn; + unsigned int mnt_flags; + unsigned long long ull, max; + aufs_bindex_t br_id; + unsigned char verbose, writer; + struct file *file, *hf, **array; + struct au_hfile *hfile; + struct inode *h_inode; + + mnt_flags = au_mntflags(sb); + verbose = !!au_opt_test(mnt_flags, VERBOSE); + + array = au_farray_alloc(sb, &max); + err = PTR_ERR(array); + if (IS_ERR(array)) + goto out; + + do_warn = 0; + br_id = au_sbr_id(sb, bindex); + for (ull = 0; ull < max; ull++) { + file = array[ull]; + if (unlikely(!file)) + break; + + /* AuDbg("%pD\n", file); */ + fi_read_lock(file); + if (unlikely(au_test_mmapped(file))) { + err = -EBUSY; + AuVerbose(verbose, "mmapped %pD\n", file); + AuDbgFile(file); + FiMustNoWaiters(file); + fi_read_unlock(file); + goto out_array; + } + + hfile = &au_fi(file)->fi_htop; + hf = hfile->hf_file; + if (!d_is_reg(file->f_path.dentry) + || !(file->f_mode & FMODE_WRITE) + || hfile->hf_br->br_id != br_id + || !(hf->f_mode & FMODE_WRITE)) + array[ull] = NULL; + else { + do_warn = 1; + get_file(file); + } + + FiMustNoWaiters(file); + fi_read_unlock(file); + fput(file); + } + + err = 0; + if (do_warn) + au_warn_ima(); + + for (ull = 0; ull < max; ull++) { + file = array[ull]; + if (!file) + continue; + + /* todo: already flushed? */ + /* + * fs/super.c:mark_files_ro() is gone, but aufs keeps its + * approach which resets f_mode and calls mnt_drop_write() and + * file_release_write() for each file, because the branch + * attribute in aufs world is totally different from the native + * fs rw/ro mode. + */ + /* fi_read_lock(file); */ + hfile = &au_fi(file)->fi_htop; + hf = hfile->hf_file; + /* fi_read_unlock(file); */ + spin_lock(&hf->f_lock); + writer = !!(hf->f_mode & FMODE_WRITER); + hf->f_mode &= ~(FMODE_WRITE | FMODE_WRITER); + spin_unlock(&hf->f_lock); + if (writer) { + h_inode = file_inode(hf); + if (hf->f_mode & FMODE_READ) + i_readcount_inc(h_inode); + put_write_access(h_inode); + __mnt_drop_write(hf->f_path.mnt); + } + } + +out_array: + au_farray_free(array, max); +out: + AuTraceErr(err); + return err; +} + +int au_br_mod(struct super_block *sb, struct au_opt_mod *mod, int remount, + int *do_refresh) +{ + int err, rerr; + aufs_bindex_t bindex; + struct dentry *root; + struct au_branch *br; + struct au_br_fhsm *bf; + + root = sb->s_root; + bindex = au_find_dbindex(root, mod->h_root); + if (bindex < 0) { + if (remount) + return 0; /* success */ + err = -ENOENT; + pr_err("%s no such branch\n", mod->path); + goto out; + } + AuDbg("bindex b%d\n", bindex); + + err = test_br(d_inode(mod->h_root), mod->perm, mod->path); + if (unlikely(err)) + goto out; + + br = au_sbr(sb, bindex); + AuDebugOn(mod->h_root != au_br_dentry(br)); + if (br->br_perm == mod->perm) + return 0; /* success */ + + /* pre-allocate for non-fhsm --> fhsm */ + bf = NULL; + if (!au_br_fhsm(br->br_perm) && au_br_fhsm(mod->perm)) { + err = au_fhsm_br_alloc(br); + if (unlikely(err)) + goto out; + bf = br->br_fhsm; + br->br_fhsm = NULL; + } + + if (au_br_writable(br->br_perm)) { + /* remove whiteout base */ + err = au_br_init_wh(sb, br, mod->perm); + if (unlikely(err)) + goto out_bf; + + if (!au_br_writable(mod->perm)) { + /* rw --> ro, file might be mmapped */ + DiMustNoWaiters(root); + IiMustNoWaiters(d_inode(root)); + di_write_unlock(root); + err = au_br_mod_files_ro(sb, bindex); + /* aufs_write_lock() calls ..._child() */ + di_write_lock_child(root); + + if (unlikely(err)) { + rerr = -ENOMEM; + br->br_wbr = kzalloc(sizeof(*br->br_wbr), + GFP_NOFS); + if (br->br_wbr) + rerr = au_wbr_init(br, sb, br->br_perm); + if (unlikely(rerr)) { + AuIOErr("nested error %d (%d)\n", + rerr, err); + br->br_perm = mod->perm; + } + } + } + } else if (au_br_writable(mod->perm)) { + /* ro --> rw */ + err = -ENOMEM; + br->br_wbr = kzalloc(sizeof(*br->br_wbr), GFP_NOFS); + if (br->br_wbr) { + err = au_wbr_init(br, sb, mod->perm); + if (unlikely(err)) { + au_kfree_rcu(br->br_wbr); + br->br_wbr = NULL; + } + } + } + if (unlikely(err)) + goto out_bf; + + if (au_br_fhsm(br->br_perm)) { + if (!au_br_fhsm(mod->perm)) { + /* fhsm --> non-fhsm */ + au_br_fhsm_fin(br->br_fhsm); + au_kfree_rcu(br->br_fhsm); + br->br_fhsm = NULL; + } + } else if (au_br_fhsm(mod->perm)) + /* non-fhsm --> fhsm */ + br->br_fhsm = bf; + + *do_refresh |= need_sigen_inc(br->br_perm, mod->perm); + br->br_perm = mod->perm; + goto out; /* success */ + +out_bf: + au_kfree_try_rcu(bf); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_br_stfs(struct au_branch *br, struct aufs_stfs *stfs) +{ + int err; + struct kstatfs kstfs; + + err = vfs_statfs(&br->br_path, &kstfs); + if (!err) { + stfs->f_blocks = kstfs.f_blocks; + stfs->f_bavail = kstfs.f_bavail; + stfs->f_files = kstfs.f_files; + stfs->f_ffree = kstfs.f_ffree; + } + + return err; +} diff --git a/fs/aufs/branch.h b/fs/aufs/branch.h new file mode 100644 index 00000000000..4c00b9183ad --- /dev/null +++ b/fs/aufs/branch.h @@ -0,0 +1,375 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * branch filesystems and xino for them + */ + +#ifndef __AUFS_BRANCH_H__ +#define __AUFS_BRANCH_H__ + +#ifdef __KERNEL__ + +#include +#include "dirren.h" +#include "dynop.h" +#include "lcnt.h" +#include "rwsem.h" +#include "super.h" + +/* ---------------------------------------------------------------------- */ + +/* a xino file */ +struct au_xino { + struct file **xi_file; + unsigned int xi_nfile; + + struct { + spinlock_t spin; + ino_t *array; + int total; + /* reserved for future use */ + /* unsigned long *bitmap; */ + wait_queue_head_t wqh; + } xi_nondir; + + struct mutex xi_mtx; /* protects xi_file array */ + struct hlist_bl_head xi_writing; + + atomic_t xi_truncating; + + struct kref xi_kref; +}; + +/* File-based Hierarchical Storage Management */ +struct au_br_fhsm { +#ifdef CONFIG_AUFS_FHSM + struct mutex bf_lock; + unsigned long bf_jiffy; + struct aufs_stfs bf_stfs; + int bf_readable; +#endif +}; + +/* members for writable branch only */ +enum {AuBrWh_BASE, AuBrWh_PLINK, AuBrWh_ORPH, AuBrWh_Last}; +struct au_wbr { + struct au_rwsem wbr_wh_rwsem; + struct dentry *wbr_wh[AuBrWh_Last]; + atomic_t wbr_wh_running; +#define wbr_whbase wbr_wh[AuBrWh_BASE] /* whiteout base */ +#define wbr_plink wbr_wh[AuBrWh_PLINK] /* pseudo-link dir */ +#define wbr_orph wbr_wh[AuBrWh_ORPH] /* dir for orphans */ + + /* mfs mode */ + unsigned long long wbr_bytes; +}; + +/* ext2 has 3 types of operations at least, ext3 has 4 */ +#define AuBrDynOp (AuDyLast * 4) + +#ifdef CONFIG_AUFS_HFSNOTIFY +/* support for asynchronous destruction */ +struct au_br_hfsnotify { + struct fsnotify_group *hfsn_group; +}; +#endif + +/* sysfs entries */ +struct au_brsysfs { + char name[16]; + struct attribute attr; +}; + +enum { + AuBrSysfs_BR, + AuBrSysfs_BRID, + AuBrSysfs_Last +}; + +/* protected by superblock rwsem */ +struct au_branch { + struct au_xino *br_xino; + + aufs_bindex_t br_id; + + int br_perm; + struct path br_path; + spinlock_t br_dykey_lock; + struct au_dykey *br_dykey[AuBrDynOp]; + au_lcnt_t br_nfiles; /* opened files */ + au_lcnt_t br_count; /* in-use for other */ + + struct au_wbr *br_wbr; + struct au_br_fhsm *br_fhsm; + +#ifdef CONFIG_AUFS_HFSNOTIFY + struct au_br_hfsnotify *br_hfsn; +#endif + +#ifdef CONFIG_SYSFS + /* entries under sysfs per mount-point */ + struct au_brsysfs br_sysfs[AuBrSysfs_Last]; +#endif + +#ifdef CONFIG_DEBUG_FS + struct dentry *br_dbgaufs; /* xino */ +#endif + + struct au_dr_br br_dirren; +}; + +/* ---------------------------------------------------------------------- */ + +static inline struct vfsmount *au_br_mnt(struct au_branch *br) +{ + return br->br_path.mnt; +} + +static inline struct dentry *au_br_dentry(struct au_branch *br) +{ + return br->br_path.dentry; +} + +static inline struct user_namespace *au_br_userns(struct au_branch *br) +{ + return mnt_user_ns(br->br_path.mnt); +} + +static inline struct super_block *au_br_sb(struct au_branch *br) +{ + return au_br_mnt(br)->mnt_sb; +} + +static inline int au_br_rdonly(struct au_branch *br) +{ + return (sb_rdonly(au_br_sb(br)) + || !au_br_writable(br->br_perm)) + ? -EROFS : 0; +} + +static inline int au_br_hnotifyable(int brperm __maybe_unused) +{ +#ifdef CONFIG_AUFS_HNOTIFY + return !(brperm & AuBrPerm_RR); +#else + return 0; +#endif +} + +static inline int au_br_test_oflag(int oflag, struct au_branch *br) +{ + int err, exec_flag; + + err = 0; + exec_flag = oflag & __FMODE_EXEC; + if (unlikely(exec_flag && path_noexec(&br->br_path))) + err = -EACCES; + + return err; +} + +static inline void au_xino_get(struct au_branch *br) +{ + struct au_xino *xi; + + xi = br->br_xino; + if (xi) + kref_get(&xi->xi_kref); +} + +static inline int au_xino_count(struct au_branch *br) +{ + int v; + struct au_xino *xi; + + v = 0; + xi = br->br_xino; + if (xi) + v = kref_read(&xi->xi_kref); + + return v; +} + +/* ---------------------------------------------------------------------- */ + +/* branch.c */ +struct au_sbinfo; +void au_br_free(struct au_sbinfo *sinfo); +int au_br_index(struct super_block *sb, aufs_bindex_t br_id); +struct au_opt_add; +int au_br_add(struct super_block *sb, struct au_opt_add *add, int remount); +struct au_opt_del; +int au_br_del(struct super_block *sb, struct au_opt_del *del, int remount); +long au_ibusy_ioctl(struct file *file, unsigned long arg); +#ifdef CONFIG_COMPAT +long au_ibusy_compat_ioctl(struct file *file, unsigned long arg); +#endif +struct au_opt_mod; +int au_br_mod(struct super_block *sb, struct au_opt_mod *mod, int remount, + int *do_refresh); +struct aufs_stfs; +int au_br_stfs(struct au_branch *br, struct aufs_stfs *stfs); + +/* xino.c */ +static const loff_t au_loff_max = LLONG_MAX; + +aufs_bindex_t au_xi_root(struct super_block *sb, struct dentry *dentry); +struct file *au_xino_create(struct super_block *sb, char *fpath, int silent, + int wbrtop); +struct file *au_xino_create2(struct super_block *sb, struct path *base, + struct file *copy_src); +struct au_xi_new { + struct au_xino *xi; /* switch between xino and xigen */ + int idx; + struct path *base; + struct file *copy_src; +}; +struct file *au_xi_new(struct super_block *sb, struct au_xi_new *xinew); + +int au_xino_read(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + ino_t *ino); +int au_xino_write(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + ino_t ino); +ssize_t xino_fread(struct file *file, void *buf, size_t size, loff_t *pos); +ssize_t xino_fwrite(struct file *file, void *buf, size_t size, loff_t *pos); + +int au_xib_trunc(struct super_block *sb); +int au_xino_trunc(struct super_block *sb, aufs_bindex_t bindex, int idx_begin); + +struct au_xino *au_xino_alloc(unsigned int nfile); +int au_xino_put(struct au_branch *br); +struct file *au_xino_file1(struct au_xino *xi); + +struct au_opt_xino; +void au_xino_clr(struct super_block *sb); +int au_xino_set(struct super_block *sb, struct au_opt_xino *xiopt, int remount); +struct file *au_xino_def(struct super_block *sb); +int au_xino_init_br(struct super_block *sb, struct au_branch *br, ino_t hino, + struct path *base); + +ino_t au_xino_new_ino(struct super_block *sb); +void au_xino_delete_inode(struct inode *inode, const int unlinked); + +void au_xinondir_leave(struct super_block *sb, aufs_bindex_t bindex, + ino_t h_ino, int idx); +int au_xinondir_enter(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + int *idx); + +int au_xino_path(struct seq_file *seq, struct file *file); + +/* ---------------------------------------------------------------------- */ + +/* @idx is signed to accept -1 meaning the first file */ +static inline struct file *au_xino_file(struct au_xino *xi, int idx) +{ + struct file *file; + + file = NULL; + if (!xi) + goto out; + + if (idx >= 0) { + if (idx < xi->xi_nfile) + file = xi->xi_file[idx]; + } else + file = au_xino_file1(xi); + +out: + return file; +} + +/* ---------------------------------------------------------------------- */ + +/* Superblock to branch */ +static inline +aufs_bindex_t au_sbr_id(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_sbr(sb, bindex)->br_id; +} + +static inline +struct vfsmount *au_sbr_mnt(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_br_mnt(au_sbr(sb, bindex)); +} + +static inline +struct user_namespace *au_sbr_userns(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_br_userns(au_sbr(sb, bindex)); +} + +static inline +struct super_block *au_sbr_sb(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_br_sb(au_sbr(sb, bindex)); +} + +static inline int au_sbr_perm(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_sbr(sb, bindex)->br_perm; +} + +static inline int au_sbr_whable(struct super_block *sb, aufs_bindex_t bindex) +{ + return au_br_whable(au_sbr_perm(sb, bindex)); +} + +/* ---------------------------------------------------------------------- */ + +#define wbr_wh_read_lock(wbr) au_rw_read_lock(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_write_lock(wbr) au_rw_write_lock(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_read_trylock(wbr) au_rw_read_trylock(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_write_trylock(wbr) au_rw_write_trylock(&(wbr)->wbr_wh_rwsem) +/* +#define wbr_wh_read_trylock_nested(wbr) \ + au_rw_read_trylock_nested(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_write_trylock_nested(wbr) \ + au_rw_write_trylock_nested(&(wbr)->wbr_wh_rwsem) +*/ + +#define wbr_wh_read_unlock(wbr) au_rw_read_unlock(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_write_unlock(wbr) au_rw_write_unlock(&(wbr)->wbr_wh_rwsem) +#define wbr_wh_downgrade_lock(wbr) au_rw_dgrade_lock(&(wbr)->wbr_wh_rwsem) + +#define WbrWhMustNoWaiters(wbr) AuRwMustNoWaiters(&(wbr)->wbr_wh_rwsem) +#define WbrWhMustAnyLock(wbr) AuRwMustAnyLock(&(wbr)->wbr_wh_rwsem) +#define WbrWhMustWriteLock(wbr) AuRwMustWriteLock(&(wbr)->wbr_wh_rwsem) + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_FHSM +static inline void au_br_fhsm_init(struct au_br_fhsm *brfhsm) +{ + mutex_init(&brfhsm->bf_lock); + brfhsm->bf_jiffy = 0; + brfhsm->bf_readable = 0; +} + +static inline void au_br_fhsm_fin(struct au_br_fhsm *brfhsm) +{ + mutex_destroy(&brfhsm->bf_lock); +} +#else +AuStubVoid(au_br_fhsm_init, struct au_br_fhsm *brfhsm) +AuStubVoid(au_br_fhsm_fin, struct au_br_fhsm *brfhsm) +#endif + +#endif /* __KERNEL__ */ +#endif /* __AUFS_BRANCH_H__ */ diff --git a/fs/aufs/conf.mk b/fs/aufs/conf.mk new file mode 100644 index 00000000000..12782f8e0f3 --- /dev/null +++ b/fs/aufs/conf.mk @@ -0,0 +1,40 @@ +# SPDX-License-Identifier: GPL-2.0 + +AuConfStr = CONFIG_AUFS_FS=${CONFIG_AUFS_FS} + +define AuConf +ifdef ${1} +AuConfStr += ${1}=${${1}} +endif +endef + +AuConfAll = BRANCH_MAX_127 BRANCH_MAX_511 BRANCH_MAX_1023 BRANCH_MAX_32767 \ + SBILIST \ + HNOTIFY HFSNOTIFY \ + EXPORT INO_T_64 \ + XATTR \ + FHSM \ + RDU \ + DIRREN \ + SHWH \ + BR_RAMFS \ + BR_FUSE POLL \ + BR_HFSPLUS \ + BDEV_LOOP \ + DEBUG MAGIC_SYSRQ +$(foreach i, ${AuConfAll}, \ + $(eval $(call AuConf,CONFIG_AUFS_${i}))) + +AuConfName = ${obj}/conf.str +${AuConfName}.tmp: FORCE + @echo ${AuConfStr} | tr ' ' '\n' | sed -e 's/^/"/' -e 's/$$/\\n"/' > $@ +${AuConfName}: ${AuConfName}.tmp + @diff -q $< $@ > /dev/null 2>&1 || { \ + echo ' GEN ' $@; \ + cp -p $< $@; \ + } +FORCE: +clean-files += ${AuConfName} ${AuConfName}.tmp +${obj}/sysfs.o: ${AuConfName} + +-include ${srctree}/${src}/conf_priv.mk diff --git a/fs/aufs/cpup.c b/fs/aufs/cpup.c new file mode 100644 index 00000000000..0513764f0c8 --- /dev/null +++ b/fs/aufs/cpup.c @@ -0,0 +1,1457 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * copy-up functions, see wbr_policy.c for copy-down + */ + +#include +#include +#include +#include "aufs.h" + +void au_cpup_attr_flags(struct inode *dst, unsigned int iflags) +{ + const unsigned int mask = S_DEAD | S_SWAPFILE | S_PRIVATE + | S_NOATIME | S_NOCMTIME | S_AUTOMOUNT; + + BUILD_BUG_ON(sizeof(iflags) != sizeof(dst->i_flags)); + + dst->i_flags |= iflags & ~mask; + if (au_test_fs_notime(dst->i_sb)) + dst->i_flags |= S_NOATIME | S_NOCMTIME; +} + +void au_cpup_attr_timesizes(struct inode *inode) +{ + struct inode *h_inode; + + h_inode = au_h_iptr(inode, au_ibtop(inode)); + fsstack_copy_attr_times(inode, h_inode); + fsstack_copy_inode_size(inode, h_inode); +} + +void au_cpup_attr_nlink(struct inode *inode, int force) +{ + struct inode *h_inode; + struct super_block *sb; + aufs_bindex_t bindex, bbot; + + sb = inode->i_sb; + bindex = au_ibtop(inode); + h_inode = au_h_iptr(inode, bindex); + if (!force + && !S_ISDIR(h_inode->i_mode) + && au_opt_test(au_mntflags(sb), PLINK) + && au_plink_test(inode)) + return; + + /* + * 0 can happen in revalidating. + * h_inode->i_mutex may not be held here, but it is harmless since once + * i_nlink reaches 0, it will never become positive except O_TMPFILE + * case. + * todo: O_TMPFILE+linkat(AT_SYMLINK_FOLLOW) bypassing aufs may cause + * the incorrect link count. + */ + set_nlink(inode, h_inode->i_nlink); + + /* + * fewer nlink makes find(1) noisy, but larger nlink doesn't. + * it may includes whplink directory. + */ + if (S_ISDIR(h_inode->i_mode)) { + bbot = au_ibbot(inode); + for (bindex++; bindex <= bbot; bindex++) { + h_inode = au_h_iptr(inode, bindex); + if (h_inode) + au_add_nlink(inode, h_inode); + } + } +} + +void au_cpup_attr_changeable(struct inode *inode) +{ + struct inode *h_inode; + + h_inode = au_h_iptr(inode, au_ibtop(inode)); + inode->i_mode = h_inode->i_mode; + inode->i_uid = h_inode->i_uid; + inode->i_gid = h_inode->i_gid; + au_cpup_attr_timesizes(inode); + au_cpup_attr_flags(inode, h_inode->i_flags); +} + +void au_cpup_igen(struct inode *inode, struct inode *h_inode) +{ + struct au_iinfo *iinfo = au_ii(inode); + + IiMustWriteLock(inode); + + iinfo->ii_higen = h_inode->i_generation; + iinfo->ii_hsb1 = h_inode->i_sb; +} + +void au_cpup_attr_all(struct inode *inode, int force) +{ + struct inode *h_inode; + + h_inode = au_h_iptr(inode, au_ibtop(inode)); + au_cpup_attr_changeable(inode); + if (inode->i_nlink > 0) + au_cpup_attr_nlink(inode, force); + inode->i_rdev = h_inode->i_rdev; + inode->i_blkbits = h_inode->i_blkbits; + au_cpup_igen(inode, h_inode); +} + +/* ---------------------------------------------------------------------- */ + +/* Note: dt_dentry and dt_h_dentry are not dget/dput-ed */ + +/* keep the timestamps of the parent dir when cpup */ +void au_dtime_store(struct au_dtime *dt, struct dentry *dentry, + struct path *h_path) +{ + struct inode *h_inode; + + dt->dt_dentry = dentry; + dt->dt_h_path = *h_path; + h_inode = d_inode(h_path->dentry); + dt->dt_atime = h_inode->i_atime; + dt->dt_mtime = h_inode->i_mtime; + /* smp_mb(); */ +} + +void au_dtime_revert(struct au_dtime *dt) +{ + struct iattr attr; + int err; + + attr.ia_atime = dt->dt_atime; + attr.ia_mtime = dt->dt_mtime; + attr.ia_valid = ATTR_FORCE | ATTR_MTIME | ATTR_MTIME_SET + | ATTR_ATIME | ATTR_ATIME_SET; + + /* no delegation since this is a directory */ + err = vfsub_notify_change(&dt->dt_h_path, &attr, /*delegated*/NULL); + if (unlikely(err)) + pr_warn("restoring timestamps failed(%d). ignored\n", err); +} + +/* ---------------------------------------------------------------------- */ + +/* internal use only */ +struct au_cpup_reg_attr { + int valid; + struct kstat st; + unsigned int iflags; /* inode->i_flags */ +}; + +static noinline_for_stack +int cpup_iattr(struct dentry *dst, aufs_bindex_t bindex, struct path *h_src, + struct au_cpup_reg_attr *h_src_attr) +{ + int err, sbits, icex; + unsigned int mnt_flags; + unsigned char verbose; + struct iattr ia; + struct path h_path; + struct inode *h_isrc, *h_idst; + struct kstat *h_st; + struct au_branch *br; + + br = au_sbr(dst->d_sb, bindex); + h_path.mnt = au_br_mnt(br); + h_path.dentry = au_h_dptr(dst, bindex); + h_idst = d_inode(h_path.dentry); + h_isrc = d_inode(h_src->dentry); + ia.ia_valid = ATTR_FORCE | ATTR_UID | ATTR_GID + | ATTR_ATIME | ATTR_MTIME + | ATTR_ATIME_SET | ATTR_MTIME_SET; + if (h_src_attr && h_src_attr->valid) { + h_st = &h_src_attr->st; + ia.ia_uid = h_st->uid; + ia.ia_gid = h_st->gid; + ia.ia_atime = h_st->atime; + ia.ia_mtime = h_st->mtime; + if (h_idst->i_mode != h_st->mode + && !S_ISLNK(h_idst->i_mode)) { + ia.ia_valid |= ATTR_MODE; + ia.ia_mode = h_st->mode; + } + sbits = !!(h_st->mode & (S_ISUID | S_ISGID)); + au_cpup_attr_flags(h_idst, h_src_attr->iflags); + } else { + ia.ia_uid = h_isrc->i_uid; + ia.ia_gid = h_isrc->i_gid; + ia.ia_atime = h_isrc->i_atime; + ia.ia_mtime = h_isrc->i_mtime; + if (h_idst->i_mode != h_isrc->i_mode + && !S_ISLNK(h_idst->i_mode)) { + ia.ia_valid |= ATTR_MODE; + ia.ia_mode = h_isrc->i_mode; + } + sbits = !!(h_isrc->i_mode & (S_ISUID | S_ISGID)); + au_cpup_attr_flags(h_idst, h_isrc->i_flags); + } + /* no delegation since it is just created */ + err = vfsub_notify_change(&h_path, &ia, /*delegated*/NULL); + + /* is this nfs only? */ + if (!err && sbits && au_test_nfs(h_path.dentry->d_sb)) { + ia.ia_valid = ATTR_FORCE | ATTR_MODE; + ia.ia_mode = h_isrc->i_mode; + err = vfsub_notify_change(&h_path, &ia, /*delegated*/NULL); + } + + icex = br->br_perm & AuBrAttr_ICEX; + if (!err) { + mnt_flags = au_mntflags(dst->d_sb); + verbose = !!au_opt_test(mnt_flags, VERBOSE); + err = au_cpup_xattr(&h_path, h_src, icex, verbose); + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_do_copy_file(struct file *dst, struct file *src, loff_t len, + char *buf, unsigned long blksize) +{ + int err; + size_t sz, rbytes, wbytes; + unsigned char all_zero; + char *p, *zp; + struct inode *h_inode; + /* reduce stack usage */ + struct iattr *ia; + + zp = page_address(ZERO_PAGE(0)); + if (unlikely(!zp)) + return -ENOMEM; /* possible? */ + + err = 0; + all_zero = 0; + while (len) { + AuDbg("len %lld\n", len); + sz = blksize; + if (len < blksize) + sz = len; + + rbytes = 0; + /* todo: signal_pending? */ + while (!rbytes || err == -EAGAIN || err == -EINTR) { + rbytes = vfsub_read_k(src, buf, sz, &src->f_pos); + err = rbytes; + } + if (unlikely(err < 0)) + break; + + all_zero = 0; + if (len >= rbytes && rbytes == blksize) + all_zero = !memcmp(buf, zp, rbytes); + if (!all_zero) { + wbytes = rbytes; + p = buf; + while (wbytes) { + size_t b; + + b = vfsub_write_k(dst, p, wbytes, &dst->f_pos); + err = b; + /* todo: signal_pending? */ + if (unlikely(err == -EAGAIN || err == -EINTR)) + continue; + if (unlikely(err < 0)) + break; + wbytes -= b; + p += b; + } + if (unlikely(err < 0)) + break; + } else { + loff_t res; + + AuLabel(hole); + res = vfsub_llseek(dst, rbytes, SEEK_CUR); + err = res; + if (unlikely(res < 0)) + break; + } + len -= rbytes; + err = 0; + } + + /* the last block may be a hole */ + if (!err && all_zero) { + AuLabel(last hole); + + err = 1; + if (au_test_nfs(dst->f_path.dentry->d_sb)) { + /* nfs requires this step to make last hole */ + /* is this only nfs? */ + do { + /* todo: signal_pending? */ + err = vfsub_write_k(dst, "\0", 1, &dst->f_pos); + } while (err == -EAGAIN || err == -EINTR); + if (err == 1) + dst->f_pos--; + } + + if (err == 1) { + ia = (void *)buf; + ia->ia_size = dst->f_pos; + ia->ia_valid = ATTR_SIZE | ATTR_FILE; + ia->ia_file = dst; + h_inode = file_inode(dst); + inode_lock_nested(h_inode, AuLsc_I_CHILD2); + /* no delegation since it is just created */ + err = vfsub_notify_change(&dst->f_path, ia, + /*delegated*/NULL); + inode_unlock(h_inode); + } + } + + return err; +} + +int au_copy_file(struct file *dst, struct file *src, loff_t len) +{ + int err; + unsigned long blksize; + unsigned char do_kfree; + char *buf; + struct super_block *h_sb; + + err = -ENOMEM; + h_sb = file_inode(dst)->i_sb; + blksize = h_sb->s_blocksize; + if (!blksize || PAGE_SIZE < blksize) + blksize = PAGE_SIZE; + AuDbg("blksize %lu\n", blksize); + do_kfree = (blksize != PAGE_SIZE && blksize >= sizeof(struct iattr *)); + if (do_kfree) + buf = kmalloc(blksize, GFP_NOFS); + else + buf = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!buf)) + goto out; + + if (len > (1 << 22)) + AuDbg("copying a large file %lld\n", (long long)len); + + src->f_pos = 0; + dst->f_pos = 0; + err = au_do_copy_file(dst, src, len, buf, blksize); + if (do_kfree) { + AuDebugOn(!au_kfree_do_sz_test(blksize)); + au_kfree_do_rcu(buf); + } else + free_page((unsigned long)buf); + +out: + return err; +} + +static int au_do_copy(struct file *dst, struct file *src, loff_t len) +{ + int err; + struct super_block *h_src_sb; + struct inode *h_src_inode; + + h_src_inode = file_inode(src); + h_src_sb = h_src_inode->i_sb; + + /* XFS acquires inode_lock */ + if (!au_test_xfs(h_src_sb)) + err = au_copy_file(dst, src, len); + else { + inode_unlock_shared(h_src_inode); + err = au_copy_file(dst, src, len); + inode_lock_shared_nested(h_src_inode, AuLsc_I_CHILD); + } + + return err; +} + +static int au_clone_or_copy(struct file *dst, struct file *src, loff_t len) +{ + int err; + loff_t lo; + struct super_block *h_src_sb; + struct inode *h_src_inode; + + h_src_inode = file_inode(src); + h_src_sb = h_src_inode->i_sb; + if (h_src_sb != file_inode(dst)->i_sb + || !dst->f_op->remap_file_range) { + err = au_do_copy(dst, src, len); + goto out; + } + + if (!au_test_nfs(h_src_sb)) { + inode_unlock_shared(h_src_inode); + lo = vfsub_clone_file_range(src, dst, len); + inode_lock_shared_nested(h_src_inode, AuLsc_I_CHILD); + } else + lo = vfsub_clone_file_range(src, dst, len); + if (lo == len) { + err = 0; + goto out; /* success */ + } else if (lo >= 0) + /* todo: possible? */ + /* paritially succeeded */ + AuDbg("lo %lld, len %lld. Retrying.\n", lo, len); + else if (lo != -EOPNOTSUPP) { + /* older XFS has a condition in cloning */ + err = lo; + goto out; + } + + /* the backend fs on NFS may not support cloning */ + err = au_do_copy(dst, src, len); + +out: + AuTraceErr(err); + return err; +} + +/* + * to support a sparse file which is opened with O_APPEND, + * we need to close the file. + */ +static int au_cp_regular(struct au_cp_generic *cpg) +{ + int err, i; + enum { SRC, DST }; + struct { + aufs_bindex_t bindex; + unsigned int flags; + struct dentry *dentry; + int force_wr; + struct file *file; + } *f, file[] = { + { + .bindex = cpg->bsrc, + .flags = O_RDONLY | O_NOATIME | O_LARGEFILE, + }, + { + .bindex = cpg->bdst, + .flags = O_WRONLY | O_NOATIME | O_LARGEFILE, + .force_wr = !!au_ftest_cpup(cpg->flags, RWDST), + } + }; + struct au_branch *br; + struct super_block *sb, *h_src_sb; + struct inode *h_src_inode; + struct task_struct *tsk = current; + + /* bsrc branch can be ro/rw. */ + sb = cpg->dentry->d_sb; + f = file; + for (i = 0; i < 2; i++, f++) { + f->dentry = au_h_dptr(cpg->dentry, f->bindex); + f->file = au_h_open(cpg->dentry, f->bindex, f->flags, + /*file*/NULL, f->force_wr); + if (IS_ERR(f->file)) { + err = PTR_ERR(f->file); + if (i == SRC) + goto out; + else + goto out_src; + } + } + + /* try stopping to update while we copyup */ + h_src_inode = d_inode(file[SRC].dentry); + h_src_sb = h_src_inode->i_sb; + if (!au_test_nfs(h_src_sb)) + IMustLock(h_src_inode); + err = au_clone_or_copy(file[DST].file, file[SRC].file, cpg->len); + + /* i wonder if we had O_NO_DELAY_FPUT flag */ + if (tsk->flags & PF_KTHREAD) + __fput_sync(file[DST].file); + else { + /* it happened actually */ + fput(file[DST].file); + /* + * too bad. + * we have to call both since we don't know which place the file + * was added to. + */ + task_work_run(); + flush_delayed_fput(); + } + br = au_sbr(sb, file[DST].bindex); + au_lcnt_dec(&br->br_nfiles); + +out_src: + fput(file[SRC].file); + br = au_sbr(sb, file[SRC].bindex); + au_lcnt_dec(&br->br_nfiles); +out: + return err; +} + +static int au_do_cpup_regular(struct au_cp_generic *cpg, + struct au_cpup_reg_attr *h_src_attr) +{ + int err, rerr; + loff_t l; + struct path h_path; + struct inode *h_src_inode, *h_dst_inode; + + err = 0; + h_src_inode = au_h_iptr(d_inode(cpg->dentry), cpg->bsrc); + l = i_size_read(h_src_inode); + if (cpg->len == -1 || l < cpg->len) + cpg->len = l; + if (cpg->len) { + /* try stopping to update while we are referencing */ + inode_lock_shared_nested(h_src_inode, AuLsc_I_CHILD); + au_pin_hdir_unlock(cpg->pin); + + h_path.dentry = au_h_dptr(cpg->dentry, cpg->bsrc); + h_path.mnt = au_sbr_mnt(cpg->dentry->d_sb, cpg->bsrc); + h_src_attr->iflags = h_src_inode->i_flags; + if (!au_test_nfs(h_src_inode->i_sb)) + err = vfsub_getattr(&h_path, &h_src_attr->st); + else { + inode_unlock_shared(h_src_inode); + err = vfsub_getattr(&h_path, &h_src_attr->st); + inode_lock_shared_nested(h_src_inode, AuLsc_I_CHILD); + } + if (unlikely(err)) { + inode_unlock_shared(h_src_inode); + goto out; + } + h_src_attr->valid = 1; + if (!au_test_nfs(h_src_inode->i_sb)) { + err = au_cp_regular(cpg); + inode_unlock_shared(h_src_inode); + } else { + inode_unlock_shared(h_src_inode); + err = au_cp_regular(cpg); + } + rerr = au_pin_hdir_relock(cpg->pin); + if (!err && rerr) + err = rerr; + } + if (!err && (h_src_inode->i_state & I_LINKABLE)) { + h_path.dentry = au_h_dptr(cpg->dentry, cpg->bdst); + h_dst_inode = d_inode(h_path.dentry); + spin_lock(&h_dst_inode->i_lock); + h_dst_inode->i_state |= I_LINKABLE; + spin_unlock(&h_dst_inode->i_lock); + } + +out: + return err; +} + +static int au_do_cpup_symlink(struct path *h_path, struct dentry *h_src, + struct inode *h_dir) +{ + int err; + DEFINE_DELAYED_CALL(done); + const char *sym; + + sym = vfs_get_link(h_src, &done); + err = PTR_ERR(sym); + if (IS_ERR(sym)) + goto out; + + err = vfsub_symlink(h_dir, h_path, sym); + +out: + do_delayed_call(&done); + return err; +} + +/* + * regardless 'acl' option, reset all ACL. + * All ACL will be copied up later from the original entry on the lower branch. + */ +static int au_reset_acl(struct inode *h_dir, struct path *h_path, umode_t mode) +{ + int err; + struct dentry *h_dentry; + struct inode *h_inode; + struct user_namespace *h_userns; + + h_userns = mnt_user_ns(h_path->mnt); + h_dentry = h_path->dentry; + h_inode = d_inode(h_dentry); + /* forget_all_cached_acls(h_inode)); */ + err = vfsub_removexattr(h_userns, h_dentry, XATTR_NAME_POSIX_ACL_ACCESS); + AuTraceErr(err); + if (err == -EOPNOTSUPP) + err = 0; + if (!err) + err = vfsub_acl_chmod(h_userns, h_inode, mode); + + AuTraceErr(err); + return err; +} + +static int au_do_cpup_dir(struct au_cp_generic *cpg, struct dentry *dst_parent, + struct inode *h_dir, struct path *h_path) +{ + int err; + struct inode *dir, *inode; + struct user_namespace *h_userns; + + h_userns = mnt_user_ns(h_path->mnt); + err = vfsub_removexattr(h_userns, h_path->dentry, + XATTR_NAME_POSIX_ACL_DEFAULT); + AuTraceErr(err); + if (err == -EOPNOTSUPP) + err = 0; + if (unlikely(err)) + goto out; + + /* + * strange behaviour from the users view, + * particularly setattr case + */ + dir = d_inode(dst_parent); + if (au_ibtop(dir) == cpg->bdst) + au_cpup_attr_nlink(dir, /*force*/1); + inode = d_inode(cpg->dentry); + au_cpup_attr_nlink(inode, /*force*/1); + +out: + return err; +} + +static noinline_for_stack +int cpup_entry(struct au_cp_generic *cpg, struct dentry *dst_parent, + struct au_cpup_reg_attr *h_src_attr) +{ + int err; + umode_t mode; + unsigned int mnt_flags; + unsigned char isdir, isreg, force; + const unsigned char do_dt = !!au_ftest_cpup(cpg->flags, DTIME); + struct au_dtime dt; + struct path h_path; + struct dentry *h_src, *h_dst, *h_parent; + struct inode *h_inode, *h_dir; + struct super_block *sb; + + /* bsrc branch can be ro/rw. */ + h_src = au_h_dptr(cpg->dentry, cpg->bsrc); + h_inode = d_inode(h_src); + AuDebugOn(h_inode != au_h_iptr(d_inode(cpg->dentry), cpg->bsrc)); + + /* try stopping to be referenced while we are creating */ + h_dst = au_h_dptr(cpg->dentry, cpg->bdst); + if (au_ftest_cpup(cpg->flags, RENAME)) + AuDebugOn(strncmp(h_dst->d_name.name, AUFS_WH_PFX, + AUFS_WH_PFX_LEN)); + h_parent = h_dst->d_parent; /* dir inode is locked */ + h_dir = d_inode(h_parent); + IMustLock(h_dir); + AuDebugOn(h_parent != h_dst->d_parent); + + sb = cpg->dentry->d_sb; + h_path.mnt = au_sbr_mnt(sb, cpg->bdst); + if (do_dt) { + h_path.dentry = h_parent; + au_dtime_store(&dt, dst_parent, &h_path); + } + h_path.dentry = h_dst; + + isreg = 0; + isdir = 0; + mode = h_inode->i_mode; + switch (mode & S_IFMT) { + case S_IFREG: + isreg = 1; + err = vfsub_create(h_dir, &h_path, 0600, /*want_excl*/true); + if (!err) + err = au_do_cpup_regular(cpg, h_src_attr); + break; + case S_IFDIR: + isdir = 1; + err = vfsub_mkdir(h_dir, &h_path, mode); + if (!err) + err = au_do_cpup_dir(cpg, dst_parent, h_dir, &h_path); + break; + case S_IFLNK: + err = au_do_cpup_symlink(&h_path, h_src, h_dir); + break; + case S_IFCHR: + case S_IFBLK: + AuDebugOn(!capable(CAP_MKNOD)); + fallthrough; + case S_IFIFO: + case S_IFSOCK: + err = vfsub_mknod(h_dir, &h_path, mode, h_inode->i_rdev); + break; + default: + AuIOErr("Unknown inode type 0%o\n", mode); + err = -EIO; + } + if (!err) + err = au_reset_acl(h_dir, &h_path, mode); + + mnt_flags = au_mntflags(sb); + if (!au_opt_test(mnt_flags, UDBA_NONE) + && !isdir + && au_opt_test(mnt_flags, XINO) + && (h_inode->i_nlink == 1 + || (h_inode->i_state & I_LINKABLE)) + /* todo: unnecessary? */ + /* && d_inode(cpg->dentry)->i_nlink == 1 */ + && cpg->bdst < cpg->bsrc + && !au_ftest_cpup(cpg->flags, KEEPLINO)) + au_xino_write(sb, cpg->bsrc, h_inode->i_ino, /*ino*/0); + /* ignore this error */ + + if (!err) { + force = 0; + if (isreg) { + force = !!cpg->len; + if (cpg->len == -1) + force = !!i_size_read(h_inode); + } + au_fhsm_wrote(sb, cpg->bdst, force); + } + + if (do_dt) + au_dtime_revert(&dt); + return err; +} + +static int au_do_ren_after_cpup(struct au_cp_generic *cpg, struct path *h_path) +{ + int err; + struct dentry *dentry, *h_dentry, *h_parent, *parent; + struct inode *h_dir; + aufs_bindex_t bdst; + + dentry = cpg->dentry; + bdst = cpg->bdst; + h_dentry = au_h_dptr(dentry, bdst); + if (!au_ftest_cpup(cpg->flags, OVERWRITE)) { + dget(h_dentry); + au_set_h_dptr(dentry, bdst, NULL); + err = au_lkup_neg(dentry, bdst, /*wh*/0); + if (!err) + h_path->dentry = dget(au_h_dptr(dentry, bdst)); + au_set_h_dptr(dentry, bdst, h_dentry); + } else { + err = 0; + parent = dget_parent(dentry); + h_parent = au_h_dptr(parent, bdst); + dput(parent); + h_path->dentry = vfsub_lkup_one(&dentry->d_name, h_parent); + if (IS_ERR(h_path->dentry)) + err = PTR_ERR(h_path->dentry); + } + if (unlikely(err)) + goto out; + + h_parent = h_dentry->d_parent; /* dir inode is locked */ + h_dir = d_inode(h_parent); + IMustLock(h_dir); + AuDbg("%pd %pd\n", h_dentry, h_path->dentry); + /* no delegation since it is just created */ + err = vfsub_rename(h_dir, h_dentry, h_dir, h_path, /*delegated*/NULL, + /*flags*/0); + dput(h_path->dentry); + +out: + return err; +} + +/* + * copyup the @dentry from @bsrc to @bdst. + * the caller must set the both of lower dentries. + * @len is for truncating when it is -1 copyup the entire file. + * in link/rename cases, @dst_parent may be different from the real one. + * basic->bsrc can be larger than basic->bdst. + * aufs doesn't touch the credential so + * security_inode_copy_up{,_xattr}() are unnecessary. + */ +static int au_cpup_single(struct au_cp_generic *cpg, struct dentry *dst_parent) +{ + int err, rerr; + aufs_bindex_t old_ibtop; + unsigned char isdir, plink; + struct dentry *h_src, *h_dst, *h_parent; + struct inode *dst_inode, *h_dir, *inode, *delegated, *src_inode; + struct super_block *sb; + struct au_branch *br; + struct path h_src_path; + /* to reduce stack size */ + struct { + struct au_dtime dt; + struct path h_path; + struct au_cpup_reg_attr h_src_attr; + } *a; + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + a->h_src_attr.valid = 0; + + sb = cpg->dentry->d_sb; + br = au_sbr(sb, cpg->bdst); + a->h_path.mnt = au_br_mnt(br); + h_dst = au_h_dptr(cpg->dentry, cpg->bdst); + h_parent = h_dst->d_parent; /* dir inode is locked */ + h_dir = d_inode(h_parent); + IMustLock(h_dir); + + h_src = au_h_dptr(cpg->dentry, cpg->bsrc); + inode = d_inode(cpg->dentry); + + if (!dst_parent) + dst_parent = dget_parent(cpg->dentry); + else + dget(dst_parent); + + plink = !!au_opt_test(au_mntflags(sb), PLINK); + dst_inode = au_h_iptr(inode, cpg->bdst); + if (dst_inode) { + if (unlikely(!plink)) { + err = -EIO; + AuIOErr("hi%lu(i%lu) exists on b%d " + "but plink is disabled\n", + dst_inode->i_ino, inode->i_ino, cpg->bdst); + goto out_parent; + } + + if (dst_inode->i_nlink) { + const int do_dt = au_ftest_cpup(cpg->flags, DTIME); + + h_src = au_plink_lkup(inode, cpg->bdst); + err = PTR_ERR(h_src); + if (IS_ERR(h_src)) + goto out_parent; + if (unlikely(d_is_negative(h_src))) { + err = -EIO; + AuIOErr("i%lu exists on b%d " + "but not pseudo-linked\n", + inode->i_ino, cpg->bdst); + dput(h_src); + goto out_parent; + } + + if (do_dt) { + a->h_path.dentry = h_parent; + au_dtime_store(&a->dt, dst_parent, &a->h_path); + } + + a->h_path.dentry = h_dst; + delegated = NULL; + err = vfsub_link(h_src, h_dir, &a->h_path, &delegated); + if (!err && au_ftest_cpup(cpg->flags, RENAME)) + err = au_do_ren_after_cpup(cpg, &a->h_path); + if (do_dt) + au_dtime_revert(&a->dt); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + dput(h_src); + goto out_parent; + } else + /* todo: cpup_wh_file? */ + /* udba work */ + au_update_ibrange(inode, /*do_put_zero*/1); + } + + isdir = S_ISDIR(inode->i_mode); + old_ibtop = au_ibtop(inode); + err = cpup_entry(cpg, dst_parent, &a->h_src_attr); + if (unlikely(err)) + goto out_rev; + dst_inode = d_inode(h_dst); + inode_lock_nested(dst_inode, AuLsc_I_CHILD2); + /* todo: necessary? */ + /* au_pin_hdir_unlock(cpg->pin); */ + + h_src_path.dentry = h_src; + h_src_path.mnt = au_sbr_mnt(sb, cpg->bsrc); + err = cpup_iattr(cpg->dentry, cpg->bdst, &h_src_path, &a->h_src_attr); + if (unlikely(err)) { + /* todo: necessary? */ + /* au_pin_hdir_relock(cpg->pin); */ /* ignore an error */ + inode_unlock(dst_inode); + goto out_rev; + } + + if (cpg->bdst < old_ibtop) { + if (S_ISREG(inode->i_mode)) { + err = au_dy_iaop(inode, cpg->bdst, dst_inode); + if (unlikely(err)) { + /* ignore an error */ + /* au_pin_hdir_relock(cpg->pin); */ + inode_unlock(dst_inode); + goto out_rev; + } + } + au_set_ibtop(inode, cpg->bdst); + } else + au_set_ibbot(inode, cpg->bdst); + au_set_h_iptr(inode, cpg->bdst, au_igrab(dst_inode), + au_hi_flags(inode, isdir)); + + /* todo: necessary? */ + /* err = au_pin_hdir_relock(cpg->pin); */ + inode_unlock(dst_inode); + if (unlikely(err)) + goto out_rev; + + src_inode = d_inode(h_src); + if (!isdir + && (src_inode->i_nlink > 1 + || src_inode->i_state & I_LINKABLE) + && plink) + au_plink_append(inode, cpg->bdst, h_dst); + + if (au_ftest_cpup(cpg->flags, RENAME)) { + a->h_path.dentry = h_dst; + err = au_do_ren_after_cpup(cpg, &a->h_path); + } + if (!err) + goto out_parent; /* success */ + + /* revert */ +out_rev: + a->h_path.dentry = h_parent; + au_dtime_store(&a->dt, dst_parent, &a->h_path); + a->h_path.dentry = h_dst; + rerr = 0; + if (d_is_positive(h_dst)) { + if (!isdir) { + /* no delegation since it is just created */ + rerr = vfsub_unlink(h_dir, &a->h_path, + /*delegated*/NULL, /*force*/0); + } else + rerr = vfsub_rmdir(h_dir, &a->h_path); + } + au_dtime_revert(&a->dt); + if (rerr) { + AuIOErr("failed removing broken entry(%d, %d)\n", err, rerr); + err = -EIO; + } +out_parent: + dput(dst_parent); + au_kfree_rcu(a); +out: + return err; +} + +#if 0 /* reserved */ +struct au_cpup_single_args { + int *errp; + struct au_cp_generic *cpg; + struct dentry *dst_parent; +}; + +static void au_call_cpup_single(void *args) +{ + struct au_cpup_single_args *a = args; + + au_pin_hdir_acquire_nest(a->cpg->pin); + *a->errp = au_cpup_single(a->cpg, a->dst_parent); + au_pin_hdir_release(a->cpg->pin); +} +#endif + +/* + * prevent SIGXFSZ in copy-up. + * testing CAP_MKNOD is for generic fs, + * but CAP_FSETID is for xfs only, currently. + */ +static int au_cpup_sio_test(struct au_pin *pin, umode_t mode) +{ + int do_sio; + struct super_block *sb; + struct inode *h_dir; + + do_sio = 0; + sb = au_pinned_parent(pin)->d_sb; + if (!au_wkq_test() + && (!au_sbi(sb)->si_plink_maint_pid + || au_plink_maint(sb, AuLock_NOPLM))) { + switch (mode & S_IFMT) { + case S_IFREG: + /* no condition about RLIMIT_FSIZE and the file size */ + do_sio = 1; + break; + case S_IFCHR: + case S_IFBLK: + do_sio = !capable(CAP_MKNOD); + break; + } + if (!do_sio) + do_sio = ((mode & (S_ISUID | S_ISGID)) + && !capable(CAP_FSETID)); + /* this workaround may be removed in the future */ + if (!do_sio) { + h_dir = au_pinned_h_dir(pin); + do_sio = h_dir->i_mode & S_ISVTX; + } + } + + return do_sio; +} + +#if 0 /* reserved */ +int au_sio_cpup_single(struct au_cp_generic *cpg, struct dentry *dst_parent) +{ + int err, wkq_err; + struct dentry *h_dentry; + + h_dentry = au_h_dptr(cpg->dentry, cpg->bsrc); + if (!au_cpup_sio_test(pin, d_inode(h_dentry)->i_mode)) + err = au_cpup_single(cpg, dst_parent); + else { + struct au_cpup_single_args args = { + .errp = &err, + .cpg = cpg, + .dst_parent = dst_parent + }; + wkq_err = au_wkq_wait(au_call_cpup_single, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + return err; +} +#endif + +/* + * copyup the @dentry from the first active lower branch to @bdst, + * using au_cpup_single(). + */ +static int au_cpup_simple(struct au_cp_generic *cpg) +{ + int err; + unsigned int flags_orig; + struct dentry *dentry; + + AuDebugOn(cpg->bsrc < 0); + + dentry = cpg->dentry; + DiMustWriteLock(dentry); + + err = au_lkup_neg(dentry, cpg->bdst, /*wh*/1); + if (!err) { + flags_orig = cpg->flags; + au_fset_cpup(cpg->flags, RENAME); + err = au_cpup_single(cpg, NULL); + cpg->flags = flags_orig; + if (!err) + return 0; /* success */ + + /* revert */ + au_set_h_dptr(dentry, cpg->bdst, NULL); + au_set_dbtop(dentry, cpg->bsrc); + } + + return err; +} + +struct au_cpup_simple_args { + int *errp; + struct au_cp_generic *cpg; +}; + +static void au_call_cpup_simple(void *args) +{ + struct au_cpup_simple_args *a = args; + + au_pin_hdir_acquire_nest(a->cpg->pin); + *a->errp = au_cpup_simple(a->cpg); + au_pin_hdir_release(a->cpg->pin); +} + +static int au_do_sio_cpup_simple(struct au_cp_generic *cpg) +{ + int err, wkq_err; + struct dentry *dentry, *parent; + struct file *h_file; + struct inode *h_dir; + struct user_namespace *h_userns; + + dentry = cpg->dentry; + h_file = NULL; + if (au_ftest_cpup(cpg->flags, HOPEN)) { + AuDebugOn(cpg->bsrc < 0); + h_file = au_h_open_pre(dentry, cpg->bsrc, /*force_wr*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + } + + parent = dget_parent(dentry); + h_dir = au_h_iptr(d_inode(parent), cpg->bdst); + h_userns = au_sbr_userns(dentry->d_sb, cpg->bdst); + if (!au_test_h_perm_sio(h_userns, h_dir, MAY_EXEC | MAY_WRITE) + && !au_cpup_sio_test(cpg->pin, d_inode(dentry)->i_mode)) + err = au_cpup_simple(cpg); + else { + struct au_cpup_simple_args args = { + .errp = &err, + .cpg = cpg + }; + wkq_err = au_wkq_wait(au_call_cpup_simple, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + dput(parent); + if (h_file) + au_h_open_post(dentry, cpg->bsrc, h_file); + +out: + return err; +} + +int au_sio_cpup_simple(struct au_cp_generic *cpg) +{ + aufs_bindex_t bsrc, bbot; + struct dentry *dentry, *h_dentry; + + if (cpg->bsrc < 0) { + dentry = cpg->dentry; + bbot = au_dbbot(dentry); + for (bsrc = cpg->bdst + 1; bsrc <= bbot; bsrc++) { + h_dentry = au_h_dptr(dentry, bsrc); + if (h_dentry) { + AuDebugOn(d_is_negative(h_dentry)); + break; + } + } + AuDebugOn(bsrc > bbot); + cpg->bsrc = bsrc; + } + AuDebugOn(cpg->bsrc <= cpg->bdst); + return au_do_sio_cpup_simple(cpg); +} + +int au_sio_cpdown_simple(struct au_cp_generic *cpg) +{ + AuDebugOn(cpg->bdst <= cpg->bsrc); + return au_do_sio_cpup_simple(cpg); +} + +/* ---------------------------------------------------------------------- */ + +/* + * copyup the deleted file for writing. + */ +static int au_do_cpup_wh(struct au_cp_generic *cpg, struct dentry *wh_dentry, + struct file *file) +{ + int err; + unsigned int flags_orig; + aufs_bindex_t bsrc_orig; + struct au_dinfo *dinfo; + struct { + struct au_hdentry *hd; + struct dentry *h_dentry; + } hdst, hsrc; + + dinfo = au_di(cpg->dentry); + AuRwMustWriteLock(&dinfo->di_rwsem); + + bsrc_orig = cpg->bsrc; + cpg->bsrc = dinfo->di_btop; + hdst.hd = au_hdentry(dinfo, cpg->bdst); + hdst.h_dentry = hdst.hd->hd_dentry; + hdst.hd->hd_dentry = wh_dentry; + dinfo->di_btop = cpg->bdst; + + hsrc.h_dentry = NULL; + if (file) { + hsrc.hd = au_hdentry(dinfo, cpg->bsrc); + hsrc.h_dentry = hsrc.hd->hd_dentry; + hsrc.hd->hd_dentry = au_hf_top(file)->f_path.dentry; + } + flags_orig = cpg->flags; + cpg->flags = !AuCpup_DTIME; + err = au_cpup_single(cpg, /*h_parent*/NULL); + cpg->flags = flags_orig; + if (file) { + if (!err) + err = au_reopen_nondir(file); + hsrc.hd->hd_dentry = hsrc.h_dentry; + } + hdst.hd->hd_dentry = hdst.h_dentry; + dinfo->di_btop = cpg->bsrc; + cpg->bsrc = bsrc_orig; + + return err; +} + +static int au_cpup_wh(struct au_cp_generic *cpg, struct file *file) +{ + int err; + aufs_bindex_t bdst; + struct au_dtime dt; + struct dentry *dentry, *parent, *h_parent, *wh_dentry; + struct au_branch *br; + struct path h_path; + + dentry = cpg->dentry; + bdst = cpg->bdst; + br = au_sbr(dentry->d_sb, bdst); + parent = dget_parent(dentry); + h_parent = au_h_dptr(parent, bdst); + wh_dentry = au_whtmp_lkup(h_parent, br, &dentry->d_name); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out; + + h_path.dentry = h_parent; + h_path.mnt = au_br_mnt(br); + au_dtime_store(&dt, parent, &h_path); + err = au_do_cpup_wh(cpg, wh_dentry, file); + if (unlikely(err)) + goto out_wh; + + dget(wh_dentry); + h_path.dentry = wh_dentry; + if (!d_is_dir(wh_dentry)) { + /* no delegation since it is just created */ + err = vfsub_unlink(d_inode(h_parent), &h_path, + /*delegated*/NULL, /*force*/0); + } else + err = vfsub_rmdir(d_inode(h_parent), &h_path); + if (unlikely(err)) { + AuIOErr("failed remove copied-up tmp file %pd(%d)\n", + wh_dentry, err); + err = -EIO; + } + au_dtime_revert(&dt); + au_set_hi_wh(d_inode(dentry), bdst, wh_dentry); + +out_wh: + dput(wh_dentry); +out: + dput(parent); + return err; +} + +struct au_cpup_wh_args { + int *errp; + struct au_cp_generic *cpg; + struct file *file; +}; + +static void au_call_cpup_wh(void *args) +{ + struct au_cpup_wh_args *a = args; + + au_pin_hdir_acquire_nest(a->cpg->pin); + *a->errp = au_cpup_wh(a->cpg, a->file); + au_pin_hdir_release(a->cpg->pin); +} + +int au_sio_cpup_wh(struct au_cp_generic *cpg, struct file *file) +{ + int err, wkq_err; + aufs_bindex_t bdst; + struct dentry *dentry, *parent, *h_orph, *h_parent; + struct inode *dir, *h_dir, *h_tmpdir; + struct au_wbr *wbr; + struct au_pin wh_pin, *pin_orig; + struct user_namespace *h_userns; + + dentry = cpg->dentry; + bdst = cpg->bdst; + parent = dget_parent(dentry); + dir = d_inode(parent); + h_orph = NULL; + h_parent = NULL; + h_dir = au_igrab(au_h_iptr(dir, bdst)); + h_tmpdir = h_dir; + pin_orig = NULL; + if (!h_dir->i_nlink) { + wbr = au_sbr(dentry->d_sb, bdst)->br_wbr; + h_orph = wbr->wbr_orph; + + h_parent = dget(au_h_dptr(parent, bdst)); + au_set_h_dptr(parent, bdst, dget(h_orph)); + h_tmpdir = d_inode(h_orph); + au_set_h_iptr(dir, bdst, au_igrab(h_tmpdir), /*flags*/0); + + inode_lock_nested(h_tmpdir, AuLsc_I_PARENT3); + /* todo: au_h_open_pre()? */ + + pin_orig = cpg->pin; + au_pin_init(&wh_pin, dentry, bdst, AuLsc_DI_PARENT, + AuLsc_I_PARENT3, cpg->pin->udba, AuPin_DI_LOCKED); + cpg->pin = &wh_pin; + } + + h_userns = au_sbr_userns(dentry->d_sb, bdst); + if (!au_test_h_perm_sio(h_userns, h_tmpdir, MAY_EXEC | MAY_WRITE) + && !au_cpup_sio_test(cpg->pin, d_inode(dentry)->i_mode)) + err = au_cpup_wh(cpg, file); + else { + struct au_cpup_wh_args args = { + .errp = &err, + .cpg = cpg, + .file = file + }; + wkq_err = au_wkq_wait(au_call_cpup_wh, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + if (h_orph) { + inode_unlock(h_tmpdir); + /* todo: au_h_open_post()? */ + au_set_h_iptr(dir, bdst, au_igrab(h_dir), /*flags*/0); + au_set_h_dptr(parent, bdst, h_parent); + AuDebugOn(!pin_orig); + cpg->pin = pin_orig; + } + iput(h_dir); + dput(parent); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * generic routine for both of copy-up and copy-down. + */ +/* cf. revalidate function in file.c */ +int au_cp_dirs(struct dentry *dentry, aufs_bindex_t bdst, + int (*cp)(struct dentry *dentry, aufs_bindex_t bdst, + struct au_pin *pin, + struct dentry *h_parent, void *arg), + void *arg) +{ + int err; + struct au_pin pin; + struct dentry *d, *parent, *h_parent, *real_parent, *h_dentry; + + err = 0; + parent = dget_parent(dentry); + if (IS_ROOT(parent)) + goto out; + + au_pin_init(&pin, dentry, bdst, AuLsc_DI_PARENT2, AuLsc_I_PARENT2, + au_opt_udba(dentry->d_sb), AuPin_MNT_WRITE); + + /* do not use au_dpage */ + real_parent = parent; + while (1) { + dput(parent); + parent = dget_parent(dentry); + h_parent = au_h_dptr(parent, bdst); + if (h_parent) + goto out; /* success */ + + /* find top dir which is necessary to cpup */ + do { + d = parent; + dput(parent); + parent = dget_parent(d); + di_read_lock_parent3(parent, !AuLock_IR); + h_parent = au_h_dptr(parent, bdst); + di_read_unlock(parent, !AuLock_IR); + } while (!h_parent); + + if (d != real_parent) + di_write_lock_child3(d); + + /* somebody else might create while we were sleeping */ + h_dentry = au_h_dptr(d, bdst); + if (!h_dentry || d_is_negative(h_dentry)) { + if (h_dentry) + au_update_dbtop(d); + + au_pin_set_dentry(&pin, d); + err = au_do_pin(&pin); + if (!err) { + err = cp(d, bdst, &pin, h_parent, arg); + au_unpin(&pin); + } + } + + if (d != real_parent) + di_write_unlock(d); + if (unlikely(err)) + break; + } + +out: + dput(parent); + return err; +} + +static int au_cpup_dir(struct dentry *dentry, aufs_bindex_t bdst, + struct au_pin *pin, + struct dentry *h_parent __maybe_unused, + void *arg __maybe_unused) +{ + struct au_cp_generic cpg = { + .dentry = dentry, + .bdst = bdst, + .bsrc = -1, + .len = 0, + .pin = pin, + .flags = AuCpup_DTIME + }; + return au_sio_cpup_simple(&cpg); +} + +int au_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst) +{ + return au_cp_dirs(dentry, bdst, au_cpup_dir, NULL); +} + +int au_test_and_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst) +{ + int err; + struct dentry *parent; + struct inode *dir; + + parent = dget_parent(dentry); + dir = d_inode(parent); + err = 0; + if (au_h_iptr(dir, bdst)) + goto out; + + di_read_unlock(parent, AuLock_IR); + di_write_lock_parent(parent); + /* someone else might change our inode while we were sleeping */ + if (!au_h_iptr(dir, bdst)) + err = au_cpup_dirs(dentry, bdst); + di_downgrade_lock(parent, AuLock_IR); + +out: + dput(parent); + return err; +} diff --git a/fs/aufs/cpup.h b/fs/aufs/cpup.h new file mode 100644 index 00000000000..d02f8150fa0 --- /dev/null +++ b/fs/aufs/cpup.h @@ -0,0 +1,100 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * copy-up/down functions + */ + +#ifndef __AUFS_CPUP_H__ +#define __AUFS_CPUP_H__ + +#ifdef __KERNEL__ + +#include + +struct inode; +struct file; +struct au_pin; + +void au_cpup_attr_flags(struct inode *dst, unsigned int iflags); +void au_cpup_attr_timesizes(struct inode *inode); +void au_cpup_attr_nlink(struct inode *inode, int force); +void au_cpup_attr_changeable(struct inode *inode); +void au_cpup_igen(struct inode *inode, struct inode *h_inode); +void au_cpup_attr_all(struct inode *inode, int force); + +/* ---------------------------------------------------------------------- */ + +struct au_cp_generic { + struct dentry *dentry; + aufs_bindex_t bdst, bsrc; + loff_t len; + struct au_pin *pin; + unsigned int flags; +}; + +/* cpup flags */ +#define AuCpup_DTIME 1 /* do dtime_store/revert */ +#define AuCpup_KEEPLINO (1 << 1) /* do not clear the lower xino, + for link(2) */ +#define AuCpup_RENAME (1 << 2) /* rename after cpup */ +#define AuCpup_HOPEN (1 << 3) /* call h_open_pre/post() in + cpup */ +#define AuCpup_OVERWRITE (1 << 4) /* allow overwriting the + existing entry */ +#define AuCpup_RWDST (1 << 5) /* force write target even if + the branch is marked as RO */ + +#ifndef CONFIG_AUFS_BR_HFSPLUS +#undef AuCpup_HOPEN +#define AuCpup_HOPEN 0 +#endif + +#define au_ftest_cpup(flags, name) ((flags) & AuCpup_##name) +#define au_fset_cpup(flags, name) \ + do { (flags) |= AuCpup_##name; } while (0) +#define au_fclr_cpup(flags, name) \ + do { (flags) &= ~AuCpup_##name; } while (0) + +int au_copy_file(struct file *dst, struct file *src, loff_t len); +int au_sio_cpup_simple(struct au_cp_generic *cpg); +int au_sio_cpdown_simple(struct au_cp_generic *cpg); +int au_sio_cpup_wh(struct au_cp_generic *cpg, struct file *file); + +int au_cp_dirs(struct dentry *dentry, aufs_bindex_t bdst, + int (*cp)(struct dentry *dentry, aufs_bindex_t bdst, + struct au_pin *pin, + struct dentry *h_parent, void *arg), + void *arg); +int au_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst); +int au_test_and_cpup_dirs(struct dentry *dentry, aufs_bindex_t bdst); + +/* ---------------------------------------------------------------------- */ + +/* keep timestamps when copyup */ +struct au_dtime { + struct dentry *dt_dentry; + struct path dt_h_path; + struct timespec64 dt_atime, dt_mtime; +}; +void au_dtime_store(struct au_dtime *dt, struct dentry *dentry, + struct path *h_path); +void au_dtime_revert(struct au_dtime *dt); + +#endif /* __KERNEL__ */ +#endif /* __AUFS_CPUP_H__ */ diff --git a/fs/aufs/dbgaufs.c b/fs/aufs/dbgaufs.c new file mode 100644 index 00000000000..80266f4fcf7 --- /dev/null +++ b/fs/aufs/dbgaufs.c @@ -0,0 +1,526 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * debugfs interface + */ + +#include +#include "aufs.h" + +#ifndef CONFIG_SYSFS +#error DEBUG_FS depends upon SYSFS +#endif + +static struct dentry *dbgaufs; +static const mode_t dbgaufs_mode = 0444; + +/* 20 is max digits length of ulong 64 */ +struct dbgaufs_arg { + int n; + char a[20 * 4]; +}; + +/* + * common function for all XINO files + */ +static int dbgaufs_xi_release(struct inode *inode __maybe_unused, + struct file *file) +{ + void *p; + + p = file->private_data; + if (p) { + /* this is struct dbgaufs_arg */ + AuDebugOn(!au_kfree_sz_test(p)); + au_kfree_do_rcu(p); + } + return 0; +} + +static int dbgaufs_xi_open(struct file *xf, struct file *file, int do_fcnt, + int cnt) +{ + int err; + struct kstat st; + struct dbgaufs_arg *p; + + err = -ENOMEM; + p = kmalloc(sizeof(*p), GFP_NOFS); + if (unlikely(!p)) + goto out; + + err = 0; + p->n = 0; + file->private_data = p; + if (!xf) + goto out; + + err = vfsub_getattr(&xf->f_path, &st); + if (!err) { + if (do_fcnt) + p->n = snprintf + (p->a, sizeof(p->a), "%d, %llux%u %lld\n", + cnt, st.blocks, st.blksize, + (long long)st.size); + else + p->n = snprintf(p->a, sizeof(p->a), "%llux%u %lld\n", + st.blocks, st.blksize, + (long long)st.size); + AuDebugOn(p->n >= sizeof(p->a)); + } else { + p->n = snprintf(p->a, sizeof(p->a), "err %d\n", err); + err = 0; + } + +out: + return err; +} + +static ssize_t dbgaufs_xi_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct dbgaufs_arg *p; + + p = file->private_data; + return simple_read_from_buffer(buf, count, ppos, p->a, p->n); +} + +/* ---------------------------------------------------------------------- */ + +struct dbgaufs_plink_arg { + int n; + char a[]; +}; + +static int dbgaufs_plink_release(struct inode *inode __maybe_unused, + struct file *file) +{ + free_page((unsigned long)file->private_data); + return 0; +} + +static int dbgaufs_plink_open(struct inode *inode, struct file *file) +{ + int err, i, limit; + unsigned long n, sum; + struct dbgaufs_plink_arg *p; + struct au_sbinfo *sbinfo; + struct super_block *sb; + struct hlist_bl_head *hbl; + + err = -ENOMEM; + p = (void *)get_zeroed_page(GFP_NOFS); + if (unlikely(!p)) + goto out; + + err = -EFBIG; + sbinfo = inode->i_private; + sb = sbinfo->si_sb; + si_noflush_read_lock(sb); + if (au_opt_test(au_mntflags(sb), PLINK)) { + limit = PAGE_SIZE - sizeof(p->n); + + /* the number of buckets */ + n = snprintf(p->a + p->n, limit, "%d\n", AuPlink_NHASH); + p->n += n; + limit -= n; + + sum = 0; + for (i = 0, hbl = sbinfo->si_plink; i < AuPlink_NHASH; + i++, hbl++) { + n = au_hbl_count(hbl); + sum += n; + + n = snprintf(p->a + p->n, limit, "%lu ", n); + p->n += n; + limit -= n; + if (unlikely(limit <= 0)) + goto out_free; + } + p->a[p->n - 1] = '\n'; + + /* the sum of plinks */ + n = snprintf(p->a + p->n, limit, "%lu\n", sum); + p->n += n; + limit -= n; + if (unlikely(limit <= 0)) + goto out_free; + } else { +#define str "1\n0\n0\n" + p->n = sizeof(str) - 1; + strcpy(p->a, str); +#undef str + } + si_read_unlock(sb); + + err = 0; + file->private_data = p; + goto out; /* success */ + +out_free: + free_page((unsigned long)p); +out: + return err; +} + +static ssize_t dbgaufs_plink_read(struct file *file, char __user *buf, + size_t count, loff_t *ppos) +{ + struct dbgaufs_plink_arg *p; + + p = file->private_data; + return simple_read_from_buffer(buf, count, ppos, p->a, p->n); +} + +static const struct file_operations dbgaufs_plink_fop = { + .owner = THIS_MODULE, + .open = dbgaufs_plink_open, + .release = dbgaufs_plink_release, + .read = dbgaufs_plink_read +}; + +/* ---------------------------------------------------------------------- */ + +static int dbgaufs_xib_open(struct inode *inode, struct file *file) +{ + int err; + struct au_sbinfo *sbinfo; + struct super_block *sb; + + sbinfo = inode->i_private; + sb = sbinfo->si_sb; + si_noflush_read_lock(sb); + err = dbgaufs_xi_open(sbinfo->si_xib, file, /*do_fcnt*/0, /*cnt*/0); + si_read_unlock(sb); + return err; +} + +static const struct file_operations dbgaufs_xib_fop = { + .owner = THIS_MODULE, + .open = dbgaufs_xib_open, + .release = dbgaufs_xi_release, + .read = dbgaufs_xi_read +}; + +/* ---------------------------------------------------------------------- */ + +#define DbgaufsXi_PREFIX "xi" + +static int dbgaufs_xino_open(struct inode *inode, struct file *file) +{ + int err, idx; + long l; + aufs_bindex_t bindex; + char *p, a[sizeof(DbgaufsXi_PREFIX) + 8]; + struct au_sbinfo *sbinfo; + struct super_block *sb; + struct au_xino *xi; + struct file *xf; + struct qstr *name; + struct au_branch *br; + + err = -ENOENT; + name = &file->f_path.dentry->d_name; + if (unlikely(name->len < sizeof(DbgaufsXi_PREFIX) + || memcmp(name->name, DbgaufsXi_PREFIX, + sizeof(DbgaufsXi_PREFIX) - 1))) + goto out; + + AuDebugOn(name->len >= sizeof(a)); + memcpy(a, name->name, name->len); + a[name->len] = '\0'; + p = strchr(a, '-'); + if (p) + *p = '\0'; + err = kstrtol(a + sizeof(DbgaufsXi_PREFIX) - 1, 10, &l); + if (unlikely(err)) + goto out; + bindex = l; + idx = 0; + if (p) { + err = kstrtol(p + 1, 10, &l); + if (unlikely(err)) + goto out; + idx = l; + } + + err = -ENOENT; + sbinfo = inode->i_private; + sb = sbinfo->si_sb; + si_noflush_read_lock(sb); + if (unlikely(bindex < 0 || bindex > au_sbbot(sb))) + goto out_si; + br = au_sbr(sb, bindex); + xi = br->br_xino; + if (unlikely(idx >= xi->xi_nfile)) + goto out_si; + xf = au_xino_file(xi, idx); + if (xf) + err = dbgaufs_xi_open(xf, file, /*do_fcnt*/1, + au_xino_count(br)); + +out_si: + si_read_unlock(sb); +out: + AuTraceErr(err); + return err; +} + +static const struct file_operations dbgaufs_xino_fop = { + .owner = THIS_MODULE, + .open = dbgaufs_xino_open, + .release = dbgaufs_xi_release, + .read = dbgaufs_xi_read +}; + +void dbgaufs_xino_del(struct au_branch *br) +{ + struct dentry *dbgaufs; + + dbgaufs = br->br_dbgaufs; + if (!dbgaufs) + return; + + br->br_dbgaufs = NULL; + /* debugfs acquires the parent i_mutex */ + lockdep_off(); + debugfs_remove(dbgaufs); + lockdep_on(); +} + +void dbgaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) +{ + aufs_bindex_t bbot; + struct au_branch *br; + + if (!au_sbi(sb)->si_dbgaufs) + return; + + bbot = au_sbbot(sb); + for (; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + dbgaufs_xino_del(br); + } +} + +static void dbgaufs_br_do_add(struct super_block *sb, aufs_bindex_t bindex, + unsigned int idx, struct dentry *parent, + struct au_sbinfo *sbinfo) +{ + struct au_branch *br; + struct dentry *d; + /* "xi" bindex(5) "-" idx(2) NULL */ + char name[sizeof(DbgaufsXi_PREFIX) + 8]; + + if (!idx) + snprintf(name, sizeof(name), DbgaufsXi_PREFIX "%d", bindex); + else + snprintf(name, sizeof(name), DbgaufsXi_PREFIX "%d-%u", + bindex, idx); + br = au_sbr(sb, bindex); + if (br->br_dbgaufs) { + struct qstr qstr = QSTR_INIT(name, strlen(name)); + + if (!au_qstreq(&br->br_dbgaufs->d_name, &qstr)) { + /* debugfs acquires the parent i_mutex */ + lockdep_off(); + d = debugfs_rename(parent, br->br_dbgaufs, parent, + name); + lockdep_on(); + if (unlikely(!d)) + pr_warn("failed renaming %pd/%s, ignored.\n", + parent, name); + } + } else { + lockdep_off(); + br->br_dbgaufs = debugfs_create_file(name, dbgaufs_mode, parent, + sbinfo, &dbgaufs_xino_fop); + lockdep_on(); + if (unlikely(!br->br_dbgaufs)) + pr_warn("failed creating %pd/%s, ignored.\n", + parent, name); + } +} + +static void dbgaufs_br_add(struct super_block *sb, aufs_bindex_t bindex, + struct dentry *parent, struct au_sbinfo *sbinfo) +{ + struct au_branch *br; + struct au_xino *xi; + unsigned int u; + + br = au_sbr(sb, bindex); + xi = br->br_xino; + for (u = 0; u < xi->xi_nfile; u++) + dbgaufs_br_do_add(sb, bindex, u, parent, sbinfo); +} + +void dbgaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex, int topdown) +{ + struct au_sbinfo *sbinfo; + struct dentry *parent; + aufs_bindex_t bbot; + + if (!au_opt_test(au_mntflags(sb), XINO)) + return; + + sbinfo = au_sbi(sb); + parent = sbinfo->si_dbgaufs; + if (!parent) + return; + + bbot = au_sbbot(sb); + if (topdown) + for (; bindex <= bbot; bindex++) + dbgaufs_br_add(sb, bindex, parent, sbinfo); + else + for (; bbot >= bindex; bbot--) + dbgaufs_br_add(sb, bbot, parent, sbinfo); +} + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_EXPORT +static int dbgaufs_xigen_open(struct inode *inode, struct file *file) +{ + int err; + struct au_sbinfo *sbinfo; + struct super_block *sb; + + sbinfo = inode->i_private; + sb = sbinfo->si_sb; + si_noflush_read_lock(sb); + err = dbgaufs_xi_open(sbinfo->si_xigen, file, /*do_fcnt*/0, /*cnt*/0); + si_read_unlock(sb); + return err; +} + +static const struct file_operations dbgaufs_xigen_fop = { + .owner = THIS_MODULE, + .open = dbgaufs_xigen_open, + .release = dbgaufs_xi_release, + .read = dbgaufs_xi_read +}; + +static int dbgaufs_xigen_init(struct au_sbinfo *sbinfo) +{ + int err; + + /* + * This function is a dynamic '__init' function actually, + * so the tiny check for si_rwsem is unnecessary. + */ + /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ + + err = -EIO; + sbinfo->si_dbgaufs_xigen = debugfs_create_file + ("xigen", dbgaufs_mode, sbinfo->si_dbgaufs, sbinfo, + &dbgaufs_xigen_fop); + if (sbinfo->si_dbgaufs_xigen) + err = 0; + + return err; +} +#else +static int dbgaufs_xigen_init(struct au_sbinfo *sbinfo) +{ + return 0; +} +#endif /* CONFIG_AUFS_EXPORT */ + +/* ---------------------------------------------------------------------- */ + +void dbgaufs_si_fin(struct au_sbinfo *sbinfo) +{ + /* + * This function is a dynamic '__fin' function actually, + * so the tiny check for si_rwsem is unnecessary. + */ + /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ + + debugfs_remove_recursive(sbinfo->si_dbgaufs); + sbinfo->si_dbgaufs = NULL; +} + +int dbgaufs_si_init(struct au_sbinfo *sbinfo) +{ + int err; + char name[SysaufsSiNameLen]; + + /* + * This function is a dynamic '__init' function actually, + * so the tiny check for si_rwsem is unnecessary. + */ + /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ + + err = -ENOENT; + if (!dbgaufs) { + AuErr1("/debug/aufs is uninitialized\n"); + goto out; + } + + err = -EIO; + sysaufs_name(sbinfo, name); + sbinfo->si_dbgaufs = debugfs_create_dir(name, dbgaufs); + if (unlikely(!sbinfo->si_dbgaufs)) + goto out; + + /* regardless plink/noplink option */ + sbinfo->si_dbgaufs_plink = debugfs_create_file + ("plink", dbgaufs_mode, sbinfo->si_dbgaufs, sbinfo, + &dbgaufs_plink_fop); + if (unlikely(!sbinfo->si_dbgaufs_plink)) + goto out_dir; + + /* regardless xino/noxino option */ + sbinfo->si_dbgaufs_xib = debugfs_create_file + ("xib", dbgaufs_mode, sbinfo->si_dbgaufs, sbinfo, + &dbgaufs_xib_fop); + if (unlikely(!sbinfo->si_dbgaufs_xib)) + goto out_dir; + + err = dbgaufs_xigen_init(sbinfo); + if (!err) + goto out; /* success */ + +out_dir: + dbgaufs_si_fin(sbinfo); +out: + if (unlikely(err)) + pr_err("debugfs/aufs failed\n"); + return err; +} + +/* ---------------------------------------------------------------------- */ + +void dbgaufs_fin(void) +{ + debugfs_remove(dbgaufs); +} + +int __init dbgaufs_init(void) +{ + int err; + + err = -EIO; + dbgaufs = debugfs_create_dir(AUFS_NAME, NULL); + if (dbgaufs) + err = 0; + return err; +} diff --git a/fs/aufs/dbgaufs.h b/fs/aufs/dbgaufs.h new file mode 100644 index 00000000000..7b4ccdebb67 --- /dev/null +++ b/fs/aufs/dbgaufs.h @@ -0,0 +1,53 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * debugfs interface + */ + +#ifndef __DBGAUFS_H__ +#define __DBGAUFS_H__ + +#ifdef __KERNEL__ + +struct super_block; +struct au_sbinfo; +struct au_branch; + +#ifdef CONFIG_DEBUG_FS +/* dbgaufs.c */ +void dbgaufs_xino_del(struct au_branch *br); +void dbgaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex); +void dbgaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex, int topdown); +void dbgaufs_si_fin(struct au_sbinfo *sbinfo); +int dbgaufs_si_init(struct au_sbinfo *sbinfo); +void dbgaufs_fin(void); +int __init dbgaufs_init(void); +#else +AuStubVoid(dbgaufs_xino_del, struct au_branch *br) +AuStubVoid(dbgaufs_brs_del, struct super_block *sb, aufs_bindex_t bindex) +AuStubVoid(dbgaufs_brs_add, struct super_block *sb, aufs_bindex_t bindex, + int topdown) +AuStubVoid(dbgaufs_si_fin, struct au_sbinfo *sbinfo) +AuStubInt0(dbgaufs_si_init, struct au_sbinfo *sbinfo) +AuStubVoid(dbgaufs_fin, void) +AuStubInt0(__init dbgaufs_init, void) +#endif /* CONFIG_DEBUG_FS */ + +#endif /* __KERNEL__ */ +#endif /* __DBGAUFS_H__ */ diff --git a/fs/aufs/dcsub.c b/fs/aufs/dcsub.c new file mode 100644 index 00000000000..0b9b1862b56 --- /dev/null +++ b/fs/aufs/dcsub.c @@ -0,0 +1,225 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sub-routines for dentry cache + */ + +#include "aufs.h" + +static void au_dpage_free(struct au_dpage *dpage) +{ + int i; + struct dentry **p; + + p = dpage->dentries; + for (i = 0; i < dpage->ndentry; i++) + dput(*p++); + free_page((unsigned long)dpage->dentries); +} + +int au_dpages_init(struct au_dcsub_pages *dpages, gfp_t gfp) +{ + int err; + void *p; + + err = -ENOMEM; + dpages->dpages = kmalloc(sizeof(*dpages->dpages), gfp); + if (unlikely(!dpages->dpages)) + goto out; + + p = (void *)__get_free_page(gfp); + if (unlikely(!p)) + goto out_dpages; + + dpages->dpages[0].ndentry = 0; + dpages->dpages[0].dentries = p; + dpages->ndpage = 1; + return 0; /* success */ + +out_dpages: + au_kfree_try_rcu(dpages->dpages); +out: + return err; +} + +void au_dpages_free(struct au_dcsub_pages *dpages) +{ + int i; + struct au_dpage *p; + + p = dpages->dpages; + for (i = 0; i < dpages->ndpage; i++) + au_dpage_free(p++); + au_kfree_try_rcu(dpages->dpages); +} + +static int au_dpages_append(struct au_dcsub_pages *dpages, + struct dentry *dentry, gfp_t gfp) +{ + int err, sz; + struct au_dpage *dpage; + void *p; + + dpage = dpages->dpages + dpages->ndpage - 1; + sz = PAGE_SIZE / sizeof(dentry); + if (unlikely(dpage->ndentry >= sz)) { + AuLabel(new dpage); + err = -ENOMEM; + sz = dpages->ndpage * sizeof(*dpages->dpages); + p = au_kzrealloc(dpages->dpages, sz, + sz + sizeof(*dpages->dpages), gfp, + /*may_shrink*/0); + if (unlikely(!p)) + goto out; + + dpages->dpages = p; + dpage = dpages->dpages + dpages->ndpage; + p = (void *)__get_free_page(gfp); + if (unlikely(!p)) + goto out; + + dpage->ndentry = 0; + dpage->dentries = p; + dpages->ndpage++; + } + + AuDebugOn(au_dcount(dentry) <= 0); + dpage->dentries[dpage->ndentry++] = dget_dlock(dentry); + return 0; /* success */ + +out: + return err; +} + +/* todo: BAD approach */ +/* copied from linux/fs/dcache.c */ +enum d_walk_ret { + D_WALK_CONTINUE, + D_WALK_QUIT, + D_WALK_NORETRY, + D_WALK_SKIP, +}; + +extern void d_walk(struct dentry *parent, void *data, + enum d_walk_ret (*enter)(void *, struct dentry *)); + +struct ac_dpages_arg { + int err; + struct au_dcsub_pages *dpages; + struct super_block *sb; + au_dpages_test test; + void *arg; +}; + +static enum d_walk_ret au_call_dpages_append(void *_arg, struct dentry *dentry) +{ + enum d_walk_ret ret; + struct ac_dpages_arg *arg = _arg; + + ret = D_WALK_CONTINUE; + if (dentry->d_sb == arg->sb + && !IS_ROOT(dentry) + && au_dcount(dentry) > 0 + && au_di(dentry) + && (!arg->test || arg->test(dentry, arg->arg))) { + arg->err = au_dpages_append(arg->dpages, dentry, GFP_ATOMIC); + if (unlikely(arg->err)) + ret = D_WALK_QUIT; + } + + return ret; +} + +int au_dcsub_pages(struct au_dcsub_pages *dpages, struct dentry *root, + au_dpages_test test, void *arg) +{ + struct ac_dpages_arg args = { + .err = 0, + .dpages = dpages, + .sb = root->d_sb, + .test = test, + .arg = arg + }; + + d_walk(root, &args, au_call_dpages_append); + + return args.err; +} + +int au_dcsub_pages_rev(struct au_dcsub_pages *dpages, struct dentry *dentry, + int do_include, au_dpages_test test, void *arg) +{ + int err; + + err = 0; + write_seqlock(&rename_lock); + spin_lock(&dentry->d_lock); + if (do_include + && au_dcount(dentry) > 0 + && (!test || test(dentry, arg))) + err = au_dpages_append(dpages, dentry, GFP_ATOMIC); + spin_unlock(&dentry->d_lock); + if (unlikely(err)) + goto out; + + /* + * RCU for vfsmount is unnecessary since this is a traverse in a single + * mount + */ + while (!IS_ROOT(dentry)) { + dentry = dentry->d_parent; /* rename_lock is locked */ + spin_lock(&dentry->d_lock); + if (au_dcount(dentry) > 0 + && (!test || test(dentry, arg))) + err = au_dpages_append(dpages, dentry, GFP_ATOMIC); + spin_unlock(&dentry->d_lock); + if (unlikely(err)) + break; + } + +out: + write_sequnlock(&rename_lock); + return err; +} + +static inline int au_dcsub_dpages_aufs(struct dentry *dentry, void *arg) +{ + return au_di(dentry) && dentry->d_sb == arg; +} + +int au_dcsub_pages_rev_aufs(struct au_dcsub_pages *dpages, + struct dentry *dentry, int do_include) +{ + return au_dcsub_pages_rev(dpages, dentry, do_include, + au_dcsub_dpages_aufs, dentry->d_sb); +} + +int au_test_subdir(struct dentry *d1, struct dentry *d2) +{ + struct path path[2] = { + { + .dentry = d1 + }, + { + .dentry = d2 + } + }; + + return path_is_under(path + 0, path + 1); +} diff --git a/fs/aufs/dcsub.h b/fs/aufs/dcsub.h new file mode 100644 index 00000000000..36f7fcdd7f2 --- /dev/null +++ b/fs/aufs/dcsub.h @@ -0,0 +1,137 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sub-routines for dentry cache + */ + +#ifndef __AUFS_DCSUB_H__ +#define __AUFS_DCSUB_H__ + +#ifdef __KERNEL__ + +#include +#include + +struct au_dpage { + int ndentry; + struct dentry **dentries; +}; + +struct au_dcsub_pages { + int ndpage; + struct au_dpage *dpages; +}; + +/* ---------------------------------------------------------------------- */ + +/* dcsub.c */ +int au_dpages_init(struct au_dcsub_pages *dpages, gfp_t gfp); +void au_dpages_free(struct au_dcsub_pages *dpages); +typedef int (*au_dpages_test)(struct dentry *dentry, void *arg); +int au_dcsub_pages(struct au_dcsub_pages *dpages, struct dentry *root, + au_dpages_test test, void *arg); +int au_dcsub_pages_rev(struct au_dcsub_pages *dpages, struct dentry *dentry, + int do_include, au_dpages_test test, void *arg); +int au_dcsub_pages_rev_aufs(struct au_dcsub_pages *dpages, + struct dentry *dentry, int do_include); +int au_test_subdir(struct dentry *d1, struct dentry *d2); + +/* ---------------------------------------------------------------------- */ + +/* + * todo: in linux-3.13, several similar (but faster) helpers are added to + * include/linux/dcache.h. Try them (in the future). + */ + +static inline int au_d_hashed_positive(struct dentry *d) +{ + int err; + struct inode *inode = d_inode(d); + + err = 0; + if (unlikely(d_unhashed(d) + || d_is_negative(d) + || !inode->i_nlink)) + err = -ENOENT; + return err; +} + +static inline int au_d_linkable(struct dentry *d) +{ + int err; + struct inode *inode = d_inode(d); + + err = au_d_hashed_positive(d); + if (err + && d_is_positive(d) + && (inode->i_state & I_LINKABLE)) + err = 0; + return err; +} + +static inline int au_d_alive(struct dentry *d) +{ + int err; + struct inode *inode; + + err = 0; + if (!IS_ROOT(d)) + err = au_d_hashed_positive(d); + else { + inode = d_inode(d); + if (unlikely(d_unlinked(d) + || d_is_negative(d) + || !inode->i_nlink)) + err = -ENOENT; + } + return err; +} + +static inline int au_alive_dir(struct dentry *d) +{ + int err; + + err = au_d_alive(d); + if (unlikely(err || IS_DEADDIR(d_inode(d)))) + err = -ENOENT; + return err; +} + +static inline int au_qstreq(struct qstr *a, struct qstr *b) +{ + return a->len == b->len + && !memcmp(a->name, b->name, a->len); +} + +/* + * by the commit + * 360f547 2015-01-25 dcache: let the dentry count go down to zero without + * taking d_lock + * the type of d_lockref.count became int, but the inlined function d_count() + * still returns unsigned int. + * I don't know why. Maybe it is for every d_count() users? + * Anyway au_dcount() lives on. + */ +static inline int au_dcount(struct dentry *d) +{ + return (int)d_count(d); +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DCSUB_H__ */ diff --git a/fs/aufs/debug.c b/fs/aufs/debug.c new file mode 100644 index 00000000000..f0c076c6125 --- /dev/null +++ b/fs/aufs/debug.c @@ -0,0 +1,441 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * debug print functions + */ + +#include +#include "aufs.h" + +/* Returns 0, or -errno. arg is in kp->arg. */ +static int param_atomic_t_set(const char *val, const struct kernel_param *kp) +{ + int err, n; + + err = kstrtoint(val, 0, &n); + if (!err) { + if (n > 0) + au_debug_on(); + else + au_debug_off(); + } + return err; +} + +/* Returns length written or -errno. Buffer is 4k (ie. be short!) */ +static int param_atomic_t_get(char *buffer, const struct kernel_param *kp) +{ + atomic_t *a; + + a = kp->arg; + return sprintf(buffer, "%d", atomic_read(a)); +} + +static struct kernel_param_ops param_ops_atomic_t = { + .set = param_atomic_t_set, + .get = param_atomic_t_get + /* void (*free)(void *arg) */ +}; + +atomic_t aufs_debug = ATOMIC_INIT(0); +MODULE_PARM_DESC(debug, "debug print"); +module_param_named(debug, aufs_debug, atomic_t, 0664); + +DEFINE_MUTEX(au_dbg_mtx); /* just to serialize the dbg msgs */ +char *au_plevel = KERN_DEBUG; +#define dpri(fmt, ...) do { \ + if ((au_plevel \ + && strcmp(au_plevel, KERN_DEBUG)) \ + || au_debug_test()) \ + printk("%s" fmt, au_plevel, ##__VA_ARGS__); \ +} while (0) + +/* ---------------------------------------------------------------------- */ + +void au_dpri_whlist(struct au_nhash *whlist) +{ + unsigned long ul, n; + struct hlist_head *head; + struct au_vdir_wh *pos; + + n = whlist->nh_num; + head = whlist->nh_head; + for (ul = 0; ul < n; ul++) { + hlist_for_each_entry(pos, head, wh_hash) + dpri("b%d, %.*s, %d\n", + pos->wh_bindex, + pos->wh_str.len, pos->wh_str.name, + pos->wh_str.len); + head++; + } +} + +void au_dpri_vdir(struct au_vdir *vdir) +{ + unsigned long ul; + union au_vdir_deblk_p p; + unsigned char *o; + + if (!vdir || IS_ERR(vdir)) { + dpri("err %ld\n", PTR_ERR(vdir)); + return; + } + + dpri("deblk %u, nblk %lu, deblk %p, last{%lu, %p}, ver %llu\n", + vdir->vd_deblk_sz, vdir->vd_nblk, vdir->vd_deblk, + vdir->vd_last.ul, vdir->vd_last.p.deblk, vdir->vd_version); + for (ul = 0; ul < vdir->vd_nblk; ul++) { + p.deblk = vdir->vd_deblk[ul]; + o = p.deblk; + dpri("[%lu]: %p\n", ul, o); + } +} + +static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode, int hn, + struct dentry *wh) +{ + char *n = NULL; + int l = 0; + + if (!inode || IS_ERR(inode)) { + dpri("i%d: err %ld\n", bindex, PTR_ERR(inode)); + return -1; + } + + /* the type of i_blocks depends upon CONFIG_LBDAF */ + BUILD_BUG_ON(sizeof(inode->i_blocks) != sizeof(unsigned long) + && sizeof(inode->i_blocks) != sizeof(u64)); + if (wh) { + n = (void *)wh->d_name.name; + l = wh->d_name.len; + } + + dpri("i%d: %p, i%lu, %s, cnt %d, nl %u, 0%o, sz %llu, blk %llu," + " hn %d, ct %lld, np %lu, st 0x%lx, f 0x%x, v %llu, g %x%s%.*s\n", + bindex, inode, + inode->i_ino, inode->i_sb ? au_sbtype(inode->i_sb) : "??", + atomic_read(&inode->i_count), inode->i_nlink, inode->i_mode, + i_size_read(inode), (unsigned long long)inode->i_blocks, + hn, (long long)timespec64_to_ns(&inode->i_ctime) & 0x0ffff, + inode->i_mapping ? inode->i_mapping->nrpages : 0, + inode->i_state, inode->i_flags, inode_peek_iversion(inode), + inode->i_generation, + l ? ", wh " : "", l, n); + return 0; +} + +void au_dpri_inode(struct inode *inode) +{ + struct au_iinfo *iinfo; + struct au_hinode *hi; + aufs_bindex_t bindex; + int err, hn; + + err = do_pri_inode(-1, inode, -1, NULL); + if (err || !au_test_aufs(inode->i_sb) || au_is_bad_inode(inode)) + return; + + iinfo = au_ii(inode); + dpri("i-1: btop %d, bbot %d, gen %d\n", + iinfo->ii_btop, iinfo->ii_bbot, au_iigen(inode, NULL)); + if (iinfo->ii_btop < 0) + return; + hn = 0; + for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot; bindex++) { + hi = au_hinode(iinfo, bindex); + hn = !!au_hn(hi); + do_pri_inode(bindex, hi->hi_inode, hn, hi->hi_whdentry); + } +} + +void au_dpri_dalias(struct inode *inode) +{ + struct dentry *d; + + spin_lock(&inode->i_lock); + hlist_for_each_entry(d, &inode->i_dentry, d_u.d_alias) + au_dpri_dentry(d); + spin_unlock(&inode->i_lock); +} + +static int do_pri_dentry(aufs_bindex_t bindex, struct dentry *dentry) +{ + struct dentry *wh = NULL; + int hn; + struct inode *inode; + struct au_iinfo *iinfo; + struct au_hinode *hi; + + if (!dentry || IS_ERR(dentry)) { + dpri("d%d: err %ld\n", bindex, PTR_ERR(dentry)); + return -1; + } + /* do not call dget_parent() here */ + /* note: access d_xxx without d_lock */ + dpri("d%d: %p, %pd2?, %s, cnt %d, flags 0x%x, %shashed\n", + bindex, dentry, dentry, + dentry->d_sb ? au_sbtype(dentry->d_sb) : "??", + au_dcount(dentry), dentry->d_flags, + d_unhashed(dentry) ? "un" : ""); + hn = -1; + inode = NULL; + if (d_is_positive(dentry)) + inode = d_inode(dentry); + if (inode + && au_test_aufs(dentry->d_sb) + && bindex >= 0 + && !au_is_bad_inode(inode)) { + iinfo = au_ii(inode); + hi = au_hinode(iinfo, bindex); + hn = !!au_hn(hi); + wh = hi->hi_whdentry; + } + do_pri_inode(bindex, inode, hn, wh); + return 0; +} + +void au_dpri_dentry(struct dentry *dentry) +{ + struct au_dinfo *dinfo; + aufs_bindex_t bindex; + int err; + + err = do_pri_dentry(-1, dentry); + if (err || !au_test_aufs(dentry->d_sb)) + return; + + dinfo = au_di(dentry); + if (!dinfo) + return; + dpri("d-1: btop %d, bbot %d, bwh %d, bdiropq %d, gen %d, tmp %d\n", + dinfo->di_btop, dinfo->di_bbot, + dinfo->di_bwh, dinfo->di_bdiropq, au_digen(dentry), + dinfo->di_tmpfile); + if (dinfo->di_btop < 0) + return; + for (bindex = dinfo->di_btop; bindex <= dinfo->di_bbot; bindex++) + do_pri_dentry(bindex, au_hdentry(dinfo, bindex)->hd_dentry); +} + +static int do_pri_file(aufs_bindex_t bindex, struct file *file) +{ + char a[32]; + + if (!file || IS_ERR(file)) { + dpri("f%d: err %ld\n", bindex, PTR_ERR(file)); + return -1; + } + a[0] = 0; + if (bindex < 0 + && !IS_ERR_OR_NULL(file->f_path.dentry) + && au_test_aufs(file->f_path.dentry->d_sb) + && au_fi(file)) + snprintf(a, sizeof(a), ", gen %d, mmapped %d", + au_figen(file), atomic_read(&au_fi(file)->fi_mmapped)); + dpri("f%d: mode 0x%x, flags 0%o, cnt %ld, v %llu, pos %llu%s\n", + bindex, file->f_mode, file->f_flags, (long)file_count(file), + file->f_version, file->f_pos, a); + if (!IS_ERR_OR_NULL(file->f_path.dentry)) + do_pri_dentry(bindex, file->f_path.dentry); + return 0; +} + +void au_dpri_file(struct file *file) +{ + struct au_finfo *finfo; + struct au_fidir *fidir; + struct au_hfile *hfile; + aufs_bindex_t bindex; + int err; + + err = do_pri_file(-1, file); + if (err + || IS_ERR_OR_NULL(file->f_path.dentry) + || !au_test_aufs(file->f_path.dentry->d_sb)) + return; + + finfo = au_fi(file); + if (!finfo) + return; + if (finfo->fi_btop < 0) + return; + fidir = finfo->fi_hdir; + if (!fidir) + do_pri_file(finfo->fi_btop, finfo->fi_htop.hf_file); + else + for (bindex = finfo->fi_btop; + bindex >= 0 && bindex <= fidir->fd_bbot; + bindex++) { + hfile = fidir->fd_hfile + bindex; + do_pri_file(bindex, hfile ? hfile->hf_file : NULL); + } +} + +static int do_pri_br(aufs_bindex_t bindex, struct au_branch *br) +{ + struct vfsmount *mnt; + struct super_block *sb; + + if (!br || IS_ERR(br)) + goto out; + mnt = au_br_mnt(br); + if (!mnt || IS_ERR(mnt)) + goto out; + sb = mnt->mnt_sb; + if (!sb || IS_ERR(sb)) + goto out; + + dpri("s%d: {perm 0x%x, id %d, wbr %p}, " + "%s, dev 0x%02x%02x, flags 0x%lx, cnt %d, active %d, " + "xino %d\n", + bindex, br->br_perm, br->br_id, br->br_wbr, + au_sbtype(sb), MAJOR(sb->s_dev), MINOR(sb->s_dev), + sb->s_flags, sb->s_count, + atomic_read(&sb->s_active), + !!au_xino_file(br->br_xino, /*idx*/-1)); + return 0; + +out: + dpri("s%d: err %ld\n", bindex, PTR_ERR(br)); + return -1; +} + +void au_dpri_sb(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + aufs_bindex_t bindex; + int err; + /* to reduce stack size */ + struct { + struct vfsmount mnt; + struct au_branch fake; + } *a; + + /* this function can be called from magic sysrq */ + a = kzalloc(sizeof(*a), GFP_ATOMIC); + if (unlikely(!a)) { + dpri("no memory\n"); + return; + } + + a->mnt.mnt_sb = sb; + a->fake.br_path.mnt = &a->mnt; + err = do_pri_br(-1, &a->fake); + au_kfree_rcu(a); + dpri("dev 0x%x\n", sb->s_dev); + if (err || !au_test_aufs(sb)) + return; + + sbinfo = au_sbi(sb); + if (!sbinfo) + return; + dpri("nw %d, gen %u, kobj %d\n", + atomic_read(&sbinfo->si_nowait.nw_len), sbinfo->si_generation, + kref_read(&sbinfo->si_kobj.kref)); + for (bindex = 0; bindex <= sbinfo->si_bbot; bindex++) + do_pri_br(bindex, sbinfo->si_branch[0 + bindex]); +} + +/* ---------------------------------------------------------------------- */ + +void __au_dbg_verify_dinode(struct dentry *dentry, const char *func, int line) +{ + struct inode *h_inode, *inode = d_inode(dentry); + struct dentry *h_dentry; + aufs_bindex_t bindex, bbot, bi; + + if (!inode /* || au_di(dentry)->di_lsc == AuLsc_DI_TMP */) + return; + + bbot = au_dbbot(dentry); + bi = au_ibbot(inode); + if (bi < bbot) + bbot = bi; + bindex = au_dbtop(dentry); + bi = au_ibtop(inode); + if (bi > bindex) + bindex = bi; + + for (; bindex <= bbot; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + h_inode = au_h_iptr(inode, bindex); + if (unlikely(h_inode != d_inode(h_dentry))) { + au_debug_on(); + AuDbg("b%d, %s:%d\n", bindex, func, line); + AuDbgDentry(dentry); + AuDbgInode(inode); + au_debug_off(); + BUG(); + } + } +} + +void au_dbg_verify_gen(struct dentry *parent, unsigned int sigen) +{ + int err, i, j; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry **dentries; + + err = au_dpages_init(&dpages, GFP_NOFS); + AuDebugOn(err); + err = au_dcsub_pages_rev_aufs(&dpages, parent, /*do_include*/1); + AuDebugOn(err); + for (i = dpages.ndpage - 1; !err && i >= 0; i--) { + dpage = dpages.dpages + i; + dentries = dpage->dentries; + for (j = dpage->ndentry - 1; !err && j >= 0; j--) + AuDebugOn(au_digen_test(dentries[j], sigen)); + } + au_dpages_free(&dpages); +} + +void au_dbg_verify_kthread(void) +{ + if (au_wkq_test()) { + au_dbg_blocked(); + /* + * It may be recursive, but udba=notify between two aufs mounts, + * where a single ro branch is shared, is not a problem. + */ + /* WARN_ON(1); */ + } +} + +/* ---------------------------------------------------------------------- */ + +int __init au_debug_init(void) +{ + aufs_bindex_t bindex; + struct au_vdir_destr destr; + + bindex = -1; + AuDebugOn(bindex >= 0); + + destr.len = -1; + AuDebugOn(destr.len < NAME_MAX); + +#ifdef CONFIG_4KSTACKS + pr_warn("CONFIG_4KSTACKS is defined.\n"); +#endif + + return 0; +} diff --git a/fs/aufs/debug.h b/fs/aufs/debug.h new file mode 100644 index 00000000000..7e46953513f --- /dev/null +++ b/fs/aufs/debug.h @@ -0,0 +1,226 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * debug print functions + */ + +#ifndef __AUFS_DEBUG_H__ +#define __AUFS_DEBUG_H__ + +#ifdef __KERNEL__ + +#include +#include +#include +#include + +#ifdef CONFIG_AUFS_DEBUG +#define AuDebugOn(a) BUG_ON(a) + +/* module parameter */ +extern atomic_t aufs_debug; +static inline void au_debug_on(void) +{ + atomic_inc(&aufs_debug); +} +static inline void au_debug_off(void) +{ + atomic_dec_if_positive(&aufs_debug); +} + +static inline int au_debug_test(void) +{ + return atomic_read(&aufs_debug) > 0; +} +#else +#define AuDebugOn(a) do {} while (0) +AuStubVoid(au_debug_on, void) +AuStubVoid(au_debug_off, void) +AuStubInt0(au_debug_test, void) +#endif /* CONFIG_AUFS_DEBUG */ + +#define param_check_atomic_t(name, p) __param_check(name, p, atomic_t) + +/* ---------------------------------------------------------------------- */ + +/* debug print */ + +#define AuDbg(fmt, ...) do { \ + if (au_debug_test()) \ + pr_debug("DEBUG: " fmt, ##__VA_ARGS__); \ +} while (0) +#define AuLabel(l) AuDbg(#l "\n") +#define AuIOErr(fmt, ...) pr_err("I/O Error, " fmt, ##__VA_ARGS__) +#define AuWarn1(fmt, ...) do { \ + static unsigned char _c; \ + if (!_c++) \ + pr_warn(fmt, ##__VA_ARGS__); \ +} while (0) + +#define AuErr1(fmt, ...) do { \ + static unsigned char _c; \ + if (!_c++) \ + pr_err(fmt, ##__VA_ARGS__); \ +} while (0) + +#define AuIOErr1(fmt, ...) do { \ + static unsigned char _c; \ + if (!_c++) \ + AuIOErr(fmt, ##__VA_ARGS__); \ +} while (0) + +#define AuUnsupportMsg "This operation is not supported." \ + " Please report this application to aufs-users ML." +#define AuUnsupport(fmt, ...) do { \ + pr_err(AuUnsupportMsg "\n" fmt, ##__VA_ARGS__); \ + dump_stack(); \ +} while (0) + +#define AuTraceErr(e) do { \ + if (unlikely((e) < 0)) \ + AuDbg("err %d\n", (int)(e)); \ +} while (0) + +#define AuTraceErrPtr(p) do { \ + if (IS_ERR(p)) \ + AuDbg("err %ld\n", PTR_ERR(p)); \ +} while (0) + +/* dirty macros for debug print, use with "%.*s" and caution */ +#define AuLNPair(qstr) (qstr)->len, (qstr)->name + +/* ---------------------------------------------------------------------- */ + +struct dentry; +#ifdef CONFIG_AUFS_DEBUG +extern struct mutex au_dbg_mtx; +extern char *au_plevel; +struct au_nhash; +void au_dpri_whlist(struct au_nhash *whlist); +struct au_vdir; +void au_dpri_vdir(struct au_vdir *vdir); +struct inode; +void au_dpri_inode(struct inode *inode); +void au_dpri_dalias(struct inode *inode); +void au_dpri_dentry(struct dentry *dentry); +struct file; +void au_dpri_file(struct file *filp); +struct super_block; +void au_dpri_sb(struct super_block *sb); + +#define au_dbg_verify_dinode(d) __au_dbg_verify_dinode(d, __func__, __LINE__) +void __au_dbg_verify_dinode(struct dentry *dentry, const char *func, int line); +void au_dbg_verify_gen(struct dentry *parent, unsigned int sigen); +void au_dbg_verify_kthread(void); + +int __init au_debug_init(void); + +#define AuDbgWhlist(w) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#w "\n"); \ + au_dpri_whlist(w); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgVdir(v) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#v "\n"); \ + au_dpri_vdir(v); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgInode(i) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#i "\n"); \ + au_dpri_inode(i); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgDAlias(i) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#i "\n"); \ + au_dpri_dalias(i); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgDentry(d) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#d "\n"); \ + au_dpri_dentry(d); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgFile(f) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#f "\n"); \ + au_dpri_file(f); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgSb(sb) do { \ + mutex_lock(&au_dbg_mtx); \ + AuDbg(#sb "\n"); \ + au_dpri_sb(sb); \ + mutex_unlock(&au_dbg_mtx); \ +} while (0) + +#define AuDbgSym(addr) do { \ + char sym[KSYM_SYMBOL_LEN]; \ + sprint_symbol(sym, (unsigned long)addr); \ + AuDbg("%s\n", sym); \ +} while (0) +#else +AuStubVoid(au_dbg_verify_dinode, struct dentry *dentry) +AuStubVoid(au_dbg_verify_gen, struct dentry *parent, unsigned int sigen) +AuStubVoid(au_dbg_verify_kthread, void) +AuStubInt0(__init au_debug_init, void) + +#define AuDbgWhlist(w) do {} while (0) +#define AuDbgVdir(v) do {} while (0) +#define AuDbgInode(i) do {} while (0) +#define AuDbgDAlias(i) do {} while (0) +#define AuDbgDentry(d) do {} while (0) +#define AuDbgFile(f) do {} while (0) +#define AuDbgSb(sb) do {} while (0) +#define AuDbgSym(addr) do {} while (0) +#endif /* CONFIG_AUFS_DEBUG */ + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_MAGIC_SYSRQ +int __init au_sysrq_init(void); +void au_sysrq_fin(void); + +#ifdef CONFIG_HW_CONSOLE +#define au_dbg_blocked() do { \ + WARN_ON(1); \ + handle_sysrq('w'); \ +} while (0) +#else +AuStubVoid(au_dbg_blocked, void) +#endif + +#else +AuStubInt0(__init au_sysrq_init, void) +AuStubVoid(au_sysrq_fin, void) +AuStubVoid(au_dbg_blocked, void) +#endif /* CONFIG_AUFS_MAGIC_SYSRQ */ + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DEBUG_H__ */ diff --git a/fs/aufs/dentry.c b/fs/aufs/dentry.c new file mode 100644 index 00000000000..16de5b2e5d0 --- /dev/null +++ b/fs/aufs/dentry.c @@ -0,0 +1,1160 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * lookup and dentry operations + */ + +#include +#include +#include "aufs.h" + +/* + * returns positive/negative dentry, NULL or an error. + * NULL means whiteout-ed or not-found. + */ +static struct dentry* +au_do_lookup(struct dentry *h_parent, struct dentry *dentry, + aufs_bindex_t bindex, struct au_do_lookup_args *args) +{ + struct dentry *h_dentry; + struct inode *h_inode; + struct au_branch *br; + struct user_namespace *h_userns; + int wh_found, opq; + unsigned char wh_able; + const unsigned char allow_neg = !!au_ftest_lkup(args->flags, ALLOW_NEG); + const unsigned char ignore_perm = !!au_ftest_lkup(args->flags, + IGNORE_PERM); + + wh_found = 0; + br = au_sbr(dentry->d_sb, bindex); + h_userns = au_br_userns(br); + wh_able = !!au_br_whable(br->br_perm); + if (wh_able) + wh_found = au_wh_test(h_userns, h_parent, &args->whname, + ignore_perm); + h_dentry = ERR_PTR(wh_found); + if (!wh_found) + goto real_lookup; + if (unlikely(wh_found < 0)) + goto out; + + /* We found a whiteout */ + /* au_set_dbbot(dentry, bindex); */ + au_set_dbwh(dentry, bindex); + if (!allow_neg) + return NULL; /* success */ + +real_lookup: + if (!ignore_perm) + h_dentry = vfsub_lkup_one(args->name, h_parent); + else + h_dentry = au_sio_lkup_one(h_userns, args->name, h_parent); + if (IS_ERR(h_dentry)) { + if (PTR_ERR(h_dentry) == -ENAMETOOLONG + && !allow_neg) + h_dentry = NULL; + goto out; + } + + h_inode = d_inode(h_dentry); + if (d_is_negative(h_dentry)) { + if (!allow_neg) + goto out_neg; + } else if (wh_found + || (args->type && args->type != (h_inode->i_mode & S_IFMT))) + goto out_neg; + else if (au_ftest_lkup(args->flags, DIRREN) + /* && h_inode */ + && !au_dr_lkup_h_ino(args, bindex, h_inode->i_ino)) { + AuDbg("b%d %pd ignored hi%llu\n", bindex, h_dentry, + (unsigned long long)h_inode->i_ino); + goto out_neg; + } + + if (au_dbbot(dentry) <= bindex) + au_set_dbbot(dentry, bindex); + if (au_dbtop(dentry) < 0 || bindex < au_dbtop(dentry)) + au_set_dbtop(dentry, bindex); + au_set_h_dptr(dentry, bindex, h_dentry); + + if (!d_is_dir(h_dentry) + || !wh_able + || (d_really_is_positive(dentry) && !d_is_dir(dentry))) + goto out; /* success */ + + inode_lock_shared_nested(h_inode, AuLsc_I_CHILD); + opq = au_diropq_test(h_userns, h_dentry); + inode_unlock_shared(h_inode); + if (opq > 0) + au_set_dbdiropq(dentry, bindex); + else if (unlikely(opq < 0)) { + au_set_h_dptr(dentry, bindex, NULL); + h_dentry = ERR_PTR(opq); + } + goto out; + +out_neg: + dput(h_dentry); + h_dentry = NULL; +out: + return h_dentry; +} + +static int au_test_shwh(struct super_block *sb, const struct qstr *name) +{ + if (unlikely(!au_opt_test(au_mntflags(sb), SHWH) + && !strncmp(name->name, AUFS_WH_PFX, AUFS_WH_PFX_LEN))) + return -EPERM; + return 0; +} + +/* + * returns the number of lower positive dentries, + * otherwise an error. + * can be called at unlinking with @type is zero. + */ +int au_lkup_dentry(struct dentry *dentry, aufs_bindex_t btop, + unsigned int flags) +{ + int npositive, err; + aufs_bindex_t bindex, btail, bdiropq; + unsigned char isdir, dirperm1, dirren; + struct au_do_lookup_args args = { + .flags = flags, + .name = &dentry->d_name + }; + struct dentry *parent; + struct super_block *sb; + + sb = dentry->d_sb; + err = au_test_shwh(sb, args.name); + if (unlikely(err)) + goto out; + + err = au_wh_name_alloc(&args.whname, args.name); + if (unlikely(err)) + goto out; + + isdir = !!d_is_dir(dentry); + dirperm1 = !!au_opt_test(au_mntflags(sb), DIRPERM1); + dirren = !!au_opt_test(au_mntflags(sb), DIRREN); + if (dirren) + au_fset_lkup(args.flags, DIRREN); + + npositive = 0; + parent = dget_parent(dentry); + btail = au_dbtaildir(parent); + for (bindex = btop; bindex <= btail; bindex++) { + struct dentry *h_parent, *h_dentry; + struct inode *h_inode, *h_dir; + struct au_branch *br; + + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry) { + if (d_is_positive(h_dentry)) + npositive++; + break; + } + h_parent = au_h_dptr(parent, bindex); + if (!h_parent || !d_is_dir(h_parent)) + continue; + + if (dirren) { + /* if the inum matches, then use the prepared name */ + err = au_dr_lkup_name(&args, bindex); + if (unlikely(err)) + goto out_parent; + } + + h_dir = d_inode(h_parent); + inode_lock_shared_nested(h_dir, AuLsc_I_PARENT); + h_dentry = au_do_lookup(h_parent, dentry, bindex, &args); + inode_unlock_shared(h_dir); + err = PTR_ERR(h_dentry); + if (IS_ERR(h_dentry)) + goto out_parent; + if (h_dentry) + au_fclr_lkup(args.flags, ALLOW_NEG); + if (dirperm1) + au_fset_lkup(args.flags, IGNORE_PERM); + + if (au_dbwh(dentry) == bindex) + break; + if (!h_dentry) + continue; + if (d_is_negative(h_dentry)) + continue; + h_inode = d_inode(h_dentry); + npositive++; + if (!args.type) + args.type = h_inode->i_mode & S_IFMT; + if (args.type != S_IFDIR) + break; + else if (isdir) { + /* the type of lower may be different */ + bdiropq = au_dbdiropq(dentry); + if (bdiropq >= 0 && bdiropq <= bindex) + break; + } + br = au_sbr(sb, bindex); + if (dirren + && au_dr_hino_test_add(&br->br_dirren, h_inode->i_ino, + /*add_ent*/NULL)) { + /* prepare next name to lookup */ + err = au_dr_lkup(&args, dentry, bindex); + if (unlikely(err)) + goto out_parent; + } + } + + if (npositive) { + AuLabel(positive); + au_update_dbtop(dentry); + } + err = npositive; + if (unlikely(!au_opt_test(au_mntflags(sb), UDBA_NONE) + && au_dbtop(dentry) < 0)) { + err = -EIO; + AuIOErr("both of real entry and whiteout found, %pd, err %d\n", + dentry, err); + } + +out_parent: + dput(parent); + au_kfree_try_rcu(args.whname.name); + if (dirren) + au_dr_lkup_fin(&args); +out: + return err; +} + +struct dentry *au_sio_lkup_one(struct user_namespace *userns, struct qstr *name, + struct dentry *parent) +{ + struct dentry *dentry; + int wkq_err; + + if (!au_test_h_perm_sio(userns, d_inode(parent), MAY_EXEC)) + dentry = vfsub_lkup_one(name, parent); + else { + struct vfsub_lkup_one_args args = { + .errp = &dentry, + .name = name, + .parent = parent + }; + + wkq_err = au_wkq_wait(vfsub_call_lkup_one, &args); + if (unlikely(wkq_err)) + dentry = ERR_PTR(wkq_err); + } + + return dentry; +} + +/* + * lookup @dentry on @bindex which should be negative. + */ +int au_lkup_neg(struct dentry *dentry, aufs_bindex_t bindex, int wh) +{ + int err; + struct dentry *parent, *h_parent, *h_dentry; + struct au_branch *br; + struct user_namespace *h_userns; + + parent = dget_parent(dentry); + h_parent = au_h_dptr(parent, bindex); + br = au_sbr(dentry->d_sb, bindex); + h_userns = au_br_userns(br); + if (wh) + h_dentry = au_whtmp_lkup(h_parent, br, &dentry->d_name); + else + h_dentry = au_sio_lkup_one(h_userns, &dentry->d_name, h_parent); + err = PTR_ERR(h_dentry); + if (IS_ERR(h_dentry)) + goto out; + if (unlikely(d_is_positive(h_dentry))) { + err = -EIO; + AuIOErr("%pd should be negative on b%d.\n", h_dentry, bindex); + dput(h_dentry); + goto out; + } + + err = 0; + if (bindex < au_dbtop(dentry)) + au_set_dbtop(dentry, bindex); + if (au_dbbot(dentry) < bindex) + au_set_dbbot(dentry, bindex); + au_set_h_dptr(dentry, bindex, h_dentry); + +out: + dput(parent); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* subset of struct inode */ +struct au_iattr { + unsigned long i_ino; + /* unsigned int i_nlink; */ + kuid_t i_uid; + kgid_t i_gid; + u64 i_version; +/* + loff_t i_size; + blkcnt_t i_blocks; +*/ + umode_t i_mode; +}; + +static void au_iattr_save(struct au_iattr *ia, struct inode *h_inode) +{ + ia->i_ino = h_inode->i_ino; + /* ia->i_nlink = h_inode->i_nlink; */ + ia->i_uid = h_inode->i_uid; + ia->i_gid = h_inode->i_gid; + ia->i_version = inode_query_iversion(h_inode); +/* + ia->i_size = h_inode->i_size; + ia->i_blocks = h_inode->i_blocks; +*/ + ia->i_mode = (h_inode->i_mode & S_IFMT); +} + +static int au_iattr_test(struct au_iattr *ia, struct inode *h_inode) +{ + return ia->i_ino != h_inode->i_ino + /* || ia->i_nlink != h_inode->i_nlink */ + || !uid_eq(ia->i_uid, h_inode->i_uid) + || !gid_eq(ia->i_gid, h_inode->i_gid) + || !inode_eq_iversion(h_inode, ia->i_version) +/* + || ia->i_size != h_inode->i_size + || ia->i_blocks != h_inode->i_blocks +*/ + || ia->i_mode != (h_inode->i_mode & S_IFMT); +} + +static int au_h_verify_dentry(struct dentry *h_dentry, struct dentry *h_parent, + struct au_branch *br) +{ + int err; + struct au_iattr ia; + struct inode *h_inode; + struct dentry *h_d; + struct super_block *h_sb; + + err = 0; + memset(&ia, -1, sizeof(ia)); + h_sb = h_dentry->d_sb; + h_inode = NULL; + if (d_is_positive(h_dentry)) { + h_inode = d_inode(h_dentry); + au_iattr_save(&ia, h_inode); + } else if (au_test_nfs(h_sb) || au_test_fuse(h_sb)) + /* nfs d_revalidate may return 0 for negative dentry */ + /* fuse d_revalidate always return 0 for negative dentry */ + goto out; + + /* main purpose is namei.c:cached_lookup() and d_revalidate */ + h_d = vfsub_lkup_one(&h_dentry->d_name, h_parent); + err = PTR_ERR(h_d); + if (IS_ERR(h_d)) + goto out; + + err = 0; + if (unlikely(h_d != h_dentry + || d_inode(h_d) != h_inode + || (h_inode && au_iattr_test(&ia, h_inode)))) + err = au_busy_or_stale(); + dput(h_d); + +out: + AuTraceErr(err); + return err; +} + +int au_h_verify(struct dentry *h_dentry, unsigned int udba, struct inode *h_dir, + struct dentry *h_parent, struct au_branch *br) +{ + int err; + + err = 0; + if (udba == AuOpt_UDBA_REVAL + && !au_test_fs_remote(h_dentry->d_sb)) { + IMustLock(h_dir); + err = (d_inode(h_dentry->d_parent) != h_dir); + } else if (udba != AuOpt_UDBA_NONE) + err = au_h_verify_dentry(h_dentry, h_parent, br); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_do_refresh_hdentry(struct dentry *dentry, struct dentry *parent) +{ + int err; + aufs_bindex_t new_bindex, bindex, bbot, bwh, bdiropq; + struct au_hdentry tmp, *p, *q; + struct au_dinfo *dinfo; + struct super_block *sb; + + DiMustWriteLock(dentry); + + sb = dentry->d_sb; + dinfo = au_di(dentry); + bbot = dinfo->di_bbot; + bwh = dinfo->di_bwh; + bdiropq = dinfo->di_bdiropq; + bindex = dinfo->di_btop; + p = au_hdentry(dinfo, bindex); + for (; bindex <= bbot; bindex++, p++) { + if (!p->hd_dentry) + continue; + + new_bindex = au_br_index(sb, p->hd_id); + if (new_bindex == bindex) + continue; + + if (dinfo->di_bwh == bindex) + bwh = new_bindex; + if (dinfo->di_bdiropq == bindex) + bdiropq = new_bindex; + if (new_bindex < 0) { + au_hdput(p); + p->hd_dentry = NULL; + continue; + } + + /* swap two lower dentries, and loop again */ + q = au_hdentry(dinfo, new_bindex); + tmp = *q; + *q = *p; + *p = tmp; + if (tmp.hd_dentry) { + bindex--; + p--; + } + } + + dinfo->di_bwh = -1; + if (bwh >= 0 && bwh <= au_sbbot(sb) && au_sbr_whable(sb, bwh)) + dinfo->di_bwh = bwh; + + dinfo->di_bdiropq = -1; + if (bdiropq >= 0 + && bdiropq <= au_sbbot(sb) + && au_sbr_whable(sb, bdiropq)) + dinfo->di_bdiropq = bdiropq; + + err = -EIO; + dinfo->di_btop = -1; + dinfo->di_bbot = -1; + bbot = au_dbbot(parent); + bindex = 0; + p = au_hdentry(dinfo, bindex); + for (; bindex <= bbot; bindex++, p++) + if (p->hd_dentry) { + dinfo->di_btop = bindex; + break; + } + + if (dinfo->di_btop >= 0) { + bindex = bbot; + p = au_hdentry(dinfo, bindex); + for (; bindex >= 0; bindex--, p--) + if (p->hd_dentry) { + dinfo->di_bbot = bindex; + err = 0; + break; + } + } + + return err; +} + +static void au_do_hide(struct dentry *dentry) +{ + struct inode *inode; + + if (d_really_is_positive(dentry)) { + inode = d_inode(dentry); + if (!d_is_dir(dentry)) { + if (inode->i_nlink && !d_unhashed(dentry)) + drop_nlink(inode); + } else { + clear_nlink(inode); + /* stop next lookup */ + inode->i_flags |= S_DEAD; + } + smp_mb(); /* necessary? */ + } + d_drop(dentry); +} + +static int au_hide_children(struct dentry *parent) +{ + int err, i, j, ndentry; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry *dentry; + + err = au_dpages_init(&dpages, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_dcsub_pages(&dpages, parent, NULL, NULL); + if (unlikely(err)) + goto out_dpages; + + /* in reverse order */ + for (i = dpages.ndpage - 1; i >= 0; i--) { + dpage = dpages.dpages + i; + ndentry = dpage->ndentry; + for (j = ndentry - 1; j >= 0; j--) { + dentry = dpage->dentries[j]; + if (dentry != parent) + au_do_hide(dentry); + } + } + +out_dpages: + au_dpages_free(&dpages); +out: + return err; +} + +static void au_hide(struct dentry *dentry) +{ + int err; + + AuDbgDentry(dentry); + if (d_is_dir(dentry)) { + /* shrink_dcache_parent(dentry); */ + err = au_hide_children(dentry); + if (unlikely(err)) + AuIOErr("%pd, failed hiding children, ignored %d\n", + dentry, err); + } + au_do_hide(dentry); +} + +/* + * By adding a dirty branch, a cached dentry may be affected in various ways. + * + * a dirty branch is added + * - on the top of layers + * - in the middle of layers + * - to the bottom of layers + * + * on the added branch there exists + * - a whiteout + * - a diropq + * - a same named entry + * + exist + * * negative --> positive + * * positive --> positive + * - type is unchanged + * - type is changed + * + doesn't exist + * * negative --> negative + * * positive --> negative (rejected by au_br_del() for non-dir case) + * - none + */ +static int au_refresh_by_dinfo(struct dentry *dentry, struct au_dinfo *dinfo, + struct au_dinfo *tmp) +{ + int err; + aufs_bindex_t bindex, bbot; + struct { + struct dentry *dentry; + struct inode *inode; + mode_t mode; + } orig_h, tmp_h = { + .dentry = NULL + }; + struct au_hdentry *hd; + struct inode *inode, *h_inode; + struct dentry *h_dentry; + + err = 0; + AuDebugOn(dinfo->di_btop < 0); + orig_h.mode = 0; + orig_h.dentry = au_hdentry(dinfo, dinfo->di_btop)->hd_dentry; + orig_h.inode = NULL; + if (d_is_positive(orig_h.dentry)) { + orig_h.inode = d_inode(orig_h.dentry); + orig_h.mode = orig_h.inode->i_mode & S_IFMT; + } + if (tmp->di_btop >= 0) { + tmp_h.dentry = au_hdentry(tmp, tmp->di_btop)->hd_dentry; + if (d_is_positive(tmp_h.dentry)) { + tmp_h.inode = d_inode(tmp_h.dentry); + tmp_h.mode = tmp_h.inode->i_mode & S_IFMT; + } + } + + inode = NULL; + if (d_really_is_positive(dentry)) + inode = d_inode(dentry); + if (!orig_h.inode) { + AuDbg("negative originally\n"); + if (inode) { + au_hide(dentry); + goto out; + } + AuDebugOn(inode); + AuDebugOn(dinfo->di_btop != dinfo->di_bbot); + AuDebugOn(dinfo->di_bdiropq != -1); + + if (!tmp_h.inode) { + AuDbg("negative --> negative\n"); + /* should have only one negative lower */ + if (tmp->di_btop >= 0 + && tmp->di_btop < dinfo->di_btop) { + AuDebugOn(tmp->di_btop != tmp->di_bbot); + AuDebugOn(dinfo->di_btop != dinfo->di_bbot); + au_set_h_dptr(dentry, dinfo->di_btop, NULL); + au_di_cp(dinfo, tmp); + hd = au_hdentry(tmp, tmp->di_btop); + au_set_h_dptr(dentry, tmp->di_btop, + dget(hd->hd_dentry)); + } + au_dbg_verify_dinode(dentry); + } else { + AuDbg("negative --> positive\n"); + /* + * similar to the behaviour of creating with bypassing + * aufs. + * unhash it in order to force an error in the + * succeeding create operation. + * we should not set S_DEAD here. + */ + d_drop(dentry); + /* au_di_swap(tmp, dinfo); */ + au_dbg_verify_dinode(dentry); + } + } else { + AuDbg("positive originally\n"); + /* inode may be NULL */ + AuDebugOn(inode && (inode->i_mode & S_IFMT) != orig_h.mode); + if (!tmp_h.inode) { + AuDbg("positive --> negative\n"); + /* or bypassing aufs */ + au_hide(dentry); + if (tmp->di_bwh >= 0 && tmp->di_bwh <= dinfo->di_btop) + dinfo->di_bwh = tmp->di_bwh; + if (inode) + err = au_refresh_hinode_self(inode); + au_dbg_verify_dinode(dentry); + } else if (orig_h.mode == tmp_h.mode) { + AuDbg("positive --> positive, same type\n"); + if (!S_ISDIR(orig_h.mode) + && dinfo->di_btop > tmp->di_btop) { + /* + * similar to the behaviour of removing and + * creating. + */ + au_hide(dentry); + if (inode) + err = au_refresh_hinode_self(inode); + au_dbg_verify_dinode(dentry); + } else { + /* fill empty slots */ + if (dinfo->di_btop > tmp->di_btop) + dinfo->di_btop = tmp->di_btop; + if (dinfo->di_bbot < tmp->di_bbot) + dinfo->di_bbot = tmp->di_bbot; + dinfo->di_bwh = tmp->di_bwh; + dinfo->di_bdiropq = tmp->di_bdiropq; + bbot = dinfo->di_bbot; + bindex = tmp->di_btop; + hd = au_hdentry(tmp, bindex); + for (; bindex <= bbot; bindex++, hd++) { + if (au_h_dptr(dentry, bindex)) + continue; + h_dentry = hd->hd_dentry; + if (!h_dentry) + continue; + AuDebugOn(d_is_negative(h_dentry)); + h_inode = d_inode(h_dentry); + AuDebugOn(orig_h.mode + != (h_inode->i_mode + & S_IFMT)); + au_set_h_dptr(dentry, bindex, + dget(h_dentry)); + } + if (inode) + err = au_refresh_hinode(inode, dentry); + au_dbg_verify_dinode(dentry); + } + } else { + AuDbg("positive --> positive, different type\n"); + /* similar to the behaviour of removing and creating */ + au_hide(dentry); + if (inode) + err = au_refresh_hinode_self(inode); + au_dbg_verify_dinode(dentry); + } + } + +out: + return err; +} + +void au_refresh_dop(struct dentry *dentry, int force_reval) +{ + const struct dentry_operations *dop + = force_reval ? &aufs_dop : dentry->d_sb->s_d_op; + static const unsigned int mask + = DCACHE_OP_REVALIDATE | DCACHE_OP_WEAK_REVALIDATE; + + BUILD_BUG_ON(sizeof(mask) != sizeof(dentry->d_flags)); + + if (dentry->d_op == dop) + return; + + AuDbg("%pd\n", dentry); + spin_lock(&dentry->d_lock); + if (dop == &aufs_dop) + dentry->d_flags |= mask; + else + dentry->d_flags &= ~mask; + dentry->d_op = dop; + spin_unlock(&dentry->d_lock); +} + +int au_refresh_dentry(struct dentry *dentry, struct dentry *parent) +{ + int err, ebrange, nbr; + unsigned int sigen; + struct au_dinfo *dinfo, *tmp; + struct super_block *sb; + struct inode *inode; + + DiMustWriteLock(dentry); + AuDebugOn(IS_ROOT(dentry)); + AuDebugOn(d_really_is_negative(parent)); + + sb = dentry->d_sb; + sigen = au_sigen(sb); + err = au_digen_test(parent, sigen); + if (unlikely(err)) + goto out; + + nbr = au_sbbot(sb) + 1; + dinfo = au_di(dentry); + err = au_di_realloc(dinfo, nbr, /*may_shrink*/0); + if (unlikely(err)) + goto out; + ebrange = au_dbrange_test(dentry); + if (!ebrange) + ebrange = au_do_refresh_hdentry(dentry, parent); + + if (d_unhashed(dentry) || ebrange /* || dinfo->di_tmpfile */) { + AuDebugOn(au_dbtop(dentry) < 0 && au_dbbot(dentry) >= 0); + if (d_really_is_positive(dentry)) { + inode = d_inode(dentry); + err = au_refresh_hinode_self(inode); + } + au_dbg_verify_dinode(dentry); + if (!err) + goto out_dgen; /* success */ + goto out; + } + + /* temporary dinfo */ + AuDbgDentry(dentry); + err = -ENOMEM; + tmp = au_di_alloc(sb, AuLsc_DI_TMP); + if (unlikely(!tmp)) + goto out; + au_di_swap(tmp, dinfo); + /* returns the number of positive dentries */ + /* + * if current working dir is removed, it returns an error. + * but the dentry is legal. + */ + err = au_lkup_dentry(dentry, /*btop*/0, AuLkup_ALLOW_NEG); + AuDbgDentry(dentry); + au_di_swap(tmp, dinfo); + if (err == -ENOENT) + err = 0; + if (err >= 0) { + /* compare/refresh by dinfo */ + AuDbgDentry(dentry); + err = au_refresh_by_dinfo(dentry, dinfo, tmp); + au_dbg_verify_dinode(dentry); + AuTraceErr(err); + } + au_di_realloc(dinfo, nbr, /*may_shrink*/1); /* harmless if err */ + au_rw_write_unlock(&tmp->di_rwsem); + au_di_free(tmp); + if (unlikely(err)) + goto out; + +out_dgen: + au_update_digen(dentry); +out: + if (unlikely(err && !(dentry->d_flags & DCACHE_NFSFS_RENAMED))) { + AuIOErr("failed refreshing %pd, %d\n", dentry, err); + AuDbgDentry(dentry); + } + AuTraceErr(err); + return err; +} + +static int au_do_h_d_reval(struct dentry *h_dentry, unsigned int flags, + struct dentry *dentry, aufs_bindex_t bindex) +{ + int err, valid; + + err = 0; + if (!(h_dentry->d_flags & DCACHE_OP_REVALIDATE)) + goto out; + + AuDbg("b%d\n", bindex); + /* + * gave up supporting LOOKUP_CREATE/OPEN for lower fs, + * due to whiteout and branch permission. + */ + flags &= ~(/*LOOKUP_PARENT |*/ LOOKUP_OPEN | LOOKUP_CREATE + | LOOKUP_FOLLOW | LOOKUP_EXCL); + /* it may return tri-state */ + valid = h_dentry->d_op->d_revalidate(h_dentry, flags); + + if (unlikely(valid < 0)) + err = valid; + else if (!valid) + err = -EINVAL; + +out: + AuTraceErr(err); + return err; +} + +/* todo: remove this */ +static int h_d_revalidate(struct dentry *dentry, struct inode *inode, + unsigned int flags, int do_udba, int dirren) +{ + int err; + umode_t mode, h_mode; + aufs_bindex_t bindex, btail, btop, ibs, ibe; + unsigned char plus, unhashed, is_root, h_plus, h_nfs, tmpfile; + struct inode *h_inode, *h_cached_inode; + struct dentry *h_dentry; + struct qstr *name, *h_name; + + err = 0; + plus = 0; + mode = 0; + ibs = -1; + ibe = -1; + unhashed = !!d_unhashed(dentry); + is_root = !!IS_ROOT(dentry); + name = &dentry->d_name; + tmpfile = au_di(dentry)->di_tmpfile; + + /* + * Theoretically, REVAL test should be unnecessary in case of + * {FS,I}NOTIFY. + * But {fs,i}notify doesn't fire some necessary events, + * IN_ATTRIB for atime/nlink/pageio + * Let's do REVAL test too. + */ + if (do_udba && inode) { + mode = (inode->i_mode & S_IFMT); + plus = (inode->i_nlink > 0); + ibs = au_ibtop(inode); + ibe = au_ibbot(inode); + } + + btop = au_dbtop(dentry); + btail = btop; + if (inode && S_ISDIR(inode->i_mode)) + btail = au_dbtaildir(dentry); + for (bindex = btop; bindex <= btail; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + + AuDbg("b%d, %pd\n", bindex, h_dentry); + h_nfs = !!au_test_nfs(h_dentry->d_sb); + spin_lock(&h_dentry->d_lock); + h_name = &h_dentry->d_name; + if (unlikely(do_udba + && !is_root + && ((!h_nfs + && (unhashed != !!d_unhashed(h_dentry) + || (!tmpfile && !dirren + && !au_qstreq(name, h_name)) + )) + || (h_nfs + && !(flags & LOOKUP_OPEN) + && (h_dentry->d_flags + & DCACHE_NFSFS_RENAMED))) + )) { + int h_unhashed; + + h_unhashed = d_unhashed(h_dentry); + spin_unlock(&h_dentry->d_lock); + AuDbg("unhash 0x%x 0x%x, %pd %pd\n", + unhashed, h_unhashed, dentry, h_dentry); + goto err; + } + spin_unlock(&h_dentry->d_lock); + + err = au_do_h_d_reval(h_dentry, flags, dentry, bindex); + if (unlikely(err)) + /* do not goto err, to keep the errno */ + break; + + /* todo: plink too? */ + if (!do_udba) + continue; + + /* UDBA tests */ + if (unlikely(!!inode != d_is_positive(h_dentry))) + goto err; + + h_inode = NULL; + if (d_is_positive(h_dentry)) + h_inode = d_inode(h_dentry); + h_plus = plus; + h_mode = mode; + h_cached_inode = h_inode; + if (h_inode) { + h_mode = (h_inode->i_mode & S_IFMT); + h_plus = (h_inode->i_nlink > 0); + } + if (inode && ibs <= bindex && bindex <= ibe) + h_cached_inode = au_h_iptr(inode, bindex); + + if (!h_nfs) { + if (unlikely(plus != h_plus && !tmpfile)) + goto err; + } else { + if (unlikely(!(h_dentry->d_flags & DCACHE_NFSFS_RENAMED) + && !is_root + && !IS_ROOT(h_dentry) + && unhashed != d_unhashed(h_dentry))) + goto err; + } + if (unlikely(mode != h_mode + || h_cached_inode != h_inode)) + goto err; + continue; + +err: + err = -EINVAL; + break; + } + + AuTraceErr(err); + return err; +} + +/* todo: consolidate with do_refresh() and au_reval_for_attr() */ +static int simple_reval_dpath(struct dentry *dentry, unsigned int sigen) +{ + int err; + struct dentry *parent; + + if (!au_digen_test(dentry, sigen)) + return 0; + + parent = dget_parent(dentry); + di_read_lock_parent(parent, AuLock_IR); + AuDebugOn(au_digen_test(parent, sigen)); + au_dbg_verify_gen(parent, sigen); + err = au_refresh_dentry(dentry, parent); + di_read_unlock(parent, AuLock_IR); + dput(parent); + AuTraceErr(err); + return err; +} + +int au_reval_dpath(struct dentry *dentry, unsigned int sigen) +{ + int err; + struct dentry *d, *parent; + + if (!au_ftest_si(au_sbi(dentry->d_sb), FAILED_REFRESH_DIR)) + return simple_reval_dpath(dentry, sigen); + + /* slow loop, keep it simple and stupid */ + /* cf: au_cpup_dirs() */ + err = 0; + parent = NULL; + while (au_digen_test(dentry, sigen)) { + d = dentry; + while (1) { + dput(parent); + parent = dget_parent(d); + if (!au_digen_test(parent, sigen)) + break; + d = parent; + } + + if (d != dentry) + di_write_lock_child2(d); + + /* someone might update our dentry while we were sleeping */ + if (au_digen_test(d, sigen)) { + /* + * todo: consolidate with simple_reval_dpath(), + * do_refresh() and au_reval_for_attr(). + */ + di_read_lock_parent(parent, AuLock_IR); + err = au_refresh_dentry(d, parent); + di_read_unlock(parent, AuLock_IR); + } + + if (d != dentry) + di_write_unlock(d); + dput(parent); + if (unlikely(err)) + break; + } + + return err; +} + +/* + * if valid returns 1, otherwise 0. + */ +static int aufs_d_revalidate(struct dentry *dentry, unsigned int flags) +{ + int valid, err; + unsigned int sigen; + unsigned char do_udba, dirren; + struct super_block *sb; + struct inode *inode; + + /* todo: support rcu-walk? */ + if (flags & LOOKUP_RCU) + return -ECHILD; + + valid = 0; + if (unlikely(!au_di(dentry))) + goto out; + + valid = 1; + sb = dentry->d_sb; + /* + * todo: very ugly + * i_mutex of parent dir may be held, + * but we should not return 'invalid' due to busy. + */ + err = aufs_read_lock(dentry, AuLock_FLUSH | AuLock_DW | AuLock_NOPLM); + if (unlikely(err)) { + valid = err; + AuTraceErr(err); + goto out; + } + inode = NULL; + if (d_really_is_positive(dentry)) + inode = d_inode(dentry); + if (unlikely(inode && au_is_bad_inode(inode))) { + err = -EINVAL; + AuTraceErr(err); + goto out_dgrade; + } + if (unlikely(au_dbrange_test(dentry))) { + err = -EINVAL; + AuTraceErr(err); + goto out_dgrade; + } + + sigen = au_sigen(sb); + if (au_digen_test(dentry, sigen)) { + AuDebugOn(IS_ROOT(dentry)); + err = au_reval_dpath(dentry, sigen); + if (unlikely(err)) { + AuTraceErr(err); + goto out_dgrade; + } + } + di_downgrade_lock(dentry, AuLock_IR); + + err = -EINVAL; + if (!(flags & (LOOKUP_OPEN | LOOKUP_EMPTY)) + && inode + && !(inode->i_state && I_LINKABLE) + && (IS_DEADDIR(inode) || !inode->i_nlink)) { + AuTraceErr(err); + goto out_inval; + } + + do_udba = !au_opt_test(au_mntflags(sb), UDBA_NONE); + if (do_udba && inode) { + aufs_bindex_t btop = au_ibtop(inode); + struct inode *h_inode; + + if (btop >= 0) { + h_inode = au_h_iptr(inode, btop); + if (h_inode && au_test_higen(inode, h_inode)) { + AuTraceErr(err); + goto out_inval; + } + } + } + + dirren = !!au_opt_test(au_mntflags(sb), DIRREN); + err = h_d_revalidate(dentry, inode, flags, do_udba, dirren); + if (unlikely(!err && do_udba && au_dbtop(dentry) < 0)) { + err = -EIO; + AuDbg("both of real entry and whiteout found, %p, err %d\n", + dentry, err); + } + goto out_inval; + +out_dgrade: + di_downgrade_lock(dentry, AuLock_IR); +out_inval: + aufs_read_unlock(dentry, AuLock_IR); + AuTraceErr(err); + valid = !err; +out: + if (!valid) { + AuDbg("%pd invalid, %d\n", dentry, valid); + d_drop(dentry); + } + return valid; +} + +static void aufs_d_release(struct dentry *dentry) +{ + if (au_di(dentry)) { + au_di_fin(dentry); + au_hn_di_reinit(dentry); + } +} + +const struct dentry_operations aufs_dop = { + .d_revalidate = aufs_d_revalidate, + .d_weak_revalidate = aufs_d_revalidate, + .d_release = aufs_d_release +}; + +/* aufs_dop without d_revalidate */ +const struct dentry_operations aufs_dop_noreval = { + .d_release = aufs_d_release +}; diff --git a/fs/aufs/dentry.h b/fs/aufs/dentry.h new file mode 100644 index 00000000000..7657906e4a3 --- /dev/null +++ b/fs/aufs/dentry.h @@ -0,0 +1,269 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * lookup and dentry operations + */ + +#ifndef __AUFS_DENTRY_H__ +#define __AUFS_DENTRY_H__ + +#ifdef __KERNEL__ + +#include +#include "dirren.h" +#include "rwsem.h" + +struct au_hdentry { + struct dentry *hd_dentry; + aufs_bindex_t hd_id; +}; + +struct au_dinfo { + atomic_t di_generation; + + struct au_rwsem di_rwsem; + aufs_bindex_t di_btop, di_bbot, di_bwh, di_bdiropq; + unsigned char di_tmpfile; /* to allow the different name */ + struct au_hdentry *di_hdentry; + struct rcu_head rcu; +} ____cacheline_aligned_in_smp; + +/* ---------------------------------------------------------------------- */ + +/* flags for au_lkup_dentry() */ +#define AuLkup_ALLOW_NEG 1 +#define AuLkup_IGNORE_PERM (1 << 1) +#define AuLkup_DIRREN (1 << 2) +#define au_ftest_lkup(flags, name) ((flags) & AuLkup_##name) +#define au_fset_lkup(flags, name) \ + do { (flags) |= AuLkup_##name; } while (0) +#define au_fclr_lkup(flags, name) \ + do { (flags) &= ~AuLkup_##name; } while (0) + +#ifndef CONFIG_AUFS_DIRREN +#undef AuLkup_DIRREN +#define AuLkup_DIRREN 0 +#endif + +struct au_do_lookup_args { + unsigned int flags; + mode_t type; + struct qstr whname, *name; + struct au_dr_lookup dirren; +}; + +/* ---------------------------------------------------------------------- */ + +/* dentry.c */ +extern const struct dentry_operations aufs_dop, aufs_dop_noreval; +struct au_branch; +struct dentry *au_sio_lkup_one(struct user_namespace *userns, struct qstr *name, + struct dentry *parent); +int au_h_verify(struct dentry *h_dentry, unsigned int udba, struct inode *h_dir, + struct dentry *h_parent, struct au_branch *br); + +int au_lkup_dentry(struct dentry *dentry, aufs_bindex_t btop, + unsigned int flags); +int au_lkup_neg(struct dentry *dentry, aufs_bindex_t bindex, int wh); +int au_refresh_dentry(struct dentry *dentry, struct dentry *parent); +int au_reval_dpath(struct dentry *dentry, unsigned int sigen); +void au_refresh_dop(struct dentry *dentry, int force_reval); + +/* dinfo.c */ +void au_di_init_once(void *_di); +struct au_dinfo *au_di_alloc(struct super_block *sb, unsigned int lsc); +void au_di_free(struct au_dinfo *dinfo); +void au_di_swap(struct au_dinfo *a, struct au_dinfo *b); +void au_di_cp(struct au_dinfo *dst, struct au_dinfo *src); +int au_di_init(struct dentry *dentry); +void au_di_fin(struct dentry *dentry); +int au_di_realloc(struct au_dinfo *dinfo, int nbr, int may_shrink); + +void di_read_lock(struct dentry *d, int flags, unsigned int lsc); +void di_read_unlock(struct dentry *d, int flags); +void di_downgrade_lock(struct dentry *d, int flags); +void di_write_lock(struct dentry *d, unsigned int lsc); +void di_write_unlock(struct dentry *d); +void di_write_lock2_child(struct dentry *d1, struct dentry *d2, int isdir); +void di_write_lock2_parent(struct dentry *d1, struct dentry *d2, int isdir); +void di_write_unlock2(struct dentry *d1, struct dentry *d2); + +struct dentry *au_h_dptr(struct dentry *dentry, aufs_bindex_t bindex); +struct dentry *au_h_d_alias(struct dentry *dentry, aufs_bindex_t bindex); +aufs_bindex_t au_dbtail(struct dentry *dentry); +aufs_bindex_t au_dbtaildir(struct dentry *dentry); + +void au_set_h_dptr(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_dentry); +int au_digen_test(struct dentry *dentry, unsigned int sigen); +int au_dbrange_test(struct dentry *dentry); +void au_update_digen(struct dentry *dentry); +void au_update_dbrange(struct dentry *dentry, int do_put_zero); +void au_update_dbtop(struct dentry *dentry); +void au_update_dbbot(struct dentry *dentry); +int au_find_dbindex(struct dentry *dentry, struct dentry *h_dentry); + +/* ---------------------------------------------------------------------- */ + +static inline struct au_dinfo *au_di(struct dentry *dentry) +{ + return dentry->d_fsdata; +} + +/* ---------------------------------------------------------------------- */ + +/* lock subclass for dinfo */ +enum { + AuLsc_DI_CHILD, /* child first */ + AuLsc_DI_CHILD2, /* rename(2), link(2), and cpup at hnotify */ + AuLsc_DI_CHILD3, /* copyup dirs */ + AuLsc_DI_PARENT, + AuLsc_DI_PARENT2, + AuLsc_DI_PARENT3, + AuLsc_DI_TMP /* temp for replacing dinfo */ +}; + +/* + * di_read_lock_child, di_write_lock_child, + * di_read_lock_child2, di_write_lock_child2, + * di_read_lock_child3, di_write_lock_child3, + * di_read_lock_parent, di_write_lock_parent, + * di_read_lock_parent2, di_write_lock_parent2, + * di_read_lock_parent3, di_write_lock_parent3, + */ +#define AuReadLockFunc(name, lsc) \ +static inline void di_read_lock_##name(struct dentry *d, int flags) \ +{ di_read_lock(d, flags, AuLsc_DI_##lsc); } + +#define AuWriteLockFunc(name, lsc) \ +static inline void di_write_lock_##name(struct dentry *d) \ +{ di_write_lock(d, AuLsc_DI_##lsc); } + +#define AuRWLockFuncs(name, lsc) \ + AuReadLockFunc(name, lsc) \ + AuWriteLockFunc(name, lsc) + +AuRWLockFuncs(child, CHILD); +AuRWLockFuncs(child2, CHILD2); +AuRWLockFuncs(child3, CHILD3); +AuRWLockFuncs(parent, PARENT); +AuRWLockFuncs(parent2, PARENT2); +AuRWLockFuncs(parent3, PARENT3); + +#undef AuReadLockFunc +#undef AuWriteLockFunc +#undef AuRWLockFuncs + +#define DiMustNoWaiters(d) AuRwMustNoWaiters(&au_di(d)->di_rwsem) +#define DiMustAnyLock(d) AuRwMustAnyLock(&au_di(d)->di_rwsem) +#define DiMustWriteLock(d) AuRwMustWriteLock(&au_di(d)->di_rwsem) + +/* ---------------------------------------------------------------------- */ + +/* todo: memory barrier? */ +static inline unsigned int au_digen(struct dentry *d) +{ + return atomic_read(&au_di(d)->di_generation); +} + +static inline void au_h_dentry_init(struct au_hdentry *hdentry) +{ + hdentry->hd_dentry = NULL; +} + +static inline struct au_hdentry *au_hdentry(struct au_dinfo *di, + aufs_bindex_t bindex) +{ + return di->di_hdentry + bindex; +} + +static inline void au_hdput(struct au_hdentry *hd) +{ + if (hd) + dput(hd->hd_dentry); +} + +static inline aufs_bindex_t au_dbtop(struct dentry *dentry) +{ + DiMustAnyLock(dentry); + return au_di(dentry)->di_btop; +} + +static inline aufs_bindex_t au_dbbot(struct dentry *dentry) +{ + DiMustAnyLock(dentry); + return au_di(dentry)->di_bbot; +} + +static inline aufs_bindex_t au_dbwh(struct dentry *dentry) +{ + DiMustAnyLock(dentry); + return au_di(dentry)->di_bwh; +} + +static inline aufs_bindex_t au_dbdiropq(struct dentry *dentry) +{ + DiMustAnyLock(dentry); + return au_di(dentry)->di_bdiropq; +} + +/* todo: hard/soft set? */ +static inline void au_set_dbtop(struct dentry *dentry, aufs_bindex_t bindex) +{ + DiMustWriteLock(dentry); + au_di(dentry)->di_btop = bindex; +} + +static inline void au_set_dbbot(struct dentry *dentry, aufs_bindex_t bindex) +{ + DiMustWriteLock(dentry); + au_di(dentry)->di_bbot = bindex; +} + +static inline void au_set_dbwh(struct dentry *dentry, aufs_bindex_t bindex) +{ + DiMustWriteLock(dentry); + /* dbwh can be outside of btop - bbot range */ + au_di(dentry)->di_bwh = bindex; +} + +static inline void au_set_dbdiropq(struct dentry *dentry, aufs_bindex_t bindex) +{ + DiMustWriteLock(dentry); + au_di(dentry)->di_bdiropq = bindex; +} + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_HNOTIFY +static inline void au_digen_dec(struct dentry *d) +{ + atomic_dec(&au_di(d)->di_generation); +} + +static inline void au_hn_di_reinit(struct dentry *dentry) +{ + dentry->d_fsdata = NULL; +} +#else +AuStubVoid(au_hn_di_reinit, struct dentry *dentry __maybe_unused) +#endif /* CONFIG_AUFS_HNOTIFY */ + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DENTRY_H__ */ diff --git a/fs/aufs/dinfo.c b/fs/aufs/dinfo.c new file mode 100644 index 00000000000..be959106d98 --- /dev/null +++ b/fs/aufs/dinfo.c @@ -0,0 +1,554 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * dentry private data + */ + +#include "aufs.h" + +void au_di_init_once(void *_dinfo) +{ + struct au_dinfo *dinfo = _dinfo; + + au_rw_init(&dinfo->di_rwsem); +} + +struct au_dinfo *au_di_alloc(struct super_block *sb, unsigned int lsc) +{ + struct au_dinfo *dinfo; + int nbr, i; + + dinfo = au_cache_alloc_dinfo(); + if (unlikely(!dinfo)) + goto out; + + nbr = au_sbbot(sb) + 1; + if (nbr <= 0) + nbr = 1; + dinfo->di_hdentry = kcalloc(nbr, sizeof(*dinfo->di_hdentry), GFP_NOFS); + if (dinfo->di_hdentry) { + au_rw_write_lock_nested(&dinfo->di_rwsem, lsc); + dinfo->di_btop = -1; + dinfo->di_bbot = -1; + dinfo->di_bwh = -1; + dinfo->di_bdiropq = -1; + dinfo->di_tmpfile = 0; + for (i = 0; i < nbr; i++) + dinfo->di_hdentry[i].hd_id = -1; + goto out; + } + + au_cache_free_dinfo(dinfo); + dinfo = NULL; + +out: + return dinfo; +} + +void au_di_free(struct au_dinfo *dinfo) +{ + struct au_hdentry *p; + aufs_bindex_t bbot, bindex; + + /* dentry may not be revalidated */ + bindex = dinfo->di_btop; + if (bindex >= 0) { + bbot = dinfo->di_bbot; + p = au_hdentry(dinfo, bindex); + while (bindex++ <= bbot) + au_hdput(p++); + } + au_kfree_try_rcu(dinfo->di_hdentry); + au_cache_free_dinfo(dinfo); +} + +void au_di_swap(struct au_dinfo *a, struct au_dinfo *b) +{ + struct au_hdentry *p; + aufs_bindex_t bi; + + AuRwMustWriteLock(&a->di_rwsem); + AuRwMustWriteLock(&b->di_rwsem); + +#define DiSwap(v, name) \ + do { \ + v = a->di_##name; \ + a->di_##name = b->di_##name; \ + b->di_##name = v; \ + } while (0) + + DiSwap(p, hdentry); + DiSwap(bi, btop); + DiSwap(bi, bbot); + DiSwap(bi, bwh); + DiSwap(bi, bdiropq); + /* smp_mb(); */ + +#undef DiSwap +} + +void au_di_cp(struct au_dinfo *dst, struct au_dinfo *src) +{ + AuRwMustWriteLock(&dst->di_rwsem); + AuRwMustWriteLock(&src->di_rwsem); + + dst->di_btop = src->di_btop; + dst->di_bbot = src->di_bbot; + dst->di_bwh = src->di_bwh; + dst->di_bdiropq = src->di_bdiropq; + /* smp_mb(); */ +} + +int au_di_init(struct dentry *dentry) +{ + int err; + struct super_block *sb; + struct au_dinfo *dinfo; + + err = 0; + sb = dentry->d_sb; + dinfo = au_di_alloc(sb, AuLsc_DI_CHILD); + if (dinfo) { + atomic_set(&dinfo->di_generation, au_sigen(sb)); + /* smp_mb(); */ /* atomic_set */ + dentry->d_fsdata = dinfo; + } else + err = -ENOMEM; + + return err; +} + +void au_di_fin(struct dentry *dentry) +{ + struct au_dinfo *dinfo; + + dinfo = au_di(dentry); + AuRwDestroy(&dinfo->di_rwsem); + au_di_free(dinfo); +} + +int au_di_realloc(struct au_dinfo *dinfo, int nbr, int may_shrink) +{ + int err, sz; + struct au_hdentry *hdp; + + AuRwMustWriteLock(&dinfo->di_rwsem); + + err = -ENOMEM; + sz = sizeof(*hdp) * (dinfo->di_bbot + 1); + if (!sz) + sz = sizeof(*hdp); + hdp = au_kzrealloc(dinfo->di_hdentry, sz, sizeof(*hdp) * nbr, GFP_NOFS, + may_shrink); + if (hdp) { + dinfo->di_hdentry = hdp; + err = 0; + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static void do_ii_write_lock(struct inode *inode, unsigned int lsc) +{ + switch (lsc) { + case AuLsc_DI_CHILD: + ii_write_lock_child(inode); + break; + case AuLsc_DI_CHILD2: + ii_write_lock_child2(inode); + break; + case AuLsc_DI_CHILD3: + ii_write_lock_child3(inode); + break; + case AuLsc_DI_PARENT: + ii_write_lock_parent(inode); + break; + case AuLsc_DI_PARENT2: + ii_write_lock_parent2(inode); + break; + case AuLsc_DI_PARENT3: + ii_write_lock_parent3(inode); + break; + default: + BUG(); + } +} + +static void do_ii_read_lock(struct inode *inode, unsigned int lsc) +{ + switch (lsc) { + case AuLsc_DI_CHILD: + ii_read_lock_child(inode); + break; + case AuLsc_DI_CHILD2: + ii_read_lock_child2(inode); + break; + case AuLsc_DI_CHILD3: + ii_read_lock_child3(inode); + break; + case AuLsc_DI_PARENT: + ii_read_lock_parent(inode); + break; + case AuLsc_DI_PARENT2: + ii_read_lock_parent2(inode); + break; + case AuLsc_DI_PARENT3: + ii_read_lock_parent3(inode); + break; + default: + BUG(); + } +} + +void di_read_lock(struct dentry *d, int flags, unsigned int lsc) +{ + struct inode *inode; + + au_rw_read_lock_nested(&au_di(d)->di_rwsem, lsc); + if (d_really_is_positive(d)) { + inode = d_inode(d); + if (au_ftest_lock(flags, IW)) + do_ii_write_lock(inode, lsc); + else if (au_ftest_lock(flags, IR)) + do_ii_read_lock(inode, lsc); + } +} + +void di_read_unlock(struct dentry *d, int flags) +{ + struct inode *inode; + + if (d_really_is_positive(d)) { + inode = d_inode(d); + if (au_ftest_lock(flags, IW)) { + au_dbg_verify_dinode(d); + ii_write_unlock(inode); + } else if (au_ftest_lock(flags, IR)) { + au_dbg_verify_dinode(d); + ii_read_unlock(inode); + } + } + au_rw_read_unlock(&au_di(d)->di_rwsem); +} + +void di_downgrade_lock(struct dentry *d, int flags) +{ + if (d_really_is_positive(d) && au_ftest_lock(flags, IR)) + ii_downgrade_lock(d_inode(d)); + au_rw_dgrade_lock(&au_di(d)->di_rwsem); +} + +void di_write_lock(struct dentry *d, unsigned int lsc) +{ + au_rw_write_lock_nested(&au_di(d)->di_rwsem, lsc); + if (d_really_is_positive(d)) + do_ii_write_lock(d_inode(d), lsc); +} + +void di_write_unlock(struct dentry *d) +{ + au_dbg_verify_dinode(d); + if (d_really_is_positive(d)) + ii_write_unlock(d_inode(d)); + au_rw_write_unlock(&au_di(d)->di_rwsem); +} + +void di_write_lock2_child(struct dentry *d1, struct dentry *d2, int isdir) +{ + AuDebugOn(d1 == d2 + || d_inode(d1) == d_inode(d2) + || d1->d_sb != d2->d_sb); + + if ((isdir && au_test_subdir(d1, d2)) + || d1 < d2) { + di_write_lock_child(d1); + di_write_lock_child2(d2); + } else { + di_write_lock_child(d2); + di_write_lock_child2(d1); + } +} + +void di_write_lock2_parent(struct dentry *d1, struct dentry *d2, int isdir) +{ + AuDebugOn(d1 == d2 + || d_inode(d1) == d_inode(d2) + || d1->d_sb != d2->d_sb); + + if ((isdir && au_test_subdir(d1, d2)) + || d1 < d2) { + di_write_lock_parent(d1); + di_write_lock_parent2(d2); + } else { + di_write_lock_parent(d2); + di_write_lock_parent2(d1); + } +} + +void di_write_unlock2(struct dentry *d1, struct dentry *d2) +{ + di_write_unlock(d1); + if (d_inode(d1) == d_inode(d2)) + au_rw_write_unlock(&au_di(d2)->di_rwsem); + else + di_write_unlock(d2); +} + +/* ---------------------------------------------------------------------- */ + +struct dentry *au_h_dptr(struct dentry *dentry, aufs_bindex_t bindex) +{ + struct dentry *d; + + DiMustAnyLock(dentry); + + if (au_dbtop(dentry) < 0 || bindex < au_dbtop(dentry)) + return NULL; + AuDebugOn(bindex < 0); + d = au_hdentry(au_di(dentry), bindex)->hd_dentry; + AuDebugOn(d && au_dcount(d) <= 0); + return d; +} + +/* + * extended version of au_h_dptr(). + * returns a hashed and positive (or linkable) h_dentry in bindex, NULL, or + * error. + */ +struct dentry *au_h_d_alias(struct dentry *dentry, aufs_bindex_t bindex) +{ + struct dentry *h_dentry; + struct inode *inode, *h_inode; + + AuDebugOn(d_really_is_negative(dentry)); + + h_dentry = NULL; + if (au_dbtop(dentry) <= bindex + && bindex <= au_dbbot(dentry)) + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry && !au_d_linkable(h_dentry)) { + dget(h_dentry); + goto out; /* success */ + } + + inode = d_inode(dentry); + AuDebugOn(bindex < au_ibtop(inode)); + AuDebugOn(au_ibbot(inode) < bindex); + h_inode = au_h_iptr(inode, bindex); + h_dentry = d_find_alias(h_inode); + if (h_dentry) { + if (!IS_ERR(h_dentry)) { + if (!au_d_linkable(h_dentry)) + goto out; /* success */ + dput(h_dentry); + } else + goto out; + } + + if (au_opt_test(au_mntflags(dentry->d_sb), PLINK)) { + h_dentry = au_plink_lkup(inode, bindex); + AuDebugOn(!h_dentry); + if (!IS_ERR(h_dentry)) { + if (!au_d_hashed_positive(h_dentry)) + goto out; /* success */ + dput(h_dentry); + h_dentry = NULL; + } + } + +out: + AuDbgDentry(h_dentry); + return h_dentry; +} + +aufs_bindex_t au_dbtail(struct dentry *dentry) +{ + aufs_bindex_t bbot, bwh; + + bbot = au_dbbot(dentry); + if (0 <= bbot) { + bwh = au_dbwh(dentry); + if (!bwh) + return bwh; + if (0 < bwh && bwh < bbot) + return bwh - 1; + } + return bbot; +} + +aufs_bindex_t au_dbtaildir(struct dentry *dentry) +{ + aufs_bindex_t bbot, bopq; + + bbot = au_dbtail(dentry); + if (0 <= bbot) { + bopq = au_dbdiropq(dentry); + if (0 <= bopq && bopq < bbot) + bbot = bopq; + } + return bbot; +} + +/* ---------------------------------------------------------------------- */ + +void au_set_h_dptr(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_dentry) +{ + struct au_dinfo *dinfo; + struct au_hdentry *hd; + struct au_branch *br; + + DiMustWriteLock(dentry); + + dinfo = au_di(dentry); + hd = au_hdentry(dinfo, bindex); + au_hdput(hd); + hd->hd_dentry = h_dentry; + if (h_dentry) { + br = au_sbr(dentry->d_sb, bindex); + hd->hd_id = br->br_id; + } +} + +int au_dbrange_test(struct dentry *dentry) +{ + int err; + aufs_bindex_t btop, bbot; + + err = 0; + btop = au_dbtop(dentry); + bbot = au_dbbot(dentry); + if (btop >= 0) + AuDebugOn(bbot < 0 && btop > bbot); + else { + err = -EIO; + AuDebugOn(bbot >= 0); + } + + return err; +} + +int au_digen_test(struct dentry *dentry, unsigned int sigen) +{ + int err; + + err = 0; + if (unlikely(au_digen(dentry) != sigen + || au_iigen_test(d_inode(dentry), sigen))) + err = -EIO; + + return err; +} + +void au_update_digen(struct dentry *dentry) +{ + atomic_set(&au_di(dentry)->di_generation, au_sigen(dentry->d_sb)); + /* smp_mb(); */ /* atomic_set */ +} + +void au_update_dbrange(struct dentry *dentry, int do_put_zero) +{ + struct au_dinfo *dinfo; + struct dentry *h_d; + struct au_hdentry *hdp; + aufs_bindex_t bindex, bbot; + + DiMustWriteLock(dentry); + + dinfo = au_di(dentry); + if (!dinfo || dinfo->di_btop < 0) + return; + + if (do_put_zero) { + bbot = dinfo->di_bbot; + bindex = dinfo->di_btop; + hdp = au_hdentry(dinfo, bindex); + for (; bindex <= bbot; bindex++, hdp++) { + h_d = hdp->hd_dentry; + if (h_d && d_is_negative(h_d)) + au_set_h_dptr(dentry, bindex, NULL); + } + } + + dinfo->di_btop = 0; + hdp = au_hdentry(dinfo, dinfo->di_btop); + for (; dinfo->di_btop <= dinfo->di_bbot; dinfo->di_btop++, hdp++) + if (hdp->hd_dentry) + break; + if (dinfo->di_btop > dinfo->di_bbot) { + dinfo->di_btop = -1; + dinfo->di_bbot = -1; + return; + } + + hdp = au_hdentry(dinfo, dinfo->di_bbot); + for (; dinfo->di_bbot >= 0; dinfo->di_bbot--, hdp--) + if (hdp->hd_dentry) + break; + AuDebugOn(dinfo->di_btop > dinfo->di_bbot || dinfo->di_bbot < 0); +} + +void au_update_dbtop(struct dentry *dentry) +{ + aufs_bindex_t bindex, bbot; + struct dentry *h_dentry; + + bbot = au_dbbot(dentry); + for (bindex = au_dbtop(dentry); bindex <= bbot; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + if (d_is_positive(h_dentry)) { + au_set_dbtop(dentry, bindex); + return; + } + au_set_h_dptr(dentry, bindex, NULL); + } +} + +void au_update_dbbot(struct dentry *dentry) +{ + aufs_bindex_t bindex, btop; + struct dentry *h_dentry; + + btop = au_dbtop(dentry); + for (bindex = au_dbbot(dentry); bindex >= btop; bindex--) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + if (d_is_positive(h_dentry)) { + au_set_dbbot(dentry, bindex); + return; + } + au_set_h_dptr(dentry, bindex, NULL); + } +} + +int au_find_dbindex(struct dentry *dentry, struct dentry *h_dentry) +{ + aufs_bindex_t bindex, bbot; + + bbot = au_dbbot(dentry); + for (bindex = au_dbtop(dentry); bindex <= bbot; bindex++) + if (au_h_dptr(dentry, bindex) == h_dentry) + return bindex; + return -1; +} diff --git a/fs/aufs/dir.c b/fs/aufs/dir.c new file mode 100644 index 00000000000..91c31e737a7 --- /dev/null +++ b/fs/aufs/dir.c @@ -0,0 +1,765 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * directory operations + */ + +#include +#include +#include "aufs.h" + +void au_add_nlink(struct inode *dir, struct inode *h_dir) +{ + unsigned int nlink; + + AuDebugOn(!S_ISDIR(dir->i_mode) || !S_ISDIR(h_dir->i_mode)); + + nlink = dir->i_nlink; + nlink += h_dir->i_nlink - 2; + if (h_dir->i_nlink < 2) + nlink += 2; + smp_mb(); /* for i_nlink */ + /* 0 can happen in revaliding */ + set_nlink(dir, nlink); +} + +void au_sub_nlink(struct inode *dir, struct inode *h_dir) +{ + unsigned int nlink; + + AuDebugOn(!S_ISDIR(dir->i_mode) || !S_ISDIR(h_dir->i_mode)); + + nlink = dir->i_nlink; + nlink -= h_dir->i_nlink - 2; + if (h_dir->i_nlink < 2) + nlink -= 2; + smp_mb(); /* for i_nlink */ + /* nlink == 0 means the branch-fs is broken */ + set_nlink(dir, nlink); +} + +loff_t au_dir_size(struct file *file, struct dentry *dentry) +{ + loff_t sz; + aufs_bindex_t bindex, bbot; + struct file *h_file; + struct dentry *h_dentry; + + sz = 0; + if (file) { + AuDebugOn(!d_is_dir(file->f_path.dentry)); + + bbot = au_fbbot_dir(file); + for (bindex = au_fbtop(file); + bindex <= bbot && sz < KMALLOC_MAX_SIZE; + bindex++) { + h_file = au_hf_dir(file, bindex); + if (h_file && file_inode(h_file)) + sz += vfsub_f_size_read(h_file); + } + } else { + AuDebugOn(!dentry); + AuDebugOn(!d_is_dir(dentry)); + + bbot = au_dbtaildir(dentry); + for (bindex = au_dbtop(dentry); + bindex <= bbot && sz < KMALLOC_MAX_SIZE; + bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry && d_is_positive(h_dentry)) + sz += i_size_read(d_inode(h_dentry)); + } + } + if (sz < KMALLOC_MAX_SIZE) + sz = roundup_pow_of_two(sz); + if (sz > KMALLOC_MAX_SIZE) + sz = KMALLOC_MAX_SIZE; + else if (sz < NAME_MAX) { + BUILD_BUG_ON(AUFS_RDBLK_DEF < NAME_MAX); + sz = AUFS_RDBLK_DEF; + } + return sz; +} + +struct au_dir_ts_arg { + struct dentry *dentry; + aufs_bindex_t brid; +}; + +static void au_do_dir_ts(void *arg) +{ + struct au_dir_ts_arg *a = arg; + struct au_dtime dt; + struct path h_path; + struct inode *dir, *h_dir; + struct super_block *sb; + struct au_branch *br; + struct au_hinode *hdir; + int err; + aufs_bindex_t btop, bindex; + + sb = a->dentry->d_sb; + if (d_really_is_negative(a->dentry)) + goto out; + /* no dir->i_mutex lock */ + aufs_read_lock(a->dentry, AuLock_DW); /* noflush */ + + dir = d_inode(a->dentry); + btop = au_ibtop(dir); + bindex = au_br_index(sb, a->brid); + if (bindex < btop) + goto out_unlock; + + br = au_sbr(sb, bindex); + h_path.dentry = au_h_dptr(a->dentry, bindex); + if (!h_path.dentry) + goto out_unlock; + h_path.mnt = au_br_mnt(br); + au_dtime_store(&dt, a->dentry, &h_path); + + br = au_sbr(sb, btop); + if (!au_br_writable(br->br_perm)) + goto out_unlock; + h_path.dentry = au_h_dptr(a->dentry, btop); + h_path.mnt = au_br_mnt(br); + err = vfsub_mnt_want_write(h_path.mnt); + if (err) + goto out_unlock; + hdir = au_hi(dir, btop); + au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT); + h_dir = au_h_iptr(dir, btop); + if (h_dir->i_nlink + && timespec64_compare(&h_dir->i_mtime, &dt.dt_mtime) < 0) { + dt.dt_h_path = h_path; + au_dtime_revert(&dt); + } + au_hn_inode_unlock(hdir); + vfsub_mnt_drop_write(h_path.mnt); + au_cpup_attr_timesizes(dir); + +out_unlock: + aufs_read_unlock(a->dentry, AuLock_DW); +out: + dput(a->dentry); + au_nwt_done(&au_sbi(sb)->si_nowait); + au_kfree_try_rcu(arg); +} + +void au_dir_ts(struct inode *dir, aufs_bindex_t bindex) +{ + int perm, wkq_err; + aufs_bindex_t btop; + struct au_dir_ts_arg *arg; + struct dentry *dentry; + struct super_block *sb; + + IMustLock(dir); + + dentry = d_find_any_alias(dir); + AuDebugOn(!dentry); + sb = dentry->d_sb; + btop = au_ibtop(dir); + if (btop == bindex) { + au_cpup_attr_timesizes(dir); + goto out; + } + + perm = au_sbr_perm(sb, btop); + if (!au_br_writable(perm)) + goto out; + + arg = kmalloc(sizeof(*arg), GFP_NOFS); + if (!arg) + goto out; + + arg->dentry = dget(dentry); /* will be dput-ted by au_do_dir_ts() */ + arg->brid = au_sbr_id(sb, bindex); + wkq_err = au_wkq_nowait(au_do_dir_ts, arg, sb, /*flags*/0); + if (unlikely(wkq_err)) { + pr_err("wkq %d\n", wkq_err); + dput(dentry); + au_kfree_try_rcu(arg); + } + +out: + dput(dentry); +} + +/* ---------------------------------------------------------------------- */ + +static int reopen_dir(struct file *file) +{ + int err; + unsigned int flags; + aufs_bindex_t bindex, btail, btop; + struct dentry *dentry, *h_dentry; + struct file *h_file; + + /* open all lower dirs */ + dentry = file->f_path.dentry; + btop = au_dbtop(dentry); + for (bindex = au_fbtop(file); bindex < btop; bindex++) + au_set_h_fptr(file, bindex, NULL); + au_set_fbtop(file, btop); + + btail = au_dbtaildir(dentry); + for (bindex = au_fbbot_dir(file); btail < bindex; bindex--) + au_set_h_fptr(file, bindex, NULL); + au_set_fbbot_dir(file, btail); + + flags = vfsub_file_flags(file); + for (bindex = btop; bindex <= btail; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + h_file = au_hf_dir(file, bindex); + if (h_file) + continue; + + h_file = au_h_open(dentry, bindex, flags, file, /*force_wr*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; /* close all? */ + au_set_h_fptr(file, bindex, h_file); + } + au_update_figen(file); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + err = 0; + +out: + return err; +} + +static int do_open_dir(struct file *file, int flags, struct file *h_file) +{ + int err; + aufs_bindex_t bindex, btail; + struct dentry *dentry, *h_dentry; + struct vfsmount *mnt; + + FiMustWriteLock(file); + AuDebugOn(h_file); + + err = 0; + mnt = file->f_path.mnt; + dentry = file->f_path.dentry; + file->f_version = inode_query_iversion(d_inode(dentry)); + bindex = au_dbtop(dentry); + au_set_fbtop(file, bindex); + btail = au_dbtaildir(dentry); + au_set_fbbot_dir(file, btail); + for (; !err && bindex <= btail; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (!h_dentry) + continue; + + err = vfsub_test_mntns(mnt, h_dentry->d_sb); + if (unlikely(err)) + break; + h_file = au_h_open(dentry, bindex, flags, file, /*force_wr*/0); + if (IS_ERR(h_file)) { + err = PTR_ERR(h_file); + break; + } + au_set_h_fptr(file, bindex, h_file); + } + au_update_figen(file); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + if (!err) + return 0; /* success */ + + /* close all */ + for (bindex = au_fbtop(file); bindex <= btail; bindex++) + au_set_h_fptr(file, bindex, NULL); + au_set_fbtop(file, -1); + au_set_fbbot_dir(file, -1); + + return err; +} + +static int aufs_open_dir(struct inode *inode __maybe_unused, + struct file *file) +{ + int err; + struct super_block *sb; + struct au_fidir *fidir; + + err = -ENOMEM; + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH); + fidir = au_fidir_alloc(sb); + if (fidir) { + struct au_do_open_args args = { + .open = do_open_dir, + .fidir = fidir + }; + err = au_do_open(file, &args); + if (unlikely(err)) + au_kfree_rcu(fidir); + } + si_read_unlock(sb); + return err; +} + +static int aufs_release_dir(struct inode *inode __maybe_unused, + struct file *file) +{ + struct au_vdir *vdir_cache; + struct au_finfo *finfo; + struct au_fidir *fidir; + struct au_hfile *hf; + aufs_bindex_t bindex, bbot; + + finfo = au_fi(file); + fidir = finfo->fi_hdir; + if (fidir) { + au_hbl_del(&finfo->fi_hlist, + &au_sbi(file->f_path.dentry->d_sb)->si_files); + vdir_cache = fidir->fd_vdir_cache; /* lock-free */ + if (vdir_cache) + au_vdir_free(vdir_cache); + + bindex = finfo->fi_btop; + if (bindex >= 0) { + hf = fidir->fd_hfile + bindex; + /* + * calls fput() instead of filp_close(), + * since no dnotify or lock for the lower file. + */ + bbot = fidir->fd_bbot; + for (; bindex <= bbot; bindex++, hf++) + if (hf->hf_file) + au_hfput(hf, /*execed*/0); + } + au_kfree_rcu(fidir); + finfo->fi_hdir = NULL; + } + au_finfo_fin(file); + return 0; +} + +/* ---------------------------------------------------------------------- */ + +static int au_do_flush_dir(struct file *file, fl_owner_t id) +{ + int err; + aufs_bindex_t bindex, bbot; + struct file *h_file; + + err = 0; + bbot = au_fbbot_dir(file); + for (bindex = au_fbtop(file); !err && bindex <= bbot; bindex++) { + h_file = au_hf_dir(file, bindex); + if (h_file) + err = vfsub_flush(h_file, id); + } + return err; +} + +static int aufs_flush_dir(struct file *file, fl_owner_t id) +{ + return au_do_flush(file, id, au_do_flush_dir); +} + +/* ---------------------------------------------------------------------- */ + +static int au_do_fsync_dir_no_file(struct dentry *dentry, int datasync) +{ + int err; + aufs_bindex_t bbot, bindex; + struct inode *inode; + struct super_block *sb; + + err = 0; + sb = dentry->d_sb; + inode = d_inode(dentry); + IMustLock(inode); + bbot = au_dbbot(dentry); + for (bindex = au_dbtop(dentry); !err && bindex <= bbot; bindex++) { + struct path h_path; + + if (au_test_ro(sb, bindex, inode)) + continue; + h_path.dentry = au_h_dptr(dentry, bindex); + if (!h_path.dentry) + continue; + + h_path.mnt = au_sbr_mnt(sb, bindex); + err = vfsub_fsync(NULL, &h_path, datasync); + } + + return err; +} + +static int au_do_fsync_dir(struct file *file, int datasync) +{ + int err; + aufs_bindex_t bbot, bindex; + struct file *h_file; + struct super_block *sb; + struct inode *inode; + + err = au_reval_and_lock_fdi(file, reopen_dir, /*wlock*/1, /*fi_lsc*/0); + if (unlikely(err)) + goto out; + + inode = file_inode(file); + sb = inode->i_sb; + bbot = au_fbbot_dir(file); + for (bindex = au_fbtop(file); !err && bindex <= bbot; bindex++) { + h_file = au_hf_dir(file, bindex); + if (!h_file || au_test_ro(sb, bindex, inode)) + continue; + + err = vfsub_fsync(h_file, &h_file->f_path, datasync); + } + +out: + return err; +} + +/* + * @file may be NULL + */ +static int aufs_fsync_dir(struct file *file, loff_t start, loff_t end, + int datasync) +{ + int err; + struct dentry *dentry; + struct inode *inode; + struct super_block *sb; + + err = 0; + dentry = file->f_path.dentry; + inode = d_inode(dentry); + inode_lock(inode); + sb = dentry->d_sb; + si_noflush_read_lock(sb); + if (file) + err = au_do_fsync_dir(file, datasync); + else { + di_write_lock_child(dentry); + err = au_do_fsync_dir_no_file(dentry, datasync); + } + au_cpup_attr_timesizes(inode); + di_write_unlock(dentry); + if (file) + fi_write_unlock(file); + + si_read_unlock(sb); + inode_unlock(inode); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int aufs_iterate_shared(struct file *file, struct dir_context *ctx) +{ + int err; + struct dentry *dentry; + struct inode *inode, *h_inode; + struct super_block *sb; + + AuDbg("%pD, ctx{%ps, %llu}\n", file, ctx->actor, ctx->pos); + + dentry = file->f_path.dentry; + inode = d_inode(dentry); + IMustLock(inode); + + sb = dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH); + err = au_reval_and_lock_fdi(file, reopen_dir, /*wlock*/1, /*fi_lsc*/0); + if (unlikely(err)) + goto out; + err = au_alive_dir(dentry); + if (!err) + err = au_vdir_init(file); + di_downgrade_lock(dentry, AuLock_IR); + if (unlikely(err)) + goto out_unlock; + + h_inode = au_h_iptr(inode, au_ibtop(inode)); + if (!au_test_nfsd()) { + err = au_vdir_fill_de(file, ctx); + fsstack_copy_attr_atime(inode, h_inode); + } else { + /* + * nfsd filldir may call lookup_one_len(), vfs_getattr(), + * encode_fh() and others. + */ + atomic_inc(&h_inode->i_count); + di_read_unlock(dentry, AuLock_IR); + si_read_unlock(sb); + err = au_vdir_fill_de(file, ctx); + fsstack_copy_attr_atime(inode, h_inode); + fi_write_unlock(file); + iput(h_inode); + + AuTraceErr(err); + return err; + } + +out_unlock: + di_read_unlock(dentry, AuLock_IR); + fi_write_unlock(file); +out: + si_read_unlock(sb); + return err; +} + +/* ---------------------------------------------------------------------- */ + +#define AuTestEmpty_WHONLY 1 +#define AuTestEmpty_CALLED (1 << 1) +#define AuTestEmpty_SHWH (1 << 2) +#define au_ftest_testempty(flags, name) ((flags) & AuTestEmpty_##name) +#define au_fset_testempty(flags, name) \ + do { (flags) |= AuTestEmpty_##name; } while (0) +#define au_fclr_testempty(flags, name) \ + do { (flags) &= ~AuTestEmpty_##name; } while (0) + +#ifndef CONFIG_AUFS_SHWH +#undef AuTestEmpty_SHWH +#define AuTestEmpty_SHWH 0 +#endif + +struct test_empty_arg { + struct dir_context ctx; + struct au_nhash *whlist; + unsigned int flags; + int err; + aufs_bindex_t bindex; +}; + +static int test_empty_cb(struct dir_context *ctx, const char *__name, + int namelen, loff_t offset __maybe_unused, u64 ino, + unsigned int d_type) +{ + struct test_empty_arg *arg = container_of(ctx, struct test_empty_arg, + ctx); + char *name = (void *)__name; + + arg->err = 0; + au_fset_testempty(arg->flags, CALLED); + /* smp_mb(); */ + if (name[0] == '.' + && (namelen == 1 || (name[1] == '.' && namelen == 2))) + goto out; /* success */ + + if (namelen <= AUFS_WH_PFX_LEN + || memcmp(name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { + if (au_ftest_testempty(arg->flags, WHONLY) + && !au_nhash_test_known_wh(arg->whlist, name, namelen)) + arg->err = -ENOTEMPTY; + goto out; + } + + name += AUFS_WH_PFX_LEN; + namelen -= AUFS_WH_PFX_LEN; + if (!au_nhash_test_known_wh(arg->whlist, name, namelen)) + arg->err = au_nhash_append_wh + (arg->whlist, name, namelen, ino, d_type, arg->bindex, + au_ftest_testempty(arg->flags, SHWH)); + +out: + /* smp_mb(); */ + AuTraceErr(arg->err); + return arg->err; +} + +static int do_test_empty(struct dentry *dentry, struct test_empty_arg *arg) +{ + int err; + struct file *h_file; + struct au_branch *br; + + h_file = au_h_open(dentry, arg->bindex, + O_RDONLY | O_NONBLOCK | O_DIRECTORY | O_LARGEFILE, + /*file*/NULL, /*force_wr*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + err = 0; + if (!au_opt_test(au_mntflags(dentry->d_sb), UDBA_NONE) + && !file_inode(h_file)->i_nlink) + goto out_put; + + do { + arg->err = 0; + au_fclr_testempty(arg->flags, CALLED); + /* smp_mb(); */ + err = vfsub_iterate_dir(h_file, &arg->ctx); + if (err >= 0) + err = arg->err; + } while (!err && au_ftest_testempty(arg->flags, CALLED)); + +out_put: + fput(h_file); + br = au_sbr(dentry->d_sb, arg->bindex); + au_lcnt_dec(&br->br_nfiles); +out: + return err; +} + +struct do_test_empty_args { + int *errp; + struct dentry *dentry; + struct test_empty_arg *arg; +}; + +static void call_do_test_empty(void *args) +{ + struct do_test_empty_args *a = args; + *a->errp = do_test_empty(a->dentry, a->arg); +} + +static int sio_test_empty(struct dentry *dentry, struct test_empty_arg *arg) +{ + int err, wkq_err; + struct dentry *h_dentry; + struct inode *h_inode; + struct user_namespace *h_userns; + + h_userns = au_sbr_userns(dentry->d_sb, arg->bindex); + h_dentry = au_h_dptr(dentry, arg->bindex); + h_inode = d_inode(h_dentry); + /* todo: i_mode changes anytime? */ + inode_lock_shared_nested(h_inode, AuLsc_I_CHILD); + err = au_test_h_perm_sio(h_userns, h_inode, MAY_EXEC | MAY_READ); + inode_unlock_shared(h_inode); + if (!err) + err = do_test_empty(dentry, arg); + else { + struct do_test_empty_args args = { + .errp = &err, + .dentry = dentry, + .arg = arg + }; + unsigned int flags = arg->flags; + + wkq_err = au_wkq_wait(call_do_test_empty, &args); + if (unlikely(wkq_err)) + err = wkq_err; + arg->flags = flags; + } + + return err; +} + +int au_test_empty_lower(struct dentry *dentry) +{ + int err; + unsigned int rdhash; + aufs_bindex_t bindex, btop, btail; + struct au_nhash whlist; + struct test_empty_arg arg = { + .ctx = { + .actor = test_empty_cb + } + }; + int (*test_empty)(struct dentry *dentry, struct test_empty_arg *arg); + + SiMustAnyLock(dentry->d_sb); + + rdhash = au_sbi(dentry->d_sb)->si_rdhash; + if (!rdhash) + rdhash = au_rdhash_est(au_dir_size(/*file*/NULL, dentry)); + err = au_nhash_alloc(&whlist, rdhash, GFP_NOFS); + if (unlikely(err)) + goto out; + + arg.flags = 0; + arg.whlist = &whlist; + btop = au_dbtop(dentry); + if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)) + au_fset_testempty(arg.flags, SHWH); + test_empty = do_test_empty; + if (au_opt_test(au_mntflags(dentry->d_sb), DIRPERM1)) + test_empty = sio_test_empty; + arg.bindex = btop; + err = test_empty(dentry, &arg); + if (unlikely(err)) + goto out_whlist; + + au_fset_testempty(arg.flags, WHONLY); + btail = au_dbtaildir(dentry); + for (bindex = btop + 1; !err && bindex <= btail; bindex++) { + struct dentry *h_dentry; + + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry && d_is_positive(h_dentry)) { + arg.bindex = bindex; + err = test_empty(dentry, &arg); + } + } + +out_whlist: + au_nhash_wh_free(&whlist); +out: + return err; +} + +int au_test_empty(struct dentry *dentry, struct au_nhash *whlist) +{ + int err; + struct test_empty_arg arg = { + .ctx = { + .actor = test_empty_cb + } + }; + aufs_bindex_t bindex, btail; + + err = 0; + arg.whlist = whlist; + arg.flags = AuTestEmpty_WHONLY; + if (au_opt_test(au_mntflags(dentry->d_sb), SHWH)) + au_fset_testempty(arg.flags, SHWH); + btail = au_dbtaildir(dentry); + for (bindex = au_dbtop(dentry); !err && bindex <= btail; bindex++) { + struct dentry *h_dentry; + + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry && d_is_positive(h_dentry)) { + arg.bindex = bindex; + err = sio_test_empty(dentry, &arg); + } + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +const struct file_operations aufs_dir_fop = { + .owner = THIS_MODULE, + .llseek = default_llseek, + .read = generic_read_dir, + .iterate_shared = aufs_iterate_shared, + .unlocked_ioctl = aufs_ioctl_dir, +#ifdef CONFIG_COMPAT + .compat_ioctl = aufs_compat_ioctl_dir, +#endif + .open = aufs_open_dir, + .release = aufs_release_dir, + .flush = aufs_flush_dir, + .fsync = aufs_fsync_dir +}; diff --git a/fs/aufs/dir.h b/fs/aufs/dir.h new file mode 100644 index 00000000000..e44c2a1a849 --- /dev/null +++ b/fs/aufs/dir.h @@ -0,0 +1,134 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * directory operations + */ + +#ifndef __AUFS_DIR_H__ +#define __AUFS_DIR_H__ + +#ifdef __KERNEL__ + +#include + +/* ---------------------------------------------------------------------- */ + +/* need to be faster and smaller */ + +struct au_nhash { + unsigned int nh_num; + struct hlist_head *nh_head; +}; + +struct au_vdir_destr { + unsigned char len; + unsigned char name[]; +} __packed; + +struct au_vdir_dehstr { + struct hlist_node hash; + struct au_vdir_destr *str; + struct rcu_head rcu; +} ____cacheline_aligned_in_smp; + +struct au_vdir_de { + ino_t de_ino; + unsigned char de_type; + /* caution: packed */ + struct au_vdir_destr de_str; +} __packed; + +struct au_vdir_wh { + struct hlist_node wh_hash; +#ifdef CONFIG_AUFS_SHWH + ino_t wh_ino; + aufs_bindex_t wh_bindex; + unsigned char wh_type; +#else + aufs_bindex_t wh_bindex; +#endif + /* caution: packed */ + struct au_vdir_destr wh_str; +} __packed; + +union au_vdir_deblk_p { + unsigned char *deblk; + struct au_vdir_de *de; +}; + +struct au_vdir { + unsigned char **vd_deblk; + unsigned long vd_nblk; + struct { + unsigned long ul; + union au_vdir_deblk_p p; + } vd_last; + + u64 vd_version; + unsigned int vd_deblk_sz; + unsigned long vd_jiffy; + struct rcu_head rcu; +} ____cacheline_aligned_in_smp; + +/* ---------------------------------------------------------------------- */ + +/* dir.c */ +extern const struct file_operations aufs_dir_fop; +void au_add_nlink(struct inode *dir, struct inode *h_dir); +void au_sub_nlink(struct inode *dir, struct inode *h_dir); +loff_t au_dir_size(struct file *file, struct dentry *dentry); +void au_dir_ts(struct inode *dir, aufs_bindex_t bsrc); +int au_test_empty_lower(struct dentry *dentry); +int au_test_empty(struct dentry *dentry, struct au_nhash *whlist); + +/* vdir.c */ +unsigned int au_rdhash_est(loff_t sz); +int au_nhash_alloc(struct au_nhash *nhash, unsigned int num_hash, gfp_t gfp); +void au_nhash_wh_free(struct au_nhash *whlist); +int au_nhash_test_longer_wh(struct au_nhash *whlist, aufs_bindex_t btgt, + int limit); +int au_nhash_test_known_wh(struct au_nhash *whlist, char *name, int nlen); +int au_nhash_append_wh(struct au_nhash *whlist, char *name, int nlen, ino_t ino, + unsigned int d_type, aufs_bindex_t bindex, + unsigned char shwh); +void au_vdir_free(struct au_vdir *vdir); +int au_vdir_init(struct file *file); +int au_vdir_fill_de(struct file *file, struct dir_context *ctx); + +/* ioctl.c */ +long aufs_ioctl_dir(struct file *file, unsigned int cmd, unsigned long arg); + +#ifdef CONFIG_AUFS_RDU +/* rdu.c */ +long au_rdu_ioctl(struct file *file, unsigned int cmd, unsigned long arg); +#ifdef CONFIG_COMPAT +long au_rdu_compat_ioctl(struct file *file, unsigned int cmd, + unsigned long arg); +#endif +#else +AuStub(long, au_rdu_ioctl, return -EINVAL, struct file *file, + unsigned int cmd, unsigned long arg) +#ifdef CONFIG_COMPAT +AuStub(long, au_rdu_compat_ioctl, return -EINVAL, struct file *file, + unsigned int cmd, unsigned long arg) +#endif +#endif + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DIR_H__ */ diff --git a/fs/aufs/dirren.c b/fs/aufs/dirren.c new file mode 100644 index 00000000000..dd530d0188e --- /dev/null +++ b/fs/aufs/dirren.c @@ -0,0 +1,1316 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2017-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * special handling in renaming a directory + * in order to support looking-up the before-renamed name on the lower readonly + * branches + */ + +#include +#include "aufs.h" + +static void au_dr_hino_del(struct au_dr_br *dr, struct au_dr_hino *ent) +{ + int idx; + + idx = au_dr_ihash(ent->dr_h_ino); + au_hbl_del(&ent->dr_hnode, dr->dr_h_ino + idx); +} + +static int au_dr_hino_test_empty(struct au_dr_br *dr) +{ + int ret, i; + struct hlist_bl_head *hbl; + + ret = 1; + for (i = 0; ret && i < AuDirren_NHASH; i++) { + hbl = dr->dr_h_ino + i; + hlist_bl_lock(hbl); + ret &= hlist_bl_empty(hbl); + hlist_bl_unlock(hbl); + } + + return ret; +} + +static struct au_dr_hino *au_dr_hino_find(struct au_dr_br *dr, ino_t ino) +{ + struct au_dr_hino *found, *ent; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + int idx; + + found = NULL; + idx = au_dr_ihash(ino); + hbl = dr->dr_h_ino + idx; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(ent, pos, hbl, dr_hnode) + if (ent->dr_h_ino == ino) { + found = ent; + break; + } + hlist_bl_unlock(hbl); + + return found; +} + +int au_dr_hino_test_add(struct au_dr_br *dr, ino_t ino, + struct au_dr_hino *add_ent) +{ + int found, idx; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_dr_hino *ent; + + found = 0; + idx = au_dr_ihash(ino); + hbl = dr->dr_h_ino + idx; +#if 0 /* debug print */ + { + struct hlist_bl_node *tmp; + + hlist_bl_for_each_entry_safe(ent, pos, tmp, hbl, dr_hnode) + AuDbg("hi%llu\n", (unsigned long long)ent->dr_h_ino); + } +#endif + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(ent, pos, hbl, dr_hnode) + if (ent->dr_h_ino == ino) { + found = 1; + break; + } + if (!found && add_ent) + hlist_bl_add_head(&add_ent->dr_hnode, hbl); + hlist_bl_unlock(hbl); + + if (!found && add_ent) + AuDbg("i%llu added\n", (unsigned long long)add_ent->dr_h_ino); + + return found; +} + +void au_dr_hino_free(struct au_dr_br *dr) +{ + int i; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos, *tmp; + struct au_dr_hino *ent; + + /* SiMustWriteLock(sb); */ + + for (i = 0; i < AuDirren_NHASH; i++) { + hbl = dr->dr_h_ino + i; + /* no spinlock since sbinfo must be write-locked */ + hlist_bl_for_each_entry_safe(ent, pos, tmp, hbl, dr_hnode) + au_kfree_rcu(ent); + INIT_HLIST_BL_HEAD(hbl); + } +} + +/* returns the number of inodes or an error */ +static int au_dr_hino_store(struct super_block *sb, struct au_branch *br, + struct file *hinofile) +{ + int err, i; + ssize_t ssz; + loff_t pos, oldsize; + __be64 u64; + struct inode *hinoinode; + struct hlist_bl_head *hbl; + struct hlist_bl_node *n1, *n2; + struct au_dr_hino *ent; + + SiMustWriteLock(sb); + AuDebugOn(!au_br_writable(br->br_perm)); + + hinoinode = file_inode(hinofile); + oldsize = i_size_read(hinoinode); + + err = 0; + pos = 0; + hbl = br->br_dirren.dr_h_ino; + for (i = 0; !err && i < AuDirren_NHASH; i++, hbl++) { + /* no bit-lock since sbinfo must be write-locked */ + hlist_bl_for_each_entry_safe(ent, n1, n2, hbl, dr_hnode) { + AuDbg("hi%llu, %pD2\n", + (unsigned long long)ent->dr_h_ino, hinofile); + u64 = cpu_to_be64(ent->dr_h_ino); + ssz = vfsub_write_k(hinofile, &u64, sizeof(u64), &pos); + if (ssz == sizeof(u64)) + continue; + + /* write error */ + pr_err("ssz %zd, %pD2\n", ssz, hinofile); + err = -ENOSPC; + if (ssz < 0) + err = ssz; + break; + } + } + /* regardless the error */ + if (pos < oldsize) { + err = vfsub_trunc(&hinofile->f_path, pos, /*attr*/0, hinofile); + AuTraceErr(err); + } + + AuTraceErr(err); + return err; +} + +static int au_dr_hino_load(struct au_dr_br *dr, struct file *hinofile) +{ + int err, hidx; + ssize_t ssz; + size_t sz, n; + loff_t pos; + uint64_t u64; + struct au_dr_hino *ent; + struct inode *hinoinode; + struct hlist_bl_head *hbl; + + err = 0; + pos = 0; + hbl = dr->dr_h_ino; + hinoinode = file_inode(hinofile); + sz = i_size_read(hinoinode); + AuDebugOn(sz % sizeof(u64)); + n = sz / sizeof(u64); + while (n--) { + ssz = vfsub_read_k(hinofile, &u64, sizeof(u64), &pos); + if (unlikely(ssz != sizeof(u64))) { + pr_err("ssz %zd, %pD2\n", ssz, hinofile); + err = -EINVAL; + if (ssz < 0) + err = ssz; + goto out_free; + } + + ent = kmalloc(sizeof(*ent), GFP_NOFS); + if (!ent) { + err = -ENOMEM; + AuTraceErr(err); + goto out_free; + } + ent->dr_h_ino = be64_to_cpu((__force __be64)u64); + AuDbg("hi%llu, %pD2\n", + (unsigned long long)ent->dr_h_ino, hinofile); + hidx = au_dr_ihash(ent->dr_h_ino); + au_hbl_add(&ent->dr_hnode, hbl + hidx); + } + goto out; /* success */ + +out_free: + au_dr_hino_free(dr); +out: + AuTraceErr(err); + return err; +} + +/* + * @bindex/@br is a switch to distinguish whether suspending hnotify or not. + * @path is a switch to distinguish load and store. + */ +static int au_dr_hino(struct super_block *sb, aufs_bindex_t bindex, + struct au_branch *br, const struct path *path) +{ + int err, flags; + unsigned char load, suspend; + struct file *hinofile; + struct au_hinode *hdir; + struct inode *dir, *delegated; + struct path hinopath; + struct qstr hinoname = QSTR_INIT(AUFS_WH_DR_BRHINO, + sizeof(AUFS_WH_DR_BRHINO) - 1); + + AuDebugOn(bindex < 0 && !br); + AuDebugOn(bindex >= 0 && br); + + err = -EINVAL; + suspend = !br; + if (suspend) + br = au_sbr(sb, bindex); + load = !!path; + if (!load) { + path = &br->br_path; + AuDebugOn(!au_br_writable(br->br_perm)); + if (unlikely(!au_br_writable(br->br_perm))) + goto out; + } + + hdir = NULL; + if (suspend) { + dir = d_inode(sb->s_root); + hdir = au_hinode(au_ii(dir), bindex); + dir = hdir->hi_inode; + au_hn_inode_lock_nested(hdir, AuLsc_I_CHILD); + } else { + dir = d_inode(path->dentry); + inode_lock_nested(dir, AuLsc_I_CHILD); + } + hinopath.dentry = vfsub_lkup_one(&hinoname, path->dentry); + err = PTR_ERR(hinopath.dentry); + if (IS_ERR(hinopath.dentry)) + goto out_unlock; + hinopath.mnt = path->mnt; + + err = 0; + flags = O_RDONLY; + if (load) { + if (d_is_negative(hinopath.dentry)) + goto out_dput; /* success */ + } else { + if (au_dr_hino_test_empty(&br->br_dirren)) { + if (d_is_positive(hinopath.dentry)) { + delegated = NULL; + err = vfsub_unlink(dir, &hinopath, &delegated, + /*force*/0); + AuTraceErr(err); + if (unlikely(err)) + pr_err("ignored err %d, %pd2\n", + err, hinopath.dentry); + if (unlikely(err == -EWOULDBLOCK)) + iput(delegated); + err = 0; + } + goto out_dput; + } else if (!d_is_positive(hinopath.dentry)) { + err = vfsub_create(dir, &hinopath, 0600, + /*want_excl*/false); + AuTraceErr(err); + if (unlikely(err)) + goto out_dput; + } + flags = O_WRONLY; + } + hinofile = vfsub_dentry_open(&hinopath, flags); + if (suspend) + au_hn_inode_unlock(hdir); + else + inode_unlock(dir); + dput(hinopath.dentry); + AuTraceErrPtr(hinofile); + if (IS_ERR(hinofile)) { + err = PTR_ERR(hinofile); + goto out; + } + + if (load) + err = au_dr_hino_load(&br->br_dirren, hinofile); + else + err = au_dr_hino_store(sb, br, hinofile); + fput(hinofile); + goto out; + +out_dput: + dput(hinopath.dentry); +out_unlock: + if (suspend) + au_hn_inode_unlock(hdir); + else + inode_unlock(dir); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_dr_brid_init(struct au_dr_brid *brid, const struct path *path) +{ + int err; + struct kstatfs kstfs; + dev_t dev; + struct dentry *dentry; + struct super_block *sb; + + err = vfs_statfs((void *)path, &kstfs); + AuTraceErr(err); + if (unlikely(err)) + goto out; + + /* todo: support for UUID */ + + if (kstfs.f_fsid.val[0] || kstfs.f_fsid.val[1]) { + brid->type = AuBrid_FSID; + brid->fsid = kstfs.f_fsid; + } else { + dentry = path->dentry; + sb = dentry->d_sb; + dev = sb->s_dev; + if (dev) { + brid->type = AuBrid_DEV; + brid->dev = dev; + } + } + +out: + return err; +} + +int au_dr_br_init(struct super_block *sb, struct au_branch *br, + const struct path *path) +{ + int err, i; + struct au_dr_br *dr; + struct hlist_bl_head *hbl; + + dr = &br->br_dirren; + hbl = dr->dr_h_ino; + for (i = 0; i < AuDirren_NHASH; i++, hbl++) + INIT_HLIST_BL_HEAD(hbl); + + err = au_dr_brid_init(&dr->dr_brid, path); + if (unlikely(err)) + goto out; + + if (au_opt_test(au_mntflags(sb), DIRREN)) + err = au_dr_hino(sb, /*bindex*/-1, br, path); + +out: + AuTraceErr(err); + return err; +} + +int au_dr_br_fin(struct super_block *sb, struct au_branch *br) +{ + int err; + + err = 0; + if (au_br_writable(br->br_perm)) + err = au_dr_hino(sb, /*bindex*/-1, br, /*path*/NULL); + if (!err) + au_dr_hino_free(&br->br_dirren); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_brid_str(struct au_dr_brid *brid, struct inode *h_inode, + char *buf, size_t sz) +{ + int err; + unsigned int major, minor; + char *p; + + p = buf; + err = snprintf(p, sz, "%d_", brid->type); + AuDebugOn(err > sz); + p += err; + sz -= err; + switch (brid->type) { + case AuBrid_Unset: + return -EINVAL; + case AuBrid_UUID: + err = snprintf(p, sz, "%pU", brid->uuid.b); + break; + case AuBrid_FSID: + err = snprintf(p, sz, "%08x-%08x", + brid->fsid.val[0], brid->fsid.val[1]); + break; + case AuBrid_DEV: + major = MAJOR(brid->dev); + minor = MINOR(brid->dev); + if (major <= 0xff && minor <= 0xff) + err = snprintf(p, sz, "%02x%02x", major, minor); + else + err = snprintf(p, sz, "%03x:%05x", major, minor); + break; + } + AuDebugOn(err > sz); + p += err; + sz -= err; + err = snprintf(p, sz, "_%llu", (unsigned long long)h_inode->i_ino); + AuDebugOn(err > sz); + p += err; + sz -= err; + + return p - buf; +} + +static int au_drinfo_name(struct au_branch *br, char *name, int len) +{ + int rlen; + struct dentry *br_dentry; + struct inode *br_inode; + + br_dentry = au_br_dentry(br); + br_inode = d_inode(br_dentry); + rlen = au_brid_str(&br->br_dirren.dr_brid, br_inode, name, len); + AuDebugOn(rlen >= AUFS_DIRREN_ENV_VAL_SZ); + AuDebugOn(rlen > len); + + return rlen; +} + +/* ---------------------------------------------------------------------- */ + +/* + * from the given @h_dentry, construct drinfo at @*fdata. + * when the size of @*fdata is not enough, reallocate and return new @fdata and + * @allocated. + */ +static int au_drinfo_construct(struct au_drinfo_fdata **fdata, + struct dentry *h_dentry, + unsigned char *allocated) +{ + int err, v; + struct au_drinfo_fdata *f, *p; + struct au_drinfo *drinfo; + struct inode *h_inode; + struct qstr *qname; + + err = 0; + f = *fdata; + h_inode = d_inode(h_dentry); + qname = &h_dentry->d_name; + drinfo = &f->drinfo; + drinfo->ino = (__force uint64_t)cpu_to_be64(h_inode->i_ino); + drinfo->oldnamelen = qname->len; + if (*allocated < sizeof(*f) + qname->len) { + v = roundup_pow_of_two(*allocated + qname->len); + p = au_krealloc(f, v, GFP_NOFS, /*may_shrink*/0); + if (unlikely(!p)) { + err = -ENOMEM; + AuTraceErr(err); + goto out; + } + f = p; + *fdata = f; + *allocated = v; + drinfo = &f->drinfo; + } + memcpy(drinfo->oldname, qname->name, qname->len); + AuDbg("i%llu, %.*s\n", + be64_to_cpu((__force __be64)drinfo->ino), drinfo->oldnamelen, + drinfo->oldname); + +out: + AuTraceErr(err); + return err; +} + +/* callers have to free the return value */ +static struct au_drinfo *au_drinfo_read_k(struct file *file, ino_t h_ino) +{ + struct au_drinfo *ret, *drinfo; + struct au_drinfo_fdata fdata; + int len; + loff_t pos; + ssize_t ssz; + + ret = ERR_PTR(-EIO); + pos = 0; + ssz = vfsub_read_k(file, &fdata, sizeof(fdata), &pos); + if (unlikely(ssz != sizeof(fdata))) { + AuIOErr("ssz %zd, %u, %pD2\n", + ssz, (unsigned int)sizeof(fdata), file); + goto out; + } + + fdata.magic = ntohl((__force __be32)fdata.magic); + switch (fdata.magic) { + case AUFS_DRINFO_MAGIC_V1: + break; + default: + AuIOErr("magic-num 0x%x, 0x%x, %pD2\n", + fdata.magic, AUFS_DRINFO_MAGIC_V1, file); + goto out; + } + + drinfo = &fdata.drinfo; + len = drinfo->oldnamelen; + if (!len) { + AuIOErr("broken drinfo %pD2\n", file); + goto out; + } + + ret = NULL; + drinfo->ino = be64_to_cpu((__force __be64)drinfo->ino); + if (unlikely(h_ino && drinfo->ino != h_ino)) { + AuDbg("ignored i%llu, i%llu, %pD2\n", + (unsigned long long)drinfo->ino, + (unsigned long long)h_ino, file); + goto out; /* success */ + } + + ret = kmalloc(sizeof(*ret) + len, GFP_NOFS); + if (unlikely(!ret)) { + ret = ERR_PTR(-ENOMEM); + AuTraceErrPtr(ret); + goto out; + } + + *ret = *drinfo; + ssz = vfsub_read_k(file, (void *)ret->oldname, len, &pos); + if (unlikely(ssz != len)) { + au_kfree_rcu(ret); + ret = ERR_PTR(-EIO); + AuIOErr("ssz %zd, %u, %pD2\n", ssz, len, file); + goto out; + } + + AuDbg("oldname %.*s\n", ret->oldnamelen, ret->oldname); + +out: + return ret; +} + +/* ---------------------------------------------------------------------- */ + +/* in order to be revertible */ +struct au_drinfo_rev_elm { + int created; + struct dentry *info_dentry; + struct au_drinfo *info_last; +}; + +struct au_drinfo_rev { + unsigned char already; + aufs_bindex_t nelm; + struct au_drinfo_rev_elm elm[]; +}; + +/* todo: isn't it too large? */ +struct au_drinfo_store { + struct path h_ppath; + struct dentry *h_dentry; + struct au_drinfo_fdata *fdata; + char *infoname; /* inside of whname, just after PFX */ + char whname[sizeof(AUFS_WH_DR_INFO_PFX) + AUFS_DIRREN_ENV_VAL_SZ]; + aufs_bindex_t btgt, btail; + unsigned char no_sio, + allocated, /* current size of *fdata */ + infonamelen, /* room size for p */ + whnamelen, /* length of the generated name */ + renameback; /* renamed back */ +}; + +/* on rename(2) error, the caller should revert it using @elm */ +static int au_drinfo_do_store(struct au_drinfo_store *w, + struct au_drinfo_rev_elm *elm) +{ + int err, len; + ssize_t ssz; + loff_t pos; + struct path infopath = { + .mnt = w->h_ppath.mnt + }; + struct inode *h_dir, *h_inode, *delegated; + struct file *infofile; + struct qstr *qname; + + AuDebugOn(elm + && memcmp(elm, page_address(ZERO_PAGE(0)), sizeof(*elm))); + + infopath.dentry = vfsub_lookup_one_len(w->whname, w->h_ppath.dentry, + w->whnamelen); + AuTraceErrPtr(infopath.dentry); + if (IS_ERR(infopath.dentry)) { + err = PTR_ERR(infopath.dentry); + goto out; + } + + err = 0; + h_dir = d_inode(w->h_ppath.dentry); + if (elm && d_is_negative(infopath.dentry)) { + err = vfsub_create(h_dir, &infopath, 0600, /*want_excl*/true); + AuTraceErr(err); + if (unlikely(err)) + goto out_dput; + elm->created = 1; + elm->info_dentry = dget(infopath.dentry); + } + + infofile = vfsub_dentry_open(&infopath, O_RDWR); + AuTraceErrPtr(infofile); + if (IS_ERR(infofile)) { + err = PTR_ERR(infofile); + goto out_dput; + } + + h_inode = d_inode(infopath.dentry); + if (elm && i_size_read(h_inode)) { + h_inode = d_inode(w->h_dentry); + elm->info_last = au_drinfo_read_k(infofile, h_inode->i_ino); + AuTraceErrPtr(elm->info_last); + if (IS_ERR(elm->info_last)) { + err = PTR_ERR(elm->info_last); + elm->info_last = NULL; + AuDebugOn(elm->info_dentry); + goto out_fput; + } + } + + if (elm && w->renameback) { + delegated = NULL; + err = vfsub_unlink(h_dir, &infopath, &delegated, /*force*/0); + AuTraceErr(err); + if (unlikely(err == -EWOULDBLOCK)) + iput(delegated); + goto out_fput; + } + + pos = 0; + qname = &w->h_dentry->d_name; + len = sizeof(*w->fdata) + qname->len; + if (!elm) + len = sizeof(*w->fdata) + w->fdata->drinfo.oldnamelen; + ssz = vfsub_write_k(infofile, w->fdata, len, &pos); + if (ssz == len) { + AuDbg("hi%llu, %.*s\n", w->fdata->drinfo.ino, + w->fdata->drinfo.oldnamelen, w->fdata->drinfo.oldname); + goto out_fput; /* success */ + } else { + err = -EIO; + if (ssz < 0) + err = ssz; + /* the caller should revert it using @elm */ + } + +out_fput: + fput(infofile); +out_dput: + dput(infopath.dentry); +out: + AuTraceErr(err); + return err; +} + +struct au_call_drinfo_do_store_args { + int *errp; + struct au_drinfo_store *w; + struct au_drinfo_rev_elm *elm; +}; + +static void au_call_drinfo_do_store(void *args) +{ + struct au_call_drinfo_do_store_args *a = args; + + *a->errp = au_drinfo_do_store(a->w, a->elm); +} + +static int au_drinfo_store_sio(struct au_drinfo_store *w, + struct au_drinfo_rev_elm *elm) +{ + int err, wkq_err; + + if (w->no_sio) + err = au_drinfo_do_store(w, elm); + else { + struct au_call_drinfo_do_store_args a = { + .errp = &err, + .w = w, + .elm = elm + }; + wkq_err = au_wkq_wait(au_call_drinfo_do_store, &a); + if (unlikely(wkq_err)) + err = wkq_err; + } + AuTraceErr(err); + + return err; +} + +static int au_drinfo_store_work_init(struct au_drinfo_store *w, + aufs_bindex_t btgt) +{ + int err; + + memset(w, 0, sizeof(*w)); + w->allocated = roundup_pow_of_two(sizeof(*w->fdata) + 40); + strcpy(w->whname, AUFS_WH_DR_INFO_PFX); + w->infoname = w->whname + sizeof(AUFS_WH_DR_INFO_PFX) - 1; + w->infonamelen = sizeof(w->whname) - sizeof(AUFS_WH_DR_INFO_PFX); + w->btgt = btgt; + w->no_sio = !!uid_eq(current_fsuid(), GLOBAL_ROOT_UID); + + err = -ENOMEM; + w->fdata = kcalloc(1, w->allocated, GFP_NOFS); + if (unlikely(!w->fdata)) { + AuTraceErr(err); + goto out; + } + w->fdata->magic = (__force uint32_t)htonl(AUFS_DRINFO_MAGIC_V1); + err = 0; + +out: + return err; +} + +static void au_drinfo_store_work_fin(struct au_drinfo_store *w) +{ + au_kfree_rcu(w->fdata); +} + +static void au_drinfo_store_rev(struct au_drinfo_rev *rev, + struct au_drinfo_store *w) +{ + struct au_drinfo_rev_elm *elm; + struct inode *h_dir, *delegated; + int err, nelm; + struct path infopath = { + .mnt = w->h_ppath.mnt + }; + + h_dir = d_inode(w->h_ppath.dentry); + IMustLock(h_dir); + + err = 0; + elm = rev->elm; + for (nelm = rev->nelm; nelm > 0; nelm--, elm++) { + AuDebugOn(elm->created && elm->info_last); + if (elm->created) { + AuDbg("here\n"); + delegated = NULL; + infopath.dentry = elm->info_dentry; + err = vfsub_unlink(h_dir, &infopath, &delegated, + !w->no_sio); + AuTraceErr(err); + if (unlikely(err == -EWOULDBLOCK)) + iput(delegated); + dput(elm->info_dentry); + } else if (elm->info_last) { + AuDbg("here\n"); + w->fdata->drinfo = *elm->info_last; + memcpy(w->fdata->drinfo.oldname, + elm->info_last->oldname, + elm->info_last->oldnamelen); + err = au_drinfo_store_sio(w, /*elm*/NULL); + au_kfree_rcu(elm->info_last); + } + if (unlikely(err)) + AuIOErr("%d, %s\n", err, w->whname); + /* go on even if err */ + } +} + +/* caller has to call au_dr_rename_fin() later */ +static int au_drinfo_store(struct dentry *dentry, aufs_bindex_t btgt, + struct qstr *dst_name, void *_rev) +{ + int err, sz, nelm; + aufs_bindex_t bindex, btail; + struct au_drinfo_store work; + struct au_drinfo_rev *rev, **p; + struct au_drinfo_rev_elm *elm; + struct super_block *sb; + struct au_branch *br; + struct au_hinode *hdir; + + err = au_drinfo_store_work_init(&work, btgt); + AuTraceErr(err); + if (unlikely(err)) + goto out; + + err = -ENOMEM; + btail = au_dbtaildir(dentry); + nelm = btail - btgt; + sz = sizeof(*rev) + sizeof(*elm) * nelm; + rev = kcalloc(1, sz, GFP_NOFS); + if (unlikely(!rev)) { + AuTraceErr(err); + goto out_args; + } + rev->nelm = nelm; + elm = rev->elm; + p = _rev; + *p = rev; + + err = 0; + sb = dentry->d_sb; + work.h_ppath.dentry = au_h_dptr(dentry, btgt); + work.h_ppath.mnt = au_sbr_mnt(sb, btgt); + hdir = au_hi(d_inode(dentry), btgt); + au_hn_inode_lock_nested(hdir, AuLsc_I_CHILD); + for (bindex = btgt + 1; bindex <= btail; bindex++, elm++) { + work.h_dentry = au_h_dptr(dentry, bindex); + if (!work.h_dentry) + continue; + + err = au_drinfo_construct(&work.fdata, work.h_dentry, + &work.allocated); + AuTraceErr(err); + if (unlikely(err)) + break; + + work.renameback = au_qstreq(&work.h_dentry->d_name, dst_name); + br = au_sbr(sb, bindex); + work.whnamelen = sizeof(AUFS_WH_DR_INFO_PFX) - 1; + work.whnamelen += au_drinfo_name(br, work.infoname, + work.infonamelen); + AuDbg("whname %.*s, i%llu, %.*s\n", + work.whnamelen, work.whname, + be64_to_cpu((__force __be64)work.fdata->drinfo.ino), + work.fdata->drinfo.oldnamelen, + work.fdata->drinfo.oldname); + + err = au_drinfo_store_sio(&work, elm); + AuTraceErr(err); + if (unlikely(err)) + break; + } + if (unlikely(err)) { + /* revert all drinfo */ + au_drinfo_store_rev(rev, &work); + au_kfree_try_rcu(rev); + *p = NULL; + } + au_hn_inode_unlock(hdir); + +out_args: + au_drinfo_store_work_fin(&work); +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_dr_rename(struct dentry *src, aufs_bindex_t bindex, + struct qstr *dst_name, void *_rev) +{ + int err, already; + ino_t ino; + struct super_block *sb; + struct au_branch *br; + struct au_dr_br *dr; + struct dentry *h_dentry; + struct inode *h_inode; + struct au_dr_hino *ent; + struct au_drinfo_rev *rev, **p; + + AuDbg("bindex %d\n", bindex); + + err = -ENOMEM; + ent = kmalloc(sizeof(*ent), GFP_NOFS); + if (unlikely(!ent)) + goto out; + + sb = src->d_sb; + br = au_sbr(sb, bindex); + dr = &br->br_dirren; + h_dentry = au_h_dptr(src, bindex); + h_inode = d_inode(h_dentry); + ino = h_inode->i_ino; + ent->dr_h_ino = ino; + already = au_dr_hino_test_add(dr, ino, ent); + AuDbg("b%d, hi%llu, already %d\n", + bindex, (unsigned long long)ino, already); + + err = au_drinfo_store(src, bindex, dst_name, _rev); + AuTraceErr(err); + if (!err) { + p = _rev; + rev = *p; + rev->already = already; + goto out; /* success */ + } + + /* revert */ + if (!already) + au_dr_hino_del(dr, ent); + au_kfree_rcu(ent); + +out: + AuTraceErr(err); + return err; +} + +void au_dr_rename_fin(struct dentry *src, aufs_bindex_t btgt, void *_rev) +{ + struct au_drinfo_rev *rev; + struct au_drinfo_rev_elm *elm; + int nelm; + + rev = _rev; + elm = rev->elm; + for (nelm = rev->nelm; nelm > 0; nelm--, elm++) { + dput(elm->info_dentry); + au_kfree_rcu(elm->info_last); + } + au_kfree_try_rcu(rev); +} + +void au_dr_rename_rev(struct dentry *src, aufs_bindex_t btgt, void *_rev) +{ + int err; + struct au_drinfo_store work; + struct au_drinfo_rev *rev = _rev; + struct super_block *sb; + struct au_branch *br; + struct inode *h_inode; + struct au_dr_br *dr; + struct au_dr_hino *ent; + + err = au_drinfo_store_work_init(&work, btgt); + if (unlikely(err)) + goto out; + + sb = src->d_sb; + br = au_sbr(sb, btgt); + work.h_ppath.dentry = au_h_dptr(src, btgt); + work.h_ppath.mnt = au_br_mnt(br); + au_drinfo_store_rev(rev, &work); + au_drinfo_store_work_fin(&work); + if (rev->already) + goto out; + + dr = &br->br_dirren; + h_inode = d_inode(work.h_ppath.dentry); + ent = au_dr_hino_find(dr, h_inode->i_ino); + BUG_ON(!ent); + au_dr_hino_del(dr, ent); + au_kfree_rcu(ent); + +out: + au_kfree_try_rcu(rev); + if (unlikely(err)) + pr_err("failed to remove dirren info\n"); +} + +/* ---------------------------------------------------------------------- */ + +static struct au_drinfo *au_drinfo_do_load(struct path *h_ppath, + char *whname, int whnamelen, + struct dentry **info_dentry) +{ + struct au_drinfo *drinfo; + struct file *f; + struct inode *h_dir; + struct path infopath; + int unlocked; + + AuDbg("%pd/%.*s\n", h_ppath->dentry, whnamelen, whname); + + *info_dentry = NULL; + drinfo = NULL; + unlocked = 0; + h_dir = d_inode(h_ppath->dentry); + inode_lock_shared_nested(h_dir, AuLsc_I_PARENT); + infopath.dentry = vfsub_lookup_one_len(whname, h_ppath->dentry, + whnamelen); + if (IS_ERR(infopath.dentry)) { + drinfo = (void *)infopath.dentry; + goto out; + } + + if (d_is_negative(infopath.dentry)) + goto out_dput; /* success */ + + infopath.mnt = h_ppath->mnt; + f = vfsub_dentry_open(&infopath, O_RDONLY); + inode_unlock_shared(h_dir); + unlocked = 1; + if (IS_ERR(f)) { + drinfo = (void *)f; + goto out_dput; + } + + drinfo = au_drinfo_read_k(f, /*h_ino*/0); + if (IS_ERR_OR_NULL(drinfo)) + goto out_fput; + + AuDbg("oldname %.*s\n", drinfo->oldnamelen, drinfo->oldname); + *info_dentry = dget(infopath.dentry); /* keep it alive */ + +out_fput: + fput(f); +out_dput: + dput(infopath.dentry); +out: + if (!unlocked) + inode_unlock_shared(h_dir); + AuTraceErrPtr(drinfo); + return drinfo; +} + +struct au_drinfo_do_load_args { + struct au_drinfo **drinfop; + struct path *h_ppath; + char *whname; + int whnamelen; + struct dentry **info_dentry; +}; + +static void au_call_drinfo_do_load(void *args) +{ + struct au_drinfo_do_load_args *a = args; + + *a->drinfop = au_drinfo_do_load(a->h_ppath, a->whname, a->whnamelen, + a->info_dentry); +} + +struct au_drinfo_load { + struct path h_ppath; + struct qstr *qname; + unsigned char no_sio; + + aufs_bindex_t ninfo; + struct au_drinfo **drinfo; +}; + +static int au_drinfo_load(struct au_drinfo_load *w, aufs_bindex_t bindex, + struct au_branch *br) +{ + int err, wkq_err, whnamelen, e; + char whname[sizeof(AUFS_WH_DR_INFO_PFX) + AUFS_DIRREN_ENV_VAL_SZ] + = AUFS_WH_DR_INFO_PFX; + struct au_drinfo *drinfo; + struct qstr oldname; + struct inode *h_dir, *delegated; + struct dentry *info_dentry; + struct path infopath; + + whnamelen = sizeof(AUFS_WH_DR_INFO_PFX) - 1; + whnamelen += au_drinfo_name(br, whname + whnamelen, + sizeof(whname) - whnamelen); + if (w->no_sio) + drinfo = au_drinfo_do_load(&w->h_ppath, whname, whnamelen, + &info_dentry); + else { + struct au_drinfo_do_load_args args = { + .drinfop = &drinfo, + .h_ppath = &w->h_ppath, + .whname = whname, + .whnamelen = whnamelen, + .info_dentry = &info_dentry + }; + wkq_err = au_wkq_wait(au_call_drinfo_do_load, &args); + if (unlikely(wkq_err)) + drinfo = ERR_PTR(wkq_err); + } + err = PTR_ERR(drinfo); + if (IS_ERR_OR_NULL(drinfo)) + goto out; + + err = 0; + oldname.len = drinfo->oldnamelen; + oldname.name = drinfo->oldname; + if (au_qstreq(w->qname, &oldname)) { + /* the name is renamed back */ + au_kfree_rcu(drinfo); + drinfo = NULL; + + infopath.dentry = info_dentry; + infopath.mnt = w->h_ppath.mnt; + h_dir = d_inode(w->h_ppath.dentry); + delegated = NULL; + inode_lock_nested(h_dir, AuLsc_I_PARENT); + e = vfsub_unlink(h_dir, &infopath, &delegated, !w->no_sio); + inode_unlock(h_dir); + if (unlikely(e)) + AuIOErr("ignored %d, %pd2\n", e, &infopath.dentry); + if (unlikely(e == -EWOULDBLOCK)) + iput(delegated); + } + au_kfree_rcu(w->drinfo[bindex]); + w->drinfo[bindex] = drinfo; + dput(info_dentry); + +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static void au_dr_lkup_free(struct au_drinfo **drinfo, int n) +{ + struct au_drinfo **p = drinfo; + + while (n-- > 0) + au_kfree_rcu(*drinfo++); + au_kfree_try_rcu(p); +} + +int au_dr_lkup(struct au_do_lookup_args *lkup, struct dentry *dentry, + aufs_bindex_t btgt) +{ + int err, ninfo; + struct au_drinfo_load w; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + struct inode *h_dir; + struct au_dr_hino *ent; + struct super_block *sb; + + AuDbg("%.*s, name %.*s, whname %.*s, b%d\n", + AuLNPair(&dentry->d_name), AuLNPair(&lkup->dirren.dr_name), + AuLNPair(&lkup->whname), btgt); + + sb = dentry->d_sb; + bbot = au_sbbot(sb); + w.ninfo = bbot + 1; + if (!lkup->dirren.drinfo) { + lkup->dirren.drinfo = kcalloc(w.ninfo, + sizeof(*lkup->dirren.drinfo), + GFP_NOFS); + if (unlikely(!lkup->dirren.drinfo)) { + err = -ENOMEM; + goto out; + } + lkup->dirren.ninfo = w.ninfo; + } + w.drinfo = lkup->dirren.drinfo; + w.no_sio = !!uid_eq(current_fsuid(), GLOBAL_ROOT_UID); + w.h_ppath.dentry = au_h_dptr(dentry, btgt); + AuDebugOn(!w.h_ppath.dentry); + w.h_ppath.mnt = au_sbr_mnt(sb, btgt); + w.qname = &dentry->d_name; + + ninfo = 0; + for (bindex = btgt + 1; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + err = au_drinfo_load(&w, bindex, br); + if (unlikely(err)) + goto out_free; + if (w.drinfo[bindex]) + ninfo++; + } + if (!ninfo) { + br = au_sbr(sb, btgt); + h_dir = d_inode(w.h_ppath.dentry); + ent = au_dr_hino_find(&br->br_dirren, h_dir->i_ino); + AuDebugOn(!ent); + au_dr_hino_del(&br->br_dirren, ent); + au_kfree_rcu(ent); + } + goto out; /* success */ + +out_free: + au_dr_lkup_free(lkup->dirren.drinfo, lkup->dirren.ninfo); + lkup->dirren.ninfo = 0; + lkup->dirren.drinfo = NULL; +out: + AuTraceErr(err); + return err; +} + +void au_dr_lkup_fin(struct au_do_lookup_args *lkup) +{ + au_dr_lkup_free(lkup->dirren.drinfo, lkup->dirren.ninfo); +} + +int au_dr_lkup_name(struct au_do_lookup_args *lkup, aufs_bindex_t btgt) +{ + int err; + struct au_drinfo *drinfo; + + err = 0; + if (!lkup->dirren.drinfo) + goto out; + AuDebugOn(lkup->dirren.ninfo <= btgt); + drinfo = lkup->dirren.drinfo[btgt]; + if (!drinfo) + goto out; + + au_kfree_try_rcu(lkup->whname.name); + lkup->whname.name = NULL; + lkup->dirren.dr_name.len = drinfo->oldnamelen; + lkup->dirren.dr_name.name = drinfo->oldname; + lkup->name = &lkup->dirren.dr_name; + err = au_wh_name_alloc(&lkup->whname, lkup->name); + if (!err) + AuDbg("name %.*s, whname %.*s, b%d\n", + AuLNPair(lkup->name), AuLNPair(&lkup->whname), + btgt); + +out: + AuTraceErr(err); + return err; +} + +int au_dr_lkup_h_ino(struct au_do_lookup_args *lkup, aufs_bindex_t bindex, + ino_t h_ino) +{ + int match; + struct au_drinfo *drinfo; + + match = 1; + if (!lkup->dirren.drinfo) + goto out; + AuDebugOn(lkup->dirren.ninfo <= bindex); + drinfo = lkup->dirren.drinfo[bindex]; + if (!drinfo) + goto out; + + match = (drinfo->ino == h_ino); + AuDbg("match %d\n", match); + +out: + return match; +} + +/* ---------------------------------------------------------------------- */ + +int au_dr_opt_set(struct super_block *sb) +{ + int err; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + err = 0; + bbot = au_sbbot(sb); + for (bindex = 0; !err && bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + err = au_dr_hino(sb, bindex, /*br*/NULL, &br->br_path); + } + + return err; +} + +int au_dr_opt_flush(struct super_block *sb) +{ + int err; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + err = 0; + bbot = au_sbbot(sb); + for (bindex = 0; !err && bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_writable(br->br_perm)) + err = au_dr_hino(sb, bindex, /*br*/NULL, /*path*/NULL); + } + + return err; +} + +int au_dr_opt_clr(struct super_block *sb, int no_flush) +{ + int err; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + err = 0; + if (!no_flush) { + err = au_dr_opt_flush(sb); + if (unlikely(err)) + goto out; + } + + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + au_dr_hino_free(&br->br_dirren); + } + +out: + return err; +} diff --git a/fs/aufs/dirren.h b/fs/aufs/dirren.h new file mode 100644 index 00000000000..1fbc8fb20de --- /dev/null +++ b/fs/aufs/dirren.h @@ -0,0 +1,140 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2017-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * renamed dir info + */ + +#ifndef __AUFS_DIRREN_H__ +#define __AUFS_DIRREN_H__ + +#ifdef __KERNEL__ + +#include +#include +#include +#include "hbl.h" + +#define AuDirren_NHASH 100 + +#ifdef CONFIG_AUFS_DIRREN +enum au_brid_type { + AuBrid_Unset, + AuBrid_UUID, + AuBrid_FSID, + AuBrid_DEV +}; + +struct au_dr_brid { + enum au_brid_type type; + union { + uuid_t uuid; /* unimplemented yet */ + fsid_t fsid; + dev_t dev; + }; +}; + +/* 20 is the max digits length of ulong 64 */ +/* brid-type "_" uuid "_" inum */ +#define AUFS_DIRREN_FNAME_SZ (1 + 1 + UUID_STRING_LEN + 20) +#define AUFS_DIRREN_ENV_VAL_SZ (AUFS_DIRREN_FNAME_SZ + 1 + 20) + +struct au_dr_hino { + struct hlist_bl_node dr_hnode; + ino_t dr_h_ino; +}; + +struct au_dr_br { + struct hlist_bl_head dr_h_ino[AuDirren_NHASH]; + struct au_dr_brid dr_brid; +}; + +struct au_dr_lookup { + /* dr_name is pointed by struct au_do_lookup_args.name */ + struct qstr dr_name; /* subset of dr_info */ + aufs_bindex_t ninfo; + struct au_drinfo **drinfo; +}; +#else +struct au_dr_hino; +/* empty */ +struct au_dr_br { }; +struct au_dr_lookup { }; +#endif + +/* ---------------------------------------------------------------------- */ + +struct au_branch; +struct au_do_lookup_args; +struct au_hinode; +#ifdef CONFIG_AUFS_DIRREN +int au_dr_hino_test_add(struct au_dr_br *dr, ino_t h_ino, + struct au_dr_hino *add_ent); +void au_dr_hino_free(struct au_dr_br *dr); +int au_dr_br_init(struct super_block *sb, struct au_branch *br, + const struct path *path); +int au_dr_br_fin(struct super_block *sb, struct au_branch *br); +int au_dr_rename(struct dentry *src, aufs_bindex_t bindex, + struct qstr *dst_name, void *_rev); +void au_dr_rename_fin(struct dentry *src, aufs_bindex_t btgt, void *rev); +void au_dr_rename_rev(struct dentry *src, aufs_bindex_t bindex, void *rev); +int au_dr_lkup(struct au_do_lookup_args *lkup, struct dentry *dentry, + aufs_bindex_t bindex); +int au_dr_lkup_name(struct au_do_lookup_args *lkup, aufs_bindex_t btgt); +int au_dr_lkup_h_ino(struct au_do_lookup_args *lkup, aufs_bindex_t bindex, + ino_t h_ino); +void au_dr_lkup_fin(struct au_do_lookup_args *lkup); +int au_dr_opt_set(struct super_block *sb); +int au_dr_opt_flush(struct super_block *sb); +int au_dr_opt_clr(struct super_block *sb, int no_flush); +#else +AuStubInt0(au_dr_hino_test_add, struct au_dr_br *dr, ino_t h_ino, + struct au_dr_hino *add_ent); +AuStubVoid(au_dr_hino_free, struct au_dr_br *dr); +AuStubInt0(au_dr_br_init, struct super_block *sb, struct au_branch *br, + const struct path *path); +AuStubInt0(au_dr_br_fin, struct super_block *sb, struct au_branch *br); +AuStubInt0(au_dr_rename, struct dentry *src, aufs_bindex_t bindex, + struct qstr *dst_name, void *_rev); +AuStubVoid(au_dr_rename_fin, struct dentry *src, aufs_bindex_t btgt, void *rev); +AuStubVoid(au_dr_rename_rev, struct dentry *src, aufs_bindex_t bindex, + void *rev); +AuStubInt0(au_dr_lkup, struct au_do_lookup_args *lkup, struct dentry *dentry, + aufs_bindex_t bindex); +AuStubInt0(au_dr_lkup_name, struct au_do_lookup_args *lkup, aufs_bindex_t btgt); +AuStubInt0(au_dr_lkup_h_ino, struct au_do_lookup_args *lkup, + aufs_bindex_t bindex, ino_t h_ino); +AuStubVoid(au_dr_lkup_fin, struct au_do_lookup_args *lkup); +AuStubInt0(au_dr_opt_set, struct super_block *sb); +AuStubInt0(au_dr_opt_flush, struct super_block *sb); +AuStubInt0(au_dr_opt_clr, struct super_block *sb, int no_flush); +#endif + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_DIRREN +static inline int au_dr_ihash(ino_t h_ino) +{ + return h_ino % AuDirren_NHASH; +} +#else +AuStubInt0(au_dr_ihash, ino_t h_ino); +#endif + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DIRREN_H__ */ diff --git a/fs/aufs/dynop.c b/fs/aufs/dynop.c new file mode 100644 index 00000000000..4732edb340d --- /dev/null +++ b/fs/aufs/dynop.c @@ -0,0 +1,368 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2010-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * dynamically customizable operations for regular files + */ + +#include "aufs.h" + +#define DyPrSym(key) AuDbgSym(key->dk_op.dy_hop) + +/* + * How large will these lists be? + * Usually just a few elements, 20-30 at most for each, I guess. + */ +static struct hlist_bl_head dynop[AuDyLast]; + +static struct au_dykey *dy_gfind_get(struct hlist_bl_head *hbl, + const void *h_op) +{ + struct au_dykey *key, *tmp; + struct hlist_bl_node *pos; + + key = NULL; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(tmp, pos, hbl, dk_hnode) + if (tmp->dk_op.dy_hop == h_op) { + if (kref_get_unless_zero(&tmp->dk_kref)) + key = tmp; + break; + } + hlist_bl_unlock(hbl); + + return key; +} + +static struct au_dykey *dy_bradd(struct au_branch *br, struct au_dykey *key) +{ + struct au_dykey **k, *found; + const void *h_op = key->dk_op.dy_hop; + int i; + + found = NULL; + k = br->br_dykey; + for (i = 0; i < AuBrDynOp; i++) + if (k[i]) { + if (k[i]->dk_op.dy_hop == h_op) { + found = k[i]; + break; + } + } else + break; + if (!found) { + spin_lock(&br->br_dykey_lock); + for (; i < AuBrDynOp; i++) + if (k[i]) { + if (k[i]->dk_op.dy_hop == h_op) { + found = k[i]; + break; + } + } else { + k[i] = key; + break; + } + spin_unlock(&br->br_dykey_lock); + BUG_ON(i == AuBrDynOp); /* expand the array */ + } + + return found; +} + +/* kref_get() if @key is already added */ +static struct au_dykey *dy_gadd(struct hlist_bl_head *hbl, struct au_dykey *key) +{ + struct au_dykey *tmp, *found; + struct hlist_bl_node *pos; + const void *h_op = key->dk_op.dy_hop; + + found = NULL; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(tmp, pos, hbl, dk_hnode) + if (tmp->dk_op.dy_hop == h_op) { + if (kref_get_unless_zero(&tmp->dk_kref)) + found = tmp; + break; + } + if (!found) + hlist_bl_add_head(&key->dk_hnode, hbl); + hlist_bl_unlock(hbl); + + if (!found) + DyPrSym(key); + return found; +} + +static void dy_free_rcu(struct rcu_head *rcu) +{ + struct au_dykey *key; + + key = container_of(rcu, struct au_dykey, dk_rcu); + DyPrSym(key); + kfree(key); +} + +static void dy_free(struct kref *kref) +{ + struct au_dykey *key; + struct hlist_bl_head *hbl; + + key = container_of(kref, struct au_dykey, dk_kref); + hbl = dynop + key->dk_op.dy_type; + au_hbl_del(&key->dk_hnode, hbl); + call_rcu(&key->dk_rcu, dy_free_rcu); +} + +void au_dy_put(struct au_dykey *key) +{ + kref_put(&key->dk_kref, dy_free); +} + +/* ---------------------------------------------------------------------- */ + +#define DyDbgSize(cnt, op) AuDebugOn(cnt != sizeof(op)/sizeof(void *)) + +#ifdef CONFIG_AUFS_DEBUG +#define DyDbgDeclare(cnt) unsigned int cnt = 0 +#define DyDbgInc(cnt) do { cnt++; } while (0) +#else +#define DyDbgDeclare(cnt) do {} while (0) +#define DyDbgInc(cnt) do {} while (0) +#endif + +#define DySet(func, dst, src, h_op, h_sb) do { \ + DyDbgInc(cnt); \ + if (h_op->func) { \ + if (src.func) \ + dst.func = src.func; \ + else \ + AuDbg("%s %s\n", au_sbtype(h_sb), #func); \ + } \ +} while (0) + +#define DySetForce(func, dst, src) do { \ + AuDebugOn(!src.func); \ + DyDbgInc(cnt); \ + dst.func = src.func; \ +} while (0) + +#define DySetAop(func) \ + DySet(func, dyaop->da_op, aufs_aop, h_aop, h_sb) +#define DySetAopForce(func) \ + DySetForce(func, dyaop->da_op, aufs_aop) + +static void dy_aop(struct au_dykey *key, const void *h_op, + struct super_block *h_sb __maybe_unused) +{ + struct au_dyaop *dyaop = (void *)key; + const struct address_space_operations *h_aop = h_op; + DyDbgDeclare(cnt); + + AuDbg("%s\n", au_sbtype(h_sb)); + + DySetAop(writepage); + DySetAopForce(readpage); /* force */ + DySetAop(writepages); + DySetAop(set_page_dirty); + DySetAop(readpages); + DySetAop(readahead); + DySetAop(write_begin); + DySetAop(write_end); + DySetAop(bmap); + DySetAop(invalidatepage); + DySetAop(releasepage); + DySetAop(freepage); + /* this one will be changed according to an aufs mount option */ + DySetAop(direct_IO); + DySetAop(migratepage); + DySetAop(isolate_page); + DySetAop(putback_page); + DySetAop(launder_page); + DySetAop(is_partially_uptodate); + DySetAop(is_dirty_writeback); + DySetAop(error_remove_page); + DySetAop(swap_activate); + DySetAop(swap_deactivate); + + DyDbgSize(cnt, *h_aop); +} + +/* ---------------------------------------------------------------------- */ + +static void dy_bug(struct kref *kref) +{ + BUG(); +} + +static struct au_dykey *dy_get(struct au_dynop *op, struct au_branch *br) +{ + struct au_dykey *key, *old; + struct hlist_bl_head *hbl; + struct op { + unsigned int sz; + void (*set)(struct au_dykey *key, const void *h_op, + struct super_block *h_sb __maybe_unused); + }; + static const struct op a[] = { + [AuDy_AOP] = { + .sz = sizeof(struct au_dyaop), + .set = dy_aop + } + }; + const struct op *p; + + hbl = dynop + op->dy_type; + key = dy_gfind_get(hbl, op->dy_hop); + if (key) + goto out_add; /* success */ + + p = a + op->dy_type; + key = kzalloc(p->sz, GFP_NOFS); + if (unlikely(!key)) { + key = ERR_PTR(-ENOMEM); + goto out; + } + + key->dk_op.dy_hop = op->dy_hop; + kref_init(&key->dk_kref); + p->set(key, op->dy_hop, au_br_sb(br)); + old = dy_gadd(hbl, key); + if (old) { + au_kfree_rcu(key); + key = old; + } + +out_add: + old = dy_bradd(br, key); + if (old) + /* its ref-count should never be zero here */ + kref_put(&key->dk_kref, dy_bug); +out: + return key; +} + +/* ---------------------------------------------------------------------- */ +/* + * Aufs prohibits O_DIRECT by default even if the branch supports it. + * This behaviour is necessary to return an error from open(O_DIRECT) instead + * of the succeeding I/O. The dio mount option enables O_DIRECT and makes + * open(O_DIRECT) always succeed, but the succeeding I/O may return an error. + * See the aufs manual in detail. + */ +static void dy_adx(struct au_dyaop *dyaop, int do_dx) +{ + if (!do_dx) + dyaop->da_op.direct_IO = NULL; + else + dyaop->da_op.direct_IO = aufs_aop.direct_IO; +} + +static struct au_dyaop *dy_aget(struct au_branch *br, + const struct address_space_operations *h_aop, + int do_dx) +{ + struct au_dyaop *dyaop; + struct au_dynop op; + + op.dy_type = AuDy_AOP; + op.dy_haop = h_aop; + dyaop = (void *)dy_get(&op, br); + if (IS_ERR(dyaop)) + goto out; + dy_adx(dyaop, do_dx); + +out: + return dyaop; +} + +int au_dy_iaop(struct inode *inode, aufs_bindex_t bindex, + struct inode *h_inode) +{ + int err, do_dx; + struct super_block *sb; + struct au_branch *br; + struct au_dyaop *dyaop; + + AuDebugOn(!S_ISREG(h_inode->i_mode)); + IiMustWriteLock(inode); + + sb = inode->i_sb; + br = au_sbr(sb, bindex); + do_dx = !!au_opt_test(au_mntflags(sb), DIO); + dyaop = dy_aget(br, h_inode->i_mapping->a_ops, do_dx); + err = PTR_ERR(dyaop); + if (IS_ERR(dyaop)) + /* unnecessary to call dy_fput() */ + goto out; + + err = 0; + inode->i_mapping->a_ops = &dyaop->da_op; + +out: + return err; +} + +/* + * Is it safe to replace a_ops during the inode/file is in operation? + * Yes, I hope so. + */ +int au_dy_irefresh(struct inode *inode) +{ + int err; + aufs_bindex_t btop; + struct inode *h_inode; + + err = 0; + if (S_ISREG(inode->i_mode)) { + btop = au_ibtop(inode); + h_inode = au_h_iptr(inode, btop); + err = au_dy_iaop(inode, btop, h_inode); + } + return err; +} + +void au_dy_arefresh(int do_dx) +{ + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_dykey *key; + + hbl = dynop + AuDy_AOP; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(key, pos, hbl, dk_hnode) + dy_adx((void *)key, do_dx); + hlist_bl_unlock(hbl); +} + +/* ---------------------------------------------------------------------- */ + +void __init au_dy_init(void) +{ + int i; + + for (i = 0; i < AuDyLast; i++) + INIT_HLIST_BL_HEAD(dynop + i); +} + +void au_dy_fin(void) +{ + int i; + + for (i = 0; i < AuDyLast; i++) + WARN_ON(!hlist_bl_empty(dynop + i)); +} diff --git a/fs/aufs/dynop.h b/fs/aufs/dynop.h new file mode 100644 index 00000000000..c0c7a5485ad --- /dev/null +++ b/fs/aufs/dynop.h @@ -0,0 +1,77 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2010-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * dynamically customizable operations (for regular files only) + */ + +#ifndef __AUFS_DYNOP_H__ +#define __AUFS_DYNOP_H__ + +#ifdef __KERNEL__ + +#include +#include + +enum {AuDy_AOP, AuDyLast}; + +struct au_dynop { + int dy_type; + union { + const void *dy_hop; + const struct address_space_operations *dy_haop; + }; +}; + +struct au_dykey { + union { + struct hlist_bl_node dk_hnode; + struct rcu_head dk_rcu; + }; + struct au_dynop dk_op; + + /* + * during I am in the branch local array, kref is gotten. when the + * branch is removed, kref is put. + */ + struct kref dk_kref; +}; + +/* stop unioning since their sizes are very different from each other */ +struct au_dyaop { + struct au_dykey da_key; + struct address_space_operations da_op; /* not const */ +}; +/* make sure that 'struct au_dykey *' can be any type */ +static_assert(!offsetof(struct au_dyaop, da_key)); + +/* ---------------------------------------------------------------------- */ + +/* dynop.c */ +struct au_branch; +void au_dy_put(struct au_dykey *key); +int au_dy_iaop(struct inode *inode, aufs_bindex_t bindex, + struct inode *h_inode); +int au_dy_irefresh(struct inode *inode); +void au_dy_arefresh(int do_dio); + +void __init au_dy_init(void); +void au_dy_fin(void); + +#endif /* __KERNEL__ */ +#endif /* __AUFS_DYNOP_H__ */ diff --git a/fs/aufs/export.c b/fs/aufs/export.c new file mode 100644 index 00000000000..f883d2bf532 --- /dev/null +++ b/fs/aufs/export.c @@ -0,0 +1,837 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * export via nfs + */ + +#include +#include +#include +#include +#include +#include +#include "aufs.h" + +union conv { +#ifdef CONFIG_AUFS_INO_T_64 + __u32 a[2]; +#else + __u32 a[1]; +#endif + ino_t ino; +}; + +static ino_t decode_ino(__u32 *a) +{ + union conv u; + + BUILD_BUG_ON(sizeof(u.ino) != sizeof(u.a)); + u.a[0] = a[0]; +#ifdef CONFIG_AUFS_INO_T_64 + u.a[1] = a[1]; +#endif + return u.ino; +} + +static void encode_ino(__u32 *a, ino_t ino) +{ + union conv u; + + u.ino = ino; + a[0] = u.a[0]; +#ifdef CONFIG_AUFS_INO_T_64 + a[1] = u.a[1]; +#endif +} + +/* NFS file handle */ +enum { + Fh_br_id, + Fh_sigen, +#ifdef CONFIG_AUFS_INO_T_64 + /* support 64bit inode number */ + Fh_ino1, + Fh_ino2, + Fh_dir_ino1, + Fh_dir_ino2, +#else + Fh_ino1, + Fh_dir_ino1, +#endif + Fh_igen, + Fh_h_type, + Fh_tail, + + Fh_ino = Fh_ino1, + Fh_dir_ino = Fh_dir_ino1 +}; + +static int au_test_anon(struct dentry *dentry) +{ + /* note: read d_flags without d_lock */ + return !!(dentry->d_flags & DCACHE_DISCONNECTED); +} + +int au_test_nfsd(void) +{ + int ret; + struct task_struct *tsk = current; + char comm[sizeof(tsk->comm)]; + + ret = 0; + if (tsk->flags & PF_KTHREAD) { + get_task_comm(comm, tsk); + ret = !strcmp(comm, "nfsd"); + } + + return ret; +} + +/* ---------------------------------------------------------------------- */ +/* inode generation external table */ + +void au_xigen_inc(struct inode *inode) +{ + loff_t pos; + ssize_t sz; + __u32 igen; + struct super_block *sb; + struct au_sbinfo *sbinfo; + + sb = inode->i_sb; + AuDebugOn(!au_opt_test(au_mntflags(sb), XINO)); + + sbinfo = au_sbi(sb); + pos = inode->i_ino; + pos *= sizeof(igen); + igen = inode->i_generation + 1; + sz = xino_fwrite(sbinfo->si_xigen, &igen, sizeof(igen), &pos); + if (sz == sizeof(igen)) + return; /* success */ + + if (unlikely(sz >= 0)) + AuIOErr("xigen error (%zd)\n", sz); +} + +int au_xigen_new(struct inode *inode) +{ + int err; + loff_t pos; + ssize_t sz; + struct super_block *sb; + struct au_sbinfo *sbinfo; + struct file *file; + + err = 0; + /* todo: dirty, at mount time */ + if (inode->i_ino == AUFS_ROOT_INO) + goto out; + sb = inode->i_sb; + SiMustAnyLock(sb); + if (unlikely(!au_opt_test(au_mntflags(sb), XINO))) + goto out; + + err = -EFBIG; + pos = inode->i_ino; + if (unlikely(au_loff_max / sizeof(inode->i_generation) - 1 < pos)) { + AuIOErr1("too large i%lld\n", pos); + goto out; + } + pos *= sizeof(inode->i_generation); + + err = 0; + sbinfo = au_sbi(sb); + file = sbinfo->si_xigen; + BUG_ON(!file); + + if (vfsub_f_size_read(file) + < pos + sizeof(inode->i_generation)) { + inode->i_generation = atomic_inc_return(&sbinfo->si_xigen_next); + sz = xino_fwrite(file, &inode->i_generation, + sizeof(inode->i_generation), &pos); + } else + sz = xino_fread(file, &inode->i_generation, + sizeof(inode->i_generation), &pos); + if (sz == sizeof(inode->i_generation)) + goto out; /* success */ + + err = sz; + if (unlikely(sz >= 0)) { + err = -EIO; + AuIOErr("xigen error (%zd)\n", sz); + } + +out: + return err; +} + +int au_xigen_set(struct super_block *sb, struct path *path) +{ + int err; + struct au_sbinfo *sbinfo; + struct file *file; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + file = au_xino_create2(sb, path, sbinfo->si_xigen); + err = PTR_ERR(file); + if (IS_ERR(file)) + goto out; + err = 0; + if (sbinfo->si_xigen) + fput(sbinfo->si_xigen); + sbinfo->si_xigen = file; + +out: + AuTraceErr(err); + return err; +} + +void au_xigen_clr(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + if (sbinfo->si_xigen) { + fput(sbinfo->si_xigen); + sbinfo->si_xigen = NULL; + } +} + +/* ---------------------------------------------------------------------- */ + +static struct dentry *decode_by_ino(struct super_block *sb, ino_t ino, + ino_t dir_ino) +{ + struct dentry *dentry, *d; + struct inode *inode; + unsigned int sigen; + + dentry = NULL; + inode = ilookup(sb, ino); + if (!inode) + goto out; + + dentry = ERR_PTR(-ESTALE); + sigen = au_sigen(sb); + if (unlikely(au_is_bad_inode(inode) + || IS_DEADDIR(inode) + || sigen != au_iigen(inode, NULL))) + goto out_iput; + + dentry = NULL; + if (!dir_ino || S_ISDIR(inode->i_mode)) + dentry = d_find_alias(inode); + else { + spin_lock(&inode->i_lock); + hlist_for_each_entry(d, &inode->i_dentry, d_u.d_alias) { + spin_lock(&d->d_lock); + if (!au_test_anon(d) + && d_inode(d->d_parent)->i_ino == dir_ino) { + dentry = dget_dlock(d); + spin_unlock(&d->d_lock); + break; + } + spin_unlock(&d->d_lock); + } + spin_unlock(&inode->i_lock); + } + if (unlikely(dentry && au_digen_test(dentry, sigen))) { + /* need to refresh */ + dput(dentry); + dentry = NULL; + } + +out_iput: + iput(inode); +out: + AuTraceErrPtr(dentry); + return dentry; +} + +/* ---------------------------------------------------------------------- */ + +/* todo: dirty? */ +/* if exportfs_decode_fh() passed vfsmount*, we could be happy */ + +struct au_compare_mnt_args { + /* input */ + struct super_block *sb; + + /* output */ + struct vfsmount *mnt; +}; + +static int au_compare_mnt(struct vfsmount *mnt, void *arg) +{ + struct au_compare_mnt_args *a = arg; + + if (mnt->mnt_sb != a->sb) + return 0; + a->mnt = mntget(mnt); + return 1; +} + +static struct vfsmount *au_mnt_get(struct super_block *sb) +{ + int err; + struct path root; + struct au_compare_mnt_args args = { + .sb = sb + }; + + get_fs_root(current->fs, &root); + rcu_read_lock(); + err = iterate_mounts(au_compare_mnt, &args, root.mnt); + rcu_read_unlock(); + path_put(&root); + AuDebugOn(!err); + AuDebugOn(!args.mnt); + return args.mnt; +} + +struct au_nfsd_si_lock { + unsigned int sigen; + aufs_bindex_t bindex, br_id; + unsigned char force_lock; +}; + +static int si_nfsd_read_lock(struct super_block *sb, + struct au_nfsd_si_lock *nsi_lock) +{ + int err; + aufs_bindex_t bindex; + + si_read_lock(sb, AuLock_FLUSH); + + /* branch id may be wrapped around */ + err = 0; + bindex = au_br_index(sb, nsi_lock->br_id); + if (bindex >= 0 && nsi_lock->sigen + AUFS_BRANCH_MAX > au_sigen(sb)) + goto out; /* success */ + + err = -ESTALE; + bindex = -1; + if (!nsi_lock->force_lock) + si_read_unlock(sb); + +out: + nsi_lock->bindex = bindex; + return err; +} + +struct find_name_by_ino { + struct dir_context ctx; + int called, found; + ino_t ino; + char *name; + int namelen; +}; + +static int +find_name_by_ino(struct dir_context *ctx, const char *name, int namelen, + loff_t offset, u64 ino, unsigned int d_type) +{ + struct find_name_by_ino *a = container_of(ctx, struct find_name_by_ino, + ctx); + + a->called++; + if (a->ino != ino) + return 0; + + memcpy(a->name, name, namelen); + a->namelen = namelen; + a->found = 1; + return 1; +} + +static struct dentry *au_lkup_by_ino(struct path *path, ino_t ino, + struct au_nfsd_si_lock *nsi_lock) +{ + struct dentry *dentry, *parent; + struct file *file; + struct inode *dir; + struct find_name_by_ino arg = { + .ctx = { + .actor = find_name_by_ino + } + }; + int err; + + parent = path->dentry; + if (nsi_lock) + si_read_unlock(parent->d_sb); + file = vfsub_dentry_open(path, au_dir_roflags); + dentry = (void *)file; + if (IS_ERR(file)) + goto out; + + dentry = ERR_PTR(-ENOMEM); + arg.name = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!arg.name)) + goto out_file; + arg.ino = ino; + arg.found = 0; + do { + arg.called = 0; + /* smp_mb(); */ + err = vfsub_iterate_dir(file, &arg.ctx); + } while (!err && !arg.found && arg.called); + dentry = ERR_PTR(err); + if (unlikely(err)) + goto out_name; + /* instead of ENOENT */ + dentry = ERR_PTR(-ESTALE); + if (!arg.found) + goto out_name; + + /* do not call vfsub_lkup_one() */ + dir = d_inode(parent); + dentry = vfsub_lookup_one_len_unlocked(arg.name, parent, arg.namelen); + AuTraceErrPtr(dentry); + if (IS_ERR(dentry)) + goto out_name; + AuDebugOn(au_test_anon(dentry)); + if (unlikely(d_really_is_negative(dentry))) { + dput(dentry); + dentry = ERR_PTR(-ENOENT); + } + +out_name: + free_page((unsigned long)arg.name); +out_file: + fput(file); +out: + if (unlikely(nsi_lock + && si_nfsd_read_lock(parent->d_sb, nsi_lock) < 0)) + if (!IS_ERR(dentry)) { + dput(dentry); + dentry = ERR_PTR(-ESTALE); + } + AuTraceErrPtr(dentry); + return dentry; +} + +static struct dentry *decode_by_dir_ino(struct super_block *sb, ino_t ino, + ino_t dir_ino, + struct au_nfsd_si_lock *nsi_lock) +{ + struct dentry *dentry; + struct path path; + + if (dir_ino != AUFS_ROOT_INO) { + path.dentry = decode_by_ino(sb, dir_ino, 0); + dentry = path.dentry; + if (!path.dentry || IS_ERR(path.dentry)) + goto out; + AuDebugOn(au_test_anon(path.dentry)); + } else + path.dentry = dget(sb->s_root); + + path.mnt = au_mnt_get(sb); + dentry = au_lkup_by_ino(&path, ino, nsi_lock); + path_put(&path); + +out: + AuTraceErrPtr(dentry); + return dentry; +} + +/* ---------------------------------------------------------------------- */ + +static int h_acceptable(void *expv, struct dentry *dentry) +{ + return 1; +} + +static char *au_build_path(struct dentry *h_parent, struct path *h_rootpath, + char *buf, int len, struct super_block *sb) +{ + char *p; + int n; + struct path path; + + p = d_path(h_rootpath, buf, len); + if (IS_ERR(p)) + goto out; + n = strlen(p); + + path.mnt = h_rootpath->mnt; + path.dentry = h_parent; + p = d_path(&path, buf, len); + if (IS_ERR(p)) + goto out; + if (n != 1) + p += n; + + path.mnt = au_mnt_get(sb); + path.dentry = sb->s_root; + p = d_path(&path, buf, len - strlen(p)); + mntput(path.mnt); + if (IS_ERR(p)) + goto out; + if (n != 1) + p[strlen(p)] = '/'; + +out: + AuTraceErrPtr(p); + return p; +} + +static +struct dentry *decode_by_path(struct super_block *sb, ino_t ino, __u32 *fh, + int fh_len, struct au_nfsd_si_lock *nsi_lock) +{ + struct dentry *dentry, *h_parent, *root; + struct super_block *h_sb; + char *pathname, *p; + struct vfsmount *h_mnt; + struct au_branch *br; + int err; + struct path path; + + br = au_sbr(sb, nsi_lock->bindex); + h_mnt = au_br_mnt(br); + h_sb = h_mnt->mnt_sb; + /* todo: call lower fh_to_dentry()? fh_to_parent()? */ + lockdep_off(); + h_parent = exportfs_decode_fh(h_mnt, (void *)(fh + Fh_tail), + fh_len - Fh_tail, fh[Fh_h_type], + h_acceptable, /*context*/NULL); + lockdep_on(); + dentry = h_parent; + if (unlikely(!h_parent || IS_ERR(h_parent))) { + AuWarn1("%s decode_fh failed, %ld\n", + au_sbtype(h_sb), PTR_ERR(h_parent)); + goto out; + } + dentry = NULL; + if (unlikely(au_test_anon(h_parent))) { + AuWarn1("%s decode_fh returned a disconnected dentry\n", + au_sbtype(h_sb)); + goto out_h_parent; + } + + dentry = ERR_PTR(-ENOMEM); + pathname = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!pathname)) + goto out_h_parent; + + root = sb->s_root; + path.mnt = h_mnt; + di_read_lock_parent(root, !AuLock_IR); + path.dentry = au_h_dptr(root, nsi_lock->bindex); + di_read_unlock(root, !AuLock_IR); + p = au_build_path(h_parent, &path, pathname, PAGE_SIZE, sb); + dentry = (void *)p; + if (IS_ERR(p)) + goto out_pathname; + + si_read_unlock(sb); + err = vfsub_kern_path(p, LOOKUP_FOLLOW | LOOKUP_DIRECTORY, &path); + dentry = ERR_PTR(err); + if (unlikely(err)) + goto out_relock; + + dentry = ERR_PTR(-ENOENT); + AuDebugOn(au_test_anon(path.dentry)); + if (unlikely(d_really_is_negative(path.dentry))) + goto out_path; + + if (ino != d_inode(path.dentry)->i_ino) + dentry = au_lkup_by_ino(&path, ino, /*nsi_lock*/NULL); + else + dentry = dget(path.dentry); + +out_path: + path_put(&path); +out_relock: + if (unlikely(si_nfsd_read_lock(sb, nsi_lock) < 0)) + if (!IS_ERR(dentry)) { + dput(dentry); + dentry = ERR_PTR(-ESTALE); + } +out_pathname: + free_page((unsigned long)pathname); +out_h_parent: + dput(h_parent); +out: + AuTraceErrPtr(dentry); + return dentry; +} + +/* ---------------------------------------------------------------------- */ + +static struct dentry * +aufs_fh_to_dentry(struct super_block *sb, struct fid *fid, int fh_len, + int fh_type) +{ + struct dentry *dentry; + __u32 *fh = fid->raw; + struct au_branch *br; + ino_t ino, dir_ino; + struct au_nfsd_si_lock nsi_lock = { + .force_lock = 0 + }; + + dentry = ERR_PTR(-ESTALE); + /* it should never happen, but the file handle is unreliable */ + if (unlikely(fh_len < Fh_tail)) + goto out; + nsi_lock.sigen = fh[Fh_sigen]; + nsi_lock.br_id = fh[Fh_br_id]; + + /* branch id may be wrapped around */ + br = NULL; + if (unlikely(si_nfsd_read_lock(sb, &nsi_lock))) + goto out; + nsi_lock.force_lock = 1; + + /* is this inode still cached? */ + ino = decode_ino(fh + Fh_ino); + /* it should never happen */ + if (unlikely(ino == AUFS_ROOT_INO)) + goto out_unlock; + + dir_ino = decode_ino(fh + Fh_dir_ino); + dentry = decode_by_ino(sb, ino, dir_ino); + if (IS_ERR(dentry)) + goto out_unlock; + if (dentry) + goto accept; + + /* is the parent dir cached? */ + br = au_sbr(sb, nsi_lock.bindex); + au_lcnt_inc(&br->br_nfiles); + dentry = decode_by_dir_ino(sb, ino, dir_ino, &nsi_lock); + if (IS_ERR(dentry)) + goto out_unlock; + if (dentry) + goto accept; + + /* lookup path */ + dentry = decode_by_path(sb, ino, fh, fh_len, &nsi_lock); + if (IS_ERR(dentry)) + goto out_unlock; + if (unlikely(!dentry)) + /* todo?: make it ESTALE */ + goto out_unlock; + +accept: + if (!au_digen_test(dentry, au_sigen(sb)) + && d_inode(dentry)->i_generation == fh[Fh_igen]) + goto out_unlock; /* success */ + + dput(dentry); + dentry = ERR_PTR(-ESTALE); +out_unlock: + if (br) + au_lcnt_dec(&br->br_nfiles); + si_read_unlock(sb); +out: + AuTraceErrPtr(dentry); + return dentry; +} + +#if 0 /* reserved for future use */ +/* support subtreecheck option */ +static struct dentry *aufs_fh_to_parent(struct super_block *sb, struct fid *fid, + int fh_len, int fh_type) +{ + struct dentry *parent; + __u32 *fh = fid->raw; + ino_t dir_ino; + + dir_ino = decode_ino(fh + Fh_dir_ino); + parent = decode_by_ino(sb, dir_ino, 0); + if (IS_ERR(parent)) + goto out; + if (!parent) + parent = decode_by_path(sb, au_br_index(sb, fh[Fh_br_id]), + dir_ino, fh, fh_len); + +out: + AuTraceErrPtr(parent); + return parent; +} +#endif + +/* ---------------------------------------------------------------------- */ + +static int aufs_encode_fh(struct inode *inode, __u32 *fh, int *max_len, + struct inode *dir) +{ + int err; + aufs_bindex_t bindex; + struct super_block *sb, *h_sb; + struct dentry *dentry, *parent, *h_parent; + struct inode *h_dir; + struct au_branch *br; + + err = -ENOSPC; + if (unlikely(*max_len <= Fh_tail)) { + AuWarn1("NFSv2 client (max_len %d)?\n", *max_len); + goto out; + } + + err = FILEID_ROOT; + if (inode->i_ino == AUFS_ROOT_INO) { + AuDebugOn(inode->i_ino != AUFS_ROOT_INO); + goto out; + } + + h_parent = NULL; + sb = inode->i_sb; + err = si_read_lock(sb, AuLock_FLUSH); + if (unlikely(err)) + goto out; + +#ifdef CONFIG_AUFS_DEBUG + if (unlikely(!au_opt_test(au_mntflags(sb), XINO))) + AuWarn1("NFS-exporting requires xino\n"); +#endif + err = -EIO; + parent = NULL; + ii_read_lock_child(inode); + bindex = au_ibtop(inode); + if (!dir) { + dentry = d_find_any_alias(inode); + if (unlikely(!dentry)) + goto out_unlock; + AuDebugOn(au_test_anon(dentry)); + parent = dget_parent(dentry); + dput(dentry); + if (unlikely(!parent)) + goto out_unlock; + if (d_really_is_positive(parent)) + dir = d_inode(parent); + } + + ii_read_lock_parent(dir); + h_dir = au_h_iptr(dir, bindex); + ii_read_unlock(dir); + if (unlikely(!h_dir)) + goto out_parent; + h_parent = d_find_any_alias(h_dir); + if (unlikely(!h_parent)) + goto out_hparent; + + err = -EPERM; + br = au_sbr(sb, bindex); + h_sb = au_br_sb(br); + if (unlikely(!h_sb->s_export_op)) { + AuErr1("%s branch is not exportable\n", au_sbtype(h_sb)); + goto out_hparent; + } + + fh[Fh_br_id] = br->br_id; + fh[Fh_sigen] = au_sigen(sb); + encode_ino(fh + Fh_ino, inode->i_ino); + encode_ino(fh + Fh_dir_ino, dir->i_ino); + fh[Fh_igen] = inode->i_generation; + + *max_len -= Fh_tail; + fh[Fh_h_type] = exportfs_encode_fh(h_parent, (void *)(fh + Fh_tail), + max_len, + /*connectable or subtreecheck*/0); + err = fh[Fh_h_type]; + *max_len += Fh_tail; + /* todo: macros? */ + if (err != FILEID_INVALID) + err = 99; + else + AuWarn1("%s encode_fh failed\n", au_sbtype(h_sb)); + +out_hparent: + dput(h_parent); +out_parent: + dput(parent); +out_unlock: + ii_read_unlock(inode); + si_read_unlock(sb); +out: + if (unlikely(err < 0)) + err = FILEID_INVALID; + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int aufs_commit_metadata(struct inode *inode) +{ + int err; + aufs_bindex_t bindex; + struct super_block *sb; + struct inode *h_inode; + int (*f)(struct inode *inode); + + sb = inode->i_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + ii_write_lock_child(inode); + bindex = au_ibtop(inode); + AuDebugOn(bindex < 0); + h_inode = au_h_iptr(inode, bindex); + + f = h_inode->i_sb->s_export_op->commit_metadata; + if (f) + err = f(h_inode); + else { + struct writeback_control wbc = { + .sync_mode = WB_SYNC_ALL, + .nr_to_write = 0 /* metadata only */ + }; + + err = sync_inode(h_inode, &wbc); + } + + au_cpup_attr_timesizes(inode); + ii_write_unlock(inode); + si_read_unlock(sb); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static struct export_operations aufs_export_op = { + .fh_to_dentry = aufs_fh_to_dentry, + /* .fh_to_parent = aufs_fh_to_parent, */ + .encode_fh = aufs_encode_fh, + .commit_metadata = aufs_commit_metadata +}; + +void au_export_init(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + __u32 u; + + BUILD_BUG_ON_MSG(IS_BUILTIN(CONFIG_AUFS_FS) + && IS_MODULE(CONFIG_EXPORTFS), + AUFS_NAME ": unsupported configuration " + "CONFIG_EXPORTFS=m and CONFIG_AUFS_FS=y"); + + sb->s_export_op = &aufs_export_op; + sbinfo = au_sbi(sb); + sbinfo->si_xigen = NULL; + get_random_bytes(&u, sizeof(u)); + BUILD_BUG_ON(sizeof(u) != sizeof(int)); + atomic_set(&sbinfo->si_xigen_next, u); +} diff --git a/fs/aufs/f_op.c b/fs/aufs/f_op.c new file mode 100644 index 00000000000..f7f9a2f0ac6 --- /dev/null +++ b/fs/aufs/f_op.c @@ -0,0 +1,771 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * file and vm operations + */ + +#include +#include +#include +#include +#include "aufs.h" + +int au_do_open_nondir(struct file *file, int flags, struct file *h_file) +{ + int err; + aufs_bindex_t bindex; + struct dentry *dentry, *h_dentry; + struct au_finfo *finfo; + struct inode *h_inode; + + FiMustWriteLock(file); + + err = 0; + dentry = file->f_path.dentry; + AuDebugOn(IS_ERR_OR_NULL(dentry)); + finfo = au_fi(file); + memset(&finfo->fi_htop, 0, sizeof(finfo->fi_htop)); + atomic_set(&finfo->fi_mmapped, 0); + bindex = au_dbtop(dentry); + if (!h_file) { + h_dentry = au_h_dptr(dentry, bindex); + err = vfsub_test_mntns(file->f_path.mnt, h_dentry->d_sb); + if (unlikely(err)) + goto out; + h_file = au_h_open(dentry, bindex, flags, file, /*force_wr*/0); + if (IS_ERR(h_file)) { + err = PTR_ERR(h_file); + goto out; + } + } else { + h_dentry = h_file->f_path.dentry; + err = vfsub_test_mntns(file->f_path.mnt, h_dentry->d_sb); + if (unlikely(err)) + goto out; + /* br ref is already inc-ed */ + } + + if ((flags & __O_TMPFILE) + && !(flags & O_EXCL)) { + h_inode = file_inode(h_file); + spin_lock(&h_inode->i_lock); + h_inode->i_state |= I_LINKABLE; + spin_unlock(&h_inode->i_lock); + } + au_set_fbtop(file, bindex); + au_set_h_fptr(file, bindex, h_file); + au_update_figen(file); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + +out: + return err; +} + +static int aufs_open_nondir(struct inode *inode __maybe_unused, + struct file *file) +{ + int err; + struct super_block *sb; + struct au_do_open_args args = { + .open = au_do_open_nondir + }; + + AuDbg("%pD, f_flags 0x%x, f_mode 0x%x\n", + file, vfsub_file_flags(file), file->f_mode); + + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH); + err = au_do_open(file, &args); + si_read_unlock(sb); + return err; +} + +int aufs_release_nondir(struct inode *inode __maybe_unused, struct file *file) +{ + struct au_finfo *finfo; + aufs_bindex_t bindex; + + finfo = au_fi(file); + au_hbl_del(&finfo->fi_hlist, + &au_sbi(file->f_path.dentry->d_sb)->si_files); + bindex = finfo->fi_btop; + if (bindex >= 0) + au_set_h_fptr(file, bindex, NULL); + + au_finfo_fin(file); + return 0; +} + +/* ---------------------------------------------------------------------- */ + +static int au_do_flush_nondir(struct file *file, fl_owner_t id) +{ + int err; + struct file *h_file; + + err = 0; + h_file = au_hf_top(file); + if (h_file) + err = vfsub_flush(h_file, id); + return err; +} + +static int aufs_flush_nondir(struct file *file, fl_owner_t id) +{ + return au_do_flush(file, id, au_do_flush_nondir); +} + +/* ---------------------------------------------------------------------- */ +/* + * read and write functions acquire [fdi]_rwsem once, but release before + * mmap_sem. This is because to stop a race condition between mmap(2). + * Releasing these aufs-rwsem should be safe, no branch-management (by keeping + * si_rwsem), no harmful copy-up should happen. Actually copy-up may happen in + * read functions after [fdi]_rwsem are released, but it should be harmless. + */ + +/* Callers should call au_read_post() or fput() in the end */ +struct file *au_read_pre(struct file *file, int keep_fi, unsigned int lsc) +{ + struct file *h_file; + int err; + + err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/0, lsc); + if (!err) { + di_read_unlock(file->f_path.dentry, AuLock_IR); + h_file = au_hf_top(file); + get_file(h_file); + if (!keep_fi) + fi_read_unlock(file); + } else + h_file = ERR_PTR(err); + + return h_file; +} + +static void au_read_post(struct inode *inode, struct file *h_file) +{ + /* update without lock, I don't think it a problem */ + fsstack_copy_attr_atime(inode, file_inode(h_file)); + fput(h_file); +} + +struct au_write_pre { + /* input */ + unsigned int lsc; + + /* output */ + blkcnt_t blks; + aufs_bindex_t btop; +}; + +/* + * return with iinfo is write-locked + * callers should call au_write_post() or iinfo_write_unlock() + fput() in the + * end + */ +static struct file *au_write_pre(struct file *file, int do_ready, + struct au_write_pre *wpre) +{ + struct file *h_file; + struct dentry *dentry; + int err; + unsigned int lsc; + struct au_pin pin; + + lsc = 0; + if (wpre) + lsc = wpre->lsc; + err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1, lsc); + h_file = ERR_PTR(err); + if (unlikely(err)) + goto out; + + dentry = file->f_path.dentry; + if (do_ready) { + err = au_ready_to_write(file, -1, &pin); + if (unlikely(err)) { + h_file = ERR_PTR(err); + di_write_unlock(dentry); + goto out_fi; + } + } + + di_downgrade_lock(dentry, /*flags*/0); + if (wpre) + wpre->btop = au_fbtop(file); + h_file = au_hf_top(file); + get_file(h_file); + if (wpre) + wpre->blks = file_inode(h_file)->i_blocks; + if (do_ready) + au_unpin(&pin); + di_read_unlock(dentry, /*flags*/0); + +out_fi: + fi_write_unlock(file); +out: + return h_file; +} + +static void au_write_post(struct inode *inode, struct file *h_file, + struct au_write_pre *wpre, ssize_t written) +{ + struct inode *h_inode; + + au_cpup_attr_timesizes(inode); + AuDebugOn(au_ibtop(inode) != wpre->btop); + h_inode = file_inode(h_file); + inode->i_mode = h_inode->i_mode; + ii_write_unlock(inode); + /* AuDbg("blks %llu, %llu\n", (u64)blks, (u64)h_inode->i_blocks); */ + if (written > 0) + au_fhsm_wrote(inode->i_sb, wpre->btop, + /*force*/h_inode->i_blocks > wpre->blks); + fput(h_file); +} + +/* + * todo: very ugly + * it locks both of i_mutex and si_rwsem for read in safe. + * if the plink maintenance mode continues forever (that is the problem), + * may loop forever. + */ +static void au_mtx_and_read_lock(struct inode *inode) +{ + int err; + struct super_block *sb = inode->i_sb; + + while (1) { + inode_lock(inode); + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (!err) + break; + inode_unlock(inode); + si_read_lock(sb, AuLock_NOPLMW); + si_read_unlock(sb); + } +} + +static ssize_t au_do_iter(struct file *h_file, int rw, struct kiocb *kio, + struct iov_iter *iov_iter) +{ + ssize_t err; + struct file *file; + ssize_t (*iter)(struct kiocb *, struct iov_iter *); + + err = security_file_permission(h_file, rw); + if (unlikely(err)) + goto out; + + err = -ENOSYS; /* the branch doesn't have its ->(read|write)_iter() */ + iter = NULL; + if (rw == MAY_READ) + iter = h_file->f_op->read_iter; + else if (rw == MAY_WRITE) + iter = h_file->f_op->write_iter; + + file = kio->ki_filp; + kio->ki_filp = h_file; + if (iter) { + lockdep_off(); + err = iter(kio, iov_iter); + lockdep_on(); + } else + /* currently there is no such fs */ + WARN_ON_ONCE(1); + kio->ki_filp = file; + +out: + return err; +} + +static ssize_t aufs_read_iter(struct kiocb *kio, struct iov_iter *iov_iter) +{ + ssize_t err; + struct file *file, *h_file; + struct inode *inode; + struct super_block *sb; + + file = kio->ki_filp; + inode = file_inode(file); + sb = inode->i_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + + h_file = au_read_pre(file, /*keep_fi*/1, /*lsc*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + if (au_test_loopback_kthread()) { + au_warn_loopback(h_file->f_path.dentry->d_sb); + if (file->f_mapping != h_file->f_mapping) { + file->f_mapping = h_file->f_mapping; + smp_mb(); /* unnecessary? */ + } + } + fi_read_unlock(file); + + err = au_do_iter(h_file, MAY_READ, kio, iov_iter); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + au_read_post(inode, h_file); + +out: + si_read_unlock(sb); + return err; +} + +static ssize_t aufs_write_iter(struct kiocb *kio, struct iov_iter *iov_iter) +{ + ssize_t err; + struct au_write_pre wpre; + struct inode *inode; + struct file *file, *h_file; + + file = kio->ki_filp; + inode = file_inode(file); + au_mtx_and_read_lock(inode); + + wpre.lsc = 0; + h_file = au_write_pre(file, /*do_ready*/1, &wpre); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + err = au_do_iter(h_file, MAY_WRITE, kio, iov_iter); + au_write_post(inode, h_file, &wpre, err); + +out: + si_read_unlock(inode->i_sb); + inode_unlock(inode); + return err; +} + +/* + * We may be able to remove aufs_splice_{read,write}() since almost all FSes + * don't have their own .splice_{read,write} implimentations, and they use + * generic_file_splice_read() and iter_file_splice_write() who can act like the + * simple converters to f_op->iter_read() and ->iter_write(). + * But we keep our own implementations because some non-mainlined FSes may have + * their own .splice_{read,write} implimentations and aufs doesn't want to take + * away an opportunity to co-work with aufs from them. + */ +static ssize_t aufs_splice_read(struct file *file, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + ssize_t err; + struct file *h_file; + struct inode *inode; + struct super_block *sb; + + inode = file_inode(file); + sb = inode->i_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + + h_file = au_read_pre(file, /*keep_fi*/0, /*lsc*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + err = vfsub_splice_to(h_file, ppos, pipe, len, flags); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + au_read_post(inode, h_file); + +out: + si_read_unlock(sb); + return err; +} + +static ssize_t +aufs_splice_write(struct pipe_inode_info *pipe, struct file *file, loff_t *ppos, + size_t len, unsigned int flags) +{ + ssize_t err; + struct au_write_pre wpre; + struct inode *inode; + struct file *h_file; + + inode = file_inode(file); + au_mtx_and_read_lock(inode); + + wpre.lsc = 0; + h_file = au_write_pre(file, /*do_ready*/1, &wpre); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + err = vfsub_splice_from(pipe, h_file, ppos, len, flags); + au_write_post(inode, h_file, &wpre, err); + +out: + si_read_unlock(inode->i_sb); + inode_unlock(inode); + return err; +} + +static long aufs_fallocate(struct file *file, int mode, loff_t offset, + loff_t len) +{ + long err; + struct au_write_pre wpre; + struct inode *inode; + struct file *h_file; + + inode = file_inode(file); + au_mtx_and_read_lock(inode); + + wpre.lsc = 0; + h_file = au_write_pre(file, /*do_ready*/1, &wpre); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + lockdep_off(); + err = vfs_fallocate(h_file, mode, offset, len); + lockdep_on(); + au_write_post(inode, h_file, &wpre, /*written*/1); + +out: + si_read_unlock(inode->i_sb); + inode_unlock(inode); + return err; +} + +static ssize_t aufs_copy_file_range(struct file *src, loff_t src_pos, + struct file *dst, loff_t dst_pos, + size_t len, unsigned int flags) +{ + ssize_t err; + struct au_write_pre wpre; + enum { SRC, DST }; + struct { + struct inode *inode; + struct file *h_file; + struct super_block *h_sb; + } a[2]; +#define a_src a[SRC] +#define a_dst a[DST] + + err = -EINVAL; + a_src.inode = file_inode(src); + if (unlikely(!S_ISREG(a_src.inode->i_mode))) + goto out; + a_dst.inode = file_inode(dst); + if (unlikely(!S_ISREG(a_dst.inode->i_mode))) + goto out; + + au_mtx_and_read_lock(a_dst.inode); + /* + * in order to match the order in di_write_lock2_{child,parent}(), + * use f_path.dentry for this comparison. + */ + if (src->f_path.dentry < dst->f_path.dentry) { + a_src.h_file = au_read_pre(src, /*keep_fi*/1, AuLsc_FI_1); + err = PTR_ERR(a_src.h_file); + if (IS_ERR(a_src.h_file)) + goto out_si; + + wpre.lsc = AuLsc_FI_2; + a_dst.h_file = au_write_pre(dst, /*do_ready*/1, &wpre); + err = PTR_ERR(a_dst.h_file); + if (IS_ERR(a_dst.h_file)) { + au_read_post(a_src.inode, a_src.h_file); + goto out_si; + } + } else { + wpre.lsc = AuLsc_FI_1; + a_dst.h_file = au_write_pre(dst, /*do_ready*/1, &wpre); + err = PTR_ERR(a_dst.h_file); + if (IS_ERR(a_dst.h_file)) + goto out_si; + + a_src.h_file = au_read_pre(src, /*keep_fi*/1, AuLsc_FI_2); + err = PTR_ERR(a_src.h_file); + if (IS_ERR(a_src.h_file)) { + au_write_post(a_dst.inode, a_dst.h_file, &wpre, + /*written*/0); + goto out_si; + } + } + + err = -EXDEV; + a_src.h_sb = file_inode(a_src.h_file)->i_sb; + a_dst.h_sb = file_inode(a_dst.h_file)->i_sb; + if (unlikely(a_src.h_sb != a_dst.h_sb)) { + AuDbgFile(src); + AuDbgFile(dst); + goto out_file; + } + + err = vfsub_copy_file_range(a_src.h_file, src_pos, a_dst.h_file, + dst_pos, len, flags); + +out_file: + au_write_post(a_dst.inode, a_dst.h_file, &wpre, err); + fi_read_unlock(src); + au_read_post(a_src.inode, a_src.h_file); +out_si: + si_read_unlock(a_dst.inode->i_sb); + inode_unlock(a_dst.inode); +out: + return err; +#undef a_src +#undef a_dst +} + +/* ---------------------------------------------------------------------- */ + +/* + * The locking order around current->mmap_sem. + * - in most and regular cases + * file I/O syscall -- aufs_read() or something + * -- si_rwsem for read -- mmap_sem + * (Note that [fdi]i_rwsem are released before mmap_sem). + * - in mmap case + * mmap(2) -- mmap_sem -- aufs_mmap() -- si_rwsem for read -- [fdi]i_rwsem + * This AB-BA order is definitely bad, but is not a problem since "si_rwsem for + * read" allows multiple processes to acquire it and [fdi]i_rwsem are not held + * in file I/O. Aufs needs to stop lockdep in aufs_mmap() though. + * It means that when aufs acquires si_rwsem for write, the process should never + * acquire mmap_sem. + * + * Actually aufs_iterate() holds [fdi]i_rwsem before mmap_sem, but this is not a + * problem either since any directory is not able to be mmap-ed. + * The similar scenario is applied to aufs_readlink() too. + */ + +#if 0 /* stop calling security_file_mmap() */ +/* cf. linux/include/linux/mman.h: calc_vm_prot_bits() */ +#define AuConv_VM_PROT(f, b) _calc_vm_trans(f, VM_##b, PROT_##b) + +static unsigned long au_arch_prot_conv(unsigned long flags) +{ + /* currently ppc64 only */ +#ifdef CONFIG_PPC64 + /* cf. linux/arch/powerpc/include/asm/mman.h */ + AuDebugOn(arch_calc_vm_prot_bits(-1) != VM_SAO); + return AuConv_VM_PROT(flags, SAO); +#else + AuDebugOn(arch_calc_vm_prot_bits(-1)); + return 0; +#endif +} + +static unsigned long au_prot_conv(unsigned long flags) +{ + return AuConv_VM_PROT(flags, READ) + | AuConv_VM_PROT(flags, WRITE) + | AuConv_VM_PROT(flags, EXEC) + | au_arch_prot_conv(flags); +} + +/* cf. linux/include/linux/mman.h: calc_vm_flag_bits() */ +#define AuConv_VM_MAP(f, b) _calc_vm_trans(f, VM_##b, MAP_##b) + +static unsigned long au_flag_conv(unsigned long flags) +{ + return AuConv_VM_MAP(flags, GROWSDOWN) + | AuConv_VM_MAP(flags, DENYWRITE) + | AuConv_VM_MAP(flags, LOCKED); +} +#endif + +static int aufs_mmap(struct file *file, struct vm_area_struct *vma) +{ + int err; + const unsigned char wlock + = (file->f_mode & FMODE_WRITE) && (vma->vm_flags & VM_SHARED); + struct super_block *sb; + struct file *h_file; + struct inode *inode; + + AuDbgVmRegion(file, vma); + + inode = file_inode(file); + sb = inode->i_sb; + lockdep_off(); + si_read_lock(sb, AuLock_NOPLMW); + + h_file = au_write_pre(file, wlock, /*wpre*/NULL); + lockdep_on(); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + err = 0; + au_set_mmapped(file); + au_vm_file_reset(vma, h_file); + /* + * we cannot call security_mmap_file() here since it may acquire + * mmap_sem or i_mutex. + * + * err = security_mmap_file(h_file, au_prot_conv(vma->vm_flags), + * au_flag_conv(vma->vm_flags)); + */ + if (!err) + err = call_mmap(h_file, vma); + if (!err) { + au_vm_prfile_set(vma, file); + fsstack_copy_attr_atime(inode, file_inode(h_file)); + goto out_fput; /* success */ + } + au_unset_mmapped(file); + au_vm_file_reset(vma, file); + +out_fput: + lockdep_off(); + ii_write_unlock(inode); + lockdep_on(); + fput(h_file); +out: + lockdep_off(); + si_read_unlock(sb); + lockdep_on(); + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int aufs_fsync_nondir(struct file *file, loff_t start, loff_t end, + int datasync) +{ + int err; + struct au_write_pre wpre; + struct inode *inode; + struct file *h_file; + + err = 0; /* -EBADF; */ /* posix? */ + if (unlikely(!(file->f_mode & FMODE_WRITE))) + goto out; + + inode = file_inode(file); + au_mtx_and_read_lock(inode); + + wpre.lsc = 0; + h_file = au_write_pre(file, /*do_ready*/1, &wpre); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out_unlock; + + err = vfsub_fsync(h_file, &h_file->f_path, datasync); + au_write_post(inode, h_file, &wpre, /*written*/0); + +out_unlock: + si_read_unlock(inode->i_sb); + inode_unlock(inode); +out: + return err; +} + +static int aufs_fasync(int fd, struct file *file, int flag) +{ + int err; + struct file *h_file; + struct super_block *sb; + + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + + h_file = au_read_pre(file, /*keep_fi*/0, /*lsc*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + if (h_file->f_op->fasync) + err = h_file->f_op->fasync(fd, h_file, flag); + fput(h_file); /* instead of au_read_post() */ + +out: + si_read_unlock(sb); + return err; +} + +static int aufs_setfl(struct file *file, unsigned long arg) +{ + int err; + struct file *h_file; + struct super_block *sb; + + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + + h_file = au_read_pre(file, /*keep_fi*/0, /*lsc*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out; + + /* stop calling h_file->fasync */ + arg |= vfsub_file_flags(file) & FASYNC; + err = setfl(/*unused fd*/-1, h_file, arg); + fput(h_file); /* instead of au_read_post() */ + +out: + si_read_unlock(sb); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* no one supports this operation, currently */ +#if 0 /* reserved for future use */ +static ssize_t aufs_sendpage(struct file *file, struct page *page, int offset, + size_t len, loff_t *pos, int more) +{ +} +#endif + +/* ---------------------------------------------------------------------- */ + +const struct file_operations aufs_file_fop = { + .owner = THIS_MODULE, + + .llseek = default_llseek, + + .read_iter = aufs_read_iter, + .write_iter = aufs_write_iter, + +#ifdef CONFIG_AUFS_POLL + .poll = aufs_poll, +#endif + .unlocked_ioctl = aufs_ioctl_nondir, +#ifdef CONFIG_COMPAT + .compat_ioctl = aufs_compat_ioctl_nondir, +#endif + .mmap = aufs_mmap, + .open = aufs_open_nondir, + .flush = aufs_flush_nondir, + .release = aufs_release_nondir, + .fsync = aufs_fsync_nondir, + .fasync = aufs_fasync, + /* .sendpage = aufs_sendpage, */ + .setfl = aufs_setfl, + .splice_write = aufs_splice_write, + .splice_read = aufs_splice_read, +#if 0 /* reserved for future use */ + .aio_splice_write = aufs_aio_splice_write, + .aio_splice_read = aufs_aio_splice_read, +#endif + .fallocate = aufs_fallocate, + .copy_file_range = aufs_copy_file_range +}; diff --git a/fs/aufs/fhsm.c b/fs/aufs/fhsm.c new file mode 100644 index 00000000000..9cef93b4299 --- /dev/null +++ b/fs/aufs/fhsm.c @@ -0,0 +1,427 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2011-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write to the Free Software + * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + */ + +/* + * File-based Hierarchy Storage Management + */ + +#include +#include +#include +#include +#include "aufs.h" + +static aufs_bindex_t au_fhsm_bottom(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + SiMustAnyLock(sb); + + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + AuDebugOn(!fhsm); + return fhsm->fhsm_bottom; +} + +void au_fhsm_set_bottom(struct super_block *sb, aufs_bindex_t bindex) +{ + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + AuDebugOn(!fhsm); + fhsm->fhsm_bottom = bindex; +} + +/* ---------------------------------------------------------------------- */ + +static int au_fhsm_test_jiffy(struct au_sbinfo *sbinfo, struct au_branch *br) +{ + struct au_br_fhsm *bf; + + bf = br->br_fhsm; + MtxMustLock(&bf->bf_lock); + + return !bf->bf_readable + || time_after(jiffies, + bf->bf_jiffy + sbinfo->si_fhsm.fhsm_expire); +} + +/* ---------------------------------------------------------------------- */ + +static void au_fhsm_notify(struct super_block *sb, int val) +{ + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + SiMustAnyLock(sb); + + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + if (au_fhsm_pid(fhsm) + && atomic_read(&fhsm->fhsm_readable) != -1) { + atomic_set(&fhsm->fhsm_readable, val); + if (val) + wake_up(&fhsm->fhsm_wqh); + } +} + +static int au_fhsm_stfs(struct super_block *sb, aufs_bindex_t bindex, + struct aufs_stfs *rstfs, int do_lock, int do_notify) +{ + int err; + struct au_branch *br; + struct au_br_fhsm *bf; + + br = au_sbr(sb, bindex); + AuDebugOn(au_br_rdonly(br)); + bf = br->br_fhsm; + AuDebugOn(!bf); + + if (do_lock) + mutex_lock(&bf->bf_lock); + else + MtxMustLock(&bf->bf_lock); + + /* sb->s_root for NFS is unreliable */ + err = au_br_stfs(br, &bf->bf_stfs); + if (unlikely(err)) { + AuErr1("FHSM failed (%d), b%d, ignored.\n", bindex, err); + goto out; + } + + bf->bf_jiffy = jiffies; + bf->bf_readable = 1; + if (do_notify) + au_fhsm_notify(sb, /*val*/1); + if (rstfs) + *rstfs = bf->bf_stfs; + +out: + if (do_lock) + mutex_unlock(&bf->bf_lock); + au_fhsm_notify(sb, /*val*/1); + + return err; +} + +void au_fhsm_wrote(struct super_block *sb, aufs_bindex_t bindex, int force) +{ + int err; + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + struct au_branch *br; + struct au_br_fhsm *bf; + + AuDbg("b%d, force %d\n", bindex, force); + SiMustAnyLock(sb); + + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + if (!au_ftest_si(sbinfo, FHSM) + || fhsm->fhsm_bottom == bindex) + return; + + br = au_sbr(sb, bindex); + bf = br->br_fhsm; + AuDebugOn(!bf); + mutex_lock(&bf->bf_lock); + if (force + || au_fhsm_pid(fhsm) + || au_fhsm_test_jiffy(sbinfo, br)) + err = au_fhsm_stfs(sb, bindex, /*rstfs*/NULL, /*do_lock*/0, + /*do_notify*/1); + mutex_unlock(&bf->bf_lock); +} + +void au_fhsm_wrote_all(struct super_block *sb, int force) +{ + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + /* exclude the bottom */ + bbot = au_fhsm_bottom(sb); + for (bindex = 0; bindex < bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_fhsm(br->br_perm)) + au_fhsm_wrote(sb, bindex, force); + } +} + +/* ---------------------------------------------------------------------- */ + +static __poll_t au_fhsm_poll(struct file *file, struct poll_table_struct *wait) +{ + __poll_t mask; + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + mask = 0; + sbinfo = file->private_data; + fhsm = &sbinfo->si_fhsm; + poll_wait(file, &fhsm->fhsm_wqh, wait); + if (atomic_read(&fhsm->fhsm_readable)) + mask = EPOLLIN /* | EPOLLRDNORM */; + + if (!mask) + AuDbg("mask 0x%x\n", mask); + return mask; +} + +static int au_fhsm_do_read_one(struct aufs_stbr __user *stbr, + struct aufs_stfs *stfs, __s16 brid) +{ + int err; + + err = copy_to_user(&stbr->stfs, stfs, sizeof(*stfs)); + if (!err) + err = __put_user(brid, &stbr->brid); + if (unlikely(err)) + err = -EFAULT; + + return err; +} + +static ssize_t au_fhsm_do_read(struct super_block *sb, + struct aufs_stbr __user *stbr, size_t count) +{ + ssize_t err; + int nstbr; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + struct au_br_fhsm *bf; + + /* except the bottom branch */ + err = 0; + nstbr = 0; + bbot = au_fhsm_bottom(sb); + for (bindex = 0; !err && bindex < bbot; bindex++) { + br = au_sbr(sb, bindex); + if (!au_br_fhsm(br->br_perm)) + continue; + + bf = br->br_fhsm; + mutex_lock(&bf->bf_lock); + if (bf->bf_readable) { + err = -EFAULT; + if (count >= sizeof(*stbr)) + err = au_fhsm_do_read_one(stbr++, &bf->bf_stfs, + br->br_id); + if (!err) { + bf->bf_readable = 0; + count -= sizeof(*stbr); + nstbr++; + } + } + mutex_unlock(&bf->bf_lock); + } + if (!err) + err = sizeof(*stbr) * nstbr; + + return err; +} + +static ssize_t au_fhsm_read(struct file *file, char __user *buf, size_t count, + loff_t *pos) +{ + ssize_t err; + int readable; + aufs_bindex_t nfhsm, bindex, bbot; + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + struct au_branch *br; + struct super_block *sb; + + err = 0; + sbinfo = file->private_data; + fhsm = &sbinfo->si_fhsm; +need_data: + spin_lock_irq(&fhsm->fhsm_wqh.lock); + if (!atomic_read(&fhsm->fhsm_readable)) { + if (vfsub_file_flags(file) & O_NONBLOCK) + err = -EAGAIN; + else + err = wait_event_interruptible_locked_irq + (fhsm->fhsm_wqh, + atomic_read(&fhsm->fhsm_readable)); + } + spin_unlock_irq(&fhsm->fhsm_wqh.lock); + if (unlikely(err)) + goto out; + + /* sb may already be dead */ + au_rw_read_lock(&sbinfo->si_rwsem); + readable = atomic_read(&fhsm->fhsm_readable); + if (readable > 0) { + sb = sbinfo->si_sb; + AuDebugOn(!sb); + /* exclude the bottom branch */ + nfhsm = 0; + bbot = au_fhsm_bottom(sb); + for (bindex = 0; bindex < bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_fhsm(br->br_perm)) + nfhsm++; + } + err = -EMSGSIZE; + if (nfhsm * sizeof(struct aufs_stbr) <= count) { + atomic_set(&fhsm->fhsm_readable, 0); + err = au_fhsm_do_read(sbinfo->si_sb, (void __user *)buf, + count); + } + } + au_rw_read_unlock(&sbinfo->si_rwsem); + if (!readable) + goto need_data; + +out: + return err; +} + +static int au_fhsm_release(struct inode *inode, struct file *file) +{ + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + /* sb may already be dead */ + sbinfo = file->private_data; + fhsm = &sbinfo->si_fhsm; + spin_lock(&fhsm->fhsm_spin); + fhsm->fhsm_pid = 0; + spin_unlock(&fhsm->fhsm_spin); + kobject_put(&sbinfo->si_kobj); + + return 0; +} + +static const struct file_operations au_fhsm_fops = { + .owner = THIS_MODULE, + .llseek = noop_llseek, + .read = au_fhsm_read, + .poll = au_fhsm_poll, + .release = au_fhsm_release +}; + +int au_fhsm_fd(struct super_block *sb, int oflags) +{ + int err, fd; + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + + err = -EPERM; + if (unlikely(!capable(CAP_SYS_ADMIN))) + goto out; + + err = -EINVAL; + if (unlikely(oflags & ~(O_CLOEXEC | O_NONBLOCK))) + goto out; + + err = 0; + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + spin_lock(&fhsm->fhsm_spin); + if (!fhsm->fhsm_pid) + fhsm->fhsm_pid = current->pid; + else + err = -EBUSY; + spin_unlock(&fhsm->fhsm_spin); + if (unlikely(err)) + goto out; + + oflags |= O_RDONLY; + /* oflags |= FMODE_NONOTIFY; */ + fd = anon_inode_getfd("[aufs_fhsm]", &au_fhsm_fops, sbinfo, oflags); + err = fd; + if (unlikely(fd < 0)) + goto out_pid; + + /* succeed regardless 'fhsm' status */ + kobject_get(&sbinfo->si_kobj); + si_noflush_read_lock(sb); + if (au_ftest_si(sbinfo, FHSM)) + au_fhsm_wrote_all(sb, /*force*/0); + si_read_unlock(sb); + goto out; /* success */ + +out_pid: + spin_lock(&fhsm->fhsm_spin); + fhsm->fhsm_pid = 0; + spin_unlock(&fhsm->fhsm_spin); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_fhsm_br_alloc(struct au_branch *br) +{ + int err; + + err = 0; + br->br_fhsm = kmalloc(sizeof(*br->br_fhsm), GFP_NOFS); + if (br->br_fhsm) + au_br_fhsm_init(br->br_fhsm); + else + err = -ENOMEM; + + return err; +} + +/* ---------------------------------------------------------------------- */ + +void au_fhsm_fin(struct super_block *sb) +{ + au_fhsm_notify(sb, /*val*/-1); +} + +void au_fhsm_init(struct au_sbinfo *sbinfo) +{ + struct au_fhsm *fhsm; + + fhsm = &sbinfo->si_fhsm; + spin_lock_init(&fhsm->fhsm_spin); + init_waitqueue_head(&fhsm->fhsm_wqh); + atomic_set(&fhsm->fhsm_readable, 0); + fhsm->fhsm_expire + = msecs_to_jiffies(AUFS_FHSM_CACHE_DEF_SEC * MSEC_PER_SEC); + fhsm->fhsm_bottom = -1; +} + +void au_fhsm_set(struct au_sbinfo *sbinfo, unsigned int sec) +{ + sbinfo->si_fhsm.fhsm_expire + = msecs_to_jiffies(sec * MSEC_PER_SEC); +} + +void au_fhsm_show(struct seq_file *seq, struct au_sbinfo *sbinfo) +{ + unsigned int u; + + if (!au_ftest_si(sbinfo, FHSM)) + return; + + u = jiffies_to_msecs(sbinfo->si_fhsm.fhsm_expire) / MSEC_PER_SEC; + if (u != AUFS_FHSM_CACHE_DEF_SEC) + seq_printf(seq, ",fhsm_sec=%u", u); +} diff --git a/fs/aufs/file.c b/fs/aufs/file.c new file mode 100644 index 00000000000..b0075b57d8b --- /dev/null +++ b/fs/aufs/file.c @@ -0,0 +1,863 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * handling file/dir, and address_space operation + */ + +#ifdef CONFIG_AUFS_DEBUG +#include +#endif +#include +#include "aufs.h" + +/* drop flags for writing */ +unsigned int au_file_roflags(unsigned int flags) +{ + flags &= ~(O_WRONLY | O_RDWR | O_APPEND | O_CREAT | O_TRUNC); + flags |= O_RDONLY | O_NOATIME; + return flags; +} + +/* common functions to regular file and dir */ +struct file *au_h_open(struct dentry *dentry, aufs_bindex_t bindex, int flags, + struct file *file, int force_wr) +{ + struct file *h_file; + struct dentry *h_dentry; + struct inode *h_inode; + struct super_block *sb; + struct au_branch *br; + struct path h_path; + int err; + + /* a race condition can happen between open and unlink/rmdir */ + h_file = ERR_PTR(-ENOENT); + h_dentry = au_h_dptr(dentry, bindex); + if (au_test_nfsd() && (!h_dentry || d_is_negative(h_dentry))) + goto out; + h_inode = d_inode(h_dentry); + spin_lock(&h_dentry->d_lock); + err = (!d_unhashed(dentry) && d_unlinked(h_dentry)) + /* || !d_inode(dentry)->i_nlink */ + ; + spin_unlock(&h_dentry->d_lock); + if (unlikely(err)) + goto out; + + sb = dentry->d_sb; + br = au_sbr(sb, bindex); + err = au_br_test_oflag(flags, br); + h_file = ERR_PTR(err); + if (unlikely(err)) + goto out; + + /* drop flags for writing */ + if (au_test_ro(sb, bindex, d_inode(dentry))) { + if (force_wr && !(flags & O_WRONLY)) + force_wr = 0; + flags = au_file_roflags(flags); + if (force_wr) { + h_file = ERR_PTR(-EROFS); + flags = au_file_roflags(flags); + if (unlikely(vfsub_native_ro(h_inode) + || IS_APPEND(h_inode))) + goto out; + flags &= ~O_ACCMODE; + flags |= O_WRONLY; + } + } + flags &= ~O_CREAT; + au_lcnt_inc(&br->br_nfiles); + h_path.dentry = h_dentry; + h_path.mnt = au_br_mnt(br); + h_file = vfsub_dentry_open(&h_path, flags); + if (IS_ERR(h_file)) + goto out_br; + + if (flags & __FMODE_EXEC) { + err = deny_write_access(h_file); + if (unlikely(err)) { + fput(h_file); + h_file = ERR_PTR(err); + goto out_br; + } + } + fsnotify_open(h_file); + goto out; /* success */ + +out_br: + au_lcnt_dec(&br->br_nfiles); +out: + return h_file; +} + +static int au_cmoo(struct dentry *dentry) +{ + int err, cmoo, matched; + unsigned int udba; + struct path h_path; + struct au_pin pin; + struct au_cp_generic cpg = { + .dentry = dentry, + .bdst = -1, + .bsrc = -1, + .len = -1, + .pin = &pin, + .flags = AuCpup_DTIME | AuCpup_HOPEN + }; + struct inode *delegated; + struct super_block *sb; + struct au_sbinfo *sbinfo; + struct au_fhsm *fhsm; + pid_t pid; + struct au_branch *br; + struct dentry *parent; + struct au_hinode *hdir; + + DiMustWriteLock(dentry); + IiMustWriteLock(d_inode(dentry)); + + err = 0; + if (IS_ROOT(dentry)) + goto out; + cpg.bsrc = au_dbtop(dentry); + if (!cpg.bsrc) + goto out; + + sb = dentry->d_sb; + sbinfo = au_sbi(sb); + fhsm = &sbinfo->si_fhsm; + pid = au_fhsm_pid(fhsm); + rcu_read_lock(); + matched = (pid + && (current->pid == pid + || rcu_dereference(current->real_parent)->pid == pid)); + rcu_read_unlock(); + if (matched) + goto out; + + br = au_sbr(sb, cpg.bsrc); + cmoo = au_br_cmoo(br->br_perm); + if (!cmoo) + goto out; + if (!d_is_reg(dentry)) + cmoo &= AuBrAttr_COO_ALL; + if (!cmoo) + goto out; + + parent = dget_parent(dentry); + di_write_lock_parent(parent); + err = au_wbr_do_copyup_bu(dentry, cpg.bsrc - 1); + cpg.bdst = err; + if (unlikely(err < 0)) { + err = 0; /* there is no upper writable branch */ + goto out_dgrade; + } + AuDbg("bsrc %d, bdst %d\n", cpg.bsrc, cpg.bdst); + + /* do not respect the coo attrib for the target branch */ + err = au_cpup_dirs(dentry, cpg.bdst); + if (unlikely(err)) + goto out_dgrade; + + di_downgrade_lock(parent, AuLock_IR); + udba = au_opt_udba(sb); + err = au_pin(&pin, dentry, cpg.bdst, udba, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (unlikely(err)) + goto out_parent; + + err = au_sio_cpup_simple(&cpg); + au_unpin(&pin); + if (unlikely(err)) + goto out_parent; + if (!(cmoo & AuBrWAttr_MOO)) + goto out_parent; /* success */ + + err = au_pin(&pin, dentry, cpg.bsrc, udba, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (unlikely(err)) + goto out_parent; + + h_path.mnt = au_br_mnt(br); + h_path.dentry = au_h_dptr(dentry, cpg.bsrc); + hdir = au_hi(d_inode(parent), cpg.bsrc); + delegated = NULL; + err = vfsub_unlink(hdir->hi_inode, &h_path, &delegated, /*force*/1); + au_unpin(&pin); + /* todo: keep h_dentry or not? */ + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + if (unlikely(err)) { + pr_err("unlink %pd after coo failed (%d), ignored\n", + dentry, err); + err = 0; + } + goto out_parent; /* success */ + +out_dgrade: + di_downgrade_lock(parent, AuLock_IR); +out_parent: + di_read_unlock(parent, AuLock_IR); + dput(parent); +out: + AuTraceErr(err); + return err; +} + +int au_do_open(struct file *file, struct au_do_open_args *args) +{ + int err, aopen = args->aopen; + struct dentry *dentry; + struct au_finfo *finfo; + + if (!aopen) + err = au_finfo_init(file, args->fidir); + else { + lockdep_off(); + err = au_finfo_init(file, args->fidir); + lockdep_on(); + } + if (unlikely(err)) + goto out; + + dentry = file->f_path.dentry; + AuDebugOn(IS_ERR_OR_NULL(dentry)); + di_write_lock_child(dentry); + err = au_cmoo(dentry); + di_downgrade_lock(dentry, AuLock_IR); + if (!err) { + if (!aopen) + err = args->open(file, vfsub_file_flags(file), NULL); + else { + lockdep_off(); + err = args->open(file, vfsub_file_flags(file), + args->h_file); + lockdep_on(); + } + } + di_read_unlock(dentry, AuLock_IR); + + finfo = au_fi(file); + if (!err) { + finfo->fi_file = file; + au_hbl_add(&finfo->fi_hlist, + &au_sbi(file->f_path.dentry->d_sb)->si_files); + } + if (!aopen) + fi_write_unlock(file); + else { + lockdep_off(); + fi_write_unlock(file); + lockdep_on(); + } + if (unlikely(err)) { + finfo->fi_hdir = NULL; + au_finfo_fin(file); + } + +out: + AuTraceErr(err); + return err; +} + +int au_reopen_nondir(struct file *file) +{ + int err; + aufs_bindex_t btop; + struct dentry *dentry; + struct au_branch *br; + struct file *h_file, *h_file_tmp; + + dentry = file->f_path.dentry; + btop = au_dbtop(dentry); + br = au_sbr(dentry->d_sb, btop); + h_file_tmp = NULL; + if (au_fbtop(file) == btop) { + h_file = au_hf_top(file); + if (file->f_mode == h_file->f_mode) + return 0; /* success */ + h_file_tmp = h_file; + get_file(h_file_tmp); + au_lcnt_inc(&br->br_nfiles); + au_set_h_fptr(file, btop, NULL); + } + AuDebugOn(au_fi(file)->fi_hdir); + /* + * it can happen + * file exists on both of rw and ro + * open --> dbtop and fbtop are both 0 + * prepend a branch as rw, "rw" become ro + * remove rw/file + * delete the top branch, "rw" becomes rw again + * --> dbtop is 1, fbtop is still 0 + * write --> fbtop is 0 but dbtop is 1 + */ + /* AuDebugOn(au_fbtop(file) < btop); */ + + h_file = au_h_open(dentry, btop, vfsub_file_flags(file) & ~O_TRUNC, + file, /*force_wr*/0); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) { + if (h_file_tmp) { + /* revert */ + au_set_h_fptr(file, btop, h_file_tmp); + h_file_tmp = NULL; + } + goto out; /* todo: close all? */ + } + + err = 0; + au_set_fbtop(file, btop); + au_set_h_fptr(file, btop, h_file); + au_update_figen(file); + /* todo: necessary? */ + /* file->f_ra = h_file->f_ra; */ + +out: + if (h_file_tmp) { + fput(h_file_tmp); + au_lcnt_dec(&br->br_nfiles); + } + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_reopen_wh(struct file *file, aufs_bindex_t btgt, + struct dentry *hi_wh) +{ + int err; + aufs_bindex_t btop; + struct au_dinfo *dinfo; + struct dentry *h_dentry; + struct au_hdentry *hdp; + + dinfo = au_di(file->f_path.dentry); + AuRwMustWriteLock(&dinfo->di_rwsem); + + btop = dinfo->di_btop; + dinfo->di_btop = btgt; + hdp = au_hdentry(dinfo, btgt); + h_dentry = hdp->hd_dentry; + hdp->hd_dentry = hi_wh; + err = au_reopen_nondir(file); + hdp->hd_dentry = h_dentry; + dinfo->di_btop = btop; + + return err; +} + +static int au_ready_to_write_wh(struct file *file, loff_t len, + aufs_bindex_t bcpup, struct au_pin *pin) +{ + int err; + struct inode *inode, *h_inode; + struct dentry *h_dentry, *hi_wh; + struct au_cp_generic cpg = { + .dentry = file->f_path.dentry, + .bdst = bcpup, + .bsrc = -1, + .len = len, + .pin = pin + }; + + au_update_dbtop(cpg.dentry); + inode = d_inode(cpg.dentry); + h_inode = NULL; + if (au_dbtop(cpg.dentry) <= bcpup + && au_dbbot(cpg.dentry) >= bcpup) { + h_dentry = au_h_dptr(cpg.dentry, bcpup); + if (h_dentry && d_is_positive(h_dentry)) + h_inode = d_inode(h_dentry); + } + hi_wh = au_hi_wh(inode, bcpup); + if (!hi_wh && !h_inode) + err = au_sio_cpup_wh(&cpg, file); + else + /* already copied-up after unlink */ + err = au_reopen_wh(file, bcpup, hi_wh); + + if (!err + && (inode->i_nlink > 1 + || (inode->i_state & I_LINKABLE)) + && au_opt_test(au_mntflags(cpg.dentry->d_sb), PLINK)) + au_plink_append(inode, bcpup, au_h_dptr(cpg.dentry, bcpup)); + + return err; +} + +/* + * prepare the @file for writing. + */ +int au_ready_to_write(struct file *file, loff_t len, struct au_pin *pin) +{ + int err; + aufs_bindex_t dbtop; + struct dentry *parent; + struct inode *inode; + struct super_block *sb; + struct file *h_file; + struct au_cp_generic cpg = { + .dentry = file->f_path.dentry, + .bdst = -1, + .bsrc = -1, + .len = len, + .pin = pin, + .flags = AuCpup_DTIME + }; + + sb = cpg.dentry->d_sb; + inode = d_inode(cpg.dentry); + cpg.bsrc = au_fbtop(file); + err = au_test_ro(sb, cpg.bsrc, inode); + if (!err && (au_hf_top(file)->f_mode & FMODE_WRITE)) { + err = au_pin(pin, cpg.dentry, cpg.bsrc, AuOpt_UDBA_NONE, + /*flags*/0); + goto out; + } + + /* need to cpup or reopen */ + parent = dget_parent(cpg.dentry); + di_write_lock_parent(parent); + err = AuWbrCopyup(au_sbi(sb), cpg.dentry); + cpg.bdst = err; + if (unlikely(err < 0)) + goto out_dgrade; + err = 0; + + if (!d_unhashed(cpg.dentry) && !au_h_dptr(parent, cpg.bdst)) { + err = au_cpup_dirs(cpg.dentry, cpg.bdst); + if (unlikely(err)) + goto out_dgrade; + } + + err = au_pin(pin, cpg.dentry, cpg.bdst, AuOpt_UDBA_NONE, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (unlikely(err)) + goto out_dgrade; + + dbtop = au_dbtop(cpg.dentry); + if (dbtop <= cpg.bdst) + cpg.bsrc = cpg.bdst; + + if (dbtop <= cpg.bdst /* just reopen */ + || !d_unhashed(cpg.dentry) /* copyup and reopen */ + ) { + h_file = au_h_open_pre(cpg.dentry, cpg.bsrc, /*force_wr*/0); + if (IS_ERR(h_file)) + err = PTR_ERR(h_file); + else { + di_downgrade_lock(parent, AuLock_IR); + if (dbtop > cpg.bdst) + err = au_sio_cpup_simple(&cpg); + if (!err) + err = au_reopen_nondir(file); + au_h_open_post(cpg.dentry, cpg.bsrc, h_file); + } + } else { /* copyup as wh and reopen */ + /* + * since writable hfsplus branch is not supported, + * h_open_pre/post() are unnecessary. + */ + err = au_ready_to_write_wh(file, len, cpg.bdst, pin); + di_downgrade_lock(parent, AuLock_IR); + } + + if (!err) { + au_pin_set_parent_lflag(pin, /*lflag*/0); + goto out_dput; /* success */ + } + au_unpin(pin); + goto out_unlock; + +out_dgrade: + di_downgrade_lock(parent, AuLock_IR); +out_unlock: + di_read_unlock(parent, AuLock_IR); +out_dput: + dput(parent); +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_do_flush(struct file *file, fl_owner_t id, + int (*flush)(struct file *file, fl_owner_t id)) +{ + int err; + struct super_block *sb; + struct inode *inode; + + inode = file_inode(file); + sb = inode->i_sb; + si_noflush_read_lock(sb); + fi_read_lock(file); + ii_read_lock_child(inode); + + err = flush(file, id); + au_cpup_attr_timesizes(inode); + + ii_read_unlock(inode); + fi_read_unlock(file); + si_read_unlock(sb); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_file_refresh_by_inode(struct file *file, int *need_reopen) +{ + int err; + struct au_pin pin; + struct au_finfo *finfo; + struct dentry *parent, *hi_wh; + struct inode *inode; + struct super_block *sb; + struct au_cp_generic cpg = { + .dentry = file->f_path.dentry, + .bdst = -1, + .bsrc = -1, + .len = -1, + .pin = &pin, + .flags = AuCpup_DTIME + }; + + FiMustWriteLock(file); + + err = 0; + finfo = au_fi(file); + sb = cpg.dentry->d_sb; + inode = d_inode(cpg.dentry); + cpg.bdst = au_ibtop(inode); + if (cpg.bdst == finfo->fi_btop || IS_ROOT(cpg.dentry)) + goto out; + + parent = dget_parent(cpg.dentry); + if (au_test_ro(sb, cpg.bdst, inode)) { + di_read_lock_parent(parent, !AuLock_IR); + err = AuWbrCopyup(au_sbi(sb), cpg.dentry); + cpg.bdst = err; + di_read_unlock(parent, !AuLock_IR); + if (unlikely(err < 0)) + goto out_parent; + err = 0; + } + + di_read_lock_parent(parent, AuLock_IR); + hi_wh = au_hi_wh(inode, cpg.bdst); + if (!S_ISDIR(inode->i_mode) + && au_opt_test(au_mntflags(sb), PLINK) + && au_plink_test(inode) + && !d_unhashed(cpg.dentry) + && cpg.bdst < au_dbtop(cpg.dentry)) { + err = au_test_and_cpup_dirs(cpg.dentry, cpg.bdst); + if (unlikely(err)) + goto out_unlock; + + /* always superio. */ + err = au_pin(&pin, cpg.dentry, cpg.bdst, AuOpt_UDBA_NONE, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (!err) { + err = au_sio_cpup_simple(&cpg); + au_unpin(&pin); + } + } else if (hi_wh) { + /* already copied-up after unlink */ + err = au_reopen_wh(file, cpg.bdst, hi_wh); + *need_reopen = 0; + } + +out_unlock: + di_read_unlock(parent, AuLock_IR); +out_parent: + dput(parent); +out: + return err; +} + +static void au_do_refresh_dir(struct file *file) +{ + aufs_bindex_t bindex, bbot, new_bindex, brid; + struct au_hfile *p, tmp, *q; + struct au_finfo *finfo; + struct super_block *sb; + struct au_fidir *fidir; + + FiMustWriteLock(file); + + sb = file->f_path.dentry->d_sb; + finfo = au_fi(file); + fidir = finfo->fi_hdir; + AuDebugOn(!fidir); + p = fidir->fd_hfile + finfo->fi_btop; + brid = p->hf_br->br_id; + bbot = fidir->fd_bbot; + for (bindex = finfo->fi_btop; bindex <= bbot; bindex++, p++) { + if (!p->hf_file) + continue; + + new_bindex = au_br_index(sb, p->hf_br->br_id); + if (new_bindex == bindex) + continue; + if (new_bindex < 0) { + au_set_h_fptr(file, bindex, NULL); + continue; + } + + /* swap two lower inode, and loop again */ + q = fidir->fd_hfile + new_bindex; + tmp = *q; + *q = *p; + *p = tmp; + if (tmp.hf_file) { + bindex--; + p--; + } + } + + p = fidir->fd_hfile; + if (!au_test_mmapped(file) && !d_unlinked(file->f_path.dentry)) { + bbot = au_sbbot(sb); + for (finfo->fi_btop = 0; finfo->fi_btop <= bbot; + finfo->fi_btop++, p++) + if (p->hf_file) { + if (file_inode(p->hf_file)) + break; + au_hfput(p, /*execed*/0); + } + } else { + bbot = au_br_index(sb, brid); + for (finfo->fi_btop = 0; finfo->fi_btop < bbot; + finfo->fi_btop++, p++) + if (p->hf_file) + au_hfput(p, /*execed*/0); + bbot = au_sbbot(sb); + } + + p = fidir->fd_hfile + bbot; + for (fidir->fd_bbot = bbot; fidir->fd_bbot >= finfo->fi_btop; + fidir->fd_bbot--, p--) + if (p->hf_file) { + if (file_inode(p->hf_file)) + break; + au_hfput(p, /*execed*/0); + } + AuDebugOn(fidir->fd_bbot < finfo->fi_btop); +} + +/* + * after branch manipulating, refresh the file. + */ +static int refresh_file(struct file *file, int (*reopen)(struct file *file)) +{ + int err, need_reopen, nbr; + aufs_bindex_t bbot, bindex; + struct dentry *dentry; + struct super_block *sb; + struct au_finfo *finfo; + struct au_hfile *hfile; + + dentry = file->f_path.dentry; + sb = dentry->d_sb; + nbr = au_sbbot(sb) + 1; + finfo = au_fi(file); + if (!finfo->fi_hdir) { + hfile = &finfo->fi_htop; + AuDebugOn(!hfile->hf_file); + bindex = au_br_index(sb, hfile->hf_br->br_id); + AuDebugOn(bindex < 0); + if (bindex != finfo->fi_btop) + au_set_fbtop(file, bindex); + } else { + err = au_fidir_realloc(finfo, nbr, /*may_shrink*/0); + if (unlikely(err)) + goto out; + au_do_refresh_dir(file); + } + + err = 0; + need_reopen = 1; + if (!au_test_mmapped(file)) + err = au_file_refresh_by_inode(file, &need_reopen); + if (finfo->fi_hdir) + /* harmless if err */ + au_fidir_realloc(finfo, nbr, /*may_shrink*/1); + if (!err && need_reopen && !d_unlinked(dentry)) + err = reopen(file); + if (!err) { + au_update_figen(file); + goto out; /* success */ + } + + /* error, close all lower files */ + if (finfo->fi_hdir) { + bbot = au_fbbot_dir(file); + for (bindex = au_fbtop(file); bindex <= bbot; bindex++) + au_set_h_fptr(file, bindex, NULL); + } + +out: + return err; +} + +/* common function to regular file and dir */ +int au_reval_and_lock_fdi(struct file *file, int (*reopen)(struct file *file), + int wlock, unsigned int fi_lsc) +{ + int err; + unsigned int sigen, figen; + aufs_bindex_t btop; + unsigned char pseudo_link; + struct dentry *dentry; + struct inode *inode; + + err = 0; + dentry = file->f_path.dentry; + inode = d_inode(dentry); + sigen = au_sigen(dentry->d_sb); + fi_write_lock_nested(file, fi_lsc); + figen = au_figen(file); + if (!fi_lsc) + di_write_lock_child(dentry); + else + di_write_lock_child2(dentry); + btop = au_dbtop(dentry); + pseudo_link = (btop != au_ibtop(inode)); + if (sigen == figen && !pseudo_link && au_fbtop(file) == btop) { + if (!wlock) { + di_downgrade_lock(dentry, AuLock_IR); + fi_downgrade_lock(file); + } + goto out; /* success */ + } + + AuDbg("sigen %d, figen %d\n", sigen, figen); + if (au_digen_test(dentry, sigen)) { + err = au_reval_dpath(dentry, sigen); + AuDebugOn(!err && au_digen_test(dentry, sigen)); + } + + if (!err) + err = refresh_file(file, reopen); + if (!err) { + if (!wlock) { + di_downgrade_lock(dentry, AuLock_IR); + fi_downgrade_lock(file); + } + } else { + di_write_unlock(dentry); + fi_write_unlock(file); + } + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* cf. aufs_nopage() */ +/* for madvise(2) */ +static int aufs_readpage(struct file *file __maybe_unused, struct page *page) +{ + unlock_page(page); + return 0; +} + +/* it will never be called, but necessary to support O_DIRECT */ +static ssize_t aufs_direct_IO(struct kiocb *iocb, struct iov_iter *iter) +{ BUG(); return 0; } + +/* they will never be called. */ +#ifdef CONFIG_AUFS_DEBUG +static int aufs_write_begin(struct file *file, struct address_space *mapping, + loff_t pos, unsigned len, unsigned flags, + struct page **pagep, void **fsdata) +{ AuUnsupport(); return 0; } +static int aufs_write_end(struct file *file, struct address_space *mapping, + loff_t pos, unsigned len, unsigned copied, + struct page *page, void *fsdata) +{ AuUnsupport(); return 0; } +static int aufs_writepage(struct page *page, struct writeback_control *wbc) +{ AuUnsupport(); return 0; } + +static int aufs_set_page_dirty(struct page *page) +{ AuUnsupport(); return 0; } +static void aufs_invalidatepage(struct page *page, unsigned int offset, + unsigned int length) +{ AuUnsupport(); } +static int aufs_releasepage(struct page *page, gfp_t gfp) +{ AuUnsupport(); return 0; } +#if 0 /* called by memory compaction regardless file */ +static int aufs_migratepage(struct address_space *mapping, struct page *newpage, + struct page *page, enum migrate_mode mode) +{ AuUnsupport(); return 0; } +#endif +static bool aufs_isolate_page(struct page *page, isolate_mode_t mode) +{ AuUnsupport(); return true; } +static void aufs_putback_page(struct page *page) +{ AuUnsupport(); } +static int aufs_launder_page(struct page *page) +{ AuUnsupport(); return 0; } +static int aufs_is_partially_uptodate(struct page *page, + unsigned long from, + unsigned long count) +{ AuUnsupport(); return 0; } +static void aufs_is_dirty_writeback(struct page *page, bool *dirty, + bool *writeback) +{ AuUnsupport(); } +static int aufs_error_remove_page(struct address_space *mapping, + struct page *page) +{ AuUnsupport(); return 0; } +static int aufs_swap_activate(struct swap_info_struct *sis, struct file *file, + sector_t *span) +{ AuUnsupport(); return 0; } +static void aufs_swap_deactivate(struct file *file) +{ AuUnsupport(); } +#endif /* CONFIG_AUFS_DEBUG */ + +const struct address_space_operations aufs_aop = { + .readpage = aufs_readpage, + .direct_IO = aufs_direct_IO, +#ifdef CONFIG_AUFS_DEBUG + .writepage = aufs_writepage, + /* no writepages, because of writepage */ + .set_page_dirty = aufs_set_page_dirty, + /* no readpages, because of readpage */ + .write_begin = aufs_write_begin, + .write_end = aufs_write_end, + /* no bmap, no block device */ + .invalidatepage = aufs_invalidatepage, + .releasepage = aufs_releasepage, + /* is fallback_migrate_page ok? */ + /* .migratepage = aufs_migratepage, */ + .isolate_page = aufs_isolate_page, + .putback_page = aufs_putback_page, + .launder_page = aufs_launder_page, + .is_partially_uptodate = aufs_is_partially_uptodate, + .is_dirty_writeback = aufs_is_dirty_writeback, + .error_remove_page = aufs_error_remove_page, + .swap_activate = aufs_swap_activate, + .swap_deactivate = aufs_swap_deactivate +#endif /* CONFIG_AUFS_DEBUG */ +}; diff --git a/fs/aufs/file.h b/fs/aufs/file.h new file mode 100644 index 00000000000..d124d9c8216 --- /dev/null +++ b/fs/aufs/file.h @@ -0,0 +1,342 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * file operations + */ + +#ifndef __AUFS_FILE_H__ +#define __AUFS_FILE_H__ + +#ifdef __KERNEL__ + +#include +#include +#include +#include +#include "rwsem.h" + +struct au_branch; +struct au_hfile { + struct file *hf_file; + struct au_branch *hf_br; +}; + +struct au_vdir; +struct au_fidir { + aufs_bindex_t fd_bbot; + aufs_bindex_t fd_nent; + struct au_vdir *fd_vdir_cache; + struct au_hfile fd_hfile[]; +}; + +static inline int au_fidir_sz(int nent) +{ + AuDebugOn(nent < 0); + return sizeof(struct au_fidir) + sizeof(struct au_hfile) * nent; +} + +struct au_finfo { + atomic_t fi_generation; + + struct au_rwsem fi_rwsem; + aufs_bindex_t fi_btop; + + /* do not union them */ + struct { /* for non-dir */ + struct au_hfile fi_htop; + atomic_t fi_mmapped; + }; + struct au_fidir *fi_hdir; /* for dir only */ + + struct hlist_bl_node fi_hlist; + struct file *fi_file; /* very ugly */ + struct rcu_head rcu; +} ____cacheline_aligned_in_smp; + +/* ---------------------------------------------------------------------- */ + +/* file.c */ +extern const struct address_space_operations aufs_aop; +unsigned int au_file_roflags(unsigned int flags); +struct file *au_h_open(struct dentry *dentry, aufs_bindex_t bindex, int flags, + struct file *file, int force_wr); +struct au_do_open_args { + int aopen; + int (*open)(struct file *file, int flags, + struct file *h_file); + struct au_fidir *fidir; + struct file *h_file; +}; +int au_do_open(struct file *file, struct au_do_open_args *args); +int au_reopen_nondir(struct file *file); +struct au_pin; +int au_ready_to_write(struct file *file, loff_t len, struct au_pin *pin); +int au_reval_and_lock_fdi(struct file *file, int (*reopen)(struct file *file), + int wlock, unsigned int fi_lsc); +int au_do_flush(struct file *file, fl_owner_t id, + int (*flush)(struct file *file, fl_owner_t id)); + +/* poll.c */ +#ifdef CONFIG_AUFS_POLL +__poll_t aufs_poll(struct file *file, struct poll_table_struct *pt); +#endif + +#ifdef CONFIG_AUFS_BR_HFSPLUS +/* hfsplus.c */ +struct file *au_h_open_pre(struct dentry *dentry, aufs_bindex_t bindex, + int force_wr); +void au_h_open_post(struct dentry *dentry, aufs_bindex_t bindex, + struct file *h_file); +#else +AuStub(struct file *, au_h_open_pre, return NULL, struct dentry *dentry, + aufs_bindex_t bindex, int force_wr) +AuStubVoid(au_h_open_post, struct dentry *dentry, aufs_bindex_t bindex, + struct file *h_file); +#endif + +/* f_op.c */ +extern const struct file_operations aufs_file_fop; +int au_do_open_nondir(struct file *file, int flags, struct file *h_file); +int aufs_release_nondir(struct inode *inode __maybe_unused, struct file *file); +struct file *au_read_pre(struct file *file, int keep_fi, unsigned int lsc); + +/* finfo.c */ +void au_hfput(struct au_hfile *hf, int execed); +void au_set_h_fptr(struct file *file, aufs_bindex_t bindex, + struct file *h_file); + +void au_update_figen(struct file *file); +struct au_fidir *au_fidir_alloc(struct super_block *sb); +int au_fidir_realloc(struct au_finfo *finfo, int nbr, int may_shrink); + +void au_fi_init_once(void *_fi); +void au_finfo_fin(struct file *file); +int au_finfo_init(struct file *file, struct au_fidir *fidir); + +/* ioctl.c */ +long aufs_ioctl_nondir(struct file *file, unsigned int cmd, unsigned long arg); +#ifdef CONFIG_COMPAT +long aufs_compat_ioctl_dir(struct file *file, unsigned int cmd, + unsigned long arg); +long aufs_compat_ioctl_nondir(struct file *file, unsigned int cmd, + unsigned long arg); +#endif + +/* ---------------------------------------------------------------------- */ + +static inline struct au_finfo *au_fi(struct file *file) +{ + return file->private_data; +} + +/* ---------------------------------------------------------------------- */ + +#define fi_read_lock(f) au_rw_read_lock(&au_fi(f)->fi_rwsem) +#define fi_write_lock(f) au_rw_write_lock(&au_fi(f)->fi_rwsem) +#define fi_read_trylock(f) au_rw_read_trylock(&au_fi(f)->fi_rwsem) +#define fi_write_trylock(f) au_rw_write_trylock(&au_fi(f)->fi_rwsem) +/* +#define fi_read_trylock_nested(f) \ + au_rw_read_trylock_nested(&au_fi(f)->fi_rwsem) +#define fi_write_trylock_nested(f) \ + au_rw_write_trylock_nested(&au_fi(f)->fi_rwsem) +*/ + +#define fi_read_unlock(f) au_rw_read_unlock(&au_fi(f)->fi_rwsem) +#define fi_write_unlock(f) au_rw_write_unlock(&au_fi(f)->fi_rwsem) +#define fi_downgrade_lock(f) au_rw_dgrade_lock(&au_fi(f)->fi_rwsem) + +/* lock subclass for finfo */ +enum { + AuLsc_FI_1, + AuLsc_FI_2 +}; + +static inline void fi_read_lock_nested(struct file *f, unsigned int lsc) +{ + au_rw_read_lock_nested(&au_fi(f)->fi_rwsem, lsc); +} + +static inline void fi_write_lock_nested(struct file *f, unsigned int lsc) +{ + au_rw_write_lock_nested(&au_fi(f)->fi_rwsem, lsc); +} + +/* + * fi_read_lock_1, fi_write_lock_1, + * fi_read_lock_2, fi_write_lock_2 + */ +#define AuReadLockFunc(name) \ +static inline void fi_read_lock_##name(struct file *f) \ +{ fi_read_lock_nested(f, AuLsc_FI_##name); } + +#define AuWriteLockFunc(name) \ +static inline void fi_write_lock_##name(struct file *f) \ +{ fi_write_lock_nested(f, AuLsc_FI_##name); } + +#define AuRWLockFuncs(name) \ + AuReadLockFunc(name) \ + AuWriteLockFunc(name) + +AuRWLockFuncs(1); +AuRWLockFuncs(2); + +#undef AuReadLockFunc +#undef AuWriteLockFunc +#undef AuRWLockFuncs + +#define FiMustNoWaiters(f) AuRwMustNoWaiters(&au_fi(f)->fi_rwsem) +#define FiMustAnyLock(f) AuRwMustAnyLock(&au_fi(f)->fi_rwsem) +#define FiMustWriteLock(f) AuRwMustWriteLock(&au_fi(f)->fi_rwsem) + +/* ---------------------------------------------------------------------- */ + +/* todo: hard/soft set? */ +static inline aufs_bindex_t au_fbtop(struct file *file) +{ + FiMustAnyLock(file); + return au_fi(file)->fi_btop; +} + +static inline aufs_bindex_t au_fbbot_dir(struct file *file) +{ + FiMustAnyLock(file); + AuDebugOn(!au_fi(file)->fi_hdir); + return au_fi(file)->fi_hdir->fd_bbot; +} + +static inline struct au_vdir *au_fvdir_cache(struct file *file) +{ + FiMustAnyLock(file); + AuDebugOn(!au_fi(file)->fi_hdir); + return au_fi(file)->fi_hdir->fd_vdir_cache; +} + +static inline void au_set_fbtop(struct file *file, aufs_bindex_t bindex) +{ + FiMustWriteLock(file); + au_fi(file)->fi_btop = bindex; +} + +static inline void au_set_fbbot_dir(struct file *file, aufs_bindex_t bindex) +{ + FiMustWriteLock(file); + AuDebugOn(!au_fi(file)->fi_hdir); + au_fi(file)->fi_hdir->fd_bbot = bindex; +} + +static inline void au_set_fvdir_cache(struct file *file, + struct au_vdir *vdir_cache) +{ + FiMustWriteLock(file); + AuDebugOn(!au_fi(file)->fi_hdir); + au_fi(file)->fi_hdir->fd_vdir_cache = vdir_cache; +} + +static inline struct file *au_hf_top(struct file *file) +{ + FiMustAnyLock(file); + AuDebugOn(au_fi(file)->fi_hdir); + return au_fi(file)->fi_htop.hf_file; +} + +static inline struct file *au_hf_dir(struct file *file, aufs_bindex_t bindex) +{ + FiMustAnyLock(file); + AuDebugOn(!au_fi(file)->fi_hdir); + return au_fi(file)->fi_hdir->fd_hfile[0 + bindex].hf_file; +} + +/* todo: memory barrier? */ +static inline unsigned int au_figen(struct file *f) +{ + return atomic_read(&au_fi(f)->fi_generation); +} + +static inline void au_set_mmapped(struct file *f) +{ + if (atomic_inc_return(&au_fi(f)->fi_mmapped)) + return; + pr_warn("fi_mmapped wrapped around\n"); + while (!atomic_inc_return(&au_fi(f)->fi_mmapped)) + ; +} + +static inline void au_unset_mmapped(struct file *f) +{ + atomic_dec(&au_fi(f)->fi_mmapped); +} + +static inline int au_test_mmapped(struct file *f) +{ + return atomic_read(&au_fi(f)->fi_mmapped); +} + +/* customize vma->vm_file */ + +static inline void au_do_vm_file_reset(struct vm_area_struct *vma, + struct file *file) +{ + struct file *f; + + f = vma->vm_file; + get_file(file); + vma->vm_file = file; + fput(f); +} + +#ifdef CONFIG_MMU +#define AuDbgVmRegion(file, vma) do {} while (0) + +static inline void au_vm_file_reset(struct vm_area_struct *vma, + struct file *file) +{ + au_do_vm_file_reset(vma, file); +} +#else +#define AuDbgVmRegion(file, vma) \ + AuDebugOn((vma)->vm_region && (vma)->vm_region->vm_file != (file)) + +static inline void au_vm_file_reset(struct vm_area_struct *vma, + struct file *file) +{ + struct file *f; + + au_do_vm_file_reset(vma, file); + f = vma->vm_region->vm_file; + get_file(file); + vma->vm_region->vm_file = file; + fput(f); +} +#endif /* CONFIG_MMU */ + +/* handle vma->vm_prfile */ +static inline void au_vm_prfile_set(struct vm_area_struct *vma, + struct file *file) +{ + get_file(file); + vma->vm_prfile = file; +#ifndef CONFIG_MMU + get_file(file); + vma->vm_region->vm_prfile = file; +#endif +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_FILE_H__ */ diff --git a/fs/aufs/finfo.c b/fs/aufs/finfo.c new file mode 100644 index 00000000000..25077670a50 --- /dev/null +++ b/fs/aufs/finfo.c @@ -0,0 +1,149 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * file private data + */ + +#include "aufs.h" + +void au_hfput(struct au_hfile *hf, int execed) +{ + if (execed) + allow_write_access(hf->hf_file); + fput(hf->hf_file); + hf->hf_file = NULL; + au_lcnt_dec(&hf->hf_br->br_nfiles); + hf->hf_br = NULL; +} + +void au_set_h_fptr(struct file *file, aufs_bindex_t bindex, struct file *val) +{ + struct au_finfo *finfo = au_fi(file); + struct au_hfile *hf; + struct au_fidir *fidir; + + fidir = finfo->fi_hdir; + if (!fidir) { + AuDebugOn(finfo->fi_btop != bindex); + hf = &finfo->fi_htop; + } else + hf = fidir->fd_hfile + bindex; + + if (hf && hf->hf_file) + au_hfput(hf, vfsub_file_execed(file)); + if (val) { + FiMustWriteLock(file); + AuDebugOn(IS_ERR_OR_NULL(file->f_path.dentry)); + hf->hf_file = val; + hf->hf_br = au_sbr(file->f_path.dentry->d_sb, bindex); + } +} + +void au_update_figen(struct file *file) +{ + atomic_set(&au_fi(file)->fi_generation, au_digen(file->f_path.dentry)); + /* smp_mb(); */ /* atomic_set */ +} + +/* ---------------------------------------------------------------------- */ + +struct au_fidir *au_fidir_alloc(struct super_block *sb) +{ + struct au_fidir *fidir; + int nbr; + + nbr = au_sbbot(sb) + 1; + if (nbr < 2) + nbr = 2; /* initial allocate for 2 branches */ + fidir = kzalloc(au_fidir_sz(nbr), GFP_NOFS); + if (fidir) { + fidir->fd_bbot = -1; + fidir->fd_nent = nbr; + } + + return fidir; +} + +int au_fidir_realloc(struct au_finfo *finfo, int nbr, int may_shrink) +{ + int err; + struct au_fidir *fidir, *p; + + AuRwMustWriteLock(&finfo->fi_rwsem); + fidir = finfo->fi_hdir; + AuDebugOn(!fidir); + + err = -ENOMEM; + p = au_kzrealloc(fidir, au_fidir_sz(fidir->fd_nent), au_fidir_sz(nbr), + GFP_NOFS, may_shrink); + if (p) { + p->fd_nent = nbr; + finfo->fi_hdir = p; + err = 0; + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +void au_finfo_fin(struct file *file) +{ + struct au_finfo *finfo; + + au_lcnt_dec(&au_sbi(file->f_path.dentry->d_sb)->si_nfiles); + + finfo = au_fi(file); + AuDebugOn(finfo->fi_hdir); + AuRwDestroy(&finfo->fi_rwsem); + au_cache_free_finfo(finfo); +} + +void au_fi_init_once(void *_finfo) +{ + struct au_finfo *finfo = _finfo; + + au_rw_init(&finfo->fi_rwsem); +} + +int au_finfo_init(struct file *file, struct au_fidir *fidir) +{ + int err; + struct au_finfo *finfo; + struct dentry *dentry; + + err = -ENOMEM; + dentry = file->f_path.dentry; + finfo = au_cache_alloc_finfo(); + if (unlikely(!finfo)) + goto out; + + err = 0; + au_lcnt_inc(&au_sbi(dentry->d_sb)->si_nfiles); + au_rw_write_lock(&finfo->fi_rwsem); + finfo->fi_btop = -1; + finfo->fi_hdir = fidir; + atomic_set(&finfo->fi_generation, au_digen(dentry)); + /* smp_mb(); */ /* atomic_set */ + + file->private_data = finfo; + +out: + return err; +} diff --git a/fs/aufs/fstype.h b/fs/aufs/fstype.h new file mode 100644 index 00000000000..af4bc6c0dd4 --- /dev/null +++ b/fs/aufs/fstype.h @@ -0,0 +1,401 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * judging filesystem type + */ + +#ifndef __AUFS_FSTYPE_H__ +#define __AUFS_FSTYPE_H__ + +#ifdef __KERNEL__ + +#include +#include +#include +#include + +static inline int au_test_aufs(struct super_block *sb) +{ + return sb->s_magic == AUFS_SUPER_MAGIC; +} + +static inline const char *au_sbtype(struct super_block *sb) +{ + return sb->s_type->name; +} + +static inline int au_test_iso9660(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_ISO9660_FS) + return sb->s_magic == ISOFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_romfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_ROMFS_FS) + return sb->s_magic == ROMFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_cramfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_CRAMFS) + return sb->s_magic == CRAMFS_MAGIC; +#endif + return 0; +} + +static inline int au_test_nfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_NFS_FS) + return sb->s_magic == NFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_fuse(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_FUSE_FS) + return sb->s_magic == FUSE_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_xfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_XFS_FS) + return sb->s_magic == XFS_SB_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_tmpfs(struct super_block *sb __maybe_unused) +{ +#ifdef CONFIG_TMPFS + return sb->s_magic == TMPFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_ecryptfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_ECRYPT_FS) + return !strcmp(au_sbtype(sb), "ecryptfs"); +#else + return 0; +#endif +} + +static inline int au_test_ramfs(struct super_block *sb) +{ + return sb->s_magic == RAMFS_MAGIC; +} + +static inline int au_test_ubifs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_UBIFS_FS) + return sb->s_magic == UBIFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_procfs(struct super_block *sb __maybe_unused) +{ +#ifdef CONFIG_PROC_FS + return sb->s_magic == PROC_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_sysfs(struct super_block *sb __maybe_unused) +{ +#ifdef CONFIG_SYSFS + return sb->s_magic == SYSFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_configfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_CONFIGFS_FS) + return sb->s_magic == CONFIGFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_minix(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_MINIX_FS) + return sb->s_magic == MINIX3_SUPER_MAGIC + || sb->s_magic == MINIX2_SUPER_MAGIC + || sb->s_magic == MINIX2_SUPER_MAGIC2 + || sb->s_magic == MINIX_SUPER_MAGIC + || sb->s_magic == MINIX_SUPER_MAGIC2; +#else + return 0; +#endif +} + +static inline int au_test_fat(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_FAT_FS) + return sb->s_magic == MSDOS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_msdos(struct super_block *sb) +{ + return au_test_fat(sb); +} + +static inline int au_test_vfat(struct super_block *sb) +{ + return au_test_fat(sb); +} + +static inline int au_test_securityfs(struct super_block *sb __maybe_unused) +{ +#ifdef CONFIG_SECURITYFS + return sb->s_magic == SECURITYFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_squashfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_SQUASHFS) + return sb->s_magic == SQUASHFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_btrfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_BTRFS_FS) + return sb->s_magic == BTRFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_xenfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_XENFS) + return sb->s_magic == XENFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_debugfs(struct super_block *sb __maybe_unused) +{ +#ifdef CONFIG_DEBUG_FS + return sb->s_magic == DEBUGFS_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_nilfs(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_NILFS) + return sb->s_magic == NILFS_SUPER_MAGIC; +#else + return 0; +#endif +} + +static inline int au_test_hfsplus(struct super_block *sb __maybe_unused) +{ +#if IS_ENABLED(CONFIG_HFSPLUS_FS) + return sb->s_magic == HFSPLUS_SUPER_MAGIC; +#else + return 0; +#endif +} + +/* ---------------------------------------------------------------------- */ +/* + * they can't be an aufs branch. + */ +static inline int au_test_fs_unsuppoted(struct super_block *sb) +{ + return +#ifndef CONFIG_AUFS_BR_RAMFS + au_test_ramfs(sb) || +#endif + au_test_procfs(sb) + || au_test_sysfs(sb) + || au_test_configfs(sb) + || au_test_debugfs(sb) + || au_test_securityfs(sb) + || au_test_xenfs(sb) + || au_test_ecryptfs(sb) + /* || !strcmp(au_sbtype(sb), "unionfs") */ + || au_test_aufs(sb); /* will be supported in next version */ +} + +static inline int au_test_fs_remote(struct super_block *sb) +{ + return !au_test_tmpfs(sb) +#ifdef CONFIG_AUFS_BR_RAMFS + && !au_test_ramfs(sb) +#endif + && !(sb->s_type->fs_flags & FS_REQUIRES_DEV); +} + +/* ---------------------------------------------------------------------- */ + +/* + * Note: these functions (below) are created after reading ->getattr() in all + * filesystems under linux/fs. it means we have to do so in every update... + */ + +/* + * some filesystems require getattr to refresh the inode attributes before + * referencing. + * in most cases, we can rely on the inode attribute in NFS (or every remote fs) + * and leave the work for d_revalidate() + */ +static inline int au_test_fs_refresh_iattr(struct super_block *sb) +{ + return au_test_nfs(sb) + || au_test_fuse(sb) + /* || au_test_btrfs(sb) */ /* untested */ + ; +} + +/* + * filesystems which don't maintain i_size or i_blocks. + */ +static inline int au_test_fs_bad_iattr_size(struct super_block *sb) +{ + return au_test_xfs(sb) + || au_test_btrfs(sb) + || au_test_ubifs(sb) + || au_test_hfsplus(sb) /* maintained, but incorrect */ + /* || au_test_minix(sb) */ /* untested */ + ; +} + +/* + * filesystems which don't store the correct value in some of their inode + * attributes. + */ +static inline int au_test_fs_bad_iattr(struct super_block *sb) +{ + return au_test_fs_bad_iattr_size(sb) + || au_test_fat(sb) + || au_test_msdos(sb) + || au_test_vfat(sb); +} + +/* they don't check i_nlink in link(2) */ +static inline int au_test_fs_no_limit_nlink(struct super_block *sb) +{ + return au_test_tmpfs(sb) +#ifdef CONFIG_AUFS_BR_RAMFS + || au_test_ramfs(sb) +#endif + || au_test_ubifs(sb) + || au_test_hfsplus(sb); +} + +/* + * filesystems which sets S_NOATIME and S_NOCMTIME. + */ +static inline int au_test_fs_notime(struct super_block *sb) +{ + return au_test_nfs(sb) + || au_test_fuse(sb) + || au_test_ubifs(sb) + ; +} + +/* temporary support for i#1 in cramfs */ +static inline int au_test_fs_unique_ino(struct inode *inode) +{ + if (au_test_cramfs(inode->i_sb)) + return inode->i_ino != 1; + return 1; +} + +/* ---------------------------------------------------------------------- */ + +/* + * the filesystem where the xino files placed must support i/o after unlink and + * maintain i_size and i_blocks. + */ +static inline int au_test_fs_bad_xino(struct super_block *sb) +{ + return au_test_fs_remote(sb) + || au_test_fs_bad_iattr_size(sb) + /* don't want unnecessary work for xino */ + || au_test_aufs(sb) + || au_test_ecryptfs(sb) + || au_test_nilfs(sb); +} + +static inline int au_test_fs_trunc_xino(struct super_block *sb) +{ + return au_test_tmpfs(sb) + || au_test_ramfs(sb); +} + +/* + * test if the @sb is real-readonly. + */ +static inline int au_test_fs_rr(struct super_block *sb) +{ + return au_test_squashfs(sb) + || au_test_iso9660(sb) + || au_test_cramfs(sb) + || au_test_romfs(sb); +} + +/* + * test if the @inode is nfs with 'noacl' option + * NFS always sets SB_POSIXACL regardless its mount option 'noacl.' + */ +static inline int au_test_nfs_noacl(struct inode *inode) +{ + return au_test_nfs(inode->i_sb) + /* && IS_POSIXACL(inode) */ + && !nfs_server_capable(inode, NFS_CAP_ACLS); +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_FSTYPE_H__ */ diff --git a/fs/aufs/hbl.h b/fs/aufs/hbl.h new file mode 100644 index 00000000000..4d5016bccc5 --- /dev/null +++ b/fs/aufs/hbl.h @@ -0,0 +1,65 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2017-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * helpers for hlist_bl.h + */ + +#ifndef __AUFS_HBL_H__ +#define __AUFS_HBL_H__ + +#ifdef __KERNEL__ + +#include + +static inline void au_hbl_add(struct hlist_bl_node *node, + struct hlist_bl_head *hbl) +{ + hlist_bl_lock(hbl); + hlist_bl_add_head(node, hbl); + hlist_bl_unlock(hbl); +} + +static inline void au_hbl_del(struct hlist_bl_node *node, + struct hlist_bl_head *hbl) +{ + hlist_bl_lock(hbl); + hlist_bl_del(node); + hlist_bl_unlock(hbl); +} + +#define au_hbl_for_each(pos, head) \ + for (pos = hlist_bl_first(head); \ + pos; \ + pos = pos->next) + +static inline unsigned long au_hbl_count(struct hlist_bl_head *hbl) +{ + unsigned long cnt; + struct hlist_bl_node *pos; + + cnt = 0; + hlist_bl_lock(hbl); + au_hbl_for_each(pos, hbl) + cnt++; + hlist_bl_unlock(hbl); + return cnt; +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_HBL_H__ */ diff --git a/fs/aufs/hfsnotify.c b/fs/aufs/hfsnotify.c new file mode 100644 index 00000000000..d0abe8ac783 --- /dev/null +++ b/fs/aufs/hfsnotify.c @@ -0,0 +1,288 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * fsnotify for the lower directories + */ + +#include "aufs.h" + +/* FS_IN_IGNORED is unnecessary */ +static const __u32 AuHfsnMask = (FS_MOVED_TO | FS_MOVED_FROM | FS_DELETE + | FS_CREATE | FS_EVENT_ON_CHILD); +static DECLARE_WAIT_QUEUE_HEAD(au_hfsn_wq); +static __cacheline_aligned_in_smp atomic64_t au_hfsn_ifree = ATOMIC64_INIT(0); + +static void au_hfsn_free_mark(struct fsnotify_mark *mark) +{ + struct au_hnotify *hn = container_of(mark, struct au_hnotify, + hn_mark); + /* AuDbg("here\n"); */ + au_cache_free_hnotify(hn); + smp_mb__before_atomic(); /* for atomic64_dec */ + if (atomic64_dec_and_test(&au_hfsn_ifree)) + wake_up(&au_hfsn_wq); +} + +static int au_hfsn_alloc(struct au_hinode *hinode) +{ + int err; + struct au_hnotify *hn; + struct super_block *sb; + struct au_branch *br; + struct fsnotify_mark *mark; + aufs_bindex_t bindex; + + hn = hinode->hi_notify; + sb = hn->hn_aufs_inode->i_sb; + bindex = au_br_index(sb, hinode->hi_id); + br = au_sbr(sb, bindex); + AuDebugOn(!br->br_hfsn); + + mark = &hn->hn_mark; + fsnotify_init_mark(mark, br->br_hfsn->hfsn_group); + mark->mask = AuHfsnMask; + /* + * by udba rename or rmdir, aufs assign a new inode to the known + * h_inode, so specify 1 to allow dups. + */ + lockdep_off(); + err = fsnotify_add_inode_mark(mark, hinode->hi_inode, /*allow_dups*/1); + lockdep_on(); + + return err; +} + +static int au_hfsn_free(struct au_hinode *hinode, struct au_hnotify *hn) +{ + struct fsnotify_mark *mark; + unsigned long long ull; + struct fsnotify_group *group; + + ull = atomic64_inc_return(&au_hfsn_ifree); + BUG_ON(!ull); + + mark = &hn->hn_mark; + spin_lock(&mark->lock); + group = mark->group; + fsnotify_get_group(group); + spin_unlock(&mark->lock); + lockdep_off(); + fsnotify_destroy_mark(mark, group); + fsnotify_put_mark(mark); + fsnotify_put_group(group); + lockdep_on(); + + /* free hn by myself */ + return 0; +} + +/* ---------------------------------------------------------------------- */ + +static void au_hfsn_ctl(struct au_hinode *hinode, int do_set) +{ + struct fsnotify_mark *mark; + + mark = &hinode->hi_notify->hn_mark; + spin_lock(&mark->lock); + if (do_set) { + AuDebugOn(mark->mask & AuHfsnMask); + mark->mask |= AuHfsnMask; + } else { + AuDebugOn(!(mark->mask & AuHfsnMask)); + mark->mask &= ~AuHfsnMask; + } + spin_unlock(&mark->lock); + /* fsnotify_recalc_inode_mask(hinode->hi_inode); */ +} + +/* ---------------------------------------------------------------------- */ + +/* #define AuDbgHnotify */ +#ifdef AuDbgHnotify +static char *au_hfsn_name(u32 mask) +{ +#ifdef CONFIG_AUFS_DEBUG +#define test_ret(flag) \ + do { \ + if (mask & flag) \ + return #flag; \ + } while (0) + test_ret(FS_ACCESS); + test_ret(FS_MODIFY); + test_ret(FS_ATTRIB); + test_ret(FS_CLOSE_WRITE); + test_ret(FS_CLOSE_NOWRITE); + test_ret(FS_OPEN); + test_ret(FS_MOVED_FROM); + test_ret(FS_MOVED_TO); + test_ret(FS_CREATE); + test_ret(FS_DELETE); + test_ret(FS_DELETE_SELF); + test_ret(FS_MOVE_SELF); + test_ret(FS_UNMOUNT); + test_ret(FS_Q_OVERFLOW); + test_ret(FS_IN_IGNORED); + test_ret(FS_ISDIR); + test_ret(FS_IN_ONESHOT); + test_ret(FS_EVENT_ON_CHILD); + return ""; +#undef test_ret +#else + return "??"; +#endif +} +#endif + +/* ---------------------------------------------------------------------- */ + +static void au_hfsn_free_group(struct fsnotify_group *group) +{ + struct au_br_hfsnotify *hfsn = group->private; + + /* AuDbg("here\n"); */ + au_kfree_try_rcu(hfsn); +} + +static int au_hfsn_handle_event(struct fsnotify_group *group, + u32 mask, const void *data, int data_type, + struct inode *dir, + const struct qstr *file_name, u32 cookie, + struct fsnotify_iter_info *iter_info) +{ + int err; + struct au_hnotify *hnotify; + struct inode *h_dir, *h_inode; + struct fsnotify_mark *inode_mark; + + AuDebugOn(data_type != FSNOTIFY_EVENT_INODE); + + err = 0; + /* if FS_UNMOUNT happens, there must be another bug */ + AuDebugOn(mask & FS_UNMOUNT); + if (mask & (FS_IN_IGNORED | FS_UNMOUNT)) + goto out; + + h_dir = dir; + h_inode = NULL; +#ifdef AuDbgHnotify + au_debug_on(); + if (1 || h_child_qstr.len != sizeof(AUFS_XINO_FNAME) - 1 + || strncmp(h_child_qstr.name, AUFS_XINO_FNAME, h_child_qstr.len)) { + AuDbg("i%lu, mask 0x%x %s, hcname %.*s, hi%lu\n", + h_dir->i_ino, mask, au_hfsn_name(mask), + AuLNPair(&h_child_qstr), h_inode ? h_inode->i_ino : 0); + /* WARN_ON(1); */ + } + au_debug_off(); +#endif + + inode_mark = fsnotify_iter_inode_mark(iter_info); + AuDebugOn(!inode_mark); + hnotify = container_of(inode_mark, struct au_hnotify, hn_mark); + err = au_hnotify(h_dir, hnotify, mask, file_name, h_inode); + +out: + return err; +} + +static struct fsnotify_ops au_hfsn_ops = { + .handle_event = au_hfsn_handle_event, + .free_group_priv = au_hfsn_free_group, + .free_mark = au_hfsn_free_mark +}; + +/* ---------------------------------------------------------------------- */ + +static void au_hfsn_fin_br(struct au_branch *br) +{ + struct au_br_hfsnotify *hfsn; + + hfsn = br->br_hfsn; + if (hfsn) { + lockdep_off(); + fsnotify_put_group(hfsn->hfsn_group); + lockdep_on(); + } +} + +static int au_hfsn_init_br(struct au_branch *br, int perm) +{ + int err; + struct fsnotify_group *group; + struct au_br_hfsnotify *hfsn; + + err = 0; + br->br_hfsn = NULL; + if (!au_br_hnotifyable(perm)) + goto out; + + err = -ENOMEM; + hfsn = kmalloc(sizeof(*hfsn), GFP_NOFS); + if (unlikely(!hfsn)) + goto out; + + err = 0; + group = fsnotify_alloc_group(&au_hfsn_ops); + if (IS_ERR(group)) { + err = PTR_ERR(group); + pr_err("fsnotify_alloc_group() failed, %d\n", err); + goto out_hfsn; + } + + group->private = hfsn; + hfsn->hfsn_group = group; + br->br_hfsn = hfsn; + goto out; /* success */ + +out_hfsn: + au_kfree_try_rcu(hfsn); +out: + return err; +} + +static int au_hfsn_reset_br(unsigned int udba, struct au_branch *br, int perm) +{ + int err; + + err = 0; + if (!br->br_hfsn) + err = au_hfsn_init_br(br, perm); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static void au_hfsn_fin(void) +{ + AuDbg("au_hfsn_ifree %lld\n", (long long)atomic64_read(&au_hfsn_ifree)); + wait_event(au_hfsn_wq, !atomic64_read(&au_hfsn_ifree)); +} + +const struct au_hnotify_op au_hnotify_op = { + .ctl = au_hfsn_ctl, + .alloc = au_hfsn_alloc, + .free = au_hfsn_free, + + .fin = au_hfsn_fin, + + .reset_br = au_hfsn_reset_br, + .fin_br = au_hfsn_fin_br, + .init_br = au_hfsn_init_br +}; diff --git a/fs/aufs/hfsplus.c b/fs/aufs/hfsplus.c new file mode 100644 index 00000000000..d250f019bed --- /dev/null +++ b/fs/aufs/hfsplus.c @@ -0,0 +1,60 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2010-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * special support for filesystems which acquires an inode mutex + * at final closing a file, eg, hfsplus. + * + * This trick is very simple and stupid, just to open the file before really + * necessary open to tell hfsplus that this is not the final closing. + * The caller should call au_h_open_pre() after acquiring the inode mutex, + * and au_h_open_post() after releasing it. + */ + +#include "aufs.h" + +struct file *au_h_open_pre(struct dentry *dentry, aufs_bindex_t bindex, + int force_wr) +{ + struct file *h_file; + struct dentry *h_dentry; + + h_dentry = au_h_dptr(dentry, bindex); + AuDebugOn(!h_dentry); + AuDebugOn(d_is_negative(h_dentry)); + + h_file = NULL; + if (au_test_hfsplus(h_dentry->d_sb) + && d_is_reg(h_dentry)) + h_file = au_h_open(dentry, bindex, + O_RDONLY | O_NOATIME | O_LARGEFILE, + /*file*/NULL, force_wr); + return h_file; +} + +void au_h_open_post(struct dentry *dentry, aufs_bindex_t bindex, + struct file *h_file) +{ + struct au_branch *br; + + if (h_file) { + fput(h_file); + br = au_sbr(dentry->d_sb, bindex); + au_lcnt_dec(&br->br_nfiles); + } +} diff --git a/fs/aufs/hnotify.c b/fs/aufs/hnotify.c new file mode 100644 index 00000000000..ff3f8b96d4e --- /dev/null +++ b/fs/aufs/hnotify.c @@ -0,0 +1,715 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * abstraction to notify the direct changes on lower directories + */ + +/* #include */ +#include "aufs.h" + +int au_hn_alloc(struct au_hinode *hinode, struct inode *inode) +{ + int err; + struct au_hnotify *hn; + + err = -ENOMEM; + hn = au_cache_alloc_hnotify(); + if (hn) { + hn->hn_aufs_inode = inode; + hinode->hi_notify = hn; + err = au_hnotify_op.alloc(hinode); + AuTraceErr(err); + if (unlikely(err)) { + hinode->hi_notify = NULL; + au_cache_free_hnotify(hn); + /* + * The upper dir was removed by udba, but the same named + * dir left. In this case, aufs assigns a new inode + * number and set the monitor again. + * For the lower dir, the old monitor is still left. + */ + if (err == -EEXIST) + err = 0; + } + } + + AuTraceErr(err); + return err; +} + +void au_hn_free(struct au_hinode *hinode) +{ + struct au_hnotify *hn; + + hn = hinode->hi_notify; + if (hn) { + hinode->hi_notify = NULL; + if (au_hnotify_op.free(hinode, hn)) + au_cache_free_hnotify(hn); + } +} + +/* ---------------------------------------------------------------------- */ + +void au_hn_ctl(struct au_hinode *hinode, int do_set) +{ + if (hinode->hi_notify) + au_hnotify_op.ctl(hinode, do_set); +} + +void au_hn_reset(struct inode *inode, unsigned int flags) +{ + aufs_bindex_t bindex, bbot; + struct inode *hi; + struct dentry *iwhdentry; + + bbot = au_ibbot(inode); + for (bindex = au_ibtop(inode); bindex <= bbot; bindex++) { + hi = au_h_iptr(inode, bindex); + if (!hi) + continue; + + /* inode_lock_nested(hi, AuLsc_I_CHILD); */ + iwhdentry = au_hi_wh(inode, bindex); + if (iwhdentry) + dget(iwhdentry); + au_igrab(hi); + au_set_h_iptr(inode, bindex, NULL, 0); + au_set_h_iptr(inode, bindex, au_igrab(hi), + flags & ~AuHi_XINO); + iput(hi); + dput(iwhdentry); + /* inode_unlock(hi); */ + } +} + +/* ---------------------------------------------------------------------- */ + +static int hn_xino(struct inode *inode, struct inode *h_inode) +{ + int err; + aufs_bindex_t bindex, bbot, bfound, btop; + struct inode *h_i; + + err = 0; + if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { + pr_warn("branch root dir was changed\n"); + goto out; + } + + bfound = -1; + bbot = au_ibbot(inode); + btop = au_ibtop(inode); +#if 0 /* reserved for future use */ + if (bindex == bbot) { + /* keep this ino in rename case */ + goto out; + } +#endif + for (bindex = btop; bindex <= bbot; bindex++) + if (au_h_iptr(inode, bindex) == h_inode) { + bfound = bindex; + break; + } + if (bfound < 0) + goto out; + + for (bindex = btop; bindex <= bbot; bindex++) { + h_i = au_h_iptr(inode, bindex); + if (!h_i) + continue; + + err = au_xino_write(inode->i_sb, bindex, h_i->i_ino, /*ino*/0); + /* ignore this error */ + /* bad action? */ + } + + /* children inode number will be broken */ + +out: + AuTraceErr(err); + return err; +} + +static int hn_gen_tree(struct dentry *dentry) +{ + int err, i, j, ndentry; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry **dentries; + + err = au_dpages_init(&dpages, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_dcsub_pages(&dpages, dentry, NULL, NULL); + if (unlikely(err)) + goto out_dpages; + + for (i = 0; i < dpages.ndpage; i++) { + dpage = dpages.dpages + i; + dentries = dpage->dentries; + ndentry = dpage->ndentry; + for (j = 0; j < ndentry; j++) { + struct dentry *d; + + d = dentries[j]; + if (IS_ROOT(d)) + continue; + + au_digen_dec(d); + if (d_really_is_positive(d)) + /* todo: reset children xino? + cached children only? */ + au_iigen_dec(d_inode(d)); + } + } + +out_dpages: + au_dpages_free(&dpages); +out: + return err; +} + +/* + * return 0 if processed. + */ +static int hn_gen_by_inode(char *name, unsigned int nlen, struct inode *inode, + const unsigned int isdir) +{ + int err; + struct dentry *d; + struct qstr *dname; + + err = 1; + if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { + pr_warn("branch root dir was changed\n"); + err = 0; + goto out; + } + + if (!isdir) { + AuDebugOn(!name); + au_iigen_dec(inode); + spin_lock(&inode->i_lock); + hlist_for_each_entry(d, &inode->i_dentry, d_u.d_alias) { + spin_lock(&d->d_lock); + dname = &d->d_name; + if (dname->len != nlen + && memcmp(dname->name, name, nlen)) { + spin_unlock(&d->d_lock); + continue; + } + err = 0; + au_digen_dec(d); + spin_unlock(&d->d_lock); + break; + } + spin_unlock(&inode->i_lock); + } else { + au_fset_si(au_sbi(inode->i_sb), FAILED_REFRESH_DIR); + d = d_find_any_alias(inode); + if (!d) { + au_iigen_dec(inode); + goto out; + } + + spin_lock(&d->d_lock); + dname = &d->d_name; + if (dname->len == nlen && !memcmp(dname->name, name, nlen)) { + spin_unlock(&d->d_lock); + err = hn_gen_tree(d); + spin_lock(&d->d_lock); + } + spin_unlock(&d->d_lock); + dput(d); + } + +out: + AuTraceErr(err); + return err; +} + +static int hn_gen_by_name(struct dentry *dentry, const unsigned int isdir) +{ + int err; + + if (IS_ROOT(dentry)) { + pr_warn("branch root dir was changed\n"); + return 0; + } + + err = 0; + if (!isdir) { + au_digen_dec(dentry); + if (d_really_is_positive(dentry)) + au_iigen_dec(d_inode(dentry)); + } else { + au_fset_si(au_sbi(dentry->d_sb), FAILED_REFRESH_DIR); + if (d_really_is_positive(dentry)) + err = hn_gen_tree(dentry); + } + + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* hnotify job flags */ +#define AuHnJob_XINO0 1 +#define AuHnJob_GEN (1 << 1) +#define AuHnJob_DIRENT (1 << 2) +#define AuHnJob_ISDIR (1 << 3) +#define AuHnJob_TRYXINO0 (1 << 4) +#define AuHnJob_MNTPNT (1 << 5) +#define au_ftest_hnjob(flags, name) ((flags) & AuHnJob_##name) +#define au_fset_hnjob(flags, name) \ + do { (flags) |= AuHnJob_##name; } while (0) +#define au_fclr_hnjob(flags, name) \ + do { (flags) &= ~AuHnJob_##name; } while (0) + +enum { + AuHn_CHILD, + AuHn_PARENT, + AuHnLast +}; + +struct au_hnotify_args { + struct inode *h_dir, *dir, *h_child_inode; + u32 mask; + unsigned int flags[AuHnLast]; + unsigned int h_child_nlen; + char h_child_name[]; +}; + +struct hn_job_args { + unsigned int flags; + struct inode *inode, *h_inode, *dir, *h_dir; + struct dentry *dentry; + char *h_name; + int h_nlen; +}; + +static int hn_job(struct hn_job_args *a) +{ + const unsigned int isdir = au_ftest_hnjob(a->flags, ISDIR); + int e; + + /* reset xino */ + if (au_ftest_hnjob(a->flags, XINO0) && a->inode) + hn_xino(a->inode, a->h_inode); /* ignore this error */ + + if (au_ftest_hnjob(a->flags, TRYXINO0) + && a->inode + && a->h_inode) { + inode_lock_shared_nested(a->h_inode, AuLsc_I_CHILD); + if (!a->h_inode->i_nlink + && !(a->h_inode->i_state & I_LINKABLE)) + hn_xino(a->inode, a->h_inode); /* ignore this error */ + inode_unlock_shared(a->h_inode); + } + + /* make the generation obsolete */ + if (au_ftest_hnjob(a->flags, GEN)) { + e = -1; + if (a->inode) + e = hn_gen_by_inode(a->h_name, a->h_nlen, a->inode, + isdir); + if (e && a->dentry) + hn_gen_by_name(a->dentry, isdir); + /* ignore this error */ + } + + /* make dir entries obsolete */ + if (au_ftest_hnjob(a->flags, DIRENT) && a->inode) { + struct au_vdir *vdir; + + vdir = au_ivdir(a->inode); + if (vdir) + vdir->vd_jiffy = 0; + /* IMustLock(a->inode); */ + /* inode_inc_iversion(a->inode); */ + } + + /* can do nothing but warn */ + if (au_ftest_hnjob(a->flags, MNTPNT) + && a->dentry + && d_mountpoint(a->dentry)) + pr_warn("mount-point %pd is removed or renamed\n", a->dentry); + + return 0; +} + +/* ---------------------------------------------------------------------- */ + +static struct dentry *lookup_wlock_by_name(char *name, unsigned int nlen, + struct inode *dir) +{ + struct dentry *dentry, *d, *parent; + struct qstr *dname; + + parent = d_find_any_alias(dir); + if (!parent) + return NULL; + + dentry = NULL; + spin_lock(&parent->d_lock); + list_for_each_entry(d, &parent->d_subdirs, d_child) { + /* AuDbg("%pd\n", d); */ + spin_lock_nested(&d->d_lock, DENTRY_D_LOCK_NESTED); + dname = &d->d_name; + if (dname->len != nlen || memcmp(dname->name, name, nlen)) + goto cont_unlock; + if (au_di(d)) + au_digen_dec(d); + else + goto cont_unlock; + if (au_dcount(d) > 0) { + dentry = dget_dlock(d); + spin_unlock(&d->d_lock); + break; + } + +cont_unlock: + spin_unlock(&d->d_lock); + } + spin_unlock(&parent->d_lock); + dput(parent); + + if (dentry) + di_write_lock_child(dentry); + + return dentry; +} + +static struct inode *lookup_wlock_by_ino(struct super_block *sb, + aufs_bindex_t bindex, ino_t h_ino) +{ + struct inode *inode; + ino_t ino; + int err; + + inode = NULL; + err = au_xino_read(sb, bindex, h_ino, &ino); + if (!err && ino) + inode = ilookup(sb, ino); + if (!inode) + goto out; + + if (unlikely(inode->i_ino == AUFS_ROOT_INO)) { + pr_warn("wrong root branch\n"); + iput(inode); + inode = NULL; + goto out; + } + + ii_write_lock_child(inode); + +out: + return inode; +} + +static void au_hn_bh(void *_args) +{ + struct au_hnotify_args *a = _args; + struct super_block *sb; + aufs_bindex_t bindex, bbot, bfound; + unsigned char xino, try_iput; + int err; + struct inode *inode; + ino_t h_ino; + struct hn_job_args args; + struct dentry *dentry; + struct au_sbinfo *sbinfo; + + AuDebugOn(!_args); + AuDebugOn(!a->h_dir); + AuDebugOn(!a->dir); + AuDebugOn(!a->mask); + AuDbg("mask 0x%x, i%lu, hi%lu, hci%lu\n", + a->mask, a->dir->i_ino, a->h_dir->i_ino, + a->h_child_inode ? a->h_child_inode->i_ino : 0); + + inode = NULL; + dentry = NULL; + /* + * do not lock a->dir->i_mutex here + * because of d_revalidate() may cause a deadlock. + */ + sb = a->dir->i_sb; + AuDebugOn(!sb); + sbinfo = au_sbi(sb); + AuDebugOn(!sbinfo); + si_write_lock(sb, AuLock_NOPLMW); + + if (au_opt_test(sbinfo->si_mntflags, DIRREN)) + switch (a->mask & FS_EVENTS_POSS_ON_CHILD) { + case FS_MOVED_FROM: + case FS_MOVED_TO: + AuWarn1("DIRREN with UDBA may not work correctly " + "for the direct rename(2)\n"); + } + + ii_read_lock_parent(a->dir); + bfound = -1; + bbot = au_ibbot(a->dir); + for (bindex = au_ibtop(a->dir); bindex <= bbot; bindex++) + if (au_h_iptr(a->dir, bindex) == a->h_dir) { + bfound = bindex; + break; + } + ii_read_unlock(a->dir); + if (unlikely(bfound < 0)) + goto out; + + xino = !!au_opt_test(au_mntflags(sb), XINO); + h_ino = 0; + if (a->h_child_inode) + h_ino = a->h_child_inode->i_ino; + + if (a->h_child_nlen + && (au_ftest_hnjob(a->flags[AuHn_CHILD], GEN) + || au_ftest_hnjob(a->flags[AuHn_CHILD], MNTPNT))) + dentry = lookup_wlock_by_name(a->h_child_name, a->h_child_nlen, + a->dir); + try_iput = 0; + if (dentry && d_really_is_positive(dentry)) + inode = d_inode(dentry); + if (xino && !inode && h_ino + && (au_ftest_hnjob(a->flags[AuHn_CHILD], XINO0) + || au_ftest_hnjob(a->flags[AuHn_CHILD], TRYXINO0) + || au_ftest_hnjob(a->flags[AuHn_CHILD], GEN))) { + inode = lookup_wlock_by_ino(sb, bfound, h_ino); + try_iput = 1; + } + + args.flags = a->flags[AuHn_CHILD]; + args.dentry = dentry; + args.inode = inode; + args.h_inode = a->h_child_inode; + args.dir = a->dir; + args.h_dir = a->h_dir; + args.h_name = a->h_child_name; + args.h_nlen = a->h_child_nlen; + err = hn_job(&args); + if (dentry) { + if (au_di(dentry)) + di_write_unlock(dentry); + dput(dentry); + } + if (inode && try_iput) { + ii_write_unlock(inode); + iput(inode); + } + + ii_write_lock_parent(a->dir); + args.flags = a->flags[AuHn_PARENT]; + args.dentry = NULL; + args.inode = a->dir; + args.h_inode = a->h_dir; + args.dir = NULL; + args.h_dir = NULL; + args.h_name = NULL; + args.h_nlen = 0; + err = hn_job(&args); + ii_write_unlock(a->dir); + +out: + iput(a->h_child_inode); + iput(a->h_dir); + iput(a->dir); + si_write_unlock(sb); + au_nwt_done(&sbinfo->si_nowait); + au_kfree_rcu(a); +} + +/* ---------------------------------------------------------------------- */ + +int au_hnotify(struct inode *h_dir, struct au_hnotify *hnotify, u32 mask, + const struct qstr *h_child_qstr, struct inode *h_child_inode) +{ + int err, len; + unsigned int flags[AuHnLast], f; + unsigned char isdir, isroot, wh; + struct inode *dir; + struct au_hnotify_args *args; + char *p, *h_child_name; + + err = 0; + AuDebugOn(!hnotify || !hnotify->hn_aufs_inode); + dir = igrab(hnotify->hn_aufs_inode); + if (!dir) + goto out; + + isroot = (dir->i_ino == AUFS_ROOT_INO); + wh = 0; + h_child_name = (void *)h_child_qstr->name; + len = h_child_qstr->len; + if (h_child_name) { + if (len > AUFS_WH_PFX_LEN + && !memcmp(h_child_name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { + h_child_name += AUFS_WH_PFX_LEN; + len -= AUFS_WH_PFX_LEN; + wh = 1; + } + } + + isdir = 0; + if (h_child_inode) + isdir = !!S_ISDIR(h_child_inode->i_mode); + flags[AuHn_PARENT] = AuHnJob_ISDIR; + flags[AuHn_CHILD] = 0; + if (isdir) + flags[AuHn_CHILD] = AuHnJob_ISDIR; + au_fset_hnjob(flags[AuHn_PARENT], DIRENT); + au_fset_hnjob(flags[AuHn_CHILD], GEN); + switch (mask & ALL_FSNOTIFY_DIRENT_EVENTS) { + case FS_MOVED_FROM: + case FS_MOVED_TO: + au_fset_hnjob(flags[AuHn_CHILD], XINO0); + au_fset_hnjob(flags[AuHn_CHILD], MNTPNT); + fallthrough; + case FS_CREATE: + AuDebugOn(!h_child_name); + break; + + case FS_DELETE: + /* + * aufs never be able to get this child inode. + * revalidation should be in d_revalidate() + * by checking i_nlink, i_generation or d_unhashed(). + */ + AuDebugOn(!h_child_name); + au_fset_hnjob(flags[AuHn_CHILD], TRYXINO0); + au_fset_hnjob(flags[AuHn_CHILD], MNTPNT); + break; + + default: + AuDebugOn(1); + } + + if (wh) + h_child_inode = NULL; + + err = -ENOMEM; + /* iput() and kfree() will be called in au_hnotify() */ + args = kmalloc(sizeof(*args) + len + 1, GFP_NOFS); + if (unlikely(!args)) { + AuErr1("no memory\n"); + iput(dir); + goto out; + } + args->flags[AuHn_PARENT] = flags[AuHn_PARENT]; + args->flags[AuHn_CHILD] = flags[AuHn_CHILD]; + args->mask = mask; + args->dir = dir; + args->h_dir = igrab(h_dir); + if (h_child_inode) + h_child_inode = igrab(h_child_inode); /* can be NULL */ + args->h_child_inode = h_child_inode; + args->h_child_nlen = len; + if (len) { + p = (void *)args; + p += sizeof(*args); + memcpy(p, h_child_name, len); + p[len] = 0; + } + + /* NFS fires the event for silly-renamed one from kworker */ + f = 0; + if (!dir->i_nlink + || (au_test_nfs(h_dir->i_sb) && (mask & FS_DELETE))) + f = AuWkq_NEST; + err = au_wkq_nowait(au_hn_bh, args, dir->i_sb, f); + if (unlikely(err)) { + pr_err("wkq %d\n", err); + iput(args->h_child_inode); + iput(args->h_dir); + iput(args->dir); + au_kfree_rcu(args); + } + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_hnotify_reset_br(unsigned int udba, struct au_branch *br, int perm) +{ + int err; + + AuDebugOn(!(udba & AuOptMask_UDBA)); + + err = 0; + if (au_hnotify_op.reset_br) + err = au_hnotify_op.reset_br(udba, br, perm); + + return err; +} + +int au_hnotify_init_br(struct au_branch *br, int perm) +{ + int err; + + err = 0; + if (au_hnotify_op.init_br) + err = au_hnotify_op.init_br(br, perm); + + return err; +} + +void au_hnotify_fin_br(struct au_branch *br) +{ + if (au_hnotify_op.fin_br) + au_hnotify_op.fin_br(br); +} + +static void au_hn_destroy_cache(void) +{ + kmem_cache_destroy(au_cache[AuCache_HNOTIFY]); + au_cache[AuCache_HNOTIFY] = NULL; +} + +int __init au_hnotify_init(void) +{ + int err; + + err = -ENOMEM; + au_cache[AuCache_HNOTIFY] = AuCache(au_hnotify); + if (au_cache[AuCache_HNOTIFY]) { + err = 0; + if (au_hnotify_op.init) + err = au_hnotify_op.init(); + if (unlikely(err)) + au_hn_destroy_cache(); + } + AuTraceErr(err); + return err; +} + +void au_hnotify_fin(void) +{ + if (au_hnotify_op.fin) + au_hnotify_op.fin(); + + /* cf. au_cache_fin() */ + if (au_cache[AuCache_HNOTIFY]) + au_hn_destroy_cache(); +} diff --git a/fs/aufs/i_op.c b/fs/aufs/i_op.c new file mode 100644 index 00000000000..9826b9d58c3 --- /dev/null +++ b/fs/aufs/i_op.c @@ -0,0 +1,1513 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode operations (except add/del/rename) + */ + +#include +#include +#include +#include +#include +#include "aufs.h" + +static int h_permission(struct inode *h_inode, int mask, + struct path *h_path, int brperm) +{ + int err; + const unsigned char write_mask = !!(mask & (MAY_WRITE | MAY_APPEND)); + struct user_namespace *h_userns; + + err = -EPERM; + if (write_mask && IS_IMMUTABLE(h_inode)) + goto out; + + err = -EACCES; + if (((mask & MAY_EXEC) + && S_ISREG(h_inode->i_mode) + && (path_noexec(h_path) + || !(h_inode->i_mode & 0111)))) + goto out; + + /* + * - skip the lower fs test in the case of write to ro branch. + * - nfs dir permission write check is optimized, but a policy for + * link/rename requires a real check. + * - nfs always sets SB_POSIXACL regardless its mount option 'noacl.' + * in this case, generic_permission() returns -EOPNOTSUPP. + */ + h_userns = mnt_user_ns(h_path->mnt); + if ((write_mask && !au_br_writable(brperm)) + || (au_test_nfs(h_inode->i_sb) && S_ISDIR(h_inode->i_mode) + && write_mask && !(mask & MAY_READ)) + || !h_inode->i_op->permission) { + /* AuLabel(generic_permission); */ + /* AuDbg("get_acl %ps\n", h_inode->i_op->get_acl); */ + err = generic_permission(h_userns, h_inode, mask); + if (err == -EOPNOTSUPP && au_test_nfs_noacl(h_inode)) + err = h_inode->i_op->permission(h_userns, h_inode, + mask); + AuTraceErr(err); + } else { + /* AuLabel(h_inode->permission); */ + err = h_inode->i_op->permission(h_userns, h_inode, mask); + AuTraceErr(err); + } + + if (!err) + err = devcgroup_inode_permission(h_inode, mask); + if (!err) + err = security_inode_permission(h_inode, mask); + +out: + return err; +} + +static int aufs_permission(struct user_namespace *userns, struct inode *inode, + int mask) +{ + int err; + aufs_bindex_t bindex, bbot; + const unsigned char isdir = !!S_ISDIR(inode->i_mode), + write_mask = !!(mask & (MAY_WRITE | MAY_APPEND)); + struct inode *h_inode; + struct super_block *sb; + struct au_branch *br; + + /* todo: support rcu-walk? */ + if (mask & MAY_NOT_BLOCK) + return -ECHILD; + + sb = inode->i_sb; + si_read_lock(sb, AuLock_FLUSH); + ii_read_lock_child(inode); +#if 0 /* reserved for future use */ + /* + * This test may be rather 'too much' since the test is essentially done + * in the aufs_lookup(). Theoretically it is possible that the inode + * generation doesn't match to the superblock's here. But it isn't a + * big deal I suppose. + */ + err = au_iigen_test(inode, au_sigen(sb)); + if (unlikely(err)) + goto out; +#endif + + if (!isdir + || write_mask + || au_opt_test(au_mntflags(sb), DIRPERM1)) { + err = au_busy_or_stale(); + h_inode = au_h_iptr(inode, au_ibtop(inode)); + if (unlikely(!h_inode + || (h_inode->i_mode & S_IFMT) + != (inode->i_mode & S_IFMT))) + goto out; + + err = 0; + bindex = au_ibtop(inode); + br = au_sbr(sb, bindex); + err = h_permission(h_inode, mask, &br->br_path, br->br_perm); + if (write_mask + && !err + && !special_file(h_inode->i_mode)) { + /* test whether the upper writable branch exists */ + err = -EROFS; + for (; bindex >= 0; bindex--) + if (!au_br_rdonly(au_sbr(sb, bindex))) { + err = 0; + break; + } + } + goto out; + } + + /* non-write to dir */ + err = 0; + bbot = au_ibbot(inode); + for (bindex = au_ibtop(inode); !err && bindex <= bbot; bindex++) { + h_inode = au_h_iptr(inode, bindex); + if (h_inode) { + err = au_busy_or_stale(); + if (unlikely(!S_ISDIR(h_inode->i_mode))) + break; + + br = au_sbr(sb, bindex); + err = h_permission(h_inode, mask, &br->br_path, + br->br_perm); + } + } + +out: + ii_read_unlock(inode); + si_read_unlock(sb); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static struct dentry *aufs_lookup(struct inode *dir, struct dentry *dentry, + unsigned int flags) +{ + struct dentry *ret, *parent; + struct inode *inode; + struct super_block *sb; + int err, npositive; + + IMustLock(dir); + + /* todo: support rcu-walk? */ + ret = ERR_PTR(-ECHILD); + if (flags & LOOKUP_RCU) + goto out; + + ret = ERR_PTR(-ENAMETOOLONG); + if (unlikely(dentry->d_name.len > AUFS_MAX_NAMELEN)) + goto out; + + sb = dir->i_sb; + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + ret = ERR_PTR(err); + if (unlikely(err)) + goto out; + + err = au_di_init(dentry); + ret = ERR_PTR(err); + if (unlikely(err)) + goto out_si; + + inode = NULL; + npositive = 0; /* suppress a warning */ + parent = dentry->d_parent; /* dir inode is locked */ + di_read_lock_parent(parent, AuLock_IR); + err = au_alive_dir(parent); + if (!err) + err = au_digen_test(parent, au_sigen(sb)); + if (!err) { + /* regardless LOOKUP_CREATE, always ALLOW_NEG */ + npositive = au_lkup_dentry(dentry, au_dbtop(parent), + AuLkup_ALLOW_NEG); + err = npositive; + } + di_read_unlock(parent, AuLock_IR); + ret = ERR_PTR(err); + if (unlikely(err < 0)) + goto out_unlock; + + if (npositive) { + inode = au_new_inode(dentry, /*must_new*/0); + if (IS_ERR(inode)) { + ret = (void *)inode; + inode = NULL; + goto out_unlock; + } + } + + if (inode) + atomic_inc(&inode->i_count); + ret = d_splice_alias(inode, dentry); +#if 0 /* reserved for future use */ + if (unlikely(d_need_lookup(dentry))) { + spin_lock(&dentry->d_lock); + dentry->d_flags &= ~DCACHE_NEED_LOOKUP; + spin_unlock(&dentry->d_lock); + } else +#endif + if (inode) { + if (!IS_ERR(ret)) { + iput(inode); + if (ret && ret != dentry) + ii_write_unlock(inode); + } else { + ii_write_unlock(inode); + iput(inode); + inode = NULL; + } + } + +out_unlock: + di_write_unlock(dentry); +out_si: + si_read_unlock(sb); +out: + return ret; +} + +/* ---------------------------------------------------------------------- */ + +/* + * very dirty and complicated aufs ->atomic_open(). + * aufs_atomic_open() + * + au_aopen_or_create() + * + add_simple() + * + vfsub_atomic_open() + * + branch fs ->atomic_open() + * may call the actual 'open' for h_file + * + inc br_nfiles only if opened + * + au_aopen_no_open() or au_aopen_do_open() + * + * au_aopen_do_open() + * + finish_open() + * + au_do_aopen() + * + au_do_open() the body of all 'open' + * + au_do_open_nondir() + * set the passed h_file + * + * au_aopen_no_open() + * + finish_no_open() + */ + +struct aopen_node { + struct hlist_bl_node hblist; + struct file *file, *h_file; +}; + +static int au_do_aopen(struct inode *inode, struct file *file) +{ + struct hlist_bl_head *aopen; + struct hlist_bl_node *pos; + struct aopen_node *node; + struct au_do_open_args args = { + .aopen = 1, + .open = au_do_open_nondir + }; + + aopen = &au_sbi(inode->i_sb)->si_aopen; + hlist_bl_lock(aopen); + hlist_bl_for_each_entry(node, pos, aopen, hblist) + if (node->file == file) { + args.h_file = node->h_file; + break; + } + hlist_bl_unlock(aopen); + /* AuDebugOn(!args.h_file); */ + + return au_do_open(file, &args); +} + +static int au_aopen_do_open(struct file *file, struct dentry *dentry, + struct aopen_node *aopen_node) +{ + int err; + struct hlist_bl_head *aopen; + + AuLabel(here); + aopen = &au_sbi(dentry->d_sb)->si_aopen; + au_hbl_add(&aopen_node->hblist, aopen); + err = finish_open(file, dentry, au_do_aopen); + au_hbl_del(&aopen_node->hblist, aopen); + /* AuDbgFile(file); */ + AuDbg("%pd%s%s\n", dentry, + (file->f_mode & FMODE_CREATED) ? " created" : "", + (file->f_mode & FMODE_OPENED) ? " opened" : ""); + + AuTraceErr(err); + return err; +} + +static int au_aopen_no_open(struct file *file, struct dentry *dentry) +{ + int err; + + AuLabel(here); + dget(dentry); + err = finish_no_open(file, dentry); + + AuTraceErr(err); + return err; +} + +static int aufs_atomic_open(struct inode *dir, struct dentry *dentry, + struct file *file, unsigned int open_flag, + umode_t create_mode) +{ + int err, did_open; + unsigned int lkup_flags; + aufs_bindex_t bindex; + struct super_block *sb; + struct dentry *parent, *d; + struct vfsub_aopen_args args = { + .open_flag = open_flag, + .create_mode = create_mode + }; + struct aopen_node aopen_node = { + .file = file + }; + + IMustLock(dir); + AuDbg("open_flag 0%o\n", open_flag); + AuDbgDentry(dentry); + + err = 0; + if (!au_di(dentry)) { + lkup_flags = LOOKUP_OPEN; + if (open_flag & O_CREAT) + lkup_flags |= LOOKUP_CREATE; + d = aufs_lookup(dir, dentry, lkup_flags); + if (IS_ERR(d)) { + err = PTR_ERR(d); + AuTraceErr(err); + goto out; + } else if (d) { + /* + * obsoleted dentry found. + * another error will be returned later. + */ + d_drop(d); + AuDbgDentry(d); + dput(d); + } + AuDbgDentry(dentry); + } + + if (d_is_positive(dentry) + || d_unhashed(dentry) + || d_unlinked(dentry) + || !(open_flag & O_CREAT)) { + err = au_aopen_no_open(file, dentry); + goto out; /* success */ + } + + err = aufs_read_lock(dentry, AuLock_DW | AuLock_FLUSH | AuLock_GEN); + if (unlikely(err)) + goto out; + + sb = dentry->d_sb; + parent = dentry->d_parent; /* dir is locked */ + di_write_lock_parent(parent); + err = au_lkup_dentry(dentry, /*btop*/0, AuLkup_ALLOW_NEG); + if (unlikely(err < 0)) + goto out_parent; + + AuDbgDentry(dentry); + if (d_is_positive(dentry)) { + err = au_aopen_no_open(file, dentry); + goto out_parent; /* success */ + } + + args.file = alloc_empty_file(file->f_flags, current_cred()); + err = PTR_ERR(args.file); + if (IS_ERR(args.file)) + goto out_parent; + + bindex = au_dbtop(dentry); + err = au_aopen_or_create(dir, dentry, &args); + AuTraceErr(err); + AuDbgFile(args.file); + file->f_mode = args.file->f_mode & ~FMODE_OPENED; + did_open = !!(args.file->f_mode & FMODE_OPENED); + if (!did_open) { + fput(args.file); + args.file = NULL; + } + di_write_unlock(parent); + di_write_unlock(dentry); + if (unlikely(err < 0)) { + if (args.file) + fput(args.file); + goto out_sb; + } + + if (!did_open) + err = au_aopen_no_open(file, dentry); + else { + aopen_node.h_file = args.file; + err = au_aopen_do_open(file, dentry, &aopen_node); + } + if (unlikely(err < 0)) { + if (args.file) + fput(args.file); + if (did_open) + au_lcnt_dec(&args.br->br_nfiles); + } + goto out_sb; /* success */ + +out_parent: + di_write_unlock(parent); + di_write_unlock(dentry); +out_sb: + si_read_unlock(sb); +out: + AuTraceErr(err); + AuDbgFile(file); + return err; +} + + +/* ---------------------------------------------------------------------- */ + +static int au_wr_dir_cpup(struct dentry *dentry, struct dentry *parent, + const unsigned char add_entry, aufs_bindex_t bcpup, + aufs_bindex_t btop) +{ + int err; + struct dentry *h_parent; + struct inode *h_dir; + + if (add_entry) + IMustLock(d_inode(parent)); + else + di_write_lock_parent(parent); + + err = 0; + if (!au_h_dptr(parent, bcpup)) { + if (btop > bcpup) + err = au_cpup_dirs(dentry, bcpup); + else if (btop < bcpup) + err = au_cpdown_dirs(dentry, bcpup); + else + BUG(); + } + if (!err && add_entry && !au_ftest_wrdir(add_entry, TMPFILE)) { + h_parent = au_h_dptr(parent, bcpup); + h_dir = d_inode(h_parent); + inode_lock_shared_nested(h_dir, AuLsc_I_PARENT); + err = au_lkup_neg(dentry, bcpup, /*wh*/0); + /* todo: no unlock here */ + inode_unlock_shared(h_dir); + + AuDbg("bcpup %d\n", bcpup); + if (!err) { + if (d_really_is_negative(dentry)) + au_set_h_dptr(dentry, btop, NULL); + au_update_dbrange(dentry, /*do_put_zero*/0); + } + } + + if (!add_entry) + di_write_unlock(parent); + if (!err) + err = bcpup; /* success */ + + AuTraceErr(err); + return err; +} + +/* + * decide the branch and the parent dir where we will create a new entry. + * returns new bindex or an error. + * copyup the parent dir if needed. + */ +int au_wr_dir(struct dentry *dentry, struct dentry *src_dentry, + struct au_wr_dir_args *args) +{ + int err; + unsigned int flags; + aufs_bindex_t bcpup, btop, src_btop; + const unsigned char add_entry + = au_ftest_wrdir(args->flags, ADD_ENTRY) + | au_ftest_wrdir(args->flags, TMPFILE); + struct super_block *sb; + struct dentry *parent; + struct au_sbinfo *sbinfo; + + sb = dentry->d_sb; + sbinfo = au_sbi(sb); + parent = dget_parent(dentry); + btop = au_dbtop(dentry); + bcpup = btop; + if (args->force_btgt < 0) { + if (src_dentry) { + src_btop = au_dbtop(src_dentry); + if (src_btop < btop) + bcpup = src_btop; + } else if (add_entry) { + flags = 0; + if (au_ftest_wrdir(args->flags, ISDIR)) + au_fset_wbr(flags, DIR); + err = AuWbrCreate(sbinfo, dentry, flags); + bcpup = err; + } + + if (bcpup < 0 || au_test_ro(sb, bcpup, d_inode(dentry))) { + if (add_entry) + err = AuWbrCopyup(sbinfo, dentry); + else { + if (!IS_ROOT(dentry)) { + di_read_lock_parent(parent, !AuLock_IR); + err = AuWbrCopyup(sbinfo, dentry); + di_read_unlock(parent, !AuLock_IR); + } else + err = AuWbrCopyup(sbinfo, dentry); + } + bcpup = err; + if (unlikely(err < 0)) + goto out; + } + } else { + bcpup = args->force_btgt; + AuDebugOn(au_test_ro(sb, bcpup, d_inode(dentry))); + } + + AuDbg("btop %d, bcpup %d\n", btop, bcpup); + err = bcpup; + if (bcpup == btop) + goto out; /* success */ + + /* copyup the new parent into the branch we process */ + err = au_wr_dir_cpup(dentry, parent, add_entry, bcpup, btop); + if (err >= 0) { + if (d_really_is_negative(dentry)) { + au_set_h_dptr(dentry, btop, NULL); + au_set_dbtop(dentry, bcpup); + au_set_dbbot(dentry, bcpup); + } + AuDebugOn(add_entry + && !au_ftest_wrdir(args->flags, TMPFILE) + && !au_h_dptr(dentry, bcpup)); + } + +out: + dput(parent); + return err; +} + +/* ---------------------------------------------------------------------- */ + +void au_pin_hdir_unlock(struct au_pin *p) +{ + if (p->hdir) + au_hn_inode_unlock(p->hdir); +} + +int au_pin_hdir_lock(struct au_pin *p) +{ + int err; + + err = 0; + if (!p->hdir) + goto out; + + /* even if an error happens later, keep this lock */ + au_hn_inode_lock_nested(p->hdir, p->lsc_hi); + + err = -EBUSY; + if (unlikely(p->hdir->hi_inode != d_inode(p->h_parent))) + goto out; + + err = 0; + if (p->h_dentry) + err = au_h_verify(p->h_dentry, p->udba, p->hdir->hi_inode, + p->h_parent, p->br); + +out: + return err; +} + +int au_pin_hdir_relock(struct au_pin *p) +{ + int err, i; + struct inode *h_i; + struct dentry *h_d[] = { + p->h_dentry, + p->h_parent + }; + + err = au_pin_hdir_lock(p); + if (unlikely(err)) + goto out; + + for (i = 0; !err && i < sizeof(h_d)/sizeof(*h_d); i++) { + if (!h_d[i]) + continue; + if (d_is_positive(h_d[i])) { + h_i = d_inode(h_d[i]); + err = !h_i->i_nlink; + } + } + +out: + return err; +} + +static void au_pin_hdir_set_owner(struct au_pin *p, struct task_struct *task) +{ + atomic_long_set(&p->hdir->hi_inode->i_rwsem.owner, (long)task); +} + +void au_pin_hdir_acquire_nest(struct au_pin *p) +{ + if (p->hdir) { + rwsem_acquire_nest(&p->hdir->hi_inode->i_rwsem.dep_map, + p->lsc_hi, 0, NULL, _RET_IP_); + au_pin_hdir_set_owner(p, current); + } +} + +void au_pin_hdir_release(struct au_pin *p) +{ + if (p->hdir) { + au_pin_hdir_set_owner(p, p->task); + rwsem_release(&p->hdir->hi_inode->i_rwsem.dep_map, _RET_IP_); + } +} + +struct dentry *au_pinned_h_parent(struct au_pin *pin) +{ + if (pin && pin->parent) + return au_h_dptr(pin->parent, pin->bindex); + return NULL; +} + +void au_unpin(struct au_pin *p) +{ + if (p->hdir) + au_pin_hdir_unlock(p); + if (p->h_mnt && au_ftest_pin(p->flags, MNT_WRITE)) + vfsub_mnt_drop_write(p->h_mnt); + if (!p->hdir) + return; + + if (!au_ftest_pin(p->flags, DI_LOCKED)) + di_read_unlock(p->parent, AuLock_IR); + iput(p->hdir->hi_inode); + dput(p->parent); + p->parent = NULL; + p->hdir = NULL; + p->h_mnt = NULL; + /* do not clear p->task */ +} + +int au_do_pin(struct au_pin *p) +{ + int err; + struct super_block *sb; + struct inode *h_dir; + + err = 0; + sb = p->dentry->d_sb; + p->br = au_sbr(sb, p->bindex); + if (IS_ROOT(p->dentry)) { + if (au_ftest_pin(p->flags, MNT_WRITE)) { + p->h_mnt = au_br_mnt(p->br); + err = vfsub_mnt_want_write(p->h_mnt); + if (unlikely(err)) { + au_fclr_pin(p->flags, MNT_WRITE); + goto out_err; + } + } + goto out; + } + + p->h_dentry = NULL; + if (p->bindex <= au_dbbot(p->dentry)) + p->h_dentry = au_h_dptr(p->dentry, p->bindex); + + p->parent = dget_parent(p->dentry); + if (!au_ftest_pin(p->flags, DI_LOCKED)) + di_read_lock(p->parent, AuLock_IR, p->lsc_di); + + h_dir = NULL; + p->h_parent = au_h_dptr(p->parent, p->bindex); + p->hdir = au_hi(d_inode(p->parent), p->bindex); + if (p->hdir) + h_dir = p->hdir->hi_inode; + + /* + * udba case, or + * if DI_LOCKED is not set, then p->parent may be different + * and h_parent can be NULL. + */ + if (unlikely(!p->hdir || !h_dir || !p->h_parent)) { + err = -EBUSY; + if (!au_ftest_pin(p->flags, DI_LOCKED)) + di_read_unlock(p->parent, AuLock_IR); + dput(p->parent); + p->parent = NULL; + goto out_err; + } + + if (au_ftest_pin(p->flags, MNT_WRITE)) { + p->h_mnt = au_br_mnt(p->br); + err = vfsub_mnt_want_write(p->h_mnt); + if (unlikely(err)) { + au_fclr_pin(p->flags, MNT_WRITE); + if (!au_ftest_pin(p->flags, DI_LOCKED)) + di_read_unlock(p->parent, AuLock_IR); + dput(p->parent); + p->parent = NULL; + goto out_err; + } + } + + au_igrab(h_dir); + err = au_pin_hdir_lock(p); + if (!err) + goto out; /* success */ + + au_unpin(p); + +out_err: + pr_err("err %d\n", err); + err = au_busy_or_stale(); +out: + return err; +} + +void au_pin_init(struct au_pin *p, struct dentry *dentry, + aufs_bindex_t bindex, int lsc_di, int lsc_hi, + unsigned int udba, unsigned char flags) +{ + p->dentry = dentry; + p->udba = udba; + p->lsc_di = lsc_di; + p->lsc_hi = lsc_hi; + p->flags = flags; + p->bindex = bindex; + + p->parent = NULL; + p->hdir = NULL; + p->h_mnt = NULL; + + p->h_dentry = NULL; + p->h_parent = NULL; + p->br = NULL; + p->task = current; +} + +int au_pin(struct au_pin *pin, struct dentry *dentry, aufs_bindex_t bindex, + unsigned int udba, unsigned char flags) +{ + au_pin_init(pin, dentry, bindex, AuLsc_DI_PARENT, AuLsc_I_PARENT2, + udba, flags); + return au_do_pin(pin); +} + +/* ---------------------------------------------------------------------- */ + +/* + * ->setattr() and ->getattr() are called in various cases. + * chmod, stat: dentry is revalidated. + * fchmod, fstat: file and dentry are not revalidated, additionally they may be + * unhashed. + * for ->setattr(), ia->ia_file is passed from ftruncate only. + */ +/* todo: consolidate with do_refresh() and simple_reval_dpath() */ +int au_reval_for_attr(struct dentry *dentry, unsigned int sigen) +{ + int err; + struct dentry *parent; + + err = 0; + if (au_digen_test(dentry, sigen)) { + parent = dget_parent(dentry); + di_read_lock_parent(parent, AuLock_IR); + err = au_refresh_dentry(dentry, parent); + di_read_unlock(parent, AuLock_IR); + dput(parent); + } + + AuTraceErr(err); + return err; +} + +int au_pin_and_icpup(struct dentry *dentry, struct iattr *ia, + struct au_icpup_args *a) +{ + int err; + loff_t sz; + aufs_bindex_t btop, ibtop; + struct dentry *hi_wh, *parent; + struct inode *inode; + struct au_wr_dir_args wr_dir_args = { + .force_btgt = -1, + .flags = 0 + }; + + if (d_is_dir(dentry)) + au_fset_wrdir(wr_dir_args.flags, ISDIR); + /* plink or hi_wh() case */ + btop = au_dbtop(dentry); + inode = d_inode(dentry); + ibtop = au_ibtop(inode); + if (btop != ibtop && !au_test_ro(inode->i_sb, ibtop, inode)) + wr_dir_args.force_btgt = ibtop; + err = au_wr_dir(dentry, /*src_dentry*/NULL, &wr_dir_args); + if (unlikely(err < 0)) + goto out; + a->btgt = err; + if (err != btop) + au_fset_icpup(a->flags, DID_CPUP); + + err = 0; + a->pin_flags = AuPin_MNT_WRITE; + parent = NULL; + if (!IS_ROOT(dentry)) { + au_fset_pin(a->pin_flags, DI_LOCKED); + parent = dget_parent(dentry); + di_write_lock_parent(parent); + } + + err = au_pin(&a->pin, dentry, a->btgt, a->udba, a->pin_flags); + if (unlikely(err)) + goto out_parent; + + sz = -1; + a->h_path.dentry = au_h_dptr(dentry, btop); + a->h_inode = d_inode(a->h_path.dentry); + if (ia && (ia->ia_valid & ATTR_SIZE)) { + inode_lock_shared_nested(a->h_inode, AuLsc_I_CHILD); + if (ia->ia_size < i_size_read(a->h_inode)) + sz = ia->ia_size; + inode_unlock_shared(a->h_inode); + } + + hi_wh = NULL; + if (au_ftest_icpup(a->flags, DID_CPUP) && d_unlinked(dentry)) { + hi_wh = au_hi_wh(inode, a->btgt); + if (!hi_wh) { + struct au_cp_generic cpg = { + .dentry = dentry, + .bdst = a->btgt, + .bsrc = -1, + .len = sz, + .pin = &a->pin + }; + err = au_sio_cpup_wh(&cpg, /*file*/NULL); + if (unlikely(err)) + goto out_unlock; + hi_wh = au_hi_wh(inode, a->btgt); + /* todo: revalidate hi_wh? */ + } + } + + if (parent) { + au_pin_set_parent_lflag(&a->pin, /*lflag*/0); + di_downgrade_lock(parent, AuLock_IR); + dput(parent); + parent = NULL; + } + if (!au_ftest_icpup(a->flags, DID_CPUP)) + goto out; /* success */ + + if (!d_unhashed(dentry)) { + struct au_cp_generic cpg = { + .dentry = dentry, + .bdst = a->btgt, + .bsrc = btop, + .len = sz, + .pin = &a->pin, + .flags = AuCpup_DTIME | AuCpup_HOPEN + }; + err = au_sio_cpup_simple(&cpg); + if (!err) + a->h_path.dentry = au_h_dptr(dentry, a->btgt); + } else if (!hi_wh) + a->h_path.dentry = au_h_dptr(dentry, a->btgt); + else + a->h_path.dentry = hi_wh; /* do not dget here */ + +out_unlock: + a->h_inode = d_inode(a->h_path.dentry); + if (!err) + goto out; /* success */ + au_unpin(&a->pin); +out_parent: + if (parent) { + di_write_unlock(parent); + dput(parent); + } +out: + if (!err) + inode_lock_nested(a->h_inode, AuLsc_I_CHILD); + return err; +} + +static int aufs_setattr(struct user_namespace *userns, struct dentry *dentry, + struct iattr *ia) +{ + int err; + struct inode *inode, *delegated; + struct super_block *sb; + struct file *file; + struct au_icpup_args *a; + struct user_namespace *h_userns; + + inode = d_inode(dentry); + IMustLock(inode); + + err = setattr_prepare(userns, dentry, ia); + if (unlikely(err)) + goto out; + + err = -ENOMEM; + a = kzalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + if (ia->ia_valid & (ATTR_KILL_SUID | ATTR_KILL_SGID)) + ia->ia_valid &= ~ATTR_MODE; + + file = NULL; + sb = dentry->d_sb; + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out_kfree; + + if (ia->ia_valid & ATTR_FILE) { + /* currently ftruncate(2) only */ + AuDebugOn(!d_is_reg(dentry)); + file = ia->ia_file; + err = au_reval_and_lock_fdi(file, au_reopen_nondir, /*wlock*/1, + /*fi_lsc*/0); + if (unlikely(err)) + goto out_si; + ia->ia_file = au_hf_top(file); + a->udba = AuOpt_UDBA_NONE; + } else { + /* fchmod() doesn't pass ia_file */ + a->udba = au_opt_udba(sb); + di_write_lock_child(dentry); + /* no d_unlinked(), to set UDBA_NONE for root */ + if (d_unhashed(dentry)) + a->udba = AuOpt_UDBA_NONE; + if (a->udba != AuOpt_UDBA_NONE) { + AuDebugOn(IS_ROOT(dentry)); + err = au_reval_for_attr(dentry, au_sigen(sb)); + if (unlikely(err)) + goto out_dentry; + } + } + + err = au_pin_and_icpup(dentry, ia, a); + if (unlikely(err < 0)) + goto out_dentry; + if (au_ftest_icpup(a->flags, DID_CPUP)) { + ia->ia_file = NULL; + ia->ia_valid &= ~ATTR_FILE; + } + + a->h_path.mnt = au_sbr_mnt(sb, a->btgt); + if ((ia->ia_valid & (ATTR_MODE | ATTR_CTIME)) + == (ATTR_MODE | ATTR_CTIME)) { + err = security_path_chmod(&a->h_path, ia->ia_mode); + if (unlikely(err)) + goto out_unlock; + } else if ((ia->ia_valid & (ATTR_UID | ATTR_GID)) + && (ia->ia_valid & ATTR_CTIME)) { + err = security_path_chown(&a->h_path, ia->ia_uid, ia->ia_gid); + if (unlikely(err)) + goto out_unlock; + } + + if (ia->ia_valid & ATTR_SIZE) { + struct file *f; + + if (ia->ia_size < i_size_read(inode)) + /* unmap only */ + truncate_setsize(inode, ia->ia_size); + + f = NULL; + if (ia->ia_valid & ATTR_FILE) + f = ia->ia_file; + inode_unlock(a->h_inode); + err = vfsub_trunc(&a->h_path, ia->ia_size, ia->ia_valid, f); + inode_lock_nested(a->h_inode, AuLsc_I_CHILD); + } else { + delegated = NULL; + while (1) { + err = vfsub_notify_change(&a->h_path, ia, &delegated); + if (delegated) { + err = break_deleg_wait(&delegated); + if (!err) + continue; + } + break; + } + } + /* + * regardless aufs 'acl' option setting. + * why don't all acl-aware fs call this func from their ->setattr()? + */ + if (!err && (ia->ia_valid & ATTR_MODE)) { + h_userns = mnt_user_ns(a->h_path.mnt); + err = vfsub_acl_chmod(h_userns, a->h_inode, ia->ia_mode); + } + if (!err) + au_cpup_attr_changeable(inode); + +out_unlock: + inode_unlock(a->h_inode); + au_unpin(&a->pin); + if (unlikely(err)) + au_update_dbtop(dentry); +out_dentry: + di_write_unlock(dentry); + if (file) { + fi_write_unlock(file); + ia->ia_file = file; + ia->ia_valid |= ATTR_FILE; + } +out_si: + si_read_unlock(sb); +out_kfree: + au_kfree_rcu(a); +out: + AuTraceErr(err); + return err; +} + +#if IS_ENABLED(CONFIG_AUFS_XATTR) || IS_ENABLED(CONFIG_FS_POSIX_ACL) +static int au_h_path_to_set_attr(struct dentry *dentry, + struct au_icpup_args *a, struct path *h_path) +{ + int err; + struct super_block *sb; + + sb = dentry->d_sb; + a->udba = au_opt_udba(sb); + /* no d_unlinked(), to set UDBA_NONE for root */ + if (d_unhashed(dentry)) + a->udba = AuOpt_UDBA_NONE; + if (a->udba != AuOpt_UDBA_NONE) { + AuDebugOn(IS_ROOT(dentry)); + err = au_reval_for_attr(dentry, au_sigen(sb)); + if (unlikely(err)) + goto out; + } + err = au_pin_and_icpup(dentry, /*ia*/NULL, a); + if (unlikely(err < 0)) + goto out; + + h_path->dentry = a->h_path.dentry; + h_path->mnt = au_sbr_mnt(sb, a->btgt); + +out: + return err; +} + +ssize_t au_sxattr(struct dentry *dentry, struct inode *inode, + struct au_sxattr *arg) +{ + int err; + struct path h_path; + struct super_block *sb; + struct au_icpup_args *a; + struct inode *h_inode; + struct user_namespace *h_userns; + + IMustLock(inode); + + err = -ENOMEM; + a = kzalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + sb = dentry->d_sb; + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out_kfree; + + h_path.dentry = NULL; /* silence gcc */ + di_write_lock_child(dentry); + err = au_h_path_to_set_attr(dentry, a, &h_path); + if (unlikely(err)) + goto out_di; + h_userns = mnt_user_ns(h_path.mnt); + + inode_unlock(a->h_inode); + switch (arg->type) { + case AU_XATTR_SET: + AuDebugOn(d_is_negative(h_path.dentry)); + err = vfsub_setxattr(h_userns, h_path.dentry, + arg->u.set.name, arg->u.set.value, + arg->u.set.size, arg->u.set.flags); + break; + case AU_ACL_SET: + err = -EOPNOTSUPP; + h_inode = d_inode(h_path.dentry); + if (h_inode->i_op->set_acl) { + /* this will call posix_acl_update_mode */ + err = h_inode->i_op->set_acl(h_userns, h_inode, + arg->u.acl_set.acl, + arg->u.acl_set.type); + } + break; + } + if (!err) + au_cpup_attr_timesizes(inode); + + au_unpin(&a->pin); + if (unlikely(err)) + au_update_dbtop(dentry); + +out_di: + di_write_unlock(dentry); + si_read_unlock(sb); +out_kfree: + au_kfree_rcu(a); +out: + AuTraceErr(err); + return err; +} +#endif + +static void au_refresh_iattr(struct inode *inode, struct kstat *st, + unsigned int nlink) +{ + unsigned int n; + + inode->i_mode = st->mode; + /* don't i_[ug]id_write() here */ + inode->i_uid = st->uid; + inode->i_gid = st->gid; + inode->i_atime = st->atime; + inode->i_mtime = st->mtime; + inode->i_ctime = st->ctime; + + au_cpup_attr_nlink(inode, /*force*/0); + if (S_ISDIR(inode->i_mode)) { + n = inode->i_nlink; + n -= nlink; + n += st->nlink; + smp_mb(); /* for i_nlink */ + /* 0 can happen */ + set_nlink(inode, n); + } + + spin_lock(&inode->i_lock); + inode->i_blocks = st->blocks; + i_size_write(inode, st->size); + spin_unlock(&inode->i_lock); +} + +/* + * common routine for aufs_getattr() and au_getxattr(). + * returns zero or negative (an error). + * @dentry will be read-locked in success. + */ +int au_h_path_getattr(struct dentry *dentry, struct inode *inode, int force, + struct path *h_path, int locked) +{ + int err; + unsigned int mnt_flags, sigen; + unsigned char udba_none; + aufs_bindex_t bindex; + struct super_block *sb, *h_sb; + + h_path->mnt = NULL; + h_path->dentry = NULL; + + err = 0; + sb = dentry->d_sb; + mnt_flags = au_mntflags(sb); + udba_none = !!au_opt_test(mnt_flags, UDBA_NONE); + + if (unlikely(locked)) + goto body; /* skip locking dinfo */ + + /* support fstat(2) */ + if (!d_unlinked(dentry) && !udba_none) { + sigen = au_sigen(sb); + err = au_digen_test(dentry, sigen); + if (!err) { + di_read_lock_child(dentry, AuLock_IR); + err = au_dbrange_test(dentry); + if (unlikely(err)) { + di_read_unlock(dentry, AuLock_IR); + goto out; + } + } else { + AuDebugOn(IS_ROOT(dentry)); + di_write_lock_child(dentry); + err = au_dbrange_test(dentry); + if (!err) + err = au_reval_for_attr(dentry, sigen); + if (!err) + di_downgrade_lock(dentry, AuLock_IR); + else { + di_write_unlock(dentry); + goto out; + } + } + } else + di_read_lock_child(dentry, AuLock_IR); + +body: + if (!inode) { + inode = d_inode(dentry); + if (unlikely(!inode)) + goto out; + } + bindex = au_ibtop(inode); + h_path->mnt = au_sbr_mnt(sb, bindex); + h_sb = h_path->mnt->mnt_sb; + if (!force + && !au_test_fs_bad_iattr(h_sb) + && udba_none) + goto out; /* success */ + + if (au_dbtop(dentry) == bindex) + h_path->dentry = au_h_dptr(dentry, bindex); + else if (au_opt_test(mnt_flags, PLINK) && au_plink_test(inode)) { + h_path->dentry = au_plink_lkup(inode, bindex); + if (IS_ERR(h_path->dentry)) + /* pretending success */ + h_path->dentry = NULL; + else + dput(h_path->dentry); + } + +out: + return err; +} + +static int aufs_getattr(struct user_namespace *userns, const struct path *path, + struct kstat *st, u32 request, unsigned int query) +{ + int err; + unsigned char positive; + struct path h_path; + struct dentry *dentry; + struct inode *inode; + struct super_block *sb; + + dentry = path->dentry; + inode = d_inode(dentry); + sb = dentry->d_sb; + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out; + err = au_h_path_getattr(dentry, /*inode*/NULL, /*force*/0, &h_path, + /*locked*/0); + if (unlikely(err)) + goto out_si; + if (unlikely(!h_path.dentry)) + /* illegally overlapped or something */ + goto out_fill; /* pretending success */ + + positive = d_is_positive(h_path.dentry); + if (positive) + /* no vfsub version */ + err = vfs_getattr(&h_path, st, request, query); + if (!err) { + if (positive) + au_refresh_iattr(inode, st, + d_inode(h_path.dentry)->i_nlink); + goto out_fill; /* success */ + } + AuTraceErr(err); + goto out_di; + +out_fill: + generic_fillattr(userns, inode, st); +out_di: + di_read_unlock(dentry, AuLock_IR); +out_si: + si_read_unlock(sb); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static const char *aufs_get_link(struct dentry *dentry, struct inode *inode, + struct delayed_call *done) +{ + const char *ret; + struct dentry *h_dentry; + struct inode *h_inode; + int err; + aufs_bindex_t bindex; + + ret = NULL; /* suppress a warning */ + err = -ECHILD; + if (!dentry) + goto out; + + err = aufs_read_lock(dentry, AuLock_IR | AuLock_GEN); + if (unlikely(err)) + goto out; + + err = au_d_hashed_positive(dentry); + if (unlikely(err)) + goto out_unlock; + + err = -EINVAL; + inode = d_inode(dentry); + bindex = au_ibtop(inode); + h_inode = au_h_iptr(inode, bindex); + if (unlikely(!h_inode->i_op->get_link)) + goto out_unlock; + + err = -EBUSY; + h_dentry = NULL; + if (au_dbtop(dentry) <= bindex) { + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry) + dget(h_dentry); + } + if (!h_dentry) { + h_dentry = d_find_any_alias(h_inode); + if (IS_ERR(h_dentry)) { + err = PTR_ERR(h_dentry); + goto out_unlock; + } + } + if (unlikely(!h_dentry)) + goto out_unlock; + + err = 0; + AuDbg("%ps\n", h_inode->i_op->get_link); + AuDbgDentry(h_dentry); + ret = vfs_get_link(h_dentry, done); + dput(h_dentry); + if (IS_ERR(ret)) + err = PTR_ERR(ret); + +out_unlock: + aufs_read_unlock(dentry, AuLock_IR); +out: + if (unlikely(err)) + ret = ERR_PTR(err); + AuTraceErrPtr(ret); + return ret; +} + +/* ---------------------------------------------------------------------- */ + +static int au_is_special(struct inode *inode) +{ + return (inode->i_mode & (S_IFBLK | S_IFCHR | S_IFIFO | S_IFSOCK)); +} + +static int aufs_update_time(struct inode *inode, struct timespec64 *ts, + int flags) +{ + int err; + aufs_bindex_t bindex; + struct super_block *sb; + struct inode *h_inode; + struct vfsmount *h_mnt; + + sb = inode->i_sb; + WARN_ONCE((flags & S_ATIME) && !IS_NOATIME(inode), + "unexpected s_flags 0x%lx", sb->s_flags); + + /* mmap_sem might be acquired already, cf. aufs_mmap() */ + lockdep_off(); + si_read_lock(sb, AuLock_FLUSH); + ii_write_lock_child(inode); + + err = 0; + bindex = au_ibtop(inode); + h_inode = au_h_iptr(inode, bindex); + if (!au_test_ro(sb, bindex, inode)) { + h_mnt = au_sbr_mnt(sb, bindex); + err = vfsub_mnt_want_write(h_mnt); + if (!err) { + err = vfsub_update_time(h_inode, ts, flags); + vfsub_mnt_drop_write(h_mnt); + } + } else if (au_is_special(h_inode)) { + /* + * Never copy-up here. + * These special files may already be opened and used for + * communicating. If we copied it up, then the communication + * would be corrupted. + */ + AuWarn1("timestamps for i%lu are ignored " + "since it is on readonly branch (hi%lu).\n", + inode->i_ino, h_inode->i_ino); + } else if (flags & ~S_ATIME) { + err = -EIO; + AuIOErr1("unexpected flags 0x%x\n", flags); + AuDebugOn(1); + } + + if (!err) + au_cpup_attr_timesizes(inode); + ii_write_unlock(inode); + si_read_unlock(sb); + lockdep_on(); + + if (!err && (flags & S_VERSION)) + inode_inc_iversion(inode); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* no getattr version will be set by module.c:aufs_init() */ +struct inode_operations aufs_iop_nogetattr[AuIop_Last], + aufs_iop[] = { + [AuIop_SYMLINK] = { + .permission = aufs_permission, +#ifdef CONFIG_FS_POSIX_ACL + .get_acl = aufs_get_acl, + .set_acl = aufs_set_acl, /* unsupport for symlink? */ +#endif + + .setattr = aufs_setattr, + .getattr = aufs_getattr, + +#ifdef CONFIG_AUFS_XATTR + .listxattr = aufs_listxattr, +#endif + + .get_link = aufs_get_link, + + /* .update_time = aufs_update_time */ + }, + [AuIop_DIR] = { + .create = aufs_create, + .lookup = aufs_lookup, + .link = aufs_link, + .unlink = aufs_unlink, + .symlink = aufs_symlink, + .mkdir = aufs_mkdir, + .rmdir = aufs_rmdir, + .mknod = aufs_mknod, + .rename = aufs_rename, + + .permission = aufs_permission, +#ifdef CONFIG_FS_POSIX_ACL + .get_acl = aufs_get_acl, + .set_acl = aufs_set_acl, +#endif + + .setattr = aufs_setattr, + .getattr = aufs_getattr, + +#ifdef CONFIG_AUFS_XATTR + .listxattr = aufs_listxattr, +#endif + + .update_time = aufs_update_time, + .atomic_open = aufs_atomic_open, + .tmpfile = aufs_tmpfile + }, + [AuIop_OTHER] = { + .permission = aufs_permission, +#ifdef CONFIG_FS_POSIX_ACL + .get_acl = aufs_get_acl, + .set_acl = aufs_set_acl, +#endif + + .setattr = aufs_setattr, + .getattr = aufs_getattr, + +#ifdef CONFIG_AUFS_XATTR + .listxattr = aufs_listxattr, +#endif + + .update_time = aufs_update_time + } +}; diff --git a/fs/aufs/i_op_add.c b/fs/aufs/i_op_add.c new file mode 100644 index 00000000000..1b4992de77d --- /dev/null +++ b/fs/aufs/i_op_add.c @@ -0,0 +1,941 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode operations (add entry) + */ + +#include +#include "aufs.h" + +/* + * final procedure of adding a new entry, except link(2). + * remove whiteout, instantiate, copyup the parent dir's times and size + * and update version. + * if it failed, re-create the removed whiteout. + */ +static int epilog(struct inode *dir, aufs_bindex_t bindex, + struct dentry *wh_dentry, struct dentry *dentry) +{ + int err, rerr; + aufs_bindex_t bwh; + struct path h_path; + struct super_block *sb; + struct inode *inode, *h_dir; + struct dentry *wh; + + bwh = -1; + sb = dir->i_sb; + if (wh_dentry) { + h_dir = d_inode(wh_dentry->d_parent); /* dir inode is locked */ + IMustLock(h_dir); + AuDebugOn(au_h_iptr(dir, bindex) != h_dir); + bwh = au_dbwh(dentry); + h_path.dentry = wh_dentry; + h_path.mnt = au_sbr_mnt(sb, bindex); + err = au_wh_unlink_dentry(au_h_iptr(dir, bindex), &h_path, + dentry); + if (unlikely(err)) + goto out; + } + + inode = au_new_inode(dentry, /*must_new*/1); + if (!IS_ERR(inode)) { + d_instantiate(dentry, inode); + dir = d_inode(dentry->d_parent); /* dir inode is locked */ + IMustLock(dir); + au_dir_ts(dir, bindex); + inode_inc_iversion(dir); + au_fhsm_wrote(sb, bindex, /*force*/0); + return 0; /* success */ + } + + err = PTR_ERR(inode); + if (!wh_dentry) + goto out; + + /* revert */ + /* dir inode is locked */ + wh = au_wh_create(dentry, bwh, wh_dentry->d_parent); + rerr = PTR_ERR(wh); + if (IS_ERR(wh)) { + AuIOErr("%pd reverting whiteout failed(%d, %d)\n", + dentry, err, rerr); + err = -EIO; + } else + dput(wh); + +out: + return err; +} + +static int au_d_may_add(struct dentry *dentry) +{ + int err; + + err = 0; + if (unlikely(d_unhashed(dentry))) + err = -ENOENT; + if (unlikely(d_really_is_positive(dentry))) + err = -EEXIST; + return err; +} + +/* + * simple tests for the adding inode operations. + * following the checks in vfs, plus the parent-child relationship. + */ +int au_may_add(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent, int isdir) +{ + int err; + umode_t h_mode; + struct dentry *h_dentry; + struct inode *h_inode; + + err = -ENAMETOOLONG; + if (unlikely(dentry->d_name.len > AUFS_MAX_NAMELEN)) + goto out; + + h_dentry = au_h_dptr(dentry, bindex); + if (d_really_is_negative(dentry)) { + err = -EEXIST; + if (unlikely(d_is_positive(h_dentry))) + goto out; + } else { + /* rename(2) case */ + err = -EIO; + if (unlikely(d_is_negative(h_dentry))) + goto out; + h_inode = d_inode(h_dentry); + if (unlikely(!h_inode->i_nlink)) + goto out; + + h_mode = h_inode->i_mode; + if (!isdir) { + err = -EISDIR; + if (unlikely(S_ISDIR(h_mode))) + goto out; + } else if (unlikely(!S_ISDIR(h_mode))) { + err = -ENOTDIR; + goto out; + } + } + + err = 0; + /* expected parent dir is locked */ + if (unlikely(h_parent != h_dentry->d_parent)) + err = -EIO; + +out: + AuTraceErr(err); + return err; +} + +/* + * initial procedure of adding a new entry. + * prepare writable branch and the parent dir, lock it, + * and lookup whiteout for the new entry. + */ +static struct dentry* +lock_hdir_lkup_wh(struct dentry *dentry, struct au_dtime *dt, + struct dentry *src_dentry, struct au_pin *pin, + struct au_wr_dir_args *wr_dir_args) +{ + struct dentry *wh_dentry, *h_parent; + struct super_block *sb; + struct au_branch *br; + int err; + unsigned int udba; + aufs_bindex_t bcpup; + + AuDbg("%pd\n", dentry); + + err = au_wr_dir(dentry, src_dentry, wr_dir_args); + bcpup = err; + wh_dentry = ERR_PTR(err); + if (unlikely(err < 0)) + goto out; + + sb = dentry->d_sb; + udba = au_opt_udba(sb); + err = au_pin(pin, dentry, bcpup, udba, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + wh_dentry = ERR_PTR(err); + if (unlikely(err)) + goto out; + + h_parent = au_pinned_h_parent(pin); + if (udba != AuOpt_UDBA_NONE + && au_dbtop(dentry) == bcpup) + err = au_may_add(dentry, bcpup, h_parent, + au_ftest_wrdir(wr_dir_args->flags, ISDIR)); + else if (unlikely(dentry->d_name.len > AUFS_MAX_NAMELEN)) + err = -ENAMETOOLONG; + wh_dentry = ERR_PTR(err); + if (unlikely(err)) + goto out_unpin; + + br = au_sbr(sb, bcpup); + if (dt) { + struct path tmp = { + .dentry = h_parent, + .mnt = au_br_mnt(br) + }; + au_dtime_store(dt, au_pinned_parent(pin), &tmp); + } + + wh_dentry = NULL; + if (bcpup != au_dbwh(dentry)) + goto out; /* success */ + + /* + * ENAMETOOLONG here means that if we allowed create such name, then it + * would not be able to removed in the future. So we don't allow such + * name here and we don't handle ENAMETOOLONG differently here. + */ + wh_dentry = au_wh_lkup(h_parent, &dentry->d_name, br); + +out_unpin: + if (IS_ERR(wh_dentry)) + au_unpin(pin); +out: + return wh_dentry; +} + +/* ---------------------------------------------------------------------- */ + +enum { Mknod, Symlink, Creat }; +struct simple_arg { + int type; + union { + struct { + umode_t mode; + bool want_excl; + bool try_aopen; + struct vfsub_aopen_args *aopen; + } c; + struct { + const char *symname; + } s; + struct { + umode_t mode; + dev_t dev; + } m; + } u; +}; + +static int add_simple(struct inode *dir, struct dentry *dentry, + struct simple_arg *arg) +{ + int err, rerr; + aufs_bindex_t btop; + unsigned char created; + const unsigned char try_aopen + = (arg->type == Creat && arg->u.c.try_aopen); + struct vfsub_aopen_args *aopen = arg->u.c.aopen; + struct dentry *wh_dentry, *parent; + struct inode *h_dir; + struct super_block *sb; + struct au_branch *br; + /* to reduce stack size */ + struct { + struct au_dtime dt; + struct au_pin pin; + struct path h_path; + struct au_wr_dir_args wr_dir_args; + } *a; + + AuDbg("%pd\n", dentry); + IMustLock(dir); + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + a->wr_dir_args.force_btgt = -1; + a->wr_dir_args.flags = AuWrDir_ADD_ENTRY; + + parent = dentry->d_parent; /* dir inode is locked */ + if (!try_aopen) { + err = aufs_read_lock(dentry, AuLock_DW | AuLock_GEN); + if (unlikely(err)) + goto out_free; + } + err = au_d_may_add(dentry); + if (unlikely(err)) + goto out_unlock; + if (!try_aopen) + di_write_lock_parent(parent); + wh_dentry = lock_hdir_lkup_wh(dentry, &a->dt, /*src_dentry*/NULL, + &a->pin, &a->wr_dir_args); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out_parent; + + btop = au_dbtop(dentry); + sb = dentry->d_sb; + br = au_sbr(sb, btop); + a->h_path.dentry = au_h_dptr(dentry, btop); + a->h_path.mnt = au_br_mnt(br); + h_dir = au_pinned_h_dir(&a->pin); + switch (arg->type) { + case Creat: + if (!try_aopen || !h_dir->i_op->atomic_open) { + err = vfsub_create(h_dir, &a->h_path, arg->u.c.mode, + arg->u.c.want_excl); + created = !err; + if (!err && try_aopen) + aopen->file->f_mode |= FMODE_CREATED; + } else { + aopen->br = br; + err = vfsub_atomic_open(h_dir, a->h_path.dentry, aopen); + AuDbg("err %d\n", err); + AuDbgFile(aopen->file); + created = err >= 0 + && !!(aopen->file->f_mode & FMODE_CREATED); + } + break; + case Symlink: + err = vfsub_symlink(h_dir, &a->h_path, arg->u.s.symname); + created = !err; + break; + case Mknod: + err = vfsub_mknod(h_dir, &a->h_path, arg->u.m.mode, + arg->u.m.dev); + created = !err; + break; + default: + BUG(); + } + if (unlikely(err < 0)) + goto out_unpin; + + err = epilog(dir, btop, wh_dentry, dentry); + if (!err) + goto out_unpin; /* success */ + + /* revert */ + if (created /* && d_is_positive(a->h_path.dentry) */) { + /* no delegation since it is just created */ + rerr = vfsub_unlink(h_dir, &a->h_path, /*delegated*/NULL, + /*force*/0); + if (rerr) { + AuIOErr("%pd revert failure(%d, %d)\n", + dentry, err, rerr); + err = -EIO; + } + au_dtime_revert(&a->dt); + } + if (try_aopen && h_dir->i_op->atomic_open + && (aopen->file->f_mode & FMODE_OPENED)) + /* aopen->file is still opened */ + au_lcnt_dec(&aopen->br->br_nfiles); + +out_unpin: + au_unpin(&a->pin); + dput(wh_dentry); +out_parent: + if (!try_aopen) + di_write_unlock(parent); +out_unlock: + if (unlikely(err)) { + au_update_dbtop(dentry); + d_drop(dentry); + } + if (!try_aopen) + aufs_read_unlock(dentry, AuLock_DW); +out_free: + au_kfree_rcu(a); +out: + return err; +} + +int aufs_mknod(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode, dev_t dev) +{ + struct simple_arg arg = { + .type = Mknod, + .u.m = { + .mode = mode, + .dev = dev + } + }; + return add_simple(dir, dentry, &arg); +} + +int aufs_symlink(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, const char *symname) +{ + struct simple_arg arg = { + .type = Symlink, + .u.s.symname = symname + }; + return add_simple(dir, dentry, &arg); +} + +int aufs_create(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode, bool want_excl) +{ + struct simple_arg arg = { + .type = Creat, + .u.c = { + .mode = mode, + .want_excl = want_excl + } + }; + return add_simple(dir, dentry, &arg); +} + +int au_aopen_or_create(struct inode *dir, struct dentry *dentry, + struct vfsub_aopen_args *aopen_args) +{ + struct simple_arg arg = { + .type = Creat, + .u.c = { + .mode = aopen_args->create_mode, + .want_excl = aopen_args->open_flag & O_EXCL, + .try_aopen = true, + .aopen = aopen_args + } + }; + return add_simple(dir, dentry, &arg); +} + +int aufs_tmpfile(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode) +{ + int err; + aufs_bindex_t bindex; + struct super_block *sb; + struct dentry *parent, *h_parent, *h_dentry; + struct inode *h_dir, *inode; + struct vfsmount *h_mnt; + struct user_namespace *h_userns; + struct au_wr_dir_args wr_dir_args = { + .force_btgt = -1, + .flags = AuWrDir_TMPFILE + }; + + /* copy-up may happen */ + inode_lock(dir); + + sb = dir->i_sb; + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out; + + err = au_di_init(dentry); + if (unlikely(err)) + goto out_si; + + err = -EBUSY; + parent = d_find_any_alias(dir); + AuDebugOn(!parent); + di_write_lock_parent(parent); + if (unlikely(d_inode(parent) != dir)) + goto out_parent; + + err = au_digen_test(parent, au_sigen(sb)); + if (unlikely(err)) + goto out_parent; + + bindex = au_dbtop(parent); + au_set_dbtop(dentry, bindex); + au_set_dbbot(dentry, bindex); + err = au_wr_dir(dentry, /*src_dentry*/NULL, &wr_dir_args); + bindex = err; + if (unlikely(err < 0)) + goto out_parent; + + err = -EOPNOTSUPP; + h_dir = au_h_iptr(dir, bindex); + if (unlikely(!h_dir->i_op->tmpfile)) + goto out_parent; + + h_mnt = au_sbr_mnt(sb, bindex); + err = vfsub_mnt_want_write(h_mnt); + if (unlikely(err)) + goto out_parent; + + h_userns = mnt_user_ns(h_mnt); + h_parent = au_h_dptr(parent, bindex); + h_dentry = vfs_tmpfile(h_userns, h_parent, mode, /*open_flag*/0); + if (IS_ERR(h_dentry)) { + err = PTR_ERR(h_dentry); + goto out_mnt; + } + + au_set_dbtop(dentry, bindex); + au_set_dbbot(dentry, bindex); + au_set_h_dptr(dentry, bindex, dget(h_dentry)); + inode = au_new_inode(dentry, /*must_new*/1); + if (IS_ERR(inode)) { + err = PTR_ERR(inode); + au_set_h_dptr(dentry, bindex, NULL); + au_set_dbtop(dentry, -1); + au_set_dbbot(dentry, -1); + } else { + if (!inode->i_nlink) + set_nlink(inode, 1); + d_tmpfile(dentry, inode); + au_di(dentry)->di_tmpfile = 1; + + /* update without i_mutex */ + if (au_ibtop(dir) == au_dbtop(dentry)) + au_cpup_attr_timesizes(dir); + } + dput(h_dentry); + +out_mnt: + vfsub_mnt_drop_write(h_mnt); +out_parent: + di_write_unlock(parent); + dput(parent); + di_write_unlock(dentry); + if (unlikely(err)) { + au_di_fin(dentry); + dentry->d_fsdata = NULL; + } +out_si: + si_read_unlock(sb); +out: + inode_unlock(dir); + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct au_link_args { + aufs_bindex_t bdst, bsrc; + struct au_pin pin; + struct path h_path; + struct dentry *src_parent, *parent; +}; + +static int au_cpup_before_link(struct dentry *src_dentry, + struct au_link_args *a) +{ + int err; + struct dentry *h_src_dentry; + struct au_cp_generic cpg = { + .dentry = src_dentry, + .bdst = a->bdst, + .bsrc = a->bsrc, + .len = -1, + .pin = &a->pin, + .flags = AuCpup_DTIME | AuCpup_HOPEN /* | AuCpup_KEEPLINO */ + }; + + di_read_lock_parent(a->src_parent, AuLock_IR); + err = au_test_and_cpup_dirs(src_dentry, a->bdst); + if (unlikely(err)) + goto out; + + h_src_dentry = au_h_dptr(src_dentry, a->bsrc); + err = au_pin(&a->pin, src_dentry, a->bdst, + au_opt_udba(src_dentry->d_sb), + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (unlikely(err)) + goto out; + + err = au_sio_cpup_simple(&cpg); + au_unpin(&a->pin); + +out: + di_read_unlock(a->src_parent, AuLock_IR); + return err; +} + +static int au_cpup_or_link(struct dentry *src_dentry, struct dentry *dentry, + struct au_link_args *a) +{ + int err; + unsigned char plink; + aufs_bindex_t bbot; + struct dentry *h_src_dentry; + struct inode *h_inode, *inode, *delegated; + struct super_block *sb; + struct file *h_file; + + plink = 0; + h_inode = NULL; + sb = src_dentry->d_sb; + inode = d_inode(src_dentry); + if (au_ibtop(inode) <= a->bdst) + h_inode = au_h_iptr(inode, a->bdst); + if (!h_inode || !h_inode->i_nlink) { + /* copyup src_dentry as the name of dentry. */ + bbot = au_dbbot(dentry); + if (bbot < a->bsrc) + au_set_dbbot(dentry, a->bsrc); + au_set_h_dptr(dentry, a->bsrc, + dget(au_h_dptr(src_dentry, a->bsrc))); + dget(a->h_path.dentry); + au_set_h_dptr(dentry, a->bdst, NULL); + AuDbg("temporary d_inode...\n"); + spin_lock(&dentry->d_lock); + dentry->d_inode = d_inode(src_dentry); /* tmp */ + spin_unlock(&dentry->d_lock); + h_file = au_h_open_pre(dentry, a->bsrc, /*force_wr*/0); + if (IS_ERR(h_file)) + err = PTR_ERR(h_file); + else { + struct au_cp_generic cpg = { + .dentry = dentry, + .bdst = a->bdst, + .bsrc = -1, + .len = -1, + .pin = &a->pin, + .flags = AuCpup_KEEPLINO + }; + err = au_sio_cpup_simple(&cpg); + au_h_open_post(dentry, a->bsrc, h_file); + if (!err) { + dput(a->h_path.dentry); + a->h_path.dentry = au_h_dptr(dentry, a->bdst); + } else + au_set_h_dptr(dentry, a->bdst, + a->h_path.dentry); + } + spin_lock(&dentry->d_lock); + dentry->d_inode = NULL; /* restore */ + spin_unlock(&dentry->d_lock); + AuDbg("temporary d_inode...done\n"); + au_set_h_dptr(dentry, a->bsrc, NULL); + au_set_dbbot(dentry, bbot); + } else { + /* the inode of src_dentry already exists on a.bdst branch */ + h_src_dentry = d_find_alias(h_inode); + if (!h_src_dentry && au_plink_test(inode)) { + plink = 1; + h_src_dentry = au_plink_lkup(inode, a->bdst); + err = PTR_ERR(h_src_dentry); + if (IS_ERR(h_src_dentry)) + goto out; + + if (unlikely(d_is_negative(h_src_dentry))) { + dput(h_src_dentry); + h_src_dentry = NULL; + } + + } + if (h_src_dentry) { + delegated = NULL; + err = vfsub_link(h_src_dentry, au_pinned_h_dir(&a->pin), + &a->h_path, &delegated); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + dput(h_src_dentry); + } else { + AuIOErr("no dentry found for hi%lu on b%d\n", + h_inode->i_ino, a->bdst); + err = -EIO; + } + } + + if (!err && !plink) + au_plink_append(inode, a->bdst, a->h_path.dentry); + +out: + AuTraceErr(err); + return err; +} + +int aufs_link(struct dentry *src_dentry, struct inode *dir, + struct dentry *dentry) +{ + int err, rerr; + struct au_dtime dt; + struct au_link_args *a; + struct dentry *wh_dentry, *h_src_dentry; + struct inode *inode, *delegated; + struct super_block *sb; + struct au_wr_dir_args wr_dir_args = { + /* .force_btgt = -1, */ + .flags = AuWrDir_ADD_ENTRY + }; + + IMustLock(dir); + inode = d_inode(src_dentry); + IMustLock(inode); + + err = -ENOMEM; + a = kzalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + a->parent = dentry->d_parent; /* dir inode is locked */ + err = aufs_read_and_write_lock2(dentry, src_dentry, + AuLock_NOPLM | AuLock_GEN); + if (unlikely(err)) + goto out_kfree; + err = au_d_linkable(src_dentry); + if (unlikely(err)) + goto out_unlock; + err = au_d_may_add(dentry); + if (unlikely(err)) + goto out_unlock; + + a->src_parent = dget_parent(src_dentry); + wr_dir_args.force_btgt = au_ibtop(inode); + + di_write_lock_parent(a->parent); + wr_dir_args.force_btgt = au_wbr(dentry, wr_dir_args.force_btgt); + wh_dentry = lock_hdir_lkup_wh(dentry, &dt, src_dentry, &a->pin, + &wr_dir_args); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out_parent; + + err = 0; + sb = dentry->d_sb; + a->bdst = au_dbtop(dentry); + a->h_path.dentry = au_h_dptr(dentry, a->bdst); + a->h_path.mnt = au_sbr_mnt(sb, a->bdst); + a->bsrc = au_ibtop(inode); + h_src_dentry = au_h_d_alias(src_dentry, a->bsrc); + if (!h_src_dentry && au_di(src_dentry)->di_tmpfile) + h_src_dentry = dget(au_hi_wh(inode, a->bsrc)); + if (!h_src_dentry) { + a->bsrc = au_dbtop(src_dentry); + h_src_dentry = au_h_d_alias(src_dentry, a->bsrc); + AuDebugOn(!h_src_dentry); + } else if (IS_ERR(h_src_dentry)) { + err = PTR_ERR(h_src_dentry); + goto out_parent; + } + + /* + * aufs doesn't touch the credential so + * security_dentry_create_files_as() is unnecessary. + */ + if (au_opt_test(au_mntflags(sb), PLINK)) { + if (a->bdst < a->bsrc + /* && h_src_dentry->d_sb != a->h_path.dentry->d_sb */) + err = au_cpup_or_link(src_dentry, dentry, a); + else { + delegated = NULL; + err = vfsub_link(h_src_dentry, au_pinned_h_dir(&a->pin), + &a->h_path, &delegated); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + } + dput(h_src_dentry); + } else { + /* + * copyup src_dentry to the branch we process, + * and then link(2) to it. + */ + dput(h_src_dentry); + if (a->bdst < a->bsrc + /* && h_src_dentry->d_sb != a->h_path.dentry->d_sb */) { + au_unpin(&a->pin); + di_write_unlock(a->parent); + err = au_cpup_before_link(src_dentry, a); + di_write_lock_parent(a->parent); + if (!err) + err = au_pin(&a->pin, dentry, a->bdst, + au_opt_udba(sb), + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (unlikely(err)) + goto out_wh; + } + if (!err) { + h_src_dentry = au_h_dptr(src_dentry, a->bdst); + err = -ENOENT; + if (h_src_dentry && d_is_positive(h_src_dentry)) { + delegated = NULL; + err = vfsub_link(h_src_dentry, + au_pinned_h_dir(&a->pin), + &a->h_path, &delegated); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry" + " for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + } + } + } + if (unlikely(err)) + goto out_unpin; + + if (wh_dentry) { + a->h_path.dentry = wh_dentry; + err = au_wh_unlink_dentry(au_pinned_h_dir(&a->pin), &a->h_path, + dentry); + if (unlikely(err)) + goto out_revert; + } + + au_dir_ts(dir, a->bdst); + inode_inc_iversion(dir); + inc_nlink(inode); + inode->i_ctime = dir->i_ctime; + d_instantiate(dentry, au_igrab(inode)); + if (d_unhashed(a->h_path.dentry)) + /* some filesystem calls d_drop() */ + d_drop(dentry); + /* some filesystems consume an inode even hardlink */ + au_fhsm_wrote(sb, a->bdst, /*force*/0); + goto out_unpin; /* success */ + +out_revert: + /* no delegation since it is just created */ + rerr = vfsub_unlink(au_pinned_h_dir(&a->pin), &a->h_path, + /*delegated*/NULL, /*force*/0); + if (unlikely(rerr)) { + AuIOErr("%pd reverting failed(%d, %d)\n", dentry, err, rerr); + err = -EIO; + } + au_dtime_revert(&dt); +out_unpin: + au_unpin(&a->pin); +out_wh: + dput(wh_dentry); +out_parent: + di_write_unlock(a->parent); + dput(a->src_parent); +out_unlock: + if (unlikely(err)) { + au_update_dbtop(dentry); + d_drop(dentry); + } + aufs_read_and_write_unlock2(dentry, src_dentry); +out_kfree: + au_kfree_rcu(a); +out: + AuTraceErr(err); + return err; +} + +int aufs_mkdir(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode) +{ + int err, rerr; + aufs_bindex_t bindex; + unsigned char diropq; + struct path h_path; + struct dentry *wh_dentry, *parent, *opq_dentry; + struct inode *h_inode; + struct super_block *sb; + struct { + struct au_pin pin; + struct au_dtime dt; + } *a; /* reduce the stack usage */ + struct au_wr_dir_args wr_dir_args = { + .force_btgt = -1, + .flags = AuWrDir_ADD_ENTRY | AuWrDir_ISDIR + }; + + IMustLock(dir); + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + err = aufs_read_lock(dentry, AuLock_DW | AuLock_GEN); + if (unlikely(err)) + goto out_free; + err = au_d_may_add(dentry); + if (unlikely(err)) + goto out_unlock; + + parent = dentry->d_parent; /* dir inode is locked */ + di_write_lock_parent(parent); + wh_dentry = lock_hdir_lkup_wh(dentry, &a->dt, /*src_dentry*/NULL, + &a->pin, &wr_dir_args); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out_parent; + + sb = dentry->d_sb; + bindex = au_dbtop(dentry); + h_path.dentry = au_h_dptr(dentry, bindex); + h_path.mnt = au_sbr_mnt(sb, bindex); + err = vfsub_mkdir(au_pinned_h_dir(&a->pin), &h_path, mode); + if (unlikely(err)) + goto out_unpin; + + /* make the dir opaque */ + diropq = 0; + h_inode = d_inode(h_path.dentry); + if (wh_dentry + || au_opt_test(au_mntflags(sb), ALWAYS_DIROPQ)) { + inode_lock_nested(h_inode, AuLsc_I_CHILD); + opq_dentry = au_diropq_create(dentry, bindex); + inode_unlock(h_inode); + err = PTR_ERR(opq_dentry); + if (IS_ERR(opq_dentry)) + goto out_dir; + dput(opq_dentry); + diropq = 1; + } + + err = epilog(dir, bindex, wh_dentry, dentry); + if (!err) { + inc_nlink(dir); + goto out_unpin; /* success */ + } + + /* revert */ + if (diropq) { + AuLabel(revert opq); + inode_lock_nested(h_inode, AuLsc_I_CHILD); + rerr = au_diropq_remove(dentry, bindex); + inode_unlock(h_inode); + if (rerr) { + AuIOErr("%pd reverting diropq failed(%d, %d)\n", + dentry, err, rerr); + err = -EIO; + } + } + +out_dir: + AuLabel(revert dir); + rerr = vfsub_rmdir(au_pinned_h_dir(&a->pin), &h_path); + if (rerr) { + AuIOErr("%pd reverting dir failed(%d, %d)\n", + dentry, err, rerr); + err = -EIO; + } + au_dtime_revert(&a->dt); +out_unpin: + au_unpin(&a->pin); + dput(wh_dentry); +out_parent: + di_write_unlock(parent); +out_unlock: + if (unlikely(err)) { + au_update_dbtop(dentry); + d_drop(dentry); + } + aufs_read_unlock(dentry, AuLock_DW); +out_free: + au_kfree_rcu(a); +out: + return err; +} diff --git a/fs/aufs/i_op_del.c b/fs/aufs/i_op_del.c new file mode 100644 index 00000000000..f7930a66ccf --- /dev/null +++ b/fs/aufs/i_op_del.c @@ -0,0 +1,515 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode operations (del entry) + */ + +#include +#include "aufs.h" + +/* + * decide if a new whiteout for @dentry is necessary or not. + * when it is necessary, prepare the parent dir for the upper branch whose + * branch index is @bcpup for creation. the actual creation of the whiteout will + * be done by caller. + * return value: + * 0: wh is unnecessary + * plus: wh is necessary + * minus: error + */ +int au_wr_dir_need_wh(struct dentry *dentry, int isdir, aufs_bindex_t *bcpup) +{ + int need_wh, err; + aufs_bindex_t btop; + struct super_block *sb; + + sb = dentry->d_sb; + btop = au_dbtop(dentry); + if (*bcpup < 0) { + *bcpup = btop; + if (au_test_ro(sb, btop, d_inode(dentry))) { + err = AuWbrCopyup(au_sbi(sb), dentry); + *bcpup = err; + if (unlikely(err < 0)) + goto out; + } + } else + AuDebugOn(btop < *bcpup + || au_test_ro(sb, *bcpup, d_inode(dentry))); + AuDbg("bcpup %d, btop %d\n", *bcpup, btop); + + if (*bcpup != btop) { + err = au_cpup_dirs(dentry, *bcpup); + if (unlikely(err)) + goto out; + need_wh = 1; + } else { + struct au_dinfo *dinfo, *tmp; + + need_wh = -ENOMEM; + dinfo = au_di(dentry); + tmp = au_di_alloc(sb, AuLsc_DI_TMP); + if (tmp) { + au_di_cp(tmp, dinfo); + au_di_swap(tmp, dinfo); + /* returns the number of positive dentries */ + need_wh = au_lkup_dentry(dentry, btop + 1, + /* AuLkup_IGNORE_PERM */ 0); + au_di_swap(tmp, dinfo); + au_rw_write_unlock(&tmp->di_rwsem); + au_di_free(tmp); + } + } + AuDbg("need_wh %d\n", need_wh); + err = need_wh; + +out: + return err; +} + +/* + * simple tests for the del-entry operations. + * following the checks in vfs, plus the parent-child relationship. + */ +int au_may_del(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent, int isdir) +{ + int err; + umode_t h_mode; + struct dentry *h_dentry, *h_latest; + struct inode *h_inode; + struct user_namespace *h_userns; + + h_dentry = au_h_dptr(dentry, bindex); + if (d_really_is_positive(dentry)) { + err = -ENOENT; + if (unlikely(d_is_negative(h_dentry))) + goto out; + h_inode = d_inode(h_dentry); + if (unlikely(!h_inode->i_nlink)) + goto out; + + h_mode = h_inode->i_mode; + if (!isdir) { + err = -EISDIR; + if (unlikely(S_ISDIR(h_mode))) + goto out; + } else if (unlikely(!S_ISDIR(h_mode))) { + err = -ENOTDIR; + goto out; + } + } else { + /* rename(2) case */ + err = -EIO; + if (unlikely(d_is_positive(h_dentry))) + goto out; + } + + err = -ENOENT; + /* expected parent dir is locked */ + if (unlikely(h_parent != h_dentry->d_parent)) + goto out; + err = 0; + + /* + * rmdir a dir may break the consistency on some filesystem. + * let's try heavy test. + */ + err = -EACCES; + h_userns = au_sbr_userns(dentry->d_sb, bindex); + if (unlikely(!au_opt_test(au_mntflags(dentry->d_sb), DIRPERM1) + && au_test_h_perm(h_userns, d_inode(h_parent), + MAY_EXEC | MAY_WRITE))) + goto out; + + h_latest = au_sio_lkup_one(h_userns, &dentry->d_name, h_parent); + err = -EIO; + if (IS_ERR(h_latest)) + goto out; + if (h_latest == h_dentry) + err = 0; + dput(h_latest); + +out: + return err; +} + +/* + * decide the branch where we operate for @dentry. the branch index will be set + * @rbcpup. after deciding it, 'pin' it and store the timestamps of the parent + * dir for reverting. + * when a new whiteout is necessary, create it. + */ +static struct dentry* +lock_hdir_create_wh(struct dentry *dentry, int isdir, aufs_bindex_t *rbcpup, + struct au_dtime *dt, struct au_pin *pin) +{ + struct dentry *wh_dentry; + struct super_block *sb; + struct path h_path; + int err, need_wh; + unsigned int udba; + aufs_bindex_t bcpup; + + need_wh = au_wr_dir_need_wh(dentry, isdir, rbcpup); + wh_dentry = ERR_PTR(need_wh); + if (unlikely(need_wh < 0)) + goto out; + + sb = dentry->d_sb; + udba = au_opt_udba(sb); + bcpup = *rbcpup; + err = au_pin(pin, dentry, bcpup, udba, + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + wh_dentry = ERR_PTR(err); + if (unlikely(err)) + goto out; + + h_path.dentry = au_pinned_h_parent(pin); + if (udba != AuOpt_UDBA_NONE + && au_dbtop(dentry) == bcpup) { + err = au_may_del(dentry, bcpup, h_path.dentry, isdir); + wh_dentry = ERR_PTR(err); + if (unlikely(err)) + goto out_unpin; + } + + h_path.mnt = au_sbr_mnt(sb, bcpup); + au_dtime_store(dt, au_pinned_parent(pin), &h_path); + wh_dentry = NULL; + if (!need_wh) + goto out; /* success, no need to create whiteout */ + + wh_dentry = au_wh_create(dentry, bcpup, h_path.dentry); + if (IS_ERR(wh_dentry)) + goto out_unpin; + + /* returns with the parent is locked and wh_dentry is dget-ed */ + goto out; /* success */ + +out_unpin: + au_unpin(pin); +out: + return wh_dentry; +} + +/* + * when removing a dir, rename it to a unique temporary whiteout-ed name first + * in order to be revertible and save time for removing many child whiteouts + * under the dir. + * returns 1 when there are too many child whiteout and caller should remove + * them asynchronously. returns 0 when the number of children is enough small to + * remove now or the branch fs is a remote fs. + * otherwise return an error. + */ +static int renwh_and_rmdir(struct dentry *dentry, aufs_bindex_t bindex, + struct au_nhash *whlist, struct inode *dir) +{ + int rmdir_later, err, dirwh; + struct dentry *h_dentry; + struct super_block *sb; + struct inode *inode; + + sb = dentry->d_sb; + SiMustAnyLock(sb); + h_dentry = au_h_dptr(dentry, bindex); + err = au_whtmp_ren(h_dentry, au_sbr(sb, bindex)); + if (unlikely(err)) + goto out; + + /* stop monitoring */ + inode = d_inode(dentry); + au_hn_free(au_hi(inode, bindex)); + + if (!au_test_fs_remote(h_dentry->d_sb)) { + dirwh = au_sbi(sb)->si_dirwh; + rmdir_later = (dirwh <= 1); + if (!rmdir_later) + rmdir_later = au_nhash_test_longer_wh(whlist, bindex, + dirwh); + if (rmdir_later) + return rmdir_later; + } + + err = au_whtmp_rmdir(dir, bindex, h_dentry, whlist); + if (unlikely(err)) { + AuIOErr("rmdir %pd, b%d failed, %d. ignored\n", + h_dentry, bindex, err); + err = 0; + } + +out: + AuTraceErr(err); + return err; +} + +/* + * final procedure for deleting a entry. + * maintain dentry and iattr. + */ +static void epilog(struct inode *dir, struct dentry *dentry, + aufs_bindex_t bindex) +{ + struct inode *inode; + + inode = d_inode(dentry); + d_drop(dentry); + inode->i_ctime = dir->i_ctime; + + au_dir_ts(dir, bindex); + inode_inc_iversion(dir); +} + +/* + * when an error happened, remove the created whiteout and revert everything. + */ +static int do_revert(int err, struct inode *dir, aufs_bindex_t bindex, + aufs_bindex_t bwh, struct dentry *wh_dentry, + struct dentry *dentry, struct au_dtime *dt) +{ + int rerr; + struct path h_path = { + .dentry = wh_dentry, + .mnt = au_sbr_mnt(dir->i_sb, bindex) + }; + + rerr = au_wh_unlink_dentry(au_h_iptr(dir, bindex), &h_path, dentry); + if (!rerr) { + au_set_dbwh(dentry, bwh); + au_dtime_revert(dt); + return 0; + } + + AuIOErr("%pd reverting whiteout failed(%d, %d)\n", dentry, err, rerr); + return -EIO; +} + +/* ---------------------------------------------------------------------- */ + +int aufs_unlink(struct inode *dir, struct dentry *dentry) +{ + int err; + aufs_bindex_t bwh, bindex, btop; + struct inode *inode, *h_dir, *delegated; + struct dentry *parent, *wh_dentry; + /* to reduce stack size */ + struct { + struct au_dtime dt; + struct au_pin pin; + struct path h_path; + } *a; + + IMustLock(dir); + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + err = aufs_read_lock(dentry, AuLock_DW | AuLock_GEN); + if (unlikely(err)) + goto out_free; + err = au_d_hashed_positive(dentry); + if (unlikely(err)) + goto out_unlock; + inode = d_inode(dentry); + IMustLock(inode); + err = -EISDIR; + if (unlikely(d_is_dir(dentry))) + goto out_unlock; /* possible? */ + + btop = au_dbtop(dentry); + bwh = au_dbwh(dentry); + bindex = -1; + parent = dentry->d_parent; /* dir inode is locked */ + di_write_lock_parent(parent); + wh_dentry = lock_hdir_create_wh(dentry, /*isdir*/0, &bindex, &a->dt, + &a->pin); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out_parent; + + a->h_path.mnt = au_sbr_mnt(dentry->d_sb, btop); + a->h_path.dentry = au_h_dptr(dentry, btop); + dget(a->h_path.dentry); + if (bindex == btop) { + h_dir = au_pinned_h_dir(&a->pin); + delegated = NULL; + err = vfsub_unlink(h_dir, &a->h_path, &delegated, /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + } else { + /* dir inode is locked */ + h_dir = d_inode(wh_dentry->d_parent); + IMustLock(h_dir); + err = 0; + } + + if (!err) { + vfsub_drop_nlink(inode); + epilog(dir, dentry, bindex); + + /* update target timestamps */ + if (bindex == btop) { + vfsub_update_h_iattr(&a->h_path, /*did*/NULL); + /*ignore*/ + inode->i_ctime = d_inode(a->h_path.dentry)->i_ctime; + } else + /* todo: this timestamp may be reverted later */ + inode->i_ctime = h_dir->i_ctime; + goto out_unpin; /* success */ + } + + /* revert */ + if (wh_dentry) { + int rerr; + + rerr = do_revert(err, dir, bindex, bwh, wh_dentry, dentry, + &a->dt); + if (rerr) + err = rerr; + } + +out_unpin: + au_unpin(&a->pin); + dput(wh_dentry); + dput(a->h_path.dentry); +out_parent: + di_write_unlock(parent); +out_unlock: + aufs_read_unlock(dentry, AuLock_DW); +out_free: + au_kfree_rcu(a); +out: + return err; +} + +int aufs_rmdir(struct inode *dir, struct dentry *dentry) +{ + int err, rmdir_later; + aufs_bindex_t bwh, bindex, btop; + struct inode *inode; + struct dentry *parent, *wh_dentry, *h_dentry; + struct au_whtmp_rmdir *args; + /* to reduce stack size */ + struct { + struct au_dtime dt; + struct au_pin pin; + } *a; + + IMustLock(dir); + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + err = aufs_read_lock(dentry, AuLock_DW | AuLock_FLUSH | AuLock_GEN); + if (unlikely(err)) + goto out_free; + err = au_alive_dir(dentry); + if (unlikely(err)) + goto out_unlock; + inode = d_inode(dentry); + IMustLock(inode); + err = -ENOTDIR; + if (unlikely(!d_is_dir(dentry))) + goto out_unlock; /* possible? */ + + err = -ENOMEM; + args = au_whtmp_rmdir_alloc(dir->i_sb, GFP_NOFS); + if (unlikely(!args)) + goto out_unlock; + + parent = dentry->d_parent; /* dir inode is locked */ + di_write_lock_parent(parent); + err = au_test_empty(dentry, &args->whlist); + if (unlikely(err)) + goto out_parent; + + btop = au_dbtop(dentry); + bwh = au_dbwh(dentry); + bindex = -1; + wh_dentry = lock_hdir_create_wh(dentry, /*isdir*/1, &bindex, &a->dt, + &a->pin); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) + goto out_parent; + + h_dentry = au_h_dptr(dentry, btop); + dget(h_dentry); + rmdir_later = 0; + if (bindex == btop) { + err = renwh_and_rmdir(dentry, btop, &args->whlist, dir); + if (err > 0) { + rmdir_later = err; + err = 0; + } + } else { + /* stop monitoring */ + au_hn_free(au_hi(inode, btop)); + + /* dir inode is locked */ + IMustLock(d_inode(wh_dentry->d_parent)); + err = 0; + } + + if (!err) { + vfsub_dead_dir(inode); + au_set_dbdiropq(dentry, -1); + epilog(dir, dentry, bindex); + + if (rmdir_later) { + au_whtmp_kick_rmdir(dir, btop, h_dentry, args); + args = NULL; + } + + goto out_unpin; /* success */ + } + + /* revert */ + AuLabel(revert); + if (wh_dentry) { + int rerr; + + rerr = do_revert(err, dir, bindex, bwh, wh_dentry, dentry, + &a->dt); + if (rerr) + err = rerr; + } + +out_unpin: + au_unpin(&a->pin); + dput(wh_dentry); + dput(h_dentry); +out_parent: + di_write_unlock(parent); + if (args) + au_whtmp_rmdir_free(args); +out_unlock: + aufs_read_unlock(dentry, AuLock_DW); +out_free: + au_kfree_rcu(a); +out: + AuTraceErr(err); + return err; +} diff --git a/fs/aufs/i_op_ren.c b/fs/aufs/i_op_ren.c new file mode 100644 index 00000000000..7bd76f8f611 --- /dev/null +++ b/fs/aufs/i_op_ren.c @@ -0,0 +1,1251 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode operation (rename entry) + * todo: this is crazy monster + */ + +#include +#include "aufs.h" + +enum { AuSRC, AuDST, AuSrcDst }; +enum { AuPARENT, AuCHILD, AuParentChild }; + +#define AuRen_ISDIR_SRC 1 +#define AuRen_ISDIR_DST (1 << 1) +#define AuRen_ISSAMEDIR (1 << 2) +#define AuRen_WHSRC (1 << 3) +#define AuRen_WHDST (1 << 4) +#define AuRen_MNT_WRITE (1 << 5) +#define AuRen_DT_DSTDIR (1 << 6) +#define AuRen_DIROPQ_SRC (1 << 7) +#define AuRen_DIROPQ_DST (1 << 8) +#define AuRen_DIRREN (1 << 9) +#define AuRen_DROPPED_SRC (1 << 10) +#define AuRen_DROPPED_DST (1 << 11) +#define au_ftest_ren(flags, name) ((flags) & AuRen_##name) +#define au_fset_ren(flags, name) \ + do { (flags) |= AuRen_##name; } while (0) +#define au_fclr_ren(flags, name) \ + do { (flags) &= ~AuRen_##name; } while (0) + +#ifndef CONFIG_AUFS_DIRREN +#undef AuRen_DIRREN +#define AuRen_DIRREN 0 +#endif + +struct au_ren_args { + struct { + struct dentry *dentry, *h_dentry, *parent, *h_parent, + *wh_dentry; + struct inode *dir, *inode; + struct au_hinode *hdir, *hinode; + struct au_dtime dt[AuParentChild]; + aufs_bindex_t btop, bdiropq; + } sd[AuSrcDst]; + +#define src_dentry sd[AuSRC].dentry +#define src_dir sd[AuSRC].dir +#define src_inode sd[AuSRC].inode +#define src_h_dentry sd[AuSRC].h_dentry +#define src_parent sd[AuSRC].parent +#define src_h_parent sd[AuSRC].h_parent +#define src_wh_dentry sd[AuSRC].wh_dentry +#define src_hdir sd[AuSRC].hdir +#define src_hinode sd[AuSRC].hinode +#define src_h_dir sd[AuSRC].hdir->hi_inode +#define src_dt sd[AuSRC].dt +#define src_btop sd[AuSRC].btop +#define src_bdiropq sd[AuSRC].bdiropq + +#define dst_dentry sd[AuDST].dentry +#define dst_dir sd[AuDST].dir +#define dst_inode sd[AuDST].inode +#define dst_h_dentry sd[AuDST].h_dentry +#define dst_parent sd[AuDST].parent +#define dst_h_parent sd[AuDST].h_parent +#define dst_wh_dentry sd[AuDST].wh_dentry +#define dst_hdir sd[AuDST].hdir +#define dst_hinode sd[AuDST].hinode +#define dst_h_dir sd[AuDST].hdir->hi_inode +#define dst_dt sd[AuDST].dt +#define dst_btop sd[AuDST].btop +#define dst_bdiropq sd[AuDST].bdiropq + + struct dentry *h_trap; + struct au_branch *br; + struct path h_path; + struct au_nhash whlist; + aufs_bindex_t btgt, src_bwh; + + struct { + unsigned short auren_flags; + unsigned char flags; /* syscall parameter */ + unsigned char exchange; + } __packed; + + struct au_whtmp_rmdir *thargs; + struct dentry *h_dst; + struct au_hinode *h_root; +}; + +/* ---------------------------------------------------------------------- */ + +/* + * functions for reverting. + * when an error happened in a single rename systemcall, we should revert + * everything as if nothing happened. + * we don't need to revert the copied-up/down the parent dir since they are + * harmless. + */ + +#define RevertFailure(fmt, ...) do { \ + AuIOErr("revert failure: " fmt " (%d, %d)\n", \ + ##__VA_ARGS__, err, rerr); \ + err = -EIO; \ +} while (0) + +static void au_ren_do_rev_diropq(int err, struct au_ren_args *a, int idx) +{ + int rerr; + struct dentry *d; +#define src_or_dst(member) a->sd[idx].member + + d = src_or_dst(dentry); /* {src,dst}_dentry */ + au_hn_inode_lock_nested(src_or_dst(hinode), AuLsc_I_CHILD); + rerr = au_diropq_remove(d, a->btgt); + au_hn_inode_unlock(src_or_dst(hinode)); + au_set_dbdiropq(d, src_or_dst(bdiropq)); + if (rerr) + RevertFailure("remove diropq %pd", d); + +#undef src_or_dst_ +} + +static void au_ren_rev_diropq(int err, struct au_ren_args *a) +{ + if (au_ftest_ren(a->auren_flags, DIROPQ_SRC)) + au_ren_do_rev_diropq(err, a, AuSRC); + if (au_ftest_ren(a->auren_flags, DIROPQ_DST)) + au_ren_do_rev_diropq(err, a, AuDST); +} + +static void au_ren_rev_rename(int err, struct au_ren_args *a) +{ + int rerr; + struct inode *delegated; + + a->h_path.dentry = vfsub_lkup_one(&a->src_dentry->d_name, + a->src_h_parent); + rerr = PTR_ERR(a->h_path.dentry); + if (IS_ERR(a->h_path.dentry)) { + RevertFailure("lkup one %pd", a->src_dentry); + return; + } + + delegated = NULL; + rerr = vfsub_rename(a->dst_h_dir, + au_h_dptr(a->src_dentry, a->btgt), + a->src_h_dir, &a->h_path, &delegated, a->flags); + if (unlikely(rerr == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal rename\n"); + iput(delegated); + } + d_drop(a->h_path.dentry); + dput(a->h_path.dentry); + /* au_set_h_dptr(a->src_dentry, a->btgt, NULL); */ + if (rerr) + RevertFailure("rename %pd", a->src_dentry); +} + +static void au_ren_rev_whtmp(int err, struct au_ren_args *a) +{ + int rerr; + struct inode *delegated; + + a->h_path.dentry = vfsub_lkup_one(&a->dst_dentry->d_name, + a->dst_h_parent); + rerr = PTR_ERR(a->h_path.dentry); + if (IS_ERR(a->h_path.dentry)) { + RevertFailure("lkup one %pd", a->dst_dentry); + return; + } + if (d_is_positive(a->h_path.dentry)) { + d_drop(a->h_path.dentry); + dput(a->h_path.dentry); + return; + } + + delegated = NULL; + rerr = vfsub_rename(a->dst_h_dir, a->h_dst, a->dst_h_dir, &a->h_path, + &delegated, a->flags); + if (unlikely(rerr == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal rename\n"); + iput(delegated); + } + d_drop(a->h_path.dentry); + dput(a->h_path.dentry); + if (!rerr) + au_set_h_dptr(a->dst_dentry, a->btgt, dget(a->h_dst)); + else + RevertFailure("rename %pd", a->h_dst); +} + +static void au_ren_rev_whsrc(int err, struct au_ren_args *a) +{ + int rerr; + + a->h_path.dentry = a->src_wh_dentry; + rerr = au_wh_unlink_dentry(a->src_h_dir, &a->h_path, a->src_dentry); + au_set_dbwh(a->src_dentry, a->src_bwh); + if (rerr) + RevertFailure("unlink %pd", a->src_wh_dentry); +} +#undef RevertFailure + +/* ---------------------------------------------------------------------- */ + +/* + * when we have to copyup the renaming entry, do it with the rename-target name + * in order to minimize the cost (the later actual rename is unnecessary). + * otherwise rename it on the target branch. + */ +static int au_ren_or_cpup(struct au_ren_args *a) +{ + int err; + struct dentry *d; + struct inode *delegated; + + d = a->src_dentry; + if (au_dbtop(d) == a->btgt) { + a->h_path.dentry = a->dst_h_dentry; + AuDebugOn(au_dbtop(d) != a->btgt); + delegated = NULL; + err = vfsub_rename(a->src_h_dir, au_h_dptr(d, a->btgt), + a->dst_h_dir, &a->h_path, &delegated, + a->flags); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal rename\n"); + iput(delegated); + } + } else + BUG(); + + if (!err && a->h_dst) + /* it will be set to dinfo later */ + dget(a->h_dst); + + return err; +} + +/* cf. aufs_rmdir() */ +static int au_ren_del_whtmp(struct au_ren_args *a) +{ + int err; + struct inode *dir; + + dir = a->dst_dir; + SiMustAnyLock(dir->i_sb); + if (!au_nhash_test_longer_wh(&a->whlist, a->btgt, + au_sbi(dir->i_sb)->si_dirwh) + || au_test_fs_remote(a->h_dst->d_sb)) { + err = au_whtmp_rmdir(dir, a->btgt, a->h_dst, &a->whlist); + if (unlikely(err)) + pr_warn("failed removing whtmp dir %pd (%d), " + "ignored.\n", a->h_dst, err); + } else { + au_nhash_wh_free(&a->thargs->whlist); + a->thargs->whlist = a->whlist; + a->whlist.nh_num = 0; + au_whtmp_kick_rmdir(dir, a->btgt, a->h_dst, a->thargs); + dput(a->h_dst); + a->thargs = NULL; + } + + return 0; +} + +/* make it 'opaque' dir. */ +static int au_ren_do_diropq(struct au_ren_args *a, int idx) +{ + int err; + struct dentry *d, *diropq; +#define src_or_dst(member) a->sd[idx].member + + err = 0; + d = src_or_dst(dentry); /* {src,dst}_dentry */ + src_or_dst(bdiropq) = au_dbdiropq(d); + src_or_dst(hinode) = au_hi(src_or_dst(inode), a->btgt); + au_hn_inode_lock_nested(src_or_dst(hinode), AuLsc_I_CHILD); + diropq = au_diropq_create(d, a->btgt); + au_hn_inode_unlock(src_or_dst(hinode)); + if (IS_ERR(diropq)) + err = PTR_ERR(diropq); + else + dput(diropq); + +#undef src_or_dst_ + return err; +} + +static int au_ren_diropq(struct au_ren_args *a) +{ + int err; + unsigned char always; + struct dentry *d; + + err = 0; + d = a->dst_dentry; /* already renamed on the branch */ + always = !!au_opt_test(au_mntflags(d->d_sb), ALWAYS_DIROPQ); + if (au_ftest_ren(a->auren_flags, ISDIR_SRC) + && !au_ftest_ren(a->auren_flags, DIRREN) + && a->btgt != au_dbdiropq(a->src_dentry) + && (a->dst_wh_dentry + || a->btgt <= au_dbdiropq(d) + /* hide the lower to keep xino */ + /* the lowers may not be a dir, but we hide them anyway */ + || a->btgt < au_dbbot(d) + || always)) { + AuDbg("here\n"); + err = au_ren_do_diropq(a, AuSRC); + if (unlikely(err)) + goto out; + au_fset_ren(a->auren_flags, DIROPQ_SRC); + } + if (!a->exchange) + goto out; /* success */ + + d = a->src_dentry; /* already renamed on the branch */ + if (au_ftest_ren(a->auren_flags, ISDIR_DST) + && a->btgt != au_dbdiropq(a->dst_dentry) + && (a->btgt < au_dbdiropq(d) + || a->btgt < au_dbbot(d) + || always)) { + AuDbgDentry(a->src_dentry); + AuDbgDentry(a->dst_dentry); + err = au_ren_do_diropq(a, AuDST); + if (unlikely(err)) + goto out_rev_src; + au_fset_ren(a->auren_flags, DIROPQ_DST); + } + goto out; /* success */ + +out_rev_src: + AuDbg("err %d, reverting src\n", err); + au_ren_rev_diropq(err, a); +out: + return err; +} + +static int do_rename(struct au_ren_args *a) +{ + int err; + struct dentry *d, *h_d; + + if (!a->exchange) { + /* prepare workqueue args for asynchronous rmdir */ + h_d = a->dst_h_dentry; + if (au_ftest_ren(a->auren_flags, ISDIR_DST) + /* && !au_ftest_ren(a->auren_flags, DIRREN) */ + && d_is_positive(h_d)) { + err = -ENOMEM; + a->thargs = au_whtmp_rmdir_alloc(a->src_dentry->d_sb, + GFP_NOFS); + if (unlikely(!a->thargs)) + goto out; + a->h_dst = dget(h_d); + } + + /* create whiteout for src_dentry */ + if (au_ftest_ren(a->auren_flags, WHSRC)) { + a->src_bwh = au_dbwh(a->src_dentry); + AuDebugOn(a->src_bwh >= 0); + a->src_wh_dentry = au_wh_create(a->src_dentry, a->btgt, + a->src_h_parent); + err = PTR_ERR(a->src_wh_dentry); + if (IS_ERR(a->src_wh_dentry)) + goto out_thargs; + } + + /* lookup whiteout for dentry */ + if (au_ftest_ren(a->auren_flags, WHDST)) { + h_d = au_wh_lkup(a->dst_h_parent, + &a->dst_dentry->d_name, a->br); + err = PTR_ERR(h_d); + if (IS_ERR(h_d)) + goto out_whsrc; + if (d_is_negative(h_d)) + dput(h_d); + else + a->dst_wh_dentry = h_d; + } + + /* rename dentry to tmpwh */ + if (a->thargs) { + err = au_whtmp_ren(a->dst_h_dentry, a->br); + if (unlikely(err)) + goto out_whdst; + + d = a->dst_dentry; + au_set_h_dptr(d, a->btgt, NULL); + err = au_lkup_neg(d, a->btgt, /*wh*/0); + if (unlikely(err)) + goto out_whtmp; + a->dst_h_dentry = au_h_dptr(d, a->btgt); + } + } + + BUG_ON(d_is_positive(a->dst_h_dentry) && a->src_btop != a->btgt); +#if 0 /* debugging */ + BUG_ON(!au_ftest_ren(a->auren_flags, DIRREN) + && d_is_positive(a->dst_h_dentry) + && a->src_btop != a->btgt); +#endif + + /* rename by vfs_rename or cpup */ + err = au_ren_or_cpup(a); + if (unlikely(err)) + /* leave the copied-up one */ + goto out_whtmp; + + /* make dir opaque */ + err = au_ren_diropq(a); + if (unlikely(err)) + goto out_rename; + + /* update target timestamps */ + if (a->exchange) { + AuDebugOn(au_dbtop(a->dst_dentry) != a->btgt); + a->h_path.dentry = au_h_dptr(a->dst_dentry, a->btgt); + vfsub_update_h_iattr(&a->h_path, /*did*/NULL); /*ignore*/ + a->dst_inode->i_ctime = d_inode(a->h_path.dentry)->i_ctime; + } + AuDebugOn(au_dbtop(a->src_dentry) != a->btgt); + a->h_path.dentry = au_h_dptr(a->src_dentry, a->btgt); + vfsub_update_h_iattr(&a->h_path, /*did*/NULL); /*ignore*/ + a->src_inode->i_ctime = d_inode(a->h_path.dentry)->i_ctime; + + if (!a->exchange) { + /* remove whiteout for dentry */ + if (a->dst_wh_dentry) { + a->h_path.dentry = a->dst_wh_dentry; + err = au_wh_unlink_dentry(a->dst_h_dir, &a->h_path, + a->dst_dentry); + if (unlikely(err)) + goto out_diropq; + } + + /* remove whtmp */ + if (a->thargs) + au_ren_del_whtmp(a); /* ignore this error */ + + au_fhsm_wrote(a->src_dentry->d_sb, a->btgt, /*force*/0); + } + err = 0; + goto out_success; + +out_diropq: + au_ren_rev_diropq(err, a); +out_rename: + au_ren_rev_rename(err, a); + dput(a->h_dst); +out_whtmp: + if (a->thargs) + au_ren_rev_whtmp(err, a); +out_whdst: + dput(a->dst_wh_dentry); + a->dst_wh_dentry = NULL; +out_whsrc: + if (a->src_wh_dentry) + au_ren_rev_whsrc(err, a); +out_success: + dput(a->src_wh_dentry); + dput(a->dst_wh_dentry); +out_thargs: + if (a->thargs) { + dput(a->h_dst); + au_whtmp_rmdir_free(a->thargs); + a->thargs = NULL; + } +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * test if @dentry dir can be rename destination or not. + * success means, it is a logically empty dir. + */ +static int may_rename_dstdir(struct dentry *dentry, struct au_nhash *whlist) +{ + return au_test_empty(dentry, whlist); +} + +/* + * test if @a->src_dentry dir can be rename source or not. + * if it can, return 0. + * success means, + * - it is a logically empty dir. + * - or, it exists on writable branch and has no children including whiteouts + * on the lower branch unless DIRREN is on. + */ +static int may_rename_srcdir(struct au_ren_args *a) +{ + int err; + unsigned int rdhash; + aufs_bindex_t btop, btgt; + struct dentry *dentry; + struct super_block *sb; + struct au_sbinfo *sbinfo; + + dentry = a->src_dentry; + sb = dentry->d_sb; + sbinfo = au_sbi(sb); + if (au_opt_test(sbinfo->si_mntflags, DIRREN)) + au_fset_ren(a->auren_flags, DIRREN); + + btgt = a->btgt; + btop = au_dbtop(dentry); + if (btop != btgt) { + struct au_nhash whlist; + + SiMustAnyLock(sb); + rdhash = sbinfo->si_rdhash; + if (!rdhash) + rdhash = au_rdhash_est(au_dir_size(/*file*/NULL, + dentry)); + err = au_nhash_alloc(&whlist, rdhash, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_test_empty(dentry, &whlist); + au_nhash_wh_free(&whlist); + goto out; + } + + if (btop == au_dbtaildir(dentry)) + return 0; /* success */ + + err = au_test_empty_lower(dentry); + +out: + if (err == -ENOTEMPTY) { + if (au_ftest_ren(a->auren_flags, DIRREN)) { + err = 0; + } else { + AuWarn1("renaming dir who has child(ren) on multiple " + "branches, is not supported\n"); + err = -EXDEV; + } + } + return err; +} + +/* side effect: sets whlist and h_dentry */ +static int au_ren_may_dir(struct au_ren_args *a) +{ + int err; + unsigned int rdhash; + struct dentry *d; + + d = a->dst_dentry; + SiMustAnyLock(d->d_sb); + + err = 0; + if (au_ftest_ren(a->auren_flags, ISDIR_DST) && a->dst_inode) { + rdhash = au_sbi(d->d_sb)->si_rdhash; + if (!rdhash) + rdhash = au_rdhash_est(au_dir_size(/*file*/NULL, d)); + err = au_nhash_alloc(&a->whlist, rdhash, GFP_NOFS); + if (unlikely(err)) + goto out; + + if (!a->exchange) { + au_set_dbtop(d, a->dst_btop); + err = may_rename_dstdir(d, &a->whlist); + au_set_dbtop(d, a->btgt); + } else + err = may_rename_srcdir(a); + } + a->dst_h_dentry = au_h_dptr(d, au_dbtop(d)); + if (unlikely(err)) + goto out; + + d = a->src_dentry; + a->src_h_dentry = au_h_dptr(d, au_dbtop(d)); + if (au_ftest_ren(a->auren_flags, ISDIR_SRC)) { + err = may_rename_srcdir(a); + if (unlikely(err)) { + au_nhash_wh_free(&a->whlist); + a->whlist.nh_num = 0; + } + } +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * simple tests for rename. + * following the checks in vfs, plus the parent-child relationship. + */ +static int au_may_ren(struct au_ren_args *a) +{ + int err, isdir; + struct inode *h_inode; + + if (a->src_btop == a->btgt) { + err = au_may_del(a->src_dentry, a->btgt, a->src_h_parent, + au_ftest_ren(a->auren_flags, ISDIR_SRC)); + if (unlikely(err)) + goto out; + err = -EINVAL; + if (unlikely(a->src_h_dentry == a->h_trap)) + goto out; + } + + err = 0; + if (a->dst_btop != a->btgt) + goto out; + + err = -ENOTEMPTY; + if (unlikely(a->dst_h_dentry == a->h_trap)) + goto out; + + err = -EIO; + isdir = !!au_ftest_ren(a->auren_flags, ISDIR_DST); + if (d_really_is_negative(a->dst_dentry)) { + if (d_is_negative(a->dst_h_dentry)) + err = au_may_add(a->dst_dentry, a->btgt, + a->dst_h_parent, isdir); + } else { + if (unlikely(d_is_negative(a->dst_h_dentry))) + goto out; + h_inode = d_inode(a->dst_h_dentry); + if (h_inode->i_nlink) + err = au_may_del(a->dst_dentry, a->btgt, + a->dst_h_parent, isdir); + } + +out: + if (unlikely(err == -ENOENT || err == -EEXIST)) + err = -EIO; + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * locking order + * (VFS) + * - src_dir and dir by lock_rename() + * - inode if exists + * (aufs) + * - lock all + * + src_dentry and dentry by aufs_read_and_write_lock2() which calls, + * + si_read_lock + * + di_write_lock2_child() + * + di_write_lock_child() + * + ii_write_lock_child() + * + di_write_lock_child2() + * + ii_write_lock_child2() + * + src_parent and parent + * + di_write_lock_parent() + * + ii_write_lock_parent() + * + di_write_lock_parent2() + * + ii_write_lock_parent2() + * + lower src_dir and dir by vfsub_lock_rename() + * + verify the every relationships between child and parent. if any + * of them failed, unlock all and return -EBUSY. + */ +static void au_ren_unlock(struct au_ren_args *a) +{ + vfsub_unlock_rename(a->src_h_parent, a->src_hdir, + a->dst_h_parent, a->dst_hdir); + if (au_ftest_ren(a->auren_flags, DIRREN) + && a->h_root) + au_hn_inode_unlock(a->h_root); + if (au_ftest_ren(a->auren_flags, MNT_WRITE)) + vfsub_mnt_drop_write(au_br_mnt(a->br)); +} + +static int au_ren_lock(struct au_ren_args *a) +{ + int err; + unsigned int udba; + + err = 0; + a->src_h_parent = au_h_dptr(a->src_parent, a->btgt); + a->src_hdir = au_hi(a->src_dir, a->btgt); + a->dst_h_parent = au_h_dptr(a->dst_parent, a->btgt); + a->dst_hdir = au_hi(a->dst_dir, a->btgt); + + err = vfsub_mnt_want_write(au_br_mnt(a->br)); + if (unlikely(err)) + goto out; + au_fset_ren(a->auren_flags, MNT_WRITE); + if (au_ftest_ren(a->auren_flags, DIRREN)) { + struct dentry *root; + struct inode *dir; + + /* + * sbinfo is already locked, so this ii_read_lock is + * unnecessary. but our debugging feature checks it. + */ + root = a->src_inode->i_sb->s_root; + if (root != a->src_parent && root != a->dst_parent) { + dir = d_inode(root); + ii_read_lock_parent3(dir); + a->h_root = au_hi(dir, a->btgt); + ii_read_unlock(dir); + au_hn_inode_lock_nested(a->h_root, AuLsc_I_PARENT3); + } + } + a->h_trap = vfsub_lock_rename(a->src_h_parent, a->src_hdir, + a->dst_h_parent, a->dst_hdir); + udba = au_opt_udba(a->src_dentry->d_sb); + if (unlikely(a->src_hdir->hi_inode != d_inode(a->src_h_parent) + || a->dst_hdir->hi_inode != d_inode(a->dst_h_parent))) + err = au_busy_or_stale(); + if (!err && au_dbtop(a->src_dentry) == a->btgt) + err = au_h_verify(a->src_h_dentry, udba, + d_inode(a->src_h_parent), a->src_h_parent, + a->br); + if (!err && au_dbtop(a->dst_dentry) == a->btgt) + err = au_h_verify(a->dst_h_dentry, udba, + d_inode(a->dst_h_parent), a->dst_h_parent, + a->br); + if (!err) + goto out; /* success */ + + err = au_busy_or_stale(); + au_ren_unlock(a); + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +static void au_ren_refresh_dir(struct au_ren_args *a) +{ + struct inode *dir; + + dir = a->dst_dir; + inode_inc_iversion(dir); + if (au_ftest_ren(a->auren_flags, ISDIR_SRC)) { + /* is this updating defined in POSIX? */ + au_cpup_attr_timesizes(a->src_inode); + au_cpup_attr_nlink(dir, /*force*/1); + } + au_dir_ts(dir, a->btgt); + + if (a->exchange) { + dir = a->src_dir; + inode_inc_iversion(dir); + if (au_ftest_ren(a->auren_flags, ISDIR_DST)) { + /* is this updating defined in POSIX? */ + au_cpup_attr_timesizes(a->dst_inode); + au_cpup_attr_nlink(dir, /*force*/1); + } + au_dir_ts(dir, a->btgt); + } + + if (au_ftest_ren(a->auren_flags, ISSAMEDIR)) + return; + + dir = a->src_dir; + inode_inc_iversion(dir); + if (au_ftest_ren(a->auren_flags, ISDIR_SRC)) + au_cpup_attr_nlink(dir, /*force*/1); + au_dir_ts(dir, a->btgt); +} + +static void au_ren_refresh(struct au_ren_args *a) +{ + aufs_bindex_t bbot, bindex; + struct dentry *d, *h_d; + struct inode *i, *h_i; + struct super_block *sb; + + d = a->dst_dentry; + d_drop(d); + if (a->h_dst) + /* already dget-ed by au_ren_or_cpup() */ + au_set_h_dptr(d, a->btgt, a->h_dst); + + i = a->dst_inode; + if (i) { + if (!a->exchange) { + if (!au_ftest_ren(a->auren_flags, ISDIR_DST)) + vfsub_drop_nlink(i); + else { + vfsub_dead_dir(i); + au_cpup_attr_timesizes(i); + } + au_update_dbrange(d, /*do_put_zero*/1); + } else + au_cpup_attr_nlink(i, /*force*/1); + } else { + bbot = a->btgt; + for (bindex = au_dbtop(d); bindex < bbot; bindex++) + au_set_h_dptr(d, bindex, NULL); + bbot = au_dbbot(d); + for (bindex = a->btgt + 1; bindex <= bbot; bindex++) + au_set_h_dptr(d, bindex, NULL); + au_update_dbrange(d, /*do_put_zero*/0); + } + + if (a->exchange + || au_ftest_ren(a->auren_flags, DIRREN)) { + d_drop(a->src_dentry); + if (au_ftest_ren(a->auren_flags, DIRREN)) + au_set_dbwh(a->src_dentry, -1); + return; + } + + d = a->src_dentry; + au_set_dbwh(d, -1); + bbot = au_dbbot(d); + for (bindex = a->btgt + 1; bindex <= bbot; bindex++) { + h_d = au_h_dptr(d, bindex); + if (h_d) + au_set_h_dptr(d, bindex, NULL); + } + au_set_dbbot(d, a->btgt); + + sb = d->d_sb; + i = a->src_inode; + if (au_opt_test(au_mntflags(sb), PLINK) && au_plink_test(i)) + return; /* success */ + + bbot = au_ibbot(i); + for (bindex = a->btgt + 1; bindex <= bbot; bindex++) { + h_i = au_h_iptr(i, bindex); + if (h_i) { + au_xino_write(sb, bindex, h_i->i_ino, /*ino*/0); + /* ignore this error */ + au_set_h_iptr(i, bindex, NULL, 0); + } + } + au_set_ibbot(i, a->btgt); +} + +/* ---------------------------------------------------------------------- */ + +/* mainly for link(2) and rename(2) */ +int au_wbr(struct dentry *dentry, aufs_bindex_t btgt) +{ + aufs_bindex_t bdiropq, bwh; + struct dentry *parent; + struct au_branch *br; + + parent = dentry->d_parent; + IMustLock(d_inode(parent)); /* dir is locked */ + + bdiropq = au_dbdiropq(parent); + bwh = au_dbwh(dentry); + br = au_sbr(dentry->d_sb, btgt); + if (au_br_rdonly(br) + || (0 <= bdiropq && bdiropq < btgt) + || (0 <= bwh && bwh < btgt)) + btgt = -1; + + AuDbg("btgt %d\n", btgt); + return btgt; +} + +/* sets src_btop, dst_btop and btgt */ +static int au_ren_wbr(struct au_ren_args *a) +{ + int err; + struct au_wr_dir_args wr_dir_args = { + /* .force_btgt = -1, */ + .flags = AuWrDir_ADD_ENTRY + }; + + a->src_btop = au_dbtop(a->src_dentry); + a->dst_btop = au_dbtop(a->dst_dentry); + if (au_ftest_ren(a->auren_flags, ISDIR_SRC) + || au_ftest_ren(a->auren_flags, ISDIR_DST)) + au_fset_wrdir(wr_dir_args.flags, ISDIR); + wr_dir_args.force_btgt = a->src_btop; + if (a->dst_inode && a->dst_btop < a->src_btop) + wr_dir_args.force_btgt = a->dst_btop; + wr_dir_args.force_btgt = au_wbr(a->dst_dentry, wr_dir_args.force_btgt); + err = au_wr_dir(a->dst_dentry, a->src_dentry, &wr_dir_args); + a->btgt = err; + if (a->exchange) + au_update_dbtop(a->dst_dentry); + + return err; +} + +static void au_ren_dt(struct au_ren_args *a) +{ + a->h_path.dentry = a->src_h_parent; + au_dtime_store(a->src_dt + AuPARENT, a->src_parent, &a->h_path); + if (!au_ftest_ren(a->auren_flags, ISSAMEDIR)) { + a->h_path.dentry = a->dst_h_parent; + au_dtime_store(a->dst_dt + AuPARENT, a->dst_parent, &a->h_path); + } + + au_fclr_ren(a->auren_flags, DT_DSTDIR); + if (!au_ftest_ren(a->auren_flags, ISDIR_SRC) + && !a->exchange) + return; + + a->h_path.dentry = a->src_h_dentry; + au_dtime_store(a->src_dt + AuCHILD, a->src_dentry, &a->h_path); + if (d_is_positive(a->dst_h_dentry)) { + au_fset_ren(a->auren_flags, DT_DSTDIR); + a->h_path.dentry = a->dst_h_dentry; + au_dtime_store(a->dst_dt + AuCHILD, a->dst_dentry, &a->h_path); + } +} + +static void au_ren_rev_dt(int err, struct au_ren_args *a) +{ + struct dentry *h_d; + struct inode *h_inode; + + au_dtime_revert(a->src_dt + AuPARENT); + if (!au_ftest_ren(a->auren_flags, ISSAMEDIR)) + au_dtime_revert(a->dst_dt + AuPARENT); + + if (au_ftest_ren(a->auren_flags, ISDIR_SRC) && err != -EIO) { + h_d = a->src_dt[AuCHILD].dt_h_path.dentry; + h_inode = d_inode(h_d); + inode_lock_nested(h_inode, AuLsc_I_CHILD); + au_dtime_revert(a->src_dt + AuCHILD); + inode_unlock(h_inode); + + if (au_ftest_ren(a->auren_flags, DT_DSTDIR)) { + h_d = a->dst_dt[AuCHILD].dt_h_path.dentry; + h_inode = d_inode(h_d); + inode_lock_nested(h_inode, AuLsc_I_CHILD); + au_dtime_revert(a->dst_dt + AuCHILD); + inode_unlock(h_inode); + } + } +} + +/* ---------------------------------------------------------------------- */ + +int aufs_rename(struct user_namespace *userns, + struct inode *_src_dir, struct dentry *_src_dentry, + struct inode *_dst_dir, struct dentry *_dst_dentry, + unsigned int _flags) +{ + int err, lock_flags; + void *rev; + /* reduce stack space */ + struct au_ren_args *a; + struct au_pin pin; + + AuDbg("%pd, %pd, 0x%x\n", _src_dentry, _dst_dentry, _flags); + IMustLock(_src_dir); + IMustLock(_dst_dir); + + err = -EINVAL; + if (unlikely(_flags & RENAME_WHITEOUT)) + goto out; + + err = -ENOMEM; + BUILD_BUG_ON(sizeof(*a) > PAGE_SIZE); + a = kzalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + a->flags = _flags; + BUILD_BUG_ON(sizeof(a->exchange) == sizeof(u8) + && RENAME_EXCHANGE > U8_MAX); + a->exchange = _flags & RENAME_EXCHANGE; + a->src_dir = _src_dir; + a->src_dentry = _src_dentry; + a->src_inode = NULL; + if (d_really_is_positive(a->src_dentry)) + a->src_inode = d_inode(a->src_dentry); + a->src_parent = a->src_dentry->d_parent; /* dir inode is locked */ + a->dst_dir = _dst_dir; + a->dst_dentry = _dst_dentry; + a->dst_inode = NULL; + if (d_really_is_positive(a->dst_dentry)) + a->dst_inode = d_inode(a->dst_dentry); + a->dst_parent = a->dst_dentry->d_parent; /* dir inode is locked */ + if (a->dst_inode) { + /* + * if EXCHANGE && src is non-dir && dst is dir, + * dst is not locked. + */ + /* IMustLock(a->dst_inode); */ + au_igrab(a->dst_inode); + } + + err = -ENOTDIR; + lock_flags = AuLock_FLUSH | AuLock_NOPLM | AuLock_GEN; + if (d_is_dir(a->src_dentry)) { + au_fset_ren(a->auren_flags, ISDIR_SRC); + if (unlikely(!a->exchange + && d_really_is_positive(a->dst_dentry) + && !d_is_dir(a->dst_dentry))) + goto out_free; + lock_flags |= AuLock_DIRS; + } + if (a->dst_inode && d_is_dir(a->dst_dentry)) { + au_fset_ren(a->auren_flags, ISDIR_DST); + if (unlikely(!a->exchange + && d_really_is_positive(a->src_dentry) + && !d_is_dir(a->src_dentry))) + goto out_free; + lock_flags |= AuLock_DIRS; + } + err = aufs_read_and_write_lock2(a->dst_dentry, a->src_dentry, + lock_flags); + if (unlikely(err)) + goto out_free; + + err = au_d_hashed_positive(a->src_dentry); + if (unlikely(err)) + goto out_unlock; + err = -ENOENT; + if (a->dst_inode) { + /* + * If it is a dir, VFS unhash it before this + * function. It means we cannot rely upon d_unhashed(). + */ + if (unlikely(!a->dst_inode->i_nlink)) + goto out_unlock; + if (!au_ftest_ren(a->auren_flags, ISDIR_DST)) { + err = au_d_hashed_positive(a->dst_dentry); + if (unlikely(err && !a->exchange)) + goto out_unlock; + } else if (unlikely(IS_DEADDIR(a->dst_inode))) + goto out_unlock; + } else if (unlikely(d_unhashed(a->dst_dentry))) + goto out_unlock; + + /* + * is it possible? + * yes, it happened (in linux-3.3-rcN) but I don't know why. + * there may exist a problem somewhere else. + */ + err = -EINVAL; + if (unlikely(d_inode(a->dst_parent) == d_inode(a->src_dentry))) + goto out_unlock; + + au_fset_ren(a->auren_flags, ISSAMEDIR); /* temporary */ + di_write_lock_parent(a->dst_parent); + + /* which branch we process */ + err = au_ren_wbr(a); + if (unlikely(err < 0)) + goto out_parent; + a->br = au_sbr(a->dst_dentry->d_sb, a->btgt); + a->h_path.mnt = au_br_mnt(a->br); + + /* are they available to be renamed */ + err = au_ren_may_dir(a); + if (unlikely(err)) + goto out_children; + + /* prepare the writable parent dir on the same branch */ + if (a->dst_btop == a->btgt) { + au_fset_ren(a->auren_flags, WHDST); + } else { + err = au_cpup_dirs(a->dst_dentry, a->btgt); + if (unlikely(err)) + goto out_children; + } + + err = 0; + if (!a->exchange) { + if (a->src_dir != a->dst_dir) { + /* + * this temporary unlock is safe, + * because both dir->i_mutex are locked. + */ + di_write_unlock(a->dst_parent); + di_write_lock_parent(a->src_parent); + err = au_wr_dir_need_wh(a->src_dentry, + au_ftest_ren(a->auren_flags, + ISDIR_SRC), + &a->btgt); + di_write_unlock(a->src_parent); + di_write_lock2_parent(a->src_parent, a->dst_parent, + /*isdir*/1); + au_fclr_ren(a->auren_flags, ISSAMEDIR); + } else + err = au_wr_dir_need_wh(a->src_dentry, + au_ftest_ren(a->auren_flags, + ISDIR_SRC), + &a->btgt); + } + if (unlikely(err < 0)) + goto out_children; + if (err) + au_fset_ren(a->auren_flags, WHSRC); + + /* cpup src */ + if (a->src_btop != a->btgt) { + err = au_pin(&pin, a->src_dentry, a->btgt, + au_opt_udba(a->src_dentry->d_sb), + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (!err) { + struct au_cp_generic cpg = { + .dentry = a->src_dentry, + .bdst = a->btgt, + .bsrc = a->src_btop, + .len = -1, + .pin = &pin, + .flags = AuCpup_DTIME | AuCpup_HOPEN + }; + AuDebugOn(au_dbtop(a->src_dentry) != a->src_btop); + err = au_sio_cpup_simple(&cpg); + au_unpin(&pin); + } + if (unlikely(err)) + goto out_children; + a->src_btop = a->btgt; + a->src_h_dentry = au_h_dptr(a->src_dentry, a->btgt); + if (!a->exchange) + au_fset_ren(a->auren_flags, WHSRC); + } + + /* cpup dst */ + if (a->exchange && a->dst_inode + && a->dst_btop != a->btgt) { + err = au_pin(&pin, a->dst_dentry, a->btgt, + au_opt_udba(a->dst_dentry->d_sb), + AuPin_DI_LOCKED | AuPin_MNT_WRITE); + if (!err) { + struct au_cp_generic cpg = { + .dentry = a->dst_dentry, + .bdst = a->btgt, + .bsrc = a->dst_btop, + .len = -1, + .pin = &pin, + .flags = AuCpup_DTIME | AuCpup_HOPEN + }; + err = au_sio_cpup_simple(&cpg); + au_unpin(&pin); + } + if (unlikely(err)) + goto out_children; + a->dst_btop = a->btgt; + a->dst_h_dentry = au_h_dptr(a->dst_dentry, a->btgt); + } + + /* lock them all */ + err = au_ren_lock(a); + if (unlikely(err)) + /* leave the copied-up one */ + goto out_children; + + if (!a->exchange) { + if (!au_opt_test(au_mntflags(a->dst_dir->i_sb), UDBA_NONE)) + err = au_may_ren(a); + else if (unlikely(a->dst_dentry->d_name.len > AUFS_MAX_NAMELEN)) + err = -ENAMETOOLONG; + if (unlikely(err)) + goto out_hdir; + } + + /* store timestamps to be revertible */ + au_ren_dt(a); + + /* store dirren info */ + if (au_ftest_ren(a->auren_flags, DIRREN)) { + err = au_dr_rename(a->src_dentry, a->btgt, + &a->dst_dentry->d_name, &rev); + AuTraceErr(err); + if (unlikely(err)) + goto out_dt; + } + + /* here we go */ + err = do_rename(a); + if (unlikely(err)) + goto out_dirren; + + if (au_ftest_ren(a->auren_flags, DIRREN)) + au_dr_rename_fin(a->src_dentry, a->btgt, rev); + + /* update dir attributes */ + au_ren_refresh_dir(a); + + /* dput/iput all lower dentries */ + au_ren_refresh(a); + + goto out_hdir; /* success */ + +out_dirren: + if (au_ftest_ren(a->auren_flags, DIRREN)) + au_dr_rename_rev(a->src_dentry, a->btgt, rev); +out_dt: + au_ren_rev_dt(err, a); +out_hdir: + au_ren_unlock(a); +out_children: + au_nhash_wh_free(&a->whlist); + if (err && a->dst_inode && a->dst_btop != a->btgt) { + AuDbg("btop %d, btgt %d\n", a->dst_btop, a->btgt); + au_set_h_dptr(a->dst_dentry, a->btgt, NULL); + au_set_dbtop(a->dst_dentry, a->dst_btop); + } +out_parent: + if (!err) { + if (d_unhashed(a->src_dentry)) + au_fset_ren(a->auren_flags, DROPPED_SRC); + if (d_unhashed(a->dst_dentry)) + au_fset_ren(a->auren_flags, DROPPED_DST); + if (!a->exchange) + d_move(a->src_dentry, a->dst_dentry); + else { + d_exchange(a->src_dentry, a->dst_dentry); + if (au_ftest_ren(a->auren_flags, DROPPED_DST)) + d_drop(a->dst_dentry); + } + if (au_ftest_ren(a->auren_flags, DROPPED_SRC)) + d_drop(a->src_dentry); + } else { + au_update_dbtop(a->dst_dentry); + if (!a->dst_inode) + d_drop(a->dst_dentry); + } + if (au_ftest_ren(a->auren_flags, ISSAMEDIR)) + di_write_unlock(a->dst_parent); + else + di_write_unlock2(a->src_parent, a->dst_parent); +out_unlock: + aufs_read_and_write_unlock2(a->dst_dentry, a->src_dentry); +out_free: + iput(a->dst_inode); + if (a->thargs) + au_whtmp_rmdir_free(a->thargs); + au_kfree_rcu(a); +out: + AuTraceErr(err); + return err; +} diff --git a/fs/aufs/iinfo.c b/fs/aufs/iinfo.c new file mode 100644 index 00000000000..016db39451c --- /dev/null +++ b/fs/aufs/iinfo.c @@ -0,0 +1,286 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode private data + */ + +#include "aufs.h" + +struct inode *au_h_iptr(struct inode *inode, aufs_bindex_t bindex) +{ + struct inode *h_inode; + struct au_hinode *hinode; + + IiMustAnyLock(inode); + + hinode = au_hinode(au_ii(inode), bindex); + h_inode = hinode->hi_inode; + AuDebugOn(h_inode && atomic_read(&h_inode->i_count) <= 0); + return h_inode; +} + +/* todo: hard/soft set? */ +void au_hiput(struct au_hinode *hinode) +{ + au_hn_free(hinode); + dput(hinode->hi_whdentry); + iput(hinode->hi_inode); +} + +unsigned int au_hi_flags(struct inode *inode, int isdir) +{ + unsigned int flags; + const unsigned int mnt_flags = au_mntflags(inode->i_sb); + + flags = 0; + if (au_opt_test(mnt_flags, XINO)) + au_fset_hi(flags, XINO); + if (isdir && au_opt_test(mnt_flags, UDBA_HNOTIFY)) + au_fset_hi(flags, HNOTIFY); + return flags; +} + +void au_set_h_iptr(struct inode *inode, aufs_bindex_t bindex, + struct inode *h_inode, unsigned int flags) +{ + struct au_hinode *hinode; + struct inode *hi; + struct au_iinfo *iinfo = au_ii(inode); + + IiMustWriteLock(inode); + + hinode = au_hinode(iinfo, bindex); + hi = hinode->hi_inode; + AuDebugOn(h_inode && atomic_read(&h_inode->i_count) <= 0); + + if (hi) + au_hiput(hinode); + hinode->hi_inode = h_inode; + if (h_inode) { + int err; + struct super_block *sb = inode->i_sb; + struct au_branch *br; + + AuDebugOn(inode->i_mode + && (h_inode->i_mode & S_IFMT) + != (inode->i_mode & S_IFMT)); + if (bindex == iinfo->ii_btop) + au_cpup_igen(inode, h_inode); + br = au_sbr(sb, bindex); + hinode->hi_id = br->br_id; + if (au_ftest_hi(flags, XINO)) { + err = au_xino_write(sb, bindex, h_inode->i_ino, + inode->i_ino); + if (unlikely(err)) + AuIOErr1("failed au_xino_write() %d\n", err); + } + + if (au_ftest_hi(flags, HNOTIFY) + && au_br_hnotifyable(br->br_perm)) { + err = au_hn_alloc(hinode, inode); + if (unlikely(err)) + AuIOErr1("au_hn_alloc() %d\n", err); + } + } +} + +void au_set_hi_wh(struct inode *inode, aufs_bindex_t bindex, + struct dentry *h_wh) +{ + struct au_hinode *hinode; + + IiMustWriteLock(inode); + + hinode = au_hinode(au_ii(inode), bindex); + AuDebugOn(hinode->hi_whdentry); + hinode->hi_whdentry = h_wh; +} + +void au_update_iigen(struct inode *inode, int half) +{ + struct au_iinfo *iinfo; + struct au_iigen *iigen; + unsigned int sigen; + + sigen = au_sigen(inode->i_sb); + iinfo = au_ii(inode); + iigen = &iinfo->ii_generation; + spin_lock(&iigen->ig_spin); + iigen->ig_generation = sigen; + if (half) + au_ig_fset(iigen->ig_flags, HALF_REFRESHED); + else + au_ig_fclr(iigen->ig_flags, HALF_REFRESHED); + spin_unlock(&iigen->ig_spin); +} + +/* it may be called at remount time, too */ +void au_update_ibrange(struct inode *inode, int do_put_zero) +{ + struct au_iinfo *iinfo; + aufs_bindex_t bindex, bbot; + + AuDebugOn(au_is_bad_inode(inode)); + IiMustWriteLock(inode); + + iinfo = au_ii(inode); + if (do_put_zero && iinfo->ii_btop >= 0) { + for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot; + bindex++) { + struct inode *h_i; + + h_i = au_hinode(iinfo, bindex)->hi_inode; + if (h_i + && !h_i->i_nlink + && !(h_i->i_state & I_LINKABLE)) + au_set_h_iptr(inode, bindex, NULL, 0); + } + } + + iinfo->ii_btop = -1; + iinfo->ii_bbot = -1; + bbot = au_sbbot(inode->i_sb); + for (bindex = 0; bindex <= bbot; bindex++) + if (au_hinode(iinfo, bindex)->hi_inode) { + iinfo->ii_btop = bindex; + break; + } + if (iinfo->ii_btop >= 0) + for (bindex = bbot; bindex >= iinfo->ii_btop; bindex--) + if (au_hinode(iinfo, bindex)->hi_inode) { + iinfo->ii_bbot = bindex; + break; + } + AuDebugOn(iinfo->ii_btop > iinfo->ii_bbot); +} + +/* ---------------------------------------------------------------------- */ + +void au_icntnr_init_once(void *_c) +{ + struct au_icntnr *c = _c; + struct au_iinfo *iinfo = &c->iinfo; + + spin_lock_init(&iinfo->ii_generation.ig_spin); + au_rw_init(&iinfo->ii_rwsem); + inode_init_once(&c->vfs_inode); +} + +void au_hinode_init(struct au_hinode *hinode) +{ + hinode->hi_inode = NULL; + hinode->hi_id = -1; + au_hn_init(hinode); + hinode->hi_whdentry = NULL; +} + +int au_iinfo_init(struct inode *inode) +{ + struct au_iinfo *iinfo; + struct super_block *sb; + struct au_hinode *hi; + int nbr, i; + + sb = inode->i_sb; + iinfo = &(container_of(inode, struct au_icntnr, vfs_inode)->iinfo); + nbr = au_sbbot(sb) + 1; + if (unlikely(nbr <= 0)) + nbr = 1; + hi = kmalloc_array(nbr, sizeof(*iinfo->ii_hinode), GFP_NOFS); + if (hi) { + au_lcnt_inc(&au_sbi(sb)->si_ninodes); + + iinfo->ii_hinode = hi; + for (i = 0; i < nbr; i++, hi++) + au_hinode_init(hi); + + iinfo->ii_generation.ig_generation = au_sigen(sb); + iinfo->ii_btop = -1; + iinfo->ii_bbot = -1; + iinfo->ii_vdir = NULL; + return 0; + } + return -ENOMEM; +} + +int au_hinode_realloc(struct au_iinfo *iinfo, int nbr, int may_shrink) +{ + int err, i; + struct au_hinode *hip; + + AuRwMustWriteLock(&iinfo->ii_rwsem); + + err = -ENOMEM; + hip = au_krealloc(iinfo->ii_hinode, sizeof(*hip) * nbr, GFP_NOFS, + may_shrink); + if (hip) { + iinfo->ii_hinode = hip; + i = iinfo->ii_bbot + 1; + hip += i; + for (; i < nbr; i++, hip++) + au_hinode_init(hip); + err = 0; + } + + return err; +} + +void au_iinfo_fin(struct inode *inode) +{ + struct au_iinfo *iinfo; + struct au_hinode *hi; + struct super_block *sb; + aufs_bindex_t bindex, bbot; + const unsigned char unlinked = !inode->i_nlink; + + AuDebugOn(au_is_bad_inode(inode)); + + sb = inode->i_sb; + au_lcnt_dec(&au_sbi(sb)->si_ninodes); + if (si_pid_test(sb)) + au_xino_delete_inode(inode, unlinked); + else { + /* + * it is safe to hide the dependency between sbinfo and + * sb->s_umount. + */ + lockdep_off(); + si_noflush_read_lock(sb); + au_xino_delete_inode(inode, unlinked); + si_read_unlock(sb); + lockdep_on(); + } + + iinfo = au_ii(inode); + if (iinfo->ii_vdir) + au_vdir_free(iinfo->ii_vdir); + + bindex = iinfo->ii_btop; + if (bindex >= 0) { + hi = au_hinode(iinfo, bindex); + bbot = iinfo->ii_bbot; + while (bindex++ <= bbot) { + if (hi->hi_inode) + au_hiput(hi); + hi++; + } + } + au_kfree_rcu(iinfo->ii_hinode); + AuRwDestroy(&iinfo->ii_rwsem); +} diff --git a/fs/aufs/inode.c b/fs/aufs/inode.c new file mode 100644 index 00000000000..e1b513aaa58 --- /dev/null +++ b/fs/aufs/inode.c @@ -0,0 +1,531 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode functions + */ + +#include +#include "aufs.h" + +struct inode *au_igrab(struct inode *inode) +{ + if (inode) { + AuDebugOn(!atomic_read(&inode->i_count)); + ihold(inode); + } + return inode; +} + +static void au_refresh_hinode_attr(struct inode *inode, int do_version) +{ + au_cpup_attr_all(inode, /*force*/0); + au_update_iigen(inode, /*half*/1); + if (do_version) + inode_inc_iversion(inode); +} + +static int au_ii_refresh(struct inode *inode, int *update) +{ + int err, e, nbr; + umode_t type; + aufs_bindex_t bindex, new_bindex; + struct super_block *sb; + struct au_iinfo *iinfo; + struct au_hinode *p, *q, tmp; + + AuDebugOn(au_is_bad_inode(inode)); + IiMustWriteLock(inode); + + *update = 0; + sb = inode->i_sb; + nbr = au_sbbot(sb) + 1; + type = inode->i_mode & S_IFMT; + iinfo = au_ii(inode); + err = au_hinode_realloc(iinfo, nbr, /*may_shrink*/0); + if (unlikely(err)) + goto out; + + AuDebugOn(iinfo->ii_btop < 0); + p = au_hinode(iinfo, iinfo->ii_btop); + for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot; + bindex++, p++) { + if (!p->hi_inode) + continue; + + AuDebugOn(type != (p->hi_inode->i_mode & S_IFMT)); + new_bindex = au_br_index(sb, p->hi_id); + if (new_bindex == bindex) + continue; + + if (new_bindex < 0) { + *update = 1; + au_hiput(p); + p->hi_inode = NULL; + continue; + } + + if (new_bindex < iinfo->ii_btop) + iinfo->ii_btop = new_bindex; + if (iinfo->ii_bbot < new_bindex) + iinfo->ii_bbot = new_bindex; + /* swap two lower inode, and loop again */ + q = au_hinode(iinfo, new_bindex); + tmp = *q; + *q = *p; + *p = tmp; + if (tmp.hi_inode) { + bindex--; + p--; + } + } + au_update_ibrange(inode, /*do_put_zero*/0); + au_hinode_realloc(iinfo, nbr, /*may_shrink*/1); /* harmless if err */ + e = au_dy_irefresh(inode); + if (unlikely(e && !err)) + err = e; + +out: + AuTraceErr(err); + return err; +} + +void au_refresh_iop(struct inode *inode, int force_getattr) +{ + int type; + struct au_sbinfo *sbi = au_sbi(inode->i_sb); + const struct inode_operations *iop + = force_getattr ? aufs_iop : sbi->si_iop_array; + + if (inode->i_op == iop) + return; + + switch (inode->i_mode & S_IFMT) { + case S_IFDIR: + type = AuIop_DIR; + break; + case S_IFLNK: + type = AuIop_SYMLINK; + break; + default: + type = AuIop_OTHER; + break; + } + + inode->i_op = iop + type; + /* unnecessary smp_wmb() */ +} + +int au_refresh_hinode_self(struct inode *inode) +{ + int err, update; + + err = au_ii_refresh(inode, &update); + if (!err) + au_refresh_hinode_attr(inode, update && S_ISDIR(inode->i_mode)); + + AuTraceErr(err); + return err; +} + +int au_refresh_hinode(struct inode *inode, struct dentry *dentry) +{ + int err, e, update; + unsigned int flags; + umode_t mode; + aufs_bindex_t bindex, bbot; + unsigned char isdir; + struct au_hinode *p; + struct au_iinfo *iinfo; + + err = au_ii_refresh(inode, &update); + if (unlikely(err)) + goto out; + + update = 0; + iinfo = au_ii(inode); + p = au_hinode(iinfo, iinfo->ii_btop); + mode = (inode->i_mode & S_IFMT); + isdir = S_ISDIR(mode); + flags = au_hi_flags(inode, isdir); + bbot = au_dbbot(dentry); + for (bindex = au_dbtop(dentry); bindex <= bbot; bindex++) { + struct inode *h_i, *h_inode; + struct dentry *h_d; + + h_d = au_h_dptr(dentry, bindex); + if (!h_d || d_is_negative(h_d)) + continue; + + h_inode = d_inode(h_d); + AuDebugOn(mode != (h_inode->i_mode & S_IFMT)); + if (iinfo->ii_btop <= bindex && bindex <= iinfo->ii_bbot) { + h_i = au_h_iptr(inode, bindex); + if (h_i) { + if (h_i == h_inode) + continue; + err = -EIO; + break; + } + } + if (bindex < iinfo->ii_btop) + iinfo->ii_btop = bindex; + if (iinfo->ii_bbot < bindex) + iinfo->ii_bbot = bindex; + au_set_h_iptr(inode, bindex, au_igrab(h_inode), flags); + update = 1; + } + au_update_ibrange(inode, /*do_put_zero*/0); + e = au_dy_irefresh(inode); + if (unlikely(e && !err)) + err = e; + if (!err) + au_refresh_hinode_attr(inode, update && isdir); + +out: + AuTraceErr(err); + return err; +} + +static int set_inode(struct inode *inode, struct dentry *dentry) +{ + int err; + unsigned int flags; + umode_t mode; + aufs_bindex_t bindex, btop, btail; + unsigned char isdir; + struct dentry *h_dentry; + struct inode *h_inode; + struct au_iinfo *iinfo; + const struct inode_operations *iop; + + IiMustWriteLock(inode); + + err = 0; + isdir = 0; + iop = au_sbi(inode->i_sb)->si_iop_array; + btop = au_dbtop(dentry); + h_dentry = au_h_dptr(dentry, btop); + h_inode = d_inode(h_dentry); + mode = h_inode->i_mode; + switch (mode & S_IFMT) { + case S_IFREG: + btail = au_dbtail(dentry); + inode->i_op = iop + AuIop_OTHER; + inode->i_fop = &aufs_file_fop; + err = au_dy_iaop(inode, btop, h_inode); + if (unlikely(err)) + goto out; + break; + case S_IFDIR: + isdir = 1; + btail = au_dbtaildir(dentry); + inode->i_op = iop + AuIop_DIR; + inode->i_fop = &aufs_dir_fop; + break; + case S_IFLNK: + btail = au_dbtail(dentry); + inode->i_op = iop + AuIop_SYMLINK; + break; + case S_IFBLK: + case S_IFCHR: + case S_IFIFO: + case S_IFSOCK: + btail = au_dbtail(dentry); + inode->i_op = iop + AuIop_OTHER; + init_special_inode(inode, mode, h_inode->i_rdev); + break; + default: + AuIOErr("Unknown file type 0%o\n", mode); + err = -EIO; + goto out; + } + + /* do not set hnotify for whiteouted dirs (SHWH mode) */ + flags = au_hi_flags(inode, isdir); + if (au_opt_test(au_mntflags(dentry->d_sb), SHWH) + && au_ftest_hi(flags, HNOTIFY) + && dentry->d_name.len > AUFS_WH_PFX_LEN + && !memcmp(dentry->d_name.name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) + au_fclr_hi(flags, HNOTIFY); + iinfo = au_ii(inode); + iinfo->ii_btop = btop; + iinfo->ii_bbot = btail; + for (bindex = btop; bindex <= btail; bindex++) { + h_dentry = au_h_dptr(dentry, bindex); + if (h_dentry) + au_set_h_iptr(inode, bindex, + au_igrab(d_inode(h_dentry)), flags); + } + au_cpup_attr_all(inode, /*force*/1); + /* + * to force calling aufs_get_acl() every time, + * do not call cache_no_acl() for aufs inode. + */ + +out: + return err; +} + +/* + * successful returns with iinfo write_locked + * minus: errno + * zero: success, matched + * plus: no error, but unmatched + */ +static int reval_inode(struct inode *inode, struct dentry *dentry) +{ + int err; + unsigned int gen, igflags; + aufs_bindex_t bindex, bbot; + struct inode *h_inode, *h_dinode; + struct dentry *h_dentry; + + /* + * before this function, if aufs got any iinfo lock, it must be only + * one, the parent dir. + * it can happen by UDBA and the obsoleted inode number. + */ + err = -EIO; + if (unlikely(inode->i_ino == parent_ino(dentry))) + goto out; + + err = 1; + ii_write_lock_new_child(inode); + h_dentry = au_h_dptr(dentry, au_dbtop(dentry)); + h_dinode = d_inode(h_dentry); + bbot = au_ibbot(inode); + for (bindex = au_ibtop(inode); bindex <= bbot; bindex++) { + h_inode = au_h_iptr(inode, bindex); + if (!h_inode || h_inode != h_dinode) + continue; + + err = 0; + gen = au_iigen(inode, &igflags); + if (gen == au_digen(dentry) + && !au_ig_ftest(igflags, HALF_REFRESHED)) + break; + + /* fully refresh inode using dentry */ + err = au_refresh_hinode(inode, dentry); + if (!err) + au_update_iigen(inode, /*half*/0); + break; + } + + if (unlikely(err)) + ii_write_unlock(inode); +out: + return err; +} + +int au_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + unsigned int d_type, ino_t *ino) +{ + int err, idx; + const int isnondir = d_type != DT_DIR; + + /* prevent hardlinked inode number from race condition */ + if (isnondir) { + err = au_xinondir_enter(sb, bindex, h_ino, &idx); + if (unlikely(err)) + goto out; + } + + err = au_xino_read(sb, bindex, h_ino, ino); + if (unlikely(err)) + goto out_xinondir; + + if (!*ino) { + err = -EIO; + *ino = au_xino_new_ino(sb); + if (unlikely(!*ino)) + goto out_xinondir; + err = au_xino_write(sb, bindex, h_ino, *ino); + if (unlikely(err)) + goto out_xinondir; + } + +out_xinondir: + if (isnondir && idx >= 0) + au_xinondir_leave(sb, bindex, h_ino, idx); +out: + return err; +} + +/* successful returns with iinfo write_locked */ +/* todo: return with unlocked? */ +struct inode *au_new_inode(struct dentry *dentry, int must_new) +{ + struct inode *inode, *h_inode; + struct dentry *h_dentry; + struct super_block *sb; + ino_t h_ino, ino; + int err, idx, hlinked; + aufs_bindex_t btop; + + sb = dentry->d_sb; + btop = au_dbtop(dentry); + h_dentry = au_h_dptr(dentry, btop); + h_inode = d_inode(h_dentry); + h_ino = h_inode->i_ino; + hlinked = !d_is_dir(h_dentry) && h_inode->i_nlink > 1; + +new_ino: + /* + * stop 'race'-ing between hardlinks under different + * parents. + */ + if (hlinked) { + err = au_xinondir_enter(sb, btop, h_ino, &idx); + inode = ERR_PTR(err); + if (unlikely(err)) + goto out; + } + + err = au_xino_read(sb, btop, h_ino, &ino); + inode = ERR_PTR(err); + if (unlikely(err)) + goto out_xinondir; + + if (!ino) { + ino = au_xino_new_ino(sb); + if (unlikely(!ino)) { + inode = ERR_PTR(-EIO); + goto out_xinondir; + } + } + + AuDbg("i%lu\n", (unsigned long)ino); + inode = au_iget_locked(sb, ino); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out_xinondir; + + AuDbg("%lx, new %d\n", inode->i_state, !!(inode->i_state & I_NEW)); + if (inode->i_state & I_NEW) { + ii_write_lock_new_child(inode); + err = set_inode(inode, dentry); + if (!err) { + unlock_new_inode(inode); + goto out_xinondir; /* success */ + } + + /* + * iget_failed() calls iput(), but we need to call + * ii_write_unlock() after iget_failed(). so dirty hack for + * i_count. + */ + atomic_inc(&inode->i_count); + iget_failed(inode); + ii_write_unlock(inode); + au_xino_write(sb, btop, h_ino, /*ino*/0); + /* ignore this error */ + goto out_iput; + } else if (!must_new && !IS_DEADDIR(inode) && inode->i_nlink) { + /* + * horrible race condition between lookup, readdir and copyup + * (or something). + */ + if (hlinked && idx >= 0) + au_xinondir_leave(sb, btop, h_ino, idx); + err = reval_inode(inode, dentry); + if (unlikely(err < 0)) { + hlinked = 0; + goto out_iput; + } + if (!err) + goto out; /* success */ + else if (hlinked && idx >= 0) { + err = au_xinondir_enter(sb, btop, h_ino, &idx); + if (unlikely(err)) { + iput(inode); + inode = ERR_PTR(err); + goto out; + } + } + } + + if (unlikely(au_test_fs_unique_ino(h_inode))) + AuWarn1("Warning: Un-notified UDBA or repeatedly renamed dir," + " b%d, %s, %pd, hi%lu, i%lu.\n", + btop, au_sbtype(h_dentry->d_sb), dentry, + (unsigned long)h_ino, (unsigned long)ino); + ino = 0; + err = au_xino_write(sb, btop, h_ino, /*ino*/0); + if (!err) { + iput(inode); + if (hlinked && idx >= 0) + au_xinondir_leave(sb, btop, h_ino, idx); + goto new_ino; + } + +out_iput: + iput(inode); + inode = ERR_PTR(err); +out_xinondir: + if (hlinked && idx >= 0) + au_xinondir_leave(sb, btop, h_ino, idx); +out: + return inode; +} + +/* ---------------------------------------------------------------------- */ + +int au_test_ro(struct super_block *sb, aufs_bindex_t bindex, + struct inode *inode) +{ + int err; + struct inode *hi; + + err = au_br_rdonly(au_sbr(sb, bindex)); + + /* pseudo-link after flushed may happen out of bounds */ + if (!err + && inode + && au_ibtop(inode) <= bindex + && bindex <= au_ibbot(inode)) { + /* + * permission check is unnecessary since vfsub routine + * will be called later + */ + hi = au_h_iptr(inode, bindex); + if (hi) + err = IS_IMMUTABLE(hi) ? -EROFS : 0; + } + + return err; +} + +int au_test_h_perm(struct user_namespace *h_userns, struct inode *h_inode, + int mask) +{ + if (uid_eq(current_fsuid(), GLOBAL_ROOT_UID)) + return 0; + return inode_permission(h_userns, h_inode, mask); +} + +int au_test_h_perm_sio(struct user_namespace *h_userns, struct inode *h_inode, + int mask) +{ + if (au_test_nfs(h_inode->i_sb) + && (mask & MAY_WRITE) + && S_ISDIR(h_inode->i_mode)) + mask |= MAY_READ; /* force permission check */ + return au_test_h_perm(h_userns, h_inode, mask); +} diff --git a/fs/aufs/inode.h b/fs/aufs/inode.h new file mode 100644 index 00000000000..8008824bf2b --- /dev/null +++ b/fs/aufs/inode.h @@ -0,0 +1,705 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * inode operations + */ + +#ifndef __AUFS_INODE_H__ +#define __AUFS_INODE_H__ + +#ifdef __KERNEL__ + +#include +#include "rwsem.h" + +struct vfsmount; + +struct au_hnotify { +#ifdef CONFIG_AUFS_HNOTIFY +#ifdef CONFIG_AUFS_HFSNOTIFY + /* never use fsnotify_add_vfsmount_mark() */ + struct fsnotify_mark hn_mark; +#endif + struct inode *hn_aufs_inode; /* no get/put */ + struct rcu_head rcu; +#endif +} ____cacheline_aligned_in_smp; + +struct au_hinode { + struct inode *hi_inode; + aufs_bindex_t hi_id; +#ifdef CONFIG_AUFS_HNOTIFY + struct au_hnotify *hi_notify; +#endif + + /* reference to the copied-up whiteout with get/put */ + struct dentry *hi_whdentry; +}; + +/* ig_flags */ +#define AuIG_HALF_REFRESHED 1 +#define au_ig_ftest(flags, name) ((flags) & AuIG_##name) +#define au_ig_fset(flags, name) \ + do { (flags) |= AuIG_##name; } while (0) +#define au_ig_fclr(flags, name) \ + do { (flags) &= ~AuIG_##name; } while (0) + +struct au_iigen { + spinlock_t ig_spin; + __u32 ig_generation, ig_flags; +}; + +struct au_vdir; +struct au_iinfo { + struct au_iigen ii_generation; + struct super_block *ii_hsb1; /* no get/put */ + + struct au_rwsem ii_rwsem; + aufs_bindex_t ii_btop, ii_bbot; + __u32 ii_higen; + struct au_hinode *ii_hinode; + struct au_vdir *ii_vdir; +}; + +struct au_icntnr { + struct au_iinfo iinfo; + struct inode vfs_inode; + struct hlist_bl_node plink; + struct rcu_head rcu; +} ____cacheline_aligned_in_smp; + +/* au_pin flags */ +#define AuPin_DI_LOCKED 1 +#define AuPin_MNT_WRITE (1 << 1) +#define au_ftest_pin(flags, name) ((flags) & AuPin_##name) +#define au_fset_pin(flags, name) \ + do { (flags) |= AuPin_##name; } while (0) +#define au_fclr_pin(flags, name) \ + do { (flags) &= ~AuPin_##name; } while (0) + +struct au_pin { + /* input */ + struct dentry *dentry; + unsigned int udba; + unsigned char lsc_di, lsc_hi, flags; + aufs_bindex_t bindex; + + /* output */ + struct dentry *parent; + struct au_hinode *hdir; + struct vfsmount *h_mnt; + + /* temporary unlock/relock for copyup */ + struct dentry *h_dentry, *h_parent; + struct au_branch *br; + struct task_struct *task; +}; + +void au_pin_hdir_unlock(struct au_pin *p); +int au_pin_hdir_lock(struct au_pin *p); +int au_pin_hdir_relock(struct au_pin *p); +void au_pin_hdir_acquire_nest(struct au_pin *p); +void au_pin_hdir_release(struct au_pin *p); + +/* ---------------------------------------------------------------------- */ + +static inline struct au_iinfo *au_ii(struct inode *inode) +{ + BUG_ON(is_bad_inode(inode)); + return &(container_of(inode, struct au_icntnr, vfs_inode)->iinfo); +} + +/* ---------------------------------------------------------------------- */ + +/* inode.c */ +struct inode *au_igrab(struct inode *inode); +void au_refresh_iop(struct inode *inode, int force_getattr); +int au_refresh_hinode_self(struct inode *inode); +int au_refresh_hinode(struct inode *inode, struct dentry *dentry); +int au_ino(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + unsigned int d_type, ino_t *ino); +struct inode *au_new_inode(struct dentry *dentry, int must_new); +int au_test_ro(struct super_block *sb, aufs_bindex_t bindex, + struct inode *inode); +int au_test_h_perm(struct user_namespace *h_userns, struct inode *h_inode, + int mask); +int au_test_h_perm_sio(struct user_namespace *h_userns, struct inode *h_inode, + int mask); + +static inline int au_wh_ino(struct super_block *sb, aufs_bindex_t bindex, + ino_t h_ino, unsigned int d_type, ino_t *ino) +{ +#ifdef CONFIG_AUFS_SHWH + return au_ino(sb, bindex, h_ino, d_type, ino); +#else + return 0; +#endif +} + +/* i_op.c */ +enum { + AuIop_SYMLINK, + AuIop_DIR, + AuIop_OTHER, + AuIop_Last +}; +extern struct inode_operations aufs_iop[AuIop_Last], /* not const */ + aufs_iop_nogetattr[AuIop_Last]; + +/* au_wr_dir flags */ +#define AuWrDir_ADD_ENTRY 1 +#define AuWrDir_ISDIR (1 << 1) +#define AuWrDir_TMPFILE (1 << 2) +#define au_ftest_wrdir(flags, name) ((flags) & AuWrDir_##name) +#define au_fset_wrdir(flags, name) \ + do { (flags) |= AuWrDir_##name; } while (0) +#define au_fclr_wrdir(flags, name) \ + do { (flags) &= ~AuWrDir_##name; } while (0) + +struct au_wr_dir_args { + aufs_bindex_t force_btgt; + unsigned char flags; +}; +int au_wr_dir(struct dentry *dentry, struct dentry *src_dentry, + struct au_wr_dir_args *args); + +struct dentry *au_pinned_h_parent(struct au_pin *pin); +void au_pin_init(struct au_pin *pin, struct dentry *dentry, + aufs_bindex_t bindex, int lsc_di, int lsc_hi, + unsigned int udba, unsigned char flags); +int au_pin(struct au_pin *pin, struct dentry *dentry, aufs_bindex_t bindex, + unsigned int udba, unsigned char flags) __must_check; +int au_do_pin(struct au_pin *pin) __must_check; +void au_unpin(struct au_pin *pin); +int au_reval_for_attr(struct dentry *dentry, unsigned int sigen); + +#define AuIcpup_DID_CPUP 1 +#define au_ftest_icpup(flags, name) ((flags) & AuIcpup_##name) +#define au_fset_icpup(flags, name) \ + do { (flags) |= AuIcpup_##name; } while (0) +#define au_fclr_icpup(flags, name) \ + do { (flags) &= ~AuIcpup_##name; } while (0) + +struct au_icpup_args { + unsigned char flags; + unsigned char pin_flags; + aufs_bindex_t btgt; + unsigned int udba; + struct au_pin pin; + struct path h_path; + struct inode *h_inode; +}; + +int au_pin_and_icpup(struct dentry *dentry, struct iattr *ia, + struct au_icpup_args *a); + +int au_h_path_getattr(struct dentry *dentry, struct inode *inode, int force, + struct path *h_path, int locked); + +/* i_op_add.c */ +int au_may_add(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent, int isdir); +int aufs_mknod(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode, dev_t dev); +int aufs_symlink(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, const char *symname); +int aufs_create(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode, bool want_excl); +struct vfsub_aopen_args; +int au_aopen_or_create(struct inode *dir, struct dentry *dentry, + struct vfsub_aopen_args *args); +int aufs_tmpfile(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode); +int aufs_link(struct dentry *src_dentry, struct inode *dir, + struct dentry *dentry); +int aufs_mkdir(struct user_namespace *userns, struct inode *dir, + struct dentry *dentry, umode_t mode); + +/* i_op_del.c */ +int au_wr_dir_need_wh(struct dentry *dentry, int isdir, aufs_bindex_t *bcpup); +int au_may_del(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent, int isdir); +int aufs_unlink(struct inode *dir, struct dentry *dentry); +int aufs_rmdir(struct inode *dir, struct dentry *dentry); + +/* i_op_ren.c */ +int au_wbr(struct dentry *dentry, aufs_bindex_t btgt); +int aufs_rename(struct user_namespace *userns, + struct inode *_src_dir, struct dentry *_src_dentry, + struct inode *_dst_dir, struct dentry *_dst_dentry, + unsigned int _flags); + +/* iinfo.c */ +struct inode *au_h_iptr(struct inode *inode, aufs_bindex_t bindex); +void au_hiput(struct au_hinode *hinode); +void au_set_hi_wh(struct inode *inode, aufs_bindex_t bindex, + struct dentry *h_wh); +unsigned int au_hi_flags(struct inode *inode, int isdir); + +/* hinode flags */ +#define AuHi_XINO 1 +#define AuHi_HNOTIFY (1 << 1) +#define au_ftest_hi(flags, name) ((flags) & AuHi_##name) +#define au_fset_hi(flags, name) \ + do { (flags) |= AuHi_##name; } while (0) +#define au_fclr_hi(flags, name) \ + do { (flags) &= ~AuHi_##name; } while (0) + +#ifndef CONFIG_AUFS_HNOTIFY +#undef AuHi_HNOTIFY +#define AuHi_HNOTIFY 0 +#endif + +void au_set_h_iptr(struct inode *inode, aufs_bindex_t bindex, + struct inode *h_inode, unsigned int flags); + +void au_update_iigen(struct inode *inode, int half); +void au_update_ibrange(struct inode *inode, int do_put_zero); + +void au_icntnr_init_once(void *_c); +void au_hinode_init(struct au_hinode *hinode); +int au_iinfo_init(struct inode *inode); +void au_iinfo_fin(struct inode *inode); +int au_hinode_realloc(struct au_iinfo *iinfo, int nbr, int may_shrink); + +#ifdef CONFIG_PROC_FS +/* plink.c */ +int au_plink_maint(struct super_block *sb, int flags); +struct au_sbinfo; +void au_plink_maint_leave(struct au_sbinfo *sbinfo); +int au_plink_maint_enter(struct super_block *sb); +#ifdef CONFIG_AUFS_DEBUG +void au_plink_list(struct super_block *sb); +#else +AuStubVoid(au_plink_list, struct super_block *sb) +#endif +int au_plink_test(struct inode *inode); +struct dentry *au_plink_lkup(struct inode *inode, aufs_bindex_t bindex); +void au_plink_append(struct inode *inode, aufs_bindex_t bindex, + struct dentry *h_dentry); +void au_plink_put(struct super_block *sb, int verbose); +void au_plink_clean(struct super_block *sb, int verbose); +void au_plink_half_refresh(struct super_block *sb, aufs_bindex_t br_id); +#else +AuStubInt0(au_plink_maint, struct super_block *sb, int flags); +AuStubVoid(au_plink_maint_leave, struct au_sbinfo *sbinfo); +AuStubInt0(au_plink_maint_enter, struct super_block *sb); +AuStubVoid(au_plink_list, struct super_block *sb); +AuStubInt0(au_plink_test, struct inode *inode); +AuStub(struct dentry *, au_plink_lkup, return NULL, + struct inode *inode, aufs_bindex_t bindex); +AuStubVoid(au_plink_append, struct inode *inode, aufs_bindex_t bindex, + struct dentry *h_dentry); +AuStubVoid(au_plink_put, struct super_block *sb, int verbose); +AuStubVoid(au_plink_clean, struct super_block *sb, int verbose); +AuStubVoid(au_plink_half_refresh, struct super_block *sb, aufs_bindex_t br_id); +#endif /* CONFIG_PROC_FS */ + +#ifdef CONFIG_AUFS_XATTR +/* xattr.c */ +int au_cpup_xattr(struct path *h_dst, struct path *h_src, int ignore_flags, + unsigned int verbose); +ssize_t aufs_listxattr(struct dentry *dentry, char *list, size_t size); +void au_xattr_init(struct super_block *sb); +#else +AuStubInt0(au_cpup_xattr, struct path *h_dst, struct path *h_src, + int ignore_flags, unsigned int verbose); +AuStubVoid(au_xattr_init, struct super_block *sb); +#endif + +#ifdef CONFIG_FS_POSIX_ACL +struct posix_acl *aufs_get_acl(struct inode *inode, int type); +int aufs_set_acl(struct user_namespace *userns, struct inode *inode, + struct posix_acl *acl, int type); +#endif + +#if IS_ENABLED(CONFIG_AUFS_XATTR) || IS_ENABLED(CONFIG_FS_POSIX_ACL) +enum { + AU_XATTR_SET, + AU_ACL_SET +}; + +struct au_sxattr { + int type; + union { + struct { + const char *name; + const void *value; + size_t size; + int flags; + } set; + struct { + struct posix_acl *acl; + int type; + } acl_set; + } u; +}; +ssize_t au_sxattr(struct dentry *dentry, struct inode *inode, + struct au_sxattr *arg); +#endif + +/* ---------------------------------------------------------------------- */ + +/* lock subclass for iinfo */ +enum { + AuLsc_II_CHILD, /* child first */ + AuLsc_II_CHILD2, /* rename(2), link(2), and cpup at hnotify */ + AuLsc_II_CHILD3, /* copyup dirs */ + AuLsc_II_PARENT, /* see AuLsc_I_PARENT in vfsub.h */ + AuLsc_II_PARENT2, + AuLsc_II_PARENT3, /* copyup dirs */ + AuLsc_II_NEW_CHILD +}; + +/* + * ii_read_lock_child, ii_write_lock_child, + * ii_read_lock_child2, ii_write_lock_child2, + * ii_read_lock_child3, ii_write_lock_child3, + * ii_read_lock_parent, ii_write_lock_parent, + * ii_read_lock_parent2, ii_write_lock_parent2, + * ii_read_lock_parent3, ii_write_lock_parent3, + * ii_read_lock_new_child, ii_write_lock_new_child, + */ +#define AuReadLockFunc(name, lsc) \ +static inline void ii_read_lock_##name(struct inode *i) \ +{ \ + au_rw_read_lock_nested(&au_ii(i)->ii_rwsem, AuLsc_II_##lsc); \ +} + +#define AuWriteLockFunc(name, lsc) \ +static inline void ii_write_lock_##name(struct inode *i) \ +{ \ + au_rw_write_lock_nested(&au_ii(i)->ii_rwsem, AuLsc_II_##lsc); \ +} + +#define AuRWLockFuncs(name, lsc) \ + AuReadLockFunc(name, lsc) \ + AuWriteLockFunc(name, lsc) + +AuRWLockFuncs(child, CHILD); +AuRWLockFuncs(child2, CHILD2); +AuRWLockFuncs(child3, CHILD3); +AuRWLockFuncs(parent, PARENT); +AuRWLockFuncs(parent2, PARENT2); +AuRWLockFuncs(parent3, PARENT3); +AuRWLockFuncs(new_child, NEW_CHILD); + +#undef AuReadLockFunc +#undef AuWriteLockFunc +#undef AuRWLockFuncs + +#define ii_read_unlock(i) au_rw_read_unlock(&au_ii(i)->ii_rwsem) +#define ii_write_unlock(i) au_rw_write_unlock(&au_ii(i)->ii_rwsem) +#define ii_downgrade_lock(i) au_rw_dgrade_lock(&au_ii(i)->ii_rwsem) + +#define IiMustNoWaiters(i) AuRwMustNoWaiters(&au_ii(i)->ii_rwsem) +#define IiMustAnyLock(i) AuRwMustAnyLock(&au_ii(i)->ii_rwsem) +#define IiMustWriteLock(i) AuRwMustWriteLock(&au_ii(i)->ii_rwsem) + +/* ---------------------------------------------------------------------- */ + +static inline void au_icntnr_init(struct au_icntnr *c) +{ +#ifdef CONFIG_AUFS_DEBUG + c->vfs_inode.i_mode = 0; +#endif +} + +static inline unsigned int au_iigen(struct inode *inode, unsigned int *igflags) +{ + unsigned int gen; + struct au_iinfo *iinfo; + struct au_iigen *iigen; + + iinfo = au_ii(inode); + iigen = &iinfo->ii_generation; + spin_lock(&iigen->ig_spin); + if (igflags) + *igflags = iigen->ig_flags; + gen = iigen->ig_generation; + spin_unlock(&iigen->ig_spin); + + return gen; +} + +/* tiny test for inode number */ +/* tmpfs generation is too rough */ +static inline int au_test_higen(struct inode *inode, struct inode *h_inode) +{ + struct au_iinfo *iinfo; + + iinfo = au_ii(inode); + AuRwMustAnyLock(&iinfo->ii_rwsem); + return !(iinfo->ii_hsb1 == h_inode->i_sb + && iinfo->ii_higen == h_inode->i_generation); +} + +static inline void au_iigen_dec(struct inode *inode) +{ + struct au_iinfo *iinfo; + struct au_iigen *iigen; + + iinfo = au_ii(inode); + iigen = &iinfo->ii_generation; + spin_lock(&iigen->ig_spin); + iigen->ig_generation--; + spin_unlock(&iigen->ig_spin); +} + +static inline int au_iigen_test(struct inode *inode, unsigned int sigen) +{ + int err; + + err = 0; + if (unlikely(inode && au_iigen(inode, NULL) != sigen)) + err = -EIO; + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static inline struct au_hinode *au_hinode(struct au_iinfo *iinfo, + aufs_bindex_t bindex) +{ + return iinfo->ii_hinode + bindex; +} + +static inline int au_is_bad_inode(struct inode *inode) +{ + return !!(is_bad_inode(inode) || !au_hinode(au_ii(inode), 0)); +} + +static inline aufs_bindex_t au_ii_br_id(struct inode *inode, + aufs_bindex_t bindex) +{ + IiMustAnyLock(inode); + return au_hinode(au_ii(inode), bindex)->hi_id; +} + +static inline aufs_bindex_t au_ibtop(struct inode *inode) +{ + IiMustAnyLock(inode); + return au_ii(inode)->ii_btop; +} + +static inline aufs_bindex_t au_ibbot(struct inode *inode) +{ + IiMustAnyLock(inode); + return au_ii(inode)->ii_bbot; +} + +static inline struct au_vdir *au_ivdir(struct inode *inode) +{ + IiMustAnyLock(inode); + return au_ii(inode)->ii_vdir; +} + +static inline struct dentry *au_hi_wh(struct inode *inode, aufs_bindex_t bindex) +{ + IiMustAnyLock(inode); + return au_hinode(au_ii(inode), bindex)->hi_whdentry; +} + +static inline void au_set_ibtop(struct inode *inode, aufs_bindex_t bindex) +{ + IiMustWriteLock(inode); + au_ii(inode)->ii_btop = bindex; +} + +static inline void au_set_ibbot(struct inode *inode, aufs_bindex_t bindex) +{ + IiMustWriteLock(inode); + au_ii(inode)->ii_bbot = bindex; +} + +static inline void au_set_ivdir(struct inode *inode, struct au_vdir *vdir) +{ + IiMustWriteLock(inode); + au_ii(inode)->ii_vdir = vdir; +} + +static inline struct au_hinode *au_hi(struct inode *inode, aufs_bindex_t bindex) +{ + IiMustAnyLock(inode); + return au_hinode(au_ii(inode), bindex); +} + +/* ---------------------------------------------------------------------- */ + +static inline struct dentry *au_pinned_parent(struct au_pin *pin) +{ + if (pin) + return pin->parent; + return NULL; +} + +static inline struct inode *au_pinned_h_dir(struct au_pin *pin) +{ + if (pin && pin->hdir) + return pin->hdir->hi_inode; + return NULL; +} + +static inline struct au_hinode *au_pinned_hdir(struct au_pin *pin) +{ + if (pin) + return pin->hdir; + return NULL; +} + +static inline void au_pin_set_dentry(struct au_pin *pin, struct dentry *dentry) +{ + if (pin) + pin->dentry = dentry; +} + +static inline void au_pin_set_parent_lflag(struct au_pin *pin, + unsigned char lflag) +{ + if (pin) { + if (lflag) + au_fset_pin(pin->flags, DI_LOCKED); + else + au_fclr_pin(pin->flags, DI_LOCKED); + } +} + +#if 0 /* reserved */ +static inline void au_pin_set_parent(struct au_pin *pin, struct dentry *parent) +{ + if (pin) { + dput(pin->parent); + pin->parent = dget(parent); + } +} +#endif + +/* ---------------------------------------------------------------------- */ + +struct au_branch; +#ifdef CONFIG_AUFS_HNOTIFY +struct au_hnotify_op { + void (*ctl)(struct au_hinode *hinode, int do_set); + int (*alloc)(struct au_hinode *hinode); + + /* + * if it returns true, the caller should free hinode->hi_notify, + * otherwise ->free() frees it. + */ + int (*free)(struct au_hinode *hinode, + struct au_hnotify *hn) __must_check; + + void (*fin)(void); + int (*init)(void); + + int (*reset_br)(unsigned int udba, struct au_branch *br, int perm); + void (*fin_br)(struct au_branch *br); + int (*init_br)(struct au_branch *br, int perm); +}; + +/* hnotify.c */ +int au_hn_alloc(struct au_hinode *hinode, struct inode *inode); +void au_hn_free(struct au_hinode *hinode); +void au_hn_ctl(struct au_hinode *hinode, int do_set); +void au_hn_reset(struct inode *inode, unsigned int flags); +int au_hnotify(struct inode *h_dir, struct au_hnotify *hnotify, u32 mask, + const struct qstr *h_child_qstr, struct inode *h_child_inode); +int au_hnotify_reset_br(unsigned int udba, struct au_branch *br, int perm); +int au_hnotify_init_br(struct au_branch *br, int perm); +void au_hnotify_fin_br(struct au_branch *br); +int __init au_hnotify_init(void); +void au_hnotify_fin(void); + +/* hfsnotify.c */ +extern const struct au_hnotify_op au_hnotify_op; + +static inline +void au_hn_init(struct au_hinode *hinode) +{ + hinode->hi_notify = NULL; +} + +static inline struct au_hnotify *au_hn(struct au_hinode *hinode) +{ + return hinode->hi_notify; +} + +#else +AuStub(int, au_hn_alloc, return -EOPNOTSUPP, + struct au_hinode *hinode __maybe_unused, + struct inode *inode __maybe_unused) +AuStub(struct au_hnotify *, au_hn, return NULL, struct au_hinode *hinode) +AuStubVoid(au_hn_free, struct au_hinode *hinode __maybe_unused) +AuStubVoid(au_hn_ctl, struct au_hinode *hinode __maybe_unused, + int do_set __maybe_unused) +AuStubVoid(au_hn_reset, struct inode *inode __maybe_unused, + unsigned int flags __maybe_unused) +AuStubInt0(au_hnotify_reset_br, unsigned int udba __maybe_unused, + struct au_branch *br __maybe_unused, + int perm __maybe_unused) +AuStubInt0(au_hnotify_init_br, struct au_branch *br __maybe_unused, + int perm __maybe_unused) +AuStubVoid(au_hnotify_fin_br, struct au_branch *br __maybe_unused) +AuStubInt0(__init au_hnotify_init, void) +AuStubVoid(au_hnotify_fin, void) +AuStubVoid(au_hn_init, struct au_hinode *hinode __maybe_unused) +#endif /* CONFIG_AUFS_HNOTIFY */ + +static inline void au_hn_suspend(struct au_hinode *hdir) +{ + au_hn_ctl(hdir, /*do_set*/0); +} + +static inline void au_hn_resume(struct au_hinode *hdir) +{ + au_hn_ctl(hdir, /*do_set*/1); +} + +static inline void au_hn_inode_lock(struct au_hinode *hdir) +{ + inode_lock(hdir->hi_inode); + au_hn_suspend(hdir); +} + +static inline void au_hn_inode_lock_nested(struct au_hinode *hdir, + unsigned int sc __maybe_unused) +{ + inode_lock_nested(hdir->hi_inode, sc); + au_hn_suspend(hdir); +} + +#if 0 /* unused */ +#include "vfsub.h" +static inline void au_hn_inode_lock_shared_nested(struct au_hinode *hdir, + unsigned int sc) +{ + inode_lock_shared_nested(hdir->hi_inode, sc); + au_hn_suspend(hdir); +} +#endif + +static inline void au_hn_inode_unlock(struct au_hinode *hdir) +{ + au_hn_resume(hdir); + inode_unlock(hdir->hi_inode); +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_INODE_H__ */ diff --git a/fs/aufs/ioctl.c b/fs/aufs/ioctl.c new file mode 100644 index 00000000000..ae4f73386b9 --- /dev/null +++ b/fs/aufs/ioctl.c @@ -0,0 +1,220 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * ioctl + * plink-management and readdir in userspace. + * assist the pathconf(3) wrapper library. + * move-down + * File-based Hierarchical Storage Management. + */ + +#include +#include +#include "aufs.h" + +static int au_wbr_fd(struct path *path, struct aufs_wbr_fd __user *arg) +{ + int err, fd; + aufs_bindex_t wbi, bindex, bbot; + struct file *h_file; + struct super_block *sb; + struct dentry *root; + struct au_branch *br; + struct aufs_wbr_fd wbrfd = { + .oflags = au_dir_roflags, + .brid = -1 + }; + const int valid = O_RDONLY | O_NONBLOCK | O_LARGEFILE | O_DIRECTORY + | O_NOATIME | O_CLOEXEC; + + AuDebugOn(wbrfd.oflags & ~valid); + + if (arg) { + err = copy_from_user(&wbrfd, arg, sizeof(wbrfd)); + if (unlikely(err)) { + err = -EFAULT; + goto out; + } + + err = -EINVAL; + AuDbg("wbrfd{0%o, %d}\n", wbrfd.oflags, wbrfd.brid); + wbrfd.oflags |= au_dir_roflags; + AuDbg("0%o\n", wbrfd.oflags); + if (unlikely(wbrfd.oflags & ~valid)) + goto out; + } + + fd = get_unused_fd_flags(0); + err = fd; + if (unlikely(fd < 0)) + goto out; + + h_file = ERR_PTR(-EINVAL); + wbi = 0; + br = NULL; + sb = path->dentry->d_sb; + root = sb->s_root; + aufs_read_lock(root, AuLock_IR); + bbot = au_sbbot(sb); + if (wbrfd.brid >= 0) { + wbi = au_br_index(sb, wbrfd.brid); + if (unlikely(wbi < 0 || wbi > bbot)) + goto out_unlock; + } + + h_file = ERR_PTR(-ENOENT); + br = au_sbr(sb, wbi); + if (!au_br_writable(br->br_perm)) { + if (arg) + goto out_unlock; + + bindex = wbi + 1; + wbi = -1; + for (; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_writable(br->br_perm)) { + wbi = bindex; + br = au_sbr(sb, wbi); + break; + } + } + } + AuDbg("wbi %d\n", wbi); + if (wbi >= 0) + h_file = au_h_open(root, wbi, wbrfd.oflags, NULL, + /*force_wr*/0); + +out_unlock: + aufs_read_unlock(root, AuLock_IR); + err = PTR_ERR(h_file); + if (IS_ERR(h_file)) + goto out_fd; + + au_lcnt_dec(&br->br_nfiles); /* cf. au_h_open() */ + fd_install(fd, h_file); + err = fd; + goto out; /* success */ + +out_fd: + put_unused_fd(fd); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +long aufs_ioctl_dir(struct file *file, unsigned int cmd, unsigned long arg) +{ + long err; + struct dentry *dentry; + + switch (cmd) { + case AUFS_CTL_RDU: + case AUFS_CTL_RDU_INO: + err = au_rdu_ioctl(file, cmd, arg); + break; + + case AUFS_CTL_WBR_FD: + err = au_wbr_fd(&file->f_path, (void __user *)arg); + break; + + case AUFS_CTL_IBUSY: + err = au_ibusy_ioctl(file, arg); + break; + + case AUFS_CTL_BRINFO: + err = au_brinfo_ioctl(file, arg); + break; + + case AUFS_CTL_FHSM_FD: + dentry = file->f_path.dentry; + if (IS_ROOT(dentry)) + err = au_fhsm_fd(dentry->d_sb, arg); + else + err = -ENOTTY; + break; + + default: + /* do not call the lower */ + AuDbg("0x%x\n", cmd); + err = -ENOTTY; + } + + AuTraceErr(err); + return err; +} + +long aufs_ioctl_nondir(struct file *file, unsigned int cmd, unsigned long arg) +{ + long err; + + switch (cmd) { + case AUFS_CTL_MVDOWN: + err = au_mvdown(file->f_path.dentry, (void __user *)arg); + break; + + case AUFS_CTL_WBR_FD: + err = au_wbr_fd(&file->f_path, (void __user *)arg); + break; + + default: + /* do not call the lower */ + AuDbg("0x%x\n", cmd); + err = -ENOTTY; + } + + AuTraceErr(err); + return err; +} + +#ifdef CONFIG_COMPAT +long aufs_compat_ioctl_dir(struct file *file, unsigned int cmd, + unsigned long arg) +{ + long err; + + switch (cmd) { + case AUFS_CTL_RDU: + case AUFS_CTL_RDU_INO: + err = au_rdu_compat_ioctl(file, cmd, arg); + break; + + case AUFS_CTL_IBUSY: + err = au_ibusy_compat_ioctl(file, arg); + break; + + case AUFS_CTL_BRINFO: + err = au_brinfo_compat_ioctl(file, arg); + break; + + default: + err = aufs_ioctl_dir(file, cmd, arg); + } + + AuTraceErr(err); + return err; +} + +long aufs_compat_ioctl_nondir(struct file *file, unsigned int cmd, + unsigned long arg) +{ + return aufs_ioctl_nondir(file, cmd, (unsigned long)compat_ptr(arg)); +} +#endif diff --git a/fs/aufs/lcnt.h b/fs/aufs/lcnt.h new file mode 100644 index 00000000000..8afcabe5582 --- /dev/null +++ b/fs/aufs/lcnt.h @@ -0,0 +1,186 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2018-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * simple long counter wrapper + */ + +#ifndef __AUFS_LCNT_H__ +#define __AUFS_LCNT_H__ + +#ifdef __KERNEL__ + +#include "debug.h" + +#define AuLCntATOMIC 1 +#define AuLCntPCPUCNT 2 +/* + * why does percpu_refcount require extra synchronize_rcu()s in + * au_br_do_free() + */ +#define AuLCntPCPUREF 3 + +/* #define AuLCntChosen AuLCntATOMIC */ +#define AuLCntChosen AuLCntPCPUCNT +/* #define AuLCntChosen AuLCntPCPUREF */ + +#if AuLCntChosen == AuLCntATOMIC +#include + +typedef atomic_long_t au_lcnt_t; + +static inline int au_lcnt_init(au_lcnt_t *cnt, void *release __maybe_unused) +{ + atomic_long_set(cnt, 0); + return 0; +} + +static inline void au_lcnt_wait_for_fin(au_lcnt_t *cnt __maybe_unused) +{ + /* empty */ +} + +static inline void au_lcnt_fin(au_lcnt_t *cnt __maybe_unused, + int do_sync __maybe_unused) +{ + /* empty */ +} + +static inline void au_lcnt_inc(au_lcnt_t *cnt) +{ + atomic_long_inc(cnt); +} + +static inline void au_lcnt_dec(au_lcnt_t *cnt) +{ + atomic_long_dec(cnt); +} + +static inline long au_lcnt_read(au_lcnt_t *cnt, int do_rev __maybe_unused) +{ + return atomic_long_read(cnt); +} +#endif + +#if AuLCntChosen == AuLCntPCPUCNT +#include + +typedef struct percpu_counter au_lcnt_t; + +static inline int au_lcnt_init(au_lcnt_t *cnt, void *release __maybe_unused) +{ + return percpu_counter_init(cnt, 0, GFP_NOFS); +} + +static inline void au_lcnt_wait_for_fin(au_lcnt_t *cnt __maybe_unused) +{ + /* empty */ +} + +static inline void au_lcnt_fin(au_lcnt_t *cnt, int do_sync __maybe_unused) +{ + percpu_counter_destroy(cnt); +} + +static inline void au_lcnt_inc(au_lcnt_t *cnt) +{ + percpu_counter_inc(cnt); +} + +static inline void au_lcnt_dec(au_lcnt_t *cnt) +{ + percpu_counter_dec(cnt); +} + +static inline long au_lcnt_read(au_lcnt_t *cnt, int do_rev __maybe_unused) +{ + s64 n; + + n = percpu_counter_sum(cnt); + BUG_ON(n < 0); + if (LONG_MAX != LLONG_MAX + && n > LONG_MAX) + AuWarn1("%s\n", "wrap-around"); + + return n; +} +#endif + +#if AuLCntChosen == AuLCntPCPUREF +#include + +typedef struct percpu_ref au_lcnt_t; + +static inline int au_lcnt_init(au_lcnt_t *cnt, percpu_ref_func_t *release) +{ + if (!release) + release = percpu_ref_exit; + return percpu_ref_init(cnt, release, /*percpu mode*/0, GFP_NOFS); +} + +static inline void au_lcnt_wait_for_fin(au_lcnt_t *cnt __maybe_unused) +{ + synchronize_rcu(); +} + +static inline void au_lcnt_fin(au_lcnt_t *cnt, int do_sync) +{ + percpu_ref_kill(cnt); + if (do_sync) + au_lcnt_wait_for_fin(cnt); +} + +static inline void au_lcnt_inc(au_lcnt_t *cnt) +{ + percpu_ref_get(cnt); +} + +static inline void au_lcnt_dec(au_lcnt_t *cnt) +{ + percpu_ref_put(cnt); +} + +/* + * avoid calling this func as possible. + */ +static inline long au_lcnt_read(au_lcnt_t *cnt, int do_rev) +{ + long l; + + percpu_ref_switch_to_atomic_sync(cnt); + l = atomic_long_read(&cnt->count); + if (do_rev) + percpu_ref_switch_to_percpu(cnt); + + /* percpu_ref is initialized by 1 instead of 0 */ + return l - 1; +} +#endif + +#ifdef CONFIG_AUFS_DEBUG +#define AuLCntZero(val) do { \ + long l = val; \ + if (l) \ + AuDbg("%s = %ld\n", #val, l); \ +} while (0) +#else +#define AuLCntZero(val) do {} while (0) +#endif + +#endif /* __KERNEL__ */ +#endif /* __AUFS_LCNT_H__ */ diff --git a/fs/aufs/loop.c b/fs/aufs/loop.c new file mode 100644 index 00000000000..63732e93868 --- /dev/null +++ b/fs/aufs/loop.c @@ -0,0 +1,148 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * support for loopback block device as a branch + */ + +#include "aufs.h" + +/* added into drivers/block/loop.c */ +static struct file *(*backing_file_func)(struct super_block *sb); + +/* + * test if two lower dentries have overlapping branches. + */ +int au_test_loopback_overlap(struct super_block *sb, struct dentry *h_adding) +{ + struct super_block *h_sb; + struct file *backing_file; + + if (unlikely(!backing_file_func)) { + /* don't load "loop" module here */ + backing_file_func = symbol_get(loop_backing_file); + if (unlikely(!backing_file_func)) + /* "loop" module is not loaded */ + return 0; + } + + h_sb = h_adding->d_sb; + backing_file = backing_file_func(h_sb); + if (!backing_file) + return 0; + + h_adding = backing_file->f_path.dentry; + /* + * h_adding can be local NFS. + * in this case aufs cannot detect the loop. + */ + if (unlikely(h_adding->d_sb == sb)) + return 1; + return !!au_test_subdir(h_adding, sb->s_root); +} + +/* true if a kernel thread named 'loop[0-9].*' accesses a file */ +int au_test_loopback_kthread(void) +{ + int ret; + struct task_struct *tsk = current; + char c, comm[sizeof(tsk->comm)]; + + ret = 0; + if (tsk->flags & PF_KTHREAD) { + get_task_comm(comm, tsk); + c = comm[4]; + ret = ('0' <= c && c <= '9' + && !strncmp(comm, "loop", 4)); + } + + return ret; +} + +/* ---------------------------------------------------------------------- */ + +#define au_warn_loopback_step 16 +static int au_warn_loopback_nelem = au_warn_loopback_step; +static unsigned long *au_warn_loopback_array; + +void au_warn_loopback(struct super_block *h_sb) +{ + int i, new_nelem; + unsigned long *a, magic; + static DEFINE_SPINLOCK(spin); + + magic = h_sb->s_magic; + spin_lock(&spin); + a = au_warn_loopback_array; + for (i = 0; i < au_warn_loopback_nelem && *a; i++) + if (a[i] == magic) { + spin_unlock(&spin); + return; + } + + /* h_sb is new to us, print it */ + if (i < au_warn_loopback_nelem) { + a[i] = magic; + goto pr; + } + + /* expand the array */ + new_nelem = au_warn_loopback_nelem + au_warn_loopback_step; + a = au_kzrealloc(au_warn_loopback_array, + au_warn_loopback_nelem * sizeof(unsigned long), + new_nelem * sizeof(unsigned long), GFP_ATOMIC, + /*may_shrink*/0); + if (a) { + au_warn_loopback_nelem = new_nelem; + au_warn_loopback_array = a; + a[i] = magic; + goto pr; + } + + spin_unlock(&spin); + AuWarn1("realloc failed, ignored\n"); + return; + +pr: + spin_unlock(&spin); + pr_warn("you may want to try another patch for loopback file " + "on %s(0x%lx) branch\n", au_sbtype(h_sb), magic); +} + +int au_loopback_init(void) +{ + int err; + struct super_block *sb __maybe_unused; + + BUILD_BUG_ON(sizeof(sb->s_magic) != sizeof(*au_warn_loopback_array)); + + err = 0; + au_warn_loopback_array = kcalloc(au_warn_loopback_step, + sizeof(unsigned long), GFP_NOFS); + if (unlikely(!au_warn_loopback_array)) + err = -ENOMEM; + + return err; +} + +void au_loopback_fin(void) +{ + if (backing_file_func) + symbol_put(loop_backing_file); + au_kfree_try_rcu(au_warn_loopback_array); +} diff --git a/fs/aufs/loop.h b/fs/aufs/loop.h new file mode 100644 index 00000000000..65c38bbb5d5 --- /dev/null +++ b/fs/aufs/loop.h @@ -0,0 +1,55 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * support for loopback mount as a branch + */ + +#ifndef __AUFS_LOOP_H__ +#define __AUFS_LOOP_H__ + +#ifdef __KERNEL__ + +struct dentry; +struct super_block; + +#ifdef CONFIG_AUFS_BDEV_LOOP +/* drivers/block/loop.c */ +struct file *loop_backing_file(struct super_block *sb); + +/* loop.c */ +int au_test_loopback_overlap(struct super_block *sb, struct dentry *h_adding); +int au_test_loopback_kthread(void); +void au_warn_loopback(struct super_block *h_sb); + +int au_loopback_init(void); +void au_loopback_fin(void); +#else +AuStub(struct file *, loop_backing_file, return NULL, struct super_block *sb) + +AuStubInt0(au_test_loopback_overlap, struct super_block *sb, + struct dentry *h_adding) +AuStubInt0(au_test_loopback_kthread, void) +AuStubVoid(au_warn_loopback, struct super_block *h_sb) + +AuStubInt0(au_loopback_init, void) +AuStubVoid(au_loopback_fin, void) +#endif /* BLK_DEV_LOOP */ + +#endif /* __KERNEL__ */ +#endif /* __AUFS_LOOP_H__ */ diff --git a/fs/aufs/magic.mk b/fs/aufs/magic.mk new file mode 100644 index 00000000000..7bc9eef3ffe --- /dev/null +++ b/fs/aufs/magic.mk @@ -0,0 +1,31 @@ +# SPDX-License-Identifier: GPL-2.0 + +# defined in ${srctree}/fs/fuse/inode.c +# tristate +ifdef CONFIG_FUSE_FS +ccflags-y += -DFUSE_SUPER_MAGIC=0x65735546 +endif + +# defined in ${srctree}/fs/xfs/xfs_sb.h +# tristate +ifdef CONFIG_XFS_FS +ccflags-y += -DXFS_SB_MAGIC=0x58465342 +endif + +# defined in ${srctree}/fs/configfs/mount.c +# tristate +ifdef CONFIG_CONFIGFS_FS +ccflags-y += -DCONFIGFS_MAGIC=0x62656570 +endif + +# defined in ${srctree}/fs/ubifs/ubifs.h +# tristate +ifdef CONFIG_UBIFS_FS +ccflags-y += -DUBIFS_SUPER_MAGIC=0x24051905 +endif + +# defined in ${srctree}/fs/hfsplus/hfsplus_raw.h +# tristate +ifdef CONFIG_HFSPLUS_FS +ccflags-y += -DHFSPLUS_SUPER_MAGIC=0x482b +endif diff --git a/fs/aufs/module.c b/fs/aufs/module.c new file mode 100644 index 00000000000..7245197a6cb --- /dev/null +++ b/fs/aufs/module.c @@ -0,0 +1,273 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * module global variables and operations + */ + +#include +#include +#include "aufs.h" + +/* shrinkable realloc */ +void *au_krealloc(void *p, unsigned int new_sz, gfp_t gfp, int may_shrink) +{ + size_t sz; + int diff; + + sz = 0; + diff = -1; + if (p) { +#if 0 /* unused */ + if (!new_sz) { + au_kfree_rcu(p); + p = NULL; + goto out; + } +#else + AuDebugOn(!new_sz); +#endif + sz = ksize(p); + diff = au_kmidx_sub(sz, new_sz); + } + if (sz && !diff) + goto out; + + if (sz < new_sz) + /* expand or SLOB */ + p = krealloc(p, new_sz, gfp); + else if (new_sz < sz && may_shrink) { + /* shrink */ + void *q; + + q = kmalloc(new_sz, gfp); + if (q) { + if (p) { + memcpy(q, p, new_sz); + au_kfree_try_rcu(p); + } + p = q; + } else + p = NULL; + } + +out: + return p; +} + +void *au_kzrealloc(void *p, unsigned int nused, unsigned int new_sz, gfp_t gfp, + int may_shrink) +{ + p = au_krealloc(p, new_sz, gfp, may_shrink); + if (p && new_sz > nused) + memset(p + nused, 0, new_sz - nused); + return p; +} + +/* ---------------------------------------------------------------------- */ +/* + * aufs caches + */ +struct kmem_cache *au_cache[AuCache_Last]; + +static void au_cache_fin(void) +{ + int i; + + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); + + /* excluding AuCache_HNOTIFY */ + BUILD_BUG_ON(AuCache_HNOTIFY + 1 != AuCache_Last); + for (i = 0; i < AuCache_HNOTIFY; i++) { + kmem_cache_destroy(au_cache[i]); + au_cache[i] = NULL; + } +} + +static int __init au_cache_init(void) +{ + au_cache[AuCache_DINFO] = AuCacheCtor(au_dinfo, au_di_init_once); + if (au_cache[AuCache_DINFO]) + /* SLAB_DESTROY_BY_RCU */ + au_cache[AuCache_ICNTNR] = AuCacheCtor(au_icntnr, + au_icntnr_init_once); + if (au_cache[AuCache_ICNTNR]) + au_cache[AuCache_FINFO] = AuCacheCtor(au_finfo, + au_fi_init_once); + if (au_cache[AuCache_FINFO]) + au_cache[AuCache_VDIR] = AuCache(au_vdir); + if (au_cache[AuCache_VDIR]) + au_cache[AuCache_DEHSTR] = AuCache(au_vdir_dehstr); + if (au_cache[AuCache_DEHSTR]) + return 0; + + au_cache_fin(); + return -ENOMEM; +} + +/* ---------------------------------------------------------------------- */ + +int au_dir_roflags; + +#ifdef CONFIG_AUFS_SBILIST +/* + * iterate_supers_type() doesn't protect us from + * remounting (branch management) + */ +struct hlist_bl_head au_sbilist; +#endif + +/* + * functions for module interface. + */ +MODULE_LICENSE("GPL"); +/* MODULE_LICENSE("GPL v2"); */ +MODULE_AUTHOR("Junjiro R. Okajima "); +MODULE_DESCRIPTION(AUFS_NAME + " -- Advanced multi layered unification filesystem"); +MODULE_VERSION(AUFS_VERSION); +MODULE_ALIAS_FS(AUFS_NAME); + +/* this module parameter has no meaning when SYSFS is disabled */ +int sysaufs_brs = 1; +MODULE_PARM_DESC(brs, "use /fs/aufs/si_*/brN"); +module_param_named(brs, sysaufs_brs, int, 0444); + +/* this module parameter has no meaning when USER_NS is disabled */ +bool au_userns; +MODULE_PARM_DESC(allow_userns, "allow unprivileged to mount under userns"); +module_param_named(allow_userns, au_userns, bool, 0444); + +/* ---------------------------------------------------------------------- */ + +static char au_esc_chars[0x20 + 3]; /* 0x01-0x20, backslash, del, and NULL */ + +int au_seq_path(struct seq_file *seq, struct path *path) +{ + int err; + + err = seq_path(seq, path, au_esc_chars); + if (err >= 0) + err = 0; + else + err = -ENOMEM; + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int __init aufs_init(void) +{ + int err, i; + char *p; + + p = au_esc_chars; + for (i = 1; i <= ' '; i++) + *p++ = i; + *p++ = '\\'; + *p++ = '\x7f'; + *p = 0; + + au_dir_roflags = au_file_roflags(O_DIRECTORY | O_LARGEFILE); + + memcpy(aufs_iop_nogetattr, aufs_iop, sizeof(aufs_iop)); + for (i = 0; i < AuIop_Last; i++) + aufs_iop_nogetattr[i].getattr = NULL; + + memset(au_cache, 0, sizeof(au_cache)); /* including hnotify */ + + au_sbilist_init(); + sysaufs_brs_init(); + au_debug_init(); + au_dy_init(); + err = sysaufs_init(); + if (unlikely(err)) + goto out; + err = dbgaufs_init(); + if (unlikely(err)) + goto out_sysaufs; + err = au_procfs_init(); + if (unlikely(err)) + goto out_dbgaufs; + err = au_wkq_init(); + if (unlikely(err)) + goto out_procfs; + err = au_loopback_init(); + if (unlikely(err)) + goto out_wkq; + err = au_hnotify_init(); + if (unlikely(err)) + goto out_loopback; + err = au_sysrq_init(); + if (unlikely(err)) + goto out_hin; + err = au_cache_init(); + if (unlikely(err)) + goto out_sysrq; + + aufs_fs_type.fs_flags |= au_userns ? FS_USERNS_MOUNT : 0; + err = register_filesystem(&aufs_fs_type); + if (unlikely(err)) + goto out_cache; + + /* since we define pr_fmt, call printk directly */ + printk(KERN_INFO AUFS_NAME " " AUFS_VERSION "\n"); + goto out; /* success */ + +out_cache: + au_cache_fin(); +out_sysrq: + au_sysrq_fin(); +out_hin: + au_hnotify_fin(); +out_loopback: + au_loopback_fin(); +out_wkq: + au_wkq_fin(); +out_procfs: + au_procfs_fin(); +out_dbgaufs: + dbgaufs_fin(); +out_sysaufs: + sysaufs_fin(); + au_dy_fin(); +out: + return err; +} + +static void __exit aufs_exit(void) +{ + unregister_filesystem(&aufs_fs_type); + au_cache_fin(); + au_sysrq_fin(); + au_hnotify_fin(); + au_loopback_fin(); + au_wkq_fin(); + au_procfs_fin(); + dbgaufs_fin(); + sysaufs_fin(); + au_dy_fin(); +} + +module_init(aufs_init); +module_exit(aufs_exit); diff --git a/fs/aufs/module.h b/fs/aufs/module.h new file mode 100644 index 00000000000..d6a76788d2f --- /dev/null +++ b/fs/aufs/module.h @@ -0,0 +1,166 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * module initialization and module-global + */ + +#ifndef __AUFS_MODULE_H__ +#define __AUFS_MODULE_H__ + +#ifdef __KERNEL__ + +#include +#include "debug.h" +#include "dentry.h" +#include "dir.h" +#include "file.h" +#include "inode.h" + +struct path; +struct seq_file; + +/* module parameters */ +extern int sysaufs_brs; +extern bool au_userns; + +/* ---------------------------------------------------------------------- */ + +extern int au_dir_roflags; + +void *au_krealloc(void *p, unsigned int new_sz, gfp_t gfp, int may_shrink); +void *au_kzrealloc(void *p, unsigned int nused, unsigned int new_sz, gfp_t gfp, + int may_shrink); + +/* + * Comparing the size of the object with sizeof(struct rcu_head) + * case 1: object is always larger + * --> au_kfree_rcu() or au_kfree_do_rcu() + * case 2: object is always smaller + * --> au_kfree_small() + * case 3: object can be any size + * --> au_kfree_try_rcu() + */ + +static inline void au_kfree_do_rcu(const void *p) +{ + struct { + struct rcu_head rcu; + } *a = (void *)p; + + kfree_rcu(a, rcu); +} + +#define au_kfree_rcu(_p) do { \ + typeof(_p) p = (_p); \ + BUILD_BUG_ON(sizeof(*p) < sizeof(struct rcu_head)); \ + if (p) \ + au_kfree_do_rcu(p); \ + } while (0) + +#define au_kfree_do_sz_test(sz) (sz >= sizeof(struct rcu_head)) +#define au_kfree_sz_test(p) (p && au_kfree_do_sz_test(ksize(p))) + +static inline void au_kfree_try_rcu(const void *p) +{ + if (!p) + return; + if (au_kfree_sz_test(p)) + au_kfree_do_rcu(p); + else + kfree(p); +} + +static inline void au_kfree_small(const void *p) +{ + if (!p) + return; + AuDebugOn(au_kfree_sz_test(p)); + kfree(p); +} + +static inline int au_kmidx_sub(size_t sz, size_t new_sz) +{ +#ifndef CONFIG_SLOB + return kmalloc_index(sz) - kmalloc_index(new_sz); +#else + return -1; /* SLOB is untested */ +#endif +} + +int au_seq_path(struct seq_file *seq, struct path *path); + +#ifdef CONFIG_PROC_FS +/* procfs.c */ +int __init au_procfs_init(void); +void au_procfs_fin(void); +#else +AuStubInt0(au_procfs_init, void); +AuStubVoid(au_procfs_fin, void); +#endif + +/* ---------------------------------------------------------------------- */ + +/* kmem cache */ +enum { + AuCache_DINFO, + AuCache_ICNTNR, + AuCache_FINFO, + AuCache_VDIR, + AuCache_DEHSTR, + AuCache_HNOTIFY, /* must be last */ + AuCache_Last +}; + +extern struct kmem_cache *au_cache[AuCache_Last]; + +#define AuCacheFlags (SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD) +#define AuCache(type) KMEM_CACHE(type, AuCacheFlags) +#define AuCacheCtor(type, ctor) \ + kmem_cache_create(#type, sizeof(struct type), \ + __alignof__(struct type), AuCacheFlags, ctor) + +#define AuCacheFuncs(name, index) \ + static inline struct au_##name *au_cache_alloc_##name(void) \ + { return kmem_cache_alloc(au_cache[AuCache_##index], GFP_NOFS); } \ + static inline void au_cache_free_##name##_norcu(struct au_##name *p) \ + { kmem_cache_free(au_cache[AuCache_##index], p); } \ + \ + static inline void au_cache_free_##name##_rcu_cb(struct rcu_head *rcu) \ + { void *p = rcu; \ + p -= offsetof(struct au_##name, rcu); \ + kmem_cache_free(au_cache[AuCache_##index], p); } \ + static inline void au_cache_free_##name##_rcu(struct au_##name *p) \ + { BUILD_BUG_ON(sizeof(struct au_##name) < sizeof(struct rcu_head)); \ + call_rcu(&p->rcu, au_cache_free_##name##_rcu_cb); } \ + \ + static inline void au_cache_free_##name(struct au_##name *p) \ + { /* au_cache_free_##name##_norcu(p); */ \ + au_cache_free_##name##_rcu(p); } + +AuCacheFuncs(dinfo, DINFO); +AuCacheFuncs(icntnr, ICNTNR); +AuCacheFuncs(finfo, FINFO); +AuCacheFuncs(vdir, VDIR); +AuCacheFuncs(vdir_dehstr, DEHSTR); +#ifdef CONFIG_AUFS_HNOTIFY +AuCacheFuncs(hnotify, HNOTIFY); +#endif + +#endif /* __KERNEL__ */ +#endif /* __AUFS_MODULE_H__ */ diff --git a/fs/aufs/mvdown.c b/fs/aufs/mvdown.c new file mode 100644 index 00000000000..c39a39e52eb --- /dev/null +++ b/fs/aufs/mvdown.c @@ -0,0 +1,706 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2011-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * move-down, opposite of copy-up + */ + +#include "aufs.h" + +struct au_mvd_args { + struct { + struct super_block *h_sb; + struct dentry *h_parent; + struct au_hinode *hdir; + struct inode *h_dir, *h_inode; + struct au_pin pin; + } info[AUFS_MVDOWN_NARRAY]; + + struct aufs_mvdown mvdown; + struct dentry *dentry, *parent; + struct inode *inode, *dir; + struct super_block *sb; + aufs_bindex_t bopq, bwh, bfound; + unsigned char rename_lock; +}; + +#define mvd_errno mvdown.au_errno +#define mvd_bsrc mvdown.stbr[AUFS_MVDOWN_UPPER].bindex +#define mvd_src_brid mvdown.stbr[AUFS_MVDOWN_UPPER].brid +#define mvd_bdst mvdown.stbr[AUFS_MVDOWN_LOWER].bindex +#define mvd_dst_brid mvdown.stbr[AUFS_MVDOWN_LOWER].brid + +#define mvd_h_src_sb info[AUFS_MVDOWN_UPPER].h_sb +#define mvd_h_src_parent info[AUFS_MVDOWN_UPPER].h_parent +#define mvd_hdir_src info[AUFS_MVDOWN_UPPER].hdir +#define mvd_h_src_dir info[AUFS_MVDOWN_UPPER].h_dir +#define mvd_h_src_inode info[AUFS_MVDOWN_UPPER].h_inode +#define mvd_pin_src info[AUFS_MVDOWN_UPPER].pin + +#define mvd_h_dst_sb info[AUFS_MVDOWN_LOWER].h_sb +#define mvd_h_dst_parent info[AUFS_MVDOWN_LOWER].h_parent +#define mvd_hdir_dst info[AUFS_MVDOWN_LOWER].hdir +#define mvd_h_dst_dir info[AUFS_MVDOWN_LOWER].h_dir +#define mvd_h_dst_inode info[AUFS_MVDOWN_LOWER].h_inode +#define mvd_pin_dst info[AUFS_MVDOWN_LOWER].pin + +#define AU_MVD_PR(flag, ...) do { \ + if (flag) \ + pr_err(__VA_ARGS__); \ + } while (0) + +static int find_lower_writable(struct au_mvd_args *a) +{ + struct super_block *sb; + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + sb = a->sb; + bindex = a->mvd_bsrc; + bbot = au_sbbot(sb); + if (a->mvdown.flags & AUFS_MVDOWN_FHSM_LOWER) + for (bindex++; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_fhsm(br->br_perm) + && !sb_rdonly(au_br_sb(br))) + return bindex; + } + else if (!(a->mvdown.flags & AUFS_MVDOWN_ROLOWER)) + for (bindex++; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (!au_br_rdonly(br)) + return bindex; + } + else + for (bindex++; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (!sb_rdonly(au_br_sb(br))) { + if (au_br_rdonly(br)) + a->mvdown.flags + |= AUFS_MVDOWN_ROLOWER_R; + return bindex; + } + } + + return -1; +} + +/* make the parent dir on bdst */ +static int au_do_mkdir(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + + err = 0; + a->mvd_hdir_src = au_hi(a->dir, a->mvd_bsrc); + a->mvd_hdir_dst = au_hi(a->dir, a->mvd_bdst); + a->mvd_h_src_parent = au_h_dptr(a->parent, a->mvd_bsrc); + a->mvd_h_dst_parent = NULL; + if (au_dbbot(a->parent) >= a->mvd_bdst) + a->mvd_h_dst_parent = au_h_dptr(a->parent, a->mvd_bdst); + if (!a->mvd_h_dst_parent) { + err = au_cpdown_dirs(a->dentry, a->mvd_bdst); + if (unlikely(err)) { + AU_MVD_PR(dmsg, "cpdown_dirs failed\n"); + goto out; + } + a->mvd_h_dst_parent = au_h_dptr(a->parent, a->mvd_bdst); + } + +out: + AuTraceErr(err); + return err; +} + +/* lock them all */ +static int au_do_lock(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct dentry *h_trap; + + a->mvd_h_src_sb = au_sbr_sb(a->sb, a->mvd_bsrc); + a->mvd_h_dst_sb = au_sbr_sb(a->sb, a->mvd_bdst); + err = au_pin(&a->mvd_pin_dst, a->dentry, a->mvd_bdst, + au_opt_udba(a->sb), + AuPin_MNT_WRITE | AuPin_DI_LOCKED); + AuTraceErr(err); + if (unlikely(err)) { + AU_MVD_PR(dmsg, "pin_dst failed\n"); + goto out; + } + + if (a->mvd_h_src_sb != a->mvd_h_dst_sb) { + a->rename_lock = 0; + au_pin_init(&a->mvd_pin_src, a->dentry, a->mvd_bsrc, + AuLsc_DI_PARENT, AuLsc_I_PARENT3, + au_opt_udba(a->sb), + AuPin_MNT_WRITE | AuPin_DI_LOCKED); + err = au_do_pin(&a->mvd_pin_src); + AuTraceErr(err); + a->mvd_h_src_dir = d_inode(a->mvd_h_src_parent); + if (unlikely(err)) { + AU_MVD_PR(dmsg, "pin_src failed\n"); + goto out_dst; + } + goto out; /* success */ + } + + a->rename_lock = 1; + au_pin_hdir_unlock(&a->mvd_pin_dst); + err = au_pin(&a->mvd_pin_src, a->dentry, a->mvd_bsrc, + au_opt_udba(a->sb), + AuPin_MNT_WRITE | AuPin_DI_LOCKED); + AuTraceErr(err); + a->mvd_h_src_dir = d_inode(a->mvd_h_src_parent); + if (unlikely(err)) { + AU_MVD_PR(dmsg, "pin_src failed\n"); + au_pin_hdir_lock(&a->mvd_pin_dst); + goto out_dst; + } + au_pin_hdir_unlock(&a->mvd_pin_src); + h_trap = vfsub_lock_rename(a->mvd_h_src_parent, a->mvd_hdir_src, + a->mvd_h_dst_parent, a->mvd_hdir_dst); + if (h_trap) { + err = (h_trap != a->mvd_h_src_parent); + if (err) + err = (h_trap != a->mvd_h_dst_parent); + } + BUG_ON(err); /* it should never happen */ + if (unlikely(a->mvd_h_src_dir != au_pinned_h_dir(&a->mvd_pin_src))) { + err = -EBUSY; + AuTraceErr(err); + vfsub_unlock_rename(a->mvd_h_src_parent, a->mvd_hdir_src, + a->mvd_h_dst_parent, a->mvd_hdir_dst); + au_pin_hdir_lock(&a->mvd_pin_src); + au_unpin(&a->mvd_pin_src); + au_pin_hdir_lock(&a->mvd_pin_dst); + goto out_dst; + } + goto out; /* success */ + +out_dst: + au_unpin(&a->mvd_pin_dst); +out: + AuTraceErr(err); + return err; +} + +static void au_do_unlock(const unsigned char dmsg, struct au_mvd_args *a) +{ + if (!a->rename_lock) + au_unpin(&a->mvd_pin_src); + else { + vfsub_unlock_rename(a->mvd_h_src_parent, a->mvd_hdir_src, + a->mvd_h_dst_parent, a->mvd_hdir_dst); + au_pin_hdir_lock(&a->mvd_pin_src); + au_unpin(&a->mvd_pin_src); + au_pin_hdir_lock(&a->mvd_pin_dst); + } + au_unpin(&a->mvd_pin_dst); +} + +/* copy-down the file */ +static int au_do_cpdown(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct au_cp_generic cpg = { + .dentry = a->dentry, + .bdst = a->mvd_bdst, + .bsrc = a->mvd_bsrc, + .len = -1, + .pin = &a->mvd_pin_dst, + .flags = AuCpup_DTIME | AuCpup_HOPEN + }; + + AuDbg("b%d, b%d\n", cpg.bsrc, cpg.bdst); + if (a->mvdown.flags & AUFS_MVDOWN_OWLOWER) + au_fset_cpup(cpg.flags, OVERWRITE); + if (a->mvdown.flags & AUFS_MVDOWN_ROLOWER) + au_fset_cpup(cpg.flags, RWDST); + err = au_sio_cpdown_simple(&cpg); + if (unlikely(err)) + AU_MVD_PR(dmsg, "cpdown failed\n"); + + AuTraceErr(err); + return err; +} + +/* + * unlink the whiteout on bdst if exist which may be created by UDBA while we + * were sleeping + */ +static int au_do_unlink_wh(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct path h_path; + struct au_branch *br; + struct inode *delegated; + + br = au_sbr(a->sb, a->mvd_bdst); + h_path.dentry = au_wh_lkup(a->mvd_h_dst_parent, &a->dentry->d_name, br); + err = PTR_ERR(h_path.dentry); + if (IS_ERR(h_path.dentry)) { + AU_MVD_PR(dmsg, "wh_lkup failed\n"); + goto out; + } + + err = 0; + if (d_is_positive(h_path.dentry)) { + h_path.mnt = au_br_mnt(br); + delegated = NULL; + err = vfsub_unlink(d_inode(a->mvd_h_dst_parent), &h_path, + &delegated, /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + if (unlikely(err)) + AU_MVD_PR(dmsg, "wh_unlink failed\n"); + } + dput(h_path.dentry); + +out: + AuTraceErr(err); + return err; +} + +/* + * unlink the topmost h_dentry + */ +static int au_do_unlink(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct path h_path; + struct inode *delegated; + + h_path.mnt = au_sbr_mnt(a->sb, a->mvd_bsrc); + h_path.dentry = au_h_dptr(a->dentry, a->mvd_bsrc); + delegated = NULL; + err = vfsub_unlink(a->mvd_h_src_dir, &h_path, &delegated, /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + if (unlikely(err)) + AU_MVD_PR(dmsg, "unlink failed\n"); + + AuTraceErr(err); + return err; +} + +/* Since mvdown succeeded, we ignore an error of this function */ +static void au_do_stfs(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct au_branch *br; + + a->mvdown.flags |= AUFS_MVDOWN_STFS_FAILED; + br = au_sbr(a->sb, a->mvd_bsrc); + err = au_br_stfs(br, &a->mvdown.stbr[AUFS_MVDOWN_UPPER].stfs); + if (!err) { + br = au_sbr(a->sb, a->mvd_bdst); + a->mvdown.stbr[AUFS_MVDOWN_LOWER].brid = br->br_id; + err = au_br_stfs(br, &a->mvdown.stbr[AUFS_MVDOWN_LOWER].stfs); + } + if (!err) + a->mvdown.flags &= ~AUFS_MVDOWN_STFS_FAILED; + else + AU_MVD_PR(dmsg, "statfs failed (%d), ignored\n", err); +} + +/* + * copy-down the file and unlink the bsrc file. + * - unlink the bdst whout if exist + * - copy-down the file (with whtmp name and rename) + * - unlink the bsrc file + */ +static int au_do_mvdown(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + + err = au_do_mkdir(dmsg, a); + if (!err) + err = au_do_lock(dmsg, a); + if (unlikely(err)) + goto out; + + /* + * do not revert the activities we made on bdst since they should be + * harmless in aufs. + */ + + err = au_do_cpdown(dmsg, a); + if (!err) + err = au_do_unlink_wh(dmsg, a); + if (!err && !(a->mvdown.flags & AUFS_MVDOWN_KUPPER)) + err = au_do_unlink(dmsg, a); + if (unlikely(err)) + goto out_unlock; + + AuDbg("%pd2, 0x%x, %d --> %d\n", + a->dentry, a->mvdown.flags, a->mvd_bsrc, a->mvd_bdst); + if (find_lower_writable(a) < 0) + a->mvdown.flags |= AUFS_MVDOWN_BOTTOM; + + if (a->mvdown.flags & AUFS_MVDOWN_STFS) + au_do_stfs(dmsg, a); + + /* maintain internal array */ + if (!(a->mvdown.flags & AUFS_MVDOWN_KUPPER)) { + au_set_h_dptr(a->dentry, a->mvd_bsrc, NULL); + au_set_dbtop(a->dentry, a->mvd_bdst); + au_set_h_iptr(a->inode, a->mvd_bsrc, NULL, /*flags*/0); + au_set_ibtop(a->inode, a->mvd_bdst); + } else { + /* hide the lower */ + au_set_h_dptr(a->dentry, a->mvd_bdst, NULL); + au_set_dbbot(a->dentry, a->mvd_bsrc); + au_set_h_iptr(a->inode, a->mvd_bdst, NULL, /*flags*/0); + au_set_ibbot(a->inode, a->mvd_bsrc); + } + if (au_dbbot(a->dentry) < a->mvd_bdst) + au_set_dbbot(a->dentry, a->mvd_bdst); + if (au_ibbot(a->inode) < a->mvd_bdst) + au_set_ibbot(a->inode, a->mvd_bdst); + +out_unlock: + au_do_unlock(dmsg, a); +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* make sure the file is idle */ +static int au_mvd_args_busy(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err, plinked; + + err = 0; + plinked = !!au_opt_test(au_mntflags(a->sb), PLINK); + if (au_dbtop(a->dentry) == a->mvd_bsrc + && au_dcount(a->dentry) == 1 + && atomic_read(&a->inode->i_count) == 1 + /* && a->mvd_h_src_inode->i_nlink == 1 */ + && (!plinked || !au_plink_test(a->inode)) + && a->inode->i_nlink == 1) + goto out; + + err = -EBUSY; + AU_MVD_PR(dmsg, + "b%d, d{b%d, c%d?}, i{c%d?, l%u}, hi{l%u}, p{%d, %d}\n", + a->mvd_bsrc, au_dbtop(a->dentry), au_dcount(a->dentry), + atomic_read(&a->inode->i_count), a->inode->i_nlink, + a->mvd_h_src_inode->i_nlink, + plinked, plinked ? au_plink_test(a->inode) : 0); + +out: + AuTraceErr(err); + return err; +} + +/* make sure the parent dir is fine */ +static int au_mvd_args_parent(const unsigned char dmsg, + struct au_mvd_args *a) +{ + int err; + aufs_bindex_t bindex; + + err = 0; + if (unlikely(au_alive_dir(a->parent))) { + err = -ENOENT; + AU_MVD_PR(dmsg, "parent dir is dead\n"); + goto out; + } + + a->bopq = au_dbdiropq(a->parent); + bindex = au_wbr_nonopq(a->dentry, a->mvd_bdst); + AuDbg("b%d\n", bindex); + if (unlikely((bindex >= 0 && bindex < a->mvd_bdst) + || (a->bopq != -1 && a->bopq < a->mvd_bdst))) { + err = -EINVAL; + a->mvd_errno = EAU_MVDOWN_OPAQUE; + AU_MVD_PR(dmsg, "ancestor is opaque b%d, b%d\n", + a->bopq, a->mvd_bdst); + } + +out: + AuTraceErr(err); + return err; +} + +static int au_mvd_args_intermediate(const unsigned char dmsg, + struct au_mvd_args *a) +{ + int err; + struct au_dinfo *dinfo, *tmp; + + /* lookup the next lower positive entry */ + err = -ENOMEM; + tmp = au_di_alloc(a->sb, AuLsc_DI_TMP); + if (unlikely(!tmp)) + goto out; + + a->bfound = -1; + a->bwh = -1; + dinfo = au_di(a->dentry); + au_di_cp(tmp, dinfo); + au_di_swap(tmp, dinfo); + + /* returns the number of positive dentries */ + err = au_lkup_dentry(a->dentry, a->mvd_bsrc + 1, + /* AuLkup_IGNORE_PERM */ 0); + if (!err) + a->bwh = au_dbwh(a->dentry); + else if (err > 0) + a->bfound = au_dbtop(a->dentry); + + au_di_swap(tmp, dinfo); + au_rw_write_unlock(&tmp->di_rwsem); + au_di_free(tmp); + if (unlikely(err < 0)) + AU_MVD_PR(dmsg, "failed look-up lower\n"); + + /* + * here, we have these cases. + * bfound == -1 + * no positive dentry under bsrc. there are more sub-cases. + * bwh < 0 + * there no whiteout, we can safely move-down. + * bwh <= bsrc + * impossible + * bsrc < bwh && bwh < bdst + * there is a whiteout on RO branch. cannot proceed. + * bwh == bdst + * there is a whiteout on the RW target branch. it should + * be removed. + * bdst < bwh + * there is a whiteout somewhere unrelated branch. + * -1 < bfound && bfound <= bsrc + * impossible. + * bfound < bdst + * found, but it is on RO branch between bsrc and bdst. cannot + * proceed. + * bfound == bdst + * found, replace it if AUFS_MVDOWN_FORCE is set. otherwise return + * error. + * bdst < bfound + * found, after we create the file on bdst, it will be hidden. + */ + + AuDebugOn(a->bfound == -1 + && a->bwh != -1 + && a->bwh <= a->mvd_bsrc); + AuDebugOn(-1 < a->bfound + && a->bfound <= a->mvd_bsrc); + + err = -EINVAL; + if (a->bfound == -1 + && a->mvd_bsrc < a->bwh + && a->bwh != -1 + && a->bwh < a->mvd_bdst) { + a->mvd_errno = EAU_MVDOWN_WHITEOUT; + AU_MVD_PR(dmsg, "bsrc %d, bdst %d, bfound %d, bwh %d\n", + a->mvd_bsrc, a->mvd_bdst, a->bfound, a->bwh); + goto out; + } else if (a->bfound != -1 && a->bfound < a->mvd_bdst) { + a->mvd_errno = EAU_MVDOWN_UPPER; + AU_MVD_PR(dmsg, "bdst %d, bfound %d\n", + a->mvd_bdst, a->bfound); + goto out; + } + + err = 0; /* success */ + +out: + AuTraceErr(err); + return err; +} + +static int au_mvd_args_exist(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + + err = 0; + if (!(a->mvdown.flags & AUFS_MVDOWN_OWLOWER) + && a->bfound == a->mvd_bdst) + err = -EEXIST; + AuTraceErr(err); + return err; +} + +static int au_mvd_args(const unsigned char dmsg, struct au_mvd_args *a) +{ + int err; + struct au_branch *br; + + err = -EISDIR; + if (unlikely(S_ISDIR(a->inode->i_mode))) + goto out; + + err = -EINVAL; + if (!(a->mvdown.flags & AUFS_MVDOWN_BRID_UPPER)) + a->mvd_bsrc = au_ibtop(a->inode); + else { + a->mvd_bsrc = au_br_index(a->sb, a->mvd_src_brid); + if (unlikely(a->mvd_bsrc < 0 + || (a->mvd_bsrc < au_dbtop(a->dentry) + || au_dbbot(a->dentry) < a->mvd_bsrc + || !au_h_dptr(a->dentry, a->mvd_bsrc)) + || (a->mvd_bsrc < au_ibtop(a->inode) + || au_ibbot(a->inode) < a->mvd_bsrc + || !au_h_iptr(a->inode, a->mvd_bsrc)))) { + a->mvd_errno = EAU_MVDOWN_NOUPPER; + AU_MVD_PR(dmsg, "no upper\n"); + goto out; + } + } + if (unlikely(a->mvd_bsrc == au_sbbot(a->sb))) { + a->mvd_errno = EAU_MVDOWN_BOTTOM; + AU_MVD_PR(dmsg, "on the bottom\n"); + goto out; + } + a->mvd_h_src_inode = au_h_iptr(a->inode, a->mvd_bsrc); + br = au_sbr(a->sb, a->mvd_bsrc); + err = au_br_rdonly(br); + if (!(a->mvdown.flags & AUFS_MVDOWN_ROUPPER)) { + if (unlikely(err)) + goto out; + } else if (!(vfsub_native_ro(a->mvd_h_src_inode) + || IS_APPEND(a->mvd_h_src_inode))) { + if (err) + a->mvdown.flags |= AUFS_MVDOWN_ROUPPER_R; + /* go on */ + } else + goto out; + + err = -EINVAL; + if (!(a->mvdown.flags & AUFS_MVDOWN_BRID_LOWER)) { + a->mvd_bdst = find_lower_writable(a); + if (unlikely(a->mvd_bdst < 0)) { + a->mvd_errno = EAU_MVDOWN_BOTTOM; + AU_MVD_PR(dmsg, "no writable lower branch\n"); + goto out; + } + } else { + a->mvd_bdst = au_br_index(a->sb, a->mvd_dst_brid); + if (unlikely(a->mvd_bdst < 0 + || au_sbbot(a->sb) < a->mvd_bdst)) { + a->mvd_errno = EAU_MVDOWN_NOLOWERBR; + AU_MVD_PR(dmsg, "no lower brid\n"); + goto out; + } + } + + err = au_mvd_args_busy(dmsg, a); + if (!err) + err = au_mvd_args_parent(dmsg, a); + if (!err) + err = au_mvd_args_intermediate(dmsg, a); + if (!err) + err = au_mvd_args_exist(dmsg, a); + if (!err) + AuDbg("b%d, b%d\n", a->mvd_bsrc, a->mvd_bdst); + +out: + AuTraceErr(err); + return err; +} + +int au_mvdown(struct dentry *dentry, struct aufs_mvdown __user *uarg) +{ + int err, e; + unsigned char dmsg; + struct au_mvd_args *args; + struct inode *inode; + + inode = d_inode(dentry); + err = -EPERM; + if (unlikely(!capable(CAP_SYS_ADMIN))) + goto out; + + err = -ENOMEM; + args = kmalloc(sizeof(*args), GFP_NOFS); + if (unlikely(!args)) + goto out; + + err = copy_from_user(&args->mvdown, uarg, sizeof(args->mvdown)); + if (!err) + /* VERIFY_WRITE */ + err = !access_ok(uarg, sizeof(*uarg)); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + goto out_free; + } + AuDbg("flags 0x%x\n", args->mvdown.flags); + args->mvdown.flags &= ~(AUFS_MVDOWN_ROLOWER_R | AUFS_MVDOWN_ROUPPER_R); + args->mvdown.au_errno = 0; + args->dentry = dentry; + args->inode = inode; + args->sb = dentry->d_sb; + + err = -ENOENT; + dmsg = !!(args->mvdown.flags & AUFS_MVDOWN_DMSG); + args->parent = dget_parent(dentry); + args->dir = d_inode(args->parent); + inode_lock_nested(args->dir, I_MUTEX_PARENT); + dput(args->parent); + if (unlikely(args->parent != dentry->d_parent)) { + AU_MVD_PR(dmsg, "parent dir is moved\n"); + goto out_dir; + } + + inode_lock_nested(inode, I_MUTEX_CHILD); + err = aufs_read_lock(dentry, AuLock_DW | AuLock_FLUSH | AuLock_NOPLMW); + if (unlikely(err)) + goto out_inode; + + di_write_lock_parent(args->parent); + err = au_mvd_args(dmsg, args); + if (unlikely(err)) + goto out_parent; + + err = au_do_mvdown(dmsg, args); + if (unlikely(err)) + goto out_parent; + + au_cpup_attr_timesizes(args->dir); + au_cpup_attr_timesizes(inode); + if (!(args->mvdown.flags & AUFS_MVDOWN_KUPPER)) + au_cpup_igen(inode, au_h_iptr(inode, args->mvd_bdst)); + /* au_digen_dec(dentry); */ + +out_parent: + di_write_unlock(args->parent); + aufs_read_unlock(dentry, AuLock_DW); +out_inode: + inode_unlock(inode); +out_dir: + inode_unlock(args->dir); +out_free: + e = copy_to_user(uarg, &args->mvdown, sizeof(args->mvdown)); + if (unlikely(e)) + err = -EFAULT; + au_kfree_rcu(args); +out: + AuTraceErr(err); + return err; +} diff --git a/fs/aufs/opts.c b/fs/aufs/opts.c new file mode 100644 index 00000000000..84646345116 --- /dev/null +++ b/fs/aufs/opts.c @@ -0,0 +1,1880 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * mount options/flags + */ + +#include +#include /* a distribution requires */ +#include +#include "aufs.h" + +/* ---------------------------------------------------------------------- */ + +enum { + Opt_br, + Opt_add, Opt_del, Opt_mod, Opt_append, Opt_prepend, + Opt_idel, Opt_imod, + Opt_dirwh, Opt_rdcache, Opt_rdblk, Opt_rdhash, + Opt_rdblk_def, Opt_rdhash_def, + Opt_xino, Opt_noxino, + Opt_trunc_xino, Opt_trunc_xino_v, Opt_notrunc_xino, + Opt_trunc_xino_path, Opt_itrunc_xino, + Opt_trunc_xib, Opt_notrunc_xib, + Opt_shwh, Opt_noshwh, + Opt_plink, Opt_noplink, Opt_list_plink, + Opt_udba, + Opt_dio, Opt_nodio, + Opt_diropq_a, Opt_diropq_w, + Opt_warn_perm, Opt_nowarn_perm, + Opt_wbr_copyup, Opt_wbr_create, + Opt_fhsm_sec, + Opt_verbose, Opt_noverbose, + Opt_sum, Opt_nosum, Opt_wsum, + Opt_dirperm1, Opt_nodirperm1, + Opt_dirren, Opt_nodirren, + Opt_acl, Opt_noacl, + Opt_tail, Opt_ignore, Opt_ignore_silent, Opt_err +}; + +static match_table_t options = { + {Opt_br, "br=%s"}, + {Opt_br, "br:%s"}, + + {Opt_add, "add=%d:%s"}, + {Opt_add, "add:%d:%s"}, + {Opt_add, "ins=%d:%s"}, + {Opt_add, "ins:%d:%s"}, + {Opt_append, "append=%s"}, + {Opt_append, "append:%s"}, + {Opt_prepend, "prepend=%s"}, + {Opt_prepend, "prepend:%s"}, + + {Opt_del, "del=%s"}, + {Opt_del, "del:%s"}, + /* {Opt_idel, "idel:%d"}, */ + {Opt_mod, "mod=%s"}, + {Opt_mod, "mod:%s"}, + /* {Opt_imod, "imod:%d:%s"}, */ + + {Opt_dirwh, "dirwh=%d"}, + + {Opt_xino, "xino=%s"}, + {Opt_noxino, "noxino"}, + {Opt_trunc_xino, "trunc_xino"}, + {Opt_trunc_xino_v, "trunc_xino_v=%d:%d"}, + {Opt_notrunc_xino, "notrunc_xino"}, + {Opt_trunc_xino_path, "trunc_xino=%s"}, + {Opt_itrunc_xino, "itrunc_xino=%d"}, + /* {Opt_zxino, "zxino=%s"}, */ + {Opt_trunc_xib, "trunc_xib"}, + {Opt_notrunc_xib, "notrunc_xib"}, + +#ifdef CONFIG_PROC_FS + {Opt_plink, "plink"}, +#else + {Opt_ignore_silent, "plink"}, +#endif + + {Opt_noplink, "noplink"}, + +#ifdef CONFIG_AUFS_DEBUG + {Opt_list_plink, "list_plink"}, +#endif + + {Opt_udba, "udba=%s"}, + + {Opt_dio, "dio"}, + {Opt_nodio, "nodio"}, + +#ifdef CONFIG_AUFS_DIRREN + {Opt_dirren, "dirren"}, + {Opt_nodirren, "nodirren"}, +#else + {Opt_ignore, "dirren"}, + {Opt_ignore_silent, "nodirren"}, +#endif + +#ifdef CONFIG_AUFS_FHSM + {Opt_fhsm_sec, "fhsm_sec=%d"}, +#else + {Opt_ignore, "fhsm_sec=%d"}, +#endif + + {Opt_diropq_a, "diropq=always"}, + {Opt_diropq_a, "diropq=a"}, + {Opt_diropq_w, "diropq=whiteouted"}, + {Opt_diropq_w, "diropq=w"}, + + {Opt_warn_perm, "warn_perm"}, + {Opt_nowarn_perm, "nowarn_perm"}, + + /* keep them temporary */ + {Opt_ignore_silent, "nodlgt"}, + {Opt_ignore, "clean_plink"}, + +#ifdef CONFIG_AUFS_SHWH + {Opt_shwh, "shwh"}, +#endif + {Opt_noshwh, "noshwh"}, + + {Opt_dirperm1, "dirperm1"}, + {Opt_nodirperm1, "nodirperm1"}, + + {Opt_verbose, "verbose"}, + {Opt_verbose, "v"}, + {Opt_noverbose, "noverbose"}, + {Opt_noverbose, "quiet"}, + {Opt_noverbose, "q"}, + {Opt_noverbose, "silent"}, + + {Opt_sum, "sum"}, + {Opt_nosum, "nosum"}, + {Opt_wsum, "wsum"}, + + {Opt_rdcache, "rdcache=%d"}, + {Opt_rdblk, "rdblk=%d"}, + {Opt_rdblk_def, "rdblk=def"}, + {Opt_rdhash, "rdhash=%d"}, + {Opt_rdhash_def, "rdhash=def"}, + + {Opt_wbr_create, "create=%s"}, + {Opt_wbr_create, "create_policy=%s"}, + {Opt_wbr_copyup, "cpup=%s"}, + {Opt_wbr_copyup, "copyup=%s"}, + {Opt_wbr_copyup, "copyup_policy=%s"}, + + /* generic VFS flag */ +#ifdef CONFIG_FS_POSIX_ACL + {Opt_acl, "acl"}, + {Opt_noacl, "noacl"}, +#else + {Opt_ignore, "acl"}, + {Opt_ignore_silent, "noacl"}, +#endif + + /* internal use for the scripts */ + {Opt_ignore_silent, "si=%s"}, + + {Opt_br, "dirs=%s"}, + {Opt_ignore, "debug=%d"}, + {Opt_ignore, "delete=whiteout"}, + {Opt_ignore, "delete=all"}, + {Opt_ignore, "imap=%s"}, + + /* temporary workaround, due to old mount(8)? */ + {Opt_ignore_silent, "relatime"}, + + {Opt_err, NULL} +}; + +/* ---------------------------------------------------------------------- */ + +static const char *au_parser_pattern(int val, match_table_t tbl) +{ + struct match_token *p; + + p = tbl; + while (p->pattern) { + if (p->token == val) + return p->pattern; + p++; + } + BUG(); + return "??"; +} + +static const char *au_optstr(int *val, match_table_t tbl) +{ + struct match_token *p; + int v; + + v = *val; + if (!v) + goto out; + p = tbl; + while (p->pattern) { + if (p->token + && (v & p->token) == p->token) { + *val &= ~p->token; + return p->pattern; + } + p++; + } + +out: + return NULL; +} + +/* ---------------------------------------------------------------------- */ + +static match_table_t brperm = { + {AuBrPerm_RO, AUFS_BRPERM_RO}, + {AuBrPerm_RR, AUFS_BRPERM_RR}, + {AuBrPerm_RW, AUFS_BRPERM_RW}, + {0, NULL} +}; + +static match_table_t brattr = { + /* general */ + {AuBrAttr_COO_REG, AUFS_BRATTR_COO_REG}, + {AuBrAttr_COO_ALL, AUFS_BRATTR_COO_ALL}, + /* 'unpin' attrib is meaningless since linux-3.18-rc1 */ + {AuBrAttr_UNPIN, AUFS_BRATTR_UNPIN}, +#ifdef CONFIG_AUFS_FHSM + {AuBrAttr_FHSM, AUFS_BRATTR_FHSM}, +#endif +#ifdef CONFIG_AUFS_XATTR + {AuBrAttr_ICEX, AUFS_BRATTR_ICEX}, + {AuBrAttr_ICEX_SEC, AUFS_BRATTR_ICEX_SEC}, + {AuBrAttr_ICEX_SYS, AUFS_BRATTR_ICEX_SYS}, + {AuBrAttr_ICEX_TR, AUFS_BRATTR_ICEX_TR}, + {AuBrAttr_ICEX_USR, AUFS_BRATTR_ICEX_USR}, + {AuBrAttr_ICEX_OTH, AUFS_BRATTR_ICEX_OTH}, +#endif + + /* ro/rr branch */ + {AuBrRAttr_WH, AUFS_BRRATTR_WH}, + + /* rw branch */ + {AuBrWAttr_MOO, AUFS_BRWATTR_MOO}, + {AuBrWAttr_NoLinkWH, AUFS_BRWATTR_NLWH}, + + {0, NULL} +}; + +static int br_attr_val(char *str, match_table_t table, substring_t args[]) +{ + int attr, v; + char *p; + + attr = 0; + do { + p = strchr(str, '+'); + if (p) + *p = 0; + v = match_token(str, table, args); + if (v) { + if (v & AuBrAttr_CMOO_Mask) + attr &= ~AuBrAttr_CMOO_Mask; + attr |= v; + } else { + if (p) + *p = '+'; + pr_warn("ignored branch attribute %s\n", str); + break; + } + if (p) + str = p + 1; + } while (p); + + return attr; +} + +static int au_do_optstr_br_attr(au_br_perm_str_t *str, int perm) +{ + int sz; + const char *p; + char *q; + + q = str->a; + *q = 0; + p = au_optstr(&perm, brattr); + if (p) { + sz = strlen(p); + memcpy(q, p, sz + 1); + q += sz; + } else + goto out; + + do { + p = au_optstr(&perm, brattr); + if (p) { + *q++ = '+'; + sz = strlen(p); + memcpy(q, p, sz + 1); + q += sz; + } + } while (p); + +out: + return q - str->a; +} + +static int noinline_for_stack br_perm_val(char *perm) +{ + int val, bad, sz; + char *p; + substring_t args[MAX_OPT_ARGS]; + au_br_perm_str_t attr; + + p = strchr(perm, '+'); + if (p) + *p = 0; + val = match_token(perm, brperm, args); + if (!val) { + if (p) + *p = '+'; + pr_warn("ignored branch permission %s\n", perm); + val = AuBrPerm_RO; + goto out; + } + if (!p) + goto out; + + val |= br_attr_val(p + 1, brattr, args); + + bad = 0; + switch (val & AuBrPerm_Mask) { + case AuBrPerm_RO: + case AuBrPerm_RR: + bad = val & AuBrWAttr_Mask; + val &= ~AuBrWAttr_Mask; + break; + case AuBrPerm_RW: + bad = val & AuBrRAttr_Mask; + val &= ~AuBrRAttr_Mask; + break; + } + + /* + * 'unpin' attrib becomes meaningless since linux-3.18-rc1, but aufs + * does not treat it as an error, just warning. + * this is a tiny guard for the user operation. + */ + if (val & AuBrAttr_UNPIN) { + bad |= AuBrAttr_UNPIN; + val &= ~AuBrAttr_UNPIN; + } + + if (unlikely(bad)) { + sz = au_do_optstr_br_attr(&attr, bad); + AuDebugOn(!sz); + pr_warn("ignored branch attribute %s\n", attr.a); + } + +out: + return val; +} + +void au_optstr_br_perm(au_br_perm_str_t *str, int perm) +{ + au_br_perm_str_t attr; + const char *p; + char *q; + int sz; + + q = str->a; + p = au_optstr(&perm, brperm); + AuDebugOn(!p || !*p); + sz = strlen(p); + memcpy(q, p, sz + 1); + q += sz; + + sz = au_do_optstr_br_attr(&attr, perm); + if (sz) { + *q++ = '+'; + memcpy(q, attr.a, sz + 1); + } + + AuDebugOn(strlen(str->a) >= sizeof(str->a)); +} + +/* ---------------------------------------------------------------------- */ + +static match_table_t udbalevel = { + {AuOpt_UDBA_REVAL, "reval"}, + {AuOpt_UDBA_NONE, "none"}, +#ifdef CONFIG_AUFS_HNOTIFY + {AuOpt_UDBA_HNOTIFY, "notify"}, /* abstraction */ +#ifdef CONFIG_AUFS_HFSNOTIFY + {AuOpt_UDBA_HNOTIFY, "fsnotify"}, +#endif +#endif + {-1, NULL} +}; + +static int noinline_for_stack udba_val(char *str) +{ + substring_t args[MAX_OPT_ARGS]; + + return match_token(str, udbalevel, args); +} + +const char *au_optstr_udba(int udba) +{ + return au_parser_pattern(udba, udbalevel); +} + +/* ---------------------------------------------------------------------- */ + +static match_table_t au_wbr_create_policy = { + {AuWbrCreate_TDP, "tdp"}, + {AuWbrCreate_TDP, "top-down-parent"}, + {AuWbrCreate_RR, "rr"}, + {AuWbrCreate_RR, "round-robin"}, + {AuWbrCreate_MFS, "mfs"}, + {AuWbrCreate_MFS, "most-free-space"}, + {AuWbrCreate_MFSV, "mfs:%d"}, + {AuWbrCreate_MFSV, "most-free-space:%d"}, + + /* top-down regardless the parent, and then mfs */ + {AuWbrCreate_TDMFS, "tdmfs:%d"}, + {AuWbrCreate_TDMFSV, "tdmfs:%d:%d"}, + + {AuWbrCreate_MFSRR, "mfsrr:%d"}, + {AuWbrCreate_MFSRRV, "mfsrr:%d:%d"}, + {AuWbrCreate_PMFS, "pmfs"}, + {AuWbrCreate_PMFSV, "pmfs:%d"}, + {AuWbrCreate_PMFSRR, "pmfsrr:%d"}, + {AuWbrCreate_PMFSRRV, "pmfsrr:%d:%d"}, + + {-1, NULL} +}; + +static int au_wbr_mfs_wmark(substring_t *arg, char *str, + struct au_opt_wbr_create *create) +{ + int err; + unsigned long long ull; + + err = 0; + if (!match_u64(arg, &ull)) + create->mfsrr_watermark = ull; + else { + pr_err("bad integer in %s\n", str); + err = -EINVAL; + } + + return err; +} + +static int au_wbr_mfs_sec(substring_t *arg, char *str, + struct au_opt_wbr_create *create) +{ + int n, err; + + err = 0; + if (!match_int(arg, &n) && 0 <= n && n <= AUFS_MFS_MAX_SEC) + create->mfs_second = n; + else { + pr_err("bad integer in %s\n", str); + err = -EINVAL; + } + + return err; +} + +static int noinline_for_stack +au_wbr_create_val(char *str, struct au_opt_wbr_create *create) +{ + int err, e; + substring_t args[MAX_OPT_ARGS]; + + err = match_token(str, au_wbr_create_policy, args); + create->wbr_create = err; + switch (err) { + case AuWbrCreate_MFSRRV: + case AuWbrCreate_TDMFSV: + case AuWbrCreate_PMFSRRV: + e = au_wbr_mfs_wmark(&args[0], str, create); + if (!e) + e = au_wbr_mfs_sec(&args[1], str, create); + if (unlikely(e)) + err = e; + break; + case AuWbrCreate_MFSRR: + case AuWbrCreate_TDMFS: + case AuWbrCreate_PMFSRR: + e = au_wbr_mfs_wmark(&args[0], str, create); + if (unlikely(e)) { + err = e; + break; + } + fallthrough; + case AuWbrCreate_MFS: + case AuWbrCreate_PMFS: + create->mfs_second = AUFS_MFS_DEF_SEC; + break; + case AuWbrCreate_MFSV: + case AuWbrCreate_PMFSV: + e = au_wbr_mfs_sec(&args[0], str, create); + if (unlikely(e)) + err = e; + break; + } + + return err; +} + +const char *au_optstr_wbr_create(int wbr_create) +{ + return au_parser_pattern(wbr_create, au_wbr_create_policy); +} + +static match_table_t au_wbr_copyup_policy = { + {AuWbrCopyup_TDP, "tdp"}, + {AuWbrCopyup_TDP, "top-down-parent"}, + {AuWbrCopyup_BUP, "bup"}, + {AuWbrCopyup_BUP, "bottom-up-parent"}, + {AuWbrCopyup_BU, "bu"}, + {AuWbrCopyup_BU, "bottom-up"}, + {-1, NULL} +}; + +static int noinline_for_stack au_wbr_copyup_val(char *str) +{ + substring_t args[MAX_OPT_ARGS]; + + return match_token(str, au_wbr_copyup_policy, args); +} + +const char *au_optstr_wbr_copyup(int wbr_copyup) +{ + return au_parser_pattern(wbr_copyup, au_wbr_copyup_policy); +} + +/* ---------------------------------------------------------------------- */ + +static const int lkup_dirflags = LOOKUP_FOLLOW | LOOKUP_DIRECTORY; + +static void dump_opts(struct au_opts *opts) +{ +#ifdef CONFIG_AUFS_DEBUG + /* reduce stack space */ + union { + struct au_opt_add *add; + struct au_opt_del *del; + struct au_opt_mod *mod; + struct au_opt_xino *xino; + struct au_opt_xino_itrunc *xino_itrunc; + struct au_opt_wbr_create *create; + } u; + struct au_opt *opt; + + opt = opts->opt; + while (opt->type != Opt_tail) { + switch (opt->type) { + case Opt_add: + u.add = &opt->add; + AuDbg("add {b%d, %s, 0x%x, %p}\n", + u.add->bindex, u.add->pathname, u.add->perm, + u.add->path.dentry); + break; + case Opt_del: + case Opt_idel: + u.del = &opt->del; + AuDbg("del {%s, %p}\n", + u.del->pathname, u.del->h_path.dentry); + break; + case Opt_mod: + case Opt_imod: + u.mod = &opt->mod; + AuDbg("mod {%s, 0x%x, %p}\n", + u.mod->path, u.mod->perm, u.mod->h_root); + break; + case Opt_append: + u.add = &opt->add; + AuDbg("append {b%d, %s, 0x%x, %p}\n", + u.add->bindex, u.add->pathname, u.add->perm, + u.add->path.dentry); + break; + case Opt_prepend: + u.add = &opt->add; + AuDbg("prepend {b%d, %s, 0x%x, %p}\n", + u.add->bindex, u.add->pathname, u.add->perm, + u.add->path.dentry); + break; + case Opt_dirwh: + AuDbg("dirwh %d\n", opt->dirwh); + break; + case Opt_rdcache: + AuDbg("rdcache %d\n", opt->rdcache); + break; + case Opt_rdblk: + AuDbg("rdblk %u\n", opt->rdblk); + break; + case Opt_rdblk_def: + AuDbg("rdblk_def\n"); + break; + case Opt_rdhash: + AuDbg("rdhash %u\n", opt->rdhash); + break; + case Opt_rdhash_def: + AuDbg("rdhash_def\n"); + break; + case Opt_xino: + u.xino = &opt->xino; + AuDbg("xino {%s %pD}\n", u.xino->path, u.xino->file); + break; + case Opt_trunc_xino: + AuLabel(trunc_xino); + break; + case Opt_notrunc_xino: + AuLabel(notrunc_xino); + break; + case Opt_trunc_xino_path: + case Opt_itrunc_xino: + u.xino_itrunc = &opt->xino_itrunc; + AuDbg("trunc_xino %d\n", u.xino_itrunc->bindex); + break; + case Opt_noxino: + AuLabel(noxino); + break; + case Opt_trunc_xib: + AuLabel(trunc_xib); + break; + case Opt_notrunc_xib: + AuLabel(notrunc_xib); + break; + case Opt_shwh: + AuLabel(shwh); + break; + case Opt_noshwh: + AuLabel(noshwh); + break; + case Opt_dirperm1: + AuLabel(dirperm1); + break; + case Opt_nodirperm1: + AuLabel(nodirperm1); + break; + case Opt_plink: + AuLabel(plink); + break; + case Opt_noplink: + AuLabel(noplink); + break; + case Opt_list_plink: + AuLabel(list_plink); + break; + case Opt_udba: + AuDbg("udba %d, %s\n", + opt->udba, au_optstr_udba(opt->udba)); + break; + case Opt_dio: + AuLabel(dio); + break; + case Opt_nodio: + AuLabel(nodio); + break; + case Opt_diropq_a: + AuLabel(diropq_a); + break; + case Opt_diropq_w: + AuLabel(diropq_w); + break; + case Opt_warn_perm: + AuLabel(warn_perm); + break; + case Opt_nowarn_perm: + AuLabel(nowarn_perm); + break; + case Opt_verbose: + AuLabel(verbose); + break; + case Opt_noverbose: + AuLabel(noverbose); + break; + case Opt_sum: + AuLabel(sum); + break; + case Opt_nosum: + AuLabel(nosum); + break; + case Opt_wsum: + AuLabel(wsum); + break; + case Opt_wbr_create: + u.create = &opt->wbr_create; + AuDbg("create %d, %s\n", u.create->wbr_create, + au_optstr_wbr_create(u.create->wbr_create)); + switch (u.create->wbr_create) { + case AuWbrCreate_MFSV: + case AuWbrCreate_PMFSV: + AuDbg("%d sec\n", u.create->mfs_second); + break; + case AuWbrCreate_MFSRR: + case AuWbrCreate_TDMFS: + AuDbg("%llu watermark\n", + u.create->mfsrr_watermark); + break; + case AuWbrCreate_MFSRRV: + case AuWbrCreate_TDMFSV: + case AuWbrCreate_PMFSRRV: + AuDbg("%llu watermark, %d sec\n", + u.create->mfsrr_watermark, + u.create->mfs_second); + break; + } + break; + case Opt_wbr_copyup: + AuDbg("copyup %d, %s\n", opt->wbr_copyup, + au_optstr_wbr_copyup(opt->wbr_copyup)); + break; + case Opt_fhsm_sec: + AuDbg("fhsm_sec %u\n", opt->fhsm_second); + break; + case Opt_dirren: + AuLabel(dirren); + break; + case Opt_nodirren: + AuLabel(nodirren); + break; + case Opt_acl: + AuLabel(acl); + break; + case Opt_noacl: + AuLabel(noacl); + break; + default: + BUG(); + } + opt++; + } +#endif +} + +void au_opts_free(struct au_opts *opts) +{ + struct au_opt *opt; + + opt = opts->opt; + while (opt->type != Opt_tail) { + switch (opt->type) { + case Opt_add: + case Opt_append: + case Opt_prepend: + path_put(&opt->add.path); + break; + case Opt_del: + case Opt_idel: + path_put(&opt->del.h_path); + break; + case Opt_mod: + case Opt_imod: + dput(opt->mod.h_root); + break; + case Opt_xino: + fput(opt->xino.file); + break; + } + opt++; + } +} + +static int opt_add(struct au_opt *opt, char *opt_str, unsigned long sb_flags, + aufs_bindex_t bindex) +{ + int err; + struct au_opt_add *add = &opt->add; + char *p; + + add->bindex = bindex; + add->perm = AuBrPerm_RO; + add->pathname = opt_str; + p = strchr(opt_str, '='); + if (p) { + *p++ = 0; + if (*p) + add->perm = br_perm_val(p); + } + + err = vfsub_kern_path(add->pathname, lkup_dirflags, &add->path); + if (!err) { + if (!p) { + add->perm = AuBrPerm_RO; + if (au_test_fs_rr(add->path.dentry->d_sb)) + add->perm = AuBrPerm_RR; + else if (!bindex && !(sb_flags & SB_RDONLY)) + add->perm = AuBrPerm_RW; + } + opt->type = Opt_add; + goto out; + } + pr_err("lookup failed %s (%d)\n", add->pathname, err); + err = -EINVAL; + +out: + return err; +} + +static int au_opts_parse_del(struct au_opt_del *del, substring_t args[]) +{ + int err; + + del->pathname = args[0].from; + AuDbg("del path %s\n", del->pathname); + + err = vfsub_kern_path(del->pathname, lkup_dirflags, &del->h_path); + if (unlikely(err)) + pr_err("lookup failed %s (%d)\n", del->pathname, err); + + return err; +} + +#if 0 /* reserved for future use */ +static int au_opts_parse_idel(struct super_block *sb, aufs_bindex_t bindex, + struct au_opt_del *del, substring_t args[]) +{ + int err; + struct dentry *root; + + err = -EINVAL; + root = sb->s_root; + aufs_read_lock(root, AuLock_FLUSH); + if (bindex < 0 || au_sbbot(sb) < bindex) { + pr_err("out of bounds, %d\n", bindex); + goto out; + } + + err = 0; + del->h_path.dentry = dget(au_h_dptr(root, bindex)); + del->h_path.mnt = mntget(au_sbr_mnt(sb, bindex)); + +out: + aufs_read_unlock(root, !AuLock_IR); + return err; +} +#endif + +static int noinline_for_stack +au_opts_parse_mod(struct au_opt_mod *mod, substring_t args[]) +{ + int err; + struct path path; + char *p; + + err = -EINVAL; + mod->path = args[0].from; + p = strchr(mod->path, '='); + if (unlikely(!p)) { + pr_err("no permission %s\n", args[0].from); + goto out; + } + + *p++ = 0; + err = vfsub_kern_path(mod->path, lkup_dirflags, &path); + if (unlikely(err)) { + pr_err("lookup failed %s (%d)\n", mod->path, err); + goto out; + } + + mod->perm = br_perm_val(p); + AuDbg("mod path %s, perm 0x%x, %s\n", mod->path, mod->perm, p); + mod->h_root = dget(path.dentry); + path_put(&path); + +out: + return err; +} + +#if 0 /* reserved for future use */ +static int au_opts_parse_imod(struct super_block *sb, aufs_bindex_t bindex, + struct au_opt_mod *mod, substring_t args[]) +{ + int err; + struct dentry *root; + + err = -EINVAL; + root = sb->s_root; + aufs_read_lock(root, AuLock_FLUSH); + if (bindex < 0 || au_sbbot(sb) < bindex) { + pr_err("out of bounds, %d\n", bindex); + goto out; + } + + err = 0; + mod->perm = br_perm_val(args[1].from); + AuDbg("mod path %s, perm 0x%x, %s\n", + mod->path, mod->perm, args[1].from); + mod->h_root = dget(au_h_dptr(root, bindex)); + +out: + aufs_read_unlock(root, !AuLock_IR); + return err; +} +#endif + +static int au_opts_parse_xino(struct super_block *sb, struct au_opt_xino *xino, + substring_t args[]) +{ + int err; + struct file *file; + + file = au_xino_create(sb, args[0].from, /*silent*/0, /*wbrtop*/0); + err = PTR_ERR(file); + if (IS_ERR(file)) + goto out; + + err = -EINVAL; + if (unlikely(file->f_path.dentry->d_sb == sb)) { + fput(file); + pr_err("%s must be outside\n", args[0].from); + goto out; + } + + err = 0; + xino->file = file; + xino->path = args[0].from; + +out: + return err; +} + +static int noinline_for_stack +au_opts_parse_xino_itrunc_path(struct super_block *sb, + struct au_opt_xino_itrunc *xino_itrunc, + substring_t args[]) +{ + int err; + aufs_bindex_t bbot, bindex; + struct path path; + struct dentry *root; + + err = vfsub_kern_path(args[0].from, lkup_dirflags, &path); + if (unlikely(err)) { + pr_err("lookup failed %s (%d)\n", args[0].from, err); + goto out; + } + + xino_itrunc->bindex = -1; + root = sb->s_root; + aufs_read_lock(root, AuLock_FLUSH); + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + if (au_h_dptr(root, bindex) == path.dentry) { + xino_itrunc->bindex = bindex; + break; + } + } + aufs_read_unlock(root, !AuLock_IR); + path_put(&path); + + if (unlikely(xino_itrunc->bindex < 0)) { + pr_err("no such branch %s\n", args[0].from); + err = -EINVAL; + } + +out: + return err; +} + +/* called without aufs lock */ +int au_opts_parse(struct super_block *sb, char *str, struct au_opts *opts) +{ + int err, n, token; + aufs_bindex_t bindex; + unsigned char skipped; + struct dentry *root; + struct au_opt *opt, *opt_tail; + char *opt_str; + /* reduce the stack space */ + union { + struct au_opt_xino_itrunc *xino_itrunc; + struct au_opt_wbr_create *create; + } u; + struct { + substring_t args[MAX_OPT_ARGS]; + } *a; + + err = -ENOMEM; + a = kmalloc(sizeof(*a), GFP_NOFS); + if (unlikely(!a)) + goto out; + + root = sb->s_root; + err = 0; + bindex = 0; + opt = opts->opt; + opt_tail = opt + opts->max_opt - 1; + opt->type = Opt_tail; + while (!err && (opt_str = strsep(&str, ",")) && *opt_str) { + err = -EINVAL; + skipped = 0; + token = match_token(opt_str, options, a->args); + switch (token) { + case Opt_br: + err = 0; + while (!err && (opt_str = strsep(&a->args[0].from, ":")) + && *opt_str) { + err = opt_add(opt, opt_str, opts->sb_flags, + bindex++); + if (unlikely(!err && ++opt > opt_tail)) { + err = -E2BIG; + break; + } + opt->type = Opt_tail; + skipped = 1; + } + break; + case Opt_add: + if (unlikely(match_int(&a->args[0], &n))) { + pr_err("bad integer in %s\n", opt_str); + break; + } + bindex = n; + err = opt_add(opt, a->args[1].from, opts->sb_flags, + bindex); + if (!err) + opt->type = token; + break; + case Opt_append: + err = opt_add(opt, a->args[0].from, opts->sb_flags, + /*dummy bindex*/1); + if (!err) + opt->type = token; + break; + case Opt_prepend: + err = opt_add(opt, a->args[0].from, opts->sb_flags, + /*bindex*/0); + if (!err) + opt->type = token; + break; + case Opt_del: + err = au_opts_parse_del(&opt->del, a->args); + if (!err) + opt->type = token; + break; +#if 0 /* reserved for future use */ + case Opt_idel: + del->pathname = "(indexed)"; + if (unlikely(match_int(&args[0], &n))) { + pr_err("bad integer in %s\n", opt_str); + break; + } + err = au_opts_parse_idel(sb, n, &opt->del, a->args); + if (!err) + opt->type = token; + break; +#endif + case Opt_mod: + err = au_opts_parse_mod(&opt->mod, a->args); + if (!err) + opt->type = token; + break; +#ifdef IMOD /* reserved for future use */ + case Opt_imod: + u.mod->path = "(indexed)"; + if (unlikely(match_int(&a->args[0], &n))) { + pr_err("bad integer in %s\n", opt_str); + break; + } + err = au_opts_parse_imod(sb, n, &opt->mod, a->args); + if (!err) + opt->type = token; + break; +#endif + case Opt_xino: + err = au_opts_parse_xino(sb, &opt->xino, a->args); + if (!err) + opt->type = token; + break; + + case Opt_trunc_xino_path: + err = au_opts_parse_xino_itrunc_path + (sb, &opt->xino_itrunc, a->args); + if (!err) + opt->type = token; + break; + + case Opt_itrunc_xino: + u.xino_itrunc = &opt->xino_itrunc; + if (unlikely(match_int(&a->args[0], &n))) { + pr_err("bad integer in %s\n", opt_str); + break; + } + u.xino_itrunc->bindex = n; + aufs_read_lock(root, AuLock_FLUSH); + if (n < 0 || au_sbbot(sb) < n) { + pr_err("out of bounds, %d\n", n); + aufs_read_unlock(root, !AuLock_IR); + break; + } + aufs_read_unlock(root, !AuLock_IR); + err = 0; + opt->type = token; + break; + + case Opt_dirwh: + if (unlikely(match_int(&a->args[0], &opt->dirwh))) + break; + err = 0; + opt->type = token; + break; + + case Opt_rdcache: + if (unlikely(match_int(&a->args[0], &n))) { + pr_err("bad integer in %s\n", opt_str); + break; + } + if (unlikely(n > AUFS_RDCACHE_MAX)) { + pr_err("rdcache must be smaller than %d\n", + AUFS_RDCACHE_MAX); + break; + } + opt->rdcache = n; + err = 0; + opt->type = token; + break; + case Opt_rdblk: + if (unlikely(match_int(&a->args[0], &n) + || n < 0 + || n > KMALLOC_MAX_SIZE)) { + pr_err("bad integer in %s\n", opt_str); + break; + } + if (unlikely(n && n < NAME_MAX)) { + pr_err("rdblk must be larger than %d\n", + NAME_MAX); + break; + } + opt->rdblk = n; + err = 0; + opt->type = token; + break; + case Opt_rdhash: + if (unlikely(match_int(&a->args[0], &n) + || n < 0 + || n * sizeof(struct hlist_head) + > KMALLOC_MAX_SIZE)) { + pr_err("bad integer in %s\n", opt_str); + break; + } + opt->rdhash = n; + err = 0; + opt->type = token; + break; + + case Opt_trunc_xino: + case Opt_notrunc_xino: + case Opt_noxino: + case Opt_trunc_xib: + case Opt_notrunc_xib: + case Opt_shwh: + case Opt_noshwh: + case Opt_dirperm1: + case Opt_nodirperm1: + case Opt_plink: + case Opt_noplink: + case Opt_list_plink: + case Opt_dio: + case Opt_nodio: + case Opt_diropq_a: + case Opt_diropq_w: + case Opt_warn_perm: + case Opt_nowarn_perm: + case Opt_verbose: + case Opt_noverbose: + case Opt_sum: + case Opt_nosum: + case Opt_wsum: + case Opt_rdblk_def: + case Opt_rdhash_def: + case Opt_dirren: + case Opt_nodirren: + case Opt_acl: + case Opt_noacl: + err = 0; + opt->type = token; + break; + + case Opt_udba: + opt->udba = udba_val(a->args[0].from); + if (opt->udba >= 0) { + err = 0; + opt->type = token; + } else + pr_err("wrong value, %s\n", opt_str); + break; + + case Opt_wbr_create: + u.create = &opt->wbr_create; + u.create->wbr_create + = au_wbr_create_val(a->args[0].from, u.create); + if (u.create->wbr_create >= 0) { + err = 0; + opt->type = token; + } else + pr_err("wrong value, %s\n", opt_str); + break; + case Opt_wbr_copyup: + opt->wbr_copyup = au_wbr_copyup_val(a->args[0].from); + if (opt->wbr_copyup >= 0) { + err = 0; + opt->type = token; + } else + pr_err("wrong value, %s\n", opt_str); + break; + + case Opt_fhsm_sec: + if (unlikely(match_int(&a->args[0], &n) + || n < 0)) { + pr_err("bad integer in %s\n", opt_str); + break; + } + if (sysaufs_brs) { + opt->fhsm_second = n; + opt->type = token; + } else + pr_warn("ignored %s\n", opt_str); + err = 0; + break; + + case Opt_ignore: + pr_warn("ignored %s\n", opt_str); + fallthrough; + case Opt_ignore_silent: + skipped = 1; + err = 0; + break; + case Opt_err: + pr_err("unknown option %s\n", opt_str); + break; + } + + if (!err && !skipped) { + if (unlikely(++opt > opt_tail)) { + err = -E2BIG; + opt--; + opt->type = Opt_tail; + break; + } + opt->type = Opt_tail; + } + } + + au_kfree_rcu(a); + dump_opts(opts); + if (unlikely(err)) + au_opts_free(opts); + +out: + return err; +} + +static int au_opt_wbr_create(struct super_block *sb, + struct au_opt_wbr_create *create) +{ + int err; + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + err = 1; /* handled */ + sbinfo = au_sbi(sb); + if (sbinfo->si_wbr_create_ops->fin) { + err = sbinfo->si_wbr_create_ops->fin(sb); + if (!err) + err = 1; + } + + sbinfo->si_wbr_create = create->wbr_create; + sbinfo->si_wbr_create_ops = au_wbr_create_ops + create->wbr_create; + switch (create->wbr_create) { + case AuWbrCreate_MFSRRV: + case AuWbrCreate_MFSRR: + case AuWbrCreate_TDMFS: + case AuWbrCreate_TDMFSV: + case AuWbrCreate_PMFSRR: + case AuWbrCreate_PMFSRRV: + sbinfo->si_wbr_mfs.mfsrr_watermark = create->mfsrr_watermark; + fallthrough; + case AuWbrCreate_MFS: + case AuWbrCreate_MFSV: + case AuWbrCreate_PMFS: + case AuWbrCreate_PMFSV: + sbinfo->si_wbr_mfs.mfs_expire + = msecs_to_jiffies(create->mfs_second * MSEC_PER_SEC); + break; + } + + if (sbinfo->si_wbr_create_ops->init) + sbinfo->si_wbr_create_ops->init(sb); /* ignore */ + + return err; +} + +/* + * returns, + * plus: processed without an error + * zero: unprocessed + */ +static int au_opt_simple(struct super_block *sb, struct au_opt *opt, + struct au_opts *opts) +{ + int err; + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + err = 1; /* handled */ + sbinfo = au_sbi(sb); + switch (opt->type) { + case Opt_udba: + sbinfo->si_mntflags &= ~AuOptMask_UDBA; + sbinfo->si_mntflags |= opt->udba; + opts->given_udba |= opt->udba; + break; + + case Opt_plink: + au_opt_set(sbinfo->si_mntflags, PLINK); + break; + case Opt_noplink: + if (au_opt_test(sbinfo->si_mntflags, PLINK)) + au_plink_put(sb, /*verbose*/1); + au_opt_clr(sbinfo->si_mntflags, PLINK); + break; + case Opt_list_plink: + if (au_opt_test(sbinfo->si_mntflags, PLINK)) + au_plink_list(sb); + break; + + case Opt_dio: + au_opt_set(sbinfo->si_mntflags, DIO); + au_fset_opts(opts->flags, REFRESH_DYAOP); + break; + case Opt_nodio: + au_opt_clr(sbinfo->si_mntflags, DIO); + au_fset_opts(opts->flags, REFRESH_DYAOP); + break; + + case Opt_fhsm_sec: + au_fhsm_set(sbinfo, opt->fhsm_second); + break; + + case Opt_diropq_a: + au_opt_set(sbinfo->si_mntflags, ALWAYS_DIROPQ); + break; + case Opt_diropq_w: + au_opt_clr(sbinfo->si_mntflags, ALWAYS_DIROPQ); + break; + + case Opt_warn_perm: + au_opt_set(sbinfo->si_mntflags, WARN_PERM); + break; + case Opt_nowarn_perm: + au_opt_clr(sbinfo->si_mntflags, WARN_PERM); + break; + + case Opt_verbose: + au_opt_set(sbinfo->si_mntflags, VERBOSE); + break; + case Opt_noverbose: + au_opt_clr(sbinfo->si_mntflags, VERBOSE); + break; + + case Opt_sum: + au_opt_set(sbinfo->si_mntflags, SUM); + break; + case Opt_wsum: + au_opt_clr(sbinfo->si_mntflags, SUM); + au_opt_set(sbinfo->si_mntflags, SUM_W); + break; + case Opt_nosum: + au_opt_clr(sbinfo->si_mntflags, SUM); + au_opt_clr(sbinfo->si_mntflags, SUM_W); + break; + + case Opt_wbr_create: + err = au_opt_wbr_create(sb, &opt->wbr_create); + break; + case Opt_wbr_copyup: + sbinfo->si_wbr_copyup = opt->wbr_copyup; + sbinfo->si_wbr_copyup_ops = au_wbr_copyup_ops + opt->wbr_copyup; + break; + + case Opt_dirwh: + sbinfo->si_dirwh = opt->dirwh; + break; + + case Opt_rdcache: + sbinfo->si_rdcache + = msecs_to_jiffies(opt->rdcache * MSEC_PER_SEC); + break; + case Opt_rdblk: + sbinfo->si_rdblk = opt->rdblk; + break; + case Opt_rdblk_def: + sbinfo->si_rdblk = AUFS_RDBLK_DEF; + break; + case Opt_rdhash: + sbinfo->si_rdhash = opt->rdhash; + break; + case Opt_rdhash_def: + sbinfo->si_rdhash = AUFS_RDHASH_DEF; + break; + + case Opt_shwh: + au_opt_set(sbinfo->si_mntflags, SHWH); + break; + case Opt_noshwh: + au_opt_clr(sbinfo->si_mntflags, SHWH); + break; + + case Opt_dirperm1: + au_opt_set(sbinfo->si_mntflags, DIRPERM1); + break; + case Opt_nodirperm1: + au_opt_clr(sbinfo->si_mntflags, DIRPERM1); + break; + + case Opt_trunc_xino: + au_opt_set(sbinfo->si_mntflags, TRUNC_XINO); + break; + case Opt_notrunc_xino: + au_opt_clr(sbinfo->si_mntflags, TRUNC_XINO); + break; + + case Opt_trunc_xino_path: + case Opt_itrunc_xino: + err = au_xino_trunc(sb, opt->xino_itrunc.bindex, + /*idx_begin*/0); + if (!err) + err = 1; + break; + + case Opt_trunc_xib: + au_fset_opts(opts->flags, TRUNC_XIB); + break; + case Opt_notrunc_xib: + au_fclr_opts(opts->flags, TRUNC_XIB); + break; + + case Opt_dirren: + err = 1; + if (!au_opt_test(sbinfo->si_mntflags, DIRREN)) { + err = au_dr_opt_set(sb); + if (!err) + err = 1; + } + if (err == 1) + au_opt_set(sbinfo->si_mntflags, DIRREN); + break; + case Opt_nodirren: + err = 1; + if (au_opt_test(sbinfo->si_mntflags, DIRREN)) { + err = au_dr_opt_clr(sb, au_ftest_opts(opts->flags, + DR_FLUSHED)); + if (!err) + err = 1; + } + if (err == 1) + au_opt_clr(sbinfo->si_mntflags, DIRREN); + break; + + case Opt_acl: + sb->s_flags |= SB_POSIXACL; + break; + case Opt_noacl: + sb->s_flags &= ~SB_POSIXACL; + break; + + default: + err = 0; + break; + } + + return err; +} + +/* + * returns tri-state. + * plus: processed without an error + * zero: unprocessed + * minus: error + */ +static int au_opt_br(struct super_block *sb, struct au_opt *opt, + struct au_opts *opts) +{ + int err, do_refresh; + + err = 0; + switch (opt->type) { + case Opt_append: + opt->add.bindex = au_sbbot(sb) + 1; + if (opt->add.bindex < 0) + opt->add.bindex = 0; + goto add; + /* Always goto add, not fallthrough */ + case Opt_prepend: + opt->add.bindex = 0; + fallthrough; + add: /* indented label */ + case Opt_add: + err = au_br_add(sb, &opt->add, + au_ftest_opts(opts->flags, REMOUNT)); + if (!err) { + err = 1; + au_fset_opts(opts->flags, REFRESH); + } + break; + + case Opt_del: + case Opt_idel: + err = au_br_del(sb, &opt->del, + au_ftest_opts(opts->flags, REMOUNT)); + if (!err) { + err = 1; + au_fset_opts(opts->flags, TRUNC_XIB); + au_fset_opts(opts->flags, REFRESH); + } + break; + + case Opt_mod: + case Opt_imod: + err = au_br_mod(sb, &opt->mod, + au_ftest_opts(opts->flags, REMOUNT), + &do_refresh); + if (!err) { + err = 1; + if (do_refresh) + au_fset_opts(opts->flags, REFRESH); + } + break; + } + return err; +} + +static int au_opt_xino(struct super_block *sb, struct au_opt *opt, + struct au_opt_xino **opt_xino, + struct au_opts *opts) +{ + int err; + + err = 0; + switch (opt->type) { + case Opt_xino: + err = au_xino_set(sb, &opt->xino, + !!au_ftest_opts(opts->flags, REMOUNT)); + if (unlikely(err)) + break; + + *opt_xino = &opt->xino; + break; + + case Opt_noxino: + au_xino_clr(sb); + *opt_xino = (void *)-1; + break; + } + + return err; +} + +int au_opts_verify(struct super_block *sb, unsigned long sb_flags, + unsigned int pending) +{ + int err, fhsm; + aufs_bindex_t bindex, bbot; + unsigned char do_plink, skip, do_free, can_no_dreval; + struct au_branch *br; + struct au_wbr *wbr; + struct dentry *root, *dentry; + struct inode *dir, *h_dir; + struct au_sbinfo *sbinfo; + struct au_hinode *hdir; + + SiMustAnyLock(sb); + + sbinfo = au_sbi(sb); + AuDebugOn(!(sbinfo->si_mntflags & AuOptMask_UDBA)); + + if (!(sb_flags & SB_RDONLY)) { + if (unlikely(!au_br_writable(au_sbr_perm(sb, 0)))) + pr_warn("first branch should be rw\n"); + if (unlikely(au_opt_test(sbinfo->si_mntflags, SHWH))) + pr_warn_once("shwh should be used with ro\n"); + } + + if (au_opt_test((sbinfo->si_mntflags | pending), UDBA_HNOTIFY) + && !au_opt_test(sbinfo->si_mntflags, XINO)) + pr_warn_once("udba=*notify requires xino\n"); + + if (au_opt_test(sbinfo->si_mntflags, DIRPERM1)) + pr_warn_once("dirperm1 breaks the protection" + " by the permission bits on the lower branch\n"); + + err = 0; + fhsm = 0; + root = sb->s_root; + dir = d_inode(root); + do_plink = !!au_opt_test(sbinfo->si_mntflags, PLINK); + can_no_dreval = !!au_opt_test((sbinfo->si_mntflags | pending), + UDBA_NONE); + bbot = au_sbbot(sb); + for (bindex = 0; !err && bindex <= bbot; bindex++) { + skip = 0; + h_dir = au_h_iptr(dir, bindex); + br = au_sbr(sb, bindex); + + if ((br->br_perm & AuBrAttr_ICEX) + && !h_dir->i_op->listxattr) + br->br_perm &= ~AuBrAttr_ICEX; +#if 0 /* untested */ + if ((br->br_perm & AuBrAttr_ICEX_SEC) + && (au_br_sb(br)->s_flags & SB_NOSEC)) + br->br_perm &= ~AuBrAttr_ICEX_SEC; +#endif + + do_free = 0; + wbr = br->br_wbr; + if (wbr) + wbr_wh_read_lock(wbr); + + if (!au_br_writable(br->br_perm)) { + do_free = !!wbr; + skip = (!wbr + || (!wbr->wbr_whbase + && !wbr->wbr_plink + && !wbr->wbr_orph)); + } else if (!au_br_wh_linkable(br->br_perm)) { + /* skip = (!br->br_whbase && !br->br_orph); */ + skip = (!wbr || !wbr->wbr_whbase); + if (skip && wbr) { + if (do_plink) + skip = !!wbr->wbr_plink; + else + skip = !wbr->wbr_plink; + } + } else { + /* skip = (br->br_whbase && br->br_ohph); */ + skip = (wbr && wbr->wbr_whbase); + if (skip) { + if (do_plink) + skip = !!wbr->wbr_plink; + else + skip = !wbr->wbr_plink; + } + } + if (wbr) + wbr_wh_read_unlock(wbr); + + if (can_no_dreval) { + dentry = br->br_path.dentry; + spin_lock(&dentry->d_lock); + if (dentry->d_flags & + (DCACHE_OP_REVALIDATE | DCACHE_OP_WEAK_REVALIDATE)) + can_no_dreval = 0; + spin_unlock(&dentry->d_lock); + } + + if (au_br_fhsm(br->br_perm)) { + fhsm++; + AuDebugOn(!br->br_fhsm); + } + + if (skip) + continue; + + hdir = au_hi(dir, bindex); + au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT); + if (wbr) + wbr_wh_write_lock(wbr); + err = au_wh_init(br, sb); + if (wbr) + wbr_wh_write_unlock(wbr); + au_hn_inode_unlock(hdir); + + if (!err && do_free) { + au_kfree_rcu(wbr); + br->br_wbr = NULL; + } + } + + if (can_no_dreval) + au_fset_si(sbinfo, NO_DREVAL); + else + au_fclr_si(sbinfo, NO_DREVAL); + + if (fhsm >= 2) { + au_fset_si(sbinfo, FHSM); + for (bindex = bbot; bindex >= 0; bindex--) { + br = au_sbr(sb, bindex); + if (au_br_fhsm(br->br_perm)) { + au_fhsm_set_bottom(sb, bindex); + break; + } + } + } else { + au_fclr_si(sbinfo, FHSM); + au_fhsm_set_bottom(sb, -1); + } + + return err; +} + +int au_opts_mount(struct super_block *sb, struct au_opts *opts) +{ + int err; + unsigned int tmp; + aufs_bindex_t bindex, bbot; + struct au_opt *opt; + struct au_opt_xino *opt_xino, xino; + struct au_sbinfo *sbinfo; + struct au_branch *br; + struct inode *dir; + + SiMustWriteLock(sb); + + err = 0; + opt_xino = NULL; + opt = opts->opt; + while (err >= 0 && opt->type != Opt_tail) + err = au_opt_simple(sb, opt++, opts); + if (err > 0) + err = 0; + else if (unlikely(err < 0)) + goto out; + + /* disable xino and udba temporary */ + sbinfo = au_sbi(sb); + tmp = sbinfo->si_mntflags; + au_opt_clr(sbinfo->si_mntflags, XINO); + au_opt_set_udba(sbinfo->si_mntflags, UDBA_REVAL); + + opt = opts->opt; + while (err >= 0 && opt->type != Opt_tail) + err = au_opt_br(sb, opt++, opts); + if (err > 0) + err = 0; + else if (unlikely(err < 0)) + goto out; + + bbot = au_sbbot(sb); + if (unlikely(bbot < 0)) { + err = -EINVAL; + pr_err("no branches\n"); + goto out; + } + + if (au_opt_test(tmp, XINO)) + au_opt_set(sbinfo->si_mntflags, XINO); + opt = opts->opt; + while (!err && opt->type != Opt_tail) + err = au_opt_xino(sb, opt++, &opt_xino, opts); + if (unlikely(err)) + goto out; + + err = au_opts_verify(sb, sb->s_flags, tmp); + if (unlikely(err)) + goto out; + + /* restore xino */ + if (au_opt_test(tmp, XINO) && !opt_xino) { + xino.file = au_xino_def(sb); + err = PTR_ERR(xino.file); + if (IS_ERR(xino.file)) + goto out; + + err = au_xino_set(sb, &xino, /*remount*/0); + fput(xino.file); + if (unlikely(err)) + goto out; + } + + /* restore udba */ + tmp &= AuOptMask_UDBA; + sbinfo->si_mntflags &= ~AuOptMask_UDBA; + sbinfo->si_mntflags |= tmp; + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + err = au_hnotify_reset_br(tmp, br, br->br_perm); + if (unlikely(err)) + AuIOErr("hnotify failed on br %d, %d, ignored\n", + bindex, err); + /* go on even if err */ + } + if (au_opt_test(tmp, UDBA_HNOTIFY)) { + dir = d_inode(sb->s_root); + au_hn_reset(dir, au_hi_flags(dir, /*isdir*/1) & ~AuHi_XINO); + } + +out: + return err; +} + +int au_opts_remount(struct super_block *sb, struct au_opts *opts) +{ + int err, rerr; + unsigned char no_dreval; + struct inode *dir; + struct au_opt_xino *opt_xino; + struct au_opt *opt; + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + err = au_dr_opt_flush(sb); + if (unlikely(err)) + goto out; + au_fset_opts(opts->flags, DR_FLUSHED); + + dir = d_inode(sb->s_root); + sbinfo = au_sbi(sb); + opt_xino = NULL; + opt = opts->opt; + while (err >= 0 && opt->type != Opt_tail) { + err = au_opt_simple(sb, opt, opts); + if (!err) + err = au_opt_br(sb, opt, opts); + if (!err) + err = au_opt_xino(sb, opt, &opt_xino, opts); + opt++; + } + if (err > 0) + err = 0; + AuTraceErr(err); + /* go on even err */ + + no_dreval = !!au_ftest_si(sbinfo, NO_DREVAL); + rerr = au_opts_verify(sb, opts->sb_flags, /*pending*/0); + if (unlikely(rerr && !err)) + err = rerr; + + if (no_dreval != !!au_ftest_si(sbinfo, NO_DREVAL)) + au_fset_opts(opts->flags, REFRESH_IDOP); + + if (au_ftest_opts(opts->flags, TRUNC_XIB)) { + rerr = au_xib_trunc(sb); + if (unlikely(rerr && !err)) + err = rerr; + } + + /* will be handled by the caller */ + if (!au_ftest_opts(opts->flags, REFRESH) + && (opts->given_udba + || au_opt_test(sbinfo->si_mntflags, XINO) + || au_ftest_opts(opts->flags, REFRESH_IDOP) + )) + au_fset_opts(opts->flags, REFRESH); + + AuDbg("status 0x%x\n", opts->flags); + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +unsigned int au_opt_udba(struct super_block *sb) +{ + return au_mntflags(sb) & AuOptMask_UDBA; +} diff --git a/fs/aufs/opts.h b/fs/aufs/opts.h new file mode 100644 index 00000000000..79f3ea56f8b --- /dev/null +++ b/fs/aufs/opts.h @@ -0,0 +1,225 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * mount options/flags + */ + +#ifndef __AUFS_OPTS_H__ +#define __AUFS_OPTS_H__ + +#ifdef __KERNEL__ + +#include + +struct file; + +/* ---------------------------------------------------------------------- */ + +/* mount flags */ +#define AuOpt_XINO 1 /* external inode number bitmap + and translation table */ +#define AuOpt_TRUNC_XINO (1 << 1) /* truncate xino files */ +#define AuOpt_UDBA_NONE (1 << 2) /* users direct branch access */ +#define AuOpt_UDBA_REVAL (1 << 3) +#define AuOpt_UDBA_HNOTIFY (1 << 4) +#define AuOpt_SHWH (1 << 5) /* show whiteout */ +#define AuOpt_PLINK (1 << 6) /* pseudo-link */ +#define AuOpt_DIRPERM1 (1 << 7) /* ignore the lower dir's perm + bits */ +#define AuOpt_ALWAYS_DIROPQ (1 << 9) /* policy to creating diropq */ +#define AuOpt_SUM (1 << 10) /* summation for statfs(2) */ +#define AuOpt_SUM_W (1 << 11) /* unimplemented */ +#define AuOpt_WARN_PERM (1 << 12) /* warn when add-branch */ +#define AuOpt_VERBOSE (1 << 13) /* print the cause of error */ +#define AuOpt_DIO (1 << 14) /* direct io */ +#define AuOpt_DIRREN (1 << 15) /* directory rename */ + +#ifndef CONFIG_AUFS_HNOTIFY +#undef AuOpt_UDBA_HNOTIFY +#define AuOpt_UDBA_HNOTIFY 0 +#endif +#ifndef CONFIG_AUFS_DIRREN +#undef AuOpt_DIRREN +#define AuOpt_DIRREN 0 +#endif +#ifndef CONFIG_AUFS_SHWH +#undef AuOpt_SHWH +#define AuOpt_SHWH 0 +#endif + +#define AuOpt_Def (AuOpt_XINO \ + | AuOpt_UDBA_REVAL \ + | AuOpt_PLINK \ + /* | AuOpt_DIRPERM1 */ \ + | AuOpt_WARN_PERM) +#define AuOptMask_UDBA (AuOpt_UDBA_NONE \ + | AuOpt_UDBA_REVAL \ + | AuOpt_UDBA_HNOTIFY) + +#define au_opt_test(flags, name) (flags & AuOpt_##name) +#define au_opt_set(flags, name) do { \ + BUILD_BUG_ON(AuOpt_##name & AuOptMask_UDBA); \ + ((flags) |= AuOpt_##name); \ +} while (0) +#define au_opt_set_udba(flags, name) do { \ + (flags) &= ~AuOptMask_UDBA; \ + ((flags) |= AuOpt_##name); \ +} while (0) +#define au_opt_clr(flags, name) do { \ + ((flags) &= ~AuOpt_##name); \ +} while (0) + +static inline unsigned int au_opts_plink(unsigned int mntflags) +{ +#ifdef CONFIG_PROC_FS + return mntflags; +#else + return mntflags & ~AuOpt_PLINK; +#endif +} + +/* ---------------------------------------------------------------------- */ + +/* policies to select one among multiple writable branches */ +enum { + AuWbrCreate_TDP, /* top down parent */ + AuWbrCreate_RR, /* round robin */ + AuWbrCreate_MFS, /* most free space */ + AuWbrCreate_MFSV, /* mfs with seconds */ + AuWbrCreate_MFSRR, /* mfs then rr */ + AuWbrCreate_MFSRRV, /* mfs then rr with seconds */ + AuWbrCreate_TDMFS, /* top down regardless parent and mfs */ + AuWbrCreate_TDMFSV, /* top down regardless parent and mfs */ + AuWbrCreate_PMFS, /* parent and mfs */ + AuWbrCreate_PMFSV, /* parent and mfs with seconds */ + AuWbrCreate_PMFSRR, /* parent, mfs and round-robin */ + AuWbrCreate_PMFSRRV, /* plus seconds */ + + AuWbrCreate_Def = AuWbrCreate_TDP +}; + +enum { + AuWbrCopyup_TDP, /* top down parent */ + AuWbrCopyup_BUP, /* bottom up parent */ + AuWbrCopyup_BU, /* bottom up */ + + AuWbrCopyup_Def = AuWbrCopyup_TDP +}; + +/* ---------------------------------------------------------------------- */ + +struct au_opt_add { + aufs_bindex_t bindex; + char *pathname; + int perm; + struct path path; +}; + +struct au_opt_del { + char *pathname; + struct path h_path; +}; + +struct au_opt_mod { + char *path; + int perm; + struct dentry *h_root; +}; + +struct au_opt_xino { + char *path; + struct file *file; +}; + +struct au_opt_xino_itrunc { + aufs_bindex_t bindex; +}; + +struct au_opt_wbr_create { + int wbr_create; + int mfs_second; + unsigned long long mfsrr_watermark; +}; + +struct au_opt { + int type; + union { + struct au_opt_xino xino; + struct au_opt_xino_itrunc xino_itrunc; + struct au_opt_add add; + struct au_opt_del del; + struct au_opt_mod mod; + int dirwh; + int rdcache; + unsigned int rdblk; + unsigned int rdhash; + int udba; + struct au_opt_wbr_create wbr_create; + int wbr_copyup; + unsigned int fhsm_second; + }; +}; + +/* opts flags */ +#define AuOpts_REMOUNT 1 +#define AuOpts_REFRESH (1 << 1) +#define AuOpts_TRUNC_XIB (1 << 2) +#define AuOpts_REFRESH_DYAOP (1 << 3) +#define AuOpts_REFRESH_IDOP (1 << 4) +#define AuOpts_DR_FLUSHED (1 << 5) +#define au_ftest_opts(flags, name) ((flags) & AuOpts_##name) +#define au_fset_opts(flags, name) \ + do { (flags) |= AuOpts_##name; } while (0) +#define au_fclr_opts(flags, name) \ + do { (flags) &= ~AuOpts_##name; } while (0) + +#ifndef CONFIG_AUFS_DIRREN +#undef AuOpts_DR_FLUSHED +#define AuOpts_DR_FLUSHED 0 +#endif + +struct au_opts { + struct au_opt *opt; + int max_opt; + + unsigned int given_udba; + unsigned int flags; + unsigned long sb_flags; +}; + +/* ---------------------------------------------------------------------- */ + +/* opts.c */ +void au_optstr_br_perm(au_br_perm_str_t *str, int perm); +const char *au_optstr_udba(int udba); +const char *au_optstr_wbr_copyup(int wbr_copyup); +const char *au_optstr_wbr_create(int wbr_create); + +void au_opts_free(struct au_opts *opts); +struct super_block; +int au_opts_parse(struct super_block *sb, char *str, struct au_opts *opts); +int au_opts_verify(struct super_block *sb, unsigned long sb_flags, + unsigned int pending); +int au_opts_mount(struct super_block *sb, struct au_opts *opts); +int au_opts_remount(struct super_block *sb, struct au_opts *opts); + +unsigned int au_opt_udba(struct super_block *sb); + +#endif /* __KERNEL__ */ +#endif /* __AUFS_OPTS_H__ */ diff --git a/fs/aufs/plink.c b/fs/aufs/plink.c new file mode 100644 index 00000000000..850829b2188 --- /dev/null +++ b/fs/aufs/plink.c @@ -0,0 +1,516 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * pseudo-link + */ + +#include "aufs.h" + +/* + * the pseudo-link maintenance mode. + * during a user process maintains the pseudo-links, + * prohibit adding a new plink and branch manipulation. + * + * Flags + * NOPLM: + * For entry functions which will handle plink, and i_mutex is already held + * in VFS. + * They cannot wait and should return an error at once. + * Callers has to check the error. + * NOPLMW: + * For entry functions which will handle plink, but i_mutex is not held + * in VFS. + * They can wait the plink maintenance mode to finish. + * + * They behave like F_SETLK and F_SETLKW. + * If the caller never handle plink, then both flags are unnecessary. + */ + +int au_plink_maint(struct super_block *sb, int flags) +{ + int err; + pid_t pid, ppid; + struct task_struct *parent, *prev; + struct au_sbinfo *sbi; + + SiMustAnyLock(sb); + + err = 0; + if (!au_opt_test(au_mntflags(sb), PLINK)) + goto out; + + sbi = au_sbi(sb); + pid = sbi->si_plink_maint_pid; + if (!pid || pid == current->pid) + goto out; + + /* todo: it highly depends upon /sbin/mount.aufs */ + prev = NULL; + parent = current; + ppid = 0; + rcu_read_lock(); + while (1) { + parent = rcu_dereference(parent->real_parent); + if (parent == prev) + break; + ppid = task_pid_vnr(parent); + if (pid == ppid) { + rcu_read_unlock(); + goto out; + } + prev = parent; + } + rcu_read_unlock(); + + if (au_ftest_lock(flags, NOPLMW)) { + /* if there is no i_mutex lock in VFS, we don't need to wait */ + /* AuDebugOn(!lockdep_depth(current)); */ + while (sbi->si_plink_maint_pid) { + si_read_unlock(sb); + /* gave up wake_up_bit() */ + wait_event(sbi->si_plink_wq, !sbi->si_plink_maint_pid); + + if (au_ftest_lock(flags, FLUSH)) + au_nwt_flush(&sbi->si_nowait); + si_noflush_read_lock(sb); + } + } else if (au_ftest_lock(flags, NOPLM)) { + AuDbg("ppid %d, pid %d\n", ppid, pid); + err = -EAGAIN; + } + +out: + return err; +} + +void au_plink_maint_leave(struct au_sbinfo *sbinfo) +{ + spin_lock(&sbinfo->si_plink_maint_lock); + sbinfo->si_plink_maint_pid = 0; + spin_unlock(&sbinfo->si_plink_maint_lock); + wake_up_all(&sbinfo->si_plink_wq); +} + +int au_plink_maint_enter(struct super_block *sb) +{ + int err; + struct au_sbinfo *sbinfo; + + err = 0; + sbinfo = au_sbi(sb); + /* make sure i am the only one in this fs */ + si_write_lock(sb, AuLock_FLUSH); + if (au_opt_test(au_mntflags(sb), PLINK)) { + spin_lock(&sbinfo->si_plink_maint_lock); + if (!sbinfo->si_plink_maint_pid) + sbinfo->si_plink_maint_pid = current->pid; + else + err = -EBUSY; + spin_unlock(&sbinfo->si_plink_maint_lock); + } + si_write_unlock(sb); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_DEBUG +void au_plink_list(struct super_block *sb) +{ + int i; + struct au_sbinfo *sbinfo; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_icntnr *icntnr; + + SiMustAnyLock(sb); + + sbinfo = au_sbi(sb); + AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); + AuDebugOn(au_plink_maint(sb, AuLock_NOPLM)); + + for (i = 0; i < AuPlink_NHASH; i++) { + hbl = sbinfo->si_plink + i; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(icntnr, pos, hbl, plink) + AuDbg("%lu\n", icntnr->vfs_inode.i_ino); + hlist_bl_unlock(hbl); + } +} +#endif + +/* is the inode pseudo-linked? */ +int au_plink_test(struct inode *inode) +{ + int found, i; + struct au_sbinfo *sbinfo; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_icntnr *icntnr; + + sbinfo = au_sbi(inode->i_sb); + AuRwMustAnyLock(&sbinfo->si_rwsem); + AuDebugOn(!au_opt_test(au_mntflags(inode->i_sb), PLINK)); + AuDebugOn(au_plink_maint(inode->i_sb, AuLock_NOPLM)); + + found = 0; + i = au_plink_hash(inode->i_ino); + hbl = sbinfo->si_plink + i; + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(icntnr, pos, hbl, plink) + if (&icntnr->vfs_inode == inode) { + found = 1; + break; + } + hlist_bl_unlock(hbl); + return found; +} + +/* ---------------------------------------------------------------------- */ + +/* + * generate a name for plink. + * the file will be stored under AUFS_WH_PLINKDIR. + */ +/* 20 is max digits length of ulong 64 */ +#define PLINK_NAME_LEN ((20 + 1) * 2) + +static int plink_name(char *name, int len, struct inode *inode, + aufs_bindex_t bindex) +{ + int rlen; + struct inode *h_inode; + + h_inode = au_h_iptr(inode, bindex); + rlen = snprintf(name, len, "%lu.%lu", inode->i_ino, h_inode->i_ino); + return rlen; +} + +struct au_do_plink_lkup_args { + struct dentry **errp; + struct qstr *tgtname; + struct dentry *h_parent; + struct au_branch *br; +}; + +static struct dentry *au_do_plink_lkup(struct qstr *tgtname, + struct dentry *h_parent, + struct au_branch *br) +{ + struct dentry *h_dentry; + struct inode *h_inode; + + h_inode = d_inode(h_parent); + inode_lock_shared_nested(h_inode, AuLsc_I_CHILD2); + h_dentry = vfsub_lkup_one(tgtname, h_parent); + inode_unlock_shared(h_inode); + return h_dentry; +} + +static void au_call_do_plink_lkup(void *args) +{ + struct au_do_plink_lkup_args *a = args; + *a->errp = au_do_plink_lkup(a->tgtname, a->h_parent, a->br); +} + +/* lookup the plink-ed @inode under the branch at @bindex */ +struct dentry *au_plink_lkup(struct inode *inode, aufs_bindex_t bindex) +{ + struct dentry *h_dentry, *h_parent; + struct au_branch *br; + int wkq_err; + char a[PLINK_NAME_LEN]; + struct qstr tgtname = QSTR_INIT(a, 0); + + AuDebugOn(au_plink_maint(inode->i_sb, AuLock_NOPLM)); + + br = au_sbr(inode->i_sb, bindex); + h_parent = br->br_wbr->wbr_plink; + tgtname.len = plink_name(a, sizeof(a), inode, bindex); + + if (!uid_eq(current_fsuid(), GLOBAL_ROOT_UID)) { + struct au_do_plink_lkup_args args = { + .errp = &h_dentry, + .tgtname = &tgtname, + .h_parent = h_parent, + .br = br + }; + + wkq_err = au_wkq_wait(au_call_do_plink_lkup, &args); + if (unlikely(wkq_err)) + h_dentry = ERR_PTR(wkq_err); + } else + h_dentry = au_do_plink_lkup(&tgtname, h_parent, br); + + return h_dentry; +} + +/* create a pseudo-link */ +static int do_whplink(struct qstr *tgt, struct dentry *h_parent, + struct dentry *h_dentry, struct au_branch *br) +{ + int err; + struct path h_path = { + .mnt = au_br_mnt(br) + }; + struct inode *h_dir, *delegated; + + h_dir = d_inode(h_parent); + inode_lock_nested(h_dir, AuLsc_I_CHILD2); +again: + h_path.dentry = vfsub_lkup_one(tgt, h_parent); + err = PTR_ERR(h_path.dentry); + if (IS_ERR(h_path.dentry)) + goto out; + + err = 0; + /* wh.plink dir is not monitored */ + /* todo: is it really safe? */ + if (d_is_positive(h_path.dentry) + && d_inode(h_path.dentry) != d_inode(h_dentry)) { + delegated = NULL; + err = vfsub_unlink(h_dir, &h_path, &delegated, /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + dput(h_path.dentry); + h_path.dentry = NULL; + if (!err) + goto again; + } + if (!err && d_is_negative(h_path.dentry)) { + delegated = NULL; + err = vfsub_link(h_dentry, h_dir, &h_path, &delegated); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + } + dput(h_path.dentry); + +out: + inode_unlock(h_dir); + return err; +} + +struct do_whplink_args { + int *errp; + struct qstr *tgt; + struct dentry *h_parent; + struct dentry *h_dentry; + struct au_branch *br; +}; + +static void call_do_whplink(void *args) +{ + struct do_whplink_args *a = args; + *a->errp = do_whplink(a->tgt, a->h_parent, a->h_dentry, a->br); +} + +static int whplink(struct dentry *h_dentry, struct inode *inode, + aufs_bindex_t bindex, struct au_branch *br) +{ + int err, wkq_err; + struct au_wbr *wbr; + struct dentry *h_parent; + char a[PLINK_NAME_LEN]; + struct qstr tgtname = QSTR_INIT(a, 0); + + wbr = au_sbr(inode->i_sb, bindex)->br_wbr; + h_parent = wbr->wbr_plink; + tgtname.len = plink_name(a, sizeof(a), inode, bindex); + + /* always superio. */ + if (!uid_eq(current_fsuid(), GLOBAL_ROOT_UID)) { + struct do_whplink_args args = { + .errp = &err, + .tgt = &tgtname, + .h_parent = h_parent, + .h_dentry = h_dentry, + .br = br + }; + wkq_err = au_wkq_wait(call_do_whplink, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } else + err = do_whplink(&tgtname, h_parent, h_dentry, br); + + return err; +} + +/* + * create a new pseudo-link for @h_dentry on @bindex. + * the linked inode is held in aufs @inode. + */ +void au_plink_append(struct inode *inode, aufs_bindex_t bindex, + struct dentry *h_dentry) +{ + struct super_block *sb; + struct au_sbinfo *sbinfo; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_icntnr *icntnr; + int found, err, cnt, i; + + sb = inode->i_sb; + sbinfo = au_sbi(sb); + AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); + AuDebugOn(au_plink_maint(sb, AuLock_NOPLM)); + + found = au_plink_test(inode); + if (found) + return; + + i = au_plink_hash(inode->i_ino); + hbl = sbinfo->si_plink + i; + au_igrab(inode); + + hlist_bl_lock(hbl); + hlist_bl_for_each_entry(icntnr, pos, hbl, plink) { + if (&icntnr->vfs_inode == inode) { + found = 1; + break; + } + } + if (!found) { + icntnr = container_of(inode, struct au_icntnr, vfs_inode); + hlist_bl_add_head(&icntnr->plink, hbl); + } + hlist_bl_unlock(hbl); + if (!found) { + cnt = au_hbl_count(hbl); +#define msg "unexpectedly unbalanced or too many pseudo-links" + if (cnt > AUFS_PLINK_WARN) + AuWarn1(msg ", %d\n", cnt); +#undef msg + err = whplink(h_dentry, inode, bindex, au_sbr(sb, bindex)); + if (unlikely(err)) { + pr_warn("err %d, damaged pseudo link.\n", err); + au_hbl_del(&icntnr->plink, hbl); + iput(&icntnr->vfs_inode); + } + } else + iput(&icntnr->vfs_inode); +} + +/* free all plinks */ +void au_plink_put(struct super_block *sb, int verbose) +{ + int i, warned; + struct au_sbinfo *sbinfo; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos, *tmp; + struct au_icntnr *icntnr; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); + AuDebugOn(au_plink_maint(sb, AuLock_NOPLM)); + + /* no spin_lock since sbinfo is write-locked */ + warned = 0; + for (i = 0; i < AuPlink_NHASH; i++) { + hbl = sbinfo->si_plink + i; + if (!warned && verbose && !hlist_bl_empty(hbl)) { + pr_warn("pseudo-link is not flushed"); + warned = 1; + } + hlist_bl_for_each_entry_safe(icntnr, pos, tmp, hbl, plink) + iput(&icntnr->vfs_inode); + INIT_HLIST_BL_HEAD(hbl); + } +} + +void au_plink_clean(struct super_block *sb, int verbose) +{ + struct dentry *root; + + root = sb->s_root; + aufs_write_lock(root); + if (au_opt_test(au_mntflags(sb), PLINK)) + au_plink_put(sb, verbose); + aufs_write_unlock(root); +} + +static int au_plink_do_half_refresh(struct inode *inode, aufs_bindex_t br_id) +{ + int do_put; + aufs_bindex_t btop, bbot, bindex; + + do_put = 0; + btop = au_ibtop(inode); + bbot = au_ibbot(inode); + if (btop >= 0) { + for (bindex = btop; bindex <= bbot; bindex++) { + if (!au_h_iptr(inode, bindex) + || au_ii_br_id(inode, bindex) != br_id) + continue; + au_set_h_iptr(inode, bindex, NULL, 0); + do_put = 1; + break; + } + if (do_put) + for (bindex = btop; bindex <= bbot; bindex++) + if (au_h_iptr(inode, bindex)) { + do_put = 0; + break; + } + } else + do_put = 1; + + return do_put; +} + +/* free the plinks on a branch specified by @br_id */ +void au_plink_half_refresh(struct super_block *sb, aufs_bindex_t br_id) +{ + struct au_sbinfo *sbinfo; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos, *tmp; + struct au_icntnr *icntnr; + struct inode *inode; + int i, do_put; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + AuDebugOn(!au_opt_test(au_mntflags(sb), PLINK)); + AuDebugOn(au_plink_maint(sb, AuLock_NOPLM)); + + /* no bit_lock since sbinfo is write-locked */ + for (i = 0; i < AuPlink_NHASH; i++) { + hbl = sbinfo->si_plink + i; + hlist_bl_for_each_entry_safe(icntnr, pos, tmp, hbl, plink) { + inode = au_igrab(&icntnr->vfs_inode); + ii_write_lock_child(inode); + do_put = au_plink_do_half_refresh(inode, br_id); + if (do_put) { + hlist_bl_del(&icntnr->plink); + iput(inode); + } + ii_write_unlock(inode); + iput(inode); + } + } +} diff --git a/fs/aufs/poll.c b/fs/aufs/poll.c new file mode 100644 index 00000000000..6975d3ab2d2 --- /dev/null +++ b/fs/aufs/poll.c @@ -0,0 +1,51 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * poll operation + * There is only one filesystem which implements ->poll operation, currently. + */ + +#include "aufs.h" + +__poll_t aufs_poll(struct file *file, struct poll_table_struct *pt) +{ + __poll_t mask; + struct file *h_file; + struct super_block *sb; + + /* We should pretend an error happened. */ + mask = EPOLLERR /* | EPOLLIN | EPOLLOUT */; + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLMW); + + h_file = au_read_pre(file, /*keep_fi*/0, /*lsc*/0); + if (IS_ERR(h_file)) { + AuDbg("h_file %ld\n", PTR_ERR(h_file)); + goto out; + } + + mask = vfs_poll(h_file, pt); + fput(h_file); /* instead of au_read_post() */ + +out: + si_read_unlock(sb); + if (mask & EPOLLERR) + AuDbg("mask 0x%x\n", mask); + return mask; +} diff --git a/fs/aufs/posix_acl.c b/fs/aufs/posix_acl.c new file mode 100644 index 00000000000..198196f6e79 --- /dev/null +++ b/fs/aufs/posix_acl.c @@ -0,0 +1,106 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2014-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * posix acl operations + */ + +#include +#include "aufs.h" + +struct posix_acl *aufs_get_acl(struct inode *inode, int type) +{ + struct posix_acl *acl; + int err; + aufs_bindex_t bindex; + struct inode *h_inode; + struct super_block *sb; + + acl = NULL; + sb = inode->i_sb; + si_read_lock(sb, AuLock_FLUSH); + ii_read_lock_child(inode); + if (!(sb->s_flags & SB_POSIXACL)) + goto out; + + bindex = au_ibtop(inode); + h_inode = au_h_iptr(inode, bindex); + if (unlikely(!h_inode + || ((h_inode->i_mode & S_IFMT) + != (inode->i_mode & S_IFMT)))) { + err = au_busy_or_stale(); + acl = ERR_PTR(err); + goto out; + } + + /* always topmost only */ + acl = get_acl(h_inode, type); + if (IS_ERR(acl)) + forget_cached_acl(inode, type); + else + set_cached_acl(inode, type, acl); + +out: + ii_read_unlock(inode); + si_read_unlock(sb); + + AuTraceErrPtr(acl); + return acl; +} + +int aufs_set_acl(struct user_namespace *userns, struct inode *inode, + struct posix_acl *acl, int type) +{ + int err; + ssize_t ssz; + struct dentry *dentry; + struct au_sxattr arg = { + .type = AU_ACL_SET, + .u.acl_set = { + .acl = acl, + .type = type + }, + }; + + IMustLock(inode); + + if (inode->i_ino == AUFS_ROOT_INO) + dentry = dget(inode->i_sb->s_root); + else { + dentry = d_find_alias(inode); + if (!dentry) + dentry = d_find_any_alias(inode); + if (!dentry) { + pr_warn("cannot handle this inode, " + "please report to aufs-users ML\n"); + err = -ENOENT; + goto out; + } + } + + ssz = au_sxattr(dentry, inode, &arg); + /* forget even it if succeeds since the branch might set differently */ + forget_cached_acl(inode, type); + dput(dentry); + err = ssz; + if (ssz >= 0) + err = 0; + +out: + return err; +} diff --git a/fs/aufs/procfs.c b/fs/aufs/procfs.c new file mode 100644 index 00000000000..50ee0a9c1bc --- /dev/null +++ b/fs/aufs/procfs.c @@ -0,0 +1,170 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2010-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * procfs interfaces + */ + +#include +#include "aufs.h" + +static int au_procfs_plm_release(struct inode *inode, struct file *file) +{ + struct au_sbinfo *sbinfo; + + sbinfo = file->private_data; + if (sbinfo) { + au_plink_maint_leave(sbinfo); + kobject_put(&sbinfo->si_kobj); + } + + return 0; +} + +static void au_procfs_plm_write_clean(struct file *file) +{ + struct au_sbinfo *sbinfo; + + sbinfo = file->private_data; + if (sbinfo) + au_plink_clean(sbinfo->si_sb, /*verbose*/0); +} + +static int au_procfs_plm_write_si(struct file *file, unsigned long id) +{ + int err; + struct super_block *sb; + struct au_sbinfo *sbinfo; + struct hlist_bl_node *pos; + + err = -EBUSY; + if (unlikely(file->private_data)) + goto out; + + sb = NULL; + /* don't use au_sbilist_lock() here */ + hlist_bl_lock(&au_sbilist); + hlist_bl_for_each_entry(sbinfo, pos, &au_sbilist, si_list) + if (id == sysaufs_si_id(sbinfo)) { + if (kobject_get_unless_zero(&sbinfo->si_kobj)) + sb = sbinfo->si_sb; + break; + } + hlist_bl_unlock(&au_sbilist); + + err = -EINVAL; + if (unlikely(!sb)) + goto out; + + err = au_plink_maint_enter(sb); + if (!err) + /* keep kobject_get() */ + file->private_data = sbinfo; + else + kobject_put(&sbinfo->si_kobj); +out: + return err; +} + +/* + * Accept a valid "si=xxxx" only. + * Once it is accepted successfully, accept "clean" too. + */ +static ssize_t au_procfs_plm_write(struct file *file, const char __user *ubuf, + size_t count, loff_t *ppos) +{ + ssize_t err; + unsigned long id; + /* last newline is allowed */ + char buf[3 + sizeof(unsigned long) * 2 + 1]; + + err = -EACCES; + if (unlikely(!capable(CAP_SYS_ADMIN))) + goto out; + + err = -EINVAL; + if (unlikely(count > sizeof(buf))) + goto out; + + err = copy_from_user(buf, ubuf, count); + if (unlikely(err)) { + err = -EFAULT; + goto out; + } + buf[count] = 0; + + err = -EINVAL; + if (!strcmp("clean", buf)) { + au_procfs_plm_write_clean(file); + goto out_success; + } else if (unlikely(strncmp("si=", buf, 3))) + goto out; + + err = kstrtoul(buf + 3, 16, &id); + if (unlikely(err)) + goto out; + + err = au_procfs_plm_write_si(file, id); + if (unlikely(err)) + goto out; + +out_success: + err = count; /* success */ +out: + return err; +} + +static const struct proc_ops au_procfs_plm_op = { + .proc_write = au_procfs_plm_write, + .proc_release = au_procfs_plm_release +}; + +/* ---------------------------------------------------------------------- */ + +static struct proc_dir_entry *au_procfs_dir; + +void au_procfs_fin(void) +{ + remove_proc_entry(AUFS_PLINK_MAINT_NAME, au_procfs_dir); + remove_proc_entry(AUFS_PLINK_MAINT_DIR, NULL); +} + +int __init au_procfs_init(void) +{ + int err; + struct proc_dir_entry *entry; + + err = -ENOMEM; + au_procfs_dir = proc_mkdir(AUFS_PLINK_MAINT_DIR, NULL); + if (unlikely(!au_procfs_dir)) + goto out; + + entry = proc_create(AUFS_PLINK_MAINT_NAME, S_IFREG | 0200, + au_procfs_dir, &au_procfs_plm_op); + if (unlikely(!entry)) + goto out_dir; + + err = 0; + goto out; /* success */ + + +out_dir: + remove_proc_entry(AUFS_PLINK_MAINT_DIR, NULL); +out: + return err; +} diff --git a/fs/aufs/rdu.c b/fs/aufs/rdu.c new file mode 100644 index 00000000000..afd77564f3d --- /dev/null +++ b/fs/aufs/rdu.c @@ -0,0 +1,384 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * readdir in userspace. + */ + +#include +#include +#include +#include "aufs.h" + +/* bits for struct aufs_rdu.flags */ +#define AuRdu_CALLED 1 +#define AuRdu_CONT (1 << 1) +#define AuRdu_FULL (1 << 2) +#define au_ftest_rdu(flags, name) ((flags) & AuRdu_##name) +#define au_fset_rdu(flags, name) \ + do { (flags) |= AuRdu_##name; } while (0) +#define au_fclr_rdu(flags, name) \ + do { (flags) &= ~AuRdu_##name; } while (0) + +struct au_rdu_arg { + struct dir_context ctx; + struct aufs_rdu *rdu; + union au_rdu_ent_ul ent; + unsigned long end; + + struct super_block *sb; + int err; +}; + +static int au_rdu_fill(struct dir_context *ctx, const char *name, int nlen, + loff_t offset, u64 h_ino, unsigned int d_type) +{ + int err, len; + struct au_rdu_arg *arg = container_of(ctx, struct au_rdu_arg, ctx); + struct aufs_rdu *rdu = arg->rdu; + struct au_rdu_ent ent; + + err = 0; + arg->err = 0; + au_fset_rdu(rdu->cookie.flags, CALLED); + len = au_rdu_len(nlen); + if (arg->ent.ul + len < arg->end) { + ent.ino = h_ino; + ent.bindex = rdu->cookie.bindex; + ent.type = d_type; + ent.nlen = nlen; + if (unlikely(nlen > AUFS_MAX_NAMELEN)) + ent.type = DT_UNKNOWN; + + /* unnecessary to support mmap_sem since this is a dir */ + err = -EFAULT; + if (copy_to_user(arg->ent.e, &ent, sizeof(ent))) + goto out; + if (copy_to_user(arg->ent.e->name, name, nlen)) + goto out; + /* the terminating NULL */ + if (__put_user(0, arg->ent.e->name + nlen)) + goto out; + err = 0; + /* AuDbg("%p, %.*s\n", arg->ent.p, nlen, name); */ + arg->ent.ul += len; + rdu->rent++; + } else { + err = -EFAULT; + au_fset_rdu(rdu->cookie.flags, FULL); + rdu->full = 1; + rdu->tail = arg->ent; + } + +out: + /* AuTraceErr(err); */ + return err; +} + +static int au_rdu_do(struct file *h_file, struct au_rdu_arg *arg) +{ + int err; + loff_t offset; + struct au_rdu_cookie *cookie = &arg->rdu->cookie; + + /* we don't have to care (FMODE_32BITHASH | FMODE_64BITHASH) for ext4 */ + offset = vfsub_llseek(h_file, cookie->h_pos, SEEK_SET); + err = offset; + if (unlikely(offset != cookie->h_pos)) + goto out; + + err = 0; + do { + arg->err = 0; + au_fclr_rdu(cookie->flags, CALLED); + /* smp_mb(); */ + err = vfsub_iterate_dir(h_file, &arg->ctx); + if (err >= 0) + err = arg->err; + } while (!err + && au_ftest_rdu(cookie->flags, CALLED) + && !au_ftest_rdu(cookie->flags, FULL)); + cookie->h_pos = h_file->f_pos; + +out: + AuTraceErr(err); + return err; +} + +static int au_rdu(struct file *file, struct aufs_rdu *rdu) +{ + int err; + aufs_bindex_t bbot; + struct au_rdu_arg arg = { + .ctx = { + .actor = au_rdu_fill + } + }; + struct dentry *dentry; + struct inode *inode; + struct file *h_file; + struct au_rdu_cookie *cookie = &rdu->cookie; + + /* VERIFY_WRITE */ + err = !access_ok(rdu->ent.e, rdu->sz); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + goto out; + } + rdu->rent = 0; + rdu->tail = rdu->ent; + rdu->full = 0; + arg.rdu = rdu; + arg.ent = rdu->ent; + arg.end = arg.ent.ul; + arg.end += rdu->sz; + + err = -ENOTDIR; + if (unlikely(!file->f_op->iterate && !file->f_op->iterate_shared)) + goto out; + + err = security_file_permission(file, MAY_READ); + AuTraceErr(err); + if (unlikely(err)) + goto out; + + dentry = file->f_path.dentry; + inode = d_inode(dentry); + inode_lock_shared(inode); + + arg.sb = inode->i_sb; + err = si_read_lock(arg.sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out_mtx; + err = au_alive_dir(dentry); + if (unlikely(err)) + goto out_si; + /* todo: reval? */ + fi_read_lock(file); + + err = -EAGAIN; + if (unlikely(au_ftest_rdu(cookie->flags, CONT) + && cookie->generation != au_figen(file))) + goto out_unlock; + + err = 0; + if (!rdu->blk) { + rdu->blk = au_sbi(arg.sb)->si_rdblk; + if (!rdu->blk) + rdu->blk = au_dir_size(file, /*dentry*/NULL); + } + bbot = au_fbtop(file); + if (cookie->bindex < bbot) + cookie->bindex = bbot; + bbot = au_fbbot_dir(file); + /* AuDbg("b%d, b%d\n", cookie->bindex, bbot); */ + for (; !err && cookie->bindex <= bbot; + cookie->bindex++, cookie->h_pos = 0) { + h_file = au_hf_dir(file, cookie->bindex); + if (!h_file) + continue; + + au_fclr_rdu(cookie->flags, FULL); + err = au_rdu_do(h_file, &arg); + AuTraceErr(err); + if (unlikely(au_ftest_rdu(cookie->flags, FULL) || err)) + break; + } + AuDbg("rent %llu\n", rdu->rent); + + if (!err && !au_ftest_rdu(cookie->flags, CONT)) { + rdu->shwh = !!au_opt_test(au_sbi(arg.sb)->si_mntflags, SHWH); + au_fset_rdu(cookie->flags, CONT); + cookie->generation = au_figen(file); + } + + ii_read_lock_child(inode); + fsstack_copy_attr_atime(inode, au_h_iptr(inode, au_ibtop(inode))); + ii_read_unlock(inode); + +out_unlock: + fi_read_unlock(file); +out_si: + si_read_unlock(arg.sb); +out_mtx: + inode_unlock_shared(inode); +out: + AuTraceErr(err); + return err; +} + +static int au_rdu_ino(struct file *file, struct aufs_rdu *rdu) +{ + int err; + ino_t ino; + unsigned long long nent; + union au_rdu_ent_ul *u; + struct au_rdu_ent ent; + struct super_block *sb; + + err = 0; + nent = rdu->nent; + u = &rdu->ent; + sb = file->f_path.dentry->d_sb; + si_read_lock(sb, AuLock_FLUSH); + while (nent-- > 0) { + /* unnecessary to support mmap_sem since this is a dir */ + err = copy_from_user(&ent, u->e, sizeof(ent)); + if (!err) + /* VERIFY_WRITE */ + err = !access_ok(&u->e->ino, sizeof(ino)); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + break; + } + + /* AuDbg("b%d, i%llu\n", ent.bindex, ent.ino); */ + if (!ent.wh) + err = au_ino(sb, ent.bindex, ent.ino, ent.type, &ino); + else + err = au_wh_ino(sb, ent.bindex, ent.ino, ent.type, + &ino); + if (unlikely(err)) { + AuTraceErr(err); + break; + } + + err = __put_user(ino, &u->e->ino); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + break; + } + u->ul += au_rdu_len(ent.nlen); + } + si_read_unlock(sb); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_rdu_verify(struct aufs_rdu *rdu) +{ + AuDbg("rdu{%llu, %p, %u | %u | %llu, %u, %u | " + "%llu, b%d, 0x%x, g%u}\n", + rdu->sz, rdu->ent.e, rdu->verify[AufsCtlRduV_SZ], + rdu->blk, + rdu->rent, rdu->shwh, rdu->full, + rdu->cookie.h_pos, rdu->cookie.bindex, rdu->cookie.flags, + rdu->cookie.generation); + + if (rdu->verify[AufsCtlRduV_SZ] == sizeof(*rdu)) + return 0; + + AuDbg("%u:%u\n", + rdu->verify[AufsCtlRduV_SZ], (unsigned int)sizeof(*rdu)); + return -EINVAL; +} + +long au_rdu_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + long err, e; + struct aufs_rdu rdu; + void __user *p = (void __user *)arg; + + err = copy_from_user(&rdu, p, sizeof(rdu)); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + goto out; + } + err = au_rdu_verify(&rdu); + if (unlikely(err)) + goto out; + + switch (cmd) { + case AUFS_CTL_RDU: + err = au_rdu(file, &rdu); + if (unlikely(err)) + break; + + e = copy_to_user(p, &rdu, sizeof(rdu)); + if (unlikely(e)) { + err = -EFAULT; + AuTraceErr(err); + } + break; + case AUFS_CTL_RDU_INO: + err = au_rdu_ino(file, &rdu); + break; + + default: + /* err = -ENOTTY; */ + err = -EINVAL; + } + +out: + AuTraceErr(err); + return err; +} + +#ifdef CONFIG_COMPAT +long au_rdu_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg) +{ + long err, e; + struct aufs_rdu rdu; + void __user *p = compat_ptr(arg); + + /* todo: get_user()? */ + err = copy_from_user(&rdu, p, sizeof(rdu)); + if (unlikely(err)) { + err = -EFAULT; + AuTraceErr(err); + goto out; + } + rdu.ent.e = compat_ptr(rdu.ent.ul); + err = au_rdu_verify(&rdu); + if (unlikely(err)) + goto out; + + switch (cmd) { + case AUFS_CTL_RDU: + err = au_rdu(file, &rdu); + if (unlikely(err)) + break; + + rdu.ent.ul = ptr_to_compat(rdu.ent.e); + rdu.tail.ul = ptr_to_compat(rdu.tail.e); + e = copy_to_user(p, &rdu, sizeof(rdu)); + if (unlikely(e)) { + err = -EFAULT; + AuTraceErr(err); + } + break; + case AUFS_CTL_RDU_INO: + err = au_rdu_ino(file, &rdu); + break; + + default: + /* err = -ENOTTY; */ + err = -EINVAL; + } + +out: + AuTraceErr(err); + return err; +} +#endif diff --git a/fs/aufs/rwsem.h b/fs/aufs/rwsem.h new file mode 100644 index 00000000000..6df9f277043 --- /dev/null +++ b/fs/aufs/rwsem.h @@ -0,0 +1,85 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * simple read-write semaphore wrappers + */ + +#ifndef __AUFS_RWSEM_H__ +#define __AUFS_RWSEM_H__ + +#ifdef __KERNEL__ + +#include "debug.h" + +/* in the future, the name 'au_rwsem' will be totally gone */ +#define au_rwsem rw_semaphore + +/* to debug easier, do not make them inlined functions */ +#define AuRwMustNoWaiters(rw) AuDebugOn(rwsem_is_contended(rw)) + +#ifdef CONFIG_LOCKDEP +/* rwsem_is_locked() is unusable */ +#define AuRwMustReadLock(rw) AuDebugOn(IS_ENABLED(CONFIG_LOCKDEP) \ + && !lockdep_recursing(current) \ + && debug_locks \ + && !lockdep_is_held_type(rw, 1)) +#define AuRwMustWriteLock(rw) AuDebugOn(IS_ENABLED(CONFIG_LOCKDEP) \ + && !lockdep_recursing(current) \ + && debug_locks \ + && !lockdep_is_held_type(rw, 0)) +#define AuRwMustAnyLock(rw) AuDebugOn(IS_ENABLED(CONFIG_LOCKDEP) \ + && !lockdep_recursing(current) \ + && debug_locks \ + && !lockdep_is_held(rw)) +#define AuRwDestroy(rw) AuDebugOn(IS_ENABLED(CONFIG_LOCKDEP) \ + && !lockdep_recursing(current) \ + && debug_locks \ + && lockdep_is_held(rw)) +#else +#define AuRwMustReadLock(rw) do {} while (0) +#define AuRwMustWriteLock(rw) do {} while (0) +#define AuRwMustAnyLock(rw) do {} while (0) +#define AuRwDestroy(rw) do {} while (0) +#endif + +#define au_rw_init(rw) init_rwsem(rw) + +#define au_rw_init_wlock(rw) do { \ + au_rw_init(rw); \ + down_write(rw); \ + } while (0) + +#define au_rw_init_wlock_nested(rw, lsc) do { \ + au_rw_init(rw); \ + down_write_nested(rw, lsc); \ + } while (0) + +#define au_rw_read_lock(rw) down_read(rw) +#define au_rw_read_lock_nested(rw, lsc) down_read_nested(rw, lsc) +#define au_rw_read_unlock(rw) up_read(rw) +#define au_rw_dgrade_lock(rw) downgrade_write(rw) +#define au_rw_write_lock(rw) down_write(rw) +#define au_rw_write_lock_nested(rw, lsc) down_write_nested(rw, lsc) +#define au_rw_write_unlock(rw) up_write(rw) +/* why is not _nested version defined? */ +#define au_rw_read_trylock(rw) down_read_trylock(rw) +#define au_rw_write_trylock(rw) down_write_trylock(rw) + +#endif /* __KERNEL__ */ +#endif /* __AUFS_RWSEM_H__ */ diff --git a/fs/aufs/sbinfo.c b/fs/aufs/sbinfo.c new file mode 100644 index 00000000000..3835be61d97 --- /dev/null +++ b/fs/aufs/sbinfo.c @@ -0,0 +1,314 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * superblock private data + */ + +#include +#include "aufs.h" + +/* + * they are necessary regardless sysfs is disabled. + */ +void au_si_free(struct kobject *kobj) +{ + int i; + struct au_sbinfo *sbinfo; + char *locked __maybe_unused; /* debug only */ + + sbinfo = container_of(kobj, struct au_sbinfo, si_kobj); + for (i = 0; i < AuPlink_NHASH; i++) + AuDebugOn(!hlist_bl_empty(sbinfo->si_plink + i)); + AuDebugOn(atomic_read(&sbinfo->si_nowait.nw_len)); + + AuLCntZero(au_lcnt_read(&sbinfo->si_ninodes, /*do_rev*/0)); + au_lcnt_fin(&sbinfo->si_ninodes, /*do_sync*/0); + AuLCntZero(au_lcnt_read(&sbinfo->si_nfiles, /*do_rev*/0)); + au_lcnt_fin(&sbinfo->si_nfiles, /*do_sync*/0); + + dbgaufs_si_fin(sbinfo); + au_rw_write_lock(&sbinfo->si_rwsem); + au_br_free(sbinfo); + au_rw_write_unlock(&sbinfo->si_rwsem); + + au_kfree_try_rcu(sbinfo->si_branch); + mutex_destroy(&sbinfo->si_xib_mtx); + AuRwDestroy(&sbinfo->si_rwsem); + + au_lcnt_wait_for_fin(&sbinfo->si_ninodes); + /* si_nfiles is waited too */ + au_kfree_rcu(sbinfo); +} + +int au_si_alloc(struct super_block *sb) +{ + int err, i; + struct au_sbinfo *sbinfo; + + err = -ENOMEM; + sbinfo = kzalloc(sizeof(*sbinfo), GFP_NOFS); + if (unlikely(!sbinfo)) + goto out; + + /* will be reallocated separately */ + sbinfo->si_branch = kzalloc(sizeof(*sbinfo->si_branch), GFP_NOFS); + if (unlikely(!sbinfo->si_branch)) + goto out_sbinfo; + + err = sysaufs_si_init(sbinfo); + if (!err) { + dbgaufs_si_null(sbinfo); + err = dbgaufs_si_init(sbinfo); + if (unlikely(err)) + kobject_put(&sbinfo->si_kobj); + } + if (unlikely(err)) + goto out_br; + + au_nwt_init(&sbinfo->si_nowait); + au_rw_init_wlock(&sbinfo->si_rwsem); + + au_lcnt_init(&sbinfo->si_ninodes, /*release*/NULL); + au_lcnt_init(&sbinfo->si_nfiles, /*release*/NULL); + + sbinfo->si_bbot = -1; + sbinfo->si_last_br_id = AUFS_BRANCH_MAX / 2; + + sbinfo->si_wbr_copyup = AuWbrCopyup_Def; + sbinfo->si_wbr_create = AuWbrCreate_Def; + sbinfo->si_wbr_copyup_ops = au_wbr_copyup_ops + sbinfo->si_wbr_copyup; + sbinfo->si_wbr_create_ops = au_wbr_create_ops + sbinfo->si_wbr_create; + + au_fhsm_init(sbinfo); + + sbinfo->si_mntflags = au_opts_plink(AuOpt_Def); + + sbinfo->si_xino_jiffy = jiffies; + sbinfo->si_xino_expire + = msecs_to_jiffies(AUFS_XINO_DEF_SEC * MSEC_PER_SEC); + mutex_init(&sbinfo->si_xib_mtx); + /* leave si_xib_last_pindex and si_xib_next_bit */ + + INIT_HLIST_BL_HEAD(&sbinfo->si_aopen); + + sbinfo->si_rdcache = msecs_to_jiffies(AUFS_RDCACHE_DEF * MSEC_PER_SEC); + sbinfo->si_rdblk = AUFS_RDBLK_DEF; + sbinfo->si_rdhash = AUFS_RDHASH_DEF; + sbinfo->si_dirwh = AUFS_DIRWH_DEF; + + for (i = 0; i < AuPlink_NHASH; i++) + INIT_HLIST_BL_HEAD(sbinfo->si_plink + i); + init_waitqueue_head(&sbinfo->si_plink_wq); + spin_lock_init(&sbinfo->si_plink_maint_lock); + + INIT_HLIST_BL_HEAD(&sbinfo->si_files); + + /* with getattr by default */ + sbinfo->si_iop_array = aufs_iop; + + /* leave other members for sysaufs and si_mnt. */ + sbinfo->si_sb = sb; + sb->s_fs_info = sbinfo; + si_pid_set(sb); + return 0; /* success */ + +out_br: + au_kfree_try_rcu(sbinfo->si_branch); +out_sbinfo: + au_kfree_rcu(sbinfo); +out: + return err; +} + +int au_sbr_realloc(struct au_sbinfo *sbinfo, int nbr, int may_shrink) +{ + int err, sz; + struct au_branch **brp; + + AuRwMustWriteLock(&sbinfo->si_rwsem); + + err = -ENOMEM; + sz = sizeof(*brp) * (sbinfo->si_bbot + 1); + if (unlikely(!sz)) + sz = sizeof(*brp); + brp = au_kzrealloc(sbinfo->si_branch, sz, sizeof(*brp) * nbr, GFP_NOFS, + may_shrink); + if (brp) { + sbinfo->si_branch = brp; + err = 0; + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +unsigned int au_sigen_inc(struct super_block *sb) +{ + unsigned int gen; + struct inode *inode; + + SiMustWriteLock(sb); + + gen = ++au_sbi(sb)->si_generation; + au_update_digen(sb->s_root); + inode = d_inode(sb->s_root); + au_update_iigen(inode, /*half*/0); + inode_inc_iversion(inode); + return gen; +} + +aufs_bindex_t au_new_br_id(struct super_block *sb) +{ + aufs_bindex_t br_id; + int i; + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + for (i = 0; i <= AUFS_BRANCH_MAX; i++) { + br_id = ++sbinfo->si_last_br_id; + AuDebugOn(br_id < 0); + if (br_id && au_br_index(sb, br_id) < 0) + return br_id; + } + + return -1; +} + +/* ---------------------------------------------------------------------- */ + +/* it is ok that new 'nwt' tasks are appended while we are sleeping */ +int si_read_lock(struct super_block *sb, int flags) +{ + int err; + + err = 0; + if (au_ftest_lock(flags, FLUSH)) + au_nwt_flush(&au_sbi(sb)->si_nowait); + + si_noflush_read_lock(sb); + err = au_plink_maint(sb, flags); + if (unlikely(err)) + si_read_unlock(sb); + + return err; +} + +int si_write_lock(struct super_block *sb, int flags) +{ + int err; + + if (au_ftest_lock(flags, FLUSH)) + au_nwt_flush(&au_sbi(sb)->si_nowait); + + si_noflush_write_lock(sb); + err = au_plink_maint(sb, flags); + if (unlikely(err)) + si_write_unlock(sb); + + return err; +} + +/* dentry and super_block lock. call at entry point */ +int aufs_read_lock(struct dentry *dentry, int flags) +{ + int err; + struct super_block *sb; + + sb = dentry->d_sb; + err = si_read_lock(sb, flags); + if (unlikely(err)) + goto out; + + if (au_ftest_lock(flags, DW)) + di_write_lock_child(dentry); + else + di_read_lock_child(dentry, flags); + + if (au_ftest_lock(flags, GEN)) { + err = au_digen_test(dentry, au_sigen(sb)); + if (!au_opt_test(au_mntflags(sb), UDBA_NONE)) + AuDebugOn(!err && au_dbrange_test(dentry)); + else if (!err) + err = au_dbrange_test(dentry); + if (unlikely(err)) + aufs_read_unlock(dentry, flags); + } + +out: + return err; +} + +void aufs_read_unlock(struct dentry *dentry, int flags) +{ + if (au_ftest_lock(flags, DW)) + di_write_unlock(dentry); + else + di_read_unlock(dentry, flags); + si_read_unlock(dentry->d_sb); +} + +void aufs_write_lock(struct dentry *dentry) +{ + si_write_lock(dentry->d_sb, AuLock_FLUSH | AuLock_NOPLMW); + di_write_lock_child(dentry); +} + +void aufs_write_unlock(struct dentry *dentry) +{ + di_write_unlock(dentry); + si_write_unlock(dentry->d_sb); +} + +int aufs_read_and_write_lock2(struct dentry *d1, struct dentry *d2, int flags) +{ + int err; + unsigned int sigen; + struct super_block *sb; + + sb = d1->d_sb; + err = si_read_lock(sb, flags); + if (unlikely(err)) + goto out; + + di_write_lock2_child(d1, d2, au_ftest_lock(flags, DIRS)); + + if (au_ftest_lock(flags, GEN)) { + sigen = au_sigen(sb); + err = au_digen_test(d1, sigen); + AuDebugOn(!err && au_dbrange_test(d1)); + if (!err) { + err = au_digen_test(d2, sigen); + AuDebugOn(!err && au_dbrange_test(d2)); + } + if (unlikely(err)) + aufs_read_and_write_unlock2(d1, d2); + } + +out: + return err; +} + +void aufs_read_and_write_unlock2(struct dentry *d1, struct dentry *d2) +{ + di_write_unlock2(d1, d2); + si_read_unlock(d1->d_sb); +} diff --git a/fs/aufs/super.c b/fs/aufs/super.c new file mode 100644 index 00000000000..1615c0e8f23 --- /dev/null +++ b/fs/aufs/super.c @@ -0,0 +1,1050 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * mount and super_block operations + */ + +#include +#include +#include +#include +#include +#include "aufs.h" + +/* + * super_operations + */ +static struct inode *aufs_alloc_inode(struct super_block *sb __maybe_unused) +{ + struct au_icntnr *c; + + c = au_cache_alloc_icntnr(); + if (c) { + au_icntnr_init(c); + inode_set_iversion(&c->vfs_inode, 1); /* sigen(sb); */ + c->iinfo.ii_hinode = NULL; + return &c->vfs_inode; + } + return NULL; +} + +static void aufs_destroy_inode(struct inode *inode) +{ + if (!au_is_bad_inode(inode)) + au_iinfo_fin(inode); +} + +static void aufs_free_inode(struct inode *inode) +{ + au_cache_free_icntnr(container_of(inode, struct au_icntnr, vfs_inode)); +} + +struct inode *au_iget_locked(struct super_block *sb, ino_t ino) +{ + struct inode *inode; + int err; + + inode = iget_locked(sb, ino); + if (unlikely(!inode)) { + inode = ERR_PTR(-ENOMEM); + goto out; + } + if (!(inode->i_state & I_NEW)) + goto out; + + err = au_xigen_new(inode); + if (!err) + err = au_iinfo_init(inode); + if (!err) + inode_inc_iversion(inode); + else { + iget_failed(inode); + inode = ERR_PTR(err); + } + +out: + /* never return NULL */ + AuDebugOn(!inode); + AuTraceErrPtr(inode); + return inode; +} + +/* lock free root dinfo */ +static int au_show_brs(struct seq_file *seq, struct super_block *sb) +{ + int err; + aufs_bindex_t bindex, bbot; + struct path path; + struct au_hdentry *hdp; + struct au_branch *br; + au_br_perm_str_t perm; + + err = 0; + bbot = au_sbbot(sb); + bindex = 0; + hdp = au_hdentry(au_di(sb->s_root), bindex); + for (; !err && bindex <= bbot; bindex++, hdp++) { + br = au_sbr(sb, bindex); + path.mnt = au_br_mnt(br); + path.dentry = hdp->hd_dentry; + err = au_seq_path(seq, &path); + if (!err) { + au_optstr_br_perm(&perm, br->br_perm); + seq_printf(seq, "=%s", perm.a); + if (bindex != bbot) + seq_putc(seq, ':'); + } + } + if (unlikely(err || seq_has_overflowed(seq))) + err = -E2BIG; + + return err; +} + +static void au_gen_fmt(char *fmt, int len __maybe_unused, const char *pat, + const char *append) +{ + char *p; + + p = fmt; + while (*pat != ':') + *p++ = *pat++; + *p++ = *pat++; + strcpy(p, append); + AuDebugOn(strlen(fmt) >= len); +} + +static void au_show_wbr_create(struct seq_file *m, int v, + struct au_sbinfo *sbinfo) +{ + const char *pat; + char fmt[32]; + struct au_wbr_mfs *mfs; + + AuRwMustAnyLock(&sbinfo->si_rwsem); + + seq_puts(m, ",create="); + pat = au_optstr_wbr_create(v); + mfs = &sbinfo->si_wbr_mfs; + switch (v) { + case AuWbrCreate_TDP: + case AuWbrCreate_RR: + case AuWbrCreate_MFS: + case AuWbrCreate_PMFS: + seq_puts(m, pat); + break; + case AuWbrCreate_MFSRR: + case AuWbrCreate_TDMFS: + case AuWbrCreate_PMFSRR: + au_gen_fmt(fmt, sizeof(fmt), pat, "%llu"); + seq_printf(m, fmt, mfs->mfsrr_watermark); + break; + case AuWbrCreate_MFSV: + case AuWbrCreate_PMFSV: + au_gen_fmt(fmt, sizeof(fmt), pat, "%lu"); + seq_printf(m, fmt, + jiffies_to_msecs(mfs->mfs_expire) + / MSEC_PER_SEC); + break; + case AuWbrCreate_MFSRRV: + case AuWbrCreate_TDMFSV: + case AuWbrCreate_PMFSRRV: + au_gen_fmt(fmt, sizeof(fmt), pat, "%llu:%lu"); + seq_printf(m, fmt, mfs->mfsrr_watermark, + jiffies_to_msecs(mfs->mfs_expire) / MSEC_PER_SEC); + break; + default: + BUG(); + } +} + +static int au_show_xino(struct seq_file *seq, struct super_block *sb) +{ +#ifdef CONFIG_SYSFS + return 0; +#else + int err; + const int len = sizeof(AUFS_XINO_FNAME) - 1; + aufs_bindex_t bindex, brid; + struct qstr *name; + struct file *f; + struct dentry *d, *h_root; + struct au_branch *br; + + AuRwMustAnyLock(&sbinfo->si_rwsem); + + err = 0; + f = au_sbi(sb)->si_xib; + if (!f) + goto out; + + /* stop printing the default xino path on the first writable branch */ + h_root = NULL; + bindex = au_xi_root(sb, f->f_path.dentry); + if (bindex >= 0) { + br = au_sbr_sb(sb, bindex); + h_root = au_br_dentry(br); + } + + d = f->f_path.dentry; + name = &d->d_name; + /* safe ->d_parent because the file is unlinked */ + if (d->d_parent == h_root + && name->len == len + && !memcmp(name->name, AUFS_XINO_FNAME, len)) + goto out; + + seq_puts(seq, ",xino="); + err = au_xino_path(seq, f); + +out: + return err; +#endif +} + +/* seq_file will re-call me in case of too long string */ +static int aufs_show_options(struct seq_file *m, struct dentry *dentry) +{ + int err; + unsigned int mnt_flags, v; + struct super_block *sb; + struct au_sbinfo *sbinfo; + +#define AuBool(name, str) do { \ + v = au_opt_test(mnt_flags, name); \ + if (v != au_opt_test(AuOpt_Def, name)) \ + seq_printf(m, ",%s" #str, v ? "" : "no"); \ +} while (0) + +#define AuStr(name, str) do { \ + v = mnt_flags & AuOptMask_##name; \ + if (v != (AuOpt_Def & AuOptMask_##name)) \ + seq_printf(m, "," #str "=%s", au_optstr_##str(v)); \ +} while (0) + +#define AuUInt(name, str, val) do { \ + if (val != AUFS_##name##_DEF) \ + seq_printf(m, "," #str "=%u", val); \ +} while (0) + + sb = dentry->d_sb; + if (sb->s_flags & SB_POSIXACL) + seq_puts(m, ",acl"); +#if 0 /* reserved for future use */ + if (sb->s_flags & SB_I_VERSION) + seq_puts(m, ",i_version"); +#endif + + /* lock free root dinfo */ + si_noflush_read_lock(sb); + sbinfo = au_sbi(sb); + seq_printf(m, ",si=%lx", sysaufs_si_id(sbinfo)); + + mnt_flags = au_mntflags(sb); + if (au_opt_test(mnt_flags, XINO)) { + err = au_show_xino(m, sb); + if (unlikely(err)) + goto out; + } else + seq_puts(m, ",noxino"); + + AuBool(TRUNC_XINO, trunc_xino); + AuStr(UDBA, udba); + AuBool(SHWH, shwh); + AuBool(PLINK, plink); + AuBool(DIO, dio); + AuBool(DIRPERM1, dirperm1); + + v = sbinfo->si_wbr_create; + if (v != AuWbrCreate_Def) + au_show_wbr_create(m, v, sbinfo); + + v = sbinfo->si_wbr_copyup; + if (v != AuWbrCopyup_Def) + seq_printf(m, ",cpup=%s", au_optstr_wbr_copyup(v)); + + v = au_opt_test(mnt_flags, ALWAYS_DIROPQ); + if (v != au_opt_test(AuOpt_Def, ALWAYS_DIROPQ)) + seq_printf(m, ",diropq=%c", v ? 'a' : 'w'); + + AuUInt(DIRWH, dirwh, sbinfo->si_dirwh); + + v = jiffies_to_msecs(sbinfo->si_rdcache) / MSEC_PER_SEC; + AuUInt(RDCACHE, rdcache, v); + + AuUInt(RDBLK, rdblk, sbinfo->si_rdblk); + AuUInt(RDHASH, rdhash, sbinfo->si_rdhash); + + au_fhsm_show(m, sbinfo); + + AuBool(DIRREN, dirren); + AuBool(SUM, sum); + /* AuBool(SUM_W, wsum); */ + AuBool(WARN_PERM, warn_perm); + AuBool(VERBOSE, verbose); + +out: + /* be sure to print "br:" last */ + if (!sysaufs_brs) { + seq_puts(m, ",br:"); + au_show_brs(m, sb); + } + si_read_unlock(sb); + return 0; + +#undef AuBool +#undef AuStr +#undef AuUInt +} + +/* ---------------------------------------------------------------------- */ + +/* sum mode which returns the summation for statfs(2) */ + +static u64 au_add_till_max(u64 a, u64 b) +{ + u64 old; + + old = a; + a += b; + if (old <= a) + return a; + return ULLONG_MAX; +} + +static u64 au_mul_till_max(u64 a, long mul) +{ + u64 old; + + old = a; + a *= mul; + if (old <= a) + return a; + return ULLONG_MAX; +} + +static int au_statfs_sum(struct super_block *sb, struct kstatfs *buf) +{ + int err; + long bsize, factor; + u64 blocks, bfree, bavail, files, ffree; + aufs_bindex_t bbot, bindex, i; + unsigned char shared; + struct path h_path; + struct super_block *h_sb; + + err = 0; + bsize = LONG_MAX; + files = 0; + ffree = 0; + blocks = 0; + bfree = 0; + bavail = 0; + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + h_path.mnt = au_sbr_mnt(sb, bindex); + h_sb = h_path.mnt->mnt_sb; + shared = 0; + for (i = 0; !shared && i < bindex; i++) + shared = (au_sbr_sb(sb, i) == h_sb); + if (shared) + continue; + + /* sb->s_root for NFS is unreliable */ + h_path.dentry = h_path.mnt->mnt_root; + err = vfs_statfs(&h_path, buf); + if (unlikely(err)) + goto out; + + if (bsize > buf->f_bsize) { + /* + * we will reduce bsize, so we have to expand blocks + * etc. to match them again + */ + factor = (bsize / buf->f_bsize); + blocks = au_mul_till_max(blocks, factor); + bfree = au_mul_till_max(bfree, factor); + bavail = au_mul_till_max(bavail, factor); + bsize = buf->f_bsize; + } + + factor = (buf->f_bsize / bsize); + blocks = au_add_till_max(blocks, + au_mul_till_max(buf->f_blocks, factor)); + bfree = au_add_till_max(bfree, + au_mul_till_max(buf->f_bfree, factor)); + bavail = au_add_till_max(bavail, + au_mul_till_max(buf->f_bavail, factor)); + files = au_add_till_max(files, buf->f_files); + ffree = au_add_till_max(ffree, buf->f_ffree); + } + + buf->f_bsize = bsize; + buf->f_blocks = blocks; + buf->f_bfree = bfree; + buf->f_bavail = bavail; + buf->f_files = files; + buf->f_ffree = ffree; + buf->f_frsize = 0; + +out: + return err; +} + +static int aufs_statfs(struct dentry *dentry, struct kstatfs *buf) +{ + int err; + struct path h_path; + struct super_block *sb; + + /* lock free root dinfo */ + sb = dentry->d_sb; + si_noflush_read_lock(sb); + if (!au_opt_test(au_mntflags(sb), SUM)) { + /* sb->s_root for NFS is unreliable */ + h_path.mnt = au_sbr_mnt(sb, 0); + h_path.dentry = h_path.mnt->mnt_root; + err = vfs_statfs(&h_path, buf); + } else + err = au_statfs_sum(sb, buf); + si_read_unlock(sb); + + if (!err) { + buf->f_type = AUFS_SUPER_MAGIC; + buf->f_namelen = AUFS_MAX_NAMELEN; + memset(&buf->f_fsid, 0, sizeof(buf->f_fsid)); + } + /* buf->f_bsize = buf->f_blocks = buf->f_bfree = buf->f_bavail = -1; */ + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int aufs_sync_fs(struct super_block *sb, int wait) +{ + int err, e; + aufs_bindex_t bbot, bindex; + struct au_branch *br; + struct super_block *h_sb; + + err = 0; + si_noflush_read_lock(sb); + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (!au_br_writable(br->br_perm)) + continue; + + h_sb = au_sbr_sb(sb, bindex); + e = vfsub_sync_filesystem(h_sb, wait); + if (unlikely(e && !err)) + err = e; + /* go on even if an error happens */ + } + si_read_unlock(sb); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* final actions when unmounting a file system */ +static void aufs_put_super(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + + sbinfo = au_sbi(sb); + if (sbinfo) + kobject_put(&sbinfo->si_kobj); +} + +/* ---------------------------------------------------------------------- */ + +void *au_array_alloc(unsigned long long *hint, au_arraycb_t cb, + struct super_block *sb, void *arg) +{ + void *array; + unsigned long long n, sz; + + array = NULL; + n = 0; + if (!*hint) + goto out; + + if (*hint > ULLONG_MAX / sizeof(array)) { + array = ERR_PTR(-EMFILE); + pr_err("hint %llu\n", *hint); + goto out; + } + + sz = sizeof(array) * *hint; + array = kzalloc(sz, GFP_NOFS); + if (unlikely(!array)) + array = vzalloc(sz); + if (unlikely(!array)) { + array = ERR_PTR(-ENOMEM); + goto out; + } + + n = cb(sb, array, *hint, arg); + AuDebugOn(n > *hint); + +out: + *hint = n; + return array; +} + +static unsigned long long au_iarray_cb(struct super_block *sb, void *a, + unsigned long long max __maybe_unused, + void *arg) +{ + unsigned long long n; + struct inode **p, *inode; + struct list_head *head; + + n = 0; + p = a; + head = arg; + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(inode, head, i_sb_list) { + if (!au_is_bad_inode(inode) + && au_ii(inode)->ii_btop >= 0) { + spin_lock(&inode->i_lock); + if (atomic_read(&inode->i_count)) { + au_igrab(inode); + *p++ = inode; + n++; + AuDebugOn(n > max); + } + spin_unlock(&inode->i_lock); + } + } + spin_unlock(&sb->s_inode_list_lock); + + return n; +} + +struct inode **au_iarray_alloc(struct super_block *sb, unsigned long long *max) +{ + struct au_sbinfo *sbi; + + sbi = au_sbi(sb); + *max = au_lcnt_read(&sbi->si_ninodes, /*do_rev*/1); + return au_array_alloc(max, au_iarray_cb, sb, &sb->s_inodes); +} + +void au_iarray_free(struct inode **a, unsigned long long max) +{ + unsigned long long ull; + + for (ull = 0; ull < max; ull++) + iput(a[ull]); + kvfree(a); +} + +/* ---------------------------------------------------------------------- */ + +/* + * refresh dentry and inode at remount time. + */ +/* todo: consolidate with simple_reval_dpath() and au_reval_for_attr() */ +static int au_do_refresh(struct dentry *dentry, unsigned int dir_flags, + struct dentry *parent) +{ + int err; + + di_write_lock_child(dentry); + di_read_lock_parent(parent, AuLock_IR); + err = au_refresh_dentry(dentry, parent); + if (!err && dir_flags) + au_hn_reset(d_inode(dentry), dir_flags); + di_read_unlock(parent, AuLock_IR); + di_write_unlock(dentry); + + return err; +} + +static int au_do_refresh_d(struct dentry *dentry, unsigned int sigen, + struct au_sbinfo *sbinfo, + const unsigned int dir_flags, unsigned int do_idop) +{ + int err; + struct dentry *parent; + + err = 0; + parent = dget_parent(dentry); + if (!au_digen_test(parent, sigen) && au_digen_test(dentry, sigen)) { + if (d_really_is_positive(dentry)) { + if (!d_is_dir(dentry)) + err = au_do_refresh(dentry, /*dir_flags*/0, + parent); + else { + err = au_do_refresh(dentry, dir_flags, parent); + if (unlikely(err)) + au_fset_si(sbinfo, FAILED_REFRESH_DIR); + } + } else + err = au_do_refresh(dentry, /*dir_flags*/0, parent); + AuDbgDentry(dentry); + } + dput(parent); + + if (!err) { + if (do_idop) + au_refresh_dop(dentry, /*force_reval*/0); + } else + au_refresh_dop(dentry, /*force_reval*/1); + + AuTraceErr(err); + return err; +} + +static int au_refresh_d(struct super_block *sb, unsigned int do_idop) +{ + int err, i, j, ndentry, e; + unsigned int sigen; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry **dentries, *d; + struct au_sbinfo *sbinfo; + struct dentry *root = sb->s_root; + const unsigned int dir_flags = au_hi_flags(d_inode(root), /*isdir*/1); + + if (do_idop) + au_refresh_dop(root, /*force_reval*/0); + + err = au_dpages_init(&dpages, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_dcsub_pages(&dpages, root, NULL, NULL); + if (unlikely(err)) + goto out_dpages; + + sigen = au_sigen(sb); + sbinfo = au_sbi(sb); + for (i = 0; i < dpages.ndpage; i++) { + dpage = dpages.dpages + i; + dentries = dpage->dentries; + ndentry = dpage->ndentry; + for (j = 0; j < ndentry; j++) { + d = dentries[j]; + e = au_do_refresh_d(d, sigen, sbinfo, dir_flags, + do_idop); + if (unlikely(e && !err)) + err = e; + /* go on even err */ + } + } + +out_dpages: + au_dpages_free(&dpages); +out: + return err; +} + +static int au_refresh_i(struct super_block *sb, unsigned int do_idop) +{ + int err, e; + unsigned int sigen; + unsigned long long max, ull; + struct inode *inode, **array; + + array = au_iarray_alloc(sb, &max); + err = PTR_ERR(array); + if (IS_ERR(array)) + goto out; + + err = 0; + sigen = au_sigen(sb); + for (ull = 0; ull < max; ull++) { + inode = array[ull]; + if (unlikely(!inode)) + break; + + e = 0; + ii_write_lock_child(inode); + if (au_iigen(inode, NULL) != sigen) { + e = au_refresh_hinode_self(inode); + if (unlikely(e)) { + au_refresh_iop(inode, /*force_getattr*/1); + pr_err("error %d, i%lu\n", e, inode->i_ino); + if (!err) + err = e; + /* go on even if err */ + } + } + if (!e && do_idop) + au_refresh_iop(inode, /*force_getattr*/0); + ii_write_unlock(inode); + } + + au_iarray_free(array, max); + +out: + return err; +} + +static void au_remount_refresh(struct super_block *sb, unsigned int do_idop) +{ + int err, e; + unsigned int udba; + aufs_bindex_t bindex, bbot; + struct dentry *root; + struct inode *inode; + struct au_branch *br; + struct au_sbinfo *sbi; + + au_sigen_inc(sb); + sbi = au_sbi(sb); + au_fclr_si(sbi, FAILED_REFRESH_DIR); + + root = sb->s_root; + DiMustNoWaiters(root); + inode = d_inode(root); + IiMustNoWaiters(inode); + + udba = au_opt_udba(sb); + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + err = au_hnotify_reset_br(udba, br, br->br_perm); + if (unlikely(err)) + AuIOErr("hnotify failed on br %d, %d, ignored\n", + bindex, err); + /* go on even if err */ + } + au_hn_reset(inode, au_hi_flags(inode, /*isdir*/1)); + + if (do_idop) { + if (au_ftest_si(sbi, NO_DREVAL)) { + AuDebugOn(sb->s_d_op == &aufs_dop_noreval); + sb->s_d_op = &aufs_dop_noreval; + AuDebugOn(sbi->si_iop_array == aufs_iop_nogetattr); + sbi->si_iop_array = aufs_iop_nogetattr; + } else { + AuDebugOn(sb->s_d_op == &aufs_dop); + sb->s_d_op = &aufs_dop; + AuDebugOn(sbi->si_iop_array == aufs_iop); + sbi->si_iop_array = aufs_iop; + } + pr_info("reset to %ps and %ps\n", + sb->s_d_op, sbi->si_iop_array); + } + + di_write_unlock(root); + err = au_refresh_d(sb, do_idop); + e = au_refresh_i(sb, do_idop); + if (unlikely(e && !err)) + err = e; + /* aufs_write_lock() calls ..._child() */ + di_write_lock_child(root); + + au_cpup_attr_all(inode, /*force*/1); + + if (unlikely(err)) + AuIOErr("refresh failed, ignored, %d\n", err); +} + +/* stop extra interpretation of errno in mount(8), and strange error messages */ +static int cvt_err(int err) +{ + AuTraceErr(err); + + switch (err) { + case -ENOENT: + case -ENOTDIR: + case -EEXIST: + case -EIO: + err = -EINVAL; + } + return err; +} + +static int aufs_remount_fs(struct super_block *sb, int *flags, char *data) +{ + int err, do_dx; + unsigned int mntflags; + struct au_opts opts = { + .opt = NULL + }; + struct dentry *root; + struct inode *inode; + struct au_sbinfo *sbinfo; + + err = 0; + root = sb->s_root; + if (!data || !*data) { + err = si_write_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (!err) { + di_write_lock_child(root); + err = au_opts_verify(sb, *flags, /*pending*/0); + aufs_write_unlock(root); + } + goto out; + } + + err = -ENOMEM; + opts.opt = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!opts.opt)) + goto out; + opts.max_opt = PAGE_SIZE / sizeof(*opts.opt); + opts.flags = AuOpts_REMOUNT; + opts.sb_flags = *flags; + + /* parse it before aufs lock */ + err = au_opts_parse(sb, data, &opts); + if (unlikely(err)) + goto out_opts; + + sbinfo = au_sbi(sb); + inode = d_inode(root); + inode_lock(inode); + err = si_write_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out_mtx; + di_write_lock_child(root); + + /* au_opts_remount() may return an error */ + err = au_opts_remount(sb, &opts); + au_opts_free(&opts); + + if (au_ftest_opts(opts.flags, REFRESH)) + au_remount_refresh(sb, au_ftest_opts(opts.flags, REFRESH_IDOP)); + + if (au_ftest_opts(opts.flags, REFRESH_DYAOP)) { + mntflags = au_mntflags(sb); + do_dx = !!au_opt_test(mntflags, DIO); + au_dy_arefresh(do_dx); + } + + au_fhsm_wrote_all(sb, /*force*/1); /* ?? */ + aufs_write_unlock(root); + +out_mtx: + inode_unlock(inode); +out_opts: + free_page((unsigned long)opts.opt); +out: + err = cvt_err(err); + AuTraceErr(err); + return err; +} + +static const struct super_operations aufs_sop = { + .alloc_inode = aufs_alloc_inode, + .destroy_inode = aufs_destroy_inode, + .free_inode = aufs_free_inode, + /* always deleting, no clearing */ + .drop_inode = generic_delete_inode, + .show_options = aufs_show_options, + .statfs = aufs_statfs, + .put_super = aufs_put_super, + .sync_fs = aufs_sync_fs, + .remount_fs = aufs_remount_fs +}; + +/* ---------------------------------------------------------------------- */ + +static int alloc_root(struct super_block *sb) +{ + int err; + struct inode *inode; + struct dentry *root; + + err = -ENOMEM; + inode = au_iget_locked(sb, AUFS_ROOT_INO); + err = PTR_ERR(inode); + if (IS_ERR(inode)) + goto out; + + inode->i_op = aufs_iop + AuIop_DIR; /* with getattr by default */ + inode->i_fop = &aufs_dir_fop; + inode->i_mode = S_IFDIR; + set_nlink(inode, 2); + unlock_new_inode(inode); + + root = d_make_root(inode); + if (unlikely(!root)) + goto out; + err = PTR_ERR(root); + if (IS_ERR(root)) + goto out; + + err = au_di_init(root); + if (!err) { + sb->s_root = root; + return 0; /* success */ + } + dput(root); + +out: + return err; +} + +static int aufs_fill_super(struct super_block *sb, void *raw_data, + int silent __maybe_unused) +{ + int err; + struct au_opts opts = { + .opt = NULL + }; + struct au_sbinfo *sbinfo; + struct dentry *root; + struct inode *inode; + char *arg = raw_data; + + if (unlikely(!arg || !*arg)) { + err = -EINVAL; + pr_err("no arg\n"); + goto out; + } + + err = -ENOMEM; + opts.opt = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!opts.opt)) + goto out; + opts.max_opt = PAGE_SIZE / sizeof(*opts.opt); + opts.sb_flags = sb->s_flags; + + err = au_si_alloc(sb); + if (unlikely(err)) + goto out_opts; + sbinfo = au_sbi(sb); + + /* all timestamps always follow the ones on the branch */ + sb->s_flags |= SB_NOATIME | SB_NODIRATIME; + sb->s_flags |= SB_I_VERSION; /* do we really need this? */ + sb->s_op = &aufs_sop; + sb->s_d_op = &aufs_dop; + sb->s_magic = AUFS_SUPER_MAGIC; + sb->s_maxbytes = 0; + sb->s_stack_depth = 1; + au_export_init(sb); + au_xattr_init(sb); + + err = alloc_root(sb); + if (unlikely(err)) { + si_write_unlock(sb); + goto out_info; + } + root = sb->s_root; + inode = d_inode(root); + + /* + * actually we can parse options regardless aufs lock here. + * but at remount time, parsing must be done before aufs lock. + * so we follow the same rule. + */ + ii_write_lock_parent(inode); + aufs_write_unlock(root); + err = au_opts_parse(sb, arg, &opts); + if (unlikely(err)) + goto out_root; + + /* lock vfs_inode first, then aufs. */ + inode_lock(inode); + aufs_write_lock(root); + err = au_opts_mount(sb, &opts); + au_opts_free(&opts); + if (!err && au_ftest_si(sbinfo, NO_DREVAL)) { + sb->s_d_op = &aufs_dop_noreval; + pr_info("%ps\n", sb->s_d_op); + au_refresh_dop(root, /*force_reval*/0); + sbinfo->si_iop_array = aufs_iop_nogetattr; + au_refresh_iop(inode, /*force_getattr*/0); + } + aufs_write_unlock(root); + inode_unlock(inode); + if (!err) + goto out_opts; /* success */ + +out_root: + dput(root); + sb->s_root = NULL; +out_info: + kobject_put(&sbinfo->si_kobj); + sb->s_fs_info = NULL; +out_opts: + free_page((unsigned long)opts.opt); +out: + AuTraceErr(err); + err = cvt_err(err); + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static struct dentry *aufs_mount(struct file_system_type *fs_type, int flags, + const char *dev_name __maybe_unused, + void *raw_data) +{ + struct dentry *root; + + /* all timestamps always follow the ones on the branch */ + /* mnt->mnt_flags |= MNT_NOATIME | MNT_NODIRATIME; */ + root = mount_nodev(fs_type, flags, raw_data, aufs_fill_super); + if (IS_ERR(root)) + goto out; + + au_sbilist_add(root->d_sb); + +out: + return root; +} + +static void aufs_kill_sb(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + + sbinfo = au_sbi(sb); + if (sbinfo) { + au_sbilist_del(sb); + aufs_write_lock(sb->s_root); + au_fhsm_fin(sb); + if (sbinfo->si_wbr_create_ops->fin) + sbinfo->si_wbr_create_ops->fin(sb); + if (au_opt_test(sbinfo->si_mntflags, UDBA_HNOTIFY)) { + au_opt_set_udba(sbinfo->si_mntflags, UDBA_NONE); + au_remount_refresh(sb, /*do_idop*/0); + } + if (au_opt_test(sbinfo->si_mntflags, PLINK)) + au_plink_put(sb, /*verbose*/1); + au_xino_clr(sb); + au_dr_opt_flush(sb); + sbinfo->si_sb = NULL; + aufs_write_unlock(sb->s_root); + au_nwt_flush(&sbinfo->si_nowait); + } + kill_anon_super(sb); +} + +struct file_system_type aufs_fs_type = { + .name = AUFS_FSTYPE, + /* a race between rename and others */ + .fs_flags = FS_RENAME_DOES_D_MOVE + /* untested */ + /*| FS_ALLOW_IDMAP*/ + , + .mount = aufs_mount, + .kill_sb = aufs_kill_sb, + /* no need to __module_get() and module_put(). */ + .owner = THIS_MODULE, +}; diff --git a/fs/aufs/super.h b/fs/aufs/super.h new file mode 100644 index 00000000000..4638331366b --- /dev/null +++ b/fs/aufs/super.h @@ -0,0 +1,587 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * super_block operations + */ + +#ifndef __AUFS_SUPER_H__ +#define __AUFS_SUPER_H__ + +#ifdef __KERNEL__ + +#include +#include +#include "hbl.h" +#include "lcnt.h" +#include "rwsem.h" +#include "wkq.h" + +/* policies to select one among multiple writable branches */ +struct au_wbr_copyup_operations { + int (*copyup)(struct dentry *dentry); +}; + +#define AuWbr_DIR 1 /* target is a dir */ +#define AuWbr_PARENT (1 << 1) /* always require a parent */ + +#define au_ftest_wbr(flags, name) ((flags) & AuWbr_##name) +#define au_fset_wbr(flags, name) { (flags) |= AuWbr_##name; } +#define au_fclr_wbr(flags, name) { (flags) &= ~AuWbr_##name; } + +struct au_wbr_create_operations { + int (*create)(struct dentry *dentry, unsigned int flags); + int (*init)(struct super_block *sb); + int (*fin)(struct super_block *sb); +}; + +struct au_wbr_mfs { + struct mutex mfs_lock; /* protect this structure */ + unsigned long mfs_jiffy; + unsigned long mfs_expire; + aufs_bindex_t mfs_bindex; + + unsigned long long mfsrr_bytes; + unsigned long long mfsrr_watermark; +}; + +#define AuPlink_NHASH 100 +static inline int au_plink_hash(ino_t ino) +{ + return ino % AuPlink_NHASH; +} + +/* File-based Hierarchical Storage Management */ +struct au_fhsm { +#ifdef CONFIG_AUFS_FHSM + /* allow only one process who can receive the notification */ + spinlock_t fhsm_spin; + pid_t fhsm_pid; + wait_queue_head_t fhsm_wqh; + atomic_t fhsm_readable; + + /* these are protected by si_rwsem */ + unsigned long fhsm_expire; + aufs_bindex_t fhsm_bottom; +#endif +}; + +struct au_branch; +struct au_sbinfo { + /* nowait tasks in the system-wide workqueue */ + struct au_nowait_tasks si_nowait; + + /* + * tried sb->s_umount, but failed due to the dependency between i_mutex. + * rwsem for au_sbinfo is necessary. + */ + struct au_rwsem si_rwsem; + + /* + * dirty approach to protect sb->sb_inodes and ->s_files (gone) from + * remount. + */ + au_lcnt_t si_ninodes, si_nfiles; + + /* branch management */ + unsigned int si_generation; + + /* see AuSi_ flags */ + unsigned char au_si_status; + + aufs_bindex_t si_bbot; + + /* dirty trick to keep br_id plus */ + unsigned int si_last_br_id : + sizeof(aufs_bindex_t) * BITS_PER_BYTE - 1; + struct au_branch **si_branch; + + /* policy to select a writable branch */ + unsigned char si_wbr_copyup; + unsigned char si_wbr_create; + struct au_wbr_copyup_operations *si_wbr_copyup_ops; + struct au_wbr_create_operations *si_wbr_create_ops; + + /* round robin */ + atomic_t si_wbr_rr_next; + + /* most free space */ + struct au_wbr_mfs si_wbr_mfs; + + /* File-based Hierarchical Storage Management */ + struct au_fhsm si_fhsm; + + /* mount flags */ + /* include/asm-ia64/siginfo.h defines a macro named si_flags */ + unsigned int si_mntflags; + + /* external inode number (bitmap and translation table) */ + loff_t si_ximaxent; /* max entries in a xino */ + + struct file *si_xib; + struct mutex si_xib_mtx; /* protect xib members */ + unsigned long *si_xib_buf; + unsigned long si_xib_last_pindex; + int si_xib_next_bit; + + unsigned long si_xino_jiffy; + unsigned long si_xino_expire; + /* reserved for future use */ + /* unsigned long long si_xib_limit; */ /* Max xib file size */ + +#ifdef CONFIG_AUFS_EXPORT + /* i_generation */ + /* todo: make xigen file an array to support many inode numbers */ + struct file *si_xigen; + atomic_t si_xigen_next; +#endif + + /* dirty trick to support atomic_open */ + struct hlist_bl_head si_aopen; + + /* vdir parameters */ + unsigned long si_rdcache; /* max cache time in jiffies */ + unsigned int si_rdblk; /* deblk size */ + unsigned int si_rdhash; /* hash size */ + + /* + * If the number of whiteouts are larger than si_dirwh, leave all of + * them after au_whtmp_ren to reduce the cost of rmdir(2). + * future fsck.aufs or kernel thread will remove them later. + * Otherwise, remove all whiteouts and the dir in rmdir(2). + */ + unsigned int si_dirwh; + + /* pseudo_link list */ + struct hlist_bl_head si_plink[AuPlink_NHASH]; + wait_queue_head_t si_plink_wq; + spinlock_t si_plink_maint_lock; + pid_t si_plink_maint_pid; + + /* file list */ + struct hlist_bl_head si_files; + + /* with/without getattr, brother of sb->s_d_op */ + const struct inode_operations *si_iop_array; + + /* + * sysfs and lifetime management. + * this is not a small structure and it may be a waste of memory in case + * of sysfs is disabled, particularly when many aufs-es are mounted. + * but using sysfs is majority. + */ + struct kobject si_kobj; +#ifdef CONFIG_DEBUG_FS + struct dentry *si_dbgaufs; + struct dentry *si_dbgaufs_plink; + struct dentry *si_dbgaufs_xib; +#ifdef CONFIG_AUFS_EXPORT + struct dentry *si_dbgaufs_xigen; +#endif +#endif + +#ifdef CONFIG_AUFS_SBILIST + struct hlist_bl_node si_list; +#endif + + /* dirty, necessary for unmounting, sysfs and sysrq */ + struct super_block *si_sb; +}; + +/* sbinfo status flags */ +/* + * set true when refresh_dirs() failed at remount time. + * then try refreshing dirs at access time again. + * if it is false, refreshing dirs at access time is unnecessary + */ +#define AuSi_FAILED_REFRESH_DIR 1 +#define AuSi_FHSM (1 << 1) /* fhsm is active now */ +#define AuSi_NO_DREVAL (1 << 2) /* disable all d_revalidate */ + +#ifndef CONFIG_AUFS_FHSM +#undef AuSi_FHSM +#define AuSi_FHSM 0 +#endif + +static inline unsigned char au_do_ftest_si(struct au_sbinfo *sbi, + unsigned int flag) +{ + AuRwMustAnyLock(&sbi->si_rwsem); + return sbi->au_si_status & flag; +} +#define au_ftest_si(sbinfo, name) au_do_ftest_si(sbinfo, AuSi_##name) +#define au_fset_si(sbinfo, name) do { \ + AuRwMustWriteLock(&(sbinfo)->si_rwsem); \ + (sbinfo)->au_si_status |= AuSi_##name; \ +} while (0) +#define au_fclr_si(sbinfo, name) do { \ + AuRwMustWriteLock(&(sbinfo)->si_rwsem); \ + (sbinfo)->au_si_status &= ~AuSi_##name; \ +} while (0) + +/* ---------------------------------------------------------------------- */ + +/* policy to select one among writable branches */ +#define AuWbrCopyup(sbinfo, ...) \ + ((sbinfo)->si_wbr_copyup_ops->copyup(__VA_ARGS__)) +#define AuWbrCreate(sbinfo, ...) \ + ((sbinfo)->si_wbr_create_ops->create(__VA_ARGS__)) + +/* flags for si_read_lock()/aufs_read_lock()/di_read_lock() */ +#define AuLock_DW 1 /* write-lock dentry */ +#define AuLock_IR (1 << 1) /* read-lock inode */ +#define AuLock_IW (1 << 2) /* write-lock inode */ +#define AuLock_FLUSH (1 << 3) /* wait for 'nowait' tasks */ +#define AuLock_DIRS (1 << 4) /* target is a pair of dirs */ + /* except RENAME_EXCHANGE */ +#define AuLock_NOPLM (1 << 5) /* return err in plm mode */ +#define AuLock_NOPLMW (1 << 6) /* wait for plm mode ends */ +#define AuLock_GEN (1 << 7) /* test digen/iigen */ +#define au_ftest_lock(flags, name) ((flags) & AuLock_##name) +#define au_fset_lock(flags, name) \ + do { (flags) |= AuLock_##name; } while (0) +#define au_fclr_lock(flags, name) \ + do { (flags) &= ~AuLock_##name; } while (0) + +/* ---------------------------------------------------------------------- */ + +/* super.c */ +extern struct file_system_type aufs_fs_type; +struct inode *au_iget_locked(struct super_block *sb, ino_t ino); +typedef unsigned long long (*au_arraycb_t)(struct super_block *sb, void *array, + unsigned long long max, void *arg); +void *au_array_alloc(unsigned long long *hint, au_arraycb_t cb, + struct super_block *sb, void *arg); +struct inode **au_iarray_alloc(struct super_block *sb, unsigned long long *max); +void au_iarray_free(struct inode **a, unsigned long long max); + +/* sbinfo.c */ +void au_si_free(struct kobject *kobj); +int au_si_alloc(struct super_block *sb); +int au_sbr_realloc(struct au_sbinfo *sbinfo, int nbr, int may_shrink); + +unsigned int au_sigen_inc(struct super_block *sb); +aufs_bindex_t au_new_br_id(struct super_block *sb); + +int si_read_lock(struct super_block *sb, int flags); +int si_write_lock(struct super_block *sb, int flags); +int aufs_read_lock(struct dentry *dentry, int flags); +void aufs_read_unlock(struct dentry *dentry, int flags); +void aufs_write_lock(struct dentry *dentry); +void aufs_write_unlock(struct dentry *dentry); +int aufs_read_and_write_lock2(struct dentry *d1, struct dentry *d2, int flags); +void aufs_read_and_write_unlock2(struct dentry *d1, struct dentry *d2); + +/* wbr_policy.c */ +extern struct au_wbr_copyup_operations au_wbr_copyup_ops[]; +extern struct au_wbr_create_operations au_wbr_create_ops[]; +int au_cpdown_dirs(struct dentry *dentry, aufs_bindex_t bdst); +int au_wbr_nonopq(struct dentry *dentry, aufs_bindex_t bindex); +int au_wbr_do_copyup_bu(struct dentry *dentry, aufs_bindex_t btop); + +/* mvdown.c */ +int au_mvdown(struct dentry *dentry, struct aufs_mvdown __user *arg); + +#ifdef CONFIG_AUFS_FHSM +/* fhsm.c */ + +static inline pid_t au_fhsm_pid(struct au_fhsm *fhsm) +{ + pid_t pid; + + spin_lock(&fhsm->fhsm_spin); + pid = fhsm->fhsm_pid; + spin_unlock(&fhsm->fhsm_spin); + + return pid; +} + +void au_fhsm_wrote(struct super_block *sb, aufs_bindex_t bindex, int force); +void au_fhsm_wrote_all(struct super_block *sb, int force); +int au_fhsm_fd(struct super_block *sb, int oflags); +int au_fhsm_br_alloc(struct au_branch *br); +void au_fhsm_set_bottom(struct super_block *sb, aufs_bindex_t bindex); +void au_fhsm_fin(struct super_block *sb); +void au_fhsm_init(struct au_sbinfo *sbinfo); +void au_fhsm_set(struct au_sbinfo *sbinfo, unsigned int sec); +void au_fhsm_show(struct seq_file *seq, struct au_sbinfo *sbinfo); +#else +AuStubVoid(au_fhsm_wrote, struct super_block *sb, aufs_bindex_t bindex, + int force) +AuStubVoid(au_fhsm_wrote_all, struct super_block *sb, int force) +AuStub(int, au_fhsm_fd, return -EOPNOTSUPP, struct super_block *sb, int oflags) +AuStub(pid_t, au_fhsm_pid, return 0, struct au_fhsm *fhsm) +AuStubInt0(au_fhsm_br_alloc, struct au_branch *br) +AuStubVoid(au_fhsm_set_bottom, struct super_block *sb, aufs_bindex_t bindex) +AuStubVoid(au_fhsm_fin, struct super_block *sb) +AuStubVoid(au_fhsm_init, struct au_sbinfo *sbinfo) +AuStubVoid(au_fhsm_set, struct au_sbinfo *sbinfo, unsigned int sec) +AuStubVoid(au_fhsm_show, struct seq_file *seq, struct au_sbinfo *sbinfo) +#endif + +/* ---------------------------------------------------------------------- */ + +static inline struct au_sbinfo *au_sbi(struct super_block *sb) +{ + return sb->s_fs_info; +} + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_EXPORT +int au_test_nfsd(void); +void au_export_init(struct super_block *sb); +void au_xigen_inc(struct inode *inode); +int au_xigen_new(struct inode *inode); +int au_xigen_set(struct super_block *sb, struct path *path); +void au_xigen_clr(struct super_block *sb); + +static inline int au_busy_or_stale(void) +{ + if (!au_test_nfsd()) + return -EBUSY; + return -ESTALE; +} +#else +AuStubInt0(au_test_nfsd, void) +AuStubVoid(au_export_init, struct super_block *sb) +AuStubVoid(au_xigen_inc, struct inode *inode) +AuStubInt0(au_xigen_new, struct inode *inode) +AuStubInt0(au_xigen_set, struct super_block *sb, struct path *path) +AuStubVoid(au_xigen_clr, struct super_block *sb) +AuStub(int, au_busy_or_stale, return -EBUSY, void) +#endif /* CONFIG_AUFS_EXPORT */ + +/* ---------------------------------------------------------------------- */ + +#ifdef CONFIG_AUFS_SBILIST +/* module.c */ +extern struct hlist_bl_head au_sbilist; + +static inline void au_sbilist_init(void) +{ + INIT_HLIST_BL_HEAD(&au_sbilist); +} + +static inline void au_sbilist_add(struct super_block *sb) +{ + au_hbl_add(&au_sbi(sb)->si_list, &au_sbilist); +} + +static inline void au_sbilist_del(struct super_block *sb) +{ + au_hbl_del(&au_sbi(sb)->si_list, &au_sbilist); +} + +#ifdef CONFIG_AUFS_MAGIC_SYSRQ +static inline void au_sbilist_lock(void) +{ + hlist_bl_lock(&au_sbilist); +} + +static inline void au_sbilist_unlock(void) +{ + hlist_bl_unlock(&au_sbilist); +} +#define AuGFP_SBILIST GFP_ATOMIC +#else +AuStubVoid(au_sbilist_lock, void) +AuStubVoid(au_sbilist_unlock, void) +#define AuGFP_SBILIST GFP_NOFS +#endif /* CONFIG_AUFS_MAGIC_SYSRQ */ +#else +AuStubVoid(au_sbilist_init, void) +AuStubVoid(au_sbilist_add, struct super_block *sb) +AuStubVoid(au_sbilist_del, struct super_block *sb) +AuStubVoid(au_sbilist_lock, void) +AuStubVoid(au_sbilist_unlock, void) +#define AuGFP_SBILIST GFP_NOFS +#endif + +/* ---------------------------------------------------------------------- */ + +static inline void dbgaufs_si_null(struct au_sbinfo *sbinfo) +{ + /* + * This function is a dynamic '__init' function actually, + * so the tiny check for si_rwsem is unnecessary. + */ + /* AuRwMustWriteLock(&sbinfo->si_rwsem); */ +#ifdef CONFIG_DEBUG_FS + sbinfo->si_dbgaufs = NULL; + sbinfo->si_dbgaufs_plink = NULL; + sbinfo->si_dbgaufs_xib = NULL; +#ifdef CONFIG_AUFS_EXPORT + sbinfo->si_dbgaufs_xigen = NULL; +#endif +#endif +} + +/* ---------------------------------------------------------------------- */ + +/* current->atomic_flags */ +/* this value should never corrupt the ones defined in linux/sched.h */ +#define PFA_AUFS 0x10 + +TASK_PFA_TEST(AUFS, test_aufs) /* task_test_aufs */ +TASK_PFA_SET(AUFS, aufs) /* task_set_aufs */ +TASK_PFA_CLEAR(AUFS, aufs) /* task_clear_aufs */ + +static inline int si_pid_test(struct super_block *sb) +{ + return !!task_test_aufs(current); +} + +static inline void si_pid_clr(struct super_block *sb) +{ + AuDebugOn(!task_test_aufs(current)); + task_clear_aufs(current); +} + +static inline void si_pid_set(struct super_block *sb) +{ + AuDebugOn(task_test_aufs(current)); + task_set_aufs(current); +} + +/* ---------------------------------------------------------------------- */ + +/* lock superblock. mainly for entry point functions */ +#define __si_read_lock(sb) au_rw_read_lock(&au_sbi(sb)->si_rwsem) +#define __si_write_lock(sb) au_rw_write_lock(&au_sbi(sb)->si_rwsem) +#define __si_read_trylock(sb) au_rw_read_trylock(&au_sbi(sb)->si_rwsem) +#define __si_write_trylock(sb) au_rw_write_trylock(&au_sbi(sb)->si_rwsem) +/* +#define __si_read_trylock_nested(sb) \ + au_rw_read_trylock_nested(&au_sbi(sb)->si_rwsem) +#define __si_write_trylock_nested(sb) \ + au_rw_write_trylock_nested(&au_sbi(sb)->si_rwsem) +*/ + +#define __si_read_unlock(sb) au_rw_read_unlock(&au_sbi(sb)->si_rwsem) +#define __si_write_unlock(sb) au_rw_write_unlock(&au_sbi(sb)->si_rwsem) +#define __si_downgrade_lock(sb) au_rw_dgrade_lock(&au_sbi(sb)->si_rwsem) + +#define SiMustNoWaiters(sb) AuRwMustNoWaiters(&au_sbi(sb)->si_rwsem) +#define SiMustAnyLock(sb) AuRwMustAnyLock(&au_sbi(sb)->si_rwsem) +#define SiMustWriteLock(sb) AuRwMustWriteLock(&au_sbi(sb)->si_rwsem) + +static inline void si_noflush_read_lock(struct super_block *sb) +{ + __si_read_lock(sb); + si_pid_set(sb); +} + +static inline int si_noflush_read_trylock(struct super_block *sb) +{ + int locked; + + locked = __si_read_trylock(sb); + if (locked) + si_pid_set(sb); + return locked; +} + +static inline void si_noflush_write_lock(struct super_block *sb) +{ + __si_write_lock(sb); + si_pid_set(sb); +} + +static inline int si_noflush_write_trylock(struct super_block *sb) +{ + int locked; + + locked = __si_write_trylock(sb); + if (locked) + si_pid_set(sb); + return locked; +} + +#if 0 /* reserved */ +static inline int si_read_trylock(struct super_block *sb, int flags) +{ + if (au_ftest_lock(flags, FLUSH)) + au_nwt_flush(&au_sbi(sb)->si_nowait); + return si_noflush_read_trylock(sb); +} +#endif + +static inline void si_read_unlock(struct super_block *sb) +{ + si_pid_clr(sb); + __si_read_unlock(sb); +} + +#if 0 /* reserved */ +static inline int si_write_trylock(struct super_block *sb, int flags) +{ + if (au_ftest_lock(flags, FLUSH)) + au_nwt_flush(&au_sbi(sb)->si_nowait); + return si_noflush_write_trylock(sb); +} +#endif + +static inline void si_write_unlock(struct super_block *sb) +{ + si_pid_clr(sb); + __si_write_unlock(sb); +} + +#if 0 /* reserved */ +static inline void si_downgrade_lock(struct super_block *sb) +{ + __si_downgrade_lock(sb); +} +#endif + +/* ---------------------------------------------------------------------- */ + +static inline aufs_bindex_t au_sbbot(struct super_block *sb) +{ + SiMustAnyLock(sb); + return au_sbi(sb)->si_bbot; +} + +static inline unsigned int au_mntflags(struct super_block *sb) +{ + SiMustAnyLock(sb); + return au_sbi(sb)->si_mntflags; +} + +static inline unsigned int au_sigen(struct super_block *sb) +{ + SiMustAnyLock(sb); + return au_sbi(sb)->si_generation; +} + +static inline struct au_branch *au_sbr(struct super_block *sb, + aufs_bindex_t bindex) +{ + SiMustAnyLock(sb); + return au_sbi(sb)->si_branch[0 + bindex]; +} + +static inline loff_t au_xi_maxent(struct super_block *sb) +{ + SiMustAnyLock(sb); + return au_sbi(sb)->si_ximaxent; +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_SUPER_H__ */ diff --git a/fs/aufs/sysaufs.c b/fs/aufs/sysaufs.c new file mode 100644 index 00000000000..f615471231b --- /dev/null +++ b/fs/aufs/sysaufs.c @@ -0,0 +1,93 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sysfs interface and lifetime management + * they are necessary regardless sysfs is disabled. + */ + +#include +#include "aufs.h" + +unsigned long sysaufs_si_mask; +struct kset *sysaufs_kset; + +#define AuSiAttr(_name) { \ + .attr = { .name = __stringify(_name), .mode = 0444 }, \ + .show = sysaufs_si_##_name, \ +} + +static struct sysaufs_si_attr sysaufs_si_attr_xi_path = AuSiAttr(xi_path); +struct attribute *sysaufs_si_attrs[] = { + &sysaufs_si_attr_xi_path.attr, + NULL, +}; + +static const struct sysfs_ops au_sbi_ops = { + .show = sysaufs_si_show +}; + +static struct kobj_type au_sbi_ktype = { + .release = au_si_free, + .sysfs_ops = &au_sbi_ops, + .default_attrs = sysaufs_si_attrs +}; + +/* ---------------------------------------------------------------------- */ + +int sysaufs_si_init(struct au_sbinfo *sbinfo) +{ + int err; + + sbinfo->si_kobj.kset = sysaufs_kset; + /* cf. sysaufs_name() */ + err = kobject_init_and_add + (&sbinfo->si_kobj, &au_sbi_ktype, /*&sysaufs_kset->kobj*/NULL, + SysaufsSiNamePrefix "%lx", sysaufs_si_id(sbinfo)); + + return err; +} + +void sysaufs_fin(void) +{ + sysfs_remove_group(&sysaufs_kset->kobj, sysaufs_attr_group); + kset_unregister(sysaufs_kset); +} + +int __init sysaufs_init(void) +{ + int err; + + do { + get_random_bytes(&sysaufs_si_mask, sizeof(sysaufs_si_mask)); + } while (!sysaufs_si_mask); + + err = -EINVAL; + sysaufs_kset = kset_create_and_add(AUFS_NAME, NULL, fs_kobj); + if (unlikely(!sysaufs_kset)) + goto out; + err = PTR_ERR(sysaufs_kset); + if (IS_ERR(sysaufs_kset)) + goto out; + err = sysfs_create_group(&sysaufs_kset->kobj, sysaufs_attr_group); + if (unlikely(err)) + kset_unregister(sysaufs_kset); + +out: + return err; +} diff --git a/fs/aufs/sysaufs.h b/fs/aufs/sysaufs.h new file mode 100644 index 00000000000..55c5c9868f3 --- /dev/null +++ b/fs/aufs/sysaufs.h @@ -0,0 +1,102 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sysfs interface and mount lifetime management + */ + +#ifndef __SYSAUFS_H__ +#define __SYSAUFS_H__ + +#ifdef __KERNEL__ + +#include +#include "module.h" + +struct super_block; +struct au_sbinfo; + +struct sysaufs_si_attr { + struct attribute attr; + int (*show)(struct seq_file *seq, struct super_block *sb); +}; + +/* ---------------------------------------------------------------------- */ + +/* sysaufs.c */ +extern unsigned long sysaufs_si_mask; +extern struct kset *sysaufs_kset; +extern struct attribute *sysaufs_si_attrs[]; +int sysaufs_si_init(struct au_sbinfo *sbinfo); +int __init sysaufs_init(void); +void sysaufs_fin(void); + +/* ---------------------------------------------------------------------- */ + +/* some people doesn't like to show a pointer in kernel */ +static inline unsigned long sysaufs_si_id(struct au_sbinfo *sbinfo) +{ + return sysaufs_si_mask ^ (unsigned long)sbinfo; +} + +#define SysaufsSiNamePrefix "si_" +#define SysaufsSiNameLen (sizeof(SysaufsSiNamePrefix) + 16) +static inline void sysaufs_name(struct au_sbinfo *sbinfo, char *name) +{ + snprintf(name, SysaufsSiNameLen, SysaufsSiNamePrefix "%lx", + sysaufs_si_id(sbinfo)); +} + +struct au_branch; +#ifdef CONFIG_SYSFS +/* sysfs.c */ +extern struct attribute_group *sysaufs_attr_group; + +int sysaufs_si_xi_path(struct seq_file *seq, struct super_block *sb); +ssize_t sysaufs_si_show(struct kobject *kobj, struct attribute *attr, + char *buf); +long au_brinfo_ioctl(struct file *file, unsigned long arg); +#ifdef CONFIG_COMPAT +long au_brinfo_compat_ioctl(struct file *file, unsigned long arg); +#endif + +void sysaufs_br_init(struct au_branch *br); +void sysaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex); +void sysaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex); + +#define sysaufs_brs_init() do {} while (0) + +#else +#define sysaufs_attr_group NULL + +AuStubInt0(sysaufs_si_xi_path, struct seq_file *seq, struct super_block *sb) +AuStub(ssize_t, sysaufs_si_show, return 0, struct kobject *kobj, + struct attribute *attr, char *buf) +AuStubVoid(sysaufs_br_init, struct au_branch *br) +AuStubVoid(sysaufs_brs_add, struct super_block *sb, aufs_bindex_t bindex) +AuStubVoid(sysaufs_brs_del, struct super_block *sb, aufs_bindex_t bindex) + +static inline void sysaufs_brs_init(void) +{ + sysaufs_brs = 0; +} + +#endif /* CONFIG_SYSFS */ + +#endif /* __KERNEL__ */ +#endif /* __SYSAUFS_H__ */ diff --git a/fs/aufs/sysfs.c b/fs/aufs/sysfs.c new file mode 100644 index 00000000000..46638d88ef2 --- /dev/null +++ b/fs/aufs/sysfs.c @@ -0,0 +1,374 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sysfs interface + */ + +#include +#include +#include "aufs.h" + +#ifdef CONFIG_AUFS_FS_MODULE +/* this entry violates the "one line per file" policy of sysfs */ +static ssize_t config_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + ssize_t err; + static char *conf = +/* this file is generated at compiling */ +#include "conf.str" + ; + + err = snprintf(buf, PAGE_SIZE, conf); + if (unlikely(err >= PAGE_SIZE)) + err = -EFBIG; + return err; +} + +static struct kobj_attribute au_config_attr = __ATTR_RO(config); +#endif + +static struct attribute *au_attr[] = { +#ifdef CONFIG_AUFS_FS_MODULE + &au_config_attr.attr, +#endif + NULL, /* need to NULL terminate the list of attributes */ +}; + +static struct attribute_group sysaufs_attr_group_body = { + .attrs = au_attr +}; + +struct attribute_group *sysaufs_attr_group = &sysaufs_attr_group_body; + +/* ---------------------------------------------------------------------- */ + +int sysaufs_si_xi_path(struct seq_file *seq, struct super_block *sb) +{ + int err; + + SiMustAnyLock(sb); + + err = 0; + if (au_opt_test(au_mntflags(sb), XINO)) { + err = au_xino_path(seq, au_sbi(sb)->si_xib); + seq_putc(seq, '\n'); + } + return err; +} + +/* + * the lifetime of branch is independent from the entry under sysfs. + * sysfs handles the lifetime of the entry, and never call ->show() after it is + * unlinked. + */ +static int sysaufs_si_br(struct seq_file *seq, struct super_block *sb, + aufs_bindex_t bindex, int idx) +{ + int err; + struct path path; + struct dentry *root; + struct au_branch *br; + au_br_perm_str_t perm; + + AuDbg("b%d\n", bindex); + + err = 0; + root = sb->s_root; + di_read_lock_parent(root, !AuLock_IR); + br = au_sbr(sb, bindex); + + switch (idx) { + case AuBrSysfs_BR: + path.mnt = au_br_mnt(br); + path.dentry = au_h_dptr(root, bindex); + err = au_seq_path(seq, &path); + if (!err) { + au_optstr_br_perm(&perm, br->br_perm); + seq_printf(seq, "=%s\n", perm.a); + } + break; + case AuBrSysfs_BRID: + seq_printf(seq, "%d\n", br->br_id); + break; + } + di_read_unlock(root, !AuLock_IR); + if (unlikely(err || seq_has_overflowed(seq))) + err = -E2BIG; + + return err; +} + +/* ---------------------------------------------------------------------- */ + +static struct seq_file *au_seq(char *p, ssize_t len) +{ + struct seq_file *seq; + + seq = kzalloc(sizeof(*seq), GFP_NOFS); + if (seq) { + /* mutex_init(&seq.lock); */ + seq->buf = p; + seq->size = len; + return seq; /* success */ + } + + seq = ERR_PTR(-ENOMEM); + return seq; +} + +#define SysaufsBr_PREFIX "br" +#define SysaufsBrid_PREFIX "brid" + +/* todo: file size may exceed PAGE_SIZE */ +ssize_t sysaufs_si_show(struct kobject *kobj, struct attribute *attr, + char *buf) +{ + ssize_t err; + int idx; + long l; + aufs_bindex_t bbot; + struct au_sbinfo *sbinfo; + struct super_block *sb; + struct seq_file *seq; + char *name; + struct attribute **cattr; + + sbinfo = container_of(kobj, struct au_sbinfo, si_kobj); + sb = sbinfo->si_sb; + + /* + * prevent a race condition between sysfs and aufs. + * for instance, sysfs_file_read() calls sysfs_get_active_two() which + * prohibits maintaining the sysfs entries. + * hew we acquire read lock after sysfs_get_active_two(). + * on the other hand, the remount process may maintain the sysfs/aufs + * entries after acquiring write lock. + * it can cause a deadlock. + * simply we gave up processing read here. + */ + err = -EBUSY; + if (unlikely(!si_noflush_read_trylock(sb))) + goto out; + + seq = au_seq(buf, PAGE_SIZE); + err = PTR_ERR(seq); + if (IS_ERR(seq)) + goto out_unlock; + + name = (void *)attr->name; + cattr = sysaufs_si_attrs; + while (*cattr) { + if (!strcmp(name, (*cattr)->name)) { + err = container_of(*cattr, struct sysaufs_si_attr, attr) + ->show(seq, sb); + goto out_seq; + } + cattr++; + } + + if (!strncmp(name, SysaufsBrid_PREFIX, + sizeof(SysaufsBrid_PREFIX) - 1)) { + idx = AuBrSysfs_BRID; + name += sizeof(SysaufsBrid_PREFIX) - 1; + } else if (!strncmp(name, SysaufsBr_PREFIX, + sizeof(SysaufsBr_PREFIX) - 1)) { + idx = AuBrSysfs_BR; + name += sizeof(SysaufsBr_PREFIX) - 1; + } else + BUG(); + + err = kstrtol(name, 10, &l); + if (!err) { + bbot = au_sbbot(sb); + if (l <= bbot) + err = sysaufs_si_br(seq, sb, (aufs_bindex_t)l, idx); + else + err = -ENOENT; + } + +out_seq: + if (!err) { + err = seq->count; + /* sysfs limit */ + if (unlikely(err == PAGE_SIZE)) + err = -EFBIG; + } + au_kfree_rcu(seq); +out_unlock: + si_read_unlock(sb); +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_brinfo(struct super_block *sb, union aufs_brinfo __user *arg) +{ + int err; + int16_t brid; + aufs_bindex_t bindex, bbot; + size_t sz; + char *buf; + struct seq_file *seq; + struct au_branch *br; + + si_read_lock(sb, AuLock_FLUSH); + bbot = au_sbbot(sb); + err = bbot + 1; + if (!arg) + goto out; + + err = -ENOMEM; + buf = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!buf)) + goto out; + + seq = au_seq(buf, PAGE_SIZE); + err = PTR_ERR(seq); + if (IS_ERR(seq)) + goto out_buf; + + sz = sizeof(*arg) - offsetof(union aufs_brinfo, path); + for (bindex = 0; bindex <= bbot; bindex++, arg++) { + /* VERIFY_WRITE */ + err = !access_ok(arg, sizeof(*arg)); + if (unlikely(err)) + break; + + br = au_sbr(sb, bindex); + brid = br->br_id; + BUILD_BUG_ON(sizeof(brid) != sizeof(arg->id)); + err = __put_user(brid, &arg->id); + if (unlikely(err)) + break; + + BUILD_BUG_ON(sizeof(br->br_perm) != sizeof(arg->perm)); + err = __put_user(br->br_perm, &arg->perm); + if (unlikely(err)) + break; + + err = au_seq_path(seq, &br->br_path); + if (unlikely(err)) + break; + seq_putc(seq, '\0'); + if (!seq_has_overflowed(seq)) { + err = copy_to_user(arg->path, seq->buf, seq->count); + seq->count = 0; + if (unlikely(err)) + break; + } else { + err = -E2BIG; + goto out_seq; + } + } + if (unlikely(err)) + err = -EFAULT; + +out_seq: + au_kfree_rcu(seq); +out_buf: + free_page((unsigned long)buf); +out: + si_read_unlock(sb); + return err; +} + +long au_brinfo_ioctl(struct file *file, unsigned long arg) +{ + return au_brinfo(file->f_path.dentry->d_sb, (void __user *)arg); +} + +#ifdef CONFIG_COMPAT +long au_brinfo_compat_ioctl(struct file *file, unsigned long arg) +{ + return au_brinfo(file->f_path.dentry->d_sb, compat_ptr(arg)); +} +#endif + +/* ---------------------------------------------------------------------- */ + +void sysaufs_br_init(struct au_branch *br) +{ + int i; + struct au_brsysfs *br_sysfs; + struct attribute *attr; + + br_sysfs = br->br_sysfs; + for (i = 0; i < ARRAY_SIZE(br->br_sysfs); i++) { + attr = &br_sysfs->attr; + sysfs_attr_init(attr); + attr->name = br_sysfs->name; + attr->mode = 0444; + br_sysfs++; + } +} + +void sysaufs_brs_del(struct super_block *sb, aufs_bindex_t bindex) +{ + struct au_branch *br; + struct kobject *kobj; + struct au_brsysfs *br_sysfs; + int i; + aufs_bindex_t bbot; + + if (!sysaufs_brs) + return; + + kobj = &au_sbi(sb)->si_kobj; + bbot = au_sbbot(sb); + for (; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + br_sysfs = br->br_sysfs; + for (i = 0; i < ARRAY_SIZE(br->br_sysfs); i++) { + sysfs_remove_file(kobj, &br_sysfs->attr); + br_sysfs++; + } + } +} + +void sysaufs_brs_add(struct super_block *sb, aufs_bindex_t bindex) +{ + int err, i; + aufs_bindex_t bbot; + struct kobject *kobj; + struct au_branch *br; + struct au_brsysfs *br_sysfs; + + if (!sysaufs_brs) + return; + + kobj = &au_sbi(sb)->si_kobj; + bbot = au_sbbot(sb); + for (; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + br_sysfs = br->br_sysfs; + snprintf(br_sysfs[AuBrSysfs_BR].name, sizeof(br_sysfs->name), + SysaufsBr_PREFIX "%d", bindex); + snprintf(br_sysfs[AuBrSysfs_BRID].name, sizeof(br_sysfs->name), + SysaufsBrid_PREFIX "%d", bindex); + for (i = 0; i < ARRAY_SIZE(br->br_sysfs); i++) { + err = sysfs_create_file(kobj, &br_sysfs->attr); + if (unlikely(err)) + pr_warn("failed %s under sysfs(%d)\n", + br_sysfs->name, err); + br_sysfs++; + } + } +} diff --git a/fs/aufs/sysrq.c b/fs/aufs/sysrq.c new file mode 100644 index 00000000000..fc8aa4a2839 --- /dev/null +++ b/fs/aufs/sysrq.c @@ -0,0 +1,149 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * magic sysrq handler + */ + +/* #include */ +#include +#include "aufs.h" + +/* ---------------------------------------------------------------------- */ + +static void sysrq_sb(struct super_block *sb) +{ + char *plevel; + struct au_sbinfo *sbinfo; + struct file *file; + struct hlist_bl_head *files; + struct hlist_bl_node *pos; + struct au_finfo *finfo; + struct inode *i; + + plevel = au_plevel; + au_plevel = KERN_WARNING; + + /* since we define pr_fmt, call printk directly */ +#define pr(str) printk(KERN_WARNING AUFS_NAME ": " str) + + sbinfo = au_sbi(sb); + printk(KERN_WARNING "si=%lx\n", sysaufs_si_id(sbinfo)); + pr("superblock\n"); + au_dpri_sb(sb); + +#if 0 /* reserved */ + do { + int err, i, j, ndentry; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + + err = au_dpages_init(&dpages, GFP_ATOMIC); + if (unlikely(err)) + break; + err = au_dcsub_pages(&dpages, sb->s_root, NULL, NULL); + if (!err) + for (i = 0; i < dpages.ndpage; i++) { + dpage = dpages.dpages + i; + ndentry = dpage->ndentry; + for (j = 0; j < ndentry; j++) + au_dpri_dentry(dpage->dentries[j]); + } + au_dpages_free(&dpages); + } while (0); +#endif + + pr("isolated inode\n"); + spin_lock(&sb->s_inode_list_lock); + list_for_each_entry(i, &sb->s_inodes, i_sb_list) { + spin_lock(&i->i_lock); + if (hlist_empty(&i->i_dentry)) + au_dpri_inode(i); + spin_unlock(&i->i_lock); + } + spin_unlock(&sb->s_inode_list_lock); + + pr("files\n"); + files = &au_sbi(sb)->si_files; + hlist_bl_lock(files); + hlist_bl_for_each_entry(finfo, pos, files, fi_hlist) { + umode_t mode; + + file = finfo->fi_file; + mode = file_inode(file)->i_mode; + if (!special_file(mode)) + au_dpri_file(file); + } + hlist_bl_unlock(files); + pr("done\n"); + +#undef pr + au_plevel = plevel; +} + +/* ---------------------------------------------------------------------- */ + +/* module parameter */ +static char *aufs_sysrq_key = "a"; +module_param_named(sysrq, aufs_sysrq_key, charp, 0444); +MODULE_PARM_DESC(sysrq, "MagicSysRq key for " AUFS_NAME); + +static void au_sysrq(int key __maybe_unused) +{ + struct au_sbinfo *sbinfo; + struct hlist_bl_node *pos; + + lockdep_off(); + au_sbilist_lock(); + hlist_bl_for_each_entry(sbinfo, pos, &au_sbilist, si_list) + sysrq_sb(sbinfo->si_sb); + au_sbilist_unlock(); + lockdep_on(); +} + +static struct sysrq_key_op au_sysrq_op = { + .handler = au_sysrq, + .help_msg = "Aufs", + .action_msg = "Aufs", + .enable_mask = SYSRQ_ENABLE_DUMP +}; + +/* ---------------------------------------------------------------------- */ + +int __init au_sysrq_init(void) +{ + int err; + char key; + + err = -1; + key = *aufs_sysrq_key; + if ('a' <= key && key <= 'z') + err = register_sysrq_key(key, &au_sysrq_op); + if (unlikely(err)) + pr_err("err %d, sysrq=%c\n", err, key); + return err; +} + +void au_sysrq_fin(void) +{ + int err; + + err = unregister_sysrq_key(*aufs_sysrq_key, &au_sysrq_op); + if (unlikely(err)) + pr_err("err %d (ignored)\n", err); +} diff --git a/fs/aufs/vdir.c b/fs/aufs/vdir.c new file mode 100644 index 00000000000..5ba006b8072 --- /dev/null +++ b/fs/aufs/vdir.c @@ -0,0 +1,896 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * virtual or vertical directory + */ + +#include +#include "aufs.h" + +static unsigned int calc_size(int nlen) +{ + return ALIGN(sizeof(struct au_vdir_de) + nlen, sizeof(ino_t)); +} + +static int set_deblk_end(union au_vdir_deblk_p *p, + union au_vdir_deblk_p *deblk_end) +{ + if (calc_size(0) <= deblk_end->deblk - p->deblk) { + p->de->de_str.len = 0; + /* smp_mb(); */ + return 0; + } + return -1; /* error */ +} + +/* returns true or false */ +static int is_deblk_end(union au_vdir_deblk_p *p, + union au_vdir_deblk_p *deblk_end) +{ + if (calc_size(0) <= deblk_end->deblk - p->deblk) + return !p->de->de_str.len; + return 1; +} + +static unsigned char *last_deblk(struct au_vdir *vdir) +{ + return vdir->vd_deblk[vdir->vd_nblk - 1]; +} + +/* ---------------------------------------------------------------------- */ + +/* estimate the appropriate size for name hash table */ +unsigned int au_rdhash_est(loff_t sz) +{ + unsigned int n; + + n = UINT_MAX; + sz >>= 10; + if (sz < n) + n = sz; + if (sz < AUFS_RDHASH_DEF) + n = AUFS_RDHASH_DEF; + /* pr_info("n %u\n", n); */ + return n; +} + +/* + * the allocated memory has to be freed by + * au_nhash_wh_free() or au_nhash_de_free(). + */ +int au_nhash_alloc(struct au_nhash *nhash, unsigned int num_hash, gfp_t gfp) +{ + struct hlist_head *head; + unsigned int u; + size_t sz; + + sz = sizeof(*nhash->nh_head) * num_hash; + head = kmalloc(sz, gfp); + if (head) { + nhash->nh_num = num_hash; + nhash->nh_head = head; + for (u = 0; u < num_hash; u++) + INIT_HLIST_HEAD(head++); + return 0; /* success */ + } + + return -ENOMEM; +} + +static void nhash_count(struct hlist_head *head) +{ +#if 0 /* debugging */ + unsigned long n; + struct hlist_node *pos; + + n = 0; + hlist_for_each(pos, head) + n++; + pr_info("%lu\n", n); +#endif +} + +static void au_nhash_wh_do_free(struct hlist_head *head) +{ + struct au_vdir_wh *pos; + struct hlist_node *node; + + hlist_for_each_entry_safe(pos, node, head, wh_hash) + au_kfree_rcu(pos); +} + +static void au_nhash_de_do_free(struct hlist_head *head) +{ + struct au_vdir_dehstr *pos; + struct hlist_node *node; + + hlist_for_each_entry_safe(pos, node, head, hash) + au_cache_free_vdir_dehstr(pos); +} + +static void au_nhash_do_free(struct au_nhash *nhash, + void (*free)(struct hlist_head *head)) +{ + unsigned int n; + struct hlist_head *head; + + n = nhash->nh_num; + if (!n) + return; + + head = nhash->nh_head; + while (n-- > 0) { + nhash_count(head); + free(head++); + } + au_kfree_try_rcu(nhash->nh_head); +} + +void au_nhash_wh_free(struct au_nhash *whlist) +{ + au_nhash_do_free(whlist, au_nhash_wh_do_free); +} + +static void au_nhash_de_free(struct au_nhash *delist) +{ + au_nhash_do_free(delist, au_nhash_de_do_free); +} + +/* ---------------------------------------------------------------------- */ + +int au_nhash_test_longer_wh(struct au_nhash *whlist, aufs_bindex_t btgt, + int limit) +{ + int num; + unsigned int u, n; + struct hlist_head *head; + struct au_vdir_wh *pos; + + num = 0; + n = whlist->nh_num; + head = whlist->nh_head; + for (u = 0; u < n; u++, head++) + hlist_for_each_entry(pos, head, wh_hash) + if (pos->wh_bindex == btgt && ++num > limit) + return 1; + return 0; +} + +static struct hlist_head *au_name_hash(struct au_nhash *nhash, + unsigned char *name, + unsigned int len) +{ + unsigned int v; + /* const unsigned int magic_bit = 12; */ + + AuDebugOn(!nhash->nh_num || !nhash->nh_head); + + v = 0; + if (len > 8) + len = 8; + while (len--) + v += *name++; + /* v = hash_long(v, magic_bit); */ + v %= nhash->nh_num; + return nhash->nh_head + v; +} + +static int au_nhash_test_name(struct au_vdir_destr *str, const char *name, + int nlen) +{ + return str->len == nlen && !memcmp(str->name, name, nlen); +} + +/* returns found or not */ +int au_nhash_test_known_wh(struct au_nhash *whlist, char *name, int nlen) +{ + struct hlist_head *head; + struct au_vdir_wh *pos; + struct au_vdir_destr *str; + + head = au_name_hash(whlist, name, nlen); + hlist_for_each_entry(pos, head, wh_hash) { + str = &pos->wh_str; + AuDbg("%.*s\n", str->len, str->name); + if (au_nhash_test_name(str, name, nlen)) + return 1; + } + return 0; +} + +/* returns found(true) or not */ +static int test_known(struct au_nhash *delist, char *name, int nlen) +{ + struct hlist_head *head; + struct au_vdir_dehstr *pos; + struct au_vdir_destr *str; + + head = au_name_hash(delist, name, nlen); + hlist_for_each_entry(pos, head, hash) { + str = pos->str; + AuDbg("%.*s\n", str->len, str->name); + if (au_nhash_test_name(str, name, nlen)) + return 1; + } + return 0; +} + +static void au_shwh_init_wh(struct au_vdir_wh *wh, ino_t ino, + unsigned char d_type) +{ +#ifdef CONFIG_AUFS_SHWH + wh->wh_ino = ino; + wh->wh_type = d_type; +#endif +} + +/* ---------------------------------------------------------------------- */ + +int au_nhash_append_wh(struct au_nhash *whlist, char *name, int nlen, ino_t ino, + unsigned int d_type, aufs_bindex_t bindex, + unsigned char shwh) +{ + int err; + struct au_vdir_destr *str; + struct au_vdir_wh *wh; + + AuDbg("%.*s\n", nlen, name); + AuDebugOn(!whlist->nh_num || !whlist->nh_head); + + err = -ENOMEM; + wh = kmalloc(sizeof(*wh) + nlen, GFP_NOFS); + if (unlikely(!wh)) + goto out; + + err = 0; + wh->wh_bindex = bindex; + if (shwh) + au_shwh_init_wh(wh, ino, d_type); + str = &wh->wh_str; + str->len = nlen; + memcpy(str->name, name, nlen); + hlist_add_head(&wh->wh_hash, au_name_hash(whlist, name, nlen)); + /* smp_mb(); */ + +out: + return err; +} + +static int append_deblk(struct au_vdir *vdir) +{ + int err; + unsigned long ul; + const unsigned int deblk_sz = vdir->vd_deblk_sz; + union au_vdir_deblk_p p, deblk_end; + unsigned char **o; + + err = -ENOMEM; + o = au_krealloc(vdir->vd_deblk, sizeof(*o) * (vdir->vd_nblk + 1), + GFP_NOFS, /*may_shrink*/0); + if (unlikely(!o)) + goto out; + + vdir->vd_deblk = o; + p.deblk = kmalloc(deblk_sz, GFP_NOFS); + if (p.deblk) { + ul = vdir->vd_nblk++; + vdir->vd_deblk[ul] = p.deblk; + vdir->vd_last.ul = ul; + vdir->vd_last.p.deblk = p.deblk; + deblk_end.deblk = p.deblk + deblk_sz; + err = set_deblk_end(&p, &deblk_end); + } + +out: + return err; +} + +static int append_de(struct au_vdir *vdir, char *name, int nlen, ino_t ino, + unsigned int d_type, struct au_nhash *delist) +{ + int err; + unsigned int sz; + const unsigned int deblk_sz = vdir->vd_deblk_sz; + union au_vdir_deblk_p p, *room, deblk_end; + struct au_vdir_dehstr *dehstr; + + p.deblk = last_deblk(vdir); + deblk_end.deblk = p.deblk + deblk_sz; + room = &vdir->vd_last.p; + AuDebugOn(room->deblk < p.deblk || deblk_end.deblk <= room->deblk + || !is_deblk_end(room, &deblk_end)); + + sz = calc_size(nlen); + if (unlikely(sz > deblk_end.deblk - room->deblk)) { + err = append_deblk(vdir); + if (unlikely(err)) + goto out; + + p.deblk = last_deblk(vdir); + deblk_end.deblk = p.deblk + deblk_sz; + /* smp_mb(); */ + AuDebugOn(room->deblk != p.deblk); + } + + err = -ENOMEM; + dehstr = au_cache_alloc_vdir_dehstr(); + if (unlikely(!dehstr)) + goto out; + + dehstr->str = &room->de->de_str; + hlist_add_head(&dehstr->hash, au_name_hash(delist, name, nlen)); + room->de->de_ino = ino; + room->de->de_type = d_type; + room->de->de_str.len = nlen; + memcpy(room->de->de_str.name, name, nlen); + + err = 0; + room->deblk += sz; + if (unlikely(set_deblk_end(room, &deblk_end))) + err = append_deblk(vdir); + /* smp_mb(); */ + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +void au_vdir_free(struct au_vdir *vdir) +{ + unsigned char **deblk; + + deblk = vdir->vd_deblk; + while (vdir->vd_nblk--) + au_kfree_try_rcu(*deblk++); + au_kfree_try_rcu(vdir->vd_deblk); + au_cache_free_vdir(vdir); +} + +static struct au_vdir *alloc_vdir(struct file *file) +{ + struct au_vdir *vdir; + struct super_block *sb; + int err; + + sb = file->f_path.dentry->d_sb; + SiMustAnyLock(sb); + + err = -ENOMEM; + vdir = au_cache_alloc_vdir(); + if (unlikely(!vdir)) + goto out; + + vdir->vd_deblk = kzalloc(sizeof(*vdir->vd_deblk), GFP_NOFS); + if (unlikely(!vdir->vd_deblk)) + goto out_free; + + vdir->vd_deblk_sz = au_sbi(sb)->si_rdblk; + if (!vdir->vd_deblk_sz) { + /* estimate the appropriate size for deblk */ + vdir->vd_deblk_sz = au_dir_size(file, /*dentry*/NULL); + /* pr_info("vd_deblk_sz %u\n", vdir->vd_deblk_sz); */ + } + vdir->vd_nblk = 0; + vdir->vd_version = 0; + vdir->vd_jiffy = 0; + err = append_deblk(vdir); + if (!err) + return vdir; /* success */ + + au_kfree_try_rcu(vdir->vd_deblk); + +out_free: + au_cache_free_vdir(vdir); +out: + vdir = ERR_PTR(err); + return vdir; +} + +static int reinit_vdir(struct au_vdir *vdir) +{ + int err; + union au_vdir_deblk_p p, deblk_end; + + while (vdir->vd_nblk > 1) { + au_kfree_try_rcu(vdir->vd_deblk[vdir->vd_nblk - 1]); + /* vdir->vd_deblk[vdir->vd_nblk - 1] = NULL; */ + vdir->vd_nblk--; + } + p.deblk = vdir->vd_deblk[0]; + deblk_end.deblk = p.deblk + vdir->vd_deblk_sz; + err = set_deblk_end(&p, &deblk_end); + /* keep vd_dblk_sz */ + vdir->vd_last.ul = 0; + vdir->vd_last.p.deblk = vdir->vd_deblk[0]; + vdir->vd_version = 0; + vdir->vd_jiffy = 0; + /* smp_mb(); */ + return err; +} + +/* ---------------------------------------------------------------------- */ + +#define AuFillVdir_CALLED 1 +#define AuFillVdir_WHABLE (1 << 1) +#define AuFillVdir_SHWH (1 << 2) +#define au_ftest_fillvdir(flags, name) ((flags) & AuFillVdir_##name) +#define au_fset_fillvdir(flags, name) \ + do { (flags) |= AuFillVdir_##name; } while (0) +#define au_fclr_fillvdir(flags, name) \ + do { (flags) &= ~AuFillVdir_##name; } while (0) + +#ifndef CONFIG_AUFS_SHWH +#undef AuFillVdir_SHWH +#define AuFillVdir_SHWH 0 +#endif + +struct fillvdir_arg { + struct dir_context ctx; + struct file *file; + struct au_vdir *vdir; + struct au_nhash delist; + struct au_nhash whlist; + aufs_bindex_t bindex; + unsigned int flags; + int err; +}; + +static int fillvdir(struct dir_context *ctx, const char *__name, int nlen, + loff_t offset __maybe_unused, u64 h_ino, + unsigned int d_type) +{ + struct fillvdir_arg *arg = container_of(ctx, struct fillvdir_arg, ctx); + char *name = (void *)__name; + struct super_block *sb; + ino_t ino; + const unsigned char shwh = !!au_ftest_fillvdir(arg->flags, SHWH); + + arg->err = 0; + sb = arg->file->f_path.dentry->d_sb; + au_fset_fillvdir(arg->flags, CALLED); + /* smp_mb(); */ + if (nlen <= AUFS_WH_PFX_LEN + || memcmp(name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)) { + if (test_known(&arg->delist, name, nlen) + || au_nhash_test_known_wh(&arg->whlist, name, nlen)) + goto out; /* already exists or whiteouted */ + + arg->err = au_ino(sb, arg->bindex, h_ino, d_type, &ino); + if (!arg->err) { + if (unlikely(nlen > AUFS_MAX_NAMELEN)) + d_type = DT_UNKNOWN; + arg->err = append_de(arg->vdir, name, nlen, ino, + d_type, &arg->delist); + } + } else if (au_ftest_fillvdir(arg->flags, WHABLE)) { + name += AUFS_WH_PFX_LEN; + nlen -= AUFS_WH_PFX_LEN; + if (au_nhash_test_known_wh(&arg->whlist, name, nlen)) + goto out; /* already whiteouted */ + + ino = 0; /* just to suppress a warning */ + if (shwh) + arg->err = au_wh_ino(sb, arg->bindex, h_ino, d_type, + &ino); + if (!arg->err) { + if (nlen <= AUFS_MAX_NAMELEN + AUFS_WH_PFX_LEN) + d_type = DT_UNKNOWN; + arg->err = au_nhash_append_wh + (&arg->whlist, name, nlen, ino, d_type, + arg->bindex, shwh); + } + } + +out: + if (!arg->err) + arg->vdir->vd_jiffy = jiffies; + /* smp_mb(); */ + AuTraceErr(arg->err); + return arg->err; +} + +static int au_handle_shwh(struct super_block *sb, struct au_vdir *vdir, + struct au_nhash *whlist, struct au_nhash *delist) +{ +#ifdef CONFIG_AUFS_SHWH + int err; + unsigned int nh, u; + struct hlist_head *head; + struct au_vdir_wh *pos; + struct hlist_node *n; + char *p, *o; + struct au_vdir_destr *destr; + + AuDebugOn(!au_opt_test(au_mntflags(sb), SHWH)); + + err = -ENOMEM; + o = p = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!p)) + goto out; + + err = 0; + nh = whlist->nh_num; + memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); + p += AUFS_WH_PFX_LEN; + for (u = 0; u < nh; u++) { + head = whlist->nh_head + u; + hlist_for_each_entry_safe(pos, n, head, wh_hash) { + destr = &pos->wh_str; + memcpy(p, destr->name, destr->len); + err = append_de(vdir, o, destr->len + AUFS_WH_PFX_LEN, + pos->wh_ino, pos->wh_type, delist); + if (unlikely(err)) + break; + } + } + + free_page((unsigned long)o); + +out: + AuTraceErr(err); + return err; +#else + return 0; +#endif +} + +static int au_do_read_vdir(struct fillvdir_arg *arg) +{ + int err; + unsigned int rdhash; + loff_t offset; + aufs_bindex_t bbot, bindex, btop; + unsigned char shwh; + struct file *hf, *file; + struct super_block *sb; + + file = arg->file; + sb = file->f_path.dentry->d_sb; + SiMustAnyLock(sb); + + rdhash = au_sbi(sb)->si_rdhash; + if (!rdhash) + rdhash = au_rdhash_est(au_dir_size(file, /*dentry*/NULL)); + err = au_nhash_alloc(&arg->delist, rdhash, GFP_NOFS); + if (unlikely(err)) + goto out; + err = au_nhash_alloc(&arg->whlist, rdhash, GFP_NOFS); + if (unlikely(err)) + goto out_delist; + + err = 0; + arg->flags = 0; + shwh = 0; + if (au_opt_test(au_mntflags(sb), SHWH)) { + shwh = 1; + au_fset_fillvdir(arg->flags, SHWH); + } + btop = au_fbtop(file); + bbot = au_fbbot_dir(file); + for (bindex = btop; !err && bindex <= bbot; bindex++) { + hf = au_hf_dir(file, bindex); + if (!hf) + continue; + + offset = vfsub_llseek(hf, 0, SEEK_SET); + err = offset; + if (unlikely(offset)) + break; + + arg->bindex = bindex; + au_fclr_fillvdir(arg->flags, WHABLE); + if (shwh + || (bindex != bbot + && au_br_whable(au_sbr_perm(sb, bindex)))) + au_fset_fillvdir(arg->flags, WHABLE); + do { + arg->err = 0; + au_fclr_fillvdir(arg->flags, CALLED); + /* smp_mb(); */ + err = vfsub_iterate_dir(hf, &arg->ctx); + if (err >= 0) + err = arg->err; + } while (!err && au_ftest_fillvdir(arg->flags, CALLED)); + + /* + * dir_relax() may be good for concurrency, but aufs should not + * use it since it will cause a lockdep problem. + */ + } + + if (!err && shwh) + err = au_handle_shwh(sb, arg->vdir, &arg->whlist, &arg->delist); + + au_nhash_wh_free(&arg->whlist); + +out_delist: + au_nhash_de_free(&arg->delist); +out: + return err; +} + +static int read_vdir(struct file *file, int may_read) +{ + int err; + unsigned long expire; + unsigned char do_read; + struct fillvdir_arg arg = { + .ctx = { + .actor = fillvdir + } + }; + struct inode *inode; + struct au_vdir *vdir, *allocated; + + err = 0; + inode = file_inode(file); + IMustLock(inode); + IiMustWriteLock(inode); + SiMustAnyLock(inode->i_sb); + + allocated = NULL; + do_read = 0; + expire = au_sbi(inode->i_sb)->si_rdcache; + vdir = au_ivdir(inode); + if (!vdir) { + do_read = 1; + vdir = alloc_vdir(file); + err = PTR_ERR(vdir); + if (IS_ERR(vdir)) + goto out; + err = 0; + allocated = vdir; + } else if (may_read + && (!inode_eq_iversion(inode, vdir->vd_version) + || time_after(jiffies, vdir->vd_jiffy + expire))) { + do_read = 1; + err = reinit_vdir(vdir); + if (unlikely(err)) + goto out; + } + + if (!do_read) + return 0; /* success */ + + arg.file = file; + arg.vdir = vdir; + err = au_do_read_vdir(&arg); + if (!err) { + /* file->f_pos = 0; */ /* todo: ctx->pos? */ + vdir->vd_version = inode_query_iversion(inode); + vdir->vd_last.ul = 0; + vdir->vd_last.p.deblk = vdir->vd_deblk[0]; + if (allocated) + au_set_ivdir(inode, allocated); + } else if (allocated) + au_vdir_free(allocated); + +out: + return err; +} + +static int copy_vdir(struct au_vdir *tgt, struct au_vdir *src) +{ + int err, rerr; + unsigned long ul, n; + const unsigned int deblk_sz = src->vd_deblk_sz; + + AuDebugOn(tgt->vd_nblk != 1); + + err = -ENOMEM; + if (tgt->vd_nblk < src->vd_nblk) { + unsigned char **p; + + p = au_krealloc(tgt->vd_deblk, sizeof(*p) * src->vd_nblk, + GFP_NOFS, /*may_shrink*/0); + if (unlikely(!p)) + goto out; + tgt->vd_deblk = p; + } + + if (tgt->vd_deblk_sz != deblk_sz) { + unsigned char *p; + + tgt->vd_deblk_sz = deblk_sz; + p = au_krealloc(tgt->vd_deblk[0], deblk_sz, GFP_NOFS, + /*may_shrink*/1); + if (unlikely(!p)) + goto out; + tgt->vd_deblk[0] = p; + } + memcpy(tgt->vd_deblk[0], src->vd_deblk[0], deblk_sz); + tgt->vd_version = src->vd_version; + tgt->vd_jiffy = src->vd_jiffy; + + n = src->vd_nblk; + for (ul = 1; ul < n; ul++) { + tgt->vd_deblk[ul] = kmemdup(src->vd_deblk[ul], deblk_sz, + GFP_NOFS); + if (unlikely(!tgt->vd_deblk[ul])) + goto out; + tgt->vd_nblk++; + } + tgt->vd_nblk = n; + tgt->vd_last.ul = tgt->vd_last.ul; + tgt->vd_last.p.deblk = tgt->vd_deblk[tgt->vd_last.ul]; + tgt->vd_last.p.deblk += src->vd_last.p.deblk + - src->vd_deblk[src->vd_last.ul]; + /* smp_mb(); */ + return 0; /* success */ + +out: + rerr = reinit_vdir(tgt); + BUG_ON(rerr); + return err; +} + +int au_vdir_init(struct file *file) +{ + int err; + struct inode *inode; + struct au_vdir *vdir_cache, *allocated; + + /* test file->f_pos here instead of ctx->pos */ + err = read_vdir(file, !file->f_pos); + if (unlikely(err)) + goto out; + + allocated = NULL; + vdir_cache = au_fvdir_cache(file); + if (!vdir_cache) { + vdir_cache = alloc_vdir(file); + err = PTR_ERR(vdir_cache); + if (IS_ERR(vdir_cache)) + goto out; + allocated = vdir_cache; + } else if (!file->f_pos && vdir_cache->vd_version != file->f_version) { + /* test file->f_pos here instead of ctx->pos */ + err = reinit_vdir(vdir_cache); + if (unlikely(err)) + goto out; + } else + return 0; /* success */ + + inode = file_inode(file); + err = copy_vdir(vdir_cache, au_ivdir(inode)); + if (!err) { + file->f_version = inode_query_iversion(inode); + if (allocated) + au_set_fvdir_cache(file, allocated); + } else if (allocated) + au_vdir_free(allocated); + +out: + return err; +} + +static loff_t calc_offset(struct au_vdir *vdir) +{ + loff_t offset; + union au_vdir_deblk_p p; + + p.deblk = vdir->vd_deblk[vdir->vd_last.ul]; + offset = vdir->vd_last.p.deblk - p.deblk; + offset += vdir->vd_deblk_sz * vdir->vd_last.ul; + return offset; +} + +/* returns true or false */ +static int seek_vdir(struct file *file, struct dir_context *ctx) +{ + int valid; + unsigned int deblk_sz; + unsigned long ul, n; + loff_t offset; + union au_vdir_deblk_p p, deblk_end; + struct au_vdir *vdir_cache; + + valid = 1; + vdir_cache = au_fvdir_cache(file); + offset = calc_offset(vdir_cache); + AuDbg("offset %lld\n", offset); + if (ctx->pos == offset) + goto out; + + vdir_cache->vd_last.ul = 0; + vdir_cache->vd_last.p.deblk = vdir_cache->vd_deblk[0]; + if (!ctx->pos) + goto out; + + valid = 0; + deblk_sz = vdir_cache->vd_deblk_sz; + ul = div64_u64(ctx->pos, deblk_sz); + AuDbg("ul %lu\n", ul); + if (ul >= vdir_cache->vd_nblk) + goto out; + + n = vdir_cache->vd_nblk; + for (; ul < n; ul++) { + p.deblk = vdir_cache->vd_deblk[ul]; + deblk_end.deblk = p.deblk + deblk_sz; + offset = ul; + offset *= deblk_sz; + while (!is_deblk_end(&p, &deblk_end) && offset < ctx->pos) { + unsigned int l; + + l = calc_size(p.de->de_str.len); + offset += l; + p.deblk += l; + } + if (!is_deblk_end(&p, &deblk_end)) { + valid = 1; + vdir_cache->vd_last.ul = ul; + vdir_cache->vd_last.p = p; + break; + } + } + +out: + /* smp_mb(); */ + if (!valid) + AuDbg("valid %d\n", !valid); + return valid; +} + +int au_vdir_fill_de(struct file *file, struct dir_context *ctx) +{ + unsigned int l, deblk_sz; + union au_vdir_deblk_p deblk_end; + struct au_vdir *vdir_cache; + struct au_vdir_de *de; + + if (!seek_vdir(file, ctx)) + return 0; + + vdir_cache = au_fvdir_cache(file); + deblk_sz = vdir_cache->vd_deblk_sz; + while (1) { + deblk_end.deblk = vdir_cache->vd_deblk[vdir_cache->vd_last.ul]; + deblk_end.deblk += deblk_sz; + while (!is_deblk_end(&vdir_cache->vd_last.p, &deblk_end)) { + de = vdir_cache->vd_last.p.de; + AuDbg("%.*s, off%lld, i%lu, dt%d\n", + de->de_str.len, de->de_str.name, ctx->pos, + (unsigned long)de->de_ino, de->de_type); + if (unlikely(!dir_emit(ctx, de->de_str.name, + de->de_str.len, de->de_ino, + de->de_type))) { + /* todo: ignore the error caused by udba? */ + /* return err; */ + return 0; + } + + l = calc_size(de->de_str.len); + vdir_cache->vd_last.p.deblk += l; + ctx->pos += l; + } + if (vdir_cache->vd_last.ul < vdir_cache->vd_nblk - 1) { + vdir_cache->vd_last.ul++; + vdir_cache->vd_last.p.deblk + = vdir_cache->vd_deblk[vdir_cache->vd_last.ul]; + ctx->pos = deblk_sz * vdir_cache->vd_last.ul; + continue; + } + break; + } + + /* smp_mb(); */ + return 0; +} diff --git a/fs/aufs/vfsub.c b/fs/aufs/vfsub.c new file mode 100644 index 00000000000..7b2675ea043 --- /dev/null +++ b/fs/aufs/vfsub.c @@ -0,0 +1,916 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sub-routines for VFS + */ + +#include +#include +#include +#include +#include +#include "aufs.h" + +#ifdef CONFIG_AUFS_BR_FUSE +int vfsub_test_mntns(struct vfsmount *mnt, struct super_block *h_sb) +{ + if (!au_test_fuse(h_sb) || !au_userns) + return 0; + + return is_current_mnt_ns(mnt) ? 0 : -EACCES; +} +#endif + +int vfsub_sync_filesystem(struct super_block *h_sb, int wait) +{ + int err; + + lockdep_off(); + down_read(&h_sb->s_umount); + err = __sync_filesystem(h_sb, wait); + up_read(&h_sb->s_umount); + lockdep_on(); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +int vfsub_update_h_iattr(struct path *h_path, int *did) +{ + int err; + struct kstat st; + struct super_block *h_sb; + + /* for remote fs, leave work for its getattr or d_revalidate */ + /* for bad i_attr fs, handle them in aufs_getattr() */ + /* still some fs may acquire i_mutex. we need to skip them */ + err = 0; + if (!did) + did = &err; + h_sb = h_path->dentry->d_sb; + *did = (!au_test_fs_remote(h_sb) && au_test_fs_refresh_iattr(h_sb)); + if (*did) + err = vfsub_getattr(h_path, &st); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct file *vfsub_dentry_open(struct path *path, int flags) +{ + return dentry_open(path, flags /* | __FMODE_NONOTIFY */, + current_cred()); +} + +struct file *vfsub_filp_open(const char *path, int oflags, int mode) +{ + struct file *file; + + lockdep_off(); + file = filp_open(path, + oflags /* | __FMODE_NONOTIFY */, + mode); + lockdep_on(); + if (IS_ERR(file)) + goto out; + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + +out: + return file; +} + +/* + * Ideally this function should call VFS:do_last() in order to keep all its + * checkings. But it is very hard for aufs to regenerate several VFS internal + * structure such as nameidata. This is a second (or third) best approach. + * cf. linux/fs/namei.c:do_last(), lookup_open() and atomic_open(). + */ +int vfsub_atomic_open(struct inode *dir, struct dentry *dentry, + struct vfsub_aopen_args *args) +{ + int err; + struct au_branch *br = args->br; + struct file *file = args->file; + /* copied from linux/fs/namei.c:atomic_open() */ + struct dentry *const DENTRY_NOT_SET = (void *)-1UL; + + IMustLock(dir); + AuDebugOn(!dir->i_op->atomic_open); + + err = au_br_test_oflag(args->open_flag, br); + if (unlikely(err)) + goto out; + + au_lcnt_inc(&br->br_nfiles); + file->f_path.dentry = DENTRY_NOT_SET; + file->f_path.mnt = au_br_mnt(br); + AuDbg("%ps\n", dir->i_op->atomic_open); + err = dir->i_op->atomic_open(dir, dentry, file, args->open_flag, + args->create_mode); + if (unlikely(err < 0)) { + au_lcnt_dec(&br->br_nfiles); + goto out; + } + + /* temporary workaround for nfsv4 branch */ + if (au_test_nfs(dir->i_sb)) + nfs_mark_for_revalidate(dir); + + if (file->f_mode & FMODE_CREATED) + fsnotify_create(dir, dentry); + if (!(file->f_mode & FMODE_OPENED)) { + au_lcnt_dec(&br->br_nfiles); + goto out; + } + + /* todo: call VFS:may_open() here */ + /* todo: ima_file_check() too? */ + if (!err && (args->open_flag & __FMODE_EXEC)) + err = deny_write_access(file); + if (!err) + fsnotify_open(file); + else + au_lcnt_dec(&br->br_nfiles); + /* note that the file is created and still opened */ + +out: + return err; +} + +int vfsub_kern_path(const char *name, unsigned int flags, struct path *path) +{ + int err; + + err = kern_path(name, flags, path); + if (!err && d_is_positive(path->dentry)) + vfsub_update_h_iattr(path, /*did*/NULL); /*ignore*/ + return err; +} + +struct dentry *vfsub_lookup_one_len_unlocked(const char *name, + struct dentry *parent, int len) +{ + struct path path = { + .mnt = NULL + }; + + path.dentry = lookup_one_len_unlocked(name, parent, len); + if (IS_ERR(path.dentry)) + goto out; + if (d_is_positive(path.dentry)) + vfsub_update_h_iattr(&path, /*did*/NULL); /*ignore*/ + +out: + AuTraceErrPtr(path.dentry); + return path.dentry; +} + +struct dentry *vfsub_lookup_one_len(const char *name, struct dentry *parent, + int len) +{ + struct path path = { + .mnt = NULL + }; + + /* VFS checks it too, but by WARN_ON_ONCE() */ + IMustLock(d_inode(parent)); + + path.dentry = lookup_one_len(name, parent, len); + if (IS_ERR(path.dentry)) + goto out; + if (d_is_positive(path.dentry)) + vfsub_update_h_iattr(&path, /*did*/NULL); /*ignore*/ + +out: + AuTraceErrPtr(path.dentry); + return path.dentry; +} + +void vfsub_call_lkup_one(void *args) +{ + struct vfsub_lkup_one_args *a = args; + *a->errp = vfsub_lkup_one(a->name, a->parent); +} + +/* ---------------------------------------------------------------------- */ + +struct dentry *vfsub_lock_rename(struct dentry *d1, struct au_hinode *hdir1, + struct dentry *d2, struct au_hinode *hdir2) +{ + struct dentry *d; + + lockdep_off(); + d = lock_rename(d1, d2); + lockdep_on(); + au_hn_suspend(hdir1); + if (hdir1 != hdir2) + au_hn_suspend(hdir2); + + return d; +} + +void vfsub_unlock_rename(struct dentry *d1, struct au_hinode *hdir1, + struct dentry *d2, struct au_hinode *hdir2) +{ + au_hn_resume(hdir1); + if (hdir1 != hdir2) + au_hn_resume(hdir2); + lockdep_off(); + unlock_rename(d1, d2); + lockdep_on(); +} + +/* ---------------------------------------------------------------------- */ + +int vfsub_create(struct inode *dir, struct path *path, int mode, bool want_excl) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_mknod(path, d, mode, 0); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_create(userns, dir, path->dentry, mode, want_excl); + lockdep_on(); + if (!err) { + struct path tmp = *path; + int did; + + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = path->dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +int vfsub_symlink(struct inode *dir, struct path *path, const char *symname) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_symlink(path, d, symname); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_symlink(userns, dir, path->dentry, symname); + lockdep_on(); + if (!err) { + struct path tmp = *path; + int did; + + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = path->dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +int vfsub_mknod(struct inode *dir, struct path *path, int mode, dev_t dev) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_mknod(path, d, mode, new_encode_dev(dev)); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_mknod(userns, dir, path->dentry, mode, dev); + lockdep_on(); + if (!err) { + struct path tmp = *path; + int did; + + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = path->dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +static int au_test_nlink(struct inode *inode) +{ + const unsigned int link_max = UINT_MAX >> 1; /* rough margin */ + + if (!au_test_fs_no_limit_nlink(inode->i_sb) + || inode->i_nlink < link_max) + return 0; + return -EMLINK; +} + +int vfsub_link(struct dentry *src_dentry, struct inode *dir, struct path *path, + struct inode **delegated_inode) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + err = au_test_nlink(d_inode(src_dentry)); + if (unlikely(err)) + return err; + + /* we don't call may_linkat() */ + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_link(src_dentry, path, d); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_link(src_dentry, userns, dir, path->dentry, delegated_inode); + lockdep_on(); + if (!err) { + struct path tmp = *path; + int did; + + /* fuse has different memory inode for the same inumber */ + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = path->dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + tmp.dentry = src_dentry; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +int vfsub_rename(struct inode *src_dir, struct dentry *src_dentry, + struct inode *dir, struct path *path, + struct inode **delegated_inode, unsigned int flags) +{ + int err; + struct renamedata rd; + struct path tmp = { + .mnt = path->mnt + }; + struct dentry *d; + + IMustLock(dir); + IMustLock(src_dir); + + d = path->dentry; + path->dentry = d->d_parent; + tmp.dentry = src_dentry->d_parent; + err = security_path_rename(&tmp, src_dentry, path, d, /*flags*/0); + path->dentry = d; + if (unlikely(err)) + goto out; + + rd.old_mnt_userns = mnt_user_ns(path->mnt); + rd.old_dir = src_dir; + rd.old_dentry = src_dentry; + rd.new_mnt_userns = rd.old_mnt_userns; + rd.new_dir = dir; + rd.new_dentry = path->dentry; + rd.delegated_inode = delegated_inode; + rd.flags = flags; + lockdep_off(); + err = vfs_rename(&rd); + lockdep_on(); + if (!err) { + int did; + + tmp.dentry = d->d_parent; + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = src_dentry; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + tmp.dentry = src_dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +int vfsub_mkdir(struct inode *dir, struct path *path, int mode) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_mkdir(path, d, mode); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_mkdir(userns, dir, path->dentry, mode); + lockdep_on(); + if (!err) { + struct path tmp = *path; + int did; + + vfsub_update_h_iattr(&tmp, &did); + if (did) { + tmp.dentry = path->dentry->d_parent; + vfsub_update_h_iattr(&tmp, /*did*/NULL); + } + /*ignore*/ + } + +out: + return err; +} + +int vfsub_rmdir(struct inode *dir, struct path *path) +{ + int err; + struct dentry *d; + struct user_namespace *userns; + + IMustLock(dir); + + d = path->dentry; + path->dentry = d->d_parent; + err = security_path_rmdir(path, d); + path->dentry = d; + if (unlikely(err)) + goto out; + userns = mnt_user_ns(path->mnt); + + lockdep_off(); + err = vfs_rmdir(userns, dir, path->dentry); + lockdep_on(); + if (!err) { + struct path tmp = { + .dentry = path->dentry->d_parent, + .mnt = path->mnt + }; + + vfsub_update_h_iattr(&tmp, /*did*/NULL); /*ignore*/ + } + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* todo: support mmap_sem? */ +ssize_t vfsub_read_u(struct file *file, char __user *ubuf, size_t count, + loff_t *ppos) +{ + ssize_t err; + + lockdep_off(); + err = vfs_read(file, ubuf, count, ppos); + lockdep_on(); + if (err >= 0) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +ssize_t vfsub_read_k(struct file *file, void *kbuf, size_t count, + loff_t *ppos) +{ + ssize_t err; + + lockdep_off(); + err = kernel_read(file, kbuf, count, ppos); + lockdep_on(); + AuTraceErr(err); + if (err >= 0) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +ssize_t vfsub_write_u(struct file *file, const char __user *ubuf, size_t count, + loff_t *ppos) +{ + ssize_t err; + + lockdep_off(); + err = vfs_write(file, ubuf, count, ppos); + lockdep_on(); + if (err >= 0) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +ssize_t vfsub_write_k(struct file *file, void *kbuf, size_t count, loff_t *ppos) +{ + ssize_t err; + + lockdep_off(); + err = kernel_write(file, kbuf, count, ppos); + lockdep_on(); + if (err >= 0) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +int vfsub_flush(struct file *file, fl_owner_t id) +{ + int err; + + err = 0; + if (file->f_op->flush) { + if (!au_test_nfs(file->f_path.dentry->d_sb)) + err = file->f_op->flush(file, id); + else { + lockdep_off(); + err = file->f_op->flush(file, id); + lockdep_on(); + } + if (!err) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); + /*ignore*/ + } + return err; +} + +int vfsub_iterate_dir(struct file *file, struct dir_context *ctx) +{ + int err; + + AuDbg("%pD, ctx{%ps, %llu}\n", file, ctx->actor, ctx->pos); + + lockdep_off(); + err = iterate_dir(file, ctx); + lockdep_on(); + if (err >= 0) + vfsub_update_h_iattr(&file->f_path, /*did*/NULL); /*ignore*/ + + return err; +} + +long vfsub_splice_to(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) +{ + long err; + + lockdep_off(); + err = do_splice_to(in, ppos, pipe, len, flags); + lockdep_on(); + file_accessed(in); + if (err >= 0) + vfsub_update_h_iattr(&in->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +long vfsub_splice_from(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags) +{ + long err; + + lockdep_off(); + err = do_splice_from(pipe, out, ppos, len, flags); + lockdep_on(); + if (err >= 0) + vfsub_update_h_iattr(&out->f_path, /*did*/NULL); /*ignore*/ + return err; +} + +int vfsub_fsync(struct file *file, struct path *path, int datasync) +{ + int err; + + /* file can be NULL */ + lockdep_off(); + err = vfs_fsync(file, datasync); + lockdep_on(); + if (!err) { + if (!path) { + AuDebugOn(!file); + path = &file->f_path; + } + vfsub_update_h_iattr(path, /*did*/NULL); /*ignore*/ + } + return err; +} + +/* cf. open.c:do_sys_truncate() and do_sys_ftruncate() */ +int vfsub_trunc(struct path *h_path, loff_t length, unsigned int attr, + struct file *h_file) +{ + int err; + struct inode *h_inode; + struct super_block *h_sb; + struct user_namespace *h_userns; + + if (!h_file) { + err = vfsub_truncate(h_path, length); + goto out; + } + + h_inode = d_inode(h_path->dentry); + h_sb = h_inode->i_sb; + lockdep_off(); + sb_start_write(h_sb); + lockdep_on(); + err = locks_verify_truncate(h_inode, h_file, length); + if (!err) + err = security_path_truncate(h_path); + if (!err) { + h_userns = mnt_user_ns(h_path->mnt); + lockdep_off(); + err = do_truncate(h_userns, h_path->dentry, length, attr, + h_file); + lockdep_on(); + } + lockdep_off(); + sb_end_write(h_sb); + lockdep_on(); + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct au_vfsub_mkdir_args { + int *errp; + struct inode *dir; + struct path *path; + int mode; +}; + +static void au_call_vfsub_mkdir(void *args) +{ + struct au_vfsub_mkdir_args *a = args; + *a->errp = vfsub_mkdir(a->dir, a->path, a->mode); +} + +int vfsub_sio_mkdir(struct inode *dir, struct path *path, int mode) +{ + int err, do_sio, wkq_err; + struct user_namespace *userns; + + userns = mnt_user_ns(path->mnt); + do_sio = au_test_h_perm_sio(userns, dir, MAY_EXEC | MAY_WRITE); + if (!do_sio) { + lockdep_off(); + err = vfsub_mkdir(dir, path, mode); + lockdep_on(); + } else { + struct au_vfsub_mkdir_args args = { + .errp = &err, + .dir = dir, + .path = path, + .mode = mode + }; + wkq_err = au_wkq_wait(au_call_vfsub_mkdir, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + return err; +} + +struct au_vfsub_rmdir_args { + int *errp; + struct inode *dir; + struct path *path; +}; + +static void au_call_vfsub_rmdir(void *args) +{ + struct au_vfsub_rmdir_args *a = args; + *a->errp = vfsub_rmdir(a->dir, a->path); +} + +int vfsub_sio_rmdir(struct inode *dir, struct path *path) +{ + int err, do_sio, wkq_err; + struct user_namespace *userns; + + userns = mnt_user_ns(path->mnt); + do_sio = au_test_h_perm_sio(userns, dir, MAY_EXEC | MAY_WRITE); + if (!do_sio) { + lockdep_off(); + err = vfsub_rmdir(dir, path); + lockdep_on(); + } else { + struct au_vfsub_rmdir_args args = { + .errp = &err, + .dir = dir, + .path = path + }; + wkq_err = au_wkq_wait(au_call_vfsub_rmdir, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct notify_change_args { + int *errp; + struct path *path; + struct iattr *ia; + struct inode **delegated_inode; +}; + +static void call_notify_change(void *args) +{ + struct notify_change_args *a = args; + struct inode *h_inode; + struct user_namespace *userns; + + h_inode = d_inode(a->path->dentry); + IMustLock(h_inode); + + *a->errp = -EPERM; + if (!IS_IMMUTABLE(h_inode) && !IS_APPEND(h_inode)) { + userns = mnt_user_ns(a->path->mnt); + lockdep_off(); + *a->errp = notify_change(userns, a->path->dentry, a->ia, + a->delegated_inode); + lockdep_on(); + if (!*a->errp) + vfsub_update_h_iattr(a->path, /*did*/NULL); /*ignore*/ + } + AuTraceErr(*a->errp); +} + +int vfsub_notify_change(struct path *path, struct iattr *ia, + struct inode **delegated_inode) +{ + int err; + struct notify_change_args args = { + .errp = &err, + .path = path, + .ia = ia, + .delegated_inode = delegated_inode + }; + + call_notify_change(&args); + + return err; +} + +int vfsub_sio_notify_change(struct path *path, struct iattr *ia, + struct inode **delegated_inode) +{ + int err, wkq_err; + struct notify_change_args args = { + .errp = &err, + .path = path, + .ia = ia, + .delegated_inode = delegated_inode + }; + + wkq_err = au_wkq_wait(call_notify_change, &args); + if (unlikely(wkq_err)) + err = wkq_err; + + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct unlink_args { + int *errp; + struct inode *dir; + struct path *path; + struct inode **delegated_inode; +}; + +static void call_unlink(void *args) +{ + struct unlink_args *a = args; + struct dentry *d = a->path->dentry; + struct inode *h_inode; + struct user_namespace *userns; + const int stop_sillyrename = (au_test_nfs(d->d_sb) + && au_dcount(d) == 1); + + IMustLock(a->dir); + + a->path->dentry = d->d_parent; + *a->errp = security_path_unlink(a->path, d); + a->path->dentry = d; + if (unlikely(*a->errp)) + return; + + if (!stop_sillyrename) + dget(d); + h_inode = NULL; + if (d_is_positive(d)) { + h_inode = d_inode(d); + ihold(h_inode); + } + + userns = mnt_user_ns(a->path->mnt); + lockdep_off(); + *a->errp = vfs_unlink(userns, a->dir, d, a->delegated_inode); + lockdep_on(); + if (!*a->errp) { + struct path tmp = { + .dentry = d->d_parent, + .mnt = a->path->mnt + }; + vfsub_update_h_iattr(&tmp, /*did*/NULL); /*ignore*/ + } + + if (!stop_sillyrename) + dput(d); + if (h_inode) + iput(h_inode); + + AuTraceErr(*a->errp); +} + +/* + * @dir: must be locked. + * @dentry: target dentry. + */ +int vfsub_unlink(struct inode *dir, struct path *path, + struct inode **delegated_inode, int force) +{ + int err; + struct unlink_args args = { + .errp = &err, + .dir = dir, + .path = path, + .delegated_inode = delegated_inode + }; + + if (!force) + call_unlink(&args); + else { + int wkq_err; + + wkq_err = au_wkq_wait(call_unlink, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + + return err; +} diff --git a/fs/aufs/vfsub.h b/fs/aufs/vfsub.h new file mode 100644 index 00000000000..62becb91cc0 --- /dev/null +++ b/fs/aufs/vfsub.h @@ -0,0 +1,358 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * sub-routines for VFS + */ + +#ifndef __AUFS_VFSUB_H__ +#define __AUFS_VFSUB_H__ + +#ifdef __KERNEL__ + +#include +#include +#include +#include +#include "debug.h" + +/* copied from linux/fs/internal.h */ +/* todo: BAD approach!! */ +extern void __mnt_drop_write(struct vfsmount *); +extern struct file *alloc_empty_file(int, const struct cred *); + +/* ---------------------------------------------------------------------- */ + +/* lock subclass for lower inode */ +/* default MAX_LOCKDEP_SUBCLASSES(8) is not enough */ +/* reduce? gave up. */ +enum { + AuLsc_I_Begin = I_MUTEX_PARENT2, /* 5 */ + AuLsc_I_PARENT, /* lower inode, parent first */ + AuLsc_I_PARENT2, /* copyup dirs */ + AuLsc_I_PARENT3, /* copyup wh */ + AuLsc_I_CHILD, + AuLsc_I_CHILD2, + AuLsc_I_End +}; + +/* to debug easier, do not make them inlined functions */ +#define MtxMustLock(mtx) AuDebugOn(!mutex_is_locked(mtx)) +#define IMustLock(i) AuDebugOn(!inode_is_locked(i)) + +/* ---------------------------------------------------------------------- */ + +static inline void vfsub_drop_nlink(struct inode *inode) +{ + AuDebugOn(!inode->i_nlink); + drop_nlink(inode); +} + +static inline void vfsub_dead_dir(struct inode *inode) +{ + AuDebugOn(!S_ISDIR(inode->i_mode)); + inode->i_flags |= S_DEAD; + clear_nlink(inode); +} + +static inline int vfsub_native_ro(struct inode *inode) +{ + return sb_rdonly(inode->i_sb) + || IS_RDONLY(inode) + /* || IS_APPEND(inode) */ + || IS_IMMUTABLE(inode); +} + +#ifdef CONFIG_AUFS_BR_FUSE +int vfsub_test_mntns(struct vfsmount *mnt, struct super_block *h_sb); +#else +AuStubInt0(vfsub_test_mntns, struct vfsmount *mnt, struct super_block *h_sb); +#endif + +int vfsub_sync_filesystem(struct super_block *h_sb, int wait); + +/* ---------------------------------------------------------------------- */ + +int vfsub_update_h_iattr(struct path *h_path, int *did); +struct file *vfsub_dentry_open(struct path *path, int flags); +struct file *vfsub_filp_open(const char *path, int oflags, int mode); +struct au_branch; +struct vfsub_aopen_args { + struct file *file; + unsigned int open_flag; + umode_t create_mode; + struct au_branch *br; +}; +int vfsub_atomic_open(struct inode *dir, struct dentry *dentry, + struct vfsub_aopen_args *args); +int vfsub_kern_path(const char *name, unsigned int flags, struct path *path); + +struct dentry *vfsub_lookup_one_len_unlocked(const char *name, + struct dentry *parent, int len); +struct dentry *vfsub_lookup_one_len(const char *name, struct dentry *parent, + int len); + +struct vfsub_lkup_one_args { + struct dentry **errp; + struct qstr *name; + struct dentry *parent; +}; + +static inline struct dentry *vfsub_lkup_one(struct qstr *name, + struct dentry *parent) +{ + return vfsub_lookup_one_len(name->name, parent, name->len); +} + +void vfsub_call_lkup_one(void *args); + +/* ---------------------------------------------------------------------- */ + +static inline int vfsub_mnt_want_write(struct vfsmount *mnt) +{ + int err; + + lockdep_off(); + err = mnt_want_write(mnt); + lockdep_on(); + return err; +} + +static inline void vfsub_mnt_drop_write(struct vfsmount *mnt) +{ + lockdep_off(); + mnt_drop_write(mnt); + lockdep_on(); +} + +#if 0 /* reserved */ +static inline void vfsub_mnt_drop_write_file(struct file *file) +{ + lockdep_off(); + mnt_drop_write_file(file); + lockdep_on(); +} +#endif + +/* ---------------------------------------------------------------------- */ + +struct au_hinode; +struct dentry *vfsub_lock_rename(struct dentry *d1, struct au_hinode *hdir1, + struct dentry *d2, struct au_hinode *hdir2); +void vfsub_unlock_rename(struct dentry *d1, struct au_hinode *hdir1, + struct dentry *d2, struct au_hinode *hdir2); + +int vfsub_create(struct inode *dir, struct path *path, int mode, + bool want_excl); +int vfsub_symlink(struct inode *dir, struct path *path, + const char *symname); +int vfsub_mknod(struct inode *dir, struct path *path, int mode, dev_t dev); +int vfsub_link(struct dentry *src_dentry, struct inode *dir, + struct path *path, struct inode **delegated_inode); +int vfsub_rename(struct inode *src_hdir, struct dentry *src_dentry, + struct inode *hdir, struct path *path, + struct inode **delegated_inode, unsigned int flags); +int vfsub_mkdir(struct inode *dir, struct path *path, int mode); +int vfsub_rmdir(struct inode *dir, struct path *path); + +/* ---------------------------------------------------------------------- */ + +ssize_t vfsub_read_u(struct file *file, char __user *ubuf, size_t count, + loff_t *ppos); +ssize_t vfsub_read_k(struct file *file, void *kbuf, size_t count, + loff_t *ppos); +ssize_t vfsub_write_u(struct file *file, const char __user *ubuf, size_t count, + loff_t *ppos); +ssize_t vfsub_write_k(struct file *file, void *kbuf, size_t count, + loff_t *ppos); +int vfsub_flush(struct file *file, fl_owner_t id); +int vfsub_iterate_dir(struct file *file, struct dir_context *ctx); + +static inline loff_t vfsub_f_size_read(struct file *file) +{ + return i_size_read(file_inode(file)); +} + +static inline unsigned int vfsub_file_flags(struct file *file) +{ + unsigned int flags; + + spin_lock(&file->f_lock); + flags = file->f_flags; + spin_unlock(&file->f_lock); + + return flags; +} + +static inline int vfsub_file_execed(struct file *file) +{ + /* todo: direct access f_flags */ + return !!(vfsub_file_flags(file) & __FMODE_EXEC); +} + +#if 0 /* reserved */ +static inline void vfsub_file_accessed(struct file *h_file) +{ + file_accessed(h_file); + vfsub_update_h_iattr(&h_file->f_path, /*did*/NULL); /*ignore*/ +} +#endif + +#if 0 /* reserved */ +static inline void vfsub_touch_atime(struct vfsmount *h_mnt, + struct dentry *h_dentry) +{ + struct path h_path = { + .dentry = h_dentry, + .mnt = h_mnt + }; + touch_atime(&h_path); + vfsub_update_h_iattr(&h_path, /*did*/NULL); /*ignore*/ +} +#endif + +static inline int vfsub_update_time(struct inode *h_inode, + struct timespec64 *ts, int flags) +{ + return update_time(h_inode, ts, flags); + /* no vfsub_update_h_iattr() since we don't have struct path */ +} + +#ifdef CONFIG_FS_POSIX_ACL +static inline int vfsub_acl_chmod(struct user_namespace *h_userns, + struct inode *h_inode, umode_t h_mode) +{ + int err; + + err = posix_acl_chmod(h_userns, h_inode, h_mode); + if (err == -EOPNOTSUPP) + err = 0; + return err; +} +#else +AuStubInt0(vfsub_acl_chmod, struct user_namespace *h_userns, + struct inode *h_inode, umode_t h_mode); +#endif + +long vfsub_splice_to(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags); +long vfsub_splice_from(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags); + +static inline long vfsub_truncate(struct path *path, loff_t length) +{ + long err; + + lockdep_off(); + err = vfs_truncate(path, length); + lockdep_on(); + return err; +} + +int vfsub_trunc(struct path *h_path, loff_t length, unsigned int attr, + struct file *h_file); +int vfsub_fsync(struct file *file, struct path *path, int datasync); + +/* + * re-use branch fs's ioctl(FICLONE) while aufs itself doesn't support such + * ioctl. + */ +static inline loff_t vfsub_clone_file_range(struct file *src, struct file *dst, + loff_t len) +{ + loff_t err; + + lockdep_off(); + err = vfs_clone_file_range(src, 0, dst, 0, len, /*remap_flags*/0); + lockdep_on(); + + return err; +} + +/* copy_file_range(2) is a systemcall */ +static inline ssize_t vfsub_copy_file_range(struct file *src, loff_t src_pos, + struct file *dst, loff_t dst_pos, + size_t len, unsigned int flags) +{ + ssize_t ssz; + + lockdep_off(); + ssz = vfs_copy_file_range(src, src_pos, dst, dst_pos, len, flags); + lockdep_on(); + + return ssz; +} + +/* ---------------------------------------------------------------------- */ + +static inline loff_t vfsub_llseek(struct file *file, loff_t offset, int origin) +{ + loff_t err; + + lockdep_off(); + err = vfs_llseek(file, offset, origin); + lockdep_on(); + return err; +} + +/* ---------------------------------------------------------------------- */ + +int vfsub_sio_mkdir(struct inode *dir, struct path *path, int mode); +int vfsub_sio_rmdir(struct inode *dir, struct path *path); +int vfsub_sio_notify_change(struct path *path, struct iattr *ia, + struct inode **delegated_inode); +int vfsub_notify_change(struct path *path, struct iattr *ia, + struct inode **delegated_inode); +int vfsub_unlink(struct inode *dir, struct path *path, + struct inode **delegated_inode, int force); + +static inline int vfsub_getattr(const struct path *path, struct kstat *st) +{ + return vfs_getattr(path, st, STATX_BASIC_STATS, AT_STATX_SYNC_AS_STAT); +} + +/* ---------------------------------------------------------------------- */ + +static inline int vfsub_setxattr(struct user_namespace *userns, + struct dentry *dentry, const char *name, + const void *value, size_t size, int flags) +{ + int err; + + lockdep_off(); + err = vfs_setxattr(userns, dentry, name, value, size, flags); + lockdep_on(); + + return err; +} + +static inline int vfsub_removexattr(struct user_namespace *userns, + struct dentry *dentry, const char *name) +{ + int err; + + lockdep_off(); + err = vfs_removexattr(userns, dentry, name); + lockdep_on(); + + return err; +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_VFSUB_H__ */ diff --git a/fs/aufs/wbr_policy.c b/fs/aufs/wbr_policy.c new file mode 100644 index 00000000000..a7cc1109f69 --- /dev/null +++ b/fs/aufs/wbr_policy.c @@ -0,0 +1,830 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * policies for selecting one among multiple writable branches + */ + +#include +#include "aufs.h" + +/* subset of cpup_attr() */ +static noinline_for_stack +int au_cpdown_attr(struct path *h_path, struct dentry *h_src) +{ + int err, sbits; + struct iattr ia; + struct inode *h_isrc; + + h_isrc = d_inode(h_src); + ia.ia_valid = ATTR_FORCE | ATTR_MODE | ATTR_UID | ATTR_GID; + ia.ia_mode = h_isrc->i_mode; + ia.ia_uid = h_isrc->i_uid; + ia.ia_gid = h_isrc->i_gid; + sbits = !!(ia.ia_mode & (S_ISUID | S_ISGID)); + au_cpup_attr_flags(d_inode(h_path->dentry), h_isrc->i_flags); + /* no delegation since it is just created */ + err = vfsub_sio_notify_change(h_path, &ia, /*delegated*/NULL); + + /* is this nfs only? */ + if (!err && sbits && au_test_nfs(h_path->dentry->d_sb)) { + ia.ia_valid = ATTR_FORCE | ATTR_MODE; + ia.ia_mode = h_isrc->i_mode; + err = vfsub_sio_notify_change(h_path, &ia, /*delegated*/NULL); + } + + return err; +} + +#define AuCpdown_PARENT_OPQ 1 +#define AuCpdown_WHED (1 << 1) +#define AuCpdown_MADE_DIR (1 << 2) +#define AuCpdown_DIROPQ (1 << 3) +#define au_ftest_cpdown(flags, name) ((flags) & AuCpdown_##name) +#define au_fset_cpdown(flags, name) \ + do { (flags) |= AuCpdown_##name; } while (0) +#define au_fclr_cpdown(flags, name) \ + do { (flags) &= ~AuCpdown_##name; } while (0) + +static int au_cpdown_dir_opq(struct dentry *dentry, aufs_bindex_t bdst, + unsigned int *flags) +{ + int err; + struct dentry *opq_dentry; + + opq_dentry = au_diropq_create(dentry, bdst); + err = PTR_ERR(opq_dentry); + if (IS_ERR(opq_dentry)) + goto out; + dput(opq_dentry); + au_fset_cpdown(*flags, DIROPQ); + +out: + return err; +} + +static int au_cpdown_dir_wh(struct dentry *dentry, struct dentry *h_parent, + struct inode *dir, aufs_bindex_t bdst) +{ + int err; + struct path h_path; + struct au_branch *br; + + br = au_sbr(dentry->d_sb, bdst); + h_path.dentry = au_wh_lkup(h_parent, &dentry->d_name, br); + err = PTR_ERR(h_path.dentry); + if (IS_ERR(h_path.dentry)) + goto out; + + err = 0; + if (d_is_positive(h_path.dentry)) { + h_path.mnt = au_br_mnt(br); + err = au_wh_unlink_dentry(au_h_iptr(dir, bdst), &h_path, + dentry); + } + dput(h_path.dentry); + +out: + return err; +} + +static int au_cpdown_dir(struct dentry *dentry, aufs_bindex_t bdst, + struct au_pin *pin, + struct dentry *h_parent, void *arg) +{ + int err, rerr; + aufs_bindex_t bopq, btop; + struct path h_path; + struct dentry *parent; + struct inode *h_dir, *h_inode, *inode, *dir; + unsigned int *flags = arg; + + btop = au_dbtop(dentry); + /* dentry is di-locked */ + parent = dget_parent(dentry); + dir = d_inode(parent); + h_dir = d_inode(h_parent); + AuDebugOn(h_dir != au_h_iptr(dir, bdst)); + IMustLock(h_dir); + + err = au_lkup_neg(dentry, bdst, /*wh*/0); + if (unlikely(err < 0)) + goto out; + h_path.dentry = au_h_dptr(dentry, bdst); + h_path.mnt = au_sbr_mnt(dentry->d_sb, bdst); + err = vfsub_sio_mkdir(au_h_iptr(dir, bdst), &h_path, 0755); + if (unlikely(err)) + goto out_put; + au_fset_cpdown(*flags, MADE_DIR); + + bopq = au_dbdiropq(dentry); + au_fclr_cpdown(*flags, WHED); + au_fclr_cpdown(*flags, DIROPQ); + if (au_dbwh(dentry) == bdst) + au_fset_cpdown(*flags, WHED); + if (!au_ftest_cpdown(*flags, PARENT_OPQ) && bopq <= bdst) + au_fset_cpdown(*flags, PARENT_OPQ); + h_inode = d_inode(h_path.dentry); + inode_lock_nested(h_inode, AuLsc_I_CHILD); + if (au_ftest_cpdown(*flags, WHED)) { + err = au_cpdown_dir_opq(dentry, bdst, flags); + if (unlikely(err)) { + inode_unlock(h_inode); + goto out_dir; + } + } + + err = au_cpdown_attr(&h_path, au_h_dptr(dentry, btop)); + inode_unlock(h_inode); + if (unlikely(err)) + goto out_opq; + + if (au_ftest_cpdown(*flags, WHED)) { + err = au_cpdown_dir_wh(dentry, h_parent, dir, bdst); + if (unlikely(err)) + goto out_opq; + } + + inode = d_inode(dentry); + if (au_ibbot(inode) < bdst) + au_set_ibbot(inode, bdst); + au_set_h_iptr(inode, bdst, au_igrab(h_inode), + au_hi_flags(inode, /*isdir*/1)); + au_fhsm_wrote(dentry->d_sb, bdst, /*force*/0); + goto out; /* success */ + + /* revert */ +out_opq: + if (au_ftest_cpdown(*flags, DIROPQ)) { + inode_lock_nested(h_inode, AuLsc_I_CHILD); + rerr = au_diropq_remove(dentry, bdst); + inode_unlock(h_inode); + if (unlikely(rerr)) { + AuIOErr("failed removing diropq for %pd b%d (%d)\n", + dentry, bdst, rerr); + err = -EIO; + goto out; + } + } +out_dir: + if (au_ftest_cpdown(*flags, MADE_DIR)) { + rerr = vfsub_sio_rmdir(au_h_iptr(dir, bdst), &h_path); + if (unlikely(rerr)) { + AuIOErr("failed removing %pd b%d (%d)\n", + dentry, bdst, rerr); + err = -EIO; + } + } +out_put: + au_set_h_dptr(dentry, bdst, NULL); + if (au_dbbot(dentry) == bdst) + au_update_dbbot(dentry); +out: + dput(parent); + return err; +} + +int au_cpdown_dirs(struct dentry *dentry, aufs_bindex_t bdst) +{ + int err; + unsigned int flags; + + flags = 0; + err = au_cp_dirs(dentry, bdst, au_cpdown_dir, &flags); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* policies for create */ + +int au_wbr_nonopq(struct dentry *dentry, aufs_bindex_t bindex) +{ + int err, i, j, ndentry; + aufs_bindex_t bopq; + struct au_dcsub_pages dpages; + struct au_dpage *dpage; + struct dentry **dentries, *parent, *d; + + err = au_dpages_init(&dpages, GFP_NOFS); + if (unlikely(err)) + goto out; + parent = dget_parent(dentry); + err = au_dcsub_pages_rev_aufs(&dpages, parent, /*do_include*/0); + if (unlikely(err)) + goto out_free; + + err = bindex; + for (i = 0; i < dpages.ndpage; i++) { + dpage = dpages.dpages + i; + dentries = dpage->dentries; + ndentry = dpage->ndentry; + for (j = 0; j < ndentry; j++) { + d = dentries[j]; + di_read_lock_parent2(d, !AuLock_IR); + bopq = au_dbdiropq(d); + di_read_unlock(d, !AuLock_IR); + if (bopq >= 0 && bopq < err) + err = bopq; + } + } + +out_free: + dput(parent); + au_dpages_free(&dpages); +out: + return err; +} + +static int au_wbr_bu(struct super_block *sb, aufs_bindex_t bindex) +{ + for (; bindex >= 0; bindex--) + if (!au_br_rdonly(au_sbr(sb, bindex))) + return bindex; + return -EROFS; +} + +/* top down parent */ +static int au_wbr_create_tdp(struct dentry *dentry, + unsigned int flags __maybe_unused) +{ + int err; + aufs_bindex_t btop, bindex; + struct super_block *sb; + struct dentry *parent, *h_parent; + + sb = dentry->d_sb; + btop = au_dbtop(dentry); + err = btop; + if (!au_br_rdonly(au_sbr(sb, btop))) + goto out; + + err = -EROFS; + parent = dget_parent(dentry); + for (bindex = au_dbtop(parent); bindex < btop; bindex++) { + h_parent = au_h_dptr(parent, bindex); + if (!h_parent || d_is_negative(h_parent)) + continue; + + if (!au_br_rdonly(au_sbr(sb, bindex))) { + err = bindex; + break; + } + } + dput(parent); + + /* bottom up here */ + if (unlikely(err < 0)) { + err = au_wbr_bu(sb, btop - 1); + if (err >= 0) + err = au_wbr_nonopq(dentry, err); + } + +out: + AuDbg("b%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* an exception for the policy other than tdp */ +static int au_wbr_create_exp(struct dentry *dentry) +{ + int err; + aufs_bindex_t bwh, bdiropq; + struct dentry *parent; + + err = -1; + bwh = au_dbwh(dentry); + parent = dget_parent(dentry); + bdiropq = au_dbdiropq(parent); + if (bwh >= 0) { + if (bdiropq >= 0) + err = min(bdiropq, bwh); + else + err = bwh; + AuDbg("%d\n", err); + } else if (bdiropq >= 0) { + err = bdiropq; + AuDbg("%d\n", err); + } + dput(parent); + + if (err >= 0) + err = au_wbr_nonopq(dentry, err); + + if (err >= 0 && au_br_rdonly(au_sbr(dentry->d_sb, err))) + err = -1; + + AuDbg("%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* round robin */ +static int au_wbr_create_init_rr(struct super_block *sb) +{ + int err; + + err = au_wbr_bu(sb, au_sbbot(sb)); + atomic_set(&au_sbi(sb)->si_wbr_rr_next, -err); /* less important */ + /* smp_mb(); */ + + AuDbg("b%d\n", err); + return err; +} + +static int au_wbr_create_rr(struct dentry *dentry, unsigned int flags) +{ + int err, nbr; + unsigned int u; + aufs_bindex_t bindex, bbot; + struct super_block *sb; + atomic_t *next; + + err = au_wbr_create_exp(dentry); + if (err >= 0) + goto out; + + sb = dentry->d_sb; + next = &au_sbi(sb)->si_wbr_rr_next; + bbot = au_sbbot(sb); + nbr = bbot + 1; + for (bindex = 0; bindex <= bbot; bindex++) { + if (!au_ftest_wbr(flags, DIR)) { + err = atomic_dec_return(next) + 1; + /* modulo for 0 is meaningless */ + if (unlikely(!err)) + err = atomic_dec_return(next) + 1; + } else + err = atomic_read(next); + AuDbg("%d\n", err); + u = err; + err = u % nbr; + AuDbg("%d\n", err); + if (!au_br_rdonly(au_sbr(sb, err))) + break; + err = -EROFS; + } + + if (err >= 0) + err = au_wbr_nonopq(dentry, err); + +out: + AuDbg("%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* most free space */ +static void au_mfs(struct dentry *dentry, struct dentry *parent) +{ + struct super_block *sb; + struct au_branch *br; + struct au_wbr_mfs *mfs; + struct dentry *h_parent; + aufs_bindex_t bindex, bbot; + int err; + unsigned long long b, bavail; + struct path h_path; + /* reduce the stack usage */ + struct kstatfs *st; + + st = kmalloc(sizeof(*st), GFP_NOFS); + if (unlikely(!st)) { + AuWarn1("failed updating mfs(%d), ignored\n", -ENOMEM); + return; + } + + bavail = 0; + sb = dentry->d_sb; + mfs = &au_sbi(sb)->si_wbr_mfs; + MtxMustLock(&mfs->mfs_lock); + mfs->mfs_bindex = -EROFS; + mfs->mfsrr_bytes = 0; + if (!parent) { + bindex = 0; + bbot = au_sbbot(sb); + } else { + bindex = au_dbtop(parent); + bbot = au_dbtaildir(parent); + } + + for (; bindex <= bbot; bindex++) { + if (parent) { + h_parent = au_h_dptr(parent, bindex); + if (!h_parent || d_is_negative(h_parent)) + continue; + } + br = au_sbr(sb, bindex); + if (au_br_rdonly(br)) + continue; + + /* sb->s_root for NFS is unreliable */ + h_path.mnt = au_br_mnt(br); + h_path.dentry = h_path.mnt->mnt_root; + err = vfs_statfs(&h_path, st); + if (unlikely(err)) { + AuWarn1("failed statfs, b%d, %d\n", bindex, err); + continue; + } + + /* when the available size is equal, select the lower one */ + BUILD_BUG_ON(sizeof(b) < sizeof(st->f_bavail) + || sizeof(b) < sizeof(st->f_bsize)); + b = st->f_bavail * st->f_bsize; + br->br_wbr->wbr_bytes = b; + if (b >= bavail) { + bavail = b; + mfs->mfs_bindex = bindex; + mfs->mfs_jiffy = jiffies; + } + } + + mfs->mfsrr_bytes = bavail; + AuDbg("b%d\n", mfs->mfs_bindex); + au_kfree_rcu(st); +} + +static int au_wbr_create_mfs(struct dentry *dentry, unsigned int flags) +{ + int err; + struct dentry *parent; + struct super_block *sb; + struct au_wbr_mfs *mfs; + + err = au_wbr_create_exp(dentry); + if (err >= 0) + goto out; + + sb = dentry->d_sb; + parent = NULL; + if (au_ftest_wbr(flags, PARENT)) + parent = dget_parent(dentry); + mfs = &au_sbi(sb)->si_wbr_mfs; + mutex_lock(&mfs->mfs_lock); + if (time_after(jiffies, mfs->mfs_jiffy + mfs->mfs_expire) + || mfs->mfs_bindex < 0 + || au_br_rdonly(au_sbr(sb, mfs->mfs_bindex))) + au_mfs(dentry, parent); + mutex_unlock(&mfs->mfs_lock); + err = mfs->mfs_bindex; + dput(parent); + + if (err >= 0) + err = au_wbr_nonopq(dentry, err); + +out: + AuDbg("b%d\n", err); + return err; +} + +static int au_wbr_create_init_mfs(struct super_block *sb) +{ + struct au_wbr_mfs *mfs; + + mfs = &au_sbi(sb)->si_wbr_mfs; + mutex_init(&mfs->mfs_lock); + mfs->mfs_jiffy = 0; + mfs->mfs_bindex = -EROFS; + + return 0; +} + +static int au_wbr_create_fin_mfs(struct super_block *sb __maybe_unused) +{ + mutex_destroy(&au_sbi(sb)->si_wbr_mfs.mfs_lock); + return 0; +} + +/* ---------------------------------------------------------------------- */ + +/* top down regardless parent, and then mfs */ +static int au_wbr_create_tdmfs(struct dentry *dentry, + unsigned int flags __maybe_unused) +{ + int err; + aufs_bindex_t bwh, btail, bindex, bfound, bmfs; + unsigned long long watermark; + struct super_block *sb; + struct au_wbr_mfs *mfs; + struct au_branch *br; + struct dentry *parent; + + sb = dentry->d_sb; + mfs = &au_sbi(sb)->si_wbr_mfs; + mutex_lock(&mfs->mfs_lock); + if (time_after(jiffies, mfs->mfs_jiffy + mfs->mfs_expire) + || mfs->mfs_bindex < 0) + au_mfs(dentry, /*parent*/NULL); + watermark = mfs->mfsrr_watermark; + bmfs = mfs->mfs_bindex; + mutex_unlock(&mfs->mfs_lock); + + /* another style of au_wbr_create_exp() */ + bwh = au_dbwh(dentry); + parent = dget_parent(dentry); + btail = au_dbtaildir(parent); + if (bwh >= 0 && bwh < btail) + btail = bwh; + + err = au_wbr_nonopq(dentry, btail); + if (unlikely(err < 0)) + goto out; + btail = err; + bfound = -1; + for (bindex = 0; bindex <= btail; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_rdonly(br)) + continue; + if (br->br_wbr->wbr_bytes > watermark) { + bfound = bindex; + break; + } + } + err = bfound; + if (err < 0) + err = bmfs; + +out: + dput(parent); + AuDbg("b%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* most free space and then round robin */ +static int au_wbr_create_mfsrr(struct dentry *dentry, unsigned int flags) +{ + int err; + struct au_wbr_mfs *mfs; + + err = au_wbr_create_mfs(dentry, flags); + if (err >= 0) { + mfs = &au_sbi(dentry->d_sb)->si_wbr_mfs; + mutex_lock(&mfs->mfs_lock); + if (mfs->mfsrr_bytes < mfs->mfsrr_watermark) + err = au_wbr_create_rr(dentry, flags); + mutex_unlock(&mfs->mfs_lock); + } + + AuDbg("b%d\n", err); + return err; +} + +static int au_wbr_create_init_mfsrr(struct super_block *sb) +{ + int err; + + au_wbr_create_init_mfs(sb); /* ignore */ + err = au_wbr_create_init_rr(sb); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* top down parent and most free space */ +static int au_wbr_create_pmfs(struct dentry *dentry, unsigned int flags) +{ + int err, e2; + unsigned long long b; + aufs_bindex_t bindex, btop, bbot; + struct super_block *sb; + struct dentry *parent, *h_parent; + struct au_branch *br; + + err = au_wbr_create_tdp(dentry, flags); + if (unlikely(err < 0)) + goto out; + parent = dget_parent(dentry); + btop = au_dbtop(parent); + bbot = au_dbtaildir(parent); + if (btop == bbot) + goto out_parent; /* success */ + + e2 = au_wbr_create_mfs(dentry, flags); + if (e2 < 0) + goto out_parent; /* success */ + + /* when the available size is equal, select upper one */ + sb = dentry->d_sb; + br = au_sbr(sb, err); + b = br->br_wbr->wbr_bytes; + AuDbg("b%d, %llu\n", err, b); + + for (bindex = btop; bindex <= bbot; bindex++) { + h_parent = au_h_dptr(parent, bindex); + if (!h_parent || d_is_negative(h_parent)) + continue; + + br = au_sbr(sb, bindex); + if (!au_br_rdonly(br) && br->br_wbr->wbr_bytes > b) { + b = br->br_wbr->wbr_bytes; + err = bindex; + AuDbg("b%d, %llu\n", err, b); + } + } + + if (err >= 0) + err = au_wbr_nonopq(dentry, err); + +out_parent: + dput(parent); +out: + AuDbg("b%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * - top down parent + * - most free space with parent + * - most free space round-robin regardless parent + */ +static int au_wbr_create_pmfsrr(struct dentry *dentry, unsigned int flags) +{ + int err; + unsigned long long watermark; + struct super_block *sb; + struct au_branch *br; + struct au_wbr_mfs *mfs; + + err = au_wbr_create_pmfs(dentry, flags | AuWbr_PARENT); + if (unlikely(err < 0)) + goto out; + + sb = dentry->d_sb; + br = au_sbr(sb, err); + mfs = &au_sbi(sb)->si_wbr_mfs; + mutex_lock(&mfs->mfs_lock); + watermark = mfs->mfsrr_watermark; + mutex_unlock(&mfs->mfs_lock); + if (br->br_wbr->wbr_bytes < watermark) + /* regardless the parent dir */ + err = au_wbr_create_mfsrr(dentry, flags); + +out: + AuDbg("b%d\n", err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* policies for copyup */ + +/* top down parent */ +static int au_wbr_copyup_tdp(struct dentry *dentry) +{ + return au_wbr_create_tdp(dentry, /*flags, anything is ok*/0); +} + +/* bottom up parent */ +static int au_wbr_copyup_bup(struct dentry *dentry) +{ + int err; + aufs_bindex_t bindex, btop; + struct dentry *parent, *h_parent; + struct super_block *sb; + + err = -EROFS; + sb = dentry->d_sb; + parent = dget_parent(dentry); + btop = au_dbtop(parent); + for (bindex = au_dbtop(dentry); bindex >= btop; bindex--) { + h_parent = au_h_dptr(parent, bindex); + if (!h_parent || d_is_negative(h_parent)) + continue; + + if (!au_br_rdonly(au_sbr(sb, bindex))) { + err = bindex; + break; + } + } + dput(parent); + + /* bottom up here */ + if (unlikely(err < 0)) + err = au_wbr_bu(sb, btop - 1); + + AuDbg("b%d\n", err); + return err; +} + +/* bottom up */ +int au_wbr_do_copyup_bu(struct dentry *dentry, aufs_bindex_t btop) +{ + int err; + + err = au_wbr_bu(dentry->d_sb, btop); + AuDbg("b%d\n", err); + if (err > btop) + err = au_wbr_nonopq(dentry, err); + + AuDbg("b%d\n", err); + return err; +} + +static int au_wbr_copyup_bu(struct dentry *dentry) +{ + int err; + aufs_bindex_t btop; + + btop = au_dbtop(dentry); + err = au_wbr_do_copyup_bu(dentry, btop); + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct au_wbr_copyup_operations au_wbr_copyup_ops[] = { + [AuWbrCopyup_TDP] = { + .copyup = au_wbr_copyup_tdp + }, + [AuWbrCopyup_BUP] = { + .copyup = au_wbr_copyup_bup + }, + [AuWbrCopyup_BU] = { + .copyup = au_wbr_copyup_bu + } +}; + +struct au_wbr_create_operations au_wbr_create_ops[] = { + [AuWbrCreate_TDP] = { + .create = au_wbr_create_tdp + }, + [AuWbrCreate_RR] = { + .create = au_wbr_create_rr, + .init = au_wbr_create_init_rr + }, + [AuWbrCreate_MFS] = { + .create = au_wbr_create_mfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_MFSV] = { + .create = au_wbr_create_mfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_MFSRR] = { + .create = au_wbr_create_mfsrr, + .init = au_wbr_create_init_mfsrr, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_MFSRRV] = { + .create = au_wbr_create_mfsrr, + .init = au_wbr_create_init_mfsrr, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_TDMFS] = { + .create = au_wbr_create_tdmfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_TDMFSV] = { + .create = au_wbr_create_tdmfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_PMFS] = { + .create = au_wbr_create_pmfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_PMFSV] = { + .create = au_wbr_create_pmfs, + .init = au_wbr_create_init_mfs, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_PMFSRR] = { + .create = au_wbr_create_pmfsrr, + .init = au_wbr_create_init_mfsrr, + .fin = au_wbr_create_fin_mfs + }, + [AuWbrCreate_PMFSRRV] = { + .create = au_wbr_create_pmfsrr, + .init = au_wbr_create_init_mfsrr, + .fin = au_wbr_create_fin_mfs + } +}; diff --git a/fs/aufs/whout.c b/fs/aufs/whout.c new file mode 100644 index 00000000000..3150087c4f8 --- /dev/null +++ b/fs/aufs/whout.c @@ -0,0 +1,1070 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * whiteout for logical deletion and opaque directory + */ + +#include "aufs.h" + +#define WH_MASK 0444 + +/* + * If a directory contains this file, then it is opaque. We start with the + * .wh. flag so that it is blocked by lookup. + */ +static struct qstr diropq_name = QSTR_INIT(AUFS_WH_DIROPQ, + sizeof(AUFS_WH_DIROPQ) - 1); + +/* + * generate whiteout name, which is NOT terminated by NULL. + * @name: original d_name.name + * @len: original d_name.len + * @wh: whiteout qstr + * returns zero when succeeds, otherwise error. + * succeeded value as wh->name should be freed by kfree(). + */ +int au_wh_name_alloc(struct qstr *wh, const struct qstr *name) +{ + char *p; + + if (unlikely(name->len > PATH_MAX - AUFS_WH_PFX_LEN)) + return -ENAMETOOLONG; + + wh->len = name->len + AUFS_WH_PFX_LEN; + p = kmalloc(wh->len, GFP_NOFS); + wh->name = p; + if (p) { + memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); + memcpy(p + AUFS_WH_PFX_LEN, name->name, name->len); + /* smp_mb(); */ + return 0; + } + return -ENOMEM; +} + +/* ---------------------------------------------------------------------- */ + +/* + * test if the @wh_name exists under @h_parent. + * @try_sio specifies the necessary of super-io. + */ +int au_wh_test(struct user_namespace *h_userns, struct dentry *h_parent, + struct qstr *wh_name, int try_sio) +{ + int err; + struct dentry *wh_dentry; + + if (!try_sio) + wh_dentry = vfsub_lkup_one(wh_name, h_parent); + else + wh_dentry = au_sio_lkup_one(h_userns, wh_name, h_parent); + err = PTR_ERR(wh_dentry); + if (IS_ERR(wh_dentry)) { + if (err == -ENAMETOOLONG) + err = 0; + goto out; + } + + err = 0; + if (d_is_negative(wh_dentry)) + goto out_wh; /* success */ + + err = 1; + if (d_is_reg(wh_dentry)) + goto out_wh; /* success */ + + err = -EIO; + AuIOErr("%pd Invalid whiteout entry type 0%o.\n", + wh_dentry, d_inode(wh_dentry)->i_mode); + +out_wh: + dput(wh_dentry); +out: + return err; +} + +/* + * test if the @h_dentry sets opaque or not. + */ +int au_diropq_test(struct user_namespace *h_userns, struct dentry *h_dentry) +{ + int err; + struct inode *h_dir; + + h_dir = d_inode(h_dentry); + err = au_wh_test(h_userns, h_dentry, &diropq_name, + au_test_h_perm_sio(h_userns, h_dir, MAY_EXEC)); + return err; +} + +/* + * returns a negative dentry whose name is unique and temporary. + */ +struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct au_branch *br, + struct qstr *prefix) +{ + struct dentry *dentry; + int i; + char defname[NAME_MAX - AUFS_MAX_NAMELEN + DNAME_INLINE_LEN + 1], + *name, *p; + /* strict atomic_t is unnecessary here */ + static unsigned short cnt; + struct qstr qs; + struct user_namespace *h_userns; + + BUILD_BUG_ON(sizeof(cnt) * 2 > AUFS_WH_TMP_LEN); + + name = defname; + qs.len = sizeof(defname) - DNAME_INLINE_LEN + prefix->len - 1; + if (unlikely(prefix->len > DNAME_INLINE_LEN)) { + dentry = ERR_PTR(-ENAMETOOLONG); + if (unlikely(qs.len > NAME_MAX)) + goto out; + dentry = ERR_PTR(-ENOMEM); + name = kmalloc(qs.len + 1, GFP_NOFS); + if (unlikely(!name)) + goto out; + } + + /* doubly whiteout-ed */ + memcpy(name, AUFS_WH_PFX AUFS_WH_PFX, AUFS_WH_PFX_LEN * 2); + p = name + AUFS_WH_PFX_LEN * 2; + memcpy(p, prefix->name, prefix->len); + p += prefix->len; + *p++ = '.'; + AuDebugOn(name + qs.len + 1 - p <= AUFS_WH_TMP_LEN); + + h_userns = au_br_userns(br); + qs.name = name; + for (i = 0; i < 3; i++) { + sprintf(p, "%.*x", AUFS_WH_TMP_LEN, cnt++); + dentry = au_sio_lkup_one(h_userns, &qs, h_parent); + if (IS_ERR(dentry) || d_is_negative(dentry)) + goto out_name; + dput(dentry); + } + /* pr_warn("could not get random name\n"); */ + dentry = ERR_PTR(-EEXIST); + AuDbg("%.*s\n", AuLNPair(&qs)); + BUG(); + +out_name: + if (name != defname) + au_kfree_try_rcu(name); +out: + AuTraceErrPtr(dentry); + return dentry; +} + +/* + * rename the @h_dentry on @br to the whiteouted temporary name. + */ +int au_whtmp_ren(struct dentry *h_dentry, struct au_branch *br) +{ + int err; + struct path h_path = { + .mnt = au_br_mnt(br) + }; + struct inode *h_dir, *delegated; + struct dentry *h_parent; + + h_parent = h_dentry->d_parent; /* dir inode is locked */ + h_dir = d_inode(h_parent); + IMustLock(h_dir); + + h_path.dentry = au_whtmp_lkup(h_parent, br, &h_dentry->d_name); + err = PTR_ERR(h_path.dentry); + if (IS_ERR(h_path.dentry)) + goto out; + + /* under the same dir, no need to lock_rename() */ + delegated = NULL; + err = vfsub_rename(h_dir, h_dentry, h_dir, &h_path, &delegated, + /*flags*/0); + AuTraceErr(err); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal rename\n"); + iput(delegated); + } + dput(h_path.dentry); + +out: + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ +/* + * functions for removing a whiteout + */ + +static int do_unlink_wh(struct inode *h_dir, struct path *h_path) +{ + int err, force; + struct inode *delegated; + + /* + * forces superio when the dir has a sticky bit. + * this may be a violation of unix fs semantics. + */ + force = (h_dir->i_mode & S_ISVTX) + && !uid_eq(current_fsuid(), d_inode(h_path->dentry)->i_uid); + delegated = NULL; + err = vfsub_unlink(h_dir, h_path, &delegated, force); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + return err; +} + +int au_wh_unlink_dentry(struct inode *h_dir, struct path *h_path, + struct dentry *dentry) +{ + int err; + + err = do_unlink_wh(h_dir, h_path); + if (!err && dentry) + au_set_dbwh(dentry, -1); + + return err; +} + +static int unlink_wh_name(struct dentry *h_parent, struct qstr *wh, + struct au_branch *br) +{ + int err; + struct path h_path = { + .mnt = au_br_mnt(br) + }; + + err = 0; + h_path.dentry = vfsub_lkup_one(wh, h_parent); + if (IS_ERR(h_path.dentry)) + err = PTR_ERR(h_path.dentry); + else { + if (d_is_reg(h_path.dentry)) + err = do_unlink_wh(d_inode(h_parent), &h_path); + dput(h_path.dentry); + } + + return err; +} + +/* ---------------------------------------------------------------------- */ +/* + * initialize/clean whiteout for a branch + */ + +static void au_wh_clean(struct inode *h_dir, struct path *whpath, + const int isdir) +{ + int err; + struct inode *delegated; + + if (d_is_negative(whpath->dentry)) + return; + + if (isdir) + err = vfsub_rmdir(h_dir, whpath); + else { + delegated = NULL; + err = vfsub_unlink(h_dir, whpath, &delegated, /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + } + if (unlikely(err)) + pr_warn("failed removing %pd (%d), ignored.\n", + whpath->dentry, err); +} + +static int test_linkable(struct dentry *h_root) +{ + struct inode *h_dir = d_inode(h_root); + + if (h_dir->i_op->link) + return 0; + + pr_err("%pd (%s) doesn't support link(2), use noplink and rw+nolwh\n", + h_root, au_sbtype(h_root->d_sb)); + return -ENOSYS; /* the branch doesn't have its ->link() */ +} + +/* todo: should this mkdir be done in /sbin/mount.aufs helper? */ +static int au_whdir(struct inode *h_dir, struct path *path) +{ + int err; + + err = -EEXIST; + if (d_is_negative(path->dentry)) { + int mode = 0700; + + if (au_test_nfs(path->dentry->d_sb)) + mode |= 0111; + err = vfsub_mkdir(h_dir, path, mode); + } else if (d_is_dir(path->dentry)) + err = 0; + else + pr_err("unknown %pd exists\n", path->dentry); + + return err; +} + +struct au_wh_base { + const struct qstr *name; + struct dentry *dentry; +}; + +static void au_wh_init_ro(struct inode *h_dir, struct au_wh_base base[], + struct path *h_path) +{ + h_path->dentry = base[AuBrWh_BASE].dentry; + au_wh_clean(h_dir, h_path, /*isdir*/0); + h_path->dentry = base[AuBrWh_PLINK].dentry; + au_wh_clean(h_dir, h_path, /*isdir*/1); + h_path->dentry = base[AuBrWh_ORPH].dentry; + au_wh_clean(h_dir, h_path, /*isdir*/1); +} + +/* + * returns tri-state, + * minus: error, caller should print the message + * zero: success + * plus: error, caller should NOT print the message + */ +static int au_wh_init_rw_nolink(struct dentry *h_root, struct au_wbr *wbr, + int do_plink, struct au_wh_base base[], + struct path *h_path) +{ + int err; + struct inode *h_dir; + + h_dir = d_inode(h_root); + h_path->dentry = base[AuBrWh_BASE].dentry; + au_wh_clean(h_dir, h_path, /*isdir*/0); + h_path->dentry = base[AuBrWh_PLINK].dentry; + if (do_plink) { + err = test_linkable(h_root); + if (unlikely(err)) { + err = 1; + goto out; + } + + err = au_whdir(h_dir, h_path); + if (unlikely(err)) + goto out; + wbr->wbr_plink = dget(base[AuBrWh_PLINK].dentry); + } else + au_wh_clean(h_dir, h_path, /*isdir*/1); + h_path->dentry = base[AuBrWh_ORPH].dentry; + err = au_whdir(h_dir, h_path); + if (unlikely(err)) + goto out; + wbr->wbr_orph = dget(base[AuBrWh_ORPH].dentry); + +out: + return err; +} + +/* + * for the moment, aufs supports the branch filesystem which does not support + * link(2). testing on FAT which does not support i_op->setattr() fully either, + * copyup failed. finally, such filesystem will not be used as the writable + * branch. + * + * returns tri-state, see above. + */ +static int au_wh_init_rw(struct dentry *h_root, struct au_wbr *wbr, + int do_plink, struct au_wh_base base[], + struct path *h_path) +{ + int err; + struct inode *h_dir; + + WbrWhMustWriteLock(wbr); + + err = test_linkable(h_root); + if (unlikely(err)) { + err = 1; + goto out; + } + + /* + * todo: should this create be done in /sbin/mount.aufs helper? + */ + err = -EEXIST; + h_dir = d_inode(h_root); + if (d_is_negative(base[AuBrWh_BASE].dentry)) { + h_path->dentry = base[AuBrWh_BASE].dentry; + err = vfsub_create(h_dir, h_path, WH_MASK, /*want_excl*/true); + } else if (d_is_reg(base[AuBrWh_BASE].dentry)) + err = 0; + else + pr_err("unknown %pd2 exists\n", base[AuBrWh_BASE].dentry); + if (unlikely(err)) + goto out; + + h_path->dentry = base[AuBrWh_PLINK].dentry; + if (do_plink) { + err = au_whdir(h_dir, h_path); + if (unlikely(err)) + goto out; + wbr->wbr_plink = dget(base[AuBrWh_PLINK].dentry); + } else + au_wh_clean(h_dir, h_path, /*isdir*/1); + wbr->wbr_whbase = dget(base[AuBrWh_BASE].dentry); + + h_path->dentry = base[AuBrWh_ORPH].dentry; + err = au_whdir(h_dir, h_path); + if (unlikely(err)) + goto out; + wbr->wbr_orph = dget(base[AuBrWh_ORPH].dentry); + +out: + return err; +} + +/* + * initialize the whiteout base file/dir for @br. + */ +int au_wh_init(struct au_branch *br, struct super_block *sb) +{ + int err, i; + const unsigned char do_plink + = !!au_opt_test(au_mntflags(sb), PLINK); + struct inode *h_dir; + struct path path = br->br_path; + struct dentry *h_root = path.dentry; + struct au_wbr *wbr = br->br_wbr; + static const struct qstr base_name[] = { + [AuBrWh_BASE] = QSTR_INIT(AUFS_BASE_NAME, + sizeof(AUFS_BASE_NAME) - 1), + [AuBrWh_PLINK] = QSTR_INIT(AUFS_PLINKDIR_NAME, + sizeof(AUFS_PLINKDIR_NAME) - 1), + [AuBrWh_ORPH] = QSTR_INIT(AUFS_ORPHDIR_NAME, + sizeof(AUFS_ORPHDIR_NAME) - 1) + }; + struct au_wh_base base[] = { + [AuBrWh_BASE] = { + .name = base_name + AuBrWh_BASE, + .dentry = NULL + }, + [AuBrWh_PLINK] = { + .name = base_name + AuBrWh_PLINK, + .dentry = NULL + }, + [AuBrWh_ORPH] = { + .name = base_name + AuBrWh_ORPH, + .dentry = NULL + } + }; + + if (wbr) + WbrWhMustWriteLock(wbr); + + for (i = 0; i < AuBrWh_Last; i++) { + /* doubly whiteouted */ + struct dentry *d; + + d = au_wh_lkup(h_root, (void *)base[i].name, br); + err = PTR_ERR(d); + if (IS_ERR(d)) + goto out; + + base[i].dentry = d; + AuDebugOn(wbr + && wbr->wbr_wh[i] + && wbr->wbr_wh[i] != base[i].dentry); + } + + if (wbr) + for (i = 0; i < AuBrWh_Last; i++) { + dput(wbr->wbr_wh[i]); + wbr->wbr_wh[i] = NULL; + } + + err = 0; + if (!au_br_writable(br->br_perm)) { + h_dir = d_inode(h_root); + au_wh_init_ro(h_dir, base, &path); + } else if (!au_br_wh_linkable(br->br_perm)) { + err = au_wh_init_rw_nolink(h_root, wbr, do_plink, base, &path); + if (err > 0) + goto out; + else if (err) + goto out_err; + } else { + err = au_wh_init_rw(h_root, wbr, do_plink, base, &path); + if (err > 0) + goto out; + else if (err) + goto out_err; + } + goto out; /* success */ + +out_err: + pr_err("an error(%d) on the writable branch %pd(%s)\n", + err, h_root, au_sbtype(h_root->d_sb)); +out: + for (i = 0; i < AuBrWh_Last; i++) + dput(base[i].dentry); + return err; +} + +/* ---------------------------------------------------------------------- */ +/* + * whiteouts are all hard-linked usually. + * when its link count reaches a ceiling, we create a new whiteout base + * asynchronously. + */ + +struct reinit_br_wh { + struct super_block *sb; + struct au_branch *br; +}; + +static void reinit_br_wh(void *arg) +{ + int err; + aufs_bindex_t bindex; + struct path h_path; + struct reinit_br_wh *a = arg; + struct au_wbr *wbr; + struct inode *dir, *delegated; + struct dentry *h_root; + struct au_hinode *hdir; + + err = 0; + wbr = a->br->br_wbr; + /* big aufs lock */ + si_noflush_write_lock(a->sb); + if (!au_br_writable(a->br->br_perm)) + goto out; + bindex = au_br_index(a->sb, a->br->br_id); + if (unlikely(bindex < 0)) + goto out; + + di_read_lock_parent(a->sb->s_root, AuLock_IR); + dir = d_inode(a->sb->s_root); + hdir = au_hi(dir, bindex); + h_root = au_h_dptr(a->sb->s_root, bindex); + AuDebugOn(h_root != au_br_dentry(a->br)); + + au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT); + wbr_wh_write_lock(wbr); + err = au_h_verify(wbr->wbr_whbase, au_opt_udba(a->sb), hdir->hi_inode, + h_root, a->br); + if (!err) { + h_path.dentry = wbr->wbr_whbase; + h_path.mnt = au_br_mnt(a->br); + delegated = NULL; + err = vfsub_unlink(hdir->hi_inode, &h_path, &delegated, + /*force*/0); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + } else { + pr_warn("%pd is moved, ignored\n", wbr->wbr_whbase); + err = 0; + } + dput(wbr->wbr_whbase); + wbr->wbr_whbase = NULL; + if (!err) + err = au_wh_init(a->br, a->sb); + wbr_wh_write_unlock(wbr); + au_hn_inode_unlock(hdir); + di_read_unlock(a->sb->s_root, AuLock_IR); + if (!err) + au_fhsm_wrote(a->sb, bindex, /*force*/0); + +out: + if (wbr) + atomic_dec(&wbr->wbr_wh_running); + au_lcnt_dec(&a->br->br_count); + si_write_unlock(a->sb); + au_nwt_done(&au_sbi(a->sb)->si_nowait); + au_kfree_rcu(a); + if (unlikely(err)) + AuIOErr("err %d\n", err); +} + +static void kick_reinit_br_wh(struct super_block *sb, struct au_branch *br) +{ + int do_dec, wkq_err; + struct reinit_br_wh *arg; + + do_dec = 1; + if (atomic_inc_return(&br->br_wbr->wbr_wh_running) != 1) + goto out; + + /* ignore ENOMEM */ + arg = kmalloc(sizeof(*arg), GFP_NOFS); + if (arg) { + /* + * dec(wh_running), kfree(arg) and dec(br_count) + * in reinit function + */ + arg->sb = sb; + arg->br = br; + au_lcnt_inc(&br->br_count); + wkq_err = au_wkq_nowait(reinit_br_wh, arg, sb, /*flags*/0); + if (unlikely(wkq_err)) { + atomic_dec(&br->br_wbr->wbr_wh_running); + au_lcnt_dec(&br->br_count); + au_kfree_rcu(arg); + } + do_dec = 0; + } + +out: + if (do_dec) + atomic_dec(&br->br_wbr->wbr_wh_running); +} + +/* ---------------------------------------------------------------------- */ + +/* + * create the whiteout @wh. + */ +static int link_or_create_wh(struct super_block *sb, aufs_bindex_t bindex, + struct dentry *wh) +{ + int err; + struct path h_path = { + .dentry = wh + }; + struct au_branch *br; + struct au_wbr *wbr; + struct dentry *h_parent; + struct inode *h_dir, *delegated; + + h_parent = wh->d_parent; /* dir inode is locked */ + h_dir = d_inode(h_parent); + IMustLock(h_dir); + + br = au_sbr(sb, bindex); + h_path.mnt = au_br_mnt(br); + wbr = br->br_wbr; + wbr_wh_read_lock(wbr); + if (wbr->wbr_whbase) { + delegated = NULL; + err = vfsub_link(wbr->wbr_whbase, h_dir, &h_path, &delegated); + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal link\n"); + iput(delegated); + } + if (!err || err != -EMLINK) + goto out; + + /* link count full. re-initialize br_whbase. */ + kick_reinit_br_wh(sb, br); + } + + /* return this error in this context */ + err = vfsub_create(h_dir, &h_path, WH_MASK, /*want_excl*/true); + if (!err) + au_fhsm_wrote(sb, bindex, /*force*/0); + +out: + wbr_wh_read_unlock(wbr); + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * create or remove the diropq. + */ +static struct dentry *do_diropq(struct dentry *dentry, aufs_bindex_t bindex, + unsigned int flags) +{ + struct dentry *opq_dentry, *h_dentry; + struct super_block *sb; + struct au_branch *br; + int err; + + sb = dentry->d_sb; + br = au_sbr(sb, bindex); + h_dentry = au_h_dptr(dentry, bindex); + opq_dentry = vfsub_lkup_one(&diropq_name, h_dentry); + if (IS_ERR(opq_dentry)) + goto out; + + if (au_ftest_diropq(flags, CREATE)) { + err = link_or_create_wh(sb, bindex, opq_dentry); + if (!err) { + au_set_dbdiropq(dentry, bindex); + goto out; /* success */ + } + } else { + struct path tmp = { + .dentry = opq_dentry, + .mnt = au_br_mnt(br) + }; + err = do_unlink_wh(au_h_iptr(d_inode(dentry), bindex), &tmp); + if (!err) + au_set_dbdiropq(dentry, -1); + } + dput(opq_dentry); + opq_dentry = ERR_PTR(err); + +out: + return opq_dentry; +} + +struct do_diropq_args { + struct dentry **errp; + struct dentry *dentry; + aufs_bindex_t bindex; + unsigned int flags; +}; + +static void call_do_diropq(void *args) +{ + struct do_diropq_args *a = args; + *a->errp = do_diropq(a->dentry, a->bindex, a->flags); +} + +struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex, + unsigned int flags) +{ + struct dentry *diropq, *h_dentry; + struct user_namespace *h_userns; + + h_userns = au_sbr_userns(dentry->d_sb, bindex); + h_dentry = au_h_dptr(dentry, bindex); + if (!au_test_h_perm_sio(h_userns, d_inode(h_dentry), + MAY_EXEC | MAY_WRITE)) + diropq = do_diropq(dentry, bindex, flags); + else { + int wkq_err; + struct do_diropq_args args = { + .errp = &diropq, + .dentry = dentry, + .bindex = bindex, + .flags = flags + }; + + wkq_err = au_wkq_wait(call_do_diropq, &args); + if (unlikely(wkq_err)) + diropq = ERR_PTR(wkq_err); + } + + return diropq; +} + +/* ---------------------------------------------------------------------- */ + +/* + * lookup whiteout dentry. + * @h_parent: lower parent dentry which must exist and be locked + * @base_name: name of dentry which will be whiteouted + * returns dentry for whiteout. + */ +struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name, + struct au_branch *br) +{ + int err; + struct qstr wh_name; + struct dentry *wh_dentry; + + err = au_wh_name_alloc(&wh_name, base_name); + wh_dentry = ERR_PTR(err); + if (!err) { + wh_dentry = vfsub_lkup_one(&wh_name, h_parent); + au_kfree_try_rcu(wh_name.name); + } + return wh_dentry; +} + +/* + * link/create a whiteout for @dentry on @bindex. + */ +struct dentry *au_wh_create(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent) +{ + struct dentry *wh_dentry; + struct super_block *sb; + int err; + + sb = dentry->d_sb; + wh_dentry = au_wh_lkup(h_parent, &dentry->d_name, au_sbr(sb, bindex)); + if (!IS_ERR(wh_dentry) && d_is_negative(wh_dentry)) { + err = link_or_create_wh(sb, bindex, wh_dentry); + if (!err) { + au_set_dbwh(dentry, bindex); + au_fhsm_wrote(sb, bindex, /*force*/0); + } else { + dput(wh_dentry); + wh_dentry = ERR_PTR(err); + } + } + + return wh_dentry; +} + +/* ---------------------------------------------------------------------- */ + +/* Delete all whiteouts in this directory on branch bindex. */ +static int del_wh_children(struct dentry *h_dentry, struct au_nhash *whlist, + aufs_bindex_t bindex, struct au_branch *br) +{ + int err; + unsigned long ul, n; + struct qstr wh_name; + char *p; + struct hlist_head *head; + struct au_vdir_wh *pos; + struct au_vdir_destr *str; + + err = -ENOMEM; + p = (void *)__get_free_page(GFP_NOFS); + wh_name.name = p; + if (unlikely(!wh_name.name)) + goto out; + + err = 0; + memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN); + p += AUFS_WH_PFX_LEN; + n = whlist->nh_num; + head = whlist->nh_head; + for (ul = 0; !err && ul < n; ul++, head++) { + hlist_for_each_entry(pos, head, wh_hash) { + if (pos->wh_bindex != bindex) + continue; + + str = &pos->wh_str; + if (str->len + AUFS_WH_PFX_LEN <= PATH_MAX) { + memcpy(p, str->name, str->len); + wh_name.len = AUFS_WH_PFX_LEN + str->len; + err = unlink_wh_name(h_dentry, &wh_name, br); + if (!err) + continue; + break; + } + AuIOErr("whiteout name too long %.*s\n", + str->len, str->name); + err = -EIO; + break; + } + } + free_page((unsigned long)wh_name.name); + +out: + return err; +} + +struct del_wh_children_args { + int *errp; + struct dentry *h_dentry; + struct au_nhash *whlist; + aufs_bindex_t bindex; + struct au_branch *br; +}; + +static void call_del_wh_children(void *args) +{ + struct del_wh_children_args *a = args; + *a->errp = del_wh_children(a->h_dentry, a->whlist, a->bindex, a->br); +} + +/* ---------------------------------------------------------------------- */ + +struct au_whtmp_rmdir *au_whtmp_rmdir_alloc(struct super_block *sb, gfp_t gfp) +{ + struct au_whtmp_rmdir *whtmp; + int err; + unsigned int rdhash; + + SiMustAnyLock(sb); + + whtmp = kzalloc(sizeof(*whtmp), gfp); + if (unlikely(!whtmp)) { + whtmp = ERR_PTR(-ENOMEM); + goto out; + } + + /* no estimation for dir size */ + rdhash = au_sbi(sb)->si_rdhash; + if (!rdhash) + rdhash = AUFS_RDHASH_DEF; + err = au_nhash_alloc(&whtmp->whlist, rdhash, gfp); + if (unlikely(err)) { + au_kfree_rcu(whtmp); + whtmp = ERR_PTR(err); + } + +out: + return whtmp; +} + +void au_whtmp_rmdir_free(struct au_whtmp_rmdir *whtmp) +{ + if (whtmp->br) + au_lcnt_dec(&whtmp->br->br_count); + dput(whtmp->wh_dentry); + iput(whtmp->dir); + au_nhash_wh_free(&whtmp->whlist); + au_kfree_rcu(whtmp); +} + +/* + * rmdir the whiteouted temporary named dir @h_dentry. + * @whlist: whiteouted children. + */ +int au_whtmp_rmdir(struct inode *dir, aufs_bindex_t bindex, + struct dentry *wh_dentry, struct au_nhash *whlist) +{ + int err; + unsigned int h_nlink; + struct path h_tmp; + struct inode *wh_inode, *h_dir; + struct au_branch *br; + struct user_namespace *h_userns; + + h_dir = d_inode(wh_dentry->d_parent); /* dir inode is locked */ + IMustLock(h_dir); + + br = au_sbr(dir->i_sb, bindex); + h_userns = au_br_userns(br); + wh_inode = d_inode(wh_dentry); + inode_lock_nested(wh_inode, AuLsc_I_CHILD); + + /* + * someone else might change some whiteouts while we were sleeping. + * it means this whlist may have an obsoleted entry. + */ + if (!au_test_h_perm_sio(h_userns, wh_inode, MAY_EXEC | MAY_WRITE)) + err = del_wh_children(wh_dentry, whlist, bindex, br); + else { + int wkq_err; + struct del_wh_children_args args = { + .errp = &err, + .h_dentry = wh_dentry, + .whlist = whlist, + .bindex = bindex, + .br = br + }; + + wkq_err = au_wkq_wait(call_del_wh_children, &args); + if (unlikely(wkq_err)) + err = wkq_err; + } + inode_unlock(wh_inode); + + if (!err) { + h_tmp.dentry = wh_dentry; + h_tmp.mnt = au_br_mnt(br); + h_nlink = h_dir->i_nlink; + err = vfsub_rmdir(h_dir, &h_tmp); + /* some fs doesn't change the parent nlink in some cases */ + h_nlink -= h_dir->i_nlink; + } + + if (!err) { + if (au_ibtop(dir) == bindex) { + /* todo: dir->i_mutex is necessary */ + au_cpup_attr_timesizes(dir); + if (h_nlink) + vfsub_drop_nlink(dir); + } + return 0; /* success */ + } + + pr_warn("failed removing %pd(%d), ignored\n", wh_dentry, err); + return err; +} + +static void call_rmdir_whtmp(void *args) +{ + int err; + aufs_bindex_t bindex; + struct au_whtmp_rmdir *a = args; + struct super_block *sb; + struct dentry *h_parent; + struct inode *h_dir; + struct au_hinode *hdir; + + /* rmdir by nfsd may cause deadlock with this i_mutex */ + /* inode_lock(a->dir); */ + err = -EROFS; + sb = a->dir->i_sb; + si_read_lock(sb, !AuLock_FLUSH); + if (!au_br_writable(a->br->br_perm)) + goto out; + bindex = au_br_index(sb, a->br->br_id); + if (unlikely(bindex < 0)) + goto out; + + err = -EIO; + ii_write_lock_parent(a->dir); + h_parent = dget_parent(a->wh_dentry); + h_dir = d_inode(h_parent); + hdir = au_hi(a->dir, bindex); + err = vfsub_mnt_want_write(au_br_mnt(a->br)); + if (unlikely(err)) + goto out_mnt; + au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT); + err = au_h_verify(a->wh_dentry, au_opt_udba(sb), h_dir, h_parent, + a->br); + if (!err) + err = au_whtmp_rmdir(a->dir, bindex, a->wh_dentry, &a->whlist); + au_hn_inode_unlock(hdir); + vfsub_mnt_drop_write(au_br_mnt(a->br)); + +out_mnt: + dput(h_parent); + ii_write_unlock(a->dir); +out: + /* inode_unlock(a->dir); */ + au_whtmp_rmdir_free(a); + si_read_unlock(sb); + au_nwt_done(&au_sbi(sb)->si_nowait); + if (unlikely(err)) + AuIOErr("err %d\n", err); +} + +void au_whtmp_kick_rmdir(struct inode *dir, aufs_bindex_t bindex, + struct dentry *wh_dentry, struct au_whtmp_rmdir *args) +{ + int wkq_err; + struct super_block *sb; + + IMustLock(dir); + + /* all post-process will be done in do_rmdir_whtmp(). */ + sb = dir->i_sb; + args->dir = au_igrab(dir); + args->br = au_sbr(sb, bindex); + au_lcnt_inc(&args->br->br_count); + args->wh_dentry = dget(wh_dentry); + wkq_err = au_wkq_nowait(call_rmdir_whtmp, args, sb, /*flags*/0); + if (unlikely(wkq_err)) { + pr_warn("rmdir error %pd (%d), ignored\n", wh_dentry, wkq_err); + au_whtmp_rmdir_free(args); + } +} diff --git a/fs/aufs/whout.h b/fs/aufs/whout.h new file mode 100644 index 00000000000..df85cf5e76c --- /dev/null +++ b/fs/aufs/whout.h @@ -0,0 +1,87 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * whiteout for logical deletion and opaque directory + */ + +#ifndef __AUFS_WHOUT_H__ +#define __AUFS_WHOUT_H__ + +#ifdef __KERNEL__ + +#include "dir.h" + +/* whout.c */ +int au_wh_name_alloc(struct qstr *wh, const struct qstr *name); +int au_wh_test(struct user_namespace *h_userns, struct dentry *h_parent, + struct qstr *wh_name, int try_sio); +int au_diropq_test(struct user_namespace *h_userns, struct dentry *h_dentry); +struct au_branch; +struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct au_branch *br, + struct qstr *prefix); +int au_whtmp_ren(struct dentry *h_dentry, struct au_branch *br); +int au_wh_unlink_dentry(struct inode *h_dir, struct path *h_path, + struct dentry *dentry); +int au_wh_init(struct au_branch *br, struct super_block *sb); + +/* diropq flags */ +#define AuDiropq_CREATE 1 +#define au_ftest_diropq(flags, name) ((flags) & AuDiropq_##name) +#define au_fset_diropq(flags, name) \ + do { (flags) |= AuDiropq_##name; } while (0) +#define au_fclr_diropq(flags, name) \ + do { (flags) &= ~AuDiropq_##name; } while (0) + +struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex, + unsigned int flags); +struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name, + struct au_branch *br); +struct dentry *au_wh_create(struct dentry *dentry, aufs_bindex_t bindex, + struct dentry *h_parent); + +/* real rmdir for the whiteout-ed dir */ +struct au_whtmp_rmdir { + struct inode *dir; + struct au_branch *br; + struct dentry *wh_dentry; + struct au_nhash whlist; +}; + +struct au_whtmp_rmdir *au_whtmp_rmdir_alloc(struct super_block *sb, gfp_t gfp); +void au_whtmp_rmdir_free(struct au_whtmp_rmdir *whtmp); +int au_whtmp_rmdir(struct inode *dir, aufs_bindex_t bindex, + struct dentry *wh_dentry, struct au_nhash *whlist); +void au_whtmp_kick_rmdir(struct inode *dir, aufs_bindex_t bindex, + struct dentry *wh_dentry, struct au_whtmp_rmdir *args); + +/* ---------------------------------------------------------------------- */ + +static inline struct dentry *au_diropq_create(struct dentry *dentry, + aufs_bindex_t bindex) +{ + return au_diropq_sio(dentry, bindex, AuDiropq_CREATE); +} + +static inline int au_diropq_remove(struct dentry *dentry, aufs_bindex_t bindex) +{ + return PTR_ERR(au_diropq_sio(dentry, bindex, !AuDiropq_CREATE)); +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_WHOUT_H__ */ diff --git a/fs/aufs/wkq.c b/fs/aufs/wkq.c new file mode 100644 index 00000000000..4d66bb2dc65 --- /dev/null +++ b/fs/aufs/wkq.c @@ -0,0 +1,372 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * workqueue for asynchronous/super-io operations + * todo: try new credential scheme + */ + +#include +#include "aufs.h" + +/* internal workqueue named AUFS_WKQ_NAME */ + +static struct workqueue_struct *au_wkq; + +struct au_wkinfo { + struct work_struct wk; + struct kobject *kobj; + + unsigned int flags; /* see wkq.h */ + + au_wkq_func_t func; + void *args; + +#ifdef CONFIG_LOCKDEP + int dont_check; + struct held_lock **hlock; +#endif + + struct completion *comp; +}; + +/* ---------------------------------------------------------------------- */ +/* + * Aufs passes some operations to the workqueue such as the internal copyup. + * This scheme looks rather unnatural for LOCKDEP debugging feature, since the + * job run by workqueue depends upon the locks acquired in the other task. + * Delegating a small operation to the workqueue, aufs passes its lockdep + * information too. And the job in the workqueue restores the info in order to + * pretend as if it acquired those locks. This is just to make LOCKDEP work + * correctly and expectedly. + */ + +#ifndef CONFIG_LOCKDEP +AuStubInt0(au_wkq_lockdep_alloc, struct au_wkinfo *wkinfo); +AuStubVoid(au_wkq_lockdep_free, struct au_wkinfo *wkinfo); +AuStubVoid(au_wkq_lockdep_pre, struct au_wkinfo *wkinfo); +AuStubVoid(au_wkq_lockdep_post, struct au_wkinfo *wkinfo); +AuStubVoid(au_wkq_lockdep_init, struct au_wkinfo *wkinfo); +#else +static void au_wkq_lockdep_init(struct au_wkinfo *wkinfo) +{ + wkinfo->hlock = NULL; + wkinfo->dont_check = 0; +} + +/* + * 1: matched + * 0: unmatched + */ +static int au_wkq_lockdep_test(struct lock_class_key *key, const char *name) +{ + static DEFINE_SPINLOCK(spin); + static struct { + char *name; + struct lock_class_key *key; + } a[] = { + { .name = "&sbinfo->si_rwsem" }, + { .name = "&finfo->fi_rwsem" }, + { .name = "&dinfo->di_rwsem" }, + { .name = "&iinfo->ii_rwsem" } + }; + static int set; + int i; + + /* lockless read from 'set.' see below */ + if (set == ARRAY_SIZE(a)) { + for (i = 0; i < ARRAY_SIZE(a); i++) + if (a[i].key == key) + goto match; + goto unmatch; + } + + spin_lock(&spin); + if (set) + for (i = 0; i < ARRAY_SIZE(a); i++) + if (a[i].key == key) { + spin_unlock(&spin); + goto match; + } + for (i = 0; i < ARRAY_SIZE(a); i++) { + if (a[i].key) { + if (unlikely(a[i].key == key)) { /* rare but possible */ + spin_unlock(&spin); + goto match; + } else + continue; + } + if (strstr(a[i].name, name)) { + /* + * the order of these three lines is important for the + * lockless read above. + */ + a[i].key = key; + spin_unlock(&spin); + set++; + /* AuDbg("%d, %s\n", set, name); */ + goto match; + } + } + spin_unlock(&spin); + goto unmatch; + +match: + return 1; +unmatch: + return 0; +} + +static int au_wkq_lockdep_alloc(struct au_wkinfo *wkinfo) +{ + int err, n; + struct task_struct *curr; + struct held_lock **hl, *held_locks, *p; + + err = 0; + curr = current; + wkinfo->dont_check = lockdep_recursing(curr); + if (wkinfo->dont_check) + goto out; + n = curr->lockdep_depth; + if (!n) + goto out; + + err = -ENOMEM; + wkinfo->hlock = kmalloc_array(n + 1, sizeof(*wkinfo->hlock), GFP_NOFS); + if (unlikely(!wkinfo->hlock)) + goto out; + + err = 0; +#if 0 /* left for debugging */ + if (0 && au_debug_test()) + lockdep_print_held_locks(curr); +#endif + held_locks = curr->held_locks; + hl = wkinfo->hlock; + while (n--) { + p = held_locks++; + if (au_wkq_lockdep_test(p->instance->key, p->instance->name)) + *hl++ = p; + } + *hl = NULL; + +out: + return err; +} + +static void au_wkq_lockdep_free(struct au_wkinfo *wkinfo) +{ + au_kfree_try_rcu(wkinfo->hlock); +} + +static void au_wkq_lockdep_pre(struct au_wkinfo *wkinfo) +{ + struct held_lock *p, **hl = wkinfo->hlock; + int subclass; + + if (wkinfo->dont_check) + lockdep_off(); + if (!hl) + return; + while ((p = *hl++)) { /* assignment */ + subclass = lockdep_hlock_class(p)->subclass; + /* AuDbg("%s, %d\n", p->instance->name, subclass); */ + if (p->read) + rwsem_acquire_read(p->instance, subclass, 0, + /*p->acquire_ip*/_RET_IP_); + else + rwsem_acquire(p->instance, subclass, 0, + /*p->acquire_ip*/_RET_IP_); + } +} + +static void au_wkq_lockdep_post(struct au_wkinfo *wkinfo) +{ + struct held_lock *p, **hl = wkinfo->hlock; + + if (wkinfo->dont_check) + lockdep_on(); + if (!hl) + return; + while ((p = *hl++)) /* assignment */ + rwsem_release(p->instance, /*p->acquire_ip*/_RET_IP_); +} +#endif + +static void wkq_func(struct work_struct *wk) +{ + struct au_wkinfo *wkinfo = container_of(wk, struct au_wkinfo, wk); + + AuDebugOn(!uid_eq(current_fsuid(), GLOBAL_ROOT_UID)); + AuDebugOn(rlimit(RLIMIT_FSIZE) != RLIM_INFINITY); + + au_wkq_lockdep_pre(wkinfo); + wkinfo->func(wkinfo->args); + au_wkq_lockdep_post(wkinfo); + if (au_ftest_wkq(wkinfo->flags, WAIT)) + complete(wkinfo->comp); + else { + kobject_put(wkinfo->kobj); + module_put(THIS_MODULE); /* todo: ?? */ + au_kfree_rcu(wkinfo); + } +} + +/* + * Since struct completion is large, try allocating it dynamically. + */ +#define AuWkqCompDeclare(name) struct completion *comp = NULL + +static int au_wkq_comp_alloc(struct au_wkinfo *wkinfo, struct completion **comp) +{ + *comp = kmalloc(sizeof(**comp), GFP_NOFS); + if (*comp) { + init_completion(*comp); + wkinfo->comp = *comp; + return 0; + } + return -ENOMEM; +} + +static void au_wkq_comp_free(struct completion *comp) +{ + au_kfree_rcu(comp); +} + +static void au_wkq_run(struct au_wkinfo *wkinfo) +{ + if (au_ftest_wkq(wkinfo->flags, NEST)) { + if (au_wkq_test()) { + AuWarn1("wkq from wkq, unless silly-rename on NFS," + " due to a dead dir by UDBA," + " or async xino write?\n"); + AuDebugOn(au_ftest_wkq(wkinfo->flags, WAIT)); + } + } else + au_dbg_verify_kthread(); + + if (au_ftest_wkq(wkinfo->flags, WAIT)) { + INIT_WORK_ONSTACK(&wkinfo->wk, wkq_func); + queue_work(au_wkq, &wkinfo->wk); + } else { + INIT_WORK(&wkinfo->wk, wkq_func); + schedule_work(&wkinfo->wk); + } +} + +/* + * Be careful. It is easy to make deadlock happen. + * processA: lock, wkq and wait + * processB: wkq and wait, lock in wkq + * --> deadlock + */ +int au_wkq_do_wait(unsigned int flags, au_wkq_func_t func, void *args) +{ + int err; + AuWkqCompDeclare(comp); + struct au_wkinfo wkinfo = { + .flags = flags, + .func = func, + .args = args + }; + + err = au_wkq_comp_alloc(&wkinfo, &comp); + if (unlikely(err)) + goto out; + err = au_wkq_lockdep_alloc(&wkinfo); + if (unlikely(err)) + goto out_comp; + if (!err) { + au_wkq_run(&wkinfo); + /* no timeout, no interrupt */ + wait_for_completion(wkinfo.comp); + } + au_wkq_lockdep_free(&wkinfo); + +out_comp: + au_wkq_comp_free(comp); +out: + destroy_work_on_stack(&wkinfo.wk); + return err; +} + +/* + * Note: dget/dput() in func for aufs dentries are not supported. It will be a + * problem in a concurrent umounting. + */ +int au_wkq_nowait(au_wkq_func_t func, void *args, struct super_block *sb, + unsigned int flags) +{ + int err; + struct au_wkinfo *wkinfo; + + atomic_inc(&au_sbi(sb)->si_nowait.nw_len); + + /* + * wkq_func() must free this wkinfo. + * it highly depends upon the implementation of workqueue. + */ + err = 0; + wkinfo = kmalloc(sizeof(*wkinfo), GFP_NOFS); + if (wkinfo) { + wkinfo->kobj = &au_sbi(sb)->si_kobj; + wkinfo->flags = flags & ~AuWkq_WAIT; + wkinfo->func = func; + wkinfo->args = args; + wkinfo->comp = NULL; + au_wkq_lockdep_init(wkinfo); + kobject_get(wkinfo->kobj); + __module_get(THIS_MODULE); /* todo: ?? */ + + au_wkq_run(wkinfo); + } else { + err = -ENOMEM; + au_nwt_done(&au_sbi(sb)->si_nowait); + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +void au_nwt_init(struct au_nowait_tasks *nwt) +{ + atomic_set(&nwt->nw_len, 0); + /* smp_mb(); */ /* atomic_set */ + init_waitqueue_head(&nwt->nw_wq); +} + +void au_wkq_fin(void) +{ + destroy_workqueue(au_wkq); +} + +int __init au_wkq_init(void) +{ + int err; + + err = 0; + au_wkq = alloc_workqueue(AUFS_WKQ_NAME, 0, WQ_DFL_ACTIVE); + if (IS_ERR(au_wkq)) + err = PTR_ERR(au_wkq); + else if (!au_wkq) + err = -ENOMEM; + + return err; +} diff --git a/fs/aufs/wkq.h b/fs/aufs/wkq.h new file mode 100644 index 00000000000..5be76b69d8f --- /dev/null +++ b/fs/aufs/wkq.h @@ -0,0 +1,89 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * workqueue for asynchronous/super-io operations + * todo: try new credentials management scheme + */ + +#ifndef __AUFS_WKQ_H__ +#define __AUFS_WKQ_H__ + +#ifdef __KERNEL__ + +#include + +struct super_block; + +/* ---------------------------------------------------------------------- */ + +/* + * in the next operation, wait for the 'nowait' tasks in system-wide workqueue + */ +struct au_nowait_tasks { + atomic_t nw_len; + wait_queue_head_t nw_wq; +}; + +/* ---------------------------------------------------------------------- */ + +typedef void (*au_wkq_func_t)(void *args); + +/* wkq flags */ +#define AuWkq_WAIT 1 +#define AuWkq_NEST (1 << 1) +#define au_ftest_wkq(flags, name) ((flags) & AuWkq_##name) +#define au_fset_wkq(flags, name) \ + do { (flags) |= AuWkq_##name; } while (0) +#define au_fclr_wkq(flags, name) \ + do { (flags) &= ~AuWkq_##name; } while (0) + +/* wkq.c */ +int au_wkq_do_wait(unsigned int flags, au_wkq_func_t func, void *args); +int au_wkq_nowait(au_wkq_func_t func, void *args, struct super_block *sb, + unsigned int flags); +void au_nwt_init(struct au_nowait_tasks *nwt); +int __init au_wkq_init(void); +void au_wkq_fin(void); + +/* ---------------------------------------------------------------------- */ + +static inline int au_wkq_test(void) +{ + return current->flags & PF_WQ_WORKER; +} + +static inline int au_wkq_wait(au_wkq_func_t func, void *args) +{ + return au_wkq_do_wait(AuWkq_WAIT, func, args); +} + +static inline void au_nwt_done(struct au_nowait_tasks *nwt) +{ + if (atomic_dec_and_test(&nwt->nw_len)) + wake_up_all(&nwt->nw_wq); +} + +static inline int au_nwt_flush(struct au_nowait_tasks *nwt) +{ + wait_event(nwt->nw_wq, !atomic_read(&nwt->nw_len)); + return 0; +} + +#endif /* __KERNEL__ */ +#endif /* __AUFS_WKQ_H__ */ diff --git a/fs/aufs/xattr.c b/fs/aufs/xattr.c new file mode 100644 index 00000000000..429580aa95e --- /dev/null +++ b/fs/aufs/xattr.c @@ -0,0 +1,368 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2014-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * handling xattr functions + */ + +#include +#include +#include +#include "aufs.h" + +static int au_xattr_ignore(int err, char *name, unsigned int ignore_flags) +{ + if (!ignore_flags) + goto out; + switch (err) { + case -ENOMEM: + case -EDQUOT: + goto out; + } + + if ((ignore_flags & AuBrAttr_ICEX) == AuBrAttr_ICEX) { + err = 0; + goto out; + } + +#define cmp(brattr, prefix) do { \ + if (!strncmp(name, XATTR_##prefix##_PREFIX, \ + XATTR_##prefix##_PREFIX_LEN)) { \ + if (ignore_flags & AuBrAttr_ICEX_##brattr) \ + err = 0; \ + goto out; \ + } \ + } while (0) + + cmp(SEC, SECURITY); + cmp(SYS, SYSTEM); + cmp(TR, TRUSTED); + cmp(USR, USER); +#undef cmp + + if (ignore_flags & AuBrAttr_ICEX_OTH) + err = 0; + +out: + return err; +} + +static const int au_xattr_out_of_list = AuBrAttr_ICEX_OTH << 1; + +static int au_do_cpup_xattr(struct path *h_dst, struct path *h_src, + char *name, char **buf, unsigned int ignore_flags, + unsigned int verbose) +{ + int err; + ssize_t ssz; + struct inode *h_idst; + struct dentry *h_dst_dentry, *h_src_dentry; + struct user_namespace *h_dst_userns, *h_src_userns; + + h_src_userns = mnt_user_ns(h_src->mnt); + h_src_dentry = h_src->dentry; + ssz = vfs_getxattr_alloc(h_src_userns, h_src_dentry, name, buf, 0, + GFP_NOFS); + err = ssz; + if (unlikely(err <= 0)) { + if (err == -ENODATA + || (err == -EOPNOTSUPP + && ((ignore_flags & au_xattr_out_of_list) + || (au_test_nfs_noacl(d_inode(h_src_dentry)) + && (!strcmp(name, XATTR_NAME_POSIX_ACL_ACCESS) + || !strcmp(name, + XATTR_NAME_POSIX_ACL_DEFAULT)))) + )) + err = 0; + if (err && (verbose || au_debug_test())) + pr_err("%s, err %d\n", name, err); + goto out; + } + + /* unlock it temporary */ + h_dst_userns = mnt_user_ns(h_dst->mnt); + h_dst_dentry = h_dst->dentry; + h_idst = d_inode(h_dst_dentry); + inode_unlock(h_idst); + err = vfsub_setxattr(h_dst_userns, h_dst_dentry, name, *buf, ssz, + /*flags*/0); + inode_lock_nested(h_idst, AuLsc_I_CHILD2); + if (unlikely(err)) { + if (verbose || au_debug_test()) + pr_err("%s, err %d\n", name, err); + err = au_xattr_ignore(err, name, ignore_flags); + } + +out: + return err; +} + +int au_cpup_xattr(struct path *h_dst, struct path *h_src, int ignore_flags, + unsigned int verbose) +{ + int err, unlocked, acl_access, acl_default; + ssize_t ssz; + struct dentry *h_dst_dentry, *h_src_dentry; + struct inode *h_isrc, *h_idst; + char *value, *p, *o, *e; + + /* try stopping to update the source inode while we are referencing */ + /* there should not be the parent-child relationship between them */ + h_dst_dentry = h_dst->dentry; + h_idst = d_inode(h_dst_dentry); + h_src_dentry = h_src->dentry; + h_isrc = d_inode(h_src_dentry); + inode_unlock(h_idst); + inode_lock_shared_nested(h_isrc, AuLsc_I_CHILD); + inode_lock_nested(h_idst, AuLsc_I_CHILD2); + unlocked = 0; + + /* some filesystems don't list POSIX ACL, for example tmpfs */ + ssz = vfs_listxattr(h_src_dentry, NULL, 0); + err = ssz; + if (unlikely(err < 0)) { + AuTraceErr(err); + if (err == -ENODATA + || err == -EOPNOTSUPP) + err = 0; /* ignore */ + goto out; + } + + err = 0; + p = NULL; + o = NULL; + if (ssz) { + err = -ENOMEM; + p = kmalloc(ssz, GFP_NOFS); + o = p; + if (unlikely(!p)) + goto out; + err = vfs_listxattr(h_src_dentry, p, ssz); + } + inode_unlock_shared(h_isrc); + unlocked = 1; + AuDbg("err %d, ssz %zd\n", err, ssz); + if (unlikely(err < 0)) + goto out_free; + + err = 0; + e = p + ssz; + value = NULL; + acl_access = 0; + acl_default = 0; + while (!err && p < e) { + acl_access |= !strncmp(p, XATTR_NAME_POSIX_ACL_ACCESS, + sizeof(XATTR_NAME_POSIX_ACL_ACCESS) - 1); + acl_default |= !strncmp(p, XATTR_NAME_POSIX_ACL_DEFAULT, + sizeof(XATTR_NAME_POSIX_ACL_DEFAULT) + - 1); + err = au_do_cpup_xattr(h_dst, h_src, p, &value, ignore_flags, + verbose); + p += strlen(p) + 1; + } + AuTraceErr(err); + ignore_flags |= au_xattr_out_of_list; + if (!err && !acl_access) { + err = au_do_cpup_xattr(h_dst, h_src, + XATTR_NAME_POSIX_ACL_ACCESS, &value, + ignore_flags, verbose); + AuTraceErr(err); + } + if (!err && !acl_default) { + err = au_do_cpup_xattr(h_dst, h_src, + XATTR_NAME_POSIX_ACL_DEFAULT, &value, + ignore_flags, verbose); + AuTraceErr(err); + } + + au_kfree_try_rcu(value); + +out_free: + au_kfree_try_rcu(o); +out: + if (!unlocked) + inode_unlock_shared(h_isrc); + AuTraceErr(err); + return err; +} + +/* ---------------------------------------------------------------------- */ + +static int au_smack_reentering(struct super_block *sb) +{ +#if IS_ENABLED(CONFIG_SECURITY_SMACK) || IS_ENABLED(CONFIG_SECURITY_SELINUX) + /* + * as a part of lookup, smack_d_instantiate() is called, and it calls + * i_op->getxattr(). ouch. + */ + return si_pid_test(sb); +#else + return 0; +#endif +} + +enum { + AU_XATTR_LIST, + AU_XATTR_GET +}; + +struct au_lgxattr { + int type; + union { + struct { + char *list; + size_t size; + } list; + struct { + const char *name; + void *value; + size_t size; + } get; + } u; +}; + +static ssize_t au_lgxattr(struct dentry *dentry, struct inode *inode, + struct au_lgxattr *arg) +{ + ssize_t err; + int reenter; + struct path h_path; + struct super_block *sb; + + sb = dentry->d_sb; + reenter = au_smack_reentering(sb); + if (!reenter) { + err = si_read_lock(sb, AuLock_FLUSH | AuLock_NOPLM); + if (unlikely(err)) + goto out; + } + err = au_h_path_getattr(dentry, inode, /*force*/1, &h_path, reenter); + if (unlikely(err)) + goto out_si; + if (unlikely(!h_path.dentry)) + /* illegally overlapped or something */ + goto out_di; /* pretending success */ + + /* always topmost entry only */ + switch (arg->type) { + case AU_XATTR_LIST: + err = vfs_listxattr(h_path.dentry, + arg->u.list.list, arg->u.list.size); + break; + case AU_XATTR_GET: + AuDebugOn(d_is_negative(h_path.dentry)); + err = vfs_getxattr(mnt_user_ns(h_path.mnt), h_path.dentry, + arg->u.get.name, arg->u.get.value, + arg->u.get.size); + break; + } + +out_di: + if (!reenter) + di_read_unlock(dentry, AuLock_IR); +out_si: + if (!reenter) + si_read_unlock(sb); +out: + AuTraceErr(err); + return err; +} + +ssize_t aufs_listxattr(struct dentry *dentry, char *list, size_t size) +{ + struct au_lgxattr arg = { + .type = AU_XATTR_LIST, + .u.list = { + .list = list, + .size = size + }, + }; + + return au_lgxattr(dentry, /*inode*/NULL, &arg); +} + +static ssize_t au_getxattr(struct dentry *dentry, struct inode *inode, + const char *name, void *value, size_t size) +{ + struct au_lgxattr arg = { + .type = AU_XATTR_GET, + .u.get = { + .name = name, + .value = value, + .size = size + }, + }; + + return au_lgxattr(dentry, inode, &arg); +} + +static int au_setxattr(struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, + int flags) +{ + struct au_sxattr arg = { + .type = AU_XATTR_SET, + .u.set = { + .name = name, + .value = value, + .size = size, + .flags = flags + }, + }; + + return au_sxattr(dentry, inode, &arg); +} + +/* ---------------------------------------------------------------------- */ + +static int au_xattr_get(const struct xattr_handler *handler, + struct dentry *dentry, struct inode *inode, + const char *name, void *buffer, size_t size) +{ + return au_getxattr(dentry, inode, name, buffer, size); +} + +static int au_xattr_set(const struct xattr_handler *handler, + struct user_namespace *userns, + struct dentry *dentry, struct inode *inode, + const char *name, const void *value, size_t size, + int flags) +{ + return au_setxattr(dentry, inode, name, value, size, flags); +} + +static const struct xattr_handler au_xattr_handler = { + .name = "", + .prefix = "", + .get = au_xattr_get, + .set = au_xattr_set +}; + +static const struct xattr_handler *au_xattr_handlers[] = { +#ifdef CONFIG_FS_POSIX_ACL + &posix_acl_access_xattr_handler, + &posix_acl_default_xattr_handler, +#endif + &au_xattr_handler, /* must be last */ + NULL +}; + +void au_xattr_init(struct super_block *sb) +{ + sb->s_xattr = au_xattr_handlers; +} diff --git a/fs/aufs/xino.c b/fs/aufs/xino.c new file mode 100644 index 00000000000..a9ce74d1ad0 --- /dev/null +++ b/fs/aufs/xino.c @@ -0,0 +1,1925 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +/* + * external inode number translation table and bitmap + * + * things to consider + * - the lifetime + * + au_xino object + * + XINO files (xino, xib, xigen) + * + dynamic debugfs entries (xiN) + * + static debugfs entries (xib, xigen) + * + static sysfs entry (xi_path) + * - several entry points to handle them. + * + mount(2) without xino option (default) + * + mount(2) with xino option + * + mount(2) with noxino option + * + umount(2) + * + remount with add/del branches + * + remount with xino/noxino options + */ + +#include +#include +#include "aufs.h" + +static aufs_bindex_t sbr_find_shared(struct super_block *sb, aufs_bindex_t btop, + aufs_bindex_t bbot, + struct super_block *h_sb) +{ + /* todo: try binary-search if the branches are many */ + for (; btop <= bbot; btop++) + if (h_sb == au_sbr_sb(sb, btop)) + return btop; + return -1; +} + +/* + * find another branch who is on the same filesystem of the specified + * branch{@btgt}. search until @bbot. + */ +static aufs_bindex_t is_sb_shared(struct super_block *sb, aufs_bindex_t btgt, + aufs_bindex_t bbot) +{ + aufs_bindex_t bindex; + struct super_block *tgt_sb; + + tgt_sb = au_sbr_sb(sb, btgt); + bindex = sbr_find_shared(sb, /*btop*/0, btgt - 1, tgt_sb); + if (bindex < 0) + bindex = sbr_find_shared(sb, btgt + 1, bbot, tgt_sb); + + return bindex; +} + +/* ---------------------------------------------------------------------- */ + +/* + * stop unnecessary notify events at creating xino files + */ + +aufs_bindex_t au_xi_root(struct super_block *sb, struct dentry *dentry) +{ + aufs_bindex_t bfound, bindex, bbot; + struct dentry *parent; + struct au_branch *br; + + bfound = -1; + parent = dentry->d_parent; /* safe d_parent access */ + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_dentry(br) == parent) { + bfound = bindex; + break; + } + } + + AuDbg("bfound b%d\n", bfound); + return bfound; +} + +struct au_xino_lock_dir { + struct au_hinode *hdir; + struct dentry *parent; + struct inode *dir; +}; + +static struct dentry *au_dget_parent_lock(struct dentry *dentry, + unsigned int lsc) +{ + struct dentry *parent; + struct inode *dir; + + parent = dget_parent(dentry); + dir = d_inode(parent); + inode_lock_nested(dir, lsc); +#if 0 /* it should not happen */ + spin_lock(&dentry->d_lock); + if (unlikely(dentry->d_parent != parent)) { + spin_unlock(&dentry->d_lock); + inode_unlock(dir); + dput(parent); + parent = NULL; + goto out; + } + spin_unlock(&dentry->d_lock); + +out: +#endif + return parent; +} + +static void au_xino_lock_dir(struct super_block *sb, struct path *xipath, + struct au_xino_lock_dir *ldir) +{ + aufs_bindex_t bindex; + + ldir->hdir = NULL; + bindex = au_xi_root(sb, xipath->dentry); + if (bindex >= 0) { + /* rw branch root */ + ldir->hdir = au_hi(d_inode(sb->s_root), bindex); + au_hn_inode_lock_nested(ldir->hdir, AuLsc_I_PARENT); + } else { + /* other */ + ldir->parent = au_dget_parent_lock(xipath->dentry, + AuLsc_I_PARENT); + ldir->dir = d_inode(ldir->parent); + } +} + +static void au_xino_unlock_dir(struct au_xino_lock_dir *ldir) +{ + if (ldir->hdir) + au_hn_inode_unlock(ldir->hdir); + else { + inode_unlock(ldir->dir); + dput(ldir->parent); + } +} + +/* ---------------------------------------------------------------------- */ + +/* + * create and set a new xino file + */ +struct file *au_xino_create(struct super_block *sb, char *fpath, int silent, + int wbrtop) +{ + struct file *file; + struct dentry *h_parent, *d; + struct inode *h_dir, *inode; + int err; + static DEFINE_MUTEX(mtx); + + /* + * at mount-time, and the xino file is the default path, + * hnotify is disabled so we have no notify events to ignore. + * when a user specified the xino, we cannot get au_hdir to be ignored. + */ + if (!wbrtop) + mutex_lock(&mtx); + file = vfsub_filp_open(fpath, O_RDWR | O_CREAT | O_EXCL | O_LARGEFILE + /* | __FMODE_NONOTIFY */, + 0666); + if (IS_ERR(file)) { + if (!wbrtop) + mutex_unlock(&mtx); + if (!silent) + pr_err("open %s(%ld)\n", fpath, PTR_ERR(file)); + return file; + } + + /* keep file count */ + err = 0; + d = file->f_path.dentry; + h_parent = au_dget_parent_lock(d, AuLsc_I_PARENT); + if (!wbrtop) + mutex_unlock(&mtx); + /* mnt_want_write() is unnecessary here */ + h_dir = d_inode(h_parent); + inode = file_inode(file); + /* no delegation since it is just created */ + if (inode->i_nlink) + err = vfsub_unlink(h_dir, &file->f_path, /*delegated*/NULL, + /*force*/0); + inode_unlock(h_dir); + dput(h_parent); + if (unlikely(err)) { + if (!silent) + pr_err("unlink %s(%d)\n", fpath, err); + goto out; + } + + err = -EINVAL; + if (unlikely(sb == d->d_sb)) { + if (!silent) + pr_err("%s must be outside\n", fpath); + goto out; + } + if (unlikely(au_test_fs_bad_xino(d->d_sb))) { + if (!silent) + pr_err("xino doesn't support %s(%s)\n", + fpath, au_sbtype(d->d_sb)); + goto out; + } + return file; /* success */ + +out: + fput(file); + file = ERR_PTR(err); + return file; +} + +/* + * create a new xinofile at the same place/path as @base. + */ +struct file *au_xino_create2(struct super_block *sb, struct path *base, + struct file *copy_src) +{ + struct file *file; + struct dentry *dentry, *parent; + struct inode *dir, *delegated; + struct qstr *name; + struct path path; + int err, do_unlock; + struct au_xino_lock_dir ldir; + + do_unlock = 1; + au_xino_lock_dir(sb, base, &ldir); + dentry = base->dentry; + parent = dentry->d_parent; /* dir inode is locked */ + dir = d_inode(parent); + IMustLock(dir); + + name = &dentry->d_name; + path.dentry = vfsub_lookup_one_len(name->name, parent, name->len); + if (IS_ERR(path.dentry)) { + file = (void *)path.dentry; + pr_err("%pd lookup err %ld\n", dentry, PTR_ERR(path.dentry)); + goto out; + } + + /* no need to mnt_want_write() since we call dentry_open() later */ + err = vfs_create(mnt_user_ns(base->mnt), dir, path.dentry, 0666, NULL); + if (unlikely(err)) { + file = ERR_PTR(err); + pr_err("%pd create err %d\n", dentry, err); + goto out_dput; + } + + path.mnt = base->mnt; + file = vfsub_dentry_open(&path, + O_RDWR | O_CREAT | O_EXCL | O_LARGEFILE + /* | __FMODE_NONOTIFY */); + if (IS_ERR(file)) { + pr_err("%pd open err %ld\n", dentry, PTR_ERR(file)); + goto out_dput; + } + + delegated = NULL; + err = vfsub_unlink(dir, &file->f_path, &delegated, /*force*/0); + au_xino_unlock_dir(&ldir); + do_unlock = 0; + if (unlikely(err == -EWOULDBLOCK)) { + pr_warn("cannot retry for NFSv4 delegation" + " for an internal unlink\n"); + iput(delegated); + } + if (unlikely(err)) { + pr_err("%pd unlink err %d\n", dentry, err); + goto out_fput; + } + + if (copy_src) { + /* no one can touch copy_src xino */ + err = au_copy_file(file, copy_src, vfsub_f_size_read(copy_src)); + if (unlikely(err)) { + pr_err("%pd copy err %d\n", dentry, err); + goto out_fput; + } + } + goto out_dput; /* success */ + +out_fput: + fput(file); + file = ERR_PTR(err); +out_dput: + dput(path.dentry); +out: + if (do_unlock) + au_xino_unlock_dir(&ldir); + return file; +} + +struct file *au_xino_file1(struct au_xino *xi) +{ + struct file *file; + unsigned int u, nfile; + + file = NULL; + nfile = xi->xi_nfile; + for (u = 0; u < nfile; u++) { + file = xi->xi_file[u]; + if (file) + break; + } + + return file; +} + +static int au_xino_file_set(struct au_xino *xi, int idx, struct file *file) +{ + int err; + struct file *f; + void *p; + + if (file) + get_file(file); + + err = 0; + f = NULL; + if (idx < xi->xi_nfile) { + f = xi->xi_file[idx]; + if (f) + fput(f); + } else { + p = au_kzrealloc(xi->xi_file, + sizeof(*xi->xi_file) * xi->xi_nfile, + sizeof(*xi->xi_file) * (idx + 1), + GFP_NOFS, /*may_shrink*/0); + if (p) { + MtxMustLock(&xi->xi_mtx); + xi->xi_file = p; + xi->xi_nfile = idx + 1; + } else { + err = -ENOMEM; + if (file) + fput(file); + goto out; + } + } + xi->xi_file[idx] = file; + +out: + return err; +} + +/* + * if @xinew->xi is not set, then create new xigen file. + */ +struct file *au_xi_new(struct super_block *sb, struct au_xi_new *xinew) +{ + struct file *file; + int err; + + SiMustAnyLock(sb); + + file = au_xino_create2(sb, xinew->base, xinew->copy_src); + if (IS_ERR(file)) { + err = PTR_ERR(file); + pr_err("%s[%d], err %d\n", + xinew->xi ? "xino" : "xigen", + xinew->idx, err); + goto out; + } + + if (xinew->xi) + err = au_xino_file_set(xinew->xi, xinew->idx, file); + else { + BUG(); + /* todo: make xigen file an array */ + /* err = au_xigen_file_set(sb, xinew->idx, file); */ + } + fput(file); + if (unlikely(err)) + file = ERR_PTR(err); + +out: + return file; +} + +/* ---------------------------------------------------------------------- */ + +/* + * truncate xino files + */ +static int au_xino_do_trunc(struct super_block *sb, aufs_bindex_t bindex, + int idx, struct kstatfs *st) +{ + int err; + blkcnt_t blocks; + struct file *file, *new_xino; + struct au_xi_new xinew = { + .idx = idx + }; + + err = 0; + xinew.xi = au_sbr(sb, bindex)->br_xino; + file = au_xino_file(xinew.xi, idx); + if (!file) + goto out; + + xinew.base = &file->f_path; + err = vfs_statfs(xinew.base, st); + if (unlikely(err)) { + AuErr1("statfs err %d, ignored\n", err); + err = 0; + goto out; + } + + blocks = file_inode(file)->i_blocks; + pr_info("begin truncating xino(b%d-%d), ib%llu, %llu/%llu free blks\n", + bindex, idx, (u64)blocks, st->f_bfree, st->f_blocks); + + xinew.copy_src = file; + new_xino = au_xi_new(sb, &xinew); + if (IS_ERR(new_xino)) { + err = PTR_ERR(new_xino); + pr_err("xino(b%d-%d), err %d, ignored\n", bindex, idx, err); + goto out; + } + + err = vfs_statfs(&new_xino->f_path, st); + if (!err) + pr_info("end truncating xino(b%d-%d), ib%llu, %llu/%llu free blks\n", + bindex, idx, (u64)file_inode(new_xino)->i_blocks, + st->f_bfree, st->f_blocks); + else { + AuErr1("statfs err %d, ignored\n", err); + err = 0; + } + +out: + return err; +} + +int au_xino_trunc(struct super_block *sb, aufs_bindex_t bindex, int idx_begin) +{ + int err, i; + unsigned long jiffy; + aufs_bindex_t bbot; + struct kstatfs *st; + struct au_branch *br; + struct au_xino *xi; + + err = -ENOMEM; + st = kmalloc(sizeof(*st), GFP_NOFS); + if (unlikely(!st)) + goto out; + + err = -EINVAL; + bbot = au_sbbot(sb); + if (unlikely(bindex < 0 || bbot < bindex)) + goto out_st; + + err = 0; + jiffy = jiffies; + br = au_sbr(sb, bindex); + xi = br->br_xino; + for (i = idx_begin; !err && i < xi->xi_nfile; i++) + err = au_xino_do_trunc(sb, bindex, i, st); + if (!err) + au_sbi(sb)->si_xino_jiffy = jiffy; + +out_st: + au_kfree_rcu(st); +out: + return err; +} + +struct xino_do_trunc_args { + struct super_block *sb; + struct au_branch *br; + int idx; +}; + +static void xino_do_trunc(void *_args) +{ + struct xino_do_trunc_args *args = _args; + struct super_block *sb; + struct au_branch *br; + struct inode *dir; + int err, idx; + aufs_bindex_t bindex; + + err = 0; + sb = args->sb; + dir = d_inode(sb->s_root); + br = args->br; + idx = args->idx; + + si_noflush_write_lock(sb); + ii_read_lock_parent(dir); + bindex = au_br_index(sb, br->br_id); + err = au_xino_trunc(sb, bindex, idx); + ii_read_unlock(dir); + if (unlikely(err)) + pr_warn("err b%d, (%d)\n", bindex, err); + atomic_dec(&br->br_xino->xi_truncating); + au_lcnt_dec(&br->br_count); + si_write_unlock(sb); + au_nwt_done(&au_sbi(sb)->si_nowait); + au_kfree_rcu(args); +} + +/* + * returns the index in the xi_file array whose corresponding file is necessary + * to truncate, or -1 which means no need to truncate. + */ +static int xino_trunc_test(struct super_block *sb, struct au_branch *br) +{ + int err; + unsigned int u; + struct kstatfs st; + struct au_sbinfo *sbinfo; + struct au_xino *xi; + struct file *file; + + /* todo: si_xino_expire and the ratio should be customizable */ + sbinfo = au_sbi(sb); + if (time_before(jiffies, + sbinfo->si_xino_jiffy + sbinfo->si_xino_expire)) + return -1; + + /* truncation border */ + xi = br->br_xino; + for (u = 0; u < xi->xi_nfile; u++) { + file = au_xino_file(xi, u); + if (!file) + continue; + + err = vfs_statfs(&file->f_path, &st); + if (unlikely(err)) { + AuErr1("statfs err %d, ignored\n", err); + return -1; + } + if (div64_u64(st.f_bfree * 100, st.f_blocks) + >= AUFS_XINO_DEF_TRUNC) + return u; + } + + return -1; +} + +static void xino_try_trunc(struct super_block *sb, struct au_branch *br) +{ + int idx; + struct xino_do_trunc_args *args; + int wkq_err; + + idx = xino_trunc_test(sb, br); + if (idx < 0) + return; + + if (atomic_inc_return(&br->br_xino->xi_truncating) > 1) + goto out; + + /* lock and kfree() will be called in trunc_xino() */ + args = kmalloc(sizeof(*args), GFP_NOFS); + if (unlikely(!args)) { + AuErr1("no memory\n"); + goto out; + } + + au_lcnt_inc(&br->br_count); + args->sb = sb; + args->br = br; + args->idx = idx; + wkq_err = au_wkq_nowait(xino_do_trunc, args, sb, /*flags*/0); + if (!wkq_err) + return; /* success */ + + pr_err("wkq %d\n", wkq_err); + au_lcnt_dec(&br->br_count); + au_kfree_rcu(args); + +out: + atomic_dec(&br->br_xino->xi_truncating); +} + +/* ---------------------------------------------------------------------- */ + +struct au_xi_calc { + int idx; + loff_t pos; +}; + +static void au_xi_calc(struct super_block *sb, ino_t h_ino, + struct au_xi_calc *calc) +{ + loff_t maxent; + + maxent = au_xi_maxent(sb); + calc->idx = div64_u64_rem(h_ino, maxent, &calc->pos); + calc->pos *= sizeof(ino_t); +} + +static int au_xino_do_new_async(struct super_block *sb, struct au_branch *br, + struct au_xi_calc *calc) +{ + int err; + struct file *file; + struct au_xino *xi = br->br_xino; + struct au_xi_new xinew = { + .xi = xi + }; + + SiMustAnyLock(sb); + + err = 0; + if (!xi) + goto out; + + mutex_lock(&xi->xi_mtx); + file = au_xino_file(xi, calc->idx); + if (file) + goto out_mtx; + + file = au_xino_file(xi, /*idx*/-1); + AuDebugOn(!file); + xinew.idx = calc->idx; + xinew.base = &file->f_path; + /* xinew.copy_src = NULL; */ + file = au_xi_new(sb, &xinew); + if (IS_ERR(file)) + err = PTR_ERR(file); + +out_mtx: + mutex_unlock(&xi->xi_mtx); +out: + return err; +} + +struct au_xino_do_new_async_args { + struct super_block *sb; + struct au_branch *br; + struct au_xi_calc calc; + ino_t ino; +}; + +struct au_xi_writing { + struct hlist_bl_node node; + ino_t h_ino, ino; +}; + +static int au_xino_do_write(struct file *file, struct au_xi_calc *calc, + ino_t ino); + +static void au_xino_call_do_new_async(void *args) +{ + struct au_xino_do_new_async_args *a = args; + struct au_branch *br; + struct super_block *sb; + struct au_sbinfo *sbi; + struct inode *root; + struct file *file; + struct au_xi_writing *del, *p; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + int err; + + br = a->br; + sb = a->sb; + sbi = au_sbi(sb); + si_noflush_read_lock(sb); + root = d_inode(sb->s_root); + ii_read_lock_child(root); + err = au_xino_do_new_async(sb, br, &a->calc); + if (unlikely(err)) { + AuIOErr("err %d\n", err); + goto out; + } + + file = au_xino_file(br->br_xino, a->calc.idx); + AuDebugOn(!file); + err = au_xino_do_write(file, &a->calc, a->ino); + if (unlikely(err)) { + AuIOErr("err %d\n", err); + goto out; + } + + del = NULL; + hbl = &br->br_xino->xi_writing; + hlist_bl_lock(hbl); + au_hbl_for_each(pos, hbl) { + p = container_of(pos, struct au_xi_writing, node); + if (p->ino == a->ino) { + del = p; + hlist_bl_del(&p->node); + break; + } + } + hlist_bl_unlock(hbl); + au_kfree_rcu(del); + +out: + au_lcnt_dec(&br->br_count); + ii_read_unlock(root); + si_read_unlock(sb); + au_nwt_done(&sbi->si_nowait); + au_kfree_rcu(a); +} + +/* + * create a new xino file asynchronously + */ +static int au_xino_new_async(struct super_block *sb, struct au_branch *br, + struct au_xi_calc *calc, ino_t ino) +{ + int err; + struct au_xino_do_new_async_args *arg; + + err = -ENOMEM; + arg = kmalloc(sizeof(*arg), GFP_NOFS); + if (unlikely(!arg)) + goto out; + + arg->sb = sb; + arg->br = br; + arg->calc = *calc; + arg->ino = ino; + au_lcnt_inc(&br->br_count); + err = au_wkq_nowait(au_xino_call_do_new_async, arg, sb, AuWkq_NEST); + if (unlikely(err)) { + pr_err("wkq %d\n", err); + au_lcnt_dec(&br->br_count); + au_kfree_rcu(arg); + } + +out: + return err; +} + +/* + * read @ino from xinofile for the specified branch{@sb, @bindex} + * at the position of @h_ino. + */ +int au_xino_read(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + ino_t *ino) +{ + int err; + ssize_t sz; + struct au_xi_calc calc; + struct au_sbinfo *sbinfo; + struct file *file; + struct au_xino *xi; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos; + struct au_xi_writing *p; + + *ino = 0; + if (!au_opt_test(au_mntflags(sb), XINO)) + return 0; /* no xino */ + + err = 0; + au_xi_calc(sb, h_ino, &calc); + xi = au_sbr(sb, bindex)->br_xino; + file = au_xino_file(xi, calc.idx); + if (!file) { + hbl = &xi->xi_writing; + hlist_bl_lock(hbl); + au_hbl_for_each(pos, hbl) { + p = container_of(pos, struct au_xi_writing, node); + if (p->h_ino == h_ino) { + AuDbg("hi%llu, i%llu, found\n", + (u64)p->h_ino, (u64)p->ino); + *ino = p->ino; + break; + } + } + hlist_bl_unlock(hbl); + return 0; + } else if (vfsub_f_size_read(file) < calc.pos + sizeof(*ino)) + return 0; /* no xino */ + + sbinfo = au_sbi(sb); + sz = xino_fread(file, ino, sizeof(*ino), &calc.pos); + if (sz == sizeof(*ino)) + return 0; /* success */ + + err = sz; + if (unlikely(sz >= 0)) { + err = -EIO; + AuIOErr("xino read error (%zd)\n", sz); + } + return err; +} + +static int au_xino_do_write(struct file *file, struct au_xi_calc *calc, + ino_t ino) +{ + ssize_t sz; + + sz = xino_fwrite(file, &ino, sizeof(ino), &calc->pos); + if (sz == sizeof(ino)) + return 0; /* success */ + + AuIOErr("write failed (%zd)\n", sz); + return -EIO; +} + +/* + * write @ino to the xinofile for the specified branch{@sb, @bindex} + * at the position of @h_ino. + * even if @ino is zero, it is written to the xinofile and means no entry. + * if the size of the xino file on a specific filesystem exceeds the watermark, + * try truncating it. + */ +int au_xino_write(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + ino_t ino) +{ + int err; + unsigned int mnt_flags; + struct au_xi_calc calc; + struct file *file; + struct au_branch *br; + struct au_xino *xi; + struct au_xi_writing *p; + + SiMustAnyLock(sb); + + mnt_flags = au_mntflags(sb); + if (!au_opt_test(mnt_flags, XINO)) + return 0; + + au_xi_calc(sb, h_ino, &calc); + br = au_sbr(sb, bindex); + xi = br->br_xino; + file = au_xino_file(xi, calc.idx); + if (!file) { + /* store the inum pair into the list */ + p = kmalloc(sizeof(*p), GFP_NOFS | __GFP_NOFAIL); + p->h_ino = h_ino; + p->ino = ino; + au_hbl_add(&p->node, &xi->xi_writing); + + /* create and write a new xino file asynchronously */ + err = au_xino_new_async(sb, br, &calc, ino); + if (!err) + return 0; /* success */ + goto out; + } + + err = au_xino_do_write(file, &calc, ino); + if (!err) { + br = au_sbr(sb, bindex); + if (au_opt_test(mnt_flags, TRUNC_XINO) + && au_test_fs_trunc_xino(au_br_sb(br))) + xino_try_trunc(sb, br); + return 0; /* success */ + } + +out: + AuIOErr("write failed (%d)\n", err); + return -EIO; +} + +static ssize_t xino_fread_wkq(struct file *file, void *buf, size_t size, + loff_t *pos); + +/* todo: unnecessary to support mmap_sem since kernel-space? */ +ssize_t xino_fread(struct file *file, void *kbuf, size_t size, loff_t *pos) +{ + ssize_t err; + int i; + const int prevent_endless = 10; + + i = 0; + do { + err = vfsub_read_k(file, kbuf, size, pos); + if (err == -EINTR + && !au_wkq_test() + && fatal_signal_pending(current)) { + err = xino_fread_wkq(file, kbuf, size, pos); + BUG_ON(err == -EINTR); + } + } while (i++ < prevent_endless + && (err == -EAGAIN || err == -EINTR)); + +#if 0 /* reserved for future use */ + if (err > 0) + fsnotify_access(file->f_path.dentry); +#endif + + return err; +} + +struct xino_fread_args { + ssize_t *errp; + struct file *file; + void *buf; + size_t size; + loff_t *pos; +}; + +static void call_xino_fread(void *args) +{ + struct xino_fread_args *a = args; + *a->errp = xino_fread(a->file, a->buf, a->size, a->pos); +} + +static ssize_t xino_fread_wkq(struct file *file, void *buf, size_t size, + loff_t *pos) +{ + ssize_t err; + int wkq_err; + struct xino_fread_args args = { + .errp = &err, + .file = file, + .buf = buf, + .size = size, + .pos = pos + }; + + wkq_err = au_wkq_wait(call_xino_fread, &args); + if (unlikely(wkq_err)) + err = wkq_err; + + return err; +} + +static ssize_t xino_fwrite_wkq(struct file *file, void *buf, size_t size, + loff_t *pos); + +static ssize_t do_xino_fwrite(struct file *file, void *kbuf, size_t size, + loff_t *pos) +{ + ssize_t err; + int i; + const int prevent_endless = 10; + + i = 0; + do { + err = vfsub_write_k(file, kbuf, size, pos); + if (err == -EINTR + && !au_wkq_test() + && fatal_signal_pending(current)) { + err = xino_fwrite_wkq(file, kbuf, size, pos); + BUG_ON(err == -EINTR); + } + } while (i++ < prevent_endless + && (err == -EAGAIN || err == -EINTR)); + +#if 0 /* reserved for future use */ + if (err > 0) + fsnotify_modify(file->f_path.dentry); +#endif + + return err; +} + +struct do_xino_fwrite_args { + ssize_t *errp; + struct file *file; + void *buf; + size_t size; + loff_t *pos; +}; + +static void call_do_xino_fwrite(void *args) +{ + struct do_xino_fwrite_args *a = args; + *a->errp = do_xino_fwrite(a->file, a->buf, a->size, a->pos); +} + +static ssize_t xino_fwrite_wkq(struct file *file, void *buf, size_t size, + loff_t *pos) +{ + ssize_t err; + int wkq_err; + struct do_xino_fwrite_args args = { + .errp = &err, + .file = file, + .buf = buf, + .size = size, + .pos = pos + }; + + /* + * it breaks RLIMIT_FSIZE and normal user's limit, + * users should care about quota and real 'filesystem full.' + */ + wkq_err = au_wkq_wait(call_do_xino_fwrite, &args); + if (unlikely(wkq_err)) + err = wkq_err; + + return err; +} + +ssize_t xino_fwrite(struct file *file, void *buf, size_t size, loff_t *pos) +{ + ssize_t err; + + if (rlimit(RLIMIT_FSIZE) == RLIM_INFINITY) { + lockdep_off(); + err = do_xino_fwrite(file, buf, size, pos); + lockdep_on(); + } else { + lockdep_off(); + err = xino_fwrite_wkq(file, buf, size, pos); + lockdep_on(); + } + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * inode number bitmap + */ +static const int page_bits = (int)PAGE_SIZE * BITS_PER_BYTE; +static ino_t xib_calc_ino(unsigned long pindex, int bit) +{ + ino_t ino; + + AuDebugOn(bit < 0 || page_bits <= bit); + ino = AUFS_FIRST_INO + pindex * page_bits + bit; + return ino; +} + +static void xib_calc_bit(ino_t ino, unsigned long *pindex, int *bit) +{ + AuDebugOn(ino < AUFS_FIRST_INO); + ino -= AUFS_FIRST_INO; + *pindex = ino / page_bits; + *bit = ino % page_bits; +} + +static int xib_pindex(struct super_block *sb, unsigned long pindex) +{ + int err; + loff_t pos; + ssize_t sz; + struct au_sbinfo *sbinfo; + struct file *xib; + unsigned long *p; + + sbinfo = au_sbi(sb); + MtxMustLock(&sbinfo->si_xib_mtx); + AuDebugOn(pindex > ULONG_MAX / PAGE_SIZE + || !au_opt_test(sbinfo->si_mntflags, XINO)); + + if (pindex == sbinfo->si_xib_last_pindex) + return 0; + + xib = sbinfo->si_xib; + p = sbinfo->si_xib_buf; + pos = sbinfo->si_xib_last_pindex; + pos *= PAGE_SIZE; + sz = xino_fwrite(xib, p, PAGE_SIZE, &pos); + if (unlikely(sz != PAGE_SIZE)) + goto out; + + pos = pindex; + pos *= PAGE_SIZE; + if (vfsub_f_size_read(xib) >= pos + PAGE_SIZE) + sz = xino_fread(xib, p, PAGE_SIZE, &pos); + else { + memset(p, 0, PAGE_SIZE); + sz = xino_fwrite(xib, p, PAGE_SIZE, &pos); + } + if (sz == PAGE_SIZE) { + sbinfo->si_xib_last_pindex = pindex; + return 0; /* success */ + } + +out: + AuIOErr1("write failed (%zd)\n", sz); + err = sz; + if (sz >= 0) + err = -EIO; + return err; +} + +static void au_xib_clear_bit(struct inode *inode) +{ + int err, bit; + unsigned long pindex; + struct super_block *sb; + struct au_sbinfo *sbinfo; + + AuDebugOn(inode->i_nlink); + + sb = inode->i_sb; + xib_calc_bit(inode->i_ino, &pindex, &bit); + AuDebugOn(page_bits <= bit); + sbinfo = au_sbi(sb); + mutex_lock(&sbinfo->si_xib_mtx); + err = xib_pindex(sb, pindex); + if (!err) { + clear_bit(bit, sbinfo->si_xib_buf); + sbinfo->si_xib_next_bit = bit; + } + mutex_unlock(&sbinfo->si_xib_mtx); +} + +/* ---------------------------------------------------------------------- */ + +/* + * truncate a xino bitmap file + */ + +/* todo: slow */ +static int do_xib_restore(struct super_block *sb, struct file *file, void *page) +{ + int err, bit; + ssize_t sz; + unsigned long pindex; + loff_t pos, pend; + struct au_sbinfo *sbinfo; + ino_t *ino; + unsigned long *p; + + err = 0; + sbinfo = au_sbi(sb); + MtxMustLock(&sbinfo->si_xib_mtx); + p = sbinfo->si_xib_buf; + pend = vfsub_f_size_read(file); + pos = 0; + while (pos < pend) { + sz = xino_fread(file, page, PAGE_SIZE, &pos); + err = sz; + if (unlikely(sz <= 0)) + goto out; + + err = 0; + for (ino = page; sz > 0; ino++, sz -= sizeof(ino)) { + if (unlikely(*ino < AUFS_FIRST_INO)) + continue; + + xib_calc_bit(*ino, &pindex, &bit); + AuDebugOn(page_bits <= bit); + err = xib_pindex(sb, pindex); + if (!err) + set_bit(bit, p); + else + goto out; + } + } + +out: + return err; +} + +static int xib_restore(struct super_block *sb) +{ + int err, i; + unsigned int nfile; + aufs_bindex_t bindex, bbot; + void *page; + struct au_branch *br; + struct au_xino *xi; + struct file *file; + + err = -ENOMEM; + page = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!page)) + goto out; + + err = 0; + bbot = au_sbbot(sb); + for (bindex = 0; !err && bindex <= bbot; bindex++) + if (!bindex || is_sb_shared(sb, bindex, bindex - 1) < 0) { + br = au_sbr(sb, bindex); + xi = br->br_xino; + nfile = xi->xi_nfile; + for (i = 0; i < nfile; i++) { + file = au_xino_file(xi, i); + if (file) + err = do_xib_restore(sb, file, page); + } + } else + AuDbg("skip shared b%d\n", bindex); + free_page((unsigned long)page); + +out: + return err; +} + +int au_xib_trunc(struct super_block *sb) +{ + int err; + ssize_t sz; + loff_t pos; + struct au_sbinfo *sbinfo; + unsigned long *p; + struct file *file; + + SiMustWriteLock(sb); + + err = 0; + sbinfo = au_sbi(sb); + if (!au_opt_test(sbinfo->si_mntflags, XINO)) + goto out; + + file = sbinfo->si_xib; + if (vfsub_f_size_read(file) <= PAGE_SIZE) + goto out; + + file = au_xino_create2(sb, &sbinfo->si_xib->f_path, NULL); + err = PTR_ERR(file); + if (IS_ERR(file)) + goto out; + fput(sbinfo->si_xib); + sbinfo->si_xib = file; + + p = sbinfo->si_xib_buf; + memset(p, 0, PAGE_SIZE); + pos = 0; + sz = xino_fwrite(sbinfo->si_xib, p, PAGE_SIZE, &pos); + if (unlikely(sz != PAGE_SIZE)) { + err = sz; + AuIOErr("err %d\n", err); + if (sz >= 0) + err = -EIO; + goto out; + } + + mutex_lock(&sbinfo->si_xib_mtx); + /* mnt_want_write() is unnecessary here */ + err = xib_restore(sb); + mutex_unlock(&sbinfo->si_xib_mtx); + +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +struct au_xino *au_xino_alloc(unsigned int nfile) +{ + struct au_xino *xi; + + xi = kzalloc(sizeof(*xi), GFP_NOFS); + if (unlikely(!xi)) + goto out; + xi->xi_nfile = nfile; + xi->xi_file = kcalloc(nfile, sizeof(*xi->xi_file), GFP_NOFS); + if (unlikely(!xi->xi_file)) + goto out_free; + + xi->xi_nondir.total = 8; /* initial size */ + xi->xi_nondir.array = kcalloc(xi->xi_nondir.total, sizeof(ino_t), + GFP_NOFS); + if (unlikely(!xi->xi_nondir.array)) + goto out_file; + + spin_lock_init(&xi->xi_nondir.spin); + init_waitqueue_head(&xi->xi_nondir.wqh); + mutex_init(&xi->xi_mtx); + INIT_HLIST_BL_HEAD(&xi->xi_writing); + atomic_set(&xi->xi_truncating, 0); + kref_init(&xi->xi_kref); + goto out; /* success */ + +out_file: + au_kfree_try_rcu(xi->xi_file); +out_free: + au_kfree_rcu(xi); + xi = NULL; +out: + return xi; +} + +static int au_xino_init(struct au_branch *br, int idx, struct file *file) +{ + int err; + struct au_xino *xi; + + err = 0; + xi = au_xino_alloc(idx + 1); + if (unlikely(!xi)) { + err = -ENOMEM; + goto out; + } + + if (file) + get_file(file); + xi->xi_file[idx] = file; + AuDebugOn(br->br_xino); + br->br_xino = xi; + +out: + return err; +} + +static void au_xino_release(struct kref *kref) +{ + struct au_xino *xi; + int i; + unsigned long ul; + struct hlist_bl_head *hbl; + struct hlist_bl_node *pos, *n; + struct au_xi_writing *p; + + xi = container_of(kref, struct au_xino, xi_kref); + for (i = 0; i < xi->xi_nfile; i++) + if (xi->xi_file[i]) + fput(xi->xi_file[i]); + for (i = xi->xi_nondir.total - 1; i >= 0; i--) + AuDebugOn(xi->xi_nondir.array[i]); + mutex_destroy(&xi->xi_mtx); + hbl = &xi->xi_writing; + ul = au_hbl_count(hbl); + if (unlikely(ul)) { + pr_warn("xi_writing %lu\n", ul); + hlist_bl_lock(hbl); + hlist_bl_for_each_entry_safe(p, pos, n, hbl, node) { + hlist_bl_del(&p->node); + /* kmemleak reported au_kfree_rcu() doesn't free it */ + kfree(p); + } + hlist_bl_unlock(hbl); + } + au_kfree_try_rcu(xi->xi_file); + au_kfree_try_rcu(xi->xi_nondir.array); + au_kfree_rcu(xi); +} + +int au_xino_put(struct au_branch *br) +{ + int ret; + struct au_xino *xi; + + ret = 0; + xi = br->br_xino; + if (xi) { + br->br_xino = NULL; + ret = kref_put(&xi->xi_kref, au_xino_release); + } + + return ret; +} + +/* ---------------------------------------------------------------------- */ + +/* + * xino mount option handlers + */ + +/* xino bitmap */ +static void xino_clear_xib(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + if (sbinfo->si_xib) + fput(sbinfo->si_xib); + sbinfo->si_xib = NULL; + if (sbinfo->si_xib_buf) + free_page((unsigned long)sbinfo->si_xib_buf); + sbinfo->si_xib_buf = NULL; +} + +static int au_xino_set_xib(struct super_block *sb, struct path *path) +{ + int err; + loff_t pos; + struct au_sbinfo *sbinfo; + struct file *file; + struct super_block *xi_sb; + + SiMustWriteLock(sb); + + sbinfo = au_sbi(sb); + file = au_xino_create2(sb, path, sbinfo->si_xib); + err = PTR_ERR(file); + if (IS_ERR(file)) + goto out; + if (sbinfo->si_xib) + fput(sbinfo->si_xib); + sbinfo->si_xib = file; + xi_sb = file_inode(file)->i_sb; + sbinfo->si_ximaxent = xi_sb->s_maxbytes; + if (unlikely(sbinfo->si_ximaxent < PAGE_SIZE)) { + err = -EIO; + pr_err("s_maxbytes(%llu) on %s is too small\n", + (u64)sbinfo->si_ximaxent, au_sbtype(xi_sb)); + goto out_unset; + } + sbinfo->si_ximaxent /= sizeof(ino_t); + + err = -ENOMEM; + if (!sbinfo->si_xib_buf) + sbinfo->si_xib_buf = (void *)get_zeroed_page(GFP_NOFS); + if (unlikely(!sbinfo->si_xib_buf)) + goto out_unset; + + sbinfo->si_xib_last_pindex = 0; + sbinfo->si_xib_next_bit = 0; + if (vfsub_f_size_read(file) < PAGE_SIZE) { + pos = 0; + err = xino_fwrite(file, sbinfo->si_xib_buf, PAGE_SIZE, &pos); + if (unlikely(err != PAGE_SIZE)) + goto out_free; + } + err = 0; + goto out; /* success */ + +out_free: + if (sbinfo->si_xib_buf) + free_page((unsigned long)sbinfo->si_xib_buf); + sbinfo->si_xib_buf = NULL; + if (err >= 0) + err = -EIO; +out_unset: + fput(sbinfo->si_xib); + sbinfo->si_xib = NULL; +out: + AuTraceErr(err); + return err; +} + +/* xino for each branch */ +static void xino_clear_br(struct super_block *sb) +{ + aufs_bindex_t bindex, bbot; + struct au_branch *br; + + bbot = au_sbbot(sb); + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + AuDebugOn(!br); + au_xino_put(br); + } +} + +static void au_xino_set_br_shared(struct super_block *sb, struct au_branch *br, + aufs_bindex_t bshared) +{ + struct au_branch *brshared; + + brshared = au_sbr(sb, bshared); + AuDebugOn(!brshared->br_xino); + AuDebugOn(!brshared->br_xino->xi_file); + if (br->br_xino != brshared->br_xino) { + au_xino_get(brshared); + au_xino_put(br); + br->br_xino = brshared->br_xino; + } +} + +struct au_xino_do_set_br { + struct au_branch *br; + ino_t h_ino; + aufs_bindex_t bshared; +}; + +static int au_xino_do_set_br(struct super_block *sb, struct path *path, + struct au_xino_do_set_br *args) +{ + int err; + struct au_xi_calc calc; + struct file *file; + struct au_branch *br; + struct au_xi_new xinew = { + .base = path + }; + + br = args->br; + xinew.xi = br->br_xino; + au_xi_calc(sb, args->h_ino, &calc); + xinew.copy_src = au_xino_file(xinew.xi, calc.idx); + if (args->bshared >= 0) + /* shared xino */ + au_xino_set_br_shared(sb, br, args->bshared); + else if (!xinew.xi) { + /* new xino */ + err = au_xino_init(br, calc.idx, xinew.copy_src); + if (unlikely(err)) + goto out; + } + + /* force re-creating */ + xinew.xi = br->br_xino; + xinew.idx = calc.idx; + mutex_lock(&xinew.xi->xi_mtx); + file = au_xi_new(sb, &xinew); + mutex_unlock(&xinew.xi->xi_mtx); + err = PTR_ERR(file); + if (IS_ERR(file)) + goto out; + AuDebugOn(!file); + + err = au_xino_do_write(file, &calc, AUFS_ROOT_INO); + if (unlikely(err)) + au_xino_put(br); + +out: + AuTraceErr(err); + return err; +} + +static int au_xino_set_br(struct super_block *sb, struct path *path) +{ + int err; + aufs_bindex_t bindex, bbot; + struct au_xino_do_set_br args; + struct inode *inode; + + SiMustWriteLock(sb); + + bbot = au_sbbot(sb); + inode = d_inode(sb->s_root); + for (bindex = 0; bindex <= bbot; bindex++) { + args.h_ino = au_h_iptr(inode, bindex)->i_ino; + args.br = au_sbr(sb, bindex); + args.bshared = is_sb_shared(sb, bindex, bindex - 1); + err = au_xino_do_set_br(sb, path, &args); + if (unlikely(err)) + break; + } + + AuTraceErr(err); + return err; +} + +void au_xino_clr(struct super_block *sb) +{ + struct au_sbinfo *sbinfo; + + au_xigen_clr(sb); + xino_clear_xib(sb); + xino_clear_br(sb); + dbgaufs_brs_del(sb, 0); + sbinfo = au_sbi(sb); + /* lvalue, do not call au_mntflags() */ + au_opt_clr(sbinfo->si_mntflags, XINO); +} + +int au_xino_set(struct super_block *sb, struct au_opt_xino *xiopt, int remount) +{ + int err, skip; + struct dentry *dentry, *parent, *cur_dentry, *cur_parent; + struct qstr *dname, *cur_name; + struct file *cur_xino; + struct au_sbinfo *sbinfo; + struct path *path, *cur_path; + + SiMustWriteLock(sb); + + err = 0; + sbinfo = au_sbi(sb); + path = &xiopt->file->f_path; + dentry = path->dentry; + parent = dget_parent(dentry); + if (remount) { + skip = 0; + cur_xino = sbinfo->si_xib; + if (cur_xino) { + cur_path = &cur_xino->f_path; + cur_dentry = cur_path->dentry; + cur_parent = dget_parent(cur_dentry); + cur_name = &cur_dentry->d_name; + dname = &dentry->d_name; + skip = (cur_parent == parent + && au_qstreq(dname, cur_name)); + dput(cur_parent); + } + if (skip) + goto out; + } + + au_opt_set(sbinfo->si_mntflags, XINO); + err = au_xino_set_xib(sb, path); + /* si_x{read,write} are set */ + if (!err) + err = au_xigen_set(sb, path); + if (!err) + err = au_xino_set_br(sb, path); + if (!err) { + dbgaufs_brs_add(sb, 0, /*topdown*/1); + goto out; /* success */ + } + + /* reset all */ + AuIOErr("failed setting xino(%d).\n", err); + au_xino_clr(sb); + +out: + dput(parent); + return err; +} + +/* + * create a xinofile at the default place/path. + */ +struct file *au_xino_def(struct super_block *sb) +{ + struct file *file; + char *page, *p; + struct au_branch *br; + struct super_block *h_sb; + struct path path; + aufs_bindex_t bbot, bindex, bwr; + + br = NULL; + bbot = au_sbbot(sb); + bwr = -1; + for (bindex = 0; bindex <= bbot; bindex++) { + br = au_sbr(sb, bindex); + if (au_br_writable(br->br_perm) + && !au_test_fs_bad_xino(au_br_sb(br))) { + bwr = bindex; + break; + } + } + + if (bwr >= 0) { + file = ERR_PTR(-ENOMEM); + page = (void *)__get_free_page(GFP_NOFS); + if (unlikely(!page)) + goto out; + path.mnt = au_br_mnt(br); + path.dentry = au_h_dptr(sb->s_root, bwr); + p = d_path(&path, page, PATH_MAX - sizeof(AUFS_XINO_FNAME)); + file = (void *)p; + if (!IS_ERR(p)) { + strcat(p, "/" AUFS_XINO_FNAME); + AuDbg("%s\n", p); + file = au_xino_create(sb, p, /*silent*/0, /*wbrtop*/1); + } + free_page((unsigned long)page); + } else { + file = au_xino_create(sb, AUFS_XINO_DEFPATH, /*silent*/0, + /*wbrtop*/0); + if (IS_ERR(file)) + goto out; + h_sb = file->f_path.dentry->d_sb; + if (unlikely(au_test_fs_bad_xino(h_sb))) { + pr_err("xino doesn't support %s(%s)\n", + AUFS_XINO_DEFPATH, au_sbtype(h_sb)); + fput(file); + file = ERR_PTR(-EINVAL); + } + } + +out: + return file; +} + +/* ---------------------------------------------------------------------- */ + +/* + * initialize the xinofile for the specified branch @br + * at the place/path where @base_file indicates. + * test whether another branch is on the same filesystem or not, + * if found then share the xinofile with another branch. + */ +int au_xino_init_br(struct super_block *sb, struct au_branch *br, ino_t h_ino, + struct path *base) +{ + int err; + struct au_xino_do_set_br args = { + .h_ino = h_ino, + .br = br + }; + + args.bshared = sbr_find_shared(sb, /*btop*/0, au_sbbot(sb), + au_br_sb(br)); + err = au_xino_do_set_br(sb, base, &args); + if (unlikely(err)) + au_xino_put(br); + + return err; +} + +/* ---------------------------------------------------------------------- */ + +/* + * get an unused inode number from bitmap + */ +ino_t au_xino_new_ino(struct super_block *sb) +{ + ino_t ino; + unsigned long *p, pindex, ul, pend; + struct au_sbinfo *sbinfo; + struct file *file; + int free_bit, err; + + if (!au_opt_test(au_mntflags(sb), XINO)) + return iunique(sb, AUFS_FIRST_INO); + + sbinfo = au_sbi(sb); + mutex_lock(&sbinfo->si_xib_mtx); + p = sbinfo->si_xib_buf; + free_bit = sbinfo->si_xib_next_bit; + if (free_bit < page_bits && !test_bit(free_bit, p)) + goto out; /* success */ + free_bit = find_first_zero_bit(p, page_bits); + if (free_bit < page_bits) + goto out; /* success */ + + pindex = sbinfo->si_xib_last_pindex; + for (ul = pindex - 1; ul < ULONG_MAX; ul--) { + err = xib_pindex(sb, ul); + if (unlikely(err)) + goto out_err; + free_bit = find_first_zero_bit(p, page_bits); + if (free_bit < page_bits) + goto out; /* success */ + } + + file = sbinfo->si_xib; + pend = vfsub_f_size_read(file) / PAGE_SIZE; + for (ul = pindex + 1; ul <= pend; ul++) { + err = xib_pindex(sb, ul); + if (unlikely(err)) + goto out_err; + free_bit = find_first_zero_bit(p, page_bits); + if (free_bit < page_bits) + goto out; /* success */ + } + BUG(); + +out: + set_bit(free_bit, p); + sbinfo->si_xib_next_bit = free_bit + 1; + pindex = sbinfo->si_xib_last_pindex; + mutex_unlock(&sbinfo->si_xib_mtx); + ino = xib_calc_ino(pindex, free_bit); + AuDbg("i%lu\n", (unsigned long)ino); + return ino; +out_err: + mutex_unlock(&sbinfo->si_xib_mtx); + AuDbg("i0\n"); + return 0; +} + +/* for s_op->delete_inode() */ +void au_xino_delete_inode(struct inode *inode, const int unlinked) +{ + int err; + unsigned int mnt_flags; + aufs_bindex_t bindex, bbot, bi; + unsigned char try_trunc; + struct au_iinfo *iinfo; + struct super_block *sb; + struct au_hinode *hi; + struct inode *h_inode; + struct au_branch *br; + struct au_xi_calc calc; + struct file *file; + + AuDebugOn(au_is_bad_inode(inode)); + + sb = inode->i_sb; + mnt_flags = au_mntflags(sb); + if (!au_opt_test(mnt_flags, XINO) + || inode->i_ino == AUFS_ROOT_INO) + return; + + if (unlinked) { + au_xigen_inc(inode); + au_xib_clear_bit(inode); + } + + iinfo = au_ii(inode); + bindex = iinfo->ii_btop; + if (bindex < 0) + return; + + try_trunc = !!au_opt_test(mnt_flags, TRUNC_XINO); + hi = au_hinode(iinfo, bindex); + bbot = iinfo->ii_bbot; + for (; bindex <= bbot; bindex++, hi++) { + h_inode = hi->hi_inode; + if (!h_inode + || (!unlinked && h_inode->i_nlink)) + continue; + + /* inode may not be revalidated */ + bi = au_br_index(sb, hi->hi_id); + if (bi < 0) + continue; + + br = au_sbr(sb, bi); + au_xi_calc(sb, h_inode->i_ino, &calc); + file = au_xino_file(br->br_xino, calc.idx); + if (IS_ERR_OR_NULL(file)) + continue; + + err = au_xino_do_write(file, &calc, /*ino*/0); + if (!err && try_trunc + && au_test_fs_trunc_xino(au_br_sb(br))) + xino_try_trunc(sb, br); + } +} + +/* ---------------------------------------------------------------------- */ + +static int au_xinondir_find(struct au_xino *xi, ino_t h_ino) +{ + int found, total, i; + + found = -1; + total = xi->xi_nondir.total; + for (i = 0; i < total; i++) { + if (xi->xi_nondir.array[i] != h_ino) + continue; + found = i; + break; + } + + return found; +} + +static int au_xinondir_expand(struct au_xino *xi) +{ + int err, sz; + ino_t *p; + + BUILD_BUG_ON(KMALLOC_MAX_SIZE > INT_MAX); + + err = -ENOMEM; + sz = xi->xi_nondir.total * sizeof(ino_t); + if (unlikely(sz > KMALLOC_MAX_SIZE / 2)) + goto out; + p = au_kzrealloc(xi->xi_nondir.array, sz, sz << 1, GFP_ATOMIC, + /*may_shrink*/0); + if (p) { + xi->xi_nondir.array = p; + xi->xi_nondir.total <<= 1; + AuDbg("xi_nondir.total %d\n", xi->xi_nondir.total); + err = 0; + } + +out: + return err; +} + +void au_xinondir_leave(struct super_block *sb, aufs_bindex_t bindex, + ino_t h_ino, int idx) +{ + struct au_xino *xi; + + AuDebugOn(!au_opt_test(au_mntflags(sb), XINO)); + xi = au_sbr(sb, bindex)->br_xino; + AuDebugOn(idx < 0 || xi->xi_nondir.total <= idx); + + spin_lock(&xi->xi_nondir.spin); + AuDebugOn(xi->xi_nondir.array[idx] != h_ino); + xi->xi_nondir.array[idx] = 0; + spin_unlock(&xi->xi_nondir.spin); + wake_up_all(&xi->xi_nondir.wqh); +} + +int au_xinondir_enter(struct super_block *sb, aufs_bindex_t bindex, ino_t h_ino, + int *idx) +{ + int err, found, empty; + struct au_xino *xi; + + err = 0; + *idx = -1; + if (!au_opt_test(au_mntflags(sb), XINO)) + goto out; /* no xino */ + + xi = au_sbr(sb, bindex)->br_xino; + +again: + spin_lock(&xi->xi_nondir.spin); + found = au_xinondir_find(xi, h_ino); + if (found == -1) { + empty = au_xinondir_find(xi, /*h_ino*/0); + if (empty == -1) { + empty = xi->xi_nondir.total; + err = au_xinondir_expand(xi); + if (unlikely(err)) + goto out_unlock; + } + xi->xi_nondir.array[empty] = h_ino; + *idx = empty; + } else { + spin_unlock(&xi->xi_nondir.spin); + wait_event(xi->xi_nondir.wqh, + xi->xi_nondir.array[found] != h_ino); + goto again; + } + +out_unlock: + spin_unlock(&xi->xi_nondir.spin); +out: + return err; +} + +/* ---------------------------------------------------------------------- */ + +int au_xino_path(struct seq_file *seq, struct file *file) +{ + int err; + + err = au_seq_path(seq, &file->f_path); + if (unlikely(err)) + goto out; + +#define Deleted "\\040(deleted)" + seq->count -= sizeof(Deleted) - 1; + AuDebugOn(memcmp(seq->buf + seq->count, Deleted, + sizeof(Deleted) - 1)); +#undef Deleted + +out: + return err; +} diff --git a/fs/dcache.c b/fs/dcache.c index cf871a81f4f..9508bd57a3b 100644 --- a/fs/dcache.c +++ b/fs/dcache.c @@ -1320,7 +1320,7 @@ enum d_walk_ret { * * The @enter() callbacks are called with d_lock held. */ -static void d_walk(struct dentry *parent, void *data, +void d_walk(struct dentry *parent, void *data, enum d_walk_ret (*enter)(void *, struct dentry *)) { struct dentry *this_parent; @@ -1425,6 +1425,7 @@ static void d_walk(struct dentry *parent, void *data, seq = 1; goto again; } +EXPORT_SYMBOL_GPL(d_walk); struct check_mount { struct vfsmount *mnt; @@ -2970,6 +2971,7 @@ void d_exchange(struct dentry *dentry1, struct dentry *dentry2) write_sequnlock(&rename_lock); } +EXPORT_SYMBOL_GPL(d_exchange); /** * d_ancestor - search for an ancestor diff --git a/fs/exec.c b/fs/exec.c index d7c4187ca02..22b8aa23954 100644 --- a/fs/exec.c +++ b/fs/exec.c @@ -65,6 +65,7 @@ #include #include #include +#include #include #include @@ -114,6 +115,7 @@ bool path_noexec(const struct path *path) return (path->mnt->mnt_flags & MNT_NOEXEC) || (path->mnt->mnt_sb->s_iflags & SB_I_NOEXEC); } +EXPORT_SYMBOL_GPL(path_noexec); #ifdef CONFIG_USELIB /* @@ -1008,6 +1010,7 @@ static int exec_mmap(struct mm_struct *mm) active_mm = tsk->active_mm; tsk->active_mm = mm; tsk->mm = mm; + lru_gen_add_mm(mm); /* * This prevents preemption while active_mm is being loaded and * it and mm are being updated, which could cause problems for @@ -1018,6 +1021,7 @@ static int exec_mmap(struct mm_struct *mm) if (!IS_ENABLED(CONFIG_ARCH_WANT_IRQS_OFF_ACTIVATE_MM)) local_irq_enable(); activate_mm(active_mm, mm); + lru_gen_switch_mm(active_mm, mm); if (IS_ENABLED(CONFIG_ARCH_WANT_IRQS_OFF_ACTIVATE_MM)) local_irq_enable(); tsk->mm->vmacache_seqnum = 0; diff --git a/fs/fcntl.c b/fs/fcntl.c index dfc72f15be7..05329961863 100644 --- a/fs/fcntl.c +++ b/fs/fcntl.c @@ -33,7 +33,7 @@ #define SETFL_MASK (O_APPEND | O_NONBLOCK | O_NDELAY | O_DIRECT | O_NOATIME) -static int setfl(int fd, struct file * filp, unsigned long arg) +int setfl(int fd, struct file *filp, unsigned long arg) { struct inode * inode = file_inode(filp); int error = 0; @@ -64,6 +64,8 @@ static int setfl(int fd, struct file * filp, unsigned long arg) if (filp->f_op->check_flags) error = filp->f_op->check_flags(arg); + if (!error && filp->f_op->setfl) + error = filp->f_op->setfl(filp, arg); if (error) return error; @@ -84,6 +86,7 @@ static int setfl(int fd, struct file * filp, unsigned long arg) out: return error; } +EXPORT_SYMBOL_GPL(setfl); static void f_modown(struct file *filp, struct pid *pid, enum pid_type type, int force) diff --git a/fs/file_table.c b/fs/file_table.c index 45437f8e100..786af52904f 100644 --- a/fs/file_table.c +++ b/fs/file_table.c @@ -161,6 +161,7 @@ struct file *alloc_empty_file(int flags, const struct cred *cred) } return ERR_PTR(-ENFILE); } +EXPORT_SYMBOL_GPL(alloc_empty_file); /* * Variant of alloc_empty_file() that doesn't check and modify nr_files. @@ -375,6 +376,7 @@ void __fput_sync(struct file *file) } EXPORT_SYMBOL(fput); +EXPORT_SYMBOL_GPL(__fput_sync); void __init files_init(void) { diff --git a/fs/fuse/dev.c b/fs/fuse/dev.c index b8d58aa0820..5d99a03b749 100644 --- a/fs/fuse/dev.c +++ b/fs/fuse/dev.c @@ -785,7 +785,8 @@ static int fuse_check_page(struct page *page) 1 << PG_active | 1 << PG_workingset | 1 << PG_reclaim | - 1 << PG_waiters))) { + 1 << PG_waiters | + LRU_GEN_MASK | LRU_USAGE_MASK))) { dump_page(page, "fuse: trying to steal weird page"); return 1; } diff --git a/fs/inode.c b/fs/inode.c index c93500d8426..01060086352 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -138,6 +138,7 @@ int inode_init_always(struct super_block *sb, struct inode *inode) inode->i_blkbits = sb->s_blocksize_bits; inode->i_flags = 0; atomic64_set(&inode->i_sequence, 0); + atomic64_set(&inode->i_sequence2, 0); atomic_set(&inode->i_count, 1); inode->i_op = &empty_iops; inode->i_fop = &no_open_fops; @@ -1778,12 +1779,13 @@ EXPORT_SYMBOL(generic_update_time); * This does the actual work of updating an inodes time or version. Must have * had called mnt_want_write() before calling this. */ -static int update_time(struct inode *inode, struct timespec64 *time, int flags) +int update_time(struct inode *inode, struct timespec64 *time, int flags) { if (inode->i_op->update_time) return inode->i_op->update_time(inode, time, flags); return generic_update_time(inode, time, flags); } +EXPORT_SYMBOL_GPL(update_time); /** * atime_needs_update - update the access time diff --git a/fs/namespace.c b/fs/namespace.c index c3f1a78ba36..df3e9a94be9 100644 --- a/fs/namespace.c +++ b/fs/namespace.c @@ -438,6 +438,7 @@ void __mnt_drop_write(struct vfsmount *mnt) mnt_dec_writers(real_mount(mnt)); preempt_enable(); } +EXPORT_SYMBOL_GPL(__mnt_drop_write); /** * mnt_drop_write - give up write access to a mount @@ -807,6 +808,13 @@ static inline int check_mnt(struct mount *mnt) return mnt->mnt_ns == current->nsproxy->mnt_ns; } +/* for aufs, CONFIG_AUFS_BR_FUSE */ +int is_current_mnt_ns(struct vfsmount *mnt) +{ + return check_mnt(real_mount(mnt)); +} +EXPORT_SYMBOL_GPL(is_current_mnt_ns); + /* * vfsmount lock must be held for write */ @@ -1981,6 +1989,7 @@ int iterate_mounts(int (*f)(struct vfsmount *, void *), void *arg, } return 0; } +EXPORT_SYMBOL_GPL(iterate_mounts); static void lock_mnt_tree(struct mount *mnt) { diff --git a/fs/notify/group.c b/fs/notify/group.c index fb89c351295..460ad19c257 100644 --- a/fs/notify/group.c +++ b/fs/notify/group.c @@ -100,6 +100,7 @@ void fsnotify_get_group(struct fsnotify_group *group) { refcount_inc(&group->refcnt); } +EXPORT_SYMBOL_GPL(fsnotify_get_group); /* * Drop a reference to a group. Free it if it's through. diff --git a/fs/open.c b/fs/open.c index 53bc0573c0e..4be00d1ec64 100644 --- a/fs/open.c +++ b/fs/open.c @@ -65,6 +65,7 @@ int do_truncate(struct user_namespace *mnt_userns, struct dentry *dentry, inode_unlock(dentry->d_inode); return ret; } +EXPORT_SYMBOL_GPL(do_truncate); long vfs_truncate(const struct path *path, loff_t length) { diff --git a/fs/proc/base.c b/fs/proc/base.c index a0a2fc1c9da..03b04b2e80c 100644 --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -476,7 +476,7 @@ static int proc_pid_schedstat(struct seq_file *m, struct pid_namespace *ns, seq_puts(m, "0 0 0\n"); else seq_printf(m, "%llu %llu %lu\n", - (unsigned long long)task->se.sum_exec_runtime, + (unsigned long long)tsk_seruntime(task), (unsigned long long)task->sched_info.run_delay, task->sched_info.pcount); @@ -2183,7 +2183,7 @@ static int map_files_get_link(struct dentry *dentry, struct path *path) rc = -ENOENT; vma = find_exact_vma(mm, vm_start, vm_end); if (vma && vma->vm_file) { - *path = vma->vm_file->f_path; + *path = vma_pr_or_file(vma)->f_path; path_get(path); rc = 0; } diff --git a/fs/proc/meminfo.c b/fs/proc/meminfo.c index 6fa761c9cc7..45fd59a0d85 100644 --- a/fs/proc/meminfo.c +++ b/fs/proc/meminfo.c @@ -108,6 +108,10 @@ static int meminfo_proc_show(struct seq_file *m, void *v) #endif show_val_kb(m, "PageTables: ", global_node_page_state(NR_PAGETABLE)); +#ifdef CONFIG_UKSM + show_val_kb(m, "KsmZeroPages: ", + global_zone_page_state(NR_UKSM_ZERO_PAGES)); +#endif show_val_kb(m, "NFS_Unstable: ", 0); show_val_kb(m, "Bounce: ", diff --git a/fs/proc/nommu.c b/fs/proc/nommu.c index 13452b32e2b..38acccfef9d 100644 --- a/fs/proc/nommu.c +++ b/fs/proc/nommu.c @@ -40,7 +40,10 @@ static int nommu_region_show(struct seq_file *m, struct vm_region *region) file = region->vm_file; if (file) { - struct inode *inode = file_inode(region->vm_file); + struct inode *inode; + + file = vmr_pr_or_file(region); + inode = file_inode(file); dev = inode->i_sb->s_dev; ino = inode->i_ino; } diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 7389df326ed..0c6a9fc1db5 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -280,7 +280,10 @@ show_map_vma(struct seq_file *m, struct vm_area_struct *vma) const char *name = NULL; if (file) { - struct inode *inode = file_inode(vma->vm_file); + struct inode *inode; + + file = vma_pr_or_file(vma); + inode = file_inode(file); dev = inode->i_sb->s_dev; ino = inode->i_ino; pgoff = ((loff_t)vma->vm_pgoff) << PAGE_SHIFT; @@ -1864,7 +1867,7 @@ static int show_numa_map(struct seq_file *m, void *v) struct proc_maps_private *proc_priv = &numa_priv->proc_maps; struct vm_area_struct *vma = v; struct numa_maps *md = &numa_priv->md; - struct file *file = vma->vm_file; + struct file *file = vma_pr_or_file(vma); struct mm_struct *mm = vma->vm_mm; struct mempolicy *pol; char buffer[64]; diff --git a/fs/proc/task_nommu.c b/fs/proc/task_nommu.c index a6d21fc0033..02c2de31196 100644 --- a/fs/proc/task_nommu.c +++ b/fs/proc/task_nommu.c @@ -155,7 +155,10 @@ static int nommu_vma_show(struct seq_file *m, struct vm_area_struct *vma) file = vma->vm_file; if (file) { - struct inode *inode = file_inode(vma->vm_file); + struct inode *inode; + + file = vma_pr_or_file(vma); + inode = file_inode(file); dev = inode->i_sb->s_dev; ino = inode->i_ino; pgoff = (loff_t)vma->vm_pgoff << PAGE_SHIFT; diff --git a/fs/read_write.c b/fs/read_write.c index 9db7adf160d..8dc93a57a93 100644 --- a/fs/read_write.c +++ b/fs/read_write.c @@ -503,6 +503,7 @@ ssize_t vfs_read(struct file *file, char __user *buf, size_t count, loff_t *pos) inc_syscr(current); return ret; } +EXPORT_SYMBOL_GPL(vfs_read); static ssize_t new_sync_write(struct file *filp, const char __user *buf, size_t len, loff_t *ppos) { @@ -613,6 +614,7 @@ ssize_t vfs_write(struct file *file, const char __user *buf, size_t count, loff_ file_end_write(file); return ret; } +EXPORT_SYMBOL_GPL(vfs_write); /* file_ppos returns &file->f_pos or NULL if file is stream */ static inline loff_t *file_ppos(struct file *file) diff --git a/fs/splice.c b/fs/splice.c index 5dbce4dcc1a..7c1be373eb7 100644 --- a/fs/splice.c +++ b/fs/splice.c @@ -759,20 +759,21 @@ static int warn_unsupported(struct file *file, const char *op) /* * Attempt to initiate a splice from pipe to file. */ -static long do_splice_from(struct pipe_inode_info *pipe, struct file *out, - loff_t *ppos, size_t len, unsigned int flags) +long do_splice_from(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags) { if (unlikely(!out->f_op->splice_write)) return warn_unsupported(out, "write"); return out->f_op->splice_write(pipe, out, ppos, len, flags); } +EXPORT_SYMBOL_GPL(do_splice_from); /* * Attempt to initiate a splice from a file to a pipe. */ -static long do_splice_to(struct file *in, loff_t *ppos, - struct pipe_inode_info *pipe, size_t len, - unsigned int flags) +long do_splice_to(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags) { unsigned int p_space; int ret; @@ -795,6 +796,7 @@ static long do_splice_to(struct file *in, loff_t *ppos, return warn_unsupported(in, "read"); return in->f_op->splice_read(in, ppos, pipe, len, flags); } +EXPORT_SYMBOL_GPL(do_splice_to); /** * splice_direct_to_actor - splices data directly between two non-pipes diff --git a/fs/sync.c b/fs/sync.c index 1373a610dc7..fa5c7fba7f1 100644 --- a/fs/sync.c +++ b/fs/sync.c @@ -28,7 +28,7 @@ * wait == 1 case since in that case write_inode() functions do * sync_dirty_buffer() and thus effectively write one block at a time. */ -static int __sync_filesystem(struct super_block *sb, int wait) +int __sync_filesystem(struct super_block *sb, int wait) { if (wait) sync_inodes_sb(sb); @@ -39,6 +39,7 @@ static int __sync_filesystem(struct super_block *sb, int wait) sb->s_op->sync_fs(sb, wait); return __sync_blockdev(sb->s_bdev, wait); } +EXPORT_SYMBOL_GPL(__sync_filesystem); /* * Write out and wait upon all dirty data associated with this diff --git a/fs/xattr.c b/fs/xattr.c index 5c8c5175b38..ff7e9ff774b 100644 --- a/fs/xattr.c +++ b/fs/xattr.c @@ -384,6 +384,7 @@ vfs_getxattr_alloc(struct user_namespace *mnt_userns, struct dentry *dentry, *xattr_value = value; return error; } +EXPORT_SYMBOL_GPL(vfs_getxattr_alloc); ssize_t __vfs_getxattr(struct dentry *dentry, struct inode *inode, const char *name, diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index f69c75bd6d2..9bfb2f65534 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -1531,6 +1531,22 @@ static inline int queue_limit_discard_alignment(struct queue_limits *lim, sector return offset << SECTOR_SHIFT; } +/* + * Two cases of handling DISCARD merge: + * If max_discard_segments > 1, the driver takes every bio + * as a range and send them to controller together. The ranges + * needn't to be contiguous. + * Otherwise, the bios/requests will be handled as same as + * others which should be contiguous. + */ +static inline bool blk_discard_mergable(struct request *req) +{ + if (req_op(req) == REQ_OP_DISCARD && + queue_max_discard_segments(req->q) > 1) + return true; + return false; +} + static inline int bdev_discard_alignment(struct block_device *bdev) { struct request_queue *q = bdev_get_queue(bdev); diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h index 6bc9c76680b..e52e44af681 100644 --- a/include/linux/cgroup.h +++ b/include/linux/cgroup.h @@ -432,6 +432,18 @@ static inline void cgroup_put(struct cgroup *cgrp) css_put(&cgrp->self); } +extern struct mutex cgroup_mutex; + +static inline void cgroup_lock(void) +{ + mutex_lock(&cgroup_mutex); +} + +static inline void cgroup_unlock(void) +{ + mutex_unlock(&cgroup_mutex); +} + /** * task_css_set_check - obtain a task's css_set with extra access conditions * @task: the task to obtain css_set for @@ -446,7 +458,6 @@ static inline void cgroup_put(struct cgroup *cgrp) * as locks used during the cgroup_subsys::attach() methods. */ #ifdef CONFIG_PROVE_RCU -extern struct mutex cgroup_mutex; extern spinlock_t css_set_lock; #define task_css_set_check(task, __c) \ rcu_dereference_check((task)->cgroups, \ @@ -704,6 +715,8 @@ struct cgroup; static inline u64 cgroup_id(const struct cgroup *cgrp) { return 1; } static inline void css_get(struct cgroup_subsys_state *css) {} static inline void css_put(struct cgroup_subsys_state *css) {} +static inline void cgroup_lock(void) {} +static inline void cgroup_unlock(void) {} static inline int cgroup_attach_task_all(struct task_struct *from, struct task_struct *t) { return 0; } static inline int cgroupstats_build(struct cgroupstats *stats, diff --git a/include/linux/compat.h b/include/linux/compat.h index 8855b1b702b..34ad63bac18 100644 --- a/include/linux/compat.h +++ b/include/linux/compat.h @@ -368,6 +368,17 @@ struct compat_robust_list_head { compat_uptr_t list_op_pending; }; +struct compat_futex_waitv { + compat_uptr_t uaddr; + compat_uint_t val; + compat_uint_t flags; +}; + +struct compat_futex_requeue { + compat_uptr_t uaddr; + compat_uint_t flags; +}; + #ifdef CONFIG_COMPAT_OLD_SIGACTION struct compat_old_sigaction { compat_uptr_t sa_handler; @@ -692,6 +703,18 @@ asmlinkage long compat_sys_get_robust_list(int pid, compat_uptr_t __user *head_ptr, compat_size_t __user *len_ptr); +/* kernel/futex2.c */ +asmlinkage long compat_sys_futex_waitv(struct compat_futex_waitv *waiters, + compat_uint_t nr_futexes, compat_uint_t flags, + struct __kernel_timespec __user *timo); + +asmlinkage long compat_sys_futex_requeue(struct compat_futex_requeue *uaddr1, + struct compat_futex_requeue *uaddr2, + compat_uint_t nr_wake, + compat_uint_t nr_requeue, + compat_uint_t cmpval, + compat_uint_t flags); + /* kernel/itimer.c */ asmlinkage long compat_sys_getitimer(int which, struct old_itimerval32 __user *it); diff --git a/include/linux/elevator.h b/include/linux/elevator.h index dcb2f9022c1..0b3c67155d8 100644 --- a/include/linux/elevator.h +++ b/include/linux/elevator.h @@ -52,7 +52,7 @@ struct elevator_mq_ops { void (*exit_icq)(struct io_cq *); }; -#define ELV_NAME_MAX (16) +#define ELV_NAME_MAX (24) struct elv_fs_entry { struct attribute attr; diff --git a/include/linux/freezer.h b/include/linux/freezer.h index 0621c5f86c3..9a1d75742ce 100644 --- a/include/linux/freezer.h +++ b/include/linux/freezer.h @@ -310,6 +310,9 @@ static inline void set_freezable(void) {} #define wait_event_freezekillable_unsafe(wq, condition) \ wait_event_killable(wq, condition) +#ifdef CONFIG_SCHED_MUQSS +#define pm_freezing (false) +#endif #endif /* !CONFIG_FREEZER */ #endif /* FREEZER_H_INCLUDED */ diff --git a/include/linux/fs.h b/include/linux/fs.h index c3c88fdb9b2..7ab93f10f1d 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -682,6 +682,7 @@ struct inode { }; atomic64_t i_version; atomic64_t i_sequence; /* see futex */ + atomic64_t i_sequence2; /* see futex2 */ atomic_t i_count; atomic_t i_dio_count; atomic_t i_writecount; @@ -1335,6 +1336,7 @@ extern void fasync_free(struct fasync_struct *); /* can be called from interrupts */ extern void kill_fasync(struct fasync_struct **, int, int); +extern int setfl(int fd, struct file *filp, unsigned long arg); extern void __f_setown(struct file *filp, struct pid *, enum pid_type, int force); extern int f_setown(struct file *filp, unsigned long arg, int force); extern void f_delown(struct file *filp); @@ -2043,6 +2045,7 @@ struct file_operations { ssize_t (*sendpage) (struct file *, struct page *, int, size_t, loff_t *, int); unsigned long (*get_unmapped_area)(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); int (*check_flags)(int); + int (*setfl)(struct file *, unsigned long); int (*flock) (struct file *, int, struct file_lock *); ssize_t (*splice_write)(struct pipe_inode_info *, struct file *, loff_t *, size_t, unsigned int); ssize_t (*splice_read)(struct file *, loff_t *, struct pipe_inode_info *, size_t, unsigned int); @@ -2565,6 +2568,7 @@ extern int current_umask(void); extern void ihold(struct inode * inode); extern void iput(struct inode *); extern int generic_update_time(struct inode *, struct timespec64 *, int); +extern int update_time(struct inode *, struct timespec64 *, int); /* /sys/fs */ extern struct kobject *fs_kobj; @@ -2805,6 +2809,7 @@ static inline bool sb_is_blkdev_sb(struct super_block *sb) } void emergency_thaw_all(void); +extern int __sync_filesystem(struct super_block *, int); extern int sync_filesystem(struct super_block *); extern const struct file_operations def_blk_fops; extern const struct file_operations def_chr_fops; diff --git a/include/linux/init_task.h b/include/linux/init_task.h index 40fc5813cf9..c7b8dd90113 100644 --- a/include/linux/init_task.h +++ b/include/linux/init_task.h @@ -35,7 +35,11 @@ extern struct cred init_cred; #define INIT_PREV_CPUTIME(x) #endif +#ifdef CONFIG_SCHED_MUQSS +#define INIT_TASK_COMM "MuQSS" +#else #define INIT_TASK_COMM "swapper" +#endif /* Attach to the init_task data structure for proper alignment */ #ifdef CONFIG_ARCH_TASK_STRUCT_ON_STACK diff --git a/include/linux/ioprio.h b/include/linux/ioprio.h index e9bfe6972ae..16ba1c7e5bd 100644 --- a/include/linux/ioprio.h +++ b/include/linux/ioprio.h @@ -53,6 +53,8 @@ enum { */ static inline int task_nice_ioprio(struct task_struct *task) { + if (iso_task(task)) + return 0; return (task_nice(task) + 20) / 5; } diff --git a/include/linux/ksm.h b/include/linux/ksm.h index 161e8164abc..f0dbdf3c986 100644 --- a/include/linux/ksm.h +++ b/include/linux/ksm.h @@ -21,20 +21,16 @@ struct mem_cgroup; #ifdef CONFIG_KSM int ksm_madvise(struct vm_area_struct *vma, unsigned long start, unsigned long end, int advice, unsigned long *vm_flags); -int __ksm_enter(struct mm_struct *mm); -void __ksm_exit(struct mm_struct *mm); -static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm) +static inline struct stable_node *page_stable_node(struct page *page) { - if (test_bit(MMF_VM_MERGEABLE, &oldmm->flags)) - return __ksm_enter(mm); - return 0; + return PageKsm(page) ? page_rmapping(page) : NULL; } -static inline void ksm_exit(struct mm_struct *mm) +static inline void set_page_stable_node(struct page *page, + struct stable_node *stable_node) { - if (test_bit(MMF_VM_MERGEABLE, &mm->flags)) - __ksm_exit(mm); + page->mapping = (void *)((unsigned long)stable_node | PAGE_MAPPING_KSM); } /* @@ -54,6 +50,33 @@ struct page *ksm_might_need_to_copy(struct page *page, void rmap_walk_ksm(struct page *page, struct rmap_walk_control *rwc); void ksm_migrate_page(struct page *newpage, struct page *oldpage); +#ifdef CONFIG_KSM_LEGACY +int __ksm_enter(struct mm_struct *mm); +void __ksm_exit(struct mm_struct *mm); +static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm) +{ + if (test_bit(MMF_VM_MERGEABLE, &oldmm->flags)) + return __ksm_enter(mm); + return 0; +} + +static inline void ksm_exit(struct mm_struct *mm) +{ + if (test_bit(MMF_VM_MERGEABLE, &mm->flags)) + __ksm_exit(mm); +} + +#elif defined(CONFIG_UKSM) +static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm) +{ + return 0; +} + +static inline void ksm_exit(struct mm_struct *mm) +{ +} +#endif /* !CONFIG_UKSM */ + #else /* !CONFIG_KSM */ static inline int ksm_fork(struct mm_struct *mm, struct mm_struct *oldmm) @@ -89,4 +112,6 @@ static inline void ksm_migrate_page(struct page *newpage, struct page *oldpage) #endif /* CONFIG_MMU */ #endif /* !CONFIG_KSM */ +#include + #endif /* __LINUX_KSM_H */ diff --git a/include/linux/lockdep.h b/include/linux/lockdep.h index 5cf38781375..59b65e1a9e8 100644 --- a/include/linux/lockdep.h +++ b/include/linux/lockdep.h @@ -248,6 +248,8 @@ static inline int lockdep_match_key(struct lockdep_map *lock, return lock->key == key; } +struct lock_class *lockdep_hlock_class(struct held_lock *hlock); + /* * Acquire a lock. * diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index c193be76070..60601a99743 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -230,6 +230,8 @@ struct obj_cgroup { }; }; +struct lru_gen_mm_list; + /* * The memory controller data structure. The memory controller controls both * page cache and RSS per cgroup. We would eventually like to provide @@ -349,6 +351,10 @@ struct mem_cgroup { struct deferred_split deferred_split_queue; #endif +#ifdef CONFIG_LRU_GEN + struct lru_gen_mm_list *mm_list; +#endif + struct mem_cgroup_per_node *nodeinfo[0]; /* WARNING: nodeinfo must be the last member here */ }; @@ -1131,7 +1137,6 @@ static inline struct mem_cgroup *page_memcg(struct page *page) static inline struct mem_cgroup *page_memcg_rcu(struct page *page) { - WARN_ON_ONCE(!rcu_read_lock_held()); return NULL; } diff --git a/include/linux/mm.h b/include/linux/mm.h index 9afb8998e7e..d76db2ab50d 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -213,6 +213,9 @@ static inline void __mm_zero_struct_page(struct page *page) extern int sysctl_max_map_count; +extern unsigned long sysctl_clean_low_kbytes; +extern unsigned long sysctl_clean_min_kbytes; + extern unsigned long sysctl_user_reserve_kbytes; extern unsigned long sysctl_admin_reserve_kbytes; @@ -1089,6 +1092,8 @@ vm_fault_t finish_mkwrite_fault(struct vm_fault *vmf); #define ZONES_PGOFF (NODES_PGOFF - ZONES_WIDTH) #define LAST_CPUPID_PGOFF (ZONES_PGOFF - LAST_CPUPID_WIDTH) #define KASAN_TAG_PGOFF (LAST_CPUPID_PGOFF - KASAN_TAG_WIDTH) +#define LRU_GEN_PGOFF (KASAN_TAG_PGOFF - LRU_GEN_WIDTH) +#define LRU_USAGE_PGOFF (LRU_GEN_PGOFF - LRU_USAGE_WIDTH) /* * Define the bit shifts to access each section. For non-existent @@ -1801,6 +1806,28 @@ static inline void unmap_shared_mapping_range(struct address_space *mapping, unmap_mapping_range(mapping, holebegin, holelen, 0); } +extern void vma_do_file_update_time(struct vm_area_struct *, const char[], int); +extern struct file *vma_do_pr_or_file(struct vm_area_struct *, const char[], + int); +extern void vma_do_get_file(struct vm_area_struct *, const char[], int); +extern void vma_do_fput(struct vm_area_struct *, const char[], int); + +#define vma_file_update_time(vma) vma_do_file_update_time(vma, __func__, \ + __LINE__) +#define vma_pr_or_file(vma) vma_do_pr_or_file(vma, __func__, \ + __LINE__) +#define vma_get_file(vma) vma_do_get_file(vma, __func__, __LINE__) +#define vma_fput(vma) vma_do_fput(vma, __func__, __LINE__) + +#ifndef CONFIG_MMU +extern struct file *vmr_do_pr_or_file(struct vm_region *, const char[], int); +extern void vmr_do_fput(struct vm_region *, const char[], int); + +#define vmr_pr_or_file(region) vmr_do_pr_or_file(region, __func__, \ + __LINE__) +#define vmr_fput(region) vmr_do_fput(region, __func__, __LINE__) +#endif /* !CONFIG_MMU */ + extern int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, int len, unsigned int gup_flags); extern int access_remote_vm(struct mm_struct *mm, unsigned long addr, diff --git a/include/linux/mm_inline.h b/include/linux/mm_inline.h index 355ea1ee32b..f3b99f65a65 100644 --- a/include/linux/mm_inline.h +++ b/include/linux/mm_inline.h @@ -79,11 +79,239 @@ static __always_inline enum lru_list page_lru(struct page *page) return lru; } +#ifdef CONFIG_LRU_GEN + +#ifdef CONFIG_LRU_GEN_ENABLED +DECLARE_STATIC_KEY_TRUE(lru_gen_static_key); + +static inline bool lru_gen_enabled(void) +{ + return static_branch_likely(&lru_gen_static_key); +} +#else +DECLARE_STATIC_KEY_FALSE(lru_gen_static_key); + +static inline bool lru_gen_enabled(void) +{ + return static_branch_unlikely(&lru_gen_static_key); +} +#endif + +/* We track at most MAX_NR_GENS generations using the sliding window technique. */ +static inline int lru_gen_from_seq(unsigned long seq) +{ + return seq % MAX_NR_GENS; +} + +/* Convert the level of usage to a tier. See the comment on MAX_NR_TIERS. */ +static inline int lru_tier_from_usage(int usage) +{ + return order_base_2(usage + 1); +} + +/* Return a proper index regardless whether we keep a full history of stats. */ +static inline int hist_from_seq_or_gen(int seq_or_gen) +{ + return seq_or_gen % NR_STAT_GENS; +} + +/* The youngest and the second youngest generations are counted as active. */ +static inline bool lru_gen_is_active(struct lruvec *lruvec, int gen) +{ + unsigned long max_seq = READ_ONCE(lruvec->evictable.max_seq); + + VM_BUG_ON(!max_seq); + VM_BUG_ON(gen >= MAX_NR_GENS); + + return gen == lru_gen_from_seq(max_seq) || gen == lru_gen_from_seq(max_seq - 1); +} + +/* Update the sizes of the multigenerational lru lists. */ +static inline void lru_gen_update_size(struct page *page, struct lruvec *lruvec, + int old_gen, int new_gen) +{ + int type = page_is_file_lru(page); + int zone = page_zonenum(page); + int delta = thp_nr_pages(page); + enum lru_list lru = type * LRU_FILE; + struct lrugen *lrugen = &lruvec->evictable; + + lockdep_assert_held(&lruvec->lru_lock); + VM_BUG_ON(old_gen != -1 && old_gen >= MAX_NR_GENS); + VM_BUG_ON(new_gen != -1 && new_gen >= MAX_NR_GENS); + VM_BUG_ON(old_gen == -1 && new_gen == -1); + + if (old_gen >= 0) + WRITE_ONCE(lrugen->sizes[old_gen][type][zone], + lrugen->sizes[old_gen][type][zone] - delta); + if (new_gen >= 0) + WRITE_ONCE(lrugen->sizes[new_gen][type][zone], + lrugen->sizes[new_gen][type][zone] + delta); + + if (old_gen < 0) { + if (lru_gen_is_active(lruvec, new_gen)) + lru += LRU_ACTIVE; + update_lru_size(lruvec, lru, zone, delta); + return; + } + + if (new_gen < 0) { + if (lru_gen_is_active(lruvec, old_gen)) + lru += LRU_ACTIVE; + update_lru_size(lruvec, lru, zone, -delta); + return; + } + + if (!lru_gen_is_active(lruvec, old_gen) && lru_gen_is_active(lruvec, new_gen)) { + update_lru_size(lruvec, lru, zone, -delta); + update_lru_size(lruvec, lru + LRU_ACTIVE, zone, delta); + } + + VM_BUG_ON(lru_gen_is_active(lruvec, old_gen) && !lru_gen_is_active(lruvec, new_gen)); +} + +/* Add a page to one of the multigenerational lru lists. Return true on success. */ +static inline bool lru_gen_addition(struct page *page, struct lruvec *lruvec, bool front) +{ + int gen; + unsigned long old_flags, new_flags; + int type = page_is_file_lru(page); + int zone = page_zonenum(page); + struct lrugen *lrugen = &lruvec->evictable; + + if (PageUnevictable(page) || !lrugen->enabled[type]) + return false; + /* + * If a page is being faulted in, add it to the youngest generation. + * try_walk_mm_list() may look at the size of the youngest generation to + * determine if the aging is due. + * + * If a page can't be evicted immediately, i.e., an anon page not in + * swap cache, a dirty file page under reclaim, or a page rejected by + * evict_pages() due to races, dirty buffer heads, etc., add it to the + * second oldest generation. + * + * If a page could be evicted immediately, i.e., a clean file page, add + * it to the oldest generation. + */ + if (PageActive(page)) + gen = lru_gen_from_seq(lrugen->max_seq); + else if ((!type && !PageSwapCache(page)) || + (PageReclaim(page) && (PageDirty(page) || PageWriteback(page))) || + (!PageReferenced(page) && PageWorkingset(page))) + gen = lru_gen_from_seq(lrugen->min_seq[type] + 1); + else + gen = lru_gen_from_seq(lrugen->min_seq[type]); + + do { + old_flags = READ_ONCE(page->flags); + VM_BUG_ON_PAGE(old_flags & LRU_GEN_MASK, page); + + new_flags = (old_flags & ~(LRU_GEN_MASK | BIT(PG_active))) | + ((gen + 1UL) << LRU_GEN_PGOFF); + /* see the comment in evict_pages() */ + if (!(old_flags & BIT(PG_referenced))) + new_flags &= ~(LRU_USAGE_MASK | LRU_TIER_FLAGS); + } while (cmpxchg(&page->flags, old_flags, new_flags) != old_flags); + + lru_gen_update_size(page, lruvec, -1, gen); + if (front) + list_add(&page->lru, &lrugen->lists[gen][type][zone]); + else + list_add_tail(&page->lru, &lrugen->lists[gen][type][zone]); + + return true; +} + +/* Delete a page from one of the multigenerational lru lists. Return true on success. */ +static inline bool lru_gen_deletion(struct page *page, struct lruvec *lruvec) +{ + int gen; + unsigned long old_flags, new_flags; + + do { + old_flags = READ_ONCE(page->flags); + if (!(old_flags & LRU_GEN_MASK)) + return false; + + VM_BUG_ON_PAGE(PageActive(page), page); + VM_BUG_ON_PAGE(PageUnevictable(page), page); + + gen = ((old_flags & LRU_GEN_MASK) >> LRU_GEN_PGOFF) - 1; + + new_flags = old_flags & ~LRU_GEN_MASK; + /* mark page active accordingly */ + if (lru_gen_is_active(lruvec, gen)) + new_flags |= BIT(PG_active); + } while (cmpxchg(&page->flags, old_flags, new_flags) != old_flags); + + lru_gen_update_size(page, lruvec, gen, -1); + list_del(&page->lru); + + return true; +} + +/* Return the level of usage of a page. See the comment on MAX_NR_TIERS. */ +static inline int page_tier_usage(struct page *page) +{ + unsigned long flags = READ_ONCE(page->flags); + + return flags & BIT(PG_workingset) ? + ((flags & LRU_USAGE_MASK) >> LRU_USAGE_PGOFF) + 1 : 0; +} + +/* Increment the usage counter after a page is accessed via file descriptors. */ +static inline void page_inc_usage(struct page *page) +{ + unsigned long usage; + unsigned long old_flags, new_flags; + + do { + old_flags = READ_ONCE(page->flags); + + if (!(old_flags & BIT(PG_workingset))) { + new_flags = old_flags | BIT(PG_workingset); + continue; + } + + usage = (old_flags & LRU_USAGE_MASK) + BIT(LRU_USAGE_PGOFF); + + new_flags = (old_flags & ~LRU_USAGE_MASK) | min(usage, LRU_USAGE_MASK); + } while (new_flags != old_flags && + cmpxchg(&page->flags, old_flags, new_flags) != old_flags); +} + +#else /* CONFIG_LRU_GEN */ + +static inline bool lru_gen_enabled(void) +{ + return false; +} + +static inline bool lru_gen_addition(struct page *page, struct lruvec *lruvec, bool front) +{ + return false; +} + +static inline bool lru_gen_deletion(struct page *page, struct lruvec *lruvec) +{ + return false; +} + +static inline void page_inc_usage(struct page *page) +{ +} + +#endif /* CONFIG_LRU_GEN */ + static __always_inline void add_page_to_lru_list(struct page *page, struct lruvec *lruvec) { enum lru_list lru = page_lru(page); + if (lru_gen_addition(page, lruvec, true)) + return; + update_lru_size(lruvec, lru, page_zonenum(page), thp_nr_pages(page)); list_add(&page->lru, &lruvec->lists[lru]); } @@ -93,6 +321,9 @@ static __always_inline void add_page_to_lru_list_tail(struct page *page, { enum lru_list lru = page_lru(page); + if (lru_gen_addition(page, lruvec, false)) + return; + update_lru_size(lruvec, lru, page_zonenum(page), thp_nr_pages(page)); list_add_tail(&page->lru, &lruvec->lists[lru]); } @@ -100,6 +331,9 @@ static __always_inline void add_page_to_lru_list_tail(struct page *page, static __always_inline void del_page_from_lru_list(struct page *page, struct lruvec *lruvec) { + if (lru_gen_deletion(page, lruvec)) + return; + list_del(&page->lru); update_lru_size(lruvec, page_lru(page), page_zonenum(page), -thp_nr_pages(page)); diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 8f0fb62e897..b5d9e99945f 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -15,6 +15,8 @@ #include #include #include +#include +#include #include @@ -279,6 +281,7 @@ struct vm_region { unsigned long vm_top; /* region allocated to here */ unsigned long vm_pgoff; /* the offset in vm_file corresponding to vm_start */ struct file *vm_file; /* the backing file or NULL */ + struct file *vm_prfile; /* the virtual backing file or NULL */ int vm_usage; /* region usage count (access under nommu_region_sem) */ bool vm_icache_flushed : 1; /* true if the icache has been flushed for @@ -358,6 +361,7 @@ struct vm_area_struct { unsigned long vm_pgoff; /* Offset (within vm_file) in PAGE_SIZE units */ struct file * vm_file; /* File we map to (can be NULL). */ + struct file *vm_prfile; /* shadow of vm_file */ void * vm_private_data; /* was vm_pte (shared mem) */ #ifdef CONFIG_SWAP @@ -370,6 +374,9 @@ struct vm_area_struct { struct mempolicy *vm_policy; /* NUMA policy for the VMA */ #endif struct vm_userfaultfd_ctx vm_userfaultfd_ctx; +#ifdef CONFIG_UKSM + struct vma_slot *uksm_vma_slot; +#endif } __randomize_layout; struct core_thread { @@ -574,6 +581,22 @@ struct mm_struct { #ifdef CONFIG_IOMMU_SUPPORT u32 pasid; +#endif +#ifdef CONFIG_LRU_GEN + struct { + /* the node of a global or per-memcg mm_struct list */ + struct list_head list; +#ifdef CONFIG_MEMCG + /* points to the memcg of the owner task above */ + struct mem_cgroup *memcg; +#endif + /* whether this mm_struct has been used since the last walk */ + nodemask_t nodes; +#ifndef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + /* the number of CPUs using this mm_struct */ + atomic_t nr_cpus; +#endif + } lrugen; #endif } __randomize_layout; @@ -601,6 +624,95 @@ static inline cpumask_t *mm_cpumask(struct mm_struct *mm) return (struct cpumask *)&mm->cpu_bitmap; } +#ifdef CONFIG_LRU_GEN + +void lru_gen_init_mm(struct mm_struct *mm); +void lru_gen_add_mm(struct mm_struct *mm); +void lru_gen_del_mm(struct mm_struct *mm); +#ifdef CONFIG_MEMCG +int lru_gen_alloc_mm_list(struct mem_cgroup *memcg); +void lru_gen_free_mm_list(struct mem_cgroup *memcg); +void lru_gen_migrate_mm(struct mm_struct *mm); +#endif + +/* Track the usage of each mm_struct so that we can skip inactive ones. */ +static inline void lru_gen_switch_mm(struct mm_struct *old, struct mm_struct *new) +{ + /* exclude init_mm, efi_mm, etc. */ + if (!core_kernel_data((unsigned long)old)) { + VM_BUG_ON(old == &init_mm); + + nodes_setall(old->lrugen.nodes); +#ifndef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + atomic_dec(&old->lrugen.nr_cpus); + VM_BUG_ON_MM(atomic_read(&old->lrugen.nr_cpus) < 0, old); +#endif + } else + VM_BUG_ON_MM(READ_ONCE(old->lrugen.list.prev) || + READ_ONCE(old->lrugen.list.next), old); + + if (!core_kernel_data((unsigned long)new)) { + VM_BUG_ON(new == &init_mm); + +#ifndef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + atomic_inc(&new->lrugen.nr_cpus); + VM_BUG_ON_MM(atomic_read(&new->lrugen.nr_cpus) < 0, new); +#endif + } else + VM_BUG_ON_MM(READ_ONCE(new->lrugen.list.prev) || + READ_ONCE(new->lrugen.list.next), new); +} + +/* Return whether this mm_struct is being used on any CPUs. */ +static inline bool lru_gen_mm_is_active(struct mm_struct *mm) +{ +#ifdef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + return !cpumask_empty(mm_cpumask(mm)); +#else + return atomic_read(&mm->lrugen.nr_cpus); +#endif +} + +#else /* CONFIG_LRU_GEN */ + +static inline void lru_gen_init_mm(struct mm_struct *mm) +{ +} + +static inline void lru_gen_add_mm(struct mm_struct *mm) +{ +} + +static inline void lru_gen_del_mm(struct mm_struct *mm) +{ +} + +#ifdef CONFIG_MEMCG +static inline int lru_gen_alloc_mm_list(struct mem_cgroup *memcg) +{ + return 0; +} + +static inline void lru_gen_free_mm_list(struct mem_cgroup *memcg) +{ +} + +static inline void lru_gen_migrate_mm(struct mm_struct *mm) +{ +} +#endif + +static inline void lru_gen_switch_mm(struct mm_struct *old, struct mm_struct *new) +{ +} + +static inline bool lru_gen_mm_is_active(struct mm_struct *mm) +{ + return false; +} + +#endif /* CONFIG_LRU_GEN */ + struct mmu_gather; extern void tlb_gather_mmu(struct mmu_gather *tlb, struct mm_struct *mm); extern void tlb_gather_mmu_fullmm(struct mmu_gather *tlb, struct mm_struct *mm); diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 0d53eba1c38..1b1dad8ccb0 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h @@ -157,6 +157,9 @@ enum zone_stat_item { NR_ZSPAGES, /* allocated in zsmalloc */ #endif NR_FREE_CMA_PAGES, +#ifdef CONFIG_UKSM + NR_UKSM_ZERO_PAGES, +#endif NR_VM_ZONE_STAT_ITEMS }; enum node_stat_item { @@ -293,6 +296,114 @@ enum lruvec_flags { */ }; +struct lruvec; +struct page_vma_mapped_walk; + +#define LRU_GEN_MASK ((BIT(LRU_GEN_WIDTH) - 1) << LRU_GEN_PGOFF) +#define LRU_USAGE_MASK ((BIT(LRU_USAGE_WIDTH) - 1) << LRU_USAGE_PGOFF) + +#ifdef CONFIG_LRU_GEN + +/* + * For each lruvec, evictable pages are divided into multiple generations. The + * youngest and the oldest generation numbers, AKA max_seq and min_seq, are + * monotonically increasing. The sliding window technique is used to track at + * most MAX_NR_GENS and at least MIN_NR_GENS generations. An offset within the + * window, AKA gen, indexes an array of per-type and per-zone lists for the + * corresponding generation. The counter in page->flags stores gen+1 while a + * page is on one of the multigenerational lru lists. Otherwise, it stores 0. + */ +#define MAX_NR_GENS ((unsigned int)CONFIG_NR_LRU_GENS) + +/* + * Each generation is then divided into multiple tiers. Tiers represent levels + * of usage from file descriptors, i.e., mark_page_accessed(). In contrast to + * moving across generations which requires the lru lock, moving across tiers + * only involves an atomic operation on page->flags and therefore has a + * negligible cost. + * + * The purposes of tiers are to: + * 1) estimate whether pages accessed multiple times via file descriptors are + * more active than pages accessed only via page tables by separating the two + * access types into upper tiers and the base tier and comparing refault rates + * across tiers. + * 2) improve buffered io performance by deferring activations of pages + * accessed multiple times until the eviction. That is activations happen in + * the reclaim path, not the access path. + * + * Pages accessed N times via file descriptors belong to tier order_base_2(N). + * The base tier uses the following page flag: + * !PageReferenced() -- readahead pages + * PageReferenced() -- single-access pages + * All upper tiers use the following page flags: + * PageReferenced() && PageWorkingset() -- multi-access pages + * in addition to the bits storing N-2 accesses. Therefore, we can support one + * upper tier without using additional bits in page->flags. + * + * Note that + * 1) PageWorkingset() is always set for upper tiers because we want to + * maintain the existing psi behavior. + * 2) !PageReferenced() && PageWorkingset() is not a valid tier. See the + * comment in evict_pages(). + * + * Pages from the base tier are evicted regardless of its refault rate. Pages + * from upper tiers will be moved to the next generation, if their refault rates + * are higher than that of the base tier. + */ +#define MAX_NR_TIERS ((unsigned int)CONFIG_TIERS_PER_GEN) +#define LRU_TIER_FLAGS (BIT(PG_referenced) | BIT(PG_workingset)) +#define LRU_USAGE_SHIFT (CONFIG_TIERS_PER_GEN - 1) + +/* Whether to keep historical stats for each generation. */ +#ifdef CONFIG_LRU_GEN_STATS +#define NR_STAT_GENS ((unsigned int)CONFIG_NR_LRU_GENS) +#else +#define NR_STAT_GENS 1U +#endif + +struct lrugen { + /* the aging increments the max generation number */ + unsigned long max_seq; + /* the eviction increments the min generation numbers */ + unsigned long min_seq[ANON_AND_FILE]; + /* the birth time of each generation in jiffies */ + unsigned long timestamps[MAX_NR_GENS]; + /* the multigenerational lru lists */ + struct list_head lists[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES]; + /* the sizes of the multigenerational lru lists in pages */ + unsigned long sizes[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES]; + /* to determine which type and its tiers to evict */ + atomic_long_t evicted[NR_STAT_GENS][ANON_AND_FILE][MAX_NR_TIERS]; + atomic_long_t refaulted[NR_STAT_GENS][ANON_AND_FILE][MAX_NR_TIERS]; + /* the base tier won't be activated */ + unsigned long activated[NR_STAT_GENS][ANON_AND_FILE][MAX_NR_TIERS - 1]; + /* arithmetic mean weighted by geometric series 1/2, 1/4, ... */ + unsigned long avg_total[ANON_AND_FILE][MAX_NR_TIERS]; + unsigned long avg_refaulted[ANON_AND_FILE][MAX_NR_TIERS]; + /* whether the multigenerational lru is enabled */ + bool enabled[ANON_AND_FILE]; +}; + +void lru_gen_init_lruvec(struct lruvec *lruvec); +void lru_gen_set_state(bool enable, bool main, bool swap); +void lru_gen_scan_around(struct page_vma_mapped_walk *pvmw); + +#else /* CONFIG_LRU_GEN */ + +static inline void lru_gen_init_lruvec(struct lruvec *lruvec) +{ +} + +static inline void lru_gen_set_state(bool enable, bool main, bool swap) +{ +} + +static inline void lru_gen_scan_around(struct page_vma_mapped_walk *pvmw) +{ +} + +#endif /* CONFIG_LRU_GEN */ + struct lruvec { struct list_head lists[NR_LRU_LISTS]; /* per lruvec lru_lock for memcg */ @@ -310,6 +421,10 @@ struct lruvec { unsigned long refaults[ANON_AND_FILE]; /* Various lruvec state flags (enum lruvec_flags) */ unsigned long flags; +#ifdef CONFIG_LRU_GEN + /* unevictable pages are on LRU_UNEVICTABLE */ + struct lrugen evictable; +#endif #ifdef CONFIG_MEMCG struct pglist_data *pgdat; #endif @@ -751,6 +866,8 @@ struct deferred_split { }; #endif +struct mm_walk_args; + /* * On NUMA machines, each NUMA node would have a pg_data_t to describe * it's memory layout. On UMA machines there is a single pglist_data which @@ -856,6 +973,9 @@ typedef struct pglist_data { unsigned long flags; +#ifdef CONFIG_LRU_GEN + struct mm_walk_args *mm_walk_args; +#endif ZONE_PADDING(_pad2_) /* Per-node vmstats */ diff --git a/include/linux/mnt_namespace.h b/include/linux/mnt_namespace.h index 8f882f5881e..6b9808f0984 100644 --- a/include/linux/mnt_namespace.h +++ b/include/linux/mnt_namespace.h @@ -7,12 +7,15 @@ struct mnt_namespace; struct fs_struct; struct user_namespace; struct ns_common; +struct vfsmount; extern struct mnt_namespace *copy_mnt_ns(unsigned long, struct mnt_namespace *, struct user_namespace *, struct fs_struct *); extern void put_mnt_ns(struct mnt_namespace *ns); extern struct ns_common *from_mnt_ns(struct mnt_namespace *); +extern int is_current_mnt_ns(struct vfsmount *mnt); + extern const struct file_operations proc_mounts_operations; extern const struct file_operations proc_mountinfo_operations; extern const struct file_operations proc_mountstats_operations; diff --git a/include/linux/nodemask.h b/include/linux/nodemask.h index ac398e143c9..89fe4e3592f 100644 --- a/include/linux/nodemask.h +++ b/include/linux/nodemask.h @@ -486,6 +486,7 @@ static inline int num_node_state(enum node_states state) #define first_online_node 0 #define first_memory_node 0 #define next_online_node(nid) (MAX_NUMNODES) +#define next_memory_node(nid) (MAX_NUMNODES) #define nr_node_ids 1U #define nr_online_nodes 1U diff --git a/include/linux/page-flags-layout.h b/include/linux/page-flags-layout.h index ef1e3e736e1..ce8d5732a3a 100644 --- a/include/linux/page-flags-layout.h +++ b/include/linux/page-flags-layout.h @@ -26,6 +26,14 @@ #define ZONES_WIDTH ZONES_SHIFT +#ifdef CONFIG_LRU_GEN +/* LRU_GEN_WIDTH is generated from order_base_2(CONFIG_NR_LRU_GENS + 1). */ +#define LRU_USAGE_WIDTH (CONFIG_TIERS_PER_GEN - 2) +#else +#define LRU_GEN_WIDTH 0 +#define LRU_USAGE_WIDTH 0 +#endif + #ifdef CONFIG_SPARSEMEM #include #define SECTIONS_SHIFT (MAX_PHYSMEM_BITS - SECTION_SIZE_BITS) @@ -55,7 +63,8 @@ #define SECTIONS_WIDTH 0 #endif -#if ZONES_WIDTH + SECTIONS_WIDTH + NODES_SHIFT <= BITS_PER_LONG - NR_PAGEFLAGS +#if ZONES_WIDTH + LRU_GEN_WIDTH + LRU_USAGE_WIDTH + SECTIONS_WIDTH + NODES_SHIFT \ + <= BITS_PER_LONG - NR_PAGEFLAGS #define NODES_WIDTH NODES_SHIFT #elif defined(CONFIG_SPARSEMEM_VMEMMAP) #error "Vmemmap: No space for nodes field in page flags" @@ -89,8 +98,8 @@ #define LAST_CPUPID_SHIFT 0 #endif -#if ZONES_WIDTH + SECTIONS_WIDTH + NODES_WIDTH + KASAN_TAG_WIDTH + LAST_CPUPID_SHIFT \ - <= BITS_PER_LONG - NR_PAGEFLAGS +#if ZONES_WIDTH + LRU_GEN_WIDTH + LRU_USAGE_WIDTH + SECTIONS_WIDTH + NODES_WIDTH + \ + KASAN_TAG_WIDTH + LAST_CPUPID_SHIFT <= BITS_PER_LONG - NR_PAGEFLAGS #define LAST_CPUPID_WIDTH LAST_CPUPID_SHIFT #else #define LAST_CPUPID_WIDTH 0 @@ -100,8 +109,8 @@ #define LAST_CPUPID_NOT_IN_PAGE_FLAGS #endif -#if ZONES_WIDTH + SECTIONS_WIDTH + NODES_WIDTH + KASAN_TAG_WIDTH + LAST_CPUPID_WIDTH \ - > BITS_PER_LONG - NR_PAGEFLAGS +#if ZONES_WIDTH + LRU_GEN_WIDTH + LRU_USAGE_WIDTH + SECTIONS_WIDTH + NODES_WIDTH + \ + KASAN_TAG_WIDTH + LAST_CPUPID_WIDTH > BITS_PER_LONG - NR_PAGEFLAGS #error "Not enough bits in page flags" #endif diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index 04a34c08e0a..e58984fca32 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -817,7 +817,7 @@ static inline void ClearPageSlabPfmemalloc(struct page *page) 1UL << PG_private | 1UL << PG_private_2 | \ 1UL << PG_writeback | 1UL << PG_reserved | \ 1UL << PG_slab | 1UL << PG_active | \ - 1UL << PG_unevictable | __PG_MLOCKED) + 1UL << PG_unevictable | __PG_MLOCKED | LRU_GEN_MASK) /* * Flags checked when a page is prepped for return by the page allocator. @@ -828,7 +828,7 @@ static inline void ClearPageSlabPfmemalloc(struct page *page) * alloc-free cycle to prevent from reusing the page. */ #define PAGE_FLAGS_CHECK_AT_PREP \ - (((1UL << NR_PAGEFLAGS) - 1) & ~__PG_HWPOISON) + ((((1UL << NR_PAGEFLAGS) - 1) & ~__PG_HWPOISON) | LRU_GEN_MASK | LRU_USAGE_MASK) #define PAGE_FLAGS_PRIVATE \ (1UL << PG_private | 1UL << PG_private_2) diff --git a/include/linux/pgtable.h b/include/linux/pgtable.h index c32600c9e1a..164c8091f24 100644 --- a/include/linux/pgtable.h +++ b/include/linux/pgtable.h @@ -193,7 +193,7 @@ static inline int ptep_test_and_clear_young(struct vm_area_struct *vma, #endif #ifndef __HAVE_ARCH_PMDP_TEST_AND_CLEAR_YOUNG -#ifdef CONFIG_TRANSPARENT_HUGEPAGE +#if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG) static inline int pmdp_test_and_clear_young(struct vm_area_struct *vma, unsigned long address, pmd_t *pmdp) @@ -214,7 +214,7 @@ static inline int pmdp_test_and_clear_young(struct vm_area_struct *vma, BUILD_BUG(); return 0; } -#endif /* CONFIG_TRANSPARENT_HUGEPAGE */ +#endif /* CONFIG_TRANSPARENT_HUGEPAGE || CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG */ #endif #ifndef __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH @@ -1119,13 +1119,26 @@ extern void untrack_pfn(struct vm_area_struct *vma, unsigned long pfn, extern void untrack_pfn_moved(struct vm_area_struct *vma); #endif +#ifdef CONFIG_UKSM +static inline int is_uksm_zero_pfn(unsigned long pfn) +{ + extern unsigned long uksm_zero_pfn; + return pfn == uksm_zero_pfn; +} +#else +static inline int is_uksm_zero_pfn(unsigned long pfn) +{ + return 0; +} +#endif + #ifdef CONFIG_MMU #ifdef __HAVE_COLOR_ZERO_PAGE static inline int is_zero_pfn(unsigned long pfn) { extern unsigned long zero_pfn; unsigned long offset_from_zero_pfn = pfn - zero_pfn; - return offset_from_zero_pfn <= (zero_page_mask >> PAGE_SHIFT); + return offset_from_zero_pfn <= (zero_page_mask >> PAGE_SHIFT) || is_uksm_zero_pfn(pfn); } #define my_zero_pfn(addr) page_to_pfn(ZERO_PAGE(addr)) @@ -1134,7 +1147,7 @@ static inline int is_zero_pfn(unsigned long pfn) static inline int is_zero_pfn(unsigned long pfn) { extern unsigned long zero_pfn; - return pfn == zero_pfn; + return (pfn == zero_pfn) || (is_uksm_zero_pfn(pfn)); } static inline unsigned long my_zero_pfn(unsigned long addr) diff --git a/include/linux/sched.h b/include/linux/sched.h index 32813c34511..35654025752 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -36,6 +36,10 @@ #include #include +#ifdef CONFIG_SCHED_MUQSS +#include +#endif + /* task_struct member predeclarations (sorted alphabetically): */ struct audit_context; struct backing_dev_info; @@ -223,6 +227,34 @@ extern long schedule_timeout_interruptible(long timeout); extern long schedule_timeout_killable(long timeout); extern long schedule_timeout_uninterruptible(long timeout); extern long schedule_timeout_idle(long timeout); + +#if defined(CONFIG_HIGH_RES_TIMERS) && defined(CONFIG_SCHED_MUQSS) +extern long schedule_msec_hrtimeout(long timeout); +extern long schedule_min_hrtimeout(void); +extern long schedule_msec_hrtimeout_interruptible(long timeout); +extern long schedule_msec_hrtimeout_uninterruptible(long timeout); +#else +static inline long schedule_msec_hrtimeout(long timeout) +{ + return schedule_timeout(msecs_to_jiffies(timeout)); +} + +static inline long schedule_min_hrtimeout(void) +{ + return schedule_timeout(1); +} + +static inline long schedule_msec_hrtimeout_interruptible(long timeout) +{ + return schedule_timeout_interruptible(msecs_to_jiffies(timeout)); +} + +static inline long schedule_msec_hrtimeout_uninterruptible(long timeout) +{ + return schedule_timeout_uninterruptible(msecs_to_jiffies(timeout)); +} +#endif + asmlinkage void schedule(void); extern void schedule_preempt_disabled(void); asmlinkage void preempt_schedule_irq(void); @@ -677,8 +709,10 @@ struct task_struct { unsigned int flags; unsigned int ptrace; -#ifdef CONFIG_SMP +#if defined(CONFIG_SMP) || defined(CONFIG_SCHED_MUQSS) int on_cpu; +#endif +#ifdef CONFIG_SMP struct __call_single_node wake_entry; #ifdef CONFIG_THREAD_INFO_IN_TASK /* Current CPU: */ @@ -704,10 +738,25 @@ struct task_struct { int static_prio; int normal_prio; unsigned int rt_priority; +#ifdef CONFIG_SCHED_MUQSS + int time_slice; + u64 deadline; + skiplist_node node; /* Skip list node */ + u64 last_ran; + u64 sched_time; /* sched_clock time spent running */ +#ifdef CONFIG_SMT_NICE + int smt_bias; /* Policy/nice level bias across smt siblings */ +#endif +#ifdef CONFIG_HOTPLUG_CPU + bool zerobound; /* Bound to CPU0 for hotplug */ +#endif + unsigned long rt_timeout; +#else /* CONFIG_SCHED_MUQSS */ const struct sched_class *sched_class; struct sched_entity se; struct sched_rt_entity rt; +#endif #ifdef CONFIG_CGROUP_SCHED struct task_group *sched_task_group; #endif @@ -915,6 +964,10 @@ struct task_struct { #ifdef CONFIG_ARCH_HAS_SCALED_CPUTIME u64 utimescaled; u64 stimescaled; +#endif +#ifdef CONFIG_SCHED_MUQSS + /* Unbanked cpu time */ + unsigned long utime_ns, stime_ns; #endif u64 gtime; struct prev_cputime prev_cputime; @@ -1407,6 +1460,40 @@ struct task_struct { */ }; +#ifdef CONFIG_SCHED_MUQSS +#define tsk_seruntime(t) ((t)->sched_time) +#define tsk_rttimeout(t) ((t)->rt_timeout) + +static inline void tsk_cpus_current(struct task_struct *p) +{ +} + +void print_scheduler_version(void); + +static inline bool iso_task(struct task_struct *p) +{ + return (p->policy == SCHED_ISO); +} +#else /* CFS */ +#define tsk_seruntime(t) ((t)->se.sum_exec_runtime) +#define tsk_rttimeout(t) ((t)->rt.timeout) + +static inline void tsk_cpus_current(struct task_struct *p) +{ + p->nr_cpus_allowed = current->nr_cpus_allowed; +} + +static inline void print_scheduler_version(void) +{ + printk(KERN_INFO "CFS CPU scheduler.\n"); +} + +static inline bool iso_task(struct task_struct *p) +{ + return false; +} +#endif /* CONFIG_SCHED_MUQSS */ + static inline struct pid *task_pid(struct task_struct *task) { return task->thread_pid; diff --git a/include/linux/sched/deadline.h b/include/linux/sched/deadline.h index 1aff00b65f3..73d6319a856 100644 --- a/include/linux/sched/deadline.h +++ b/include/linux/sched/deadline.h @@ -28,7 +28,16 @@ static inline bool dl_time_before(u64 a, u64 b) #ifdef CONFIG_SMP struct root_domain; +#ifdef CONFIG_SCHED_MUQSS +static inline void dl_clear_root_domain(struct root_domain *rd) +{ +} +static inline void dl_add_task_root_domain(struct task_struct *p) +{ +} +#else /* CONFIG_SCHED_MUQSS */ extern void dl_add_task_root_domain(struct task_struct *p); extern void dl_clear_root_domain(struct root_domain *rd); +#endif /* CONFIG_SCHED_MUQSS */ #endif /* CONFIG_SMP */ diff --git a/include/linux/sched/nohz.h b/include/linux/sched/nohz.h index 6d67e9a5af6..101fe470aa8 100644 --- a/include/linux/sched/nohz.h +++ b/include/linux/sched/nohz.h @@ -13,7 +13,7 @@ extern int get_nohz_timer_target(void); static inline void nohz_balance_enter_idle(int cpu) { } #endif -#ifdef CONFIG_NO_HZ_COMMON +#if defined(CONFIG_NO_HZ_COMMON) && !defined(CONFIG_SCHED_MUQSS) void calc_load_nohz_start(void); void calc_load_nohz_remote(struct rq *rq); void calc_load_nohz_stop(void); diff --git a/include/linux/sched/prio.h b/include/linux/sched/prio.h index ab83d85e118..0e1898e247c 100644 --- a/include/linux/sched/prio.h +++ b/include/linux/sched/prio.h @@ -14,6 +14,12 @@ */ #define MAX_RT_PRIO 100 +#ifdef CONFIG_SCHED_MUQSS +#define ISO_PRIO (MAX_RT_PRIO) +#define NORMAL_PRIO (MAX_RT_PRIO + 1) +#define IDLE_PRIO (MAX_RT_PRIO + 2) +#define PRIO_LIMIT ((IDLE_PRIO) + 1) +#endif /* CONFIG_SCHED_MUQSS */ #define MAX_PRIO (MAX_RT_PRIO + NICE_WIDTH) #define DEFAULT_PRIO (MAX_RT_PRIO + NICE_WIDTH / 2) diff --git a/include/linux/sched/rt.h b/include/linux/sched/rt.h index e5af028c08b..010b2244e0b 100644 --- a/include/linux/sched/rt.h +++ b/include/linux/sched/rt.h @@ -24,8 +24,10 @@ static inline bool task_is_realtime(struct task_struct *tsk) if (policy == SCHED_FIFO || policy == SCHED_RR) return true; +#ifndef CONFIG_SCHED_MUQSS if (policy == SCHED_DEADLINE) return true; +#endif return false; } diff --git a/include/linux/sched/task.h b/include/linux/sched/task.h index ef02be869cf..9e504da356c 100644 --- a/include/linux/sched/task.h +++ b/include/linux/sched/task.h @@ -93,7 +93,7 @@ int kernel_wait(pid_t pid, int *stat); extern void free_task(struct task_struct *tsk); /* sched_exec is called by processes performing an exec */ -#ifdef CONFIG_SMP +#if defined(CONFIG_SMP) && !defined(CONFIG_SCHED_MUQSS) extern void sched_exec(void); #else #define sched_exec() {} diff --git a/include/linux/skip_list.h b/include/linux/skip_list.h new file mode 100644 index 00000000000..d4be84ba273 --- /dev/null +++ b/include/linux/skip_list.h @@ -0,0 +1,33 @@ +#ifndef _LINUX_SKIP_LISTS_H +#define _LINUX_SKIP_LISTS_H +typedef u64 keyType; +typedef void *valueType; + +typedef struct nodeStructure skiplist_node; + +struct nodeStructure { + int level; /* Levels in this structure */ + keyType key; + valueType value; + skiplist_node *next[8]; + skiplist_node *prev[8]; +}; + +typedef struct listStructure { + int entries; + int level; /* Maximum level of the list + (1 more than the number of levels in the list) */ + skiplist_node *header; /* pointer to header */ +} skiplist; + +void skiplist_init(skiplist_node *slnode); +skiplist *new_skiplist(skiplist_node *slnode); +void free_skiplist(skiplist *l); +void skiplist_node_init(skiplist_node *node); +void skiplist_insert(skiplist *l, skiplist_node *node, keyType key, valueType value, unsigned int randseed); +void skiplist_delete(skiplist *l, skiplist_node *node); + +static inline bool skiplist_node_empty(skiplist_node *node) { + return (!node->next[0]); +} +#endif /* _LINUX_SKIP_LISTS_H */ diff --git a/include/linux/splice.h b/include/linux/splice.h index a55179fd60f..8e21c53cf88 100644 --- a/include/linux/splice.h +++ b/include/linux/splice.h @@ -93,4 +93,10 @@ extern void splice_shrink_spd(struct splice_pipe_desc *); extern const struct pipe_buf_operations page_cache_pipe_buf_ops; extern const struct pipe_buf_operations default_pipe_buf_ops; + +extern long do_splice_from(struct pipe_inode_info *pipe, struct file *out, + loff_t *ppos, size_t len, unsigned int flags); +extern long do_splice_to(struct file *in, loff_t *ppos, + struct pipe_inode_info *pipe, size_t len, + unsigned int flags); #endif diff --git a/include/linux/sradix-tree.h b/include/linux/sradix-tree.h new file mode 100644 index 00000000000..d71edba6b63 --- /dev/null +++ b/include/linux/sradix-tree.h @@ -0,0 +1,77 @@ +#ifndef _LINUX_SRADIX_TREE_H +#define _LINUX_SRADIX_TREE_H + + +#define INIT_SRADIX_TREE(root, mask) \ +do { \ + (root)->height = 0; \ + (root)->gfp_mask = (mask); \ + (root)->rnode = NULL; \ +} while (0) + +#define ULONG_BITS (sizeof(unsigned long) * 8) +#define SRADIX_TREE_INDEX_BITS (8 /* CHAR_BIT */ * sizeof(unsigned long)) +//#define SRADIX_TREE_MAP_SHIFT 6 +//#define SRADIX_TREE_MAP_SIZE (1UL << SRADIX_TREE_MAP_SHIFT) +//#define SRADIX_TREE_MAP_MASK (SRADIX_TREE_MAP_SIZE-1) + +struct sradix_tree_node { + unsigned int height; /* Height from the bottom */ + unsigned int count; + unsigned int fulls; /* Number of full sublevel trees */ + struct sradix_tree_node *parent; + void *stores[0]; +}; + +/* A simple radix tree implementation */ +struct sradix_tree_root { + unsigned int height; + struct sradix_tree_node *rnode; + + /* Where found to have available empty stores in its sublevels */ + struct sradix_tree_node *enter_node; + unsigned int shift; + unsigned int stores_size; + unsigned int mask; + unsigned long min; /* The first hole index */ + unsigned long num; + //unsigned long *height_to_maxindex; + + /* How the node is allocated and freed. */ + struct sradix_tree_node *(*alloc)(void); + void (*free)(struct sradix_tree_node *node); + + /* When a new node is added and removed */ + void (*extend)(struct sradix_tree_node *parent, struct sradix_tree_node *child); + void (*assign)(struct sradix_tree_node *node, unsigned int index, void *item); + void (*rm)(struct sradix_tree_node *node, unsigned int offset); +}; + +struct sradix_tree_path { + struct sradix_tree_node *node; + int offset; +}; + +static inline +void init_sradix_tree_root(struct sradix_tree_root *root, unsigned long shift) +{ + root->height = 0; + root->rnode = NULL; + root->shift = shift; + root->stores_size = 1UL << shift; + root->mask = root->stores_size - 1; +} + + +extern void *sradix_tree_next(struct sradix_tree_root *root, + struct sradix_tree_node *node, unsigned long index, + int (*iter)(void *, unsigned long)); + +extern int sradix_tree_enter(struct sradix_tree_root *root, void **item, int num); + +extern void sradix_tree_delete_from_leaf(struct sradix_tree_root *root, + struct sradix_tree_node *node, unsigned long index); + +extern void *sradix_tree_lookup(struct sradix_tree_root *root, unsigned long index); + +#endif /* _LINUX_SRADIX_TREE_H */ diff --git a/include/linux/swap.h b/include/linux/swap.h index 144727041e7..30b1f15f5c6 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h @@ -365,8 +365,8 @@ extern void deactivate_page(struct page *page); extern void mark_page_lazyfree(struct page *page); extern void swap_setup(void); -extern void lru_cache_add_inactive_or_unevictable(struct page *page, - struct vm_area_struct *vma); +extern void lru_cache_add_page_vma(struct page *page, struct vm_area_struct *vma, + bool faulting); /* linux/mm/vmscan.c */ extern unsigned long zone_reclaimable_pages(struct zone *zone); diff --git a/include/linux/syscalls.h b/include/linux/syscalls.h index 050511e8f1f..a0a9748b023 100644 --- a/include/linux/syscalls.h +++ b/include/linux/syscalls.h @@ -71,6 +71,8 @@ struct open_how; struct mount_attr; struct landlock_ruleset_attr; enum landlock_rule_type; +struct futex_waitv; +struct futex_requeue; #include #include @@ -623,6 +625,20 @@ asmlinkage long sys_get_robust_list(int pid, asmlinkage long sys_set_robust_list(struct robust_list_head __user *head, size_t len); +/* kernel/futex2.c */ +asmlinkage long sys_futex_wait(void __user *uaddr, unsigned int val, + unsigned int flags, + struct __kernel_timespec __user *timo); +asmlinkage long sys_futex_wake(void __user *uaddr, unsigned int nr_wake, + unsigned int flags); +asmlinkage long sys_futex_waitv(struct futex_waitv __user *waiters, + unsigned int nr_futexes, unsigned int flags, + struct __kernel_timespec __user *timo); +asmlinkage long sys_futex_requeue(struct futex_requeue __user *uaddr1, + struct futex_requeue __user *uaddr2, + unsigned int nr_wake, unsigned int nr_requeue, + unsigned int cmpval, unsigned int flags); + /* kernel/hrtimer.c */ asmlinkage long sys_nanosleep(struct __kernel_timespec __user *rqtp, struct __kernel_timespec __user *rmtp); @@ -1309,6 +1325,8 @@ int ksys_ipc(unsigned int call, int first, unsigned long second, unsigned long third, void __user * ptr, long fifth); int compat_ksys_ipc(u32 call, int first, int second, u32 third, u32 ptr, u32 fifth); +long ksys_futex_wake(void __user *uaddr, unsigned long nr_wake, + unsigned int flags); /* * The following kernel syscall equivalents are just wrappers to fs-internal diff --git a/include/linux/tcp.h b/include/linux/tcp.h index 48d8a363319..1bd559c69e8 100644 --- a/include/linux/tcp.h +++ b/include/linux/tcp.h @@ -225,7 +225,8 @@ struct tcp_sock { u8 compressed_ack; u8 dup_ack_counter:2, tlp_retrans:1, /* TLP is a retransmission */ - unused:5; + fast_ack_mode:2, /* which fast ack mode ? */ + unused:3; u32 chrono_start; /* Start time in jiffies of a TCP chrono */ u32 chrono_stat[3]; /* Time in jiffies for chrono_stat stats */ u8 chrono_type:2, /* current chronograph type */ diff --git a/include/linux/uksm.h b/include/linux/uksm.h new file mode 100644 index 00000000000..bb8651f534f --- /dev/null +++ b/include/linux/uksm.h @@ -0,0 +1,149 @@ +#ifndef __LINUX_UKSM_H +#define __LINUX_UKSM_H +/* + * Memory merging support. + * + * This code enables dynamic sharing of identical pages found in different + * memory areas, even if they are not shared by fork(). + */ + +/* if !CONFIG_UKSM this file should not be compiled at all. */ +#ifdef CONFIG_UKSM + +#include +#include +#include +#include +#include + +extern unsigned long zero_pfn __read_mostly; +extern unsigned long uksm_zero_pfn __read_mostly; +extern struct page *empty_uksm_zero_page; + +/* must be done before linked to mm */ +extern void uksm_vma_add_new(struct vm_area_struct *vma); +extern void uksm_remove_vma(struct vm_area_struct *vma); + +#define UKSM_SLOT_NEED_SORT (1 << 0) +#define UKSM_SLOT_NEED_RERAND (1 << 1) +#define UKSM_SLOT_SCANNED (1 << 2) /* It's scanned in this round */ +#define UKSM_SLOT_FUL_SCANNED (1 << 3) +#define UKSM_SLOT_IN_UKSM (1 << 4) + +struct vma_slot { + struct sradix_tree_node *snode; + unsigned long sindex; + + struct list_head slot_list; + unsigned long fully_scanned_round; + unsigned long dedup_num; + unsigned long pages_scanned; + unsigned long this_sampled; + unsigned long last_scanned; + unsigned long pages_to_scan; + struct scan_rung *rung; + struct page **rmap_list_pool; + unsigned int *pool_counts; + unsigned long pool_size; + struct vm_area_struct *vma; + struct mm_struct *mm; + unsigned long ctime_j; + unsigned long pages; + unsigned long flags; + unsigned long pages_cowed; /* pages cowed this round */ + unsigned long pages_merged; /* pages merged this round */ + unsigned long pages_bemerged; + + /* when it has page merged in this eval round */ + struct list_head dedup_list; +}; + +static inline void uksm_unmap_zero_page(pte_t pte) +{ + if (pte_pfn(pte) == uksm_zero_pfn) + __dec_zone_page_state(empty_uksm_zero_page, NR_UKSM_ZERO_PAGES); +} + +static inline void uksm_map_zero_page(pte_t pte) +{ + if (pte_pfn(pte) == uksm_zero_pfn) + __inc_zone_page_state(empty_uksm_zero_page, NR_UKSM_ZERO_PAGES); +} + +static inline void uksm_cow_page(struct vm_area_struct *vma, struct page *page) +{ + if (vma->uksm_vma_slot && PageKsm(page)) + vma->uksm_vma_slot->pages_cowed++; +} + +static inline void uksm_cow_pte(struct vm_area_struct *vma, pte_t pte) +{ + if (vma->uksm_vma_slot && pte_pfn(pte) == uksm_zero_pfn) + vma->uksm_vma_slot->pages_cowed++; +} + +static inline int uksm_flags_can_scan(unsigned long vm_flags) +{ +#ifdef VM_SAO + if (vm_flags & VM_SAO) + return 0; +#endif + + return !(vm_flags & (VM_PFNMAP | VM_IO | VM_DONTEXPAND | + VM_HUGETLB | VM_MIXEDMAP | VM_SHARED + | VM_MAYSHARE | VM_GROWSUP | VM_GROWSDOWN)); +} + +static inline void uksm_vm_flags_mod(unsigned long *vm_flags_p) +{ + if (uksm_flags_can_scan(*vm_flags_p)) + *vm_flags_p |= VM_MERGEABLE; +} + +/* + * Just a wrapper for BUG_ON for where ksm_zeropage must not be. TODO: it will + * be removed when uksm zero page patch is stable enough. + */ +static inline void uksm_bugon_zeropage(pte_t pte) +{ + BUG_ON(pte_pfn(pte) == uksm_zero_pfn); +} +#else +static inline void uksm_vma_add_new(struct vm_area_struct *vma) +{ +} + +static inline void uksm_remove_vma(struct vm_area_struct *vma) +{ +} + +static inline void uksm_unmap_zero_page(pte_t pte) +{ +} + +static inline void uksm_map_zero_page(pte_t pte) +{ +} + +static inline void uksm_cow_page(struct vm_area_struct *vma, struct page *page) +{ +} + +static inline void uksm_cow_pte(struct vm_area_struct *vma, pte_t pte) +{ +} + +static inline int uksm_flags_can_scan(unsigned long vm_flags) +{ + return 0; +} + +static inline void uksm_vm_flags_mod(unsigned long *vm_flags_p) +{ +} + +static inline void uksm_bugon_zeropage(pte_t pte) +{ +} +#endif /* !CONFIG_UKSM */ +#endif /* __LINUX_UKSM_H */ diff --git a/include/linux/user_namespace.h b/include/linux/user_namespace.h index bfa6463f8a9..27aa2bd5e50 100644 --- a/include/linux/user_namespace.h +++ b/include/linux/user_namespace.h @@ -116,6 +116,8 @@ void put_ucounts(struct ucounts *ucounts); #ifdef CONFIG_USER_NS +extern int unprivileged_userns_clone; + static inline struct user_namespace *get_user_ns(struct user_namespace *ns) { if (ns) @@ -149,6 +151,8 @@ extern bool current_in_userns(const struct user_namespace *target_ns); struct ns_common *ns_get_owner(struct ns_common *ns); #else +#define unprivileged_userns_clone 0 + static inline struct user_namespace *get_user_ns(struct user_namespace *ns) { return &init_user_ns; diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h index 3c8c59471bc..2cdc5a0709f 100644 --- a/include/net/inet_connection_sock.h +++ b/include/net/inet_connection_sock.h @@ -134,8 +134,9 @@ struct inet_connection_sock { u32 icsk_probes_tstamp; u32 icsk_user_timeout; - u64 icsk_ca_priv[104 / sizeof(u64)]; -#define ICSK_CA_PRIV_SIZE (13 * sizeof(u64)) +/* XXX inflated by temporary internal debugging info */ +#define ICSK_CA_PRIV_SIZE (216) + u64 icsk_ca_priv[ICSK_CA_PRIV_SIZE / sizeof(u64)]; }; #define ICSK_TIME_RETRANS 1 /* Retransmit timer */ diff --git a/include/net/tcp.h b/include/net/tcp.h index b42b3e6731e..d3ba0435c73 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -803,6 +803,11 @@ static inline u32 tcp_stamp_us_delta(u64 t1, u64 t0) return max_t(s64, t1 - t0, 0); } +static inline u32 tcp_stamp32_us_delta(u32 t1, u32 t0) +{ + return max_t(s32, t1 - t0, 0); +} + static inline u32 tcp_skb_timestamp(const struct sk_buff *skb) { return tcp_ns_to_ts(skb->skb_mstamp_ns); @@ -870,16 +875,22 @@ struct tcp_skb_cb { __u32 ack_seq; /* Sequence number ACK'd */ union { struct { +#define TCPCB_DELIVERED_CE_MASK ((1U<<20) - 1) /* There is space for up to 24 bytes */ - __u32 in_flight:30,/* Bytes in flight at transmit */ - is_app_limited:1, /* cwnd not fully used? */ - unused:1; + __u32 is_app_limited:1, /* cwnd not fully used? */ + delivered_ce:20, + unused:11; /* pkts S/ACKed so far upon tx of skb, incl retrans: */ __u32 delivered; /* start of send pipeline phase */ - u64 first_tx_mstamp; + u32 first_tx_mstamp; /* when we reached the "delivered" count */ - u64 delivered_mstamp; + u32 delivered_mstamp; +#define TCPCB_IN_FLIGHT_BITS 20 +#define TCPCB_IN_FLIGHT_MAX ((1U << TCPCB_IN_FLIGHT_BITS) - 1) + u32 in_flight:20, /* packets in flight at transmit */ + unused2:12; + u32 lost; /* packets lost so far upon tx of skb */ } tx; /* only used for outgoing skbs */ union { struct inet_skb_parm h4; @@ -1004,7 +1015,11 @@ enum tcp_ca_ack_event_flags { #define TCP_CONG_NON_RESTRICTED 0x1 /* Requires ECN/ECT set on all packets */ #define TCP_CONG_NEEDS_ECN 0x2 -#define TCP_CONG_MASK (TCP_CONG_NON_RESTRICTED | TCP_CONG_NEEDS_ECN) +/* Wants notification of CE events (CA_EVENT_ECN_IS_CE, CA_EVENT_ECN_NO_CE). */ +#define TCP_CONG_WANTS_CE_EVENTS 0x4 +#define TCP_CONG_MASK (TCP_CONG_NON_RESTRICTED | \ + TCP_CONG_NEEDS_ECN | \ + TCP_CONG_WANTS_CE_EVENTS) union tcp_cc_info; @@ -1024,8 +1039,13 @@ struct ack_sample { */ struct rate_sample { u64 prior_mstamp; /* starting timestamp for interval */ + u32 prior_lost; /* tp->lost at "prior_mstamp" */ u32 prior_delivered; /* tp->delivered at "prior_mstamp" */ + u32 prior_delivered_ce;/* tp->delivered_ce at "prior_mstamp" */ + u32 tx_in_flight; /* packets in flight at starting timestamp */ + s32 lost; /* number of packets lost over interval */ s32 delivered; /* number of packets delivered over interval */ + s32 delivered_ce; /* packets delivered w/ CE mark over interval */ long interval_us; /* time for tp->delivered to incr "delivered" */ u32 snd_interval_us; /* snd interval for delivered packets */ u32 rcv_interval_us; /* rcv interval for delivered packets */ @@ -1036,6 +1056,7 @@ struct rate_sample { bool is_app_limited; /* is sample from packet with bubble in pipe? */ bool is_retrans; /* is sample from retransmission? */ bool is_ack_delayed; /* is this (likely) a delayed ACK? */ + bool is_ece; /* did this ACK have ECN marked? */ }; struct tcp_congestion_ops { @@ -1059,8 +1080,11 @@ struct tcp_congestion_ops { /* hook for packet ack accounting (optional) */ void (*pkts_acked)(struct sock *sk, const struct ack_sample *sample); - /* override sysctl_tcp_min_tso_segs */ - u32 (*min_tso_segs)(struct sock *sk); + /* pick target number of segments per TSO/GSO skb (optional): */ + u32 (*tso_segs)(struct sock *sk, unsigned int mss_now); + + /* react to a specific lost skb (optional) */ + void (*skb_marked_lost)(struct sock *sk, const struct sk_buff *skb); /* call when packets are delivered to update cwnd and pacing rate, * after all the ca_state processing. (optional) @@ -1123,6 +1147,14 @@ static inline char *tcp_ca_get_name_by_key(u32 key, char *buffer) } #endif +static inline bool tcp_ca_wants_ce_events(const struct sock *sk) +{ + const struct inet_connection_sock *icsk = inet_csk(sk); + + return icsk->icsk_ca_ops->flags & (TCP_CONG_NEEDS_ECN | + TCP_CONG_WANTS_CE_EVENTS); +} + static inline bool tcp_ca_needs_ecn(const struct sock *sk) { const struct inet_connection_sock *icsk = inet_csk(sk); @@ -1148,6 +1180,7 @@ static inline void tcp_ca_event(struct sock *sk, const enum tcp_ca_event event) } /* From tcp_rate.c */ +void tcp_set_tx_in_flight(struct sock *sk, struct sk_buff *skb); void tcp_rate_skb_sent(struct sock *sk, struct sk_buff *skb); void tcp_rate_skb_delivered(struct sock *sk, struct sk_buff *skb, struct rate_sample *rs); diff --git a/include/uapi/asm-generic/unistd.h b/include/uapi/asm-generic/unistd.h index d2a942086fc..c0d552f660a 100644 --- a/include/uapi/asm-generic/unistd.h +++ b/include/uapi/asm-generic/unistd.h @@ -872,8 +872,20 @@ __SYSCALL(__NR_landlock_add_rule, sys_landlock_add_rule) #define __NR_landlock_restrict_self 446 __SYSCALL(__NR_landlock_restrict_self, sys_landlock_restrict_self) +#define __NR_futex_wait 443 +__SYSCALL(__NR_futex_wait, sys_futex_wait) + +#define __NR_futex_wake 444 +__SYSCALL(__NR_futex_wake, sys_futex_wake) + +#define __NR_futex_waitv 445 +__SC_COMP(__NR_futex_waitv, sys_futex_waitv, compat_sys_futex_waitv) + +#define __NR_futex_requeue 446 +__SC_COMP(__NR_futex_requeue, sys_futex_requeue, compat_sys_futex_requeue) + #undef __NR_syscalls -#define __NR_syscalls 447 +#define __NR_syscalls 451 /* * 32 bit systems traditionally used different diff --git a/include/uapi/linux/Kbuild b/include/uapi/linux/Kbuild new file mode 100644 index 00000000000..e1b2bd375ea --- /dev/null +++ b/include/uapi/linux/Kbuild @@ -0,0 +1 @@ +header-y = aufs_type.h diff --git a/include/uapi/linux/aufs_type.h b/include/uapi/linux/aufs_type.h new file mode 100644 index 00000000000..fa04f623f38 --- /dev/null +++ b/include/uapi/linux/aufs_type.h @@ -0,0 +1,452 @@ +/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */ +/* + * Copyright (C) 2005-2020 Junjiro R. Okajima + * + * This program, aufs is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License as published by + * the Free Software Foundation; either version 2 of the License, or + * (at your option) any later version. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program. If not, see . + */ + +#ifndef __AUFS_TYPE_H__ +#define __AUFS_TYPE_H__ + +#define AUFS_NAME "aufs" + +#ifdef __KERNEL__ +/* + * define it before including all other headers. + * sched.h may use pr_* macros before defining "current", so define the + * no-current version first, and re-define later. + */ +#define pr_fmt(fmt) AUFS_NAME " %s:%d: " fmt, __func__, __LINE__ +#include +#undef pr_fmt +#define pr_fmt(fmt) \ + AUFS_NAME " %s:%d:%.*s[%d]: " fmt, __func__, __LINE__, \ + (int)sizeof(current->comm), current->comm, current->pid +#include +#else +#include +#include +#include +#endif /* __KERNEL__ */ + +#define AUFS_VERSION "5.13-20210705" + +/* todo? move this to linux-2.6.19/include/magic.h */ +#define AUFS_SUPER_MAGIC ('a' << 24 | 'u' << 16 | 'f' << 8 | 's') + +/* ---------------------------------------------------------------------- */ + +#ifdef __KERNEL__ +#ifdef CONFIG_AUFS_BRANCH_MAX_127 +typedef int8_t aufs_bindex_t; +#define AUFS_BRANCH_MAX 127 +#else +typedef int16_t aufs_bindex_t; +#ifdef CONFIG_AUFS_BRANCH_MAX_511 +#define AUFS_BRANCH_MAX 511 +#elif defined(CONFIG_AUFS_BRANCH_MAX_1023) +#define AUFS_BRANCH_MAX 1023 +#elif defined(CONFIG_AUFS_BRANCH_MAX_32767) +#define AUFS_BRANCH_MAX 32767 +#endif +#endif + +#ifndef AUFS_BRANCH_MAX +#error unknown CONFIG_AUFS_BRANCH_MAX value +#endif +#endif /* __KERNEL__ */ + +/* ---------------------------------------------------------------------- */ + +#define AUFS_FSTYPE AUFS_NAME + +#define AUFS_ROOT_INO 2 +#define AUFS_FIRST_INO 11 + +#define AUFS_WH_PFX ".wh." +#define AUFS_WH_PFX_LEN ((int)sizeof(AUFS_WH_PFX) - 1) +#define AUFS_WH_TMP_LEN 4 +/* a limit for rmdir/rename a dir and copyup */ +#define AUFS_MAX_NAMELEN (NAME_MAX \ + - AUFS_WH_PFX_LEN * 2 /* doubly whiteouted */\ + - 1 /* dot */\ + - AUFS_WH_TMP_LEN) /* hex */ +#define AUFS_XINO_FNAME "." AUFS_NAME ".xino" +#define AUFS_XINO_DEFPATH "/tmp/" AUFS_XINO_FNAME +#define AUFS_XINO_DEF_SEC 30 /* seconds */ +#define AUFS_XINO_DEF_TRUNC 45 /* percentage */ +#define AUFS_DIRWH_DEF 3 +#define AUFS_RDCACHE_DEF 10 /* seconds */ +#define AUFS_RDCACHE_MAX 3600 /* seconds */ +#define AUFS_RDBLK_DEF 512 /* bytes */ +#define AUFS_RDHASH_DEF 32 +#define AUFS_WKQ_NAME AUFS_NAME "d" +#define AUFS_MFS_DEF_SEC 30 /* seconds */ +#define AUFS_MFS_MAX_SEC 3600 /* seconds */ +#define AUFS_FHSM_CACHE_DEF_SEC 30 /* seconds */ +#define AUFS_PLINK_WARN 50 /* number of plinks in a single bucket */ + +/* pseudo-link maintenace under /proc */ +#define AUFS_PLINK_MAINT_NAME "plink_maint" +#define AUFS_PLINK_MAINT_DIR "fs/" AUFS_NAME +#define AUFS_PLINK_MAINT_PATH AUFS_PLINK_MAINT_DIR "/" AUFS_PLINK_MAINT_NAME + +/* dirren, renamed dir */ +#define AUFS_DR_INFO_PFX AUFS_WH_PFX ".dr." +#define AUFS_DR_BRHINO_NAME AUFS_WH_PFX "hino" +/* whiteouted doubly */ +#define AUFS_WH_DR_INFO_PFX AUFS_WH_PFX AUFS_DR_INFO_PFX +#define AUFS_WH_DR_BRHINO AUFS_WH_PFX AUFS_DR_BRHINO_NAME + +#define AUFS_DIROPQ_NAME AUFS_WH_PFX ".opq" /* whiteouted doubly */ +#define AUFS_WH_DIROPQ AUFS_WH_PFX AUFS_DIROPQ_NAME + +#define AUFS_BASE_NAME AUFS_WH_PFX AUFS_NAME +#define AUFS_PLINKDIR_NAME AUFS_WH_PFX "plnk" +#define AUFS_ORPHDIR_NAME AUFS_WH_PFX "orph" + +/* doubly whiteouted */ +#define AUFS_WH_BASE AUFS_WH_PFX AUFS_BASE_NAME +#define AUFS_WH_PLINKDIR AUFS_WH_PFX AUFS_PLINKDIR_NAME +#define AUFS_WH_ORPHDIR AUFS_WH_PFX AUFS_ORPHDIR_NAME + +/* branch permissions and attributes */ +#define AUFS_BRPERM_RW "rw" +#define AUFS_BRPERM_RO "ro" +#define AUFS_BRPERM_RR "rr" +#define AUFS_BRATTR_COO_REG "coo_reg" +#define AUFS_BRATTR_COO_ALL "coo_all" +#define AUFS_BRATTR_FHSM "fhsm" +#define AUFS_BRATTR_UNPIN "unpin" +#define AUFS_BRATTR_ICEX "icex" +#define AUFS_BRATTR_ICEX_SEC "icexsec" +#define AUFS_BRATTR_ICEX_SYS "icexsys" +#define AUFS_BRATTR_ICEX_TR "icextr" +#define AUFS_BRATTR_ICEX_USR "icexusr" +#define AUFS_BRATTR_ICEX_OTH "icexoth" +#define AUFS_BRRATTR_WH "wh" +#define AUFS_BRWATTR_NLWH "nolwh" +#define AUFS_BRWATTR_MOO "moo" + +#define AuBrPerm_RW 1 /* writable, hardlinkable wh */ +#define AuBrPerm_RO (1 << 1) /* readonly */ +#define AuBrPerm_RR (1 << 2) /* natively readonly */ +#define AuBrPerm_Mask (AuBrPerm_RW | AuBrPerm_RO | AuBrPerm_RR) + +#define AuBrAttr_COO_REG (1 << 3) /* copy-up on open */ +#define AuBrAttr_COO_ALL (1 << 4) +#define AuBrAttr_COO_Mask (AuBrAttr_COO_REG | AuBrAttr_COO_ALL) + +#define AuBrAttr_FHSM (1 << 5) /* file-based hsm */ +#define AuBrAttr_UNPIN (1 << 6) /* rename-able top dir of + branch. meaningless since + linux-3.18-rc1 */ + +/* ignore error in copying XATTR */ +#define AuBrAttr_ICEX_SEC (1 << 7) +#define AuBrAttr_ICEX_SYS (1 << 8) +#define AuBrAttr_ICEX_TR (1 << 9) +#define AuBrAttr_ICEX_USR (1 << 10) +#define AuBrAttr_ICEX_OTH (1 << 11) +#define AuBrAttr_ICEX (AuBrAttr_ICEX_SEC \ + | AuBrAttr_ICEX_SYS \ + | AuBrAttr_ICEX_TR \ + | AuBrAttr_ICEX_USR \ + | AuBrAttr_ICEX_OTH) + +#define AuBrRAttr_WH (1 << 12) /* whiteout-able */ +#define AuBrRAttr_Mask AuBrRAttr_WH + +#define AuBrWAttr_NoLinkWH (1 << 13) /* un-hardlinkable whiteouts */ +#define AuBrWAttr_MOO (1 << 14) /* move-up on open */ +#define AuBrWAttr_Mask (AuBrWAttr_NoLinkWH | AuBrWAttr_MOO) + +#define AuBrAttr_CMOO_Mask (AuBrAttr_COO_Mask | AuBrWAttr_MOO) + +/* #warning test userspace */ +#ifdef __KERNEL__ +#ifndef CONFIG_AUFS_FHSM +#undef AuBrAttr_FHSM +#define AuBrAttr_FHSM 0 +#endif +#ifndef CONFIG_AUFS_XATTR +#undef AuBrAttr_ICEX +#define AuBrAttr_ICEX 0 +#undef AuBrAttr_ICEX_SEC +#define AuBrAttr_ICEX_SEC 0 +#undef AuBrAttr_ICEX_SYS +#define AuBrAttr_ICEX_SYS 0 +#undef AuBrAttr_ICEX_TR +#define AuBrAttr_ICEX_TR 0 +#undef AuBrAttr_ICEX_USR +#define AuBrAttr_ICEX_USR 0 +#undef AuBrAttr_ICEX_OTH +#define AuBrAttr_ICEX_OTH 0 +#endif +#endif + +/* the longest combination */ +/* AUFS_BRATTR_ICEX and AUFS_BRATTR_ICEX_TR don't affect here */ +#define AuBrPermStrSz sizeof(AUFS_BRPERM_RW \ + "+" AUFS_BRATTR_COO_REG \ + "+" AUFS_BRATTR_FHSM \ + "+" AUFS_BRATTR_UNPIN \ + "+" AUFS_BRATTR_ICEX_SEC \ + "+" AUFS_BRATTR_ICEX_SYS \ + "+" AUFS_BRATTR_ICEX_USR \ + "+" AUFS_BRATTR_ICEX_OTH \ + "+" AUFS_BRWATTR_NLWH) + +typedef struct { + char a[AuBrPermStrSz]; +} au_br_perm_str_t; + +static inline int au_br_writable(int brperm) +{ + return brperm & AuBrPerm_RW; +} + +static inline int au_br_whable(int brperm) +{ + return brperm & (AuBrPerm_RW | AuBrRAttr_WH); +} + +static inline int au_br_wh_linkable(int brperm) +{ + return !(brperm & AuBrWAttr_NoLinkWH); +} + +static inline int au_br_cmoo(int brperm) +{ + return brperm & AuBrAttr_CMOO_Mask; +} + +static inline int au_br_fhsm(int brperm) +{ + return brperm & AuBrAttr_FHSM; +} + +/* ---------------------------------------------------------------------- */ + +/* ioctl */ +enum { + /* readdir in userspace */ + AuCtl_RDU, + AuCtl_RDU_INO, + + AuCtl_WBR_FD, /* pathconf wrapper */ + AuCtl_IBUSY, /* busy inode */ + AuCtl_MVDOWN, /* move-down */ + AuCtl_BR, /* info about branches */ + AuCtl_FHSM_FD /* connection for fhsm */ +}; + +/* borrowed from linux/include/linux/kernel.h */ +#ifndef ALIGN +#ifdef _GNU_SOURCE +#define ALIGN(x, a) __ALIGN_MASK(x, (typeof(x))(a)-1) +#else +#define ALIGN(x, a) (((x) + (a) - 1) & ~((a) - 1)) +#endif +#define __ALIGN_MASK(x, mask) (((x)+(mask))&~(mask)) +#endif + +/* borrowed from linux/include/linux/compiler-gcc3.h */ +#ifndef __aligned +#define __aligned(x) __attribute__((aligned(x))) +#endif + +#ifdef __KERNEL__ +#ifndef __packed +#define __packed __attribute__((packed)) +#endif +#endif + +struct au_rdu_cookie { + uint64_t h_pos; + int16_t bindex; + uint8_t flags; + uint8_t pad; + uint32_t generation; +} __aligned(8); + +struct au_rdu_ent { + uint64_t ino; + int16_t bindex; + uint8_t type; + uint8_t nlen; + uint8_t wh; + char name[]; +} __aligned(8); + +static inline int au_rdu_len(int nlen) +{ + /* include the terminating NULL */ + return ALIGN(sizeof(struct au_rdu_ent) + nlen + 1, + sizeof(uint64_t)); +} + +union au_rdu_ent_ul { + struct au_rdu_ent __user *e; + uint64_t ul; +}; + +enum { + AufsCtlRduV_SZ, + AufsCtlRduV_End +}; + +struct aufs_rdu { + /* input */ + union { + uint64_t sz; /* AuCtl_RDU */ + uint64_t nent; /* AuCtl_RDU_INO */ + }; + union au_rdu_ent_ul ent; + uint16_t verify[AufsCtlRduV_End]; + + /* input/output */ + uint32_t blk; + + /* output */ + union au_rdu_ent_ul tail; + /* number of entries which were added in a single call */ + uint64_t rent; + uint8_t full; + uint8_t shwh; + + struct au_rdu_cookie cookie; +} __aligned(8); + +/* ---------------------------------------------------------------------- */ + +/* dirren. the branch is identified by the filename who contains this */ +struct au_drinfo { + uint64_t ino; + union { + uint8_t oldnamelen; + uint64_t _padding; + }; + uint8_t oldname[]; +} __aligned(8); + +struct au_drinfo_fdata { + uint32_t magic; + struct au_drinfo drinfo; +} __aligned(8); + +#define AUFS_DRINFO_MAGIC_V1 ('a' << 24 | 'd' << 16 | 'r' << 8 | 0x01) +/* future */ +#define AUFS_DRINFO_MAGIC_V2 ('a' << 24 | 'd' << 16 | 'r' << 8 | 0x02) + +/* ---------------------------------------------------------------------- */ + +struct aufs_wbr_fd { + uint32_t oflags; + int16_t brid; +} __aligned(8); + +/* ---------------------------------------------------------------------- */ + +struct aufs_ibusy { + uint64_t ino, h_ino; + int16_t bindex; +} __aligned(8); + +/* ---------------------------------------------------------------------- */ + +/* error code for move-down */ +/* the actual message strings are implemented in aufs-util.git */ +enum { + EAU_MVDOWN_OPAQUE = 1, + EAU_MVDOWN_WHITEOUT, + EAU_MVDOWN_UPPER, + EAU_MVDOWN_BOTTOM, + EAU_MVDOWN_NOUPPER, + EAU_MVDOWN_NOLOWERBR, + EAU_Last +}; + +/* flags for move-down */ +#define AUFS_MVDOWN_DMSG 1 +#define AUFS_MVDOWN_OWLOWER (1 << 1) /* overwrite lower */ +#define AUFS_MVDOWN_KUPPER (1 << 2) /* keep upper */ +#define AUFS_MVDOWN_ROLOWER (1 << 3) /* do even if lower is RO */ +#define AUFS_MVDOWN_ROLOWER_R (1 << 4) /* did on lower RO */ +#define AUFS_MVDOWN_ROUPPER (1 << 5) /* do even if upper is RO */ +#define AUFS_MVDOWN_ROUPPER_R (1 << 6) /* did on upper RO */ +#define AUFS_MVDOWN_BRID_UPPER (1 << 7) /* upper brid */ +#define AUFS_MVDOWN_BRID_LOWER (1 << 8) /* lower brid */ +#define AUFS_MVDOWN_FHSM_LOWER (1 << 9) /* find fhsm attr for lower */ +#define AUFS_MVDOWN_STFS (1 << 10) /* req. stfs */ +#define AUFS_MVDOWN_STFS_FAILED (1 << 11) /* output: stfs is unusable */ +#define AUFS_MVDOWN_BOTTOM (1 << 12) /* output: no more lowers */ + +/* index for move-down */ +enum { + AUFS_MVDOWN_UPPER, + AUFS_MVDOWN_LOWER, + AUFS_MVDOWN_NARRAY +}; + +/* + * additional info of move-down + * number of free blocks and inodes. + * subset of struct kstatfs, but smaller and always 64bit. + */ +struct aufs_stfs { + uint64_t f_blocks; + uint64_t f_bavail; + uint64_t f_files; + uint64_t f_ffree; +}; + +struct aufs_stbr { + int16_t brid; /* optional input */ + int16_t bindex; /* output */ + struct aufs_stfs stfs; /* output when AUFS_MVDOWN_STFS set */ +} __aligned(8); + +struct aufs_mvdown { + uint32_t flags; /* input/output */ + struct aufs_stbr stbr[AUFS_MVDOWN_NARRAY]; /* input/output */ + int8_t au_errno; /* output */ +} __aligned(8); + +/* ---------------------------------------------------------------------- */ + +union aufs_brinfo { + /* PATH_MAX may differ between kernel-space and user-space */ + char _spacer[4096]; + struct { + int16_t id; + int perm; + char path[]; + }; +} __aligned(8); + +/* ---------------------------------------------------------------------- */ + +#define AuCtlType 'A' +#define AUFS_CTL_RDU _IOWR(AuCtlType, AuCtl_RDU, struct aufs_rdu) +#define AUFS_CTL_RDU_INO _IOWR(AuCtlType, AuCtl_RDU_INO, struct aufs_rdu) +#define AUFS_CTL_WBR_FD _IOW(AuCtlType, AuCtl_WBR_FD, \ + struct aufs_wbr_fd) +#define AUFS_CTL_IBUSY _IOWR(AuCtlType, AuCtl_IBUSY, struct aufs_ibusy) +#define AUFS_CTL_MVDOWN _IOWR(AuCtlType, AuCtl_MVDOWN, \ + struct aufs_mvdown) +#define AUFS_CTL_BRINFO _IOW(AuCtlType, AuCtl_BR, union aufs_brinfo) +#define AUFS_CTL_FHSM_FD _IOW(AuCtlType, AuCtl_FHSM_FD, int) + +#endif /* __AUFS_TYPE_H__ */ diff --git a/include/uapi/linux/futex.h b/include/uapi/linux/futex.h index a89eb0accd5..afc3245e572 100644 --- a/include/uapi/linux/futex.h +++ b/include/uapi/linux/futex.h @@ -21,6 +21,7 @@ #define FUTEX_WAKE_BITSET 10 #define FUTEX_WAIT_REQUEUE_PI 11 #define FUTEX_CMP_REQUEUE_PI 12 +#define FUTEX_WAIT_MULTIPLE 31 #define FUTEX_PRIVATE_FLAG 128 #define FUTEX_CLOCK_REALTIME 256 @@ -40,6 +41,39 @@ FUTEX_PRIVATE_FLAG) #define FUTEX_CMP_REQUEUE_PI_PRIVATE (FUTEX_CMP_REQUEUE_PI | \ FUTEX_PRIVATE_FLAG) +#define FUTEX_WAIT_MULTIPLE_PRIVATE (FUTEX_WAIT_MULTIPLE | \ + FUTEX_PRIVATE_FLAG) + +/* Size argument to futex2 syscall */ +#define FUTEX_32 2 + +#define FUTEX_SIZE_MASK 0x3 + +#define FUTEX_SHARED_FLAG 8 + +#define FUTEX_WAITV_MAX 128 + +/** + * struct futex_waitv - A waiter for vectorized wait + * @uaddr: User address to wait on + * @val: Expected value at uaddr + * @flags: Flags for this waiter + */ +struct futex_waitv { + void __user *uaddr; + unsigned int val; + unsigned int flags; +}; + +/** + * struct futex_requeue - Define an address and its flags for requeue operation + * @uaddr: User address of one of the requeue arguments + * @flags: Flags for this address + */ +struct futex_requeue { + void __user *uaddr; + unsigned int flags; +}; /* * Support for robust futexes: the kernel cleans up held futexes at @@ -150,4 +184,21 @@ struct robust_list_head { (((op & 0xf) << 28) | ((cmp & 0xf) << 24) \ | ((oparg & 0xfff) << 12) | (cmparg & 0xfff)) +/* + * Maximum number of multiple futexes to wait for + */ +#define FUTEX_MULTIPLE_MAX_COUNT 128 + +/** + * struct futex_wait_block - Block of futexes to be waited for + * @uaddr: User address of the futex + * @val: Futex value expected by userspace + * @bitset: Bitset for the optional bitmasked wakeup + */ +struct futex_wait_block { + __u32 __user *uaddr; + __u32 val; + __u32 bitset; +}; + #endif /* _UAPI_LINUX_FUTEX_H */ diff --git a/include/uapi/linux/inet_diag.h b/include/uapi/linux/inet_diag.h index 20ee93f0f87..96d52dd9c48 100644 --- a/include/uapi/linux/inet_diag.h +++ b/include/uapi/linux/inet_diag.h @@ -231,9 +231,42 @@ struct tcp_bbr_info { __u32 bbr_cwnd_gain; /* cwnd gain shifted left 8 bits */ }; +/* Phase as reported in netlink/ss stats. */ +enum tcp_bbr2_phase { + BBR2_PHASE_INVALID = 0, + BBR2_PHASE_STARTUP = 1, + BBR2_PHASE_DRAIN = 2, + BBR2_PHASE_PROBE_RTT = 3, + BBR2_PHASE_PROBE_BW_UP = 4, + BBR2_PHASE_PROBE_BW_DOWN = 5, + BBR2_PHASE_PROBE_BW_CRUISE = 6, + BBR2_PHASE_PROBE_BW_REFILL = 7 +}; + +struct tcp_bbr2_info { + /* u64 bw: bandwidth (app throughput) estimate in Byte per sec: */ + __u32 bbr_bw_lsb; /* lower 32 bits of bw */ + __u32 bbr_bw_msb; /* upper 32 bits of bw */ + __u32 bbr_min_rtt; /* min-filtered RTT in uSec */ + __u32 bbr_pacing_gain; /* pacing gain shifted left 8 bits */ + __u32 bbr_cwnd_gain; /* cwnd gain shifted left 8 bits */ + __u32 bbr_bw_hi_lsb; /* lower 32 bits of bw_hi */ + __u32 bbr_bw_hi_msb; /* upper 32 bits of bw_hi */ + __u32 bbr_bw_lo_lsb; /* lower 32 bits of bw_lo */ + __u32 bbr_bw_lo_msb; /* upper 32 bits of bw_lo */ + __u8 bbr_mode; /* current bbr_mode in state machine */ + __u8 bbr_phase; /* current state machine phase */ + __u8 unused1; /* alignment padding; not used yet */ + __u8 bbr_version; /* MUST be at this offset in struct */ + __u32 bbr_inflight_lo; /* lower/short-term data volume bound */ + __u32 bbr_inflight_hi; /* higher/long-term data volume bound */ + __u32 bbr_extra_acked; /* max excess packets ACKed in epoch */ +}; + union tcp_cc_info { struct tcpvegas_info vegas; struct tcp_dctcp_info dctcp; struct tcp_bbr_info bbr; + struct tcp_bbr2_info bbr2; }; #endif /* _UAPI_INET_DIAG_H_ */ diff --git a/include/uapi/linux/sched.h b/include/uapi/linux/sched.h index 3bac0a8ceab..f48c5c5da65 100644 --- a/include/uapi/linux/sched.h +++ b/include/uapi/linux/sched.h @@ -115,9 +115,16 @@ struct clone_args { #define SCHED_FIFO 1 #define SCHED_RR 2 #define SCHED_BATCH 3 -/* SCHED_ISO: reserved but not implemented yet */ +/* SCHED_ISO: Implemented on MuQSS only */ #define SCHED_IDLE 5 +#ifdef CONFIG_SCHED_MUQSS +#define SCHED_ISO 4 +#define SCHED_IDLEPRIO SCHED_IDLE +#define SCHED_MAX (SCHED_IDLEPRIO) +#define SCHED_RANGE(policy) ((policy) <= SCHED_MAX) +#else /* CONFIG_SCHED_MUQSS */ #define SCHED_DEADLINE 6 +#endif /* CONFIG_SCHED_MUQSS */ /* Can be ORed in to make sure the process is reverted back to SCHED_NORMAL on fork */ #define SCHED_RESET_ON_FORK 0x40000000 diff --git a/init/Kconfig b/init/Kconfig index a61c92066c2..4ea15478834 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -104,6 +104,48 @@ config THREAD_INFO_IN_TASK menu "General setup" +config SCHED_MUQSS + bool "MuQSS cpu scheduler" + select HIGH_RES_TIMERS + default n + help + The Multiple Queue Skiplist Scheduler for excellent interactivity and + responsiveness on the desktop and highly scalable deterministic + low latency on any hardware. + +config PCK_INTERACTIVE + bool "Tune kernel for interactivity" + default y + help + Tunes the kernel for responsiveness at the cost of throughput and power usage. + + --- Block Layer ---------------------------------------- + + Default scheduler.........: mq-deadline -> bfq + + --- Virtual Memory Subsystem --------------------------- + + Background-reclaim hugepages...: no -> yes + + --- CFS CPU Scheduler ---------------------------------- + + Scheduling latency.............: 6 -> 4 ms + Minimal granularity............: 0.75 -> 0.4 ms + Wakeup granularity.............: 1 -> 0.5 ms + CPU migration cost.............: 0.5 -> 0.25 ms + Bandwidth slice size...........: 5 -> 3 ms + Task rebalancing threshold.....: 32 -> 128 + Ondemand sampling down factor..: 1 -> 5 + + --- MuQSS CPU Scheduler -------------------------------- + + Scheduling interval............: 6 -> 4 ms + ISO task max realtime use......: 70 % -> 25 % + Yield type.....................: 1 -> 0 + Ondemand default freq up thresh: 80 -> 40 + Ondemand micro freq up thresh..: 95 -> 45 + Ondemand sampling down factor..: 1 -> 5 + config BROKEN bool @@ -522,6 +564,7 @@ config SCHED_THERMAL_PRESSURE default y if ARM64 depends on SMP depends on CPU_FREQ_THERMAL + depends on !SCHED_MUQSS help Select this option to enable thermal pressure accounting in the scheduler. Thermal pressure is the value conveyed to the scheduler @@ -871,6 +914,7 @@ config NUMA_BALANCING depends on ARCH_SUPPORTS_NUMA_BALANCING depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY depends on SMP && NUMA && MIGRATION + depends on !SCHED_MUQSS help This option adds support for automatic NUMA aware memory/task placement. The mechanism is quite primitive and is based on migrating memory when @@ -955,9 +999,13 @@ menuconfig CGROUP_SCHED help This feature lets CPU scheduler recognize task groups and control CPU bandwidth allocation to such task groups. It uses cgroups to group - tasks. + tasks. In combination with MuQSS this is purely a STUB to create the + files associated with the CPU controller cgroup but most of the + controls do nothing. This is useful for working in environments and + with applications that will only work if this control group is + present. -if CGROUP_SCHED +if CGROUP_SCHED && !SCHED_MUQSS config FAIR_GROUP_SCHED bool "Group scheduling for SCHED_OTHER" depends on CGROUP_SCHED @@ -1086,6 +1134,7 @@ config CGROUP_DEVICE config CGROUP_CPUACCT bool "Simple CPU accounting controller" + depends on !SCHED_MUQSS help Provides a simple controller for monitoring the total CPU consumed by the tasks in a cgroup. @@ -1195,6 +1244,22 @@ config USER_NS If unsure, say N. +config USER_NS_UNPRIVILEGED + bool "Allow unprivileged users to create namespaces" + default y + depends on USER_NS + help + When disabled, unprivileged users will not be able to create + new namespaces. Allowing users to create their own namespaces + has been part of several recent local privilege escalation + exploits, so if you need user namespaces but are + paranoid^Wsecurity-conscious you want to disable this. + + This setting can be overridden at runtime via the + kernel.unprivileged_userns_clone sysctl. + + If unsure, say Y. + config PID_NS bool "PID Namespaces" default y @@ -1228,6 +1293,7 @@ config CHECKPOINT_RESTORE config SCHED_AUTOGROUP bool "Automatic process group scheduling" + depends on !SCHED_MUQSS select CGROUPS select CGROUP_SCHED select FAIR_GROUP_SCHED @@ -1334,7 +1400,7 @@ config CC_OPTIMIZE_FOR_PERFORMANCE config CC_OPTIMIZE_FOR_PERFORMANCE_O3 bool "Optimize more for performance (-O3)" - depends on ARC + help Choosing this option will pass "-O3" to your compiler to optimize the kernel yet more for performance. @@ -1555,6 +1621,13 @@ config FUTEX support for "fast userspace mutexes". The resulting kernel may not run glibc-based applications correctly. +config FUTEX2 + bool "Enable futex2 support" if EXPERT + depends on FUTEX + default y + help + Support for futex2 interface. + config FUTEX_PI bool depends on FUTEX && RT_MUTEXES @@ -2349,4 +2422,4 @@ config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE # kernel/time/posix-stubs.c. All these overrides need to be available in # . config ARCH_HAS_SYSCALL_WRAPPER - def_bool n + def_bool n \ No newline at end of file diff --git a/init/init_task.c b/init/init_task.c index 8b08c2e19cb..984f6fd8a5b 100644 --- a/init/init_task.c +++ b/init/init_task.c @@ -75,9 +75,17 @@ struct task_struct init_task .stack = init_stack, .usage = REFCOUNT_INIT(2), .flags = PF_KTHREAD, +#ifdef CONFIG_SCHED_MUQSS + .prio = NORMAL_PRIO, + .static_prio = MAX_PRIO - 20, + .normal_prio = NORMAL_PRIO, + .deadline = 0, + .time_slice = 1000000, +#else .prio = MAX_PRIO - 20, .static_prio = MAX_PRIO - 20, .normal_prio = MAX_PRIO - 20, +#endif .policy = SCHED_NORMAL, .cpus_ptr = &init_task.cpus_mask, .cpus_mask = CPU_MASK_ALL, @@ -87,6 +95,7 @@ struct task_struct init_task .restart_block = { .fn = do_no_restart_syscall, }, +#ifndef CONFIG_SCHED_MUQSS .se = { .group_node = LIST_HEAD_INIT(init_task.se.group_node), }, @@ -94,6 +103,7 @@ struct task_struct init_task .run_list = LIST_HEAD_INIT(init_task.rt.run_list), .time_slice = RR_TIMESLICE, }, +#endif .tasks = LIST_HEAD_INIT(init_task.tasks), #ifdef CONFIG_SMP .pushable_tasks = PLIST_NODE_INIT(init_task.pushable_tasks, MAX_PRIO), diff --git a/init/main.c b/init/main.c index e6836a9400d..6c26f0af060 100644 --- a/init/main.c +++ b/init/main.c @@ -1462,6 +1462,8 @@ static int __ref kernel_init(void *unused) do_sysctl_args(); + print_scheduler_version(); + if (ramdisk_execute_command) { ret = run_init_process(ramdisk_execute_command); if (!ret) diff --git a/kernel/Kconfig.MuQSS b/kernel/Kconfig.MuQSS new file mode 100644 index 00000000000..91688dae437 --- /dev/null +++ b/kernel/Kconfig.MuQSS @@ -0,0 +1,106 @@ +choice + depends on SMP + prompt "CPU scheduler runqueue sharing" + default RQ_MC if SCHED_MUQSS + default RQ_NONE + +config RQ_NONE + bool "No sharing" + help + This is the default behaviour where the CPU scheduler has one runqueue + per CPU, whether it is a physical or logical CPU (hyperthread). + + This can still be enabled runtime with the boot parameter + rqshare=none + + If unsure, say N. + +config RQ_SMT + bool "SMT (hyperthread) siblings" + depends on SCHED_SMT && SCHED_MUQSS + + help + With this option enabled, the CPU scheduler will have one runqueue + shared by SMT (hyperthread) siblings. As these logical cores share + one physical core, sharing the runqueue resource can lead to decreased + overhead, lower latency and higher throughput. + + This can still be enabled runtime with the boot parameter + rqshare=smt + + If unsure, say N. + +config RQ_MC + bool "Multicore siblings" + depends on SCHED_MC && SCHED_MUQSS + help + With this option enabled, the CPU scheduler will have one runqueue + shared by multicore siblings in addition to any SMT siblings. + As these physical cores share caches, sharing the runqueue resource + will lead to lower latency, but its effects on overhead and throughput + are less predictable. As a general rule, 6 or fewer cores will likely + benefit from this, while larger CPUs will only derive a latency + benefit. If your workloads are primarily single threaded, this will + possibly worsen throughput. If you are only concerned about latency + then enable this regardless of how many cores you have. + + This can still be enabled runtime with the boot parameter + rqshare=mc + + If unsure, say Y. + +config RQ_MC_LLC + bool "Multicore siblings (LLC)" + depends on SCHED_MC && SCHED_MUQSS + help + With this option enabled, the CPU scheduler will behave similarly as + with "Multicore siblings". + This option takes LLC cache into account when scheduling tasks. + Option may benefit CPUs with multiple LLC caches, such as Ryzen + and Xeon CPUs. + + This can still be enabled runtime with the boot parameter + rqshare=llc + + If unsure, say N. + +config RQ_SMP + bool "Symmetric Multi-Processing" + depends on SMP && SCHED_MUQSS + help + With this option enabled, the CPU scheduler will have one runqueue + shared by all physical CPUs unless they are on separate NUMA nodes. + As physical CPUs usually do not share resources, sharing the runqueue + will normally worsen throughput but improve latency. If you only + care about latency enable this. + + This can still be enabled runtime with the boot parameter + rqshare=smp + + If unsure, say N. + +config RQ_ALL + bool "NUMA" + depends on SMP && SCHED_MUQSS + help + With this option enabled, the CPU scheduler will have one runqueue + regardless of the architecture configuration, including across NUMA + nodes. This can substantially decrease throughput in NUMA + configurations, but light NUMA designs will not be dramatically + affected. This option should only be chosen if latency is the prime + concern. + + This can still be enabled runtime with the boot parameter + rqshare=all + + If unsure, say N. +endchoice + +config SHARERQ + int + default 0 if RQ_NONE + default 1 if RQ_SMT + default 2 if RQ_MC + default 3 if RQ_MC_LLC + default 4 if RQ_SMP + default 5 if RQ_ALL diff --git a/kernel/Makefile b/kernel/Makefile index 4df609be42d..1444ff31c1c 100644 --- a/kernel/Makefile +++ b/kernel/Makefile @@ -10,7 +10,8 @@ obj-y = fork.o exec_domain.o panic.o \ extable.o params.o \ kthread.o sys_ni.o nsproxy.o \ notifier.o ksysfs.o cred.o reboot.o \ - async.o range.o smpboot.o ucount.o regset.o + async.o range.o smpboot.o ucount.o regset.o \ + skip_list.o obj-$(CONFIG_USERMODE_DRIVER) += usermode_driver.o obj-$(CONFIG_MODULES) += kmod.o @@ -60,6 +61,7 @@ obj-$(CONFIG_PROFILING) += profile.o obj-$(CONFIG_STACKTRACE) += stacktrace.o obj-y += time/ obj-$(CONFIG_FUTEX) += futex.o +obj-$(CONFIG_FUTEX2) += futex2.o obj-$(CONFIG_GENERIC_ISA_DMA) += dma.o obj-$(CONFIG_SMP) += smp.o ifneq ($(CONFIG_SMP),y) diff --git a/kernel/bounds.c b/kernel/bounds.c index 9795d75b09b..a8cbf2d0b11 100644 --- a/kernel/bounds.c +++ b/kernel/bounds.c @@ -22,6 +22,12 @@ int main(void) DEFINE(NR_CPUS_BITS, ilog2(CONFIG_NR_CPUS)); #endif DEFINE(SPINLOCK_SIZE, sizeof(spinlock_t)); +#ifdef CONFIG_LRU_GEN + /* bits needed to represent internal values stored in page->flags */ + DEFINE(LRU_GEN_WIDTH, order_base_2(CONFIG_NR_LRU_GENS + 1)); + /* bits needed to represent normalized values for external uses */ + DEFINE(LRU_GEN_SHIFT, order_base_2(CONFIG_NR_LRU_GENS)); +#endif /* End of constants */ return 0; diff --git a/kernel/delayacct.c b/kernel/delayacct.c index 27725754ac9..769d773c718 100644 --- a/kernel/delayacct.c +++ b/kernel/delayacct.c @@ -106,7 +106,7 @@ int __delayacct_add_tsk(struct taskstats *d, struct task_struct *tsk) */ t1 = tsk->sched_info.pcount; t2 = tsk->sched_info.run_delay; - t3 = tsk->se.sum_exec_runtime; + t3 = tsk_seruntime(tsk); d->cpu_count += t1; diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c index 6addc978031..4e93e560272 100644 --- a/kernel/events/uprobes.c +++ b/kernel/events/uprobes.c @@ -184,7 +184,7 @@ static int __replace_page(struct vm_area_struct *vma, unsigned long addr, if (new_page) { get_page(new_page); page_add_new_anon_rmap(new_page, vma, addr, false); - lru_cache_add_inactive_or_unevictable(new_page, vma); + lru_cache_add_page_vma(new_page, vma, false); } else /* no new page, just dec_mm_counter for old_page */ dec_mm_counter(mm, MM_ANONPAGES); diff --git a/kernel/exit.c b/kernel/exit.c index 65809fac303..67340b3d216 100644 --- a/kernel/exit.c +++ b/kernel/exit.c @@ -122,7 +122,7 @@ static void __exit_signal(struct task_struct *tsk) sig->curr_target = next_thread(tsk); } - add_device_randomness((const void*) &tsk->se.sum_exec_runtime, + add_device_randomness((const void*) &tsk_seruntime(tsk), sizeof(unsigned long long)); /* @@ -143,7 +143,7 @@ static void __exit_signal(struct task_struct *tsk) sig->inblock += task_io_get_inblock(tsk); sig->oublock += task_io_get_oublock(tsk); task_io_accounting_add(&sig->ioac, &tsk->ioac); - sig->sum_sched_runtime += tsk->se.sum_exec_runtime; + sig->sum_sched_runtime += tsk_seruntime(tsk); sig->nr_threads--; __unhash_process(tsk, group_dead); write_sequnlock(&sig->stats_lock); @@ -422,6 +422,7 @@ void mm_update_next_owner(struct mm_struct *mm) goto retry; } WRITE_ONCE(mm->owner, c); + lru_gen_migrate_mm(mm); task_unlock(c); put_task_struct(c); } diff --git a/kernel/fork.c b/kernel/fork.c index 567fee34050..d19da85e700 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -98,6 +98,10 @@ #include #include +#ifdef CONFIG_USER_NS +#include +#endif + #include #include #include @@ -559,7 +563,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, struct inode *inode = file_inode(file); struct address_space *mapping = file->f_mapping; - get_file(file); + vma_get_file(tmp); if (tmp->vm_flags & VM_DENYWRITE) put_write_access(inode); i_mmap_lock_write(mapping); @@ -592,7 +596,7 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm, __vma_link_rb(mm, tmp, rb_link, rb_parent); rb_link = &tmp->vm_rb.rb_right; rb_parent = &tmp->vm_rb; - + uksm_vma_add_new(tmp); mm->map_count++; if (!(tmp->vm_flags & VM_WIPEONFORK)) retval = copy_page_range(tmp, mpnt); @@ -669,6 +673,7 @@ static void check_mm(struct mm_struct *mm) #if defined(CONFIG_TRANSPARENT_HUGEPAGE) && !USE_SPLIT_PMD_PTLOCKS VM_BUG_ON_MM(mm->pmd_huge_pte, mm); #endif + VM_BUG_ON_MM(lru_gen_mm_is_active(mm), mm); } #define allocate_mm() (kmem_cache_alloc(mm_cachep, GFP_KERNEL)) @@ -1061,6 +1066,7 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, goto fail_nocontext; mm->user_ns = get_user_ns(user_ns); + lru_gen_init_mm(mm); return mm; fail_nocontext: @@ -1103,6 +1109,7 @@ static inline void __mmput(struct mm_struct *mm) } if (mm->binfmt) module_put(mm->binfmt->module); + lru_gen_del_mm(mm); mmdrop(mm); } @@ -1322,6 +1329,8 @@ static void mm_release(struct task_struct *tsk, struct mm_struct *mm) put_user(0, tsk->clear_child_tid); do_futex(tsk->clear_child_tid, FUTEX_WAKE, 1, NULL, NULL, 0, 0); + ksys_futex_wake(tsk->clear_child_tid, 1, + FUTEX_32 | FUTEX_SHARED_FLAG); } tsk->clear_child_tid = NULL; } @@ -1871,6 +1880,10 @@ static __latent_entropy struct task_struct *copy_process( if ((clone_flags & (CLONE_NEWUSER|CLONE_FS)) == (CLONE_NEWUSER|CLONE_FS)) return ERR_PTR(-EINVAL); + if ((clone_flags & CLONE_NEWUSER) && !unprivileged_userns_clone) + if (!capable(CAP_SYS_ADMIN)) + return ERR_PTR(-EPERM); + /* * Thread groups must share signals as well, and detached threads * can only be started up within the thread group. @@ -2523,6 +2536,13 @@ pid_t kernel_clone(struct kernel_clone_args *args) get_task_struct(p); } + if (IS_ENABLED(CONFIG_LRU_GEN) && !(clone_flags & CLONE_VM)) { + /* lock the task to synchronize with memcg migration */ + task_lock(p); + lru_gen_add_mm(p->mm); + task_unlock(p); + } + wake_up_new_task(p); /* forking complete and child started to run, tell ptracer */ @@ -2973,6 +2993,12 @@ int ksys_unshare(unsigned long unshare_flags) if (unshare_flags & CLONE_NEWNS) unshare_flags |= CLONE_FS; + if ((unshare_flags & CLONE_NEWUSER) && !unprivileged_userns_clone) { + err = -EPERM; + if (!capable(CAP_SYS_ADMIN)) + goto bad_unshare_out; + } + err = check_unshare_flags(unshare_flags); if (err) goto bad_unshare_out; diff --git a/kernel/futex.c b/kernel/futex.c index 408cad5e896..aa33e66df5c 100644 --- a/kernel/futex.c +++ b/kernel/futex.c @@ -197,6 +197,8 @@ struct futex_pi_state { * @rt_waiter: rt_waiter storage for use with requeue_pi * @requeue_pi_key: the requeue_pi target futex key * @bitset: bitset for the optional bitmasked wakeup + * @uaddr: userspace address of futex + * @uval: expected futex's value * * We use this hashed waitqueue, instead of a normal wait_queue_entry_t, so * we can wake only the relevant ones (hashed queues may be shared). @@ -219,6 +221,8 @@ struct futex_q { struct rt_mutex_waiter *rt_waiter; union futex_key *requeue_pi_key; u32 bitset; + u32 __user *uaddr; + u32 uval; } __randomize_layout; static const struct futex_q futex_q_init = { @@ -2316,6 +2320,29 @@ static int unqueue_me(struct futex_q *q) return ret; } +/** + * unqueue_multiple() - Remove several futexes from their futex_hash_bucket + * @q: The list of futexes to unqueue + * @count: Number of futexes in the list + * + * Helper to unqueue a list of futexes. This can't fail. + * + * Return: + * - >=0 - Index of the last futex that was awoken; + * - -1 - If no futex was awoken + */ +static int unqueue_multiple(struct futex_q *q, int count) +{ + int ret = -1; + int i; + + for (i = 0; i < count; i++) { + if (!unqueue_me(&q[i])) + ret = i; + } + return ret; +} + /* * PI futexes can not be requeued and must remove themself from the * hash bucket. The hash bucket lock (i.e. lock_ptr) is held. @@ -2679,6 +2706,205 @@ static int futex_wait_setup(u32 __user *uaddr, u32 val, unsigned int flags, return ret; } +/** + * futex_wait_multiple_setup() - Prepare to wait and enqueue multiple futexes + * @qs: The corresponding futex list + * @count: The size of the lists + * @flags: Futex flags (FLAGS_SHARED, etc.) + * @awaken: Index of the last awoken futex + * + * Prepare multiple futexes in a single step and enqueue them. This may fail if + * the futex list is invalid or if any futex was already awoken. On success the + * task is ready to interruptible sleep. + * + * Return: + * - 1 - One of the futexes was awaken by another thread + * - 0 - Success + * - <0 - -EFAULT, -EWOULDBLOCK or -EINVAL + */ +static int futex_wait_multiple_setup(struct futex_q *qs, int count, + unsigned int flags, int *awaken) +{ + struct futex_hash_bucket *hb; + int ret, i; + u32 uval; + + /* + * Enqueuing multiple futexes is tricky, because we need to + * enqueue each futex in the list before dealing with the next + * one to avoid deadlocking on the hash bucket. But, before + * enqueuing, we need to make sure that current->state is + * TASK_INTERRUPTIBLE, so we don't absorb any awake events, which + * cannot be done before the get_futex_key of the next key, + * because it calls get_user_pages, which can sleep. Thus, we + * fetch the list of futexes keys in two steps, by first pinning + * all the memory keys in the futex key, and only then we read + * each key and queue the corresponding futex. + */ +retry: + for (i = 0; i < count; i++) { + qs[i].key = FUTEX_KEY_INIT; + ret = get_futex_key(qs[i].uaddr, flags & FLAGS_SHARED, + &qs[i].key, FUTEX_READ); + if (unlikely(ret)) { + return ret; + } + } + + set_current_state(TASK_INTERRUPTIBLE); + + for (i = 0; i < count; i++) { + struct futex_q *q = &qs[i]; + + hb = queue_lock(q); + + ret = get_futex_value_locked(&uval, q->uaddr); + if (ret) { + /* + * We need to try to handle the fault, which + * cannot be done without sleep, so we need to + * undo all the work already done, to make sure + * we don't miss any wake ups. Therefore, clean + * up, handle the fault and retry from the + * beginning. + */ + queue_unlock(hb); + + /* + * Keys 0..(i-1) are implicitly put + * on unqueue_multiple. + */ + *awaken = unqueue_multiple(qs, i); + + __set_current_state(TASK_RUNNING); + + /* + * On a real fault, prioritize the error even if + * some other futex was awoken. Userspace gave + * us a bad address, -EFAULT them. + */ + ret = get_user(uval, q->uaddr); + if (ret) + return ret; + + /* + * Even if the page fault was handled, If + * something was already awaken, we can safely + * give up and succeed to give a hint for userspace to + * acquire the right futex faster. + */ + if (*awaken >= 0) + return 1; + + goto retry; + } + + if (uval != q->uval) { + queue_unlock(hb); + + /* + * If something was already awaken, we can + * safely ignore the error and succeed. + */ + *awaken = unqueue_multiple(qs, i); + __set_current_state(TASK_RUNNING); + if (*awaken >= 0) + return 1; + + return -EWOULDBLOCK; + } + + /* + * The bucket lock can't be held while dealing with the + * next futex. Queue each futex at this moment so hb can + * be unlocked. + */ + queue_me(&qs[i], hb); + } + return 0; +} + +/** + * futex_wait_multiple() - Prepare to wait on and enqueue several futexes + * @qs: The list of futexes to wait on + * @op: Operation code from futex's syscall + * @count: The number of objects + * @abs_time: Timeout before giving up and returning to userspace + * + * Entry point for the FUTEX_WAIT_MULTIPLE futex operation, this function + * sleeps on a group of futexes and returns on the first futex that + * triggered, or after the timeout has elapsed. + * + * Return: + * - >=0 - Hint to the futex that was awoken + * - <0 - On error + */ +static int futex_wait_multiple(struct futex_q *qs, int op, + u32 count, ktime_t *abs_time) +{ + struct hrtimer_sleeper timeout, *to; + int ret, flags = 0, hint = 0; + unsigned int i; + + if (!(op & FUTEX_PRIVATE_FLAG)) + flags |= FLAGS_SHARED; + + if (op & FUTEX_CLOCK_REALTIME) + flags |= FLAGS_CLOCKRT; + + to = futex_setup_timer(abs_time, &timeout, flags, 0); + while (1) { + ret = futex_wait_multiple_setup(qs, count, flags, &hint); + if (ret) { + if (ret > 0) { + /* A futex was awaken during setup */ + ret = hint; + } + break; + } + + if (to) + hrtimer_start_expires(&to->timer, HRTIMER_MODE_ABS); + + /* + * Avoid sleeping if another thread already tried to + * wake us. + */ + for (i = 0; i < count; i++) { + if (plist_node_empty(&qs[i].list)) + break; + } + + if (i == count && (!to || to->task)) + freezable_schedule(); + + ret = unqueue_multiple(qs, count); + + __set_current_state(TASK_RUNNING); + + if (ret >= 0) + break; + if (to && !to->task) { + ret = -ETIMEDOUT; + break; + } else if (signal_pending(current)) { + ret = -ERESTARTSYS; + break; + } + /* + * The final case is a spurious wakeup, for + * which just retry. + */ + } + + if (to) { + hrtimer_cancel(&to->timer); + destroy_hrtimer_on_stack(&to->timer); + } + + return ret; +} + static int futex_wait(u32 __user *uaddr, unsigned int flags, u32 val, ktime_t *abs_time, u32 bitset) { @@ -3763,6 +3989,7 @@ static __always_inline bool futex_cmd_has_timeout(u32 cmd) case FUTEX_LOCK_PI: case FUTEX_WAIT_BITSET: case FUTEX_WAIT_REQUEUE_PI: + case FUTEX_WAIT_MULTIPLE: return true; } return false; @@ -3775,13 +4002,51 @@ futex_init_timeout(u32 cmd, u32 op, struct timespec64 *ts, ktime_t *t) return -EINVAL; *t = timespec64_to_ktime(*ts); - if (cmd == FUTEX_WAIT) + if (cmd == FUTEX_WAIT || cmd == FUTEX_WAIT_MULTIPLE) *t = ktime_add_safe(ktime_get(), *t); else if (cmd != FUTEX_LOCK_PI && !(op & FUTEX_CLOCK_REALTIME)) *t = timens_ktime_to_host(CLOCK_MONOTONIC, *t); return 0; } +/** + * futex_read_wait_block - Read an array of futex_wait_block from userspace + * @uaddr: Userspace address of the block + * @count: Number of blocks to be read + * + * This function creates and allocate an array of futex_q (we zero it to + * initialize the fields) and then, for each futex_wait_block element from + * userspace, fill a futex_q element with proper values. + */ +inline struct futex_q *futex_read_wait_block(u32 __user *uaddr, u32 count) +{ + unsigned int i; + struct futex_q *qs; + struct futex_wait_block fwb; + struct futex_wait_block __user *entry = + (struct futex_wait_block __user *)uaddr; + + if (!count || count > FUTEX_MULTIPLE_MAX_COUNT) + return ERR_PTR(-EINVAL); + + qs = kcalloc(count, sizeof(*qs), GFP_KERNEL); + if (!qs) + return ERR_PTR(-ENOMEM); + + for (i = 0; i < count; i++) { + if (copy_from_user(&fwb, &entry[i], sizeof(fwb))) { + kfree(qs); + return ERR_PTR(-EFAULT); + } + + qs[i].uaddr = fwb.uaddr; + qs[i].uval = fwb.val; + qs[i].bitset = fwb.bitset; + } + + return qs; +} + SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, const struct __kernel_timespec __user *, utime, u32 __user *, uaddr2, u32, val3) @@ -3801,6 +4066,25 @@ SYSCALL_DEFINE6(futex, u32 __user *, uaddr, int, op, u32, val, tp = &t; } + if (cmd == FUTEX_WAIT_MULTIPLE) { + int ret; + struct futex_q *qs; + +#ifdef CONFIG_X86_X32 + if (unlikely(in_x32_syscall())) + return -ENOSYS; +#endif + qs = futex_read_wait_block(uaddr, val); + + if (IS_ERR(qs)) + return PTR_ERR(qs); + + ret = futex_wait_multiple(qs, op, val, tp); + kfree(qs); + + return ret; + } + return do_futex(uaddr, op, val, tp, uaddr2, (unsigned long)utime, val3); } @@ -3963,6 +4247,58 @@ COMPAT_SYSCALL_DEFINE3(get_robust_list, int, pid, #endif /* CONFIG_COMPAT */ #ifdef CONFIG_COMPAT_32BIT_TIME +/** + * struct compat_futex_wait_block - Block of futexes to be waited for + * @uaddr: User address of the futex (compatible pointer) + * @val: Futex value expected by userspace + * @bitset: Bitset for the optional bitmasked wakeup + */ +struct compat_futex_wait_block { + compat_uptr_t uaddr; + __u32 pad; + __u32 val; + __u32 bitset; +}; + +/** + * compat_futex_read_wait_block - Read an array of futex_wait_block from + * userspace + * @uaddr: Userspace address of the block + * @count: Number of blocks to be read + * + * This function does the same as futex_read_wait_block(), except that it + * converts the pointer to the futex from the compat version to the regular one. + */ +inline struct futex_q *compat_futex_read_wait_block(u32 __user *uaddr, + u32 count) +{ + unsigned int i; + struct futex_q *qs; + struct compat_futex_wait_block fwb; + struct compat_futex_wait_block __user *entry = + (struct compat_futex_wait_block __user *)uaddr; + + if (!count || count > FUTEX_MULTIPLE_MAX_COUNT) + return ERR_PTR(-EINVAL); + + qs = kcalloc(count, sizeof(*qs), GFP_KERNEL); + if (!qs) + return ERR_PTR(-ENOMEM); + + for (i = 0; i < count; i++) { + if (copy_from_user(&fwb, &entry[i], sizeof(fwb))) { + kfree(qs); + return ERR_PTR(-EFAULT); + } + + qs[i].uaddr = compat_ptr(fwb.uaddr); + qs[i].uval = fwb.val; + qs[i].bitset = fwb.bitset; + } + + return qs; +} + SYSCALL_DEFINE6(futex_time32, u32 __user *, uaddr, int, op, u32, val, const struct old_timespec32 __user *, utime, u32 __user *, uaddr2, u32, val3) @@ -3980,6 +4316,19 @@ SYSCALL_DEFINE6(futex_time32, u32 __user *, uaddr, int, op, u32, val, tp = &t; } + if (cmd == FUTEX_WAIT_MULTIPLE) { + int ret; + struct futex_q *qs = compat_futex_read_wait_block(uaddr, val); + + if (IS_ERR(qs)) + return PTR_ERR(qs); + + ret = futex_wait_multiple(qs, op, val, tp); + kfree(qs); + + return ret; + } + return do_futex(uaddr, op, val, tp, uaddr2, (unsigned long)utime, val3); } #endif /* CONFIG_COMPAT_32BIT_TIME */ diff --git a/kernel/futex2.c b/kernel/futex2.c new file mode 100644 index 00000000000..54c13b7e272 --- /dev/null +++ b/kernel/futex2.c @@ -0,0 +1,1297 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/* + * futex2 system call interface by André Almeida + * + * Copyright 2021 Collabora Ltd. + * + * Based on original futex implementation by: + * (C) 2002 Rusty Russell, IBM + * (C) 2003, 2006 Ingo Molnar, Red Hat Inc. + * (C) 2003, 2004 Jamie Lokier + * (C) 2006 Thomas Gleixner, Timesys Corp. + * (C) 2007 Eric Dumazet + * (C) 2009 Darren Hart, IBM + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef CONFIG_X86_64 +#include +#endif + +/** + * struct futex_key - Components to build unique key for a futex + * @pointer: Pointer to current->mm or inode's UUID for file backed futexes + * @index: Start address of the page containing futex or index of the page + * @offset: Address offset of uaddr in a page + */ +struct futex_key { + u64 pointer; + unsigned long index; + unsigned long offset; +}; + +/** + * struct futex_waiter - List entry for a waiter + * @uaddr: Virtual address of userspace futex + * @key: Information that uniquely identify a futex + * @list: List node struct + * @val: Expected value for this waiter + * @flags: Flags + * @bucket: Pointer to the bucket for this waiter + * @index: Index of waiter in futexv list + */ +struct futex_waiter { + void __user *uaddr; + struct futex_key key; + struct list_head list; + unsigned int val; + unsigned int flags; + struct futex_bucket *bucket; + unsigned int index; +}; + +/** + * struct futex_waiter_head - List of futexes to be waited + * @task: Task to be awaken + * @hint: Was someone on this list awakened? + * @objects: List of futexes + */ +struct futex_waiter_head { + struct task_struct *task; + bool hint; + struct futex_waiter objects[0]; +}; + +/** + * struct futex_bucket - A bucket of futex's hash table + * @waiters: Number of waiters in the bucket + * @lock: Bucket lock + * @list: List of waiters on this bucket + */ +struct futex_bucket { + atomic_t waiters; + spinlock_t lock; + struct list_head list; +}; + + +/* Mask for futex2 flag operations */ +#define FUTEX2_MASK (FUTEX_SIZE_MASK | FUTEX_CLOCK_REALTIME | FUTEX_SHARED_FLAG) + +/* Mask for sys_futex_waitv flag */ +#define FUTEXV_MASK (FUTEX_CLOCK_REALTIME) + +/* Mask for each futex in futex_waitv list */ +#define FUTEXV_WAITER_MASK (FUTEX_SIZE_MASK | FUTEX_SHARED_FLAG) + +#define is_object_shared ((futexv->objects[i].flags & FUTEX_SHARED_FLAG) ? true : false) + +#define FUT_OFF_INODE 1 /* We set bit 0 if key has a reference on inode */ +#define FUT_OFF_MMSHARED 2 /* We set bit 1 if key has a reference on mm */ + +static struct futex_bucket *futex_table; +static unsigned int futex2_hashsize; + +/* + * Reflects a new waiter being added to the waitqueue. + */ +static inline void bucket_inc_waiters(struct futex_bucket *bucket) +{ +#ifdef CONFIG_SMP + atomic_inc(&bucket->waiters); + /* + * Issue a barrier after adding so futex_wake() will see that the + * value had increased + */ + smp_mb__after_atomic(); +#endif +} + +/* + * Reflects a waiter being removed from the waitqueue by wakeup + * paths. + */ +static inline void bucket_dec_waiters(struct futex_bucket *bucket) +{ +#ifdef CONFIG_SMP + atomic_dec(&bucket->waiters); +#endif +} + +/* + * Get the number of waiters in a bucket + */ +static inline int bucket_get_waiters(struct futex_bucket *bucket) +{ +#ifdef CONFIG_SMP + /* + * Issue a barrier before reading so we get an updated value from + * futex_wait() + */ + smp_mb(); + return atomic_read(&bucket->waiters); +#else + return 1; +#endif +} + +/** + * futex_get_inode_uuid - Gets an UUID for an inode + * @inode: inode to get UUID + * + * Generate a machine wide unique identifier for this inode. + * + * This relies on u64 not wrapping in the life-time of the machine; which with + * 1ns resolution means almost 585 years. + * + * This further relies on the fact that a well formed program will not unmap + * the file while it has a (shared) futex waiting on it. This mapping will have + * a file reference which pins the mount and inode. + * + * If for some reason an inode gets evicted and read back in again, it will get + * a new sequence number and will _NOT_ match, even though it is the exact same + * file. + * + * It is important that match_futex() will never have a false-positive, esp. + * for PI futexes that can mess up the state. The above argues that false-negatives + * are only possible for malformed programs. + * + * Returns: UUID for the given inode + */ +static u64 futex_get_inode_uuid(struct inode *inode) +{ + static atomic64_t i_seq; + u64 old; + + /* Does the inode already have a sequence number? */ + old = atomic64_read(&inode->i_sequence2); + + if (likely(old)) + return old; + + for (;;) { + u64 new = atomic64_add_return(1, &i_seq); + + if (WARN_ON_ONCE(!new)) + continue; + + old = atomic64_cmpxchg_relaxed(&inode->i_sequence2, 0, new); + if (old) + return old; + return new; + } +} + +/** + * futex_get_shared_key - Get a key for a shared futex + * @address: Futex memory address + * @mm: Current process mm_struct pointer + * @key: Key struct to be filled + * + * Returns: 0 on success, error code otherwise + */ +static int futex_get_shared_key(uintptr_t address, struct mm_struct *mm, + struct futex_key *key) +{ + int ret; + struct page *page, *tail; + struct address_space *mapping; + +again: + ret = get_user_pages_fast(address, 1, 0, &page); + if (ret < 0) + return ret; + + /* + * The treatment of mapping from this point on is critical. The page + * lock protects many things but in this context the page lock + * stabilizes mapping, prevents inode freeing in the shared + * file-backed region case and guards against movement to swap cache. + * + * Strictly speaking the page lock is not needed in all cases being + * considered here and page lock forces unnecessarily serialization + * From this point on, mapping will be re-verified if necessary and + * page lock will be acquired only if it is unavoidable + * + * Mapping checks require the head page for any compound page so the + * head page and mapping is looked up now. For anonymous pages, it + * does not matter if the page splits in the future as the key is + * based on the address. For filesystem-backed pages, the tail is + * required as the index of the page determines the key. For + * base pages, there is no tail page and tail == page. + */ + tail = page; + page = compound_head(page); + mapping = READ_ONCE(page->mapping); + + /* + * If page->mapping is NULL, then it cannot be a PageAnon + * page; but it might be the ZERO_PAGE or in the gate area or + * in a special mapping (all cases which we are happy to fail); + * or it may have been a good file page when get_user_pages_fast + * found it, but truncated or holepunched or subjected to + * invalidate_complete_page2 before we got the page lock (also + * cases which we are happy to fail). And we hold a reference, + * so refcount care in invalidate_complete_page's remove_mapping + * prevents drop_caches from setting mapping to NULL beneath us. + * + * The case we do have to guard against is when memory pressure made + * shmem_writepage move it from filecache to swapcache beneath us: + * an unlikely race, but we do need to retry for page->mapping. + */ + if (unlikely(!mapping)) { + int shmem_swizzled; + + /* + * Page lock is required to identify which special case above + * applies. If this is really a shmem page then the page lock + * will prevent unexpected transitions. + */ + lock_page(page); + shmem_swizzled = PageSwapCache(page) || page->mapping; + unlock_page(page); + put_page(page); + + if (shmem_swizzled) + goto again; + + return -EFAULT; + } + + /* + * Private mappings are handled in a simple way. + * + * If the futex key is stored on an anonymous page, then the associated + * object is the mm which is implicitly pinned by the calling process. + * + * NOTE: When userspace waits on a MAP_SHARED mapping, even if + * it's a read-only handle, it's expected that futexes attach to + * the object not the particular process. + */ + if (PageAnon(page)) { + key->offset |= FUT_OFF_MMSHARED; + } else { + struct inode *inode; + + /* + * The associated futex object in this case is the inode and + * the page->mapping must be traversed. Ordinarily this should + * be stabilised under page lock but it's not strictly + * necessary in this case as we just want to pin the inode, not + * update the radix tree or anything like that. + * + * The RCU read lock is taken as the inode is finally freed + * under RCU. If the mapping still matches expectations then the + * mapping->host can be safely accessed as being a valid inode. + */ + rcu_read_lock(); + + if (READ_ONCE(page->mapping) != mapping) { + rcu_read_unlock(); + put_page(page); + + goto again; + } + + inode = READ_ONCE(mapping->host); + if (!inode) { + rcu_read_unlock(); + put_page(page); + + goto again; + } + + key->pointer = futex_get_inode_uuid(inode); + key->index = (unsigned long)page_to_pgoff(tail); + key->offset |= FUT_OFF_INODE; + + rcu_read_unlock(); + } + + put_page(page); + + return 0; +} + +/** + * futex_get_bucket - Check if the user address is valid, prepare internal + * data and calculate the hash + * @uaddr: futex user address + * @key: data that uniquely identifies a futex + * @shared: is this a shared futex? + * + * For private futexes, each uaddr will be unique for a given mm_struct, and it + * won't be freed for the life time of the process. For shared futexes, check + * futex_get_shared_key(). + * + * Return: address of bucket on success, error code otherwise + */ +static struct futex_bucket *futex_get_bucket(void __user *uaddr, + struct futex_key *key, + bool shared) +{ + uintptr_t address = (uintptr_t)uaddr; + u32 hash_key; + + /* Checking if uaddr is valid and accessible */ + if (unlikely(!IS_ALIGNED(address, sizeof(u32)))) + return ERR_PTR(-EINVAL); + if (unlikely(!access_ok(uaddr, sizeof(u32)))) + return ERR_PTR(-EFAULT); + + key->offset = address % PAGE_SIZE; + address -= key->offset; + key->pointer = (u64)address; + key->index = (unsigned long)current->mm; + + if (shared) + futex_get_shared_key(address, current->mm, key); + + /* Generate hash key for this futex using uaddr and current->mm */ + hash_key = jhash2((u32 *)key, sizeof(*key) / sizeof(u32), 0); + + /* Since HASH_SIZE is 2^n, subtracting 1 makes a perfect bit mask */ + return &futex_table[hash_key & (futex2_hashsize - 1)]; +} + +/** + * futex_get_user - Get the userspace value on this address + * @uval: variable to store the value + * @uaddr: userspace address + * + * Check the comment at futex_enqueue() for more information. + */ +static int futex_get_user(u32 *uval, u32 __user *uaddr) +{ + int ret; + + pagefault_disable(); + ret = __get_user(*uval, uaddr); + pagefault_enable(); + + return ret; +} + +/** + * futex_setup_time - Prepare the timeout mechanism and start it. + * @timo: Timeout value from userspace + * @timeout: Pointer to hrtimer handler + * @flags: Flags from userspace, to decide which clockid to use + * + * Return: 0 on success, error code otherwise + */ +static int futex_setup_time(struct __kernel_timespec __user *timo, + struct hrtimer_sleeper *timeout, + unsigned int flags) +{ + ktime_t time; + struct timespec64 ts; + clockid_t clockid = (flags & FUTEX_CLOCK_REALTIME) ? + CLOCK_REALTIME : CLOCK_MONOTONIC; + + if (get_timespec64(&ts, timo)) + return -EFAULT; + + if (!timespec64_valid(&ts)) + return -EINVAL; + + time = timespec64_to_ktime(ts); + + hrtimer_init_sleeper(timeout, clockid, HRTIMER_MODE_ABS); + + hrtimer_set_expires(&timeout->timer, time); + + hrtimer_sleeper_start_expires(timeout, HRTIMER_MODE_ABS); + + return 0; +} + +/** + * futex_dequeue_multiple - Remove multiple futexes from hash table + * @futexv: list of waiters + * @nr: number of futexes to be removed + * + * This function is used if (a) something went wrong while enqueuing, and we + * need to undo our work (then nr <= nr_futexes) or (b) we woke up, and thus + * need to remove every waiter, check if some was indeed woken and return. + * Before removing a waiter, we check if it's on the list, since we have no + * clue who have been waken. + * + * Return: + * * -1 - If no futex was woken during the removal + * * 0>= - At least one futex was found woken, index of the last one + */ +static int futex_dequeue_multiple(struct futex_waiter_head *futexv, unsigned int nr) +{ + int i, ret = -1; + + for (i = 0; i < nr; i++) { + spin_lock(&futexv->objects[i].bucket->lock); + if (!list_empty(&futexv->objects[i].list)) { + list_del_init(&futexv->objects[i].list); + bucket_dec_waiters(futexv->objects[i].bucket); + } else { + ret = i; + } + spin_unlock(&futexv->objects[i].bucket->lock); + } + + return ret; +} + +/** + * futex_enqueue - Check the value and enqueue a futex on a wait list + * + * @futexv: List of futexes + * @nr_futexes: Number of futexes in the list + * @awakened: If a futex was awakened during enqueueing, store the index here + * + * Get the value from the userspace address and compares with the expected one. + * + * Getting the value from user futex address: + * + * Since we are in a hurry, we use a spin lock and we can't sleep. + * Try to get the value with page fault disabled (when enable, we might + * sleep). + * + * If we fail, we aren't sure if the address is invalid or is just a + * page fault. Then, release the lock (so we can sleep) and try to get + * the value with page fault enabled. In order to trigger a page fault + * handling, we just call __get_user() again. If we sleep with enqueued + * futexes, we might miss a wake, so dequeue everything before sleeping. + * + * If get_user succeeds, this mean that the address is valid and we do + * the work again. Since we just handled the page fault, the page is + * likely pinned in memory and we should be luckier this time and be + * able to get the value. If we fail anyway, we will try again. + * + * If even with page faults enabled we get and error, this means that + * the address is not valid and we return from the syscall. + * + * If we got an unexpected value or need to treat a page fault and realized that + * a futex was awakened, we can priority this and return success. + * + * In success, enqueue the futex in the correct bucket + * + * Return: + * * 1 - We were awake in the process and nothing is enqueued + * * 0 - Everything is enqueued and we are ready to sleep + * * 0< - Something went wrong, nothing is enqueued, return error code + */ +static int futex_enqueue(struct futex_waiter_head *futexv, unsigned int nr_futexes, + int *awakened) +{ + int i, ret; + u32 uval, val; + u32 __user *uaddr; + bool retry = false; + struct futex_bucket *bucket; + +retry: + set_current_state(TASK_INTERRUPTIBLE); + + for (i = 0; i < nr_futexes; i++) { + uaddr = (u32 __user *)futexv->objects[i].uaddr; + val = (u32)futexv->objects[i].val; + + if (is_object_shared && retry) { + struct futex_bucket *tmp = + futex_get_bucket((void __user *)uaddr, + &futexv->objects[i].key, true); + if (IS_ERR(tmp)) { + __set_current_state(TASK_RUNNING); + futex_dequeue_multiple(futexv, i); + return PTR_ERR(tmp); + } + futexv->objects[i].bucket = tmp; + } + + bucket = futexv->objects[i].bucket; + + bucket_inc_waiters(bucket); + spin_lock(&bucket->lock); + + ret = futex_get_user(&uval, uaddr); + + if (unlikely(ret)) { + spin_unlock(&bucket->lock); + + bucket_dec_waiters(bucket); + __set_current_state(TASK_RUNNING); + *awakened = futex_dequeue_multiple(futexv, i); + + if (*awakened >= 0) + return 1; + + if (__get_user(uval, uaddr)) + return -EFAULT; + + retry = true; + goto retry; + } + + if (uval != val) { + spin_unlock(&bucket->lock); + + bucket_dec_waiters(bucket); + __set_current_state(TASK_RUNNING); + *awakened = futex_dequeue_multiple(futexv, i); + + if (*awakened >= 0) + return 1; + + return -EAGAIN; + } + + list_add_tail(&futexv->objects[i].list, &bucket->list); + spin_unlock(&bucket->lock); + } + + return 0; +} + +/** + * __futex_wait - Enqueue the list of futexes and wait to be woken + * @futexv: List of futexes to wait + * @nr_futexes: Length of futexv + * @timeout: Pointer to timeout handler + * + * Return: + * * 0 >= - Hint of which futex woke us + * * 0 < - Error code + */ +static int __futex_wait(struct futex_waiter_head *futexv, unsigned int nr_futexes, + struct hrtimer_sleeper *timeout) +{ + int ret; + + while (1) { + int awakened = -1; + + ret = futex_enqueue(futexv, nr_futexes, &awakened); + + if (ret) { + if (awakened >= 0) + return awakened; + return ret; + } + + /* Before sleeping, check if someone was woken */ + if (!futexv->hint && (!timeout || timeout->task)) + freezable_schedule(); + + __set_current_state(TASK_RUNNING); + + /* + * One of those things triggered this wake: + * + * * We have been removed from the bucket. futex_wake() woke + * us. We just need to dequeue and return 0 to userspace. + * + * However, if no futex was dequeued by a futex_wake(): + * + * * If the there's a timeout and it has expired, + * return -ETIMEDOUT. + * + * * If there is a signal pending, something wants to kill our + * thread, return -ERESTARTSYS. + * + * * If there's no signal pending, it was a spurious wake + * (scheduler gave us a chance to do some work, even if we + * don't want to). We need to remove ourselves from the + * bucket and add again, to prevent losing wakeups in the + * meantime. + */ + + ret = futex_dequeue_multiple(futexv, nr_futexes); + + /* Normal wake */ + if (ret >= 0) + return ret; + + if (timeout && !timeout->task) + return -ETIMEDOUT; + + if (signal_pending(current)) + return -ERESTARTSYS; + + /* Spurious wake, do everything again */ + } +} + +/** + * futex_wait - Setup the timer (if there's one) and wait on a list of futexes + * @futexv: List of futexes + * @nr_futexes: Length of futexv + * @timo: Timeout + * @flags: Timeout flags + * + * Return: + * * 0 >= - Hint of which futex woke us + * * 0 < - Error code + */ +static int futex_set_timer_and_wait(struct futex_waiter_head *futexv, + unsigned int nr_futexes, + struct __kernel_timespec __user *timo, + unsigned int flags) +{ + struct hrtimer_sleeper timeout; + int ret; + + if (timo) { + ret = futex_setup_time(timo, &timeout, flags); + if (ret) + return ret; + } + + ret = __futex_wait(futexv, nr_futexes, timo ? &timeout : NULL); + + if (timo) + hrtimer_cancel(&timeout.timer); + + return ret; +} + +/** + * sys_futex_wait - Wait on a futex address if (*uaddr) == val + * @uaddr: User address of futex + * @val: Expected value of futex + * @flags: Specify the size of futex and the clockid + * @timo: Optional absolute timeout. + * + * The user thread is put to sleep, waiting for a futex_wake() at uaddr, if the + * value at *uaddr is the same as val (otherwise, the syscall returns + * immediately with -EAGAIN). + * + * Returns 0 on success, error code otherwise. + */ +SYSCALL_DEFINE4(futex_wait, void __user *, uaddr, unsigned int, val, + unsigned int, flags, struct __kernel_timespec __user *, timo) +{ + bool shared = (flags & FUTEX_SHARED_FLAG) ? true : false; + unsigned int size = flags & FUTEX_SIZE_MASK; + struct futex_waiter *waiter; + struct futex_waiter_head *futexv; + + /* Wrapper for a futexv_head of one element */ + struct { + struct futex_waiter_head futexv; + struct futex_waiter waiter; + } __packed wait_single; + + if (flags & ~FUTEX2_MASK) + return -EINVAL; + + if (size != FUTEX_32) + return -EINVAL; + + futexv = &wait_single.futexv; + futexv->task = current; + futexv->hint = false; + + waiter = &wait_single.waiter; + waiter->index = 0; + waiter->val = val; + waiter->uaddr = uaddr; + memset(&wait_single.waiter.key, 0, sizeof(struct futex_key)); + + INIT_LIST_HEAD(&waiter->list); + + /* Get an unlocked hash bucket */ + waiter->bucket = futex_get_bucket(uaddr, &waiter->key, shared); + if (IS_ERR(waiter->bucket)) + return PTR_ERR(waiter->bucket); + + return futex_set_timer_and_wait(futexv, 1, timo, flags); +} + +#ifdef CONFIG_COMPAT +/** + * compat_futex_parse_waitv - Parse a waitv array from userspace + * @futexv: Kernel side list of waiters to be filled + * @uwaitv: Userspace list to be parsed + * @nr_futexes: Length of futexv + * + * Return: Error code on failure, pointer to a prepared futexv otherwise + */ +static int compat_futex_parse_waitv(struct futex_waiter_head *futexv, + struct compat_futex_waitv __user *uwaitv, + unsigned int nr_futexes) +{ + struct futex_bucket *bucket; + struct compat_futex_waitv waitv; + unsigned int i; + + for (i = 0; i < nr_futexes; i++) { + if (copy_from_user(&waitv, &uwaitv[i], sizeof(waitv))) + return -EFAULT; + + if ((waitv.flags & ~FUTEXV_WAITER_MASK) || + (waitv.flags & FUTEX_SIZE_MASK) != FUTEX_32) + return -EINVAL; + + futexv->objects[i].key.pointer = 0; + futexv->objects[i].flags = waitv.flags; + futexv->objects[i].uaddr = compat_ptr(waitv.uaddr); + futexv->objects[i].val = waitv.val; + futexv->objects[i].index = i; + + bucket = futex_get_bucket(compat_ptr(waitv.uaddr), + &futexv->objects[i].key, + is_object_shared); + + if (IS_ERR(bucket)) + return PTR_ERR(bucket); + + futexv->objects[i].bucket = bucket; + + INIT_LIST_HEAD(&futexv->objects[i].list); + } + + return 0; +} + +COMPAT_SYSCALL_DEFINE4(futex_waitv, struct compat_futex_waitv __user *, waiters, + unsigned int, nr_futexes, unsigned int, flags, + struct __kernel_timespec __user *, timo) +{ + struct futex_waiter_head *futexv; + int ret; + + if (flags & ~FUTEXV_MASK) + return -EINVAL; + + if (!nr_futexes || nr_futexes > FUTEX_WAITV_MAX || !waiters) + return -EINVAL; + + futexv = kmalloc((sizeof(struct futex_waiter) * nr_futexes) + + sizeof(*futexv), GFP_KERNEL); + if (!futexv) + return -ENOMEM; + + futexv->hint = false; + futexv->task = current; + + ret = compat_futex_parse_waitv(futexv, waiters, nr_futexes); + + if (!ret) + ret = futex_set_timer_and_wait(futexv, nr_futexes, timo, flags); + + kfree(futexv); + + return ret; +} +#endif + +/** + * futex_parse_waitv - Parse a waitv array from userspace + * @futexv: Kernel side list of waiters to be filled + * @uwaitv: Userspace list to be parsed + * @nr_futexes: Length of futexv + * + * Return: Error code on failure, pointer to a prepared futexv otherwise + */ +static int futex_parse_waitv(struct futex_waiter_head *futexv, + struct futex_waitv __user *uwaitv, + unsigned int nr_futexes) +{ + struct futex_bucket *bucket; + struct futex_waitv waitv; + unsigned int i; + + for (i = 0; i < nr_futexes; i++) { + if (copy_from_user(&waitv, &uwaitv[i], sizeof(waitv))) + return -EFAULT; + + if ((waitv.flags & ~FUTEXV_WAITER_MASK) || + (waitv.flags & FUTEX_SIZE_MASK) != FUTEX_32) + return -EINVAL; + + futexv->objects[i].key.pointer = 0; + futexv->objects[i].flags = waitv.flags; + futexv->objects[i].uaddr = waitv.uaddr; + futexv->objects[i].val = waitv.val; + futexv->objects[i].index = i; + + bucket = futex_get_bucket(waitv.uaddr, &futexv->objects[i].key, + is_object_shared); + + if (IS_ERR(bucket)) + return PTR_ERR(bucket); + + futexv->objects[i].bucket = bucket; + + INIT_LIST_HEAD(&futexv->objects[i].list); + } + + return 0; +} + +/** + * sys_futex_waitv - Wait on a list of futexes + * @waiters: List of futexes to wait on + * @nr_futexes: Length of futexv + * @flags: Flag for timeout (monotonic/realtime) + * @timo: Optional absolute timeout. + * + * Given an array of `struct futex_waitv`, wait on each uaddr. The thread wakes + * if a futex_wake() is performed at any uaddr. The syscall returns immediately + * if any waiter has *uaddr != val. *timo is an optional timeout value for the + * operation. Each waiter has individual flags. The `flags` argument for the + * syscall should be used solely for specifying the timeout as realtime, if + * needed. Flags for shared futexes, sizes, etc. should be used on the + * individual flags of each waiter. + * + * Returns the array index of one of the awaken futexes. There's no given + * information of how many were awakened, or any particular attribute of it (if + * it's the first awakened, if it is of the smaller index...). + */ +SYSCALL_DEFINE4(futex_waitv, struct futex_waitv __user *, waiters, + unsigned int, nr_futexes, unsigned int, flags, + struct __kernel_timespec __user *, timo) +{ + struct futex_waiter_head *futexv; + int ret; + + if (flags & ~FUTEXV_MASK) + return -EINVAL; + + if (!nr_futexes || nr_futexes > FUTEX_WAITV_MAX || !waiters) + return -EINVAL; + + futexv = kmalloc((sizeof(struct futex_waiter) * nr_futexes) + + sizeof(*futexv), GFP_KERNEL); + if (!futexv) + return -ENOMEM; + + futexv->hint = false; + futexv->task = current; + +#ifdef CONFIG_X86_X32_ABI + if (in_x32_syscall()) { + ret = compat_futex_parse_waitv(futexv, (struct compat_futex_waitv *)waiters, + nr_futexes); + } else +#endif + { + ret = futex_parse_waitv(futexv, waiters, nr_futexes); + } + + if (!ret) + ret = futex_set_timer_and_wait(futexv, nr_futexes, timo, flags); + + kfree(futexv); + + return ret; +} + +/** + * futex_get_parent - For a given futex in a futexv list, get a pointer to the futexv + * @waiter: Address of futex in the list + * @index: Index of futex in the list + * + * Return: A pointer to its futexv struct + */ +static inline struct futex_waiter_head *futex_get_parent(uintptr_t waiter, + unsigned int index) +{ + uintptr_t parent = waiter - sizeof(struct futex_waiter_head) + - (uintptr_t)(index * sizeof(struct futex_waiter)); + + return (struct futex_waiter_head *)parent; +} + +/** + * futex_mark_wake - Find the task to be wake and add it in wake queue + * @waiter: Waiter to be wake + * @bucket: Bucket to be decremented + * @wake_q: Wake queue to insert the task + */ +static void futex_mark_wake(struct futex_waiter *waiter, + struct futex_bucket *bucket, + struct wake_q_head *wake_q) +{ + struct task_struct *task; + struct futex_waiter_head *parent = futex_get_parent((uintptr_t)waiter, + waiter->index); + + lockdep_assert_held(&bucket->lock); + parent->hint = true; + task = parent->task; + get_task_struct(task); + list_del_init(&waiter->list); + wake_q_add_safe(wake_q, task); + bucket_dec_waiters(bucket); +} + +static inline bool futex_match(struct futex_key key1, struct futex_key key2) +{ + return (key1.index == key2.index && + key1.pointer == key2.pointer && + key1.offset == key2.offset); +} + +long ksys_futex_wake(void __user *uaddr, unsigned long nr_wake, + unsigned int flags) +{ + bool shared = (flags & FUTEX_SHARED_FLAG) ? true : false; + unsigned int size = flags & FUTEX_SIZE_MASK; + struct futex_waiter waiter, *aux, *tmp; + struct futex_bucket *bucket; + DEFINE_WAKE_Q(wake_q); + int ret = 0; + + if (flags & ~FUTEX2_MASK) + return -EINVAL; + + if (size != FUTEX_32) + return -EINVAL; + + bucket = futex_get_bucket(uaddr, &waiter.key, shared); + if (IS_ERR(bucket)) + return PTR_ERR(bucket); + + if (!bucket_get_waiters(bucket) || !nr_wake) + return 0; + + spin_lock(&bucket->lock); + list_for_each_entry_safe(aux, tmp, &bucket->list, list) { + if (futex_match(waiter.key, aux->key)) { + futex_mark_wake(aux, bucket, &wake_q); + if (++ret >= nr_wake) + break; + } + } + spin_unlock(&bucket->lock); + + wake_up_q(&wake_q); + + return ret; +} + +/** + * sys_futex_wake - Wake a number of futexes waiting on an address + * @uaddr: Address of futex to be woken up + * @nr_wake: Number of futexes waiting in uaddr to be woken up + * @flags: Flags for size and shared + * + * Wake `nr_wake` threads waiting at uaddr. + * + * Returns the number of woken threads on success, error code otherwise. + */ +SYSCALL_DEFINE3(futex_wake, void __user *, uaddr, unsigned int, nr_wake, + unsigned int, flags) +{ + return ksys_futex_wake(uaddr, nr_wake, flags); +} + +static void futex_double_unlock(struct futex_bucket *b1, struct futex_bucket *b2) +{ + spin_unlock(&b1->lock); + if (b1 != b2) + spin_unlock(&b2->lock); +} + +static inline int __futex_requeue(struct futex_requeue rq1, + struct futex_requeue rq2, unsigned int nr_wake, + unsigned int nr_requeue, unsigned int cmpval, + bool shared1, bool shared2) +{ + struct futex_waiter w1, w2, *aux, *tmp; + bool retry = false; + struct futex_bucket *b1, *b2; + DEFINE_WAKE_Q(wake_q); + u32 uval; + int ret; + + b1 = futex_get_bucket(rq1.uaddr, &w1.key, shared1); + if (IS_ERR(b1)) + return PTR_ERR(b1); + + b2 = futex_get_bucket(rq2.uaddr, &w2.key, shared2); + if (IS_ERR(b2)) + return PTR_ERR(b2); + +retry: + if (shared1 && retry) { + b1 = futex_get_bucket(rq1.uaddr, &w1.key, shared1); + if (IS_ERR(b1)) + return PTR_ERR(b1); + } + + if (shared2 && retry) { + b2 = futex_get_bucket(rq2.uaddr, &w2.key, shared2); + if (IS_ERR(b2)) + return PTR_ERR(b2); + } + + bucket_inc_waiters(b2); + /* + * To ensure the locks are taken in the same order for all threads (and + * thus avoiding deadlocks), take the "smaller" one first + */ + if (b1 <= b2) { + spin_lock(&b1->lock); + if (b1 < b2) + spin_lock_nested(&b2->lock, SINGLE_DEPTH_NESTING); + } else { + spin_lock(&b2->lock); + spin_lock_nested(&b1->lock, SINGLE_DEPTH_NESTING); + } + + ret = futex_get_user(&uval, rq1.uaddr); + + if (unlikely(ret)) { + futex_double_unlock(b1, b2); + if (__get_user(uval, (u32 __user *)rq1.uaddr)) + return -EFAULT; + + bucket_dec_waiters(b2); + retry = true; + goto retry; + } + + if (uval != cmpval) { + futex_double_unlock(b1, b2); + + bucket_dec_waiters(b2); + return -EAGAIN; + } + + list_for_each_entry_safe(aux, tmp, &b1->list, list) { + if (futex_match(w1.key, aux->key)) { + if (ret < nr_wake) { + futex_mark_wake(aux, b1, &wake_q); + ret++; + continue; + } + + if (ret >= nr_wake + nr_requeue) + break; + + aux->key.pointer = w2.key.pointer; + aux->key.index = w2.key.index; + aux->key.offset = w2.key.offset; + + if (b1 != b2) { + list_del_init_careful(&aux->list); + bucket_dec_waiters(b1); + + list_add_tail(&aux->list, &b2->list); + bucket_inc_waiters(b2); + } + ret++; + } + } + + futex_double_unlock(b1, b2); + wake_up_q(&wake_q); + bucket_dec_waiters(b2); + + return ret; +} + +#ifdef CONFIG_COMPAT +static int compat_futex_parse_requeue(struct futex_requeue *rq, + struct compat_futex_requeue __user *uaddr, + bool *shared) +{ + struct compat_futex_requeue tmp; + + if (copy_from_user(&tmp, uaddr, sizeof(tmp))) + return -EFAULT; + + if (tmp.flags & ~FUTEXV_WAITER_MASK || + (tmp.flags & FUTEX_SIZE_MASK) != FUTEX_32) + return -EINVAL; + + *shared = (tmp.flags & FUTEX_SHARED_FLAG) ? true : false; + + rq->uaddr = compat_ptr(tmp.uaddr); + rq->flags = tmp.flags; + + return 0; +} + +COMPAT_SYSCALL_DEFINE6(futex_requeue, struct compat_futex_requeue __user *, uaddr1, + struct compat_futex_requeue __user *, uaddr2, + unsigned int, nr_wake, unsigned int, nr_requeue, + unsigned int, cmpval, unsigned int, flags) +{ + struct futex_requeue rq1, rq2; + bool shared1, shared2; + int ret; + + if (flags) + return -EINVAL; + + ret = compat_futex_parse_requeue(&rq1, uaddr1, &shared1); + if (ret) + return ret; + + ret = compat_futex_parse_requeue(&rq2, uaddr2, &shared2); + if (ret) + return ret; + + return __futex_requeue(rq1, rq2, nr_wake, nr_requeue, cmpval, shared1, shared2); +} +#endif + +/** + * futex_parse_requeue - Copy a user struct futex_requeue and check it's flags + * @rq: Kernel struct + * @uaddr: Address of user struct + * @shared: Out parameter, defines if this is a shared futex + * + * Return: 0 on success, error code otherwise + */ +static int futex_parse_requeue(struct futex_requeue *rq, + struct futex_requeue __user *uaddr, bool *shared) +{ + if (copy_from_user(rq, uaddr, sizeof(*rq))) + return -EFAULT; + + if (rq->flags & ~FUTEXV_WAITER_MASK || + (rq->flags & FUTEX_SIZE_MASK) != FUTEX_32) + return -EINVAL; + + *shared = (rq->flags & FUTEX_SHARED_FLAG) ? true : false; + + return 0; +} + +/** + * sys_futex_requeue - Wake futexes at uaddr1 and requeue from uaddr1 to uaddr2 + * @uaddr1: Address of futexes to be waken/dequeued + * @uaddr2: Address for the futexes to be enqueued + * @nr_wake: Number of futexes waiting in uaddr1 to be woken up + * @nr_requeue: Number of futexes to be requeued from uaddr1 to uaddr2 + * @cmpval: Expected value at uaddr1 + * @flags: Reserved flags arg for requeue operation expansion. Must be 0. + * + * If (uaddr1->uaddr == cmpval), wake at uaddr1->uaddr a nr_wake number of + * waiters and then, remove a number of nr_requeue waiters at uaddr1->uaddr + * and add then to uaddr2->uaddr list. Each uaddr has its own set of flags, + * that must be defined at struct futex_requeue (such as size, shared, NUMA). + * + * Return the number of the woken futexes + the number of requeued ones on + * success, error code otherwise. + */ +SYSCALL_DEFINE6(futex_requeue, struct futex_requeue __user *, uaddr1, + struct futex_requeue __user *, uaddr2, + unsigned int, nr_wake, unsigned int, nr_requeue, + unsigned int, cmpval, unsigned int, flags) +{ + struct futex_requeue rq1, rq2; + bool shared1, shared2; + int ret; + + if (flags) + return -EINVAL; + +#ifdef CONFIG_X86_X32_ABI + if (in_x32_syscall()) { + ret = compat_futex_parse_requeue(&rq1, (struct compat_futex_requeue *)uaddr1, + &shared1); + if (ret) + return ret; + + ret = compat_futex_parse_requeue(&rq2, (struct compat_futex_requeue *)uaddr2, + &shared2); + if (ret) + return ret; + } else +#endif + { + ret = futex_parse_requeue(&rq1, uaddr1, &shared1); + if (ret) + return ret; + + ret = futex_parse_requeue(&rq2, uaddr2, &shared2); + if (ret) + return ret; + } + + return __futex_requeue(rq1, rq2, nr_wake, nr_requeue, cmpval, shared1, shared2); +} + +static ssize_t wait_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%u\n", __NR_futex_wait); + +} +static struct kobj_attribute futex2_wait_attr = __ATTR_RO(wait); + +static ssize_t wake_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%u\n", __NR_futex_wake); + +} +static struct kobj_attribute futex2_wake_attr = __ATTR_RO(wake); + +static ssize_t waitv_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%u\n", __NR_futex_waitv); + +} +static struct kobj_attribute futex2_waitv_attr = __ATTR_RO(waitv); + +static struct attribute *futex2_sysfs_attrs[] = { + &futex2_wait_attr.attr, + &futex2_wake_attr.attr, + &futex2_waitv_attr.attr, + NULL, +}; + +static const struct attribute_group futex2_sysfs_attr_group = { + .attrs = futex2_sysfs_attrs, + .name = "futex2", +}; + +static int __init futex2_sysfs_init(void) +{ + return sysfs_create_group(kernel_kobj, &futex2_sysfs_attr_group); +} +subsys_initcall(futex2_sysfs_init); + +static int __init futex2_init(void) +{ + int i; + unsigned int futex_shift; + +#if CONFIG_BASE_SMALL + futex2_hashsize = 16; +#else + futex2_hashsize = roundup_pow_of_two(256 * num_possible_cpus()); +#endif + + futex_table = alloc_large_system_hash("futex2", sizeof(struct futex_bucket), + futex2_hashsize, 0, + futex2_hashsize < 256 ? HASH_SMALL : 0, + &futex_shift, NULL, + futex2_hashsize, futex2_hashsize); + futex2_hashsize = 1UL << futex_shift; + + BUG_ON(!is_power_of_2(futex2_hashsize)); + + for (i = 0; i < futex2_hashsize; i++) { + INIT_LIST_HEAD(&futex_table[i].list); + spin_lock_init(&futex_table[i].lock); + atomic_set(&futex_table[i].waiters, 0); + } + + return 0; +} +core_initcall(futex2_init); diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index d79ef2493a2..8aa20871e7d 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -108,6 +108,23 @@ config GENERIC_IRQ_RESERVATION_MODE config IRQ_FORCED_THREADING bool +config FORCE_IRQ_THREADING + bool "Make IRQ threading compulsory" + depends on IRQ_FORCED_THREADING + default n + help + + Make IRQ threading mandatory for any IRQ handlers that support it + instead of being optional and requiring the threadirqs kernel + parameter. Instead they can be optionally disabled with the + nothreadirqs kernel parameter. + + Enabling this may make some architectures not boot with runqueue + sharing and MuQSS. + + Enable if you are building for a desktop or low latency system, + otherwise say N. + config SPARSE_IRQ bool "Support sparse irq numbering" if MAY_HAVE_SPARSE_IRQ help diff --git a/kernel/irq/manage.c b/kernel/irq/manage.c index 4c14356543d..8cf5cae5284 100644 --- a/kernel/irq/manage.c +++ b/kernel/irq/manage.c @@ -25,9 +25,20 @@ #include "internals.h" #if defined(CONFIG_IRQ_FORCED_THREADING) && !defined(CONFIG_PREEMPT_RT) +#ifdef CONFIG_FORCE_IRQ_THREADING +__read_mostly bool force_irqthreads = true; +#else __read_mostly bool force_irqthreads; +#endif EXPORT_SYMBOL_GPL(force_irqthreads); +static int __init setup_noforced_irqthreads(char *arg) +{ + force_irqthreads = false; + return 0; +} +early_param("nothreadirqs", setup_noforced_irqthreads); + static int __init setup_forced_irqthreads(char *arg) { force_irqthreads = true; diff --git a/kernel/kthread.c b/kernel/kthread.c index 08931e525dd..e81e6a15e03 100644 --- a/kernel/kthread.c +++ b/kernel/kthread.c @@ -498,6 +498,34 @@ void kthread_bind(struct task_struct *p, unsigned int cpu) } EXPORT_SYMBOL(kthread_bind); +#if defined(CONFIG_SCHED_MUQSS) && defined(CONFIG_SMP) +extern void __do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask); + +/* + * new_kthread_bind is a special variant of __kthread_bind_mask. + * For new threads to work on muqss we want to call do_set_cpus_allowed + * without the task_cpu being set and the task rescheduled until they're + * rescheduled on their own so we call __do_set_cpus_allowed directly which + * only changes the cpumask. This is particularly important for smpboot threads + * to work. + */ +static void new_kthread_bind(struct task_struct *p, unsigned int cpu) +{ + unsigned long flags; + + if (WARN_ON(!wait_task_inactive(p, TASK_UNINTERRUPTIBLE))) + return; + + /* It's safe because the task is inactive. */ + raw_spin_lock_irqsave(&p->pi_lock, flags); + __do_set_cpus_allowed(p, cpumask_of(cpu)); + p->flags |= PF_NO_SETAFFINITY; + raw_spin_unlock_irqrestore(&p->pi_lock, flags); +} +#else +#define new_kthread_bind(p, cpu) kthread_bind(p, cpu) +#endif + /** * kthread_create_on_cpu - Create a cpu bound kthread * @threadfn: the function to run until signal_pending(current). @@ -518,7 +546,7 @@ struct task_struct *kthread_create_on_cpu(int (*threadfn)(void *data), cpu); if (IS_ERR(p)) return p; - kthread_bind(p, cpu); + new_kthread_bind(p, cpu); /* CPU hotplug need to bind once again when unparking the thread. */ to_kthread(p)->cpu = cpu; return p; @@ -1361,6 +1389,7 @@ void kthread_use_mm(struct mm_struct *mm) tsk->mm = mm; membarrier_update_current_mm(mm); switch_mm_irqs_off(active_mm, mm, tsk); + lru_gen_switch_mm(active_mm, mm); local_irq_enable(); task_unlock(tsk); #ifdef finish_arch_post_lock_switch diff --git a/kernel/locking/lockdep.c b/kernel/locking/lockdep.c index 9125bd41921..38b2a5a512e 100644 --- a/kernel/locking/lockdep.c +++ b/kernel/locking/lockdep.c @@ -189,7 +189,7 @@ static struct lock_class lock_classes[MAX_LOCKDEP_KEYS]; static DECLARE_BITMAP(lock_classes_in_use, MAX_LOCKDEP_KEYS); -static inline struct lock_class *hlock_class(struct held_lock *hlock) +inline struct lock_class *lockdep_hlock_class(struct held_lock *hlock) { unsigned int class_idx = hlock->class_idx; @@ -210,6 +210,8 @@ static inline struct lock_class *hlock_class(struct held_lock *hlock) */ return lock_classes + class_idx; } +EXPORT_SYMBOL_GPL(lockdep_hlock_class); +#define hlock_class(hlock) lockdep_hlock_class(hlock) #ifdef CONFIG_LOCK_STAT static DEFINE_PER_CPU(struct lock_class_stats[MAX_LOCKDEP_KEYS], cpu_lock_stats); diff --git a/kernel/sched/Makefile b/kernel/sched/Makefile index 5fc9c9b7086..1ff14a21193 100644 --- a/kernel/sched/Makefile +++ b/kernel/sched/Makefile @@ -22,15 +22,23 @@ ifneq ($(CONFIG_SCHED_OMIT_FRAME_POINTER),y) CFLAGS_core.o := $(PROFILING) -fno-omit-frame-pointer endif +ifdef CONFIG_SCHED_MUQSS +obj-y += MuQSS.o clock.o cputime.o +obj-y += idle.o +obj-y += wait.o wait_bit.o swait.o completion.o + +obj-$(CONFIG_SMP) += topology.o +else obj-y += core.o loadavg.o clock.o cputime.o obj-y += idle.o fair.o rt.o deadline.o obj-y += wait.o wait_bit.o swait.o completion.o obj-$(CONFIG_SMP) += cpupri.o cpudeadline.o topology.o stop_task.o pelt.o obj-$(CONFIG_SCHED_AUTOGROUP) += autogroup.o -obj-$(CONFIG_SCHEDSTATS) += stats.o obj-$(CONFIG_SCHED_DEBUG) += debug.o obj-$(CONFIG_CGROUP_CPUACCT) += cpuacct.o +endif +obj-$(CONFIG_SCHEDSTATS) += stats.o obj-$(CONFIG_CPU_FREQ) += cpufreq.o obj-$(CONFIG_CPU_FREQ_GOV_SCHEDUTIL) += cpufreq_schedutil.o obj-$(CONFIG_MEMBARRIER) += membarrier.o diff --git a/kernel/sched/MuQSS.c b/kernel/sched/MuQSS.c new file mode 100644 index 00000000000..cd6b334bb87 --- /dev/null +++ b/kernel/sched/MuQSS.c @@ -0,0 +1,8262 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * kernel/sched/MuQSS.c, was kernel/sched.c + * + * Kernel scheduler and related syscalls + * + * Copyright (C) 1991-2002 Linus Torvalds + * + * 1996-12-23 Modified by Dave Grothe to fix bugs in semaphores and + * make semaphores SMP safe + * 1998-11-19 Implemented schedule_timeout() and related stuff + * by Andrea Arcangeli + * 2002-01-04 New ultra-scalable O(1) scheduler by Ingo Molnar: + * hybrid priority-list and round-robin design with + * an array-switch method of distributing timeslices + * and per-CPU runqueues. Cleanups and useful suggestions + * by Davide Libenzi, preemptible kernel bits by Robert Love. + * 2003-09-03 Interactivity tuning by Con Kolivas. + * 2004-04-02 Scheduler domains code by Nick Piggin + * 2007-04-15 Work begun on replacing all interactivity tuning with a + * fair scheduling design by Con Kolivas. + * 2007-05-05 Load balancing (smp-nice) and other improvements + * by Peter Williams + * 2007-05-06 Interactivity improvements to CFS by Mike Galbraith + * 2007-07-01 Group scheduling enhancements by Srivatsa Vaddagiri + * 2007-11-29 RT balancing improvements by Steven Rostedt, Gregory Haskins, + * Thomas Gleixner, Mike Kravetz + * 2009-08-13 Brainfuck deadline scheduling policy by Con Kolivas deletes + * a whole lot of those previous things. + * 2016-10-01 Multiple Queue Skiplist Scheduler scalable evolution of BFS + * scheduler by Con Kolivas. + * 2019-08-31 LLC bits by Eduards Bezverhijs + */ +#define CREATE_TRACE_POINTS +#include +#undef CREATE_TRACE_POINTS + +#include +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include +#include + +#include "../workqueue_internal.h" +#include "../../fs/io-wq.h" +#include "../smpboot.h" + +#include "MuQSS.h" +#include "smp.h" + +#define rt_prio(prio) unlikely((prio) < MAX_RT_PRIO) +#define rt_task(p) rt_prio((p)->prio) +#define batch_task(p) (unlikely((p)->policy == SCHED_BATCH)) +#define is_rt_policy(policy) ((policy) == SCHED_FIFO || \ + (policy) == SCHED_RR) +#define has_rt_policy(p) unlikely(is_rt_policy((p)->policy)) + +#define is_idle_policy(policy) ((policy) == SCHED_IDLEPRIO) +#define idleprio_task(p) unlikely(is_idle_policy((p)->policy)) +#define task_running_idle(p) unlikely((p)->prio == IDLE_PRIO) + +#define is_iso_policy(policy) ((policy) == SCHED_ISO) +#define iso_task(p) unlikely(is_iso_policy((p)->policy)) +#define task_running_iso(p) unlikely((p)->prio == ISO_PRIO) + +#define rq_idle(rq) ((rq)->rq_prio == PRIO_LIMIT) + +#define ISO_PERIOD (5 * HZ) + +/* + * 'User priority' is the nice value converted to something we + * can work with better when scaling various scheduler parameters, + * it's a [ 0 ... 39 ] range. + */ +#define USER_PRIO(p) ((p)-MAX_RT_PRIO) +#define TASK_USER_PRIO(p) USER_PRIO((p)->static_prio) +#define MAX_USER_PRIO (USER_PRIO(MAX_PRIO)) +#define STOP_PRIO (MAX_RT_PRIO - 1) + +/* + * Some helpers for converting to/from various scales. Use shifts to get + * approximate multiples of ten for less overhead. + */ +#define APPROX_NS_PS (1073741824) /* Approximate ns per second */ +#define JIFFIES_TO_NS(TIME) ((TIME) * (APPROX_NS_PS / HZ)) +#define JIFFY_NS (APPROX_NS_PS / HZ) +#define JIFFY_US (1048576 / HZ) +#define NS_TO_JIFFIES(TIME) ((TIME) / JIFFY_NS) +#define HALF_JIFFY_NS (APPROX_NS_PS / HZ / 2) +#define HALF_JIFFY_US (1048576 / HZ / 2) +#define MS_TO_NS(TIME) ((TIME) << 20) +#define MS_TO_US(TIME) ((TIME) << 10) +#define NS_TO_MS(TIME) ((TIME) >> 20) +#define NS_TO_US(TIME) ((TIME) >> 10) +#define US_TO_NS(TIME) ((TIME) << 10) +#define TICK_APPROX_NS ((APPROX_NS_PS+HZ/2)/HZ) + +#define RESCHED_US (100) /* Reschedule if less than this many μs left */ + +void print_scheduler_version(void) +{ + printk(KERN_INFO "MuQSS CPU scheduler v0.210 by Con Kolivas.\n"); +} + +/* + * This is the time all tasks within the same priority round robin. + * Value is in ms and set to a minimum of 6ms. + * Tunable via /proc interface. + */ +#ifdef CONFIG_PCK_INTERACTIVE +int rr_interval __read_mostly = 4; +#else +int rr_interval __read_mostly = 6; +#endif + +/* + * Tunable to choose whether to prioritise latency or throughput, simple + * binary yes or no + */ +int sched_interactive __read_mostly = 1; + +/* + * sched_iso_cpu - sysctl which determines the cpu percentage SCHED_ISO tasks + * are allowed to run five seconds as real time tasks. This is the total over + * all online cpus. + */ +#ifdef CONFIG_PCK_INTERACTIVE +int sched_iso_cpu __read_mostly = 25; +#else +int sched_iso_cpu __read_mostly = 70; +#endif + +/* + * sched_yield_type - Choose what sort of yield sched_yield will perform. + * 0: No yield. + * 1: Yield only to better priority/deadline tasks. (default) + * 2: Expire timeslice and recalculate deadline. + */ + +#ifdef CONFIG_PCK_INTERACTIVE +int sched_yield_type __read_mostly = 0; +#else +int sched_yield_type __read_mostly = 1; +#endif + +/* + * The relative length of deadline for each priority(nice) level. + */ +static int prio_ratios[NICE_WIDTH] __read_mostly; + + +/* + * The quota handed out to tasks of all priority levels when refilling their + * time_slice. + */ +static inline int timeslice(void) +{ + return MS_TO_US(rr_interval); +} + +DEFINE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues); + +#ifdef CONFIG_SMP + +/* Define RQ share levels */ +#define RQSHARE_NONE 0 +#define RQSHARE_SMT 1 +#define RQSHARE_MC 2 +#define RQSHARE_MC_LLC 3 +#define RQSHARE_SMP 4 +#define RQSHARE_ALL 5 + +/* Define locality levels */ +#define LOCALITY_SAME 0 +#define LOCALITY_SMT 1 +#define LOCALITY_MC_LLC 2 +#define LOCALITY_MC 3 +#define LOCALITY_SMP 4 +#define LOCALITY_DISTANT 5 + +/* + * This determines what level of runqueue sharing will be done and is + * configurable at boot time with the bootparam rqshare = + */ +static int rqshare __read_mostly = CONFIG_SHARERQ; /* Default RQSHARE_MC */ + +static int __init set_rqshare(char *str) +{ + if (!strncmp(str, "none", 4)) { + rqshare = RQSHARE_NONE; + return 1; + } + if (!strncmp(str, "smt", 3)) { + rqshare = RQSHARE_SMT; + return 1; + } + if (!strncmp(str, "mc", 2)) { + rqshare = RQSHARE_MC; + return 1; + } + if (!strncmp(str, "llc", 3)) { + rqshare = RQSHARE_MC_LLC; + return 1; + } + if (!strncmp(str, "smp", 3)) { + rqshare = RQSHARE_SMP; + return 1; + } + if (!strncmp(str, "all", 3)) { + rqshare = RQSHARE_ALL; + return 1; + } + return 0; +} +__setup("rqshare=", set_rqshare); + +/* + * Total number of runqueues. Equals number of CPUs when there is no runqueue + * sharing but is usually less with SMT/MC sharing of runqueues. + */ +static int total_runqueues __read_mostly = 1; + +static cpumask_t cpu_idle_map ____cacheline_aligned_in_smp; + +/* + * For asym packing, by default the lower numbered cpu has higher priority. + */ +int __weak arch_asym_cpu_priority(int cpu) +{ + return -cpu; +} + +int __weak arch_sd_sibling_asym_packing(void) +{ + return 0*SD_ASYM_PACKING; +} + +#ifdef CONFIG_SCHED_SMT +DEFINE_STATIC_KEY_FALSE(sched_smt_present); +EXPORT_SYMBOL_GPL(sched_smt_present); +#endif + +#else +struct rq *uprq; +#endif /* CONFIG_SMP */ + +#include "stats.h" + +/* + * All common locking functions performed on rq->lock. rq->clock is local to + * the CPU accessing it so it can be modified just with interrupts disabled + * when we're not updating niffies. + * Looking up task_rq must be done under rq->lock to be safe. + */ + +/* + * RQ-clock updating methods: + */ + +#ifdef HAVE_SCHED_AVG_IRQ +static void update_irq_load_avg(struct rq *rq, long delta); +#else +static inline void update_irq_load_avg(struct rq *rq, long delta) {} +#endif + +static void update_rq_clock_task(struct rq *rq, s64 delta) +{ +/* + * In theory, the compile should just see 0 here, and optimize out the call + * to sched_rt_avg_update. But I don't trust it... + */ + s64 __maybe_unused steal = 0, irq_delta = 0; +#ifdef CONFIG_IRQ_TIME_ACCOUNTING + irq_delta = irq_time_read(cpu_of(rq)) - rq->prev_irq_time; + + /* + * Since irq_time is only updated on {soft,}irq_exit, we might run into + * this case when a previous update_rq_clock() happened inside a + * {soft,}irq region. + * + * When this happens, we stop ->clock_task and only update the + * prev_irq_time stamp to account for the part that fit, so that a next + * update will consume the rest. This ensures ->clock_task is + * monotonic. + * + * It does however cause some slight miss-attribution of {soft,}irq + * time, a more accurate solution would be to update the irq_time using + * the current rq->clock timestamp, except that would require using + * atomic ops. + */ + if (irq_delta > delta) + irq_delta = delta; + + rq->prev_irq_time += irq_delta; + delta -= irq_delta; +#endif +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING + if (static_key_false((¶virt_steal_rq_enabled))) { + steal = paravirt_steal_clock(cpu_of(rq)); + steal -= rq->prev_steal_time_rq; + + if (unlikely(steal > delta)) + steal = delta; + + rq->prev_steal_time_rq += steal; + delta -= steal; + } +#endif + rq->clock_task += delta; + +#ifdef CONFIG_HAVE_SCHED_AVG_IRQ + if (irq_delta + steal) + update_irq_load_avg(rq, irq_delta + steal); +#endif +} + +static inline void update_rq_clock(struct rq *rq) +{ + s64 delta = sched_clock_cpu(cpu_of(rq)) - rq->clock; + + if (unlikely(delta < 0)) + return; + rq->clock += delta; + update_rq_clock_task(rq, delta); +} + +/* + * Niffies are a globally increasing nanosecond counter. They're only used by + * update_load_avg and time_slice_expired, however deadlines are based on them + * across CPUs. Update them whenever we will call one of those functions, and + * synchronise them across CPUs whenever we hold both runqueue locks. + */ +static inline void update_clocks(struct rq *rq) +{ + s64 ndiff, minndiff; + long jdiff; + + update_rq_clock(rq); + ndiff = rq->clock - rq->old_clock; + rq->old_clock = rq->clock; + jdiff = jiffies - rq->last_jiffy; + + /* Subtract any niffies added by balancing with other rqs */ + ndiff -= rq->niffies - rq->last_niffy; + minndiff = JIFFIES_TO_NS(jdiff) - rq->niffies + rq->last_jiffy_niffies; + if (minndiff < 0) + minndiff = 0; + ndiff = max(ndiff, minndiff); + rq->niffies += ndiff; + rq->last_niffy = rq->niffies; + if (jdiff) { + rq->last_jiffy += jdiff; + rq->last_jiffy_niffies = rq->niffies; + } +} + +/* + * Any time we have two runqueues locked we use that as an opportunity to + * synchronise niffies to the highest value as idle ticks may have artificially + * kept niffies low on one CPU and the truth can only be later. + */ +static inline void synchronise_niffies(struct rq *rq1, struct rq *rq2) +{ + if (rq1->niffies > rq2->niffies) + rq2->niffies = rq1->niffies; + else + rq1->niffies = rq2->niffies; +} + +/* + * double_rq_lock - safely lock two runqueues + * + * Note this does not disable interrupts like task_rq_lock, + * you need to do so manually before calling. + */ + +/* For when we know rq1 != rq2 */ +static inline void __double_rq_lock(struct rq *rq1, struct rq *rq2) + __acquires(rq1->lock) + __acquires(rq2->lock) +{ + if (rq1 < rq2) { + raw_spin_lock(rq1->lock); + raw_spin_lock_nested(rq2->lock, SINGLE_DEPTH_NESTING); + } else { + raw_spin_lock(rq2->lock); + raw_spin_lock_nested(rq1->lock, SINGLE_DEPTH_NESTING); + } +} + +static inline void double_rq_lock(struct rq *rq1, struct rq *rq2) + __acquires(rq1->lock) + __acquires(rq2->lock) +{ + BUG_ON(!irqs_disabled()); + if (rq1->lock == rq2->lock) { + raw_spin_lock(rq1->lock); + __acquire(rq2->lock); /* Fake it out ;) */ + } else + __double_rq_lock(rq1, rq2); + synchronise_niffies(rq1, rq2); +} + +/* + * double_rq_unlock - safely unlock two runqueues + * + * Note this does not restore interrupts like task_rq_unlock, + * you need to do so manually after calling. + */ +static inline void double_rq_unlock(struct rq *rq1, struct rq *rq2) + __releases(rq1->lock) + __releases(rq2->lock) +{ + raw_spin_unlock(rq1->lock); + if (rq1->lock != rq2->lock) + raw_spin_unlock(rq2->lock); + else + __release(rq2->lock); +} + +static inline void lock_all_rqs(void) +{ + int cpu; + + preempt_disable(); + for_each_possible_cpu(cpu) { + struct rq *rq = cpu_rq(cpu); + + do_raw_spin_lock(rq->lock); + } +} + +static inline void unlock_all_rqs(void) +{ + int cpu; + + for_each_possible_cpu(cpu) { + struct rq *rq = cpu_rq(cpu); + + do_raw_spin_unlock(rq->lock); + } + preempt_enable(); +} + +/* Specially nest trylock an rq */ +static inline bool trylock_rq(struct rq *this_rq, struct rq *rq) +{ + if (unlikely(!do_raw_spin_trylock(rq->lock))) + return false; + spin_acquire(&rq->lock->dep_map, SINGLE_DEPTH_NESTING, 1, _RET_IP_); + synchronise_niffies(this_rq, rq); + return true; +} + +/* Unlock a specially nested trylocked rq */ +static inline void unlock_rq(struct rq *rq) +{ + spin_release(&rq->lock->dep_map, _RET_IP_); + do_raw_spin_unlock(rq->lock); +} + +/* + * cmpxchg based fetch_or, macro so it works for different integer types + */ +#define fetch_or(ptr, mask) \ + ({ \ + typeof(ptr) _ptr = (ptr); \ + typeof(mask) _mask = (mask); \ + typeof(*_ptr) _old, _val = *_ptr; \ + \ + for (;;) { \ + _old = cmpxchg(_ptr, _val, _val | _mask); \ + if (_old == _val) \ + break; \ + _val = _old; \ + } \ + _old; \ +}) + +#if defined(CONFIG_SMP) && defined(TIF_POLLING_NRFLAG) +/* + * Atomically set TIF_NEED_RESCHED and test for TIF_POLLING_NRFLAG, + * this avoids any races wrt polling state changes and thereby avoids + * spurious IPIs. + */ +static bool set_nr_and_not_polling(struct task_struct *p) +{ + struct thread_info *ti = task_thread_info(p); + return !(fetch_or(&ti->flags, _TIF_NEED_RESCHED) & _TIF_POLLING_NRFLAG); +} + +/* + * Atomically set TIF_NEED_RESCHED if TIF_POLLING_NRFLAG is set. + * + * If this returns true, then the idle task promises to call + * sched_ttwu_pending() and reschedule soon. + */ +static bool set_nr_if_polling(struct task_struct *p) +{ + struct thread_info *ti = task_thread_info(p); + typeof(ti->flags) old, val = READ_ONCE(ti->flags); + + for (;;) { + if (!(val & _TIF_POLLING_NRFLAG)) + return false; + if (val & _TIF_NEED_RESCHED) + return true; + old = cmpxchg(&ti->flags, val, val | _TIF_NEED_RESCHED); + if (old == val) + break; + val = old; + } + return true; +} + +#else +static bool set_nr_and_not_polling(struct task_struct *p) +{ + set_tsk_need_resched(p); + return true; +} + +#ifdef CONFIG_SMP +static bool set_nr_if_polling(struct task_struct *p) +{ + return false; +} +#endif +#endif + +static bool __wake_q_add(struct wake_q_head *head, struct task_struct *task) +{ + struct wake_q_node *node = &task->wake_q; + + /* + * Atomically grab the task, if ->wake_q is !nil already it means + * it's already queued (either by us or someone else) and will get the + * wakeup due to that. + * + * In order to ensure that a pending wakeup will observe our pending + * state, even in the failed case, an explicit smp_mb() must be used. + */ + smp_mb__before_atomic(); + if (unlikely(cmpxchg_relaxed(&node->next, NULL, WAKE_Q_TAIL))) + return false; + + /* + * The head is context local, there can be no concurrency. + */ + *head->lastp = node; + head->lastp = &node->next; + return true; +} + +/** + * wake_q_add() - queue a wakeup for 'later' waking. + * @head: the wake_q_head to add @task to + * @task: the task to queue for 'later' wakeup + * + * Queue a task for later wakeup, most likely by the wake_up_q() call in the + * same context, _HOWEVER_ this is not guaranteed, the wakeup can come + * instantly. + * + * This function must be used as-if it were wake_up_process(); IOW the task + * must be ready to be woken at this location. + */ +void wake_q_add(struct wake_q_head *head, struct task_struct *task) +{ + if (__wake_q_add(head, task)) + get_task_struct(task); +} + +/** + * wake_q_add_safe() - safely queue a wakeup for 'later' waking. + * @head: the wake_q_head to add @task to + * @task: the task to queue for 'later' wakeup + * + * Queue a task for later wakeup, most likely by the wake_up_q() call in the + * same context, _HOWEVER_ this is not guaranteed, the wakeup can come + * instantly. + * + * This function must be used as-if it were wake_up_process(); IOW the task + * must be ready to be woken at this location. + * + * This function is essentially a task-safe equivalent to wake_q_add(). Callers + * that already hold reference to @task can call the 'safe' version and trust + * wake_q to do the right thing depending whether or not the @task is already + * queued for wakeup. + */ +void wake_q_add_safe(struct wake_q_head *head, struct task_struct *task) +{ + if (!__wake_q_add(head, task)) + put_task_struct(task); +} + +void wake_up_q(struct wake_q_head *head) +{ + struct wake_q_node *node = head->first; + + while (node != WAKE_Q_TAIL) { + struct task_struct *task; + + task = container_of(node, struct task_struct, wake_q); + BUG_ON(!task); + /* Task can safely be re-inserted now */ + node = node->next; + task->wake_q.next = NULL; + + /* + * wake_up_process() executes a full barrier, which pairs with + * the queueing in wake_q_add() so as not to miss wakeups. + */ + wake_up_process(task); + put_task_struct(task); + } +} + +static inline void smp_sched_reschedule(int cpu) +{ + if (likely(cpu_online(cpu))) + smp_send_reschedule(cpu); +} + +/* + * resched_task - mark a task 'to be rescheduled now'. + * + * On UP this means the setting of the need_resched flag, on SMP it + * might also involve a cross-CPU call to trigger the scheduler on + * the target CPU. + */ +void resched_task(struct task_struct *p) +{ + int cpu; +#ifdef CONFIG_LOCKDEP + /* Kernel threads call this when creating workqueues while still + * inactive from __kthread_bind_mask, holding only the pi_lock */ + if (!(p->flags & PF_KTHREAD)) { + struct rq *rq = task_rq(p); + + lockdep_assert_held(rq->lock); + } +#endif + if (test_tsk_need_resched(p)) + return; + + cpu = task_cpu(p); + if (cpu == smp_processor_id()) { + set_tsk_need_resched(p); + set_preempt_need_resched(); + return; + } + + if (set_nr_and_not_polling(p)) + smp_sched_reschedule(cpu); + else + trace_sched_wake_idle_without_ipi(cpu); +} + +/* + * A task that is not running or queued will not have a node set. + * A task that is queued but not running will have a node set. + * A task that is currently running will have ->on_cpu set but no node set. + */ +static inline bool task_queued(struct task_struct *p) +{ + return !skiplist_node_empty(&p->node); +} + +static void enqueue_task(struct rq *rq, struct task_struct *p, int flags); +static inline void resched_if_idle(struct rq *rq); + +static inline bool deadline_before(u64 deadline, u64 time) +{ + return (deadline < time); +} + +/* + * Deadline is "now" in niffies + (offset by priority). Setting the deadline + * is the key to everything. It distributes cpu fairly amongst tasks of the + * same nice value, it proportions cpu according to nice level, it means the + * task that last woke up the longest ago has the earliest deadline, thus + * ensuring that interactive tasks get low latency on wake up. The CPU + * proportion works out to the square of the virtual deadline difference, so + * this equation will give nice 19 3% CPU compared to nice 0. + */ +static inline u64 prio_deadline_diff(int user_prio) +{ + return (prio_ratios[user_prio] * rr_interval * (MS_TO_NS(1) / 128)); +} + +static inline u64 task_deadline_diff(struct task_struct *p) +{ + return prio_deadline_diff(TASK_USER_PRIO(p)); +} + +static inline u64 static_deadline_diff(int static_prio) +{ + return prio_deadline_diff(USER_PRIO(static_prio)); +} + +static inline int longest_deadline_diff(void) +{ + return prio_deadline_diff(39); +} + +static inline int ms_longest_deadline_diff(void) +{ + return NS_TO_MS(longest_deadline_diff()); +} + +static inline bool rq_local(struct rq *rq); + +#ifndef SCHED_CAPACITY_SCALE +#define SCHED_CAPACITY_SCALE 1024 +#endif + +static inline int rq_load(struct rq *rq) +{ + return rq->nr_running; +} + +/* + * Update the load average for feeding into cpu frequency governors. Use a + * rough estimate of a rolling average with ~ time constant of 32ms. + * 80/128 ~ 0.63. * 80 / 32768 / 128 == * 5 / 262144 + * Make sure a call to update_clocks has been made before calling this to get + * an updated rq->niffies. + */ +static void update_load_avg(struct rq *rq, unsigned int flags) +{ + long us_interval, load; + + us_interval = NS_TO_US(rq->niffies - rq->load_update); + if (unlikely(us_interval <= 0)) + return; + + load = rq->load_avg - (rq->load_avg * us_interval * 5 / 262144); + if (unlikely(load < 0)) + load = 0; + load += rq_load(rq) * SCHED_CAPACITY_SCALE * us_interval * 5 / 262144; + rq->load_avg = load; + + rq->load_update = rq->niffies; + update_irq_load_avg(rq, 0); + if (likely(rq_local(rq))) + cpufreq_trigger(rq, flags); +} + +#ifdef HAVE_SCHED_AVG_IRQ +/* + * IRQ variant of update_load_avg below. delta is actually time in nanoseconds + * here so we scale curload to how long it's been since the last update. + */ +static void update_irq_load_avg(struct rq *rq, long delta) +{ + long us_interval, load; + + us_interval = NS_TO_US(rq->niffies - rq->irq_load_update); + if (unlikely(us_interval <= 0)) + return; + + load = rq->irq_load_avg - (rq->irq_load_avg * us_interval * 5 / 262144); + if (unlikely(load < 0)) + load = 0; + load += NS_TO_US(delta) * SCHED_CAPACITY_SCALE * 5 / 262144; + rq->irq_load_avg = load; + + rq->irq_load_update = rq->niffies; +} +#endif + +/* + * Removing from the runqueue. Enter with rq locked. Deleting a task + * from the skip list is done via the stored node reference in the task struct + * and does not require a full look up. Thus it occurs in O(k) time where k + * is the "level" of the list the task was stored at - usually < 4, max 8. + */ +static void dequeue_task(struct rq *rq, struct task_struct *p, int flags) +{ + skiplist_delete(rq->sl, &p->node); + rq->best_key = rq->node->next[0]->key; + update_clocks(rq); + + if (!(flags & DEQUEUE_SAVE)) { + sched_info_dequeued(rq, p); + psi_dequeue(p, flags & DEQUEUE_SLEEP); + } + rq->nr_running--; + if (rt_task(p)) + rq->rt_nr_running--; + update_load_avg(rq, flags); +} + +#ifdef CONFIG_PREEMPT_RCU +static bool rcu_read_critical(struct task_struct *p) +{ + return p->rcu_read_unlock_special.b.blocked; +} +#else /* CONFIG_PREEMPT_RCU */ +#define rcu_read_critical(p) (false) +#endif /* CONFIG_PREEMPT_RCU */ + +/* + * To determine if it's safe for a task of SCHED_IDLEPRIO to actually run as + * an idle task, we ensure none of the following conditions are met. + */ +static bool idleprio_suitable(struct task_struct *p) +{ + return (!(p->sched_contributes_to_load) && !(p->flags & (PF_EXITING)) && + !signal_pending(p) && !rcu_read_critical(p) && !freezing(p)); +} + +/* + * To determine if a task of SCHED_ISO can run in pseudo-realtime, we check + * that the iso_refractory flag is not set. + */ +static inline bool isoprio_suitable(struct rq *rq) +{ + return !rq->iso_refractory; +} + +static inline void inc_nr_running(struct rq *rq) +{ + rq->nr_running++; + if (trace_sched_update_nr_running_tp_enabled()) { + call_trace_sched_update_nr_running(rq, 1); + } +} + +static inline void dec_nr_running(struct rq *rq) +{ + rq->nr_running--; + if (trace_sched_update_nr_running_tp_enabled()) { + call_trace_sched_update_nr_running(rq, -1); + } +} + +/* + * Adding to the runqueue. Enter with rq locked. + */ +static void enqueue_task(struct rq *rq, struct task_struct *p, int flags) +{ + unsigned int randseed, cflags = 0; + u64 sl_id; + + if (!rt_task(p)) { + /* Check it hasn't gotten rt from PI */ + if ((idleprio_task(p) && idleprio_suitable(p)) || + (iso_task(p) && isoprio_suitable(rq))) + p->prio = p->normal_prio; + else + p->prio = NORMAL_PRIO; + } else + rq->rt_nr_running++; + /* + * The sl_id key passed to the skiplist generates a sorted list. + * Realtime and sched iso tasks run FIFO so they only need be sorted + * according to priority. The skiplist will put tasks of the same + * key inserted later in FIFO order. Tasks of sched normal, batch + * and idleprio are sorted according to their deadlines. Idleprio + * tasks are offset by an impossibly large deadline value ensuring + * they get sorted into last positions, but still according to their + * own deadlines. This creates a "landscape" of skiplists running + * from priority 0 realtime in first place to the lowest priority + * idleprio tasks last. Skiplist insertion is an O(log n) process. + */ + if (p->prio <= ISO_PRIO) { + sl_id = p->prio; + } else { + sl_id = p->deadline; + if (idleprio_task(p)) { + if (p->prio == IDLE_PRIO) + sl_id |= 0xF000000000000000; + else + sl_id += longest_deadline_diff(); + } + } + /* + * Some architectures don't have better than microsecond resolution + * so mask out ~microseconds as the random seed for skiplist insertion. + */ + update_clocks(rq); + if (!(flags & ENQUEUE_RESTORE)) { + sched_info_queued(rq, p); + psi_enqueue(p, flags & ENQUEUE_WAKEUP); + } + + randseed = (rq->niffies >> 10) & 0xFFFFFFFF; + skiplist_insert(rq->sl, &p->node, sl_id, p, randseed); + rq->best_key = rq->node->next[0]->key; + if (p->in_iowait) + cflags |= SCHED_CPUFREQ_IOWAIT; + inc_nr_running(rq); + update_load_avg(rq, cflags); +} + +/* + * Returns the relative length of deadline all compared to the shortest + * deadline which is that of nice -20. + */ +static inline int task_prio_ratio(struct task_struct *p) +{ + return prio_ratios[TASK_USER_PRIO(p)]; +} + +/* + * task_timeslice - all tasks of all priorities get the exact same timeslice + * length. CPU distribution is handled by giving different deadlines to + * tasks of different priorities. Use 128 as the base value for fast shifts. + */ +static inline int task_timeslice(struct task_struct *p) +{ + return (rr_interval * task_prio_ratio(p) / 128); +} + +#ifdef CONFIG_SMP +/* Entered with rq locked */ +static inline void resched_if_idle(struct rq *rq) +{ + if (rq_idle(rq)) + resched_task(rq->curr); +} + +static inline bool rq_local(struct rq *rq) +{ + return (rq->cpu == smp_processor_id()); +} +#ifdef CONFIG_SMT_NICE +static const cpumask_t *thread_cpumask(int cpu); + +/* Find the best real time priority running on any SMT siblings of cpu and if + * none are running, the static priority of the best deadline task running. + * The lookups to the other runqueues is done lockless as the occasional wrong + * value would be harmless. */ +static int best_smt_bias(struct rq *this_rq) +{ + int other_cpu, best_bias = 0; + + for_each_cpu(other_cpu, &this_rq->thread_mask) { + struct rq *rq = cpu_rq(other_cpu); + + if (rq_idle(rq)) + continue; + if (unlikely(!rq->online)) + continue; + if (!rq->rq_mm) + continue; + if (likely(rq->rq_smt_bias > best_bias)) + best_bias = rq->rq_smt_bias; + } + return best_bias; +} + +static int task_prio_bias(struct task_struct *p) +{ + if (rt_task(p)) + return 1 << 30; + else if (task_running_iso(p)) + return 1 << 29; + else if (task_running_idle(p)) + return 0; + return MAX_PRIO - p->static_prio; +} + +static bool smt_always_schedule(struct task_struct __maybe_unused *p, struct rq __maybe_unused *this_rq) +{ + return true; +} + +static bool (*smt_schedule)(struct task_struct *p, struct rq *this_rq) = &smt_always_schedule; + +/* We've already decided p can run on CPU, now test if it shouldn't for SMT + * nice reasons. */ +static bool smt_should_schedule(struct task_struct *p, struct rq *this_rq) +{ + int best_bias, task_bias; + + /* Kernel threads always run */ + if (unlikely(!p->mm)) + return true; + if (rt_task(p)) + return true; + if (!idleprio_suitable(p)) + return true; + best_bias = best_smt_bias(this_rq); + /* The smt siblings are all idle or running IDLEPRIO */ + if (best_bias < 1) + return true; + task_bias = task_prio_bias(p); + if (task_bias < 1) + return false; + if (task_bias >= best_bias) + return true; + /* Dither 25% cpu of normal tasks regardless of nice difference */ + if (best_bias % 4 == 1) + return true; + /* Sorry, you lose */ + return false; +} +#else /* CONFIG_SMT_NICE */ +#define smt_schedule(p, this_rq) (true) +#endif /* CONFIG_SMT_NICE */ + +static inline void atomic_set_cpu(int cpu, cpumask_t *cpumask) +{ + set_bit(cpu, (volatile unsigned long *)cpumask); +} + +/* + * The cpu_idle_map stores a bitmap of all the CPUs currently idle to + * allow easy lookup of whether any suitable idle CPUs are available. + * It's cheaper to maintain a binary yes/no if there are any idle CPUs on the + * idle_cpus variable than to do a full bitmask check when we are busy. The + * bits are set atomically but read locklessly as occasional false positive / + * negative is harmless. + */ +static inline void set_cpuidle_map(int cpu) +{ + if (likely(cpu_online(cpu))) + atomic_set_cpu(cpu, &cpu_idle_map); +} + +static inline void atomic_clear_cpu(int cpu, cpumask_t *cpumask) +{ + clear_bit(cpu, (volatile unsigned long *)cpumask); +} + +static inline void clear_cpuidle_map(int cpu) +{ + atomic_clear_cpu(cpu, &cpu_idle_map); +} + +static bool suitable_idle_cpus(struct task_struct *p) +{ + return (cpumask_intersects(p->cpus_ptr, &cpu_idle_map)); +} + +/* + * Resched current on rq. We don't know if rq is local to this CPU nor if it + * is locked so we do not use an intermediate variable for the task to avoid + * having it dereferenced. + */ +static void resched_curr(struct rq *rq) +{ + int cpu; + + if (test_tsk_need_resched(rq->curr)) + return; + + rq->preempt = rq->curr; + cpu = rq->cpu; + + /* We're doing this without holding the rq lock if it's not task_rq */ + + if (cpu == smp_processor_id()) { + set_tsk_need_resched(rq->curr); + set_preempt_need_resched(); + return; + } + + if (set_nr_and_not_polling(rq->curr)) + smp_sched_reschedule(cpu); + else + trace_sched_wake_idle_without_ipi(cpu); +} + +#define CPUIDLE_DIFF_THREAD (1) +#define CPUIDLE_DIFF_CORE_LLC (2) +#define CPUIDLE_DIFF_CORE (4) +#define CPUIDLE_CACHE_BUSY (8) +#define CPUIDLE_DIFF_CPU (16) +#define CPUIDLE_THREAD_BUSY (32) +#define CPUIDLE_DIFF_NODE (64) + +/* + * The best idle CPU is chosen according to the CPUIDLE ranking above where the + * lowest value would give the most suitable CPU to schedule p onto next. The + * order works out to be the following: + * + * Same thread, idle or busy cache, idle or busy threads + * Other core, same cache, idle or busy cache, idle threads. + * Same node, other CPU, idle cache, idle threads. + * Same node, other CPU, busy cache, idle threads. + * Other core, same cache, busy threads. + * Same node, other CPU, busy threads. + * Other node, other CPU, idle cache, idle threads. + * Other node, other CPU, busy cache, idle threads. + * Other node, other CPU, busy threads. + */ +static int best_mask_cpu(int best_cpu, struct rq *rq, cpumask_t *tmpmask) +{ + int best_ranking = CPUIDLE_DIFF_NODE | CPUIDLE_THREAD_BUSY | + CPUIDLE_DIFF_CPU | CPUIDLE_CACHE_BUSY | CPUIDLE_DIFF_CORE | + CPUIDLE_DIFF_CORE_LLC | CPUIDLE_DIFF_THREAD; + int cpu_tmp; + + if (cpumask_test_cpu(best_cpu, tmpmask)) + goto out; + + for_each_cpu(cpu_tmp, tmpmask) { + int ranking, locality; + struct rq *tmp_rq; + + ranking = 0; + tmp_rq = cpu_rq(cpu_tmp); + + locality = rq->cpu_locality[cpu_tmp]; +#ifdef CONFIG_NUMA + if (locality > LOCALITY_SMP) + ranking |= CPUIDLE_DIFF_NODE; + else +#endif + if (locality > LOCALITY_MC) + ranking |= CPUIDLE_DIFF_CPU; +#ifdef CONFIG_SCHED_MC + else if (locality == LOCALITY_MC_LLC) + ranking |= CPUIDLE_DIFF_CORE_LLC; + else if (locality == LOCALITY_MC) + ranking |= CPUIDLE_DIFF_CORE; + if (!(tmp_rq->cache_idle(tmp_rq))) + ranking |= CPUIDLE_CACHE_BUSY; +#endif +#ifdef CONFIG_SCHED_SMT + if (locality == LOCALITY_SMT) + ranking |= CPUIDLE_DIFF_THREAD; +#endif + if (ranking < best_ranking +#ifdef CONFIG_SCHED_SMT + || (ranking == best_ranking && (tmp_rq->siblings_idle(tmp_rq))) +#endif + ) { + best_cpu = cpu_tmp; + best_ranking = ranking; + } + } +out: + return best_cpu; +} + +bool cpus_share_cache(int this_cpu, int that_cpu) +{ + struct rq *this_rq = cpu_rq(this_cpu); + + return (this_rq->cpu_locality[that_cpu] < LOCALITY_SMP); +} + +/* As per resched_curr but only will resched idle task */ +static inline void resched_idle(struct rq *rq) +{ + if (test_tsk_need_resched(rq->idle)) + return; + + rq->preempt = rq->idle; + + set_tsk_need_resched(rq->idle); + + if (rq_local(rq)) { + set_preempt_need_resched(); + return; + } + + smp_sched_reschedule(rq->cpu); +} + +DEFINE_PER_CPU(cpumask_t, idlemask); + +static struct rq *resched_best_idle(struct task_struct *p, int cpu) +{ + cpumask_t *tmpmask = &(per_cpu(idlemask, cpu)); + struct rq *rq; + int best_cpu; + + cpumask_and(tmpmask, p->cpus_ptr, &cpu_idle_map); + best_cpu = best_mask_cpu(cpu, task_rq(p), tmpmask); + rq = cpu_rq(best_cpu); + if (!smt_schedule(p, rq)) + return NULL; + rq->preempt = p; + resched_idle(rq); + return rq; +} + +static inline void resched_suitable_idle(struct task_struct *p) +{ + if (suitable_idle_cpus(p)) + resched_best_idle(p, task_cpu(p)); +} + +static inline struct rq *rq_order(struct rq *rq, int cpu) +{ + return rq->rq_order[cpu]; +} +#else /* CONFIG_SMP */ +static inline void set_cpuidle_map(int cpu) +{ +} + +static inline void clear_cpuidle_map(int cpu) +{ +} + +static inline bool suitable_idle_cpus(struct task_struct *p) +{ + return uprq->curr == uprq->idle; +} + +static inline void resched_suitable_idle(struct task_struct *p) +{ +} + +static inline void resched_curr(struct rq *rq) +{ + resched_task(rq->curr); +} + +static inline void resched_if_idle(struct rq *rq) +{ +} + +static inline bool rq_local(struct rq *rq) +{ + return true; +} + +static inline struct rq *rq_order(struct rq *rq, int cpu) +{ + return rq; +} + +static inline bool smt_schedule(struct task_struct *p, struct rq *rq) +{ + return true; +} +#endif /* CONFIG_SMP */ + +static inline int normal_prio(struct task_struct *p) +{ + if (has_rt_policy(p)) + return MAX_RT_PRIO - 1 - p->rt_priority; + if (idleprio_task(p)) + return IDLE_PRIO; + if (iso_task(p)) + return ISO_PRIO; + return NORMAL_PRIO; +} + +/* + * Calculate the current priority, i.e. the priority + * taken into account by the scheduler. This value might + * be boosted by RT tasks as it will be RT if the task got + * RT-boosted. If not then it returns p->normal_prio. + */ +static int effective_prio(struct task_struct *p) +{ + p->normal_prio = normal_prio(p); + /* + * If we are RT tasks or we were boosted to RT priority, + * keep the priority unchanged. Otherwise, update priority + * to the normal priority: + */ + if (!rt_prio(p->prio)) + return p->normal_prio; + return p->prio; +} + +/* + * activate_task - move a task to the runqueue. Enter with rq locked. + */ +static void activate_task(struct rq *rq, struct task_struct *p, int flags) +{ + resched_if_idle(rq); + + /* + * Sleep time is in units of nanosecs, so shift by 20 to get a + * milliseconds-range estimation of the amount of time that the task + * spent sleeping: + */ + if (unlikely(prof_on == SLEEP_PROFILING)) { + if (p->state == TASK_UNINTERRUPTIBLE) + profile_hits(SLEEP_PROFILING, (void *)get_wchan(p), + (rq->niffies - p->last_ran) >> 20); + } + + p->prio = effective_prio(p); + enqueue_task(rq, p, flags); + p->on_rq = TASK_ON_RQ_QUEUED; +} + +/* + * deactivate_task - If it's running, it's not on the runqueue and we can just + * decrement the nr_running. Enter with rq locked. + */ +static inline void deactivate_task(struct task_struct *p, struct rq *rq) +{ + p->on_rq = 0; + sched_info_dequeued(rq, p); + /* deactivate_task is always DEQUEUE_SLEEP in muqss */ + psi_dequeue(p, DEQUEUE_SLEEP); +} + +#ifdef CONFIG_SMP +void set_task_cpu(struct task_struct *p, unsigned int new_cpu) +{ + struct rq *rq; + + if (task_cpu(p) == new_cpu) + return; + + /* Do NOT call set_task_cpu on a currently queued task as we will not + * be reliably holding the rq lock after changing CPU. */ + BUG_ON(task_queued(p)); + rq = task_rq(p); + +#ifdef CONFIG_LOCKDEP + /* + * The caller should hold either p->pi_lock or rq->lock, when changing + * a task's CPU. ->pi_lock for waking tasks, rq->lock for runnable tasks. + * + * Furthermore, all task_rq users should acquire both locks, see + * task_rq_lock(). + */ + WARN_ON_ONCE(debug_locks && !(lockdep_is_held(&p->pi_lock) || + lockdep_is_held(rq->lock))); +#endif + + trace_sched_migrate_task(p, new_cpu); + rseq_migrate(p); + perf_event_task_migrate(p); + + /* + * After ->cpu is set up to a new value, task_rq_lock(p, ...) can be + * successfully executed on another CPU. We must ensure that updates of + * per-task data have been completed by this moment. + */ + smp_wmb(); + + p->wake_cpu = new_cpu; + + if (task_running(rq, p)) { + /* + * We should only be calling this on a running task if we're + * holding rq lock. + */ + lockdep_assert_held(rq->lock); + + /* + * We can't change the task_thread_info CPU on a running task + * as p will still be protected by the rq lock of the CPU it + * is still running on so we only set the wake_cpu for it to be + * lazily updated once off the CPU. + */ + return; + } + +#ifdef CONFIG_THREAD_INFO_IN_TASK + WRITE_ONCE(p->cpu, new_cpu); +#else + WRITE_ONCE(task_thread_info(p)->cpu, new_cpu); +#endif + /* We're no longer protecting p after this point since we're holding + * the wrong runqueue lock. */ +} +#endif /* CONFIG_SMP */ + +/* + * Move a task off the runqueue and take it to a cpu for it will + * become the running task. + */ +static inline void take_task(struct rq *rq, int cpu, struct task_struct *p) +{ + struct rq *p_rq = task_rq(p); + + dequeue_task(p_rq, p, DEQUEUE_SAVE); + if (p_rq != rq) { + sched_info_dequeued(p_rq, p); + sched_info_queued(rq, p); + } + set_task_cpu(p, cpu); +} + +/* + * Returns a descheduling task to the runqueue unless it is being + * deactivated. + */ +static inline void return_task(struct task_struct *p, struct rq *rq, + int cpu, bool deactivate) +{ + if (deactivate) + deactivate_task(p, rq); + else { +#ifdef CONFIG_SMP + /* + * set_task_cpu was called on the running task that doesn't + * want to deactivate so it has to be enqueued to a different + * CPU and we need its lock. Tag it to be moved with as the + * lock is dropped in finish_lock_switch. + */ + if (unlikely(p->wake_cpu != cpu)) + WRITE_ONCE(p->on_rq, TASK_ON_RQ_MIGRATING); + else +#endif + enqueue_task(rq, p, ENQUEUE_RESTORE); + } +} + +/* Enter with rq lock held. We know p is on the local cpu */ +static inline void __set_tsk_resched(struct task_struct *p) +{ + set_tsk_need_resched(p); + set_preempt_need_resched(); +} + +/** + * task_curr - is this task currently executing on a CPU? + * @p: the task in question. + * + * Return: 1 if the task is currently executing. 0 otherwise. + */ +inline int task_curr(const struct task_struct *p) +{ + return cpu_curr(task_cpu(p)) == p; +} + +#ifdef CONFIG_SMP +/* + * wait_task_inactive - wait for a thread to unschedule. + * + * If @match_state is nonzero, it's the @p->state value just checked and + * not expected to change. If it changes, i.e. @p might have woken up, + * then return zero. When we succeed in waiting for @p to be off its CPU, + * we return a positive number (its total switch count). If a second call + * a short while later returns the same number, the caller can be sure that + * @p has remained unscheduled the whole time. + * + * The caller must ensure that the task *will* unschedule sometime soon, + * else this function might spin for a *long* time. This function can't + * be called with interrupts off, or it may introduce deadlock with + * smp_call_function() if an IPI is sent by the same process we are + * waiting to become inactive. + */ +unsigned long wait_task_inactive(struct task_struct *p, long match_state) +{ + int running, queued; + struct rq_flags rf; + unsigned long ncsw; + struct rq *rq; + + for (;;) { + rq = task_rq(p); + + /* + * If the task is actively running on another CPU + * still, just relax and busy-wait without holding + * any locks. + * + * NOTE! Since we don't hold any locks, it's not + * even sure that "rq" stays as the right runqueue! + * But we don't care, since this will return false + * if the runqueue has changed and p is actually now + * running somewhere else! + */ + while (task_running(rq, p)) { + if (match_state && unlikely(p->state != match_state)) + return 0; + cpu_relax(); + } + + /* + * Ok, time to look more closely! We need the rq + * lock now, to be *sure*. If we're wrong, we'll + * just go back and repeat. + */ + rq = task_rq_lock(p, &rf); + trace_sched_wait_task(p); + running = task_running(rq, p); + queued = task_on_rq_queued(p); + ncsw = 0; + if (!match_state || p->state == match_state) + ncsw = p->nvcsw | LONG_MIN; /* sets MSB */ + task_rq_unlock(rq, p, &rf); + + /* + * If it changed from the expected state, bail out now. + */ + if (unlikely(!ncsw)) + break; + + /* + * Was it really running after all now that we + * checked with the proper locks actually held? + * + * Oops. Go back and try again.. + */ + if (unlikely(running)) { + cpu_relax(); + continue; + } + + /* + * It's not enough that it's not actively running, + * it must be off the runqueue _entirely_, and not + * preempted! + * + * So if it was still runnable (but just not actively + * running right now), it's preempted, and we should + * yield - it could be a while. + */ + if (unlikely(queued)) { + ktime_t to = NSEC_PER_SEC / HZ; + + set_current_state(TASK_UNINTERRUPTIBLE); + schedule_hrtimeout(&to, HRTIMER_MODE_REL); + continue; + } + + /* + * Ahh, all good. It wasn't running, and it wasn't + * runnable, which means that it will never become + * running in the future either. We're all done! + */ + break; + } + + return ncsw; +} + +/*** + * kick_process - kick a running thread to enter/exit the kernel + * @p: the to-be-kicked thread + * + * Cause a process which is running on another CPU to enter + * kernel-mode, without any delay. (to get signals handled.) + * + * NOTE: this function doesn't have to take the runqueue lock, + * because all it wants to ensure is that the remote task enters + * the kernel. If the IPI races and the task has been migrated + * to another CPU then no harm is done and the purpose has been + * achieved as well. + */ +void kick_process(struct task_struct *p) +{ + int cpu; + + preempt_disable(); + cpu = task_cpu(p); + if ((cpu != smp_processor_id()) && task_curr(p)) + smp_sched_reschedule(cpu); + preempt_enable(); +} +EXPORT_SYMBOL_GPL(kick_process); +#endif + +/* + * RT tasks preempt purely on priority. SCHED_NORMAL tasks preempt on the + * basis of earlier deadlines. SCHED_IDLEPRIO don't preempt anything else or + * between themselves, they cooperatively multitask. An idle rq scores as + * prio PRIO_LIMIT so it is always preempted. + */ +static inline bool +can_preempt(struct task_struct *p, int prio, u64 deadline) +{ + /* Better static priority RT task or better policy preemption */ + if (p->prio < prio) + return true; + if (p->prio > prio) + return false; + if (p->policy == SCHED_BATCH) + return false; + /* SCHED_NORMAL and ISO will preempt based on deadline */ + if (!deadline_before(p->deadline, deadline)) + return false; + return true; +} + +#ifdef CONFIG_SMP + +/* + * Per-CPU kthreads are allowed to run on !active && online CPUs, see + * __set_cpus_allowed_ptr(). + */ +static inline bool is_cpu_allowed(struct task_struct *p, int cpu) +{ + if (!cpumask_test_cpu(cpu, p->cpus_ptr)) + return false; + + if (!(p->flags & PF_KTHREAD)) + return cpu_active(cpu); + + /* KTHREAD_IS_PER_CPU is always allowed. */ + if (kthread_is_per_cpu(p)) + return cpu_online(cpu); + + /* Regular kernel threads don't get to stay during offline. */ + if (cpu_dying(cpu)) + return false; + + /* But are allowed during online. */ + return cpu_online(cpu); +} + +/* + * Check to see if p can run on cpu, and if not, whether there are any online + * CPUs it can run on instead. This only happens with the hotplug threads that + * bring up the CPUs. + */ +static inline bool sched_other_cpu(struct task_struct *p, int cpu) +{ + if (likely(cpumask_test_cpu(cpu, p->cpus_ptr))) + return false; + if (p->nr_cpus_allowed == 1) { + cpumask_t valid_mask; + + cpumask_and(&valid_mask, p->cpus_ptr, cpu_online_mask); + if (unlikely(cpumask_empty(&valid_mask))) + return false; + } + return true; +} + +static inline bool needs_other_cpu(struct task_struct *p, int cpu) +{ + if (cpumask_test_cpu(cpu, p->cpus_ptr)) + return false; + return true; +} + +#define cpu_online_map (*(cpumask_t *)cpu_online_mask) + +static void try_preempt(struct task_struct *p, struct rq *this_rq) +{ + int i, this_entries = rq_load(this_rq); + cpumask_t tmp; + + if (suitable_idle_cpus(p) && resched_best_idle(p, task_cpu(p))) + return; + + /* IDLEPRIO tasks never preempt anything but idle */ + if (p->policy == SCHED_IDLEPRIO) + return; + + cpumask_and(&tmp, &cpu_online_map, p->cpus_ptr); + + for (i = 0; i < num_online_cpus(); i++) { + struct rq *rq = this_rq->cpu_order[i]; + + if (!cpumask_test_cpu(rq->cpu, &tmp)) + continue; + + if (!sched_interactive && rq != this_rq && rq_load(rq) <= this_entries) + continue; + if (smt_schedule(p, rq) && can_preempt(p, rq->rq_prio, rq->rq_deadline)) { + /* We set rq->preempting lockless, it's a hint only */ + rq->preempting = p; + resched_curr(rq); + return; + } + } +} + +static int __set_cpus_allowed_ptr(struct task_struct *p, + const struct cpumask *new_mask, + u32 flags); +#else /* CONFIG_SMP */ +static inline bool needs_other_cpu(struct task_struct *p, int cpu) +{ + return false; +} + +static void try_preempt(struct task_struct *p, struct rq *this_rq) +{ + if (p->policy == SCHED_IDLEPRIO) + return; + if (can_preempt(p, uprq->rq_prio, uprq->rq_deadline)) + resched_curr(uprq); +} + +static inline int __set_cpus_allowed_ptr(struct task_struct *p, + const struct cpumask *new_mask, + u32 __always_unused flags) +{ + return set_cpus_allowed_ptr(p, new_mask); +} +#endif /* CONFIG_SMP */ + +static void +ttwu_stat(struct task_struct *p, int cpu, int wake_flags) +{ + struct rq *rq; + + if (!schedstat_enabled()) + return; + + rq = this_rq(); + +#ifdef CONFIG_SMP + if (cpu == rq->cpu) { + __schedstat_inc(rq->ttwu_local); + } else { + struct sched_domain *sd; + + rcu_read_lock(); + for_each_domain(rq->cpu, sd) { + if (cpumask_test_cpu(cpu, sched_domain_span(sd))) { + __schedstat_inc(sd->ttwu_wake_remote); + break; + } + } + rcu_read_unlock(); + } + +#endif /* CONFIG_SMP */ + + __schedstat_inc(rq->ttwu_count); +} + +/* + * Mark the task runnable and perform wakeup-preemption. + */ +static void ttwu_do_wakeup(struct rq *rq, struct task_struct *p, int wake_flags) +{ + /* + * Sync wakeups (i.e. those types of wakeups where the waker + * has indicated that it will leave the CPU in short order) + * don't trigger a preemption if there are no idle cpus, + * instead waiting for current to deschedule. + */ + if (wake_flags & WF_SYNC) + resched_suitable_idle(p); + else + try_preempt(p, rq); + p->state = TASK_RUNNING; + trace_sched_wakeup(p); +} + +static void +ttwu_do_activate(struct rq *rq, struct task_struct *p, int wake_flags) +{ + int en_flags = ENQUEUE_WAKEUP; + + lockdep_assert_held(rq->lock); + + if (p->sched_contributes_to_load) + rq->nr_uninterruptible--; + +#ifdef CONFIG_SMP + if (wake_flags & WF_MIGRATED) + en_flags |= ENQUEUE_MIGRATED; + else +#endif + if (p->in_iowait) { + delayacct_blkio_end(p); + atomic_dec(&task_rq(p)->nr_iowait); + } + + activate_task(rq, p, en_flags); + ttwu_do_wakeup(rq, p, wake_flags); +} + +/* + * Consider @p being inside a wait loop: + * + * for (;;) { + * set_current_state(TASK_UNINTERRUPTIBLE); + * + * if (CONDITION) + * break; + * + * schedule(); + * } + * __set_current_state(TASK_RUNNING); + * + * between set_current_state() and schedule(). In this case @p is still + * runnable, so all that needs doing is change p->state back to TASK_RUNNING in + * an atomic manner. + * + * By taking task_rq(p)->lock we serialize against schedule(), if @p->on_rq + * then schedule() must still happen and p->state can be changed to + * TASK_RUNNING. Otherwise we lost the race, schedule() has happened, and we + * need to do a full wakeup with enqueue. + * + * Returns: %true when the wakeup is done, + * %false otherwise. + */ +static int ttwu_runnable(struct task_struct *p, int wake_flags) +{ + struct rq *rq; + int ret = 0; + + rq = __task_rq_lock(p, NULL); + if (likely(task_on_rq_queued(p))) { + ttwu_do_wakeup(rq, p, wake_flags); + ret = 1; + } + __task_rq_unlock(rq, NULL); + + return ret; +} + +#ifdef CONFIG_SMP +void sched_ttwu_pending(void *arg) +{ + struct llist_node *llist = arg; + struct rq *rq = this_rq(); + struct task_struct *p, *t; + struct rq_flags rf; + + if (!llist) + return; + + /* + * rq::ttwu_pending racy indication of out-standing wakeups. + * Races such that false-negatives are possible, since they + * are shorter lived that false-positives would be. + */ + WRITE_ONCE(rq->ttwu_pending, 0); + + rq_lock_irqsave(rq, &rf); + + llist_for_each_entry_safe(p, t, llist, wake_entry.llist) { + if (WARN_ON_ONCE(p->on_cpu)) + smp_cond_load_acquire(&p->on_cpu, !VAL); + + if (WARN_ON_ONCE(task_cpu(p) != cpu_of(rq))) + set_task_cpu(p, cpu_of(rq)); + + ttwu_do_activate(rq, p, 0); + } + + rq_unlock_irqrestore(rq, &rf); +} + +void send_call_function_single_ipi(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + + if (!set_nr_if_polling(rq->idle)) + arch_send_call_function_single_ipi(cpu); + else + trace_sched_wake_idle_without_ipi(cpu); +} + +/* + * Queue a task on the target CPUs wake_list and wake the CPU via IPI if + * necessary. The wakee CPU on receipt of the IPI will queue the task + * via sched_ttwu_wakeup() for activation so the wakee incurs the cost + * of the wakeup instead of the waker. + */ +static void __ttwu_queue_wakelist(struct task_struct *p, int cpu, int wake_flags) +{ + struct rq *rq = cpu_rq(cpu); + + WRITE_ONCE(rq->ttwu_pending, 1); + __smp_call_single_queue(cpu, &p->wake_entry.llist); +} + +void wake_up_if_idle(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + struct rq_flags rf; + + rcu_read_lock(); + + if (!is_idle_task(rcu_dereference(rq->curr))) + goto out; + + if (set_nr_if_polling(rq->idle)) { + trace_sched_wake_idle_without_ipi(cpu); + } else { + rq_lock_irqsave(rq, &rf); + if (likely(is_idle_task(rq->curr))) + smp_sched_reschedule(cpu); + /* Else cpu is not in idle, do nothing here */ + rq_unlock_irqrestore(rq, &rf); + } + +out: + rcu_read_unlock(); +} + +static inline bool ttwu_queue_cond(int cpu, int wake_flags) +{ + /* + * Do not complicate things with the async wake_list while the CPU is + * in hotplug state. + */ + if (!cpu_active(cpu)) + return false; + + /* + * If the CPU does not share cache, then queue the task on the + * remote rqs wakelist to avoid accessing remote data. + */ + if (!cpus_share_cache(smp_processor_id(), cpu)) + return true; + + /* + * If the task is descheduling and the only running task on the + * CPU then use the wakelist to offload the task activation to + * the soon-to-be-idle CPU as the current CPU is likely busy. + * nr_running is checked to avoid unnecessary task stacking. + */ + if ((wake_flags & WF_ON_CPU) && cpu_rq(cpu)->nr_running <= 1) + return true; + + return false; +} + +static bool ttwu_queue_wakelist(struct task_struct *p, int cpu, int wake_flags) +{ + /* CFS would require sched_feat(TTWU_QUEUE) here but that is + * fixed enabled */ + if (ttwu_queue_cond(cpu, wake_flags)) { + if (WARN_ON_ONCE(cpu == smp_processor_id())) + return false; + + sched_clock_cpu(cpu); /* Sync clocks across CPUs */ + __ttwu_queue_wakelist(p, cpu, wake_flags); + return true; + } + + return false; +} + +static int valid_task_cpu(struct task_struct *p) +{ + cpumask_t valid_mask; + + if (p->flags & PF_KTHREAD) + cpumask_and(&valid_mask, p->cpus_ptr, cpu_all_mask); + else + cpumask_and(&valid_mask, p->cpus_ptr, cpu_active_mask); + + if (unlikely(!cpumask_weight(&valid_mask))) { + /* We shouldn't be hitting this any more */ + printk(KERN_WARNING "SCHED: No cpumask for %s/%d weight %d\n", p->comm, + p->pid, cpumask_weight(p->cpus_ptr)); + return cpumask_any(p->cpus_ptr); + } + return cpumask_any(&valid_mask); +} + +/* + * For a task that's just being woken up we have a valuable balancing + * opportunity so choose the nearest cache most lightly loaded runqueue. + * Entered with rq locked and returns with the chosen runqueue locked. + */ +static inline int select_best_cpu(struct task_struct *p) +{ + unsigned int idlest = ~0U; + struct rq *rq = NULL; + int i; + + if (suitable_idle_cpus(p)) { + int cpu = task_cpu(p); + + if (unlikely(needs_other_cpu(p, cpu))) + cpu = valid_task_cpu(p); + rq = resched_best_idle(p, cpu); + if (likely(rq)) + return rq->cpu; + } + + for (i = 0; i < num_online_cpus(); i++) { + struct rq *other_rq = task_rq(p)->cpu_order[i]; + int entries; + + if (!other_rq->online) + continue; + if (needs_other_cpu(p, other_rq->cpu)) + continue; + entries = rq_load(other_rq); + if (entries >= idlest) + continue; + idlest = entries; + rq = other_rq; + } + if (unlikely(!rq)) + return task_cpu(p); + return rq->cpu; +} +#else /* CONFIG_SMP */ + +static inline bool ttwu_queue_wakelist(struct task_struct *p, int cpu, int wake_flags) +{ + return false; +} + +static int valid_task_cpu(struct task_struct *p) +{ + return 0; +} + +static inline int select_best_cpu(struct task_struct *p) +{ + return 0; +} + +static struct rq *resched_best_idle(struct task_struct *p, int cpu) +{ + return NULL; +} +#endif /* CONFIG_SMP */ + +static void ttwu_queue(struct task_struct *p, int cpu, int wake_flags) +{ + struct rq *rq = cpu_rq(cpu); + + if (ttwu_queue_wakelist(p, cpu, wake_flags)) + return; + + rq_lock(rq); + update_rq_clock(rq); + ttwu_do_activate(rq, p, wake_flags); + rq_unlock(rq); +} + +/*** + * try_to_wake_up - wake up a thread + * @p: the thread to be awakened + * @state: the mask of task states that can be woken + * @wake_flags: wake modifier flags (WF_*) + * + * Put it on the run-queue if it's not already there. The "current" + * thread is always on the run-queue (except when the actual + * re-schedule is in progress), and as such you're allowed to do + * the simpler "current->state = TASK_RUNNING" to mark yourself + * runnable without the overhead of this. + * + * Return: %true if @p was woken up, %false if it was already running. + * or @state didn't match @p's state. + */ +static int +try_to_wake_up(struct task_struct *p, unsigned int state, int wake_flags) +{ + unsigned long flags; + int cpu, success = 0; + + preempt_disable(); + if (p == current) { + /* + * We're waking current, this means 'p->on_rq' and 'task_cpu(p) + * == smp_processor_id()'. Together this means we can special + * case the whole 'p->on_rq && ttwu_runnable()' case below + * without taking any locks. + * + * In particular: + * - we rely on Program-Order guarantees for all the ordering, + * - we're serialized against set_special_state() by virtue of + * it disabling IRQs (this allows not taking ->pi_lock). + */ + if (!(p->state & state)) + goto out; + + success = 1; + trace_sched_waking(p); + p->state = TASK_RUNNING; + trace_sched_wakeup(p); + goto out; + } + + /* + * If we are going to wake up a thread waiting for CONDITION we + * need to ensure that CONDITION=1 done by the caller can not be + * reordered with p->state check below. This pairs with smp_store_mb() + * in set_current_state() that the waiting thread does. + */ + raw_spin_lock_irqsave(&p->pi_lock, flags); + smp_mb__after_spinlock(); + if (!(p->state & state)) + goto unlock; + + trace_sched_waking(p); + + /* We're going to change ->state: */ + success = 1; + + /* + * Ensure we load p->on_rq _after_ p->state, otherwise it would + * be possible to, falsely, observe p->on_rq == 0 and get stuck + * in smp_cond_load_acquire() below. + * + * sched_ttwu_pending() try_to_wake_up() + * STORE p->on_rq = 1 LOAD p->state + * UNLOCK rq->lock + * + * __schedule() (switch to task 'p') + * LOCK rq->lock smp_rmb(); + * smp_mb__after_spinlock(); + * UNLOCK rq->lock + * + * [task p] + * STORE p->state = UNINTERRUPTIBLE LOAD p->on_rq + * + * Pairs with the LOCK+smp_mb__after_spinlock() on rq->lock in + * __schedule(). See the comment for smp_mb__after_spinlock(). + */ + smp_rmb(); + if (READ_ONCE(p->on_rq) && ttwu_runnable(p, wake_flags)) + goto unlock; + +#ifdef CONFIG_SMP + /* + * Ensure we load p->on_cpu _after_ p->on_rq, otherwise it would be + * possible to, falsely, observe p->on_cpu == 0. + * + * One must be running (->on_cpu == 1) in order to remove oneself + * from the runqueue. + * + * __schedule() (switch to task 'p') try_to_wake_up() + * STORE p->on_cpu = 1 LOAD p->on_rq + * UNLOCK rq->lock + * + * __schedule() (put 'p' to sleep) + * LOCK rq->lock smp_rmb(); + * smp_mb__after_spinlock(); + * STORE p->on_rq = 0 LOAD p->on_cpu + * + * Pairs with the LOCK+smp_mb__after_spinlock() on rq->lock in + * __schedule(). See the comment for smp_mb__after_spinlock(). + * + * Form a control-dep-acquire with p->on_rq == 0 above, to ensure + * schedule()'s deactivate_task() has 'happened' and p will no longer + * care about it's own p->state. See the comment in __schedule(). + */ + smp_acquire__after_ctrl_dep(); + + /* + * We're doing the wakeup (@success == 1), they did a dequeue (p->on_rq + * == 0), which means we need to do an enqueue, change p->state to + * TASK_WAKING such that we can unlock p->pi_lock before doing the + * enqueue, such as ttwu_queue_wakelist(). + */ + p->state = TASK_WAKING; + + /* + * If the owning (remote) CPU is still in the middle of schedule() with + * this task as prev, considering queueing p on the remote CPUs wake_list + * which potentially sends an IPI instead of spinning on p->on_cpu to + * let the waker make forward progress. This is safe because IRQs are + * disabled and the IPI will deliver after on_cpu is cleared. + * + * Ensure we load task_cpu(p) after p->on_cpu: + * + * set_task_cpu(p, cpu); + * STORE p->cpu = @cpu + * __schedule() (switch to task 'p') + * LOCK rq->lock + * smp_mb__after_spin_lock() smp_cond_load_acquire(&p->on_cpu) + * STORE p->on_cpu = 1 LOAD p->cpu + * + * to ensure we observe the correct CPU on which the task is currently + * scheduling. + */ + if (smp_load_acquire(&p->on_cpu) && + ttwu_queue_wakelist(p, task_cpu(p), wake_flags | WF_ON_CPU)) + goto unlock; + + /* + * If the owning (remote) CPU is still in the middle of schedule() with + * this task as prev, wait until it's done referencing the task. + * + * Pairs with the smp_store_release() in finish_task(). + * + * This ensures that tasks getting woken will be fully ordered against + * their previous state and preserve Program Order. + */ + smp_cond_load_acquire(&p->on_cpu, !VAL); + + cpu = select_best_cpu(p); + if (task_cpu(p) != cpu) { + if (p->in_iowait) { + delayacct_blkio_end(p); + atomic_dec(&task_rq(p)->nr_iowait); + } + + wake_flags |= WF_MIGRATED; + psi_ttwu_dequeue(p); + set_task_cpu(p, cpu); + } + +#else + cpu = task_cpu(p); +#endif /* CONFIG_SMP */ + + ttwu_queue(p, cpu, wake_flags); +unlock: + raw_spin_unlock_irqrestore(&p->pi_lock, flags); +out: + if (success) + ttwu_stat(p, task_cpu(p), wake_flags); + preempt_enable(); + + return success; +} + +/** + * try_invoke_on_locked_down_task - Invoke a function on task in fixed state + * @p: Process for which the function is to be invoked, can be @current. + * @func: Function to invoke. + * @arg: Argument to function. + * + * If the specified task can be quickly locked into a definite state + * (either sleeping or on a given runqueue), arrange to keep it in that + * state while invoking @func(@arg). This function can use ->on_rq and + * task_curr() to work out what the state is, if required. Given that + * @func can be invoked with a runqueue lock held, it had better be quite + * lightweight. + * + * Returns: + * @false if the task slipped out from under the locks. + * @true if the task was locked onto a runqueue or is sleeping. + * However, @func can override this by returning @false. + */ +bool try_invoke_on_locked_down_task(struct task_struct *p, bool (*func)(struct task_struct *t, void *arg), void *arg) +{ + struct rq_flags rf; + bool ret = false; + struct rq *rq; + + raw_spin_lock_irqsave(&p->pi_lock, rf.flags); + if (p->on_rq) { + rq = __task_rq_lock(p, NULL); + if (task_rq(p) == rq) + ret = func(p, arg); + rq_unlock(rq); + } else { + switch (p->state) { + case TASK_RUNNING: + case TASK_WAKING: + break; + default: + smp_rmb(); // See smp_rmb() comment in try_to_wake_up(). + if (!p->on_rq) + ret = func(p, arg); + } + } + raw_spin_unlock_irqrestore(&p->pi_lock, rf.flags); + return ret; +} + +/** + * wake_up_process - Wake up a specific process + * @p: The process to be woken up. + * + * Attempt to wake up the nominated process and move it to the set of runnable + * processes. + * + * Return: 1 if the process was woken up, 0 if it was already running. + * + * This function executes a full memory barrier before accessing the task state. + */ +int wake_up_process(struct task_struct *p) +{ + return try_to_wake_up(p, TASK_NORMAL, 0); +} +EXPORT_SYMBOL(wake_up_process); + +int wake_up_state(struct task_struct *p, unsigned int state) +{ + return try_to_wake_up(p, state, 0); +} + +static void time_slice_expired(struct task_struct *p, struct rq *rq); + +/* + * Perform scheduler related setup for a newly forked process p. + * p is forked by current. + */ +int sched_fork(unsigned long __maybe_unused clone_flags, struct task_struct *p) +{ + unsigned long flags; + +#ifdef CONFIG_PREEMPT_NOTIFIERS + INIT_HLIST_HEAD(&p->preempt_notifiers); +#endif + +#ifdef CONFIG_COMPACTION + p->capture_control = NULL; +#endif + +#ifdef CONFIG_SMP + p->wake_entry.u_flags = CSD_TYPE_TTWU; +#endif + /* + * We mark the process as NEW here. This guarantees that + * nobody will actually run it, and a signal or other external + * event cannot wake it up and insert it on the runqueue either. + */ + p->state = TASK_NEW; + + /* + * The process state is set to the same value of the process executing + * do_fork() code. That is running. This guarantees that nobody will + * actually run it, and a signal or other external event cannot wake + * it up and insert it on the runqueue either. + */ + + /* Should be reset in fork.c but done here for ease of MuQSS patching */ + p->on_cpu = + p->on_rq = + p->utime = + p->stime = + p->sched_time = + p->stime_ns = + p->utime_ns = 0; + skiplist_node_init(&p->node); + + /* + * Revert to default priority/policy on fork if requested. + */ + if (unlikely(p->sched_reset_on_fork)) { + if (p->policy == SCHED_FIFO || p->policy == SCHED_RR || p-> policy == SCHED_ISO) { + p->policy = SCHED_NORMAL; + p->normal_prio = normal_prio(p); + } + + if (PRIO_TO_NICE(p->static_prio) < 0) { + p->static_prio = NICE_TO_PRIO(0); + p->normal_prio = p->static_prio; + } + + /* + * We don't need the reset flag anymore after the fork. It has + * fulfilled its duty: + */ + p->sched_reset_on_fork = 0; + } + + /* + * Silence PROVE_RCU. + */ + raw_spin_lock_irqsave(&p->pi_lock, flags); + rseq_migrate(p); + set_task_cpu(p, smp_processor_id()); + raw_spin_unlock_irqrestore(&p->pi_lock, flags); + +#ifdef CONFIG_SCHED_INFO + if (unlikely(sched_info_on())) + memset(&p->sched_info, 0, sizeof(p->sched_info)); +#endif + init_task_preempt_count(p); + + return 0; +} + +void sched_post_fork(struct task_struct *p) +{ +} + +#ifdef CONFIG_SCHEDSTATS + +DEFINE_STATIC_KEY_FALSE(sched_schedstats); +static bool __initdata __sched_schedstats = false; + +static void set_schedstats(bool enabled) +{ + if (enabled) + static_branch_enable(&sched_schedstats); + else + static_branch_disable(&sched_schedstats); +} + +void force_schedstat_enabled(void) +{ + if (!schedstat_enabled()) { + pr_info("kernel profiling enabled schedstats, disable via kernel.sched_schedstats.\n"); + static_branch_enable(&sched_schedstats); + } +} + +static int __init setup_schedstats(char *str) +{ + int ret = 0; + if (!str) + goto out; + + /* + * This code is called before jump labels have been set up, so we can't + * change the static branch directly just yet. Instead set a temporary + * variable so init_schedstats() can do it later. + */ + if (!strcmp(str, "enable")) { + __sched_schedstats = true; + ret = 1; + } else if (!strcmp(str, "disable")) { + __sched_schedstats = false; + ret = 1; + } +out: + if (!ret) + pr_warn("Unable to parse schedstats=\n"); + + return ret; +} +__setup("schedstats=", setup_schedstats); + +static void __init init_schedstats(void) +{ + set_schedstats(__sched_schedstats); +} + +#ifdef CONFIG_PROC_SYSCTL +int sysctl_schedstats(struct ctl_table *table, int write, void *buffer, + size_t *lenp, loff_t *ppos) +{ + struct ctl_table t; + int err; + int state = static_branch_likely(&sched_schedstats); + + if (write && !capable(CAP_SYS_ADMIN)) + return -EPERM; + + t = *table; + t.data = &state; + err = proc_dointvec_minmax(&t, write, buffer, lenp, ppos); + if (err < 0) + return err; + if (write) + set_schedstats(state); + return err; +} +#endif /* CONFIG_PROC_SYSCTL */ +#else /* !CONFIG_SCHEDSTATS */ +static inline void init_schedstats(void) {} +#endif /* CONFIG_SCHEDSTATS */ + +static void update_cpu_clock_switch(struct rq *rq, struct task_struct *p); + +static void account_task_cpu(struct rq *rq, struct task_struct *p) +{ + update_clocks(rq); + /* This isn't really a context switch but accounting is the same */ + update_cpu_clock_switch(rq, p); + p->last_ran = rq->niffies; +} + +bool sched_smp_initialized __read_mostly; + +static inline int hrexpiry_enabled(struct rq *rq) +{ + if (unlikely(!cpu_active(cpu_of(rq)) || !sched_smp_initialized)) + return 0; + return hrtimer_is_hres_active(&rq->hrexpiry_timer); +} + +/* + * Use HR-timers to deliver accurate preemption points. + */ +static inline void hrexpiry_clear(struct rq *rq) +{ + if (!hrexpiry_enabled(rq)) + return; + if (hrtimer_active(&rq->hrexpiry_timer)) + hrtimer_cancel(&rq->hrexpiry_timer); +} + +/* + * High-resolution time_slice expiry. + * Runs from hardirq context with interrupts disabled. + */ +static enum hrtimer_restart hrexpiry(struct hrtimer *timer) +{ + struct rq *rq = container_of(timer, struct rq, hrexpiry_timer); + struct task_struct *p; + + /* This can happen during CPU hotplug / resume */ + if (unlikely(cpu_of(rq) != smp_processor_id())) + goto out; + + /* + * We're doing this without the runqueue lock but this should always + * be run on the local CPU. Time slice should run out in __schedule + * but we set it to zero here in case niffies is slightly less. + */ + p = rq->curr; + p->time_slice = 0; + __set_tsk_resched(p); +out: + return HRTIMER_NORESTART; +} + +/* + * Called to set the hrexpiry timer state. + * + * called with irqs disabled from the local CPU only + */ +static void hrexpiry_start(struct rq *rq, u64 delay) +{ + if (!hrexpiry_enabled(rq)) + return; + + hrtimer_start(&rq->hrexpiry_timer, ns_to_ktime(delay), + HRTIMER_MODE_REL_PINNED); +} + +static void init_rq_hrexpiry(struct rq *rq) +{ + hrtimer_init(&rq->hrexpiry_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + rq->hrexpiry_timer.function = hrexpiry; +} + +static inline int rq_dither(struct rq *rq) +{ + if (!hrexpiry_enabled(rq)) + return HALF_JIFFY_US; + return 0; +} + +/* + * wake_up_new_task - wake up a newly created task for the first time. + * + * This function will do some initial scheduler statistics housekeeping + * that must be done for every newly created context, then puts the task + * on the runqueue and wakes it. + */ +void wake_up_new_task(struct task_struct *p) +{ + struct task_struct *parent, *rq_curr; + struct rq *rq, *new_rq; + unsigned long flags; + + parent = p->parent; + + raw_spin_lock_irqsave(&p->pi_lock, flags); + p->state = TASK_RUNNING; + /* Task_rq can't change yet on a new task */ + new_rq = rq = task_rq(p); + if (unlikely(needs_other_cpu(p, task_cpu(p)))) { + set_task_cpu(p, valid_task_cpu(p)); + new_rq = task_rq(p); + } + + double_rq_lock(rq, new_rq); + rq_curr = rq->curr; + + /* + * Make sure we do not leak PI boosting priority to the child. + */ + p->prio = rq_curr->normal_prio; + + trace_sched_wakeup_new(p); + + /* + * Share the timeslice between parent and child, thus the + * total amount of pending timeslices in the system doesn't change, + * resulting in more scheduling fairness. If it's negative, it won't + * matter since that's the same as being 0. rq->rq_deadline is only + * modified within schedule() so it is always equal to + * current->deadline. + */ + account_task_cpu(rq, rq_curr); + p->last_ran = rq_curr->last_ran; + if (likely(rq_curr->policy != SCHED_FIFO)) { + rq_curr->time_slice /= 2; + if (rq_curr->time_slice < RESCHED_US) { + /* + * Forking task has run out of timeslice. Reschedule it and + * start its child with a new time slice and deadline. The + * child will end up running first because its deadline will + * be slightly earlier. + */ + __set_tsk_resched(rq_curr); + time_slice_expired(p, new_rq); + if (suitable_idle_cpus(p)) + resched_best_idle(p, task_cpu(p)); + else if (unlikely(rq != new_rq)) + try_preempt(p, new_rq); + } else { + p->time_slice = rq_curr->time_slice; + if (rq_curr == parent && rq == new_rq && !suitable_idle_cpus(p)) { + /* + * The VM isn't cloned, so we're in a good position to + * do child-runs-first in anticipation of an exec. This + * usually avoids a lot of COW overhead. + */ + __set_tsk_resched(rq_curr); + } else { + /* + * Adjust the hrexpiry since rq_curr will keep + * running and its timeslice has been shortened. + */ + hrexpiry_start(rq, US_TO_NS(rq_curr->time_slice)); + try_preempt(p, new_rq); + } + } + } else { + time_slice_expired(p, new_rq); + try_preempt(p, new_rq); + } + activate_task(new_rq, p, 0); + double_rq_unlock(rq, new_rq); + raw_spin_unlock_irqrestore(&p->pi_lock, flags); +} + +#ifdef CONFIG_PREEMPT_NOTIFIERS + +static DEFINE_STATIC_KEY_FALSE(preempt_notifier_key); + +void preempt_notifier_inc(void) +{ + static_branch_inc(&preempt_notifier_key); +} +EXPORT_SYMBOL_GPL(preempt_notifier_inc); + +void preempt_notifier_dec(void) +{ + static_branch_dec(&preempt_notifier_key); +} +EXPORT_SYMBOL_GPL(preempt_notifier_dec); + +/** + * preempt_notifier_register - tell me when current is being preempted & rescheduled + * @notifier: notifier struct to register + */ +void preempt_notifier_register(struct preempt_notifier *notifier) +{ + if (!static_branch_unlikely(&preempt_notifier_key)) + WARN(1, "registering preempt_notifier while notifiers disabled\n"); + + hlist_add_head(¬ifier->link, ¤t->preempt_notifiers); +} +EXPORT_SYMBOL_GPL(preempt_notifier_register); + +/** + * preempt_notifier_unregister - no longer interested in preemption notifications + * @notifier: notifier struct to unregister + * + * This is *not* safe to call from within a preemption notifier. + */ +void preempt_notifier_unregister(struct preempt_notifier *notifier) +{ + hlist_del(¬ifier->link); +} +EXPORT_SYMBOL_GPL(preempt_notifier_unregister); + +static void __fire_sched_in_preempt_notifiers(struct task_struct *curr) +{ + struct preempt_notifier *notifier; + + hlist_for_each_entry(notifier, &curr->preempt_notifiers, link) + notifier->ops->sched_in(notifier, raw_smp_processor_id()); +} + +static __always_inline void fire_sched_in_preempt_notifiers(struct task_struct *curr) +{ + if (static_branch_unlikely(&preempt_notifier_key)) + __fire_sched_in_preempt_notifiers(curr); +} + +static void +__fire_sched_out_preempt_notifiers(struct task_struct *curr, + struct task_struct *next) +{ + struct preempt_notifier *notifier; + + hlist_for_each_entry(notifier, &curr->preempt_notifiers, link) + notifier->ops->sched_out(notifier, next); +} + +static __always_inline void +fire_sched_out_preempt_notifiers(struct task_struct *curr, + struct task_struct *next) +{ + if (static_branch_unlikely(&preempt_notifier_key)) + __fire_sched_out_preempt_notifiers(curr, next); +} + +#else /* !CONFIG_PREEMPT_NOTIFIERS */ + +static inline void fire_sched_in_preempt_notifiers(struct task_struct *curr) +{ +} + +static inline void +fire_sched_out_preempt_notifiers(struct task_struct *curr, + struct task_struct *next) +{ +} + +#endif /* CONFIG_PREEMPT_NOTIFIERS */ + +static inline void prepare_task(struct task_struct *next) +{ + /* + * Claim the task as running, we do this before switching to it + * such that any running task will have this set. + * + * See the ttwu() WF_ON_CPU case and its ordering comment. + */ + WRITE_ONCE(next->on_cpu, 1); +} + +static inline void finish_task(struct task_struct *prev) +{ +#ifdef CONFIG_SMP + /* + * This must be the very last reference to @prev from this CPU. After + * p->on_cpu is cleared, the task can be moved to a different CPU. We + * must ensure this doesn't happen until the switch is completely + * finished. + * + * In particular, the load of prev->state in finish_task_switch() must + * happen before this. + * + * Pairs with the smp_cond_load_acquire() in try_to_wake_up(). + */ + smp_store_release(&prev->on_cpu, 0); +#endif +} + +static inline void +prepare_lock_switch(struct rq *rq, struct task_struct *next) +{ + /* + * Since the runqueue lock will be released by the next + * task (which is an invalid locking op but in the case + * of the scheduler it's an obvious special-case), so we + * do an early lockdep release here: + */ + spin_release(&rq->lock->dep_map, _THIS_IP_); +#ifdef CONFIG_DEBUG_SPINLOCK + /* this is a valid case when another task releases the spinlock */ + rq->lock->owner = next; +#endif +} + +static inline void finish_lock_switch(struct rq *rq, struct task_struct *prev) +{ + /* + * If we are tracking spinlock dependencies then we have to + * fix up the runqueue lock - which gets 'carried over' from + * prev into current: + */ + spin_acquire(&rq->lock->dep_map, 0, 0, _THIS_IP_); + +#ifdef CONFIG_SMP + /* + * If prev was marked as migrating to another CPU in return_task, drop + * the local runqueue lock but leave interrupts disabled and grab the + * remote lock we're migrating it to before enabling them. + */ + if (unlikely(task_on_rq_migrating(prev))) { + sched_info_dequeued(rq, prev); + /* + * We move the ownership of prev to the new cpu now. ttwu can't + * activate prev to the wrong cpu since it has to grab this + * runqueue in ttwu_remote. + */ +#ifdef CONFIG_THREAD_INFO_IN_TASK + prev->cpu = prev->wake_cpu; +#else + task_thread_info(prev)->cpu = prev->wake_cpu; +#endif + raw_spin_unlock(rq->lock); + + raw_spin_lock(&prev->pi_lock); + rq = __task_rq_lock(prev, NULL); + /* Check that someone else hasn't already queued prev */ + if (likely(!task_queued(prev))) { + enqueue_task(rq, prev, 0); + prev->on_rq = TASK_ON_RQ_QUEUED; + /* Wake up the CPU if it's not already running */ + resched_if_idle(rq); + } + raw_spin_unlock(&prev->pi_lock); + } +#endif + raw_spin_unlock_irq(rq->lock); +} + +#ifndef prepare_arch_switch +# define prepare_arch_switch(next) do { } while (0) +#endif +#ifndef finish_arch_switch +# define finish_arch_switch(prev) do { } while (0) +#endif +#ifndef finish_arch_post_lock_switch +# define finish_arch_post_lock_switch() do { } while (0) +#endif + +static inline void kmap_local_sched_out(void) +{ +#ifdef CONFIG_KMAP_LOCAL + if (unlikely(current->kmap_ctrl.idx)) + __kmap_local_sched_out(); +#endif +} + +static inline void kmap_local_sched_in(void) +{ +#ifdef CONFIG_KMAP_LOCAL + if (unlikely(current->kmap_ctrl.idx)) + __kmap_local_sched_in(); +#endif +} + +/** + * prepare_task_switch - prepare to switch tasks + * @rq: the runqueue preparing to switch + * @next: the task we are going to switch to. + * + * This is called with the rq lock held and interrupts off. It must + * be paired with a subsequent finish_task_switch after the context + * switch. + * + * prepare_task_switch sets up locking and calls architecture specific + * hooks. + */ +static inline void +prepare_task_switch(struct rq *rq, struct task_struct *prev, + struct task_struct *next) +{ + kcov_prepare_switch(prev); + sched_info_switch(rq, prev, next); + perf_event_task_sched_out(prev, next); + rseq_preempt(prev); + fire_sched_out_preempt_notifiers(prev, next); + kmap_local_sched_out(); + prepare_task(next); + prepare_arch_switch(next); +} + +/** + * finish_task_switch - clean up after a task-switch + * @rq: runqueue associated with task-switch + * @prev: the thread we just switched away from. + * + * finish_task_switch must be called after the context switch, paired + * with a prepare_task_switch call before the context switch. + * finish_task_switch will reconcile locking set up by prepare_task_switch, + * and do any other architecture-specific cleanup actions. + * + * Note that we may have delayed dropping an mm in context_switch(). If + * so, we finish that here outside of the runqueue lock. (Doing it + * with the lock held can cause deadlocks; see schedule() for + * details.) + * + * The context switch have flipped the stack from under us and restored the + * local variables which were saved when this task called schedule() in the + * past. prev == current is still correct but we need to recalculate this_rq + * because prev may have moved to another CPU. + */ +static void finish_task_switch(struct task_struct *prev) + __releases(rq->lock) +{ + struct rq *rq = this_rq(); + struct mm_struct *mm = rq->prev_mm; + long prev_state; + + /* + * The previous task will have left us with a preempt_count of 2 + * because it left us after: + * + * schedule() + * preempt_disable(); // 1 + * __schedule() + * raw_spin_lock_irq(rq->lock) // 2 + * + * Also, see FORK_PREEMPT_COUNT. + */ + if (WARN_ONCE(preempt_count() != 2*PREEMPT_DISABLE_OFFSET, + "corrupted preempt_count: %s/%d/0x%x\n", + current->comm, current->pid, preempt_count())) + preempt_count_set(FORK_PREEMPT_COUNT); + + rq->prev_mm = NULL; + + /* + * A task struct has one reference for the use as "current". + * If a task dies, then it sets TASK_DEAD in tsk->state and calls + * schedule one last time. The schedule call will never return, and + * the scheduled task must drop that reference. + * + * We must observe prev->state before clearing prev->on_cpu (in + * finish_task), otherwise a concurrent wakeup can get prev + * running on another CPU and we could rave with its RUNNING -> DEAD + * transition, resulting in a double drop. + */ + prev_state = prev->state; + vtime_task_switch(prev); + perf_event_task_sched_in(prev, current); + finish_task(prev); + finish_lock_switch(rq, prev); + finish_arch_post_lock_switch(); + kcov_finish_switch(current); + /* + * kmap_local_sched_out() is invoked with rq::lock held and + * interrupts disabled. There is no requirement for that, but the + * sched out code does not have an interrupt enabled section. + * Restoring the maps on sched in does not require interrupts being + * disabled either. + */ + kmap_local_sched_in(); + + fire_sched_in_preempt_notifiers(current); + /* + * When switching through a kernel thread, the loop in + * membarrier_{private,global}_expedited() may have observed that + * kernel thread and not issued an IPI. It is therefore possible to + * schedule between user->kernel->user threads without passing though + * switch_mm(). Membarrier requires a barrier after storing to + * rq->curr, before returning to userspace, so provide them here: + * + * - a full memory barrier for {PRIVATE,GLOBAL}_EXPEDITED, implicitly + * provided by mmdrop(), + * - a sync_core for SYNC_CORE. + */ + if (mm) { + membarrier_mm_sync_core_before_usermode(mm); + mmdrop(mm); + } + if (unlikely(prev_state == TASK_DEAD)) { + /* + * Remove function-return probe instances associated with this + * task and put them back on the free list. + */ + kprobe_flush_task(prev); + + /* Task is done with its stack. */ + put_task_stack(prev); + + put_task_struct_rcu_user(prev); + } +} + +/** + * schedule_tail - first thing a freshly forked thread must call. + * @prev: the thread we just switched away from. + */ +asmlinkage __visible void schedule_tail(struct task_struct *prev) +{ + /* + * New tasks start with FORK_PREEMPT_COUNT, see there and + * finish_task_switch() for details. + * + * finish_task_switch() will drop rq->lock() and lower preempt_count + * and the preempt_enable() will end up enabling preemption (on + * PREEMPT_COUNT kernels). + */ + + finish_task_switch(prev); + preempt_enable(); + + if (current->set_child_tid) + put_user(task_pid_vnr(current), current->set_child_tid); + + calculate_sigpending(); +} + +/* + * context_switch - switch to the new MM and the new thread's register state. + */ +static __always_inline void +context_switch(struct rq *rq, struct task_struct *prev, + struct task_struct *next) +{ + prepare_task_switch(rq, prev, next); + + /* + * For paravirt, this is coupled with an exit in switch_to to + * combine the page table reload and the switch backend into + * one hypercall. + */ + arch_start_context_switch(prev); + + /* + * kernel -> kernel lazy + transfer active + * user -> kernel lazy + mmgrab() active + * + * kernel -> user switch + mmdrop() active + * user -> user switch + */ + if (!next->mm) { // to kernel + enter_lazy_tlb(prev->active_mm, next); + + next->active_mm = prev->active_mm; + if (prev->mm) // from user + mmgrab(prev->active_mm); + else + prev->active_mm = NULL; + } else { // to user + membarrier_switch_mm(rq, prev->active_mm, next->mm); + /* + * sys_membarrier() requires an smp_mb() between setting + * rq->curr / membarrier_switch_mm() and returning to userspace. + * + * The below provides this either through switch_mm(), or in + * case 'prev->active_mm == next->mm' through + * finish_task_switch()'s mmdrop(). + */ + switch_mm_irqs_off(prev->active_mm, next->mm, next); + + if (!prev->mm) { // from kernel + /* will mmdrop() in finish_task_switch(). */ + rq->prev_mm = prev->active_mm; + prev->active_mm = NULL; + } + } + prepare_lock_switch(rq, next); + + /* Here we just switch the register state and the stack. */ + switch_to(prev, next, prev); + barrier(); + + finish_task_switch(prev); +} + +/* + * nr_running, nr_uninterruptible and nr_context_switches: + * + * externally visible scheduler statistics: current number of runnable + * threads, total number of context switches performed since bootup. + */ +unsigned long nr_running(void) +{ + unsigned long i, sum = 0; + + for_each_online_cpu(i) + sum += cpu_rq(i)->nr_running; + + return sum; +} + +static unsigned long nr_uninterruptible(void) +{ + unsigned long i, sum = 0; + + for_each_online_cpu(i) + sum += cpu_rq(i)->nr_uninterruptible; + + return sum; +} + +/* + * Check if only the current task is running on the CPU. + * + * Caution: this function does not check that the caller has disabled + * preemption, thus the result might have a time-of-check-to-time-of-use + * race. The caller is responsible to use it correctly, for example: + * + * - from a non-preemptible section (of course) + * + * - from a thread that is bound to a single CPU + * + * - in a loop with very short iterations (e.g. a polling loop) + */ +bool single_task_running(void) +{ + if (rq_load(raw_rq()) == 1) + return true; + else + return false; +} +EXPORT_SYMBOL(single_task_running); + +unsigned long long nr_context_switches(void) +{ + int cpu; + unsigned long long sum = 0; + + for_each_possible_cpu(cpu) + sum += cpu_rq(cpu)->nr_switches; + + return sum; +} + +/* + * Consumers of these two interfaces, like for example the cpufreq menu + * governor are using nonsensical data. Boosting frequency for a CPU that has + * IO-wait which might not even end up running the task when it does become + * runnable. + */ + +unsigned long nr_iowait_cpu(int cpu) +{ + return atomic_read(&cpu_rq(cpu)->nr_iowait); +} + +/* + * IO-wait accounting, and how it's mostly bollocks (on SMP). + * + * The idea behind IO-wait account is to account the idle time that we could + * have spend running if it were not for IO. That is, if we were to improve the + * storage performance, we'd have a proportional reduction in IO-wait time. + * + * This all works nicely on UP, where, when a task blocks on IO, we account + * idle time as IO-wait, because if the storage were faster, it could've been + * running and we'd not be idle. + * + * This has been extended to SMP, by doing the same for each CPU. This however + * is broken. + * + * Imagine for instance the case where two tasks block on one CPU, only the one + * CPU will have IO-wait accounted, while the other has regular idle. Even + * though, if the storage were faster, both could've ran at the same time, + * utilising both CPUs. + * + * This means, that when looking globally, the current IO-wait accounting on + * SMP is a lower bound, by reason of under accounting. + * + * Worse, since the numbers are provided per CPU, they are sometimes + * interpreted per CPU, and that is nonsensical. A blocked task isn't strictly + * associated with any one particular CPU, it can wake to another CPU than it + * blocked on. This means the per CPU IO-wait number is meaningless. + * + * Task CPU affinities can make all that even more 'interesting'. + */ + +unsigned long nr_iowait(void) +{ + unsigned long cpu, sum = 0; + + for_each_possible_cpu(cpu) + sum += nr_iowait_cpu(cpu); + + return sum; +} + +unsigned long nr_active(void) +{ + return nr_running() + nr_uninterruptible(); +} + +/* Variables and functions for calc_load */ +static unsigned long calc_load_update; +unsigned long avenrun[3]; +EXPORT_SYMBOL(avenrun); + +/** + * get_avenrun - get the load average array + * @loads: pointer to dest load array + * @offset: offset to add + * @shift: shift count to shift the result left + * + * These values are estimates at best, so no need for locking. + */ +void get_avenrun(unsigned long *loads, unsigned long offset, int shift) +{ + loads[0] = (avenrun[0] + offset) << shift; + loads[1] = (avenrun[1] + offset) << shift; + loads[2] = (avenrun[2] + offset) << shift; +} + +/* + * calc_load - update the avenrun load estimates every LOAD_FREQ seconds. + */ +void calc_global_load(void) +{ + long active; + + if (time_before(jiffies, READ_ONCE(calc_load_update))) + return; + active = nr_active() * FIXED_1; + + avenrun[0] = calc_load(avenrun[0], EXP_1, active); + avenrun[1] = calc_load(avenrun[1], EXP_5, active); + avenrun[2] = calc_load(avenrun[2], EXP_15, active); + + calc_load_update = jiffies + LOAD_FREQ; +} + +/** + * fixed_power_int - compute: x^n, in O(log n) time + * + * @x: base of the power + * @frac_bits: fractional bits of @x + * @n: power to raise @x to. + * + * By exploiting the relation between the definition of the natural power + * function: x^n := x*x*...*x (x multiplied by itself for n times), and + * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, + * (where: n_i \elem {0, 1}, the binary vector representing n), + * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is + * of course trivially computable in O(log_2 n), the length of our binary + * vector. + */ +static unsigned long +fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) +{ + unsigned long result = 1UL << frac_bits; + + if (n) { + for (;;) { + if (n & 1) { + result *= x; + result += 1UL << (frac_bits - 1); + result >>= frac_bits; + } + n >>= 1; + if (!n) + break; + x *= x; + x += 1UL << (frac_bits - 1); + x >>= frac_bits; + } + } + + return result; +} + +/* + * a1 = a0 * e + a * (1 - e) + * + * a2 = a1 * e + a * (1 - e) + * = (a0 * e + a * (1 - e)) * e + a * (1 - e) + * = a0 * e^2 + a * (1 - e) * (1 + e) + * + * a3 = a2 * e + a * (1 - e) + * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) + * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) + * + * ... + * + * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] + * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) + * = a0 * e^n + a * (1 - e^n) + * + * [1] application of the geometric series: + * + * n 1 - x^(n+1) + * S_n := \Sum x^i = ------------- + * i=0 1 - x + */ +unsigned long +calc_load_n(unsigned long load, unsigned long exp, + unsigned long active, unsigned int n) +{ + return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); +} + +DEFINE_PER_CPU(struct kernel_stat, kstat); +DEFINE_PER_CPU(struct kernel_cpustat, kernel_cpustat); + +EXPORT_PER_CPU_SYMBOL(kstat); +EXPORT_PER_CPU_SYMBOL(kernel_cpustat); + +#ifdef CONFIG_PARAVIRT +static inline u64 steal_ticks(u64 steal) +{ + if (unlikely(steal > NSEC_PER_SEC)) + return div_u64(steal, TICK_NSEC); + + return __iter_div_u64_rem(steal, TICK_NSEC, &steal); +} +#endif + +#ifndef nsecs_to_cputime +# define nsecs_to_cputime(__nsecs) nsecs_to_jiffies(__nsecs) +#endif + +/* + * On each tick, add the number of nanoseconds to the unbanked variables and + * once one tick's worth has accumulated, account it allowing for accurate + * sub-tick accounting and totals. Use the TICK_APPROX_NS to match the way we + * deduct nanoseconds. + */ +static void pc_idle_time(struct rq *rq, struct task_struct *idle, unsigned long ns) +{ + u64 *cpustat = kcpustat_this_cpu->cpustat; + unsigned long ticks; + + if (atomic_read(&rq->nr_iowait) > 0) { + rq->iowait_ns += ns; + if (rq->iowait_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->iowait_ns); + cpustat[CPUTIME_IOWAIT] += (__force u64)TICK_APPROX_NS * ticks; + rq->iowait_ns %= JIFFY_NS; + } + } else { + rq->idle_ns += ns; + if (rq->idle_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->idle_ns); + cpustat[CPUTIME_IDLE] += (__force u64)TICK_APPROX_NS * ticks; + rq->idle_ns %= JIFFY_NS; + } + } + acct_update_integrals(idle); +} + +static void pc_system_time(struct rq *rq, struct task_struct *p, + int hardirq_offset, unsigned long ns) +{ + u64 *cpustat = kcpustat_this_cpu->cpustat; + unsigned long ticks; + + p->stime_ns += ns; + if (p->stime_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(p->stime_ns); + p->stime_ns %= JIFFY_NS; + p->stime += (__force u64)TICK_APPROX_NS * ticks; + account_group_system_time(p, TICK_APPROX_NS * ticks); + } + p->sched_time += ns; + account_group_exec_runtime(p, ns); + + if (hardirq_count() - hardirq_offset) { + rq->irq_ns += ns; + if (rq->irq_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->irq_ns); + cpustat[CPUTIME_IRQ] += (__force u64)TICK_APPROX_NS * ticks; + rq->irq_ns %= JIFFY_NS; + } + } else if (in_serving_softirq()) { + rq->softirq_ns += ns; + if (rq->softirq_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->softirq_ns); + cpustat[CPUTIME_SOFTIRQ] += (__force u64)TICK_APPROX_NS * ticks; + rq->softirq_ns %= JIFFY_NS; + } + } else { + rq->system_ns += ns; + if (rq->system_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->system_ns); + cpustat[CPUTIME_SYSTEM] += (__force u64)TICK_APPROX_NS * ticks; + rq->system_ns %= JIFFY_NS; + } + } + acct_update_integrals(p); +} + +static void pc_user_time(struct rq *rq, struct task_struct *p, unsigned long ns) +{ + u64 *cpustat = kcpustat_this_cpu->cpustat; + unsigned long ticks; + + p->utime_ns += ns; + if (p->utime_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(p->utime_ns); + p->utime_ns %= JIFFY_NS; + p->utime += (__force u64)TICK_APPROX_NS * ticks; + account_group_user_time(p, TICK_APPROX_NS * ticks); + } + p->sched_time += ns; + account_group_exec_runtime(p, ns); + + if (this_cpu_ksoftirqd() == p) { + /* + * ksoftirqd time do not get accounted in cpu_softirq_time. + * So, we have to handle it separately here. + */ + rq->softirq_ns += ns; + if (rq->softirq_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->softirq_ns); + cpustat[CPUTIME_SOFTIRQ] += (__force u64)TICK_APPROX_NS * ticks; + rq->softirq_ns %= JIFFY_NS; + } + } + + if (task_nice(p) > 0 || idleprio_task(p)) { + rq->nice_ns += ns; + if (rq->nice_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->nice_ns); + cpustat[CPUTIME_NICE] += (__force u64)TICK_APPROX_NS * ticks; + rq->nice_ns %= JIFFY_NS; + } + } else { + rq->user_ns += ns; + if (rq->user_ns >= JIFFY_NS) { + ticks = NS_TO_JIFFIES(rq->user_ns); + cpustat[CPUTIME_USER] += (__force u64)TICK_APPROX_NS * ticks; + rq->user_ns %= JIFFY_NS; + } + } + acct_update_integrals(p); +} + +/* + * This is called on clock ticks. + * Bank in p->sched_time the ns elapsed since the last tick or switch. + * CPU scheduler quota accounting is also performed here in microseconds. + */ +static void update_cpu_clock_tick(struct rq *rq, struct task_struct *p) +{ + s64 account_ns = rq->niffies - p->last_ran; + struct task_struct *idle = rq->idle; + + /* Accurate tick timekeeping */ + if (user_mode(get_irq_regs())) + pc_user_time(rq, p, account_ns); + else if (p != idle || (irq_count() != HARDIRQ_OFFSET)) { + pc_system_time(rq, p, HARDIRQ_OFFSET, account_ns); + } else + pc_idle_time(rq, idle, account_ns); + + /* time_slice accounting is done in usecs to avoid overflow on 32bit */ + if (p->policy != SCHED_FIFO && p != idle) + p->time_slice -= NS_TO_US(account_ns); + + p->last_ran = rq->niffies; +} + +/* + * This is called on context switches. + * Bank in p->sched_time the ns elapsed since the last tick or switch. + * CPU scheduler quota accounting is also performed here in microseconds. + */ +static void update_cpu_clock_switch(struct rq *rq, struct task_struct *p) +{ + s64 account_ns = rq->niffies - p->last_ran; + struct task_struct *idle = rq->idle; + + /* Accurate subtick timekeeping */ + if (p != idle) + pc_user_time(rq, p, account_ns); + else + pc_idle_time(rq, idle, account_ns); + + /* time_slice accounting is done in usecs to avoid overflow on 32bit */ + if (p->policy != SCHED_FIFO && p != idle) + p->time_slice -= NS_TO_US(account_ns); +} + +/* + * Return any ns on the sched_clock that have not yet been accounted in + * @p in case that task is currently running. + * + * Called with task_rq_lock(p) held. + */ +static inline u64 do_task_delta_exec(struct task_struct *p, struct rq *rq) +{ + u64 ns = 0; + + /* + * Must be ->curr _and_ ->on_rq. If dequeued, we would + * project cycles that may never be accounted to this + * thread, breaking clock_gettime(). + */ + if (p == rq->curr && task_on_rq_queued(p)) { + update_clocks(rq); + ns = rq->niffies - p->last_ran; + } + + return ns; +} + +/* + * Return accounted runtime for the task. + * Return separately the current's pending runtime that have not been + * accounted yet. + */ +unsigned long long task_sched_runtime(struct task_struct *p) +{ + struct rq_flags rf; + struct rq *rq; + u64 ns; + +#if defined(CONFIG_64BIT) && defined(CONFIG_SMP) + /* + * 64-bit doesn't need locks to atomically read a 64-bit value. + * So we have a optimisation chance when the task's delta_exec is 0. + * Reading ->on_cpu is racy, but this is ok. + * + * If we race with it leaving CPU, we'll take a lock. So we're correct. + * If we race with it entering CPU, unaccounted time is 0. This is + * indistinguishable from the read occurring a few cycles earlier. + * If we see ->on_cpu without ->on_rq, the task is leaving, and has + * been accounted, so we're correct here as well. + */ + if (!p->on_cpu || !task_on_rq_queued(p)) + return tsk_seruntime(p); +#endif + + rq = task_rq_lock(p, &rf); + ns = p->sched_time + do_task_delta_exec(p, rq); + task_rq_unlock(rq, p, &rf); + + return ns; +} + +/* + * Functions to test for when SCHED_ISO tasks have used their allocated + * quota as real time scheduling and convert them back to SCHED_NORMAL. All + * data is modified only by the local runqueue during scheduler_tick with + * interrupts disabled. + */ + +/* + * Test if SCHED_ISO tasks have run longer than their alloted period as RT + * tasks and set the refractory flag if necessary. There is 10% hysteresis + * for unsetting the flag. 115/128 is ~90/100 as a fast shift instead of a + * slow division. + */ +static inline void iso_tick(struct rq *rq) +{ + rq->iso_ticks = rq->iso_ticks * (ISO_PERIOD - 1) / ISO_PERIOD; + rq->iso_ticks += 100; + if (rq->iso_ticks > ISO_PERIOD * sched_iso_cpu) { + rq->iso_refractory = true; + if (unlikely(rq->iso_ticks > ISO_PERIOD * 100)) + rq->iso_ticks = ISO_PERIOD * 100; + } +} + +/* No SCHED_ISO task was running so decrease rq->iso_ticks */ +static inline void no_iso_tick(struct rq *rq, int ticks) +{ + if (rq->iso_ticks > 0 || rq->iso_refractory) { + rq->iso_ticks = rq->iso_ticks * (ISO_PERIOD - ticks) / ISO_PERIOD; + if (rq->iso_ticks < ISO_PERIOD * (sched_iso_cpu * 115 / 128)) { + rq->iso_refractory = false; + if (unlikely(rq->iso_ticks < 0)) + rq->iso_ticks = 0; + } + } +} + +/* This manages tasks that have run out of timeslice during a scheduler_tick */ +static void task_running_tick(struct rq *rq) +{ + struct task_struct *p = rq->curr; + + /* + * If a SCHED_ISO task is running we increment the iso_ticks. In + * order to prevent SCHED_ISO tasks from causing starvation in the + * presence of true RT tasks we account those as iso_ticks as well. + */ + if (rt_task(p) || task_running_iso(p)) + iso_tick(rq); + else + no_iso_tick(rq, 1); + + /* SCHED_FIFO tasks never run out of timeslice. */ + if (p->policy == SCHED_FIFO) + return; + + if (iso_task(p)) { + if (task_running_iso(p)) { + if (rq->iso_refractory) { + /* + * SCHED_ISO task is running as RT and limit + * has been hit. Force it to reschedule as + * SCHED_NORMAL by zeroing its time_slice + */ + p->time_slice = 0; + } + } else if (!rq->iso_refractory) { + /* Can now run again ISO. Reschedule to pick up prio */ + goto out_resched; + } + } + + /* + * Tasks that were scheduled in the first half of a tick are not + * allowed to run into the 2nd half of the next tick if they will + * run out of time slice in the interim. Otherwise, if they have + * less than RESCHED_US μs of time slice left they will be rescheduled. + * Dither is used as a backup for when hrexpiry is disabled or high res + * timers not configured in. + */ + if (p->time_slice - rq->dither >= RESCHED_US) + return; +out_resched: + rq_lock(rq); + __set_tsk_resched(p); + rq_unlock(rq); +} + +static inline void task_tick(struct rq *rq) +{ + if (!rq_idle(rq)) + task_running_tick(rq); + else if (rq->last_jiffy > rq->last_scheduler_tick) + no_iso_tick(rq, rq->last_jiffy - rq->last_scheduler_tick); +} + +#ifdef CONFIG_NO_HZ_FULL +/* + * We can stop the timer tick any time highres timers are active since + * we rely entirely on highres timeouts for task expiry rescheduling. + */ +static void sched_stop_tick(struct rq *rq, int cpu) +{ + if (!hrexpiry_enabled(rq)) + return; + if (!tick_nohz_full_enabled()) + return; + if (!tick_nohz_full_cpu(cpu)) + return; + tick_nohz_dep_clear_cpu(cpu, TICK_DEP_BIT_SCHED); +} + +static inline void sched_start_tick(struct rq *rq, int cpu) +{ + tick_nohz_dep_set_cpu(cpu, TICK_DEP_BIT_SCHED); +} + +struct tick_work { + int cpu; + atomic_t state; + struct delayed_work work; +}; +/* Values for ->state, see diagram below. */ +#define TICK_SCHED_REMOTE_OFFLINE 0 +#define TICK_SCHED_REMOTE_OFFLINING 1 +#define TICK_SCHED_REMOTE_RUNNING 2 + +/* + * State diagram for ->state: + * + * + * TICK_SCHED_REMOTE_OFFLINE + * | ^ + * | | + * | | sched_tick_remote() + * | | + * | | + * +--TICK_SCHED_REMOTE_OFFLINING + * | ^ + * | | + * sched_tick_start() | | sched_tick_stop() + * | | + * V | + * TICK_SCHED_REMOTE_RUNNING + * + * + * Other transitions get WARN_ON_ONCE(), except that sched_tick_remote() + * and sched_tick_start() are happy to leave the state in RUNNING. + */ + +static struct tick_work __percpu *tick_work_cpu; + +static void sched_tick_remote(struct work_struct *work) +{ + struct delayed_work *dwork = to_delayed_work(work); + struct tick_work *twork = container_of(dwork, struct tick_work, work); + int cpu = twork->cpu; + struct rq *rq = cpu_rq(cpu); + struct task_struct *curr; + u64 delta; + int os; + + /* + * Handle the tick only if it appears the remote CPU is running in full + * dynticks mode. The check is racy by nature, but missing a tick or + * having one too much is no big deal because the scheduler tick updates + * statistics and checks timeslices in a time-independent way, regardless + * of when exactly it is running. + */ + if (!tick_nohz_tick_stopped_cpu(cpu)) + goto out_requeue; + + rq_lock_irq(rq); + if (cpu_is_offline(cpu)) + goto out_unlock; + + curr = rq->curr; + update_rq_clock(rq); + + if (!is_idle_task(curr)) { + /* + * Make sure the next tick runs within a reasonable + * amount of time. + */ + delta = rq_clock_task(rq) - curr->last_ran; + WARN_ON_ONCE(delta > (u64)NSEC_PER_SEC * 3); + } + task_tick(rq); + +out_unlock: + rq_unlock_irq(rq, NULL); + +out_requeue: + + /* + * Run the remote tick once per second (1Hz). This arbitrary + * frequency is large enough to avoid overload but short enough + * to keep scheduler internal stats reasonably up to date. But + * first update state to reflect hotplug activity if required. + */ + os = atomic_fetch_add_unless(&twork->state, -1, TICK_SCHED_REMOTE_RUNNING); + WARN_ON_ONCE(os == TICK_SCHED_REMOTE_OFFLINE); + if (os == TICK_SCHED_REMOTE_RUNNING) + queue_delayed_work(system_unbound_wq, dwork, HZ); +} + +static void sched_tick_start(int cpu) +{ + struct tick_work *twork; + int os; + + if (housekeeping_cpu(cpu, HK_FLAG_TICK)) + return; + + WARN_ON_ONCE(!tick_work_cpu); + + twork = per_cpu_ptr(tick_work_cpu, cpu); + os = atomic_xchg(&twork->state, TICK_SCHED_REMOTE_RUNNING); + WARN_ON_ONCE(os == TICK_SCHED_REMOTE_RUNNING); + if (os == TICK_SCHED_REMOTE_OFFLINE) { + twork->cpu = cpu; + INIT_DELAYED_WORK(&twork->work, sched_tick_remote); + queue_delayed_work(system_unbound_wq, &twork->work, HZ); + } +} + +#ifdef CONFIG_HOTPLUG_CPU +static void sched_tick_stop(int cpu) +{ + struct tick_work *twork; + int os; + + if (housekeeping_cpu(cpu, HK_FLAG_TICK)) + return; + + WARN_ON_ONCE(!tick_work_cpu); + + twork = per_cpu_ptr(tick_work_cpu, cpu); + /* There cannot be competing actions, but don't rely on stop-machine. */ + os = atomic_xchg(&twork->state, TICK_SCHED_REMOTE_OFFLINING); + WARN_ON_ONCE(os != TICK_SCHED_REMOTE_RUNNING); + /* Don't cancel, as this would mess up the state machine. */ +} +#endif /* CONFIG_HOTPLUG_CPU */ + +int __init sched_tick_offload_init(void) +{ + tick_work_cpu = alloc_percpu(struct tick_work); + BUG_ON(!tick_work_cpu); + return 0; +} + +#else /* !CONFIG_NO_HZ_FULL */ +static inline void sched_stop_tick(struct rq *rq, int cpu) {} +static inline void sched_start_tick(struct rq *rq, int cpu) {} +static inline void sched_tick_start(int cpu) { } +static inline void sched_tick_stop(int cpu) { } +#endif + +/* + * This function gets called by the timer code, with HZ frequency. + * We call it with interrupts disabled. + */ +void scheduler_tick(void) +{ + int cpu __maybe_unused = smp_processor_id(); + struct rq *rq = cpu_rq(cpu); + + arch_scale_freq_tick(); + sched_clock_tick(); + update_clocks(rq); + update_load_avg(rq, 0); + update_cpu_clock_tick(rq, rq->curr); + task_tick(rq); + rq->last_scheduler_tick = rq->last_jiffy; + rq->last_tick = rq->clock; + perf_event_task_tick(); + sched_stop_tick(rq, cpu); +} + +#if defined(CONFIG_PREEMPTION) && (defined(CONFIG_DEBUG_PREEMPT) || \ + defined(CONFIG_TRACE_PREEMPT_TOGGLE)) +/* + * If the value passed in is equal to the current preempt count + * then we just disabled preemption. Start timing the latency. + */ +static inline void preempt_latency_start(int val) +{ + if (preempt_count() == val) { + unsigned long ip = get_lock_parent_ip(); +#ifdef CONFIG_DEBUG_PREEMPT + current->preempt_disable_ip = ip; +#endif + trace_preempt_off(CALLER_ADDR0, ip); + } +} + +void preempt_count_add(int val) +{ +#ifdef CONFIG_DEBUG_PREEMPT + /* + * Underflow? + */ + if (DEBUG_LOCKS_WARN_ON((preempt_count() < 0))) + return; +#endif + __preempt_count_add(val); +#ifdef CONFIG_DEBUG_PREEMPT + /* + * Spinlock count overflowing soon? + */ + DEBUG_LOCKS_WARN_ON((preempt_count() & PREEMPT_MASK) >= + PREEMPT_MASK - 10); +#endif + preempt_latency_start(val); +} +EXPORT_SYMBOL(preempt_count_add); +NOKPROBE_SYMBOL(preempt_count_add); + +/* + * If the value passed in equals to the current preempt count + * then we just enabled preemption. Stop timing the latency. + */ +static inline void preempt_latency_stop(int val) +{ + if (preempt_count() == val) + trace_preempt_on(CALLER_ADDR0, get_lock_parent_ip()); +} + +void preempt_count_sub(int val) +{ +#ifdef CONFIG_DEBUG_PREEMPT + /* + * Underflow? + */ + if (DEBUG_LOCKS_WARN_ON(val > preempt_count())) + return; + /* + * Is the spinlock portion underflowing? + */ + if (DEBUG_LOCKS_WARN_ON((val < PREEMPT_MASK) && + !(preempt_count() & PREEMPT_MASK))) + return; +#endif + + preempt_latency_stop(val); + __preempt_count_sub(val); +} +EXPORT_SYMBOL(preempt_count_sub); +NOKPROBE_SYMBOL(preempt_count_sub); + +#else +static inline void preempt_latency_start(int val) { } +static inline void preempt_latency_stop(int val) { } +#endif + +static inline unsigned long get_preempt_disable_ip(struct task_struct *p) +{ +#ifdef CONFIG_DEBUG_PREEMPT + return p->preempt_disable_ip; +#else + return 0; +#endif +} + +/* + * The time_slice is only refilled when it is empty and that is when we set a + * new deadline. Make sure update_clocks has been called recently to update + * rq->niffies. + */ +static void time_slice_expired(struct task_struct *p, struct rq *rq) +{ + p->time_slice = timeslice(); + p->deadline = rq->niffies + task_deadline_diff(p); +#ifdef CONFIG_SMT_NICE + if (!p->mm) + p->smt_bias = 0; + else if (rt_task(p)) + p->smt_bias = 1 << 30; + else if (task_running_iso(p)) + p->smt_bias = 1 << 29; + else if (idleprio_task(p)) { + if (task_running_idle(p)) + p->smt_bias = 0; + else + p->smt_bias = 1; + } else if (--p->smt_bias < 1) + p->smt_bias = MAX_PRIO - p->static_prio; +#endif +} + +/* + * Timeslices below RESCHED_US are considered as good as expired as there's no + * point rescheduling when there's so little time left. SCHED_BATCH tasks + * have been flagged be not latency sensitive and likely to be fully CPU + * bound so every time they're rescheduled they have their time_slice + * refilled, but get a new later deadline to have little effect on + * SCHED_NORMAL tasks. + + */ +static inline void check_deadline(struct task_struct *p, struct rq *rq) +{ + if (p->time_slice < RESCHED_US || batch_task(p)) + time_slice_expired(p, rq); +} + +/* + * Task selection with skiplists is a simple matter of picking off the first + * task in the sorted list, an O(1) operation. The lookup is amortised O(1) + * being bound to the number of processors. + * + * Runqueues are selectively locked based on their unlocked data and then + * unlocked if not needed. At most 3 locks will be held at any time and are + * released as soon as they're no longer needed. All balancing between CPUs + * is thus done here in an extremely simple first come best fit manner. + * + * This iterates over runqueues in cache locality order. In interactive mode + * it iterates over all CPUs and finds the task with the best key/deadline. + * In non-interactive mode it will only take a task if it's from the current + * runqueue or a runqueue with more tasks than the current one with a better + * key/deadline. + */ +#ifdef CONFIG_SMP +static inline struct task_struct +*earliest_deadline_task(struct rq *rq, int cpu, struct task_struct *idle) +{ + struct rq *locked = NULL, *chosen = NULL; + struct task_struct *edt = idle; + int i, best_entries = 0; + u64 best_key = ~0ULL; + + for (i = 0; i < total_runqueues; i++) { + struct rq *other_rq = rq_order(rq, i); + skiplist_node *next; + int entries; + + entries = other_rq->sl->entries; + /* + * Check for queued entres lockless first. The local runqueue + * is locked so entries will always be accurate. + */ + if (!sched_interactive) { + /* + * Don't reschedule balance across nodes unless the CPU + * is idle. + */ + if (edt != idle && rq->cpu_locality[other_rq->cpu] > LOCALITY_SMP) + break; + if (entries <= best_entries) + continue; + } else if (!entries) + continue; + + /* if (i) implies other_rq != rq */ + if (i) { + /* Check for best id queued lockless first */ + if (other_rq->best_key >= best_key) + continue; + + if (unlikely(!trylock_rq(rq, other_rq))) + continue; + + /* Need to reevaluate entries after locking */ + entries = other_rq->sl->entries; + if (unlikely(!entries)) { + unlock_rq(other_rq); + continue; + } + } + + next = other_rq->node; + /* + * In interactive mode we check beyond the best entry on other + * runqueues if we can't get the best for smt or affinity + * reasons. + */ + while ((next = next->next[0]) != other_rq->node) { + struct task_struct *p; + u64 key = next->key; + + /* Reevaluate key after locking */ + if (key >= best_key) + break; + + p = next->value; + if (!smt_schedule(p, rq)) { + if (i && !sched_interactive) + break; + continue; + } + + if (sched_other_cpu(p, cpu)) { + if (sched_interactive || !i) + continue; + break; + } + /* Make sure affinity is ok */ + if (i) { + /* From this point on p is the best so far */ + if (locked) + unlock_rq(locked); + chosen = locked = other_rq; + } + best_entries = entries; + best_key = key; + edt = p; + break; + } + /* rq->preempting is a hint only as the state may have changed + * since it was set with the resched call but if we have met + * the condition we can break out here. */ + if (edt == rq->preempting) + break; + if (i && other_rq != chosen) + unlock_rq(other_rq); + } + + if (likely(edt != idle)) + take_task(rq, cpu, edt); + + if (locked) + unlock_rq(locked); + + rq->preempting = NULL; + + return edt; +} +#else /* CONFIG_SMP */ +static inline struct task_struct +*earliest_deadline_task(struct rq *rq, int cpu, struct task_struct *idle) +{ + struct task_struct *edt; + + if (unlikely(!rq->sl->entries)) + return idle; + edt = rq->node->next[0]->value; + take_task(rq, cpu, edt); + return edt; +} +#endif /* CONFIG_SMP */ + +/* + * Print scheduling while atomic bug: + */ +static noinline void __schedule_bug(struct task_struct *prev) +{ + /* Save this before calling printk(), since that will clobber it */ + unsigned long preempt_disable_ip = get_preempt_disable_ip(current); + + if (oops_in_progress) + return; + + printk(KERN_ERR "BUG: scheduling while atomic: %s/%d/0x%08x\n", + prev->comm, prev->pid, preempt_count()); + + debug_show_held_locks(prev); + print_modules(); + if (irqs_disabled()) + print_irqtrace_events(prev); + if (IS_ENABLED(CONFIG_DEBUG_PREEMPT) + && in_atomic_preempt_off()) { + pr_err("Preemption disabled at:"); + print_ip_sym(KERN_ERR, preempt_disable_ip); + } + dump_stack(); + add_taint(TAINT_WARN, LOCKDEP_STILL_OK); +} + +/* + * Various schedule()-time debugging checks and statistics: + */ +static inline void schedule_debug(struct task_struct *prev, bool preempt) +{ +#ifdef CONFIG_SCHED_STACK_END_CHECK + if (task_stack_end_corrupted(prev)) + panic("corrupted stack end detected inside scheduler\n"); + + if (task_scs_end_corrupted(prev)) + panic("corrupted shadow stack detected inside scheduler\n"); +#endif + +#ifdef CONFIG_DEBUG_ATOMIC_SLEEP + if (!preempt && prev->state && prev->non_block_count) { + printk(KERN_ERR "BUG: scheduling in a non-blocking section: %s/%d/%i\n", + prev->comm, prev->pid, prev->non_block_count); + dump_stack(); + add_taint(TAINT_WARN, LOCKDEP_STILL_OK); + } +#endif + + if (unlikely(in_atomic_preempt_off())) { + __schedule_bug(prev); + preempt_count_set(PREEMPT_DISABLED); + } + rcu_sleep_check(); + SCHED_WARN_ON(ct_state() == CONTEXT_USER); + + profile_hit(SCHED_PROFILING, __builtin_return_address(0)); + + schedstat_inc(this_rq()->sched_count); +} + +/* + * The currently running task's information is all stored in rq local data + * which is only modified by the local CPU. + */ +static inline void set_rq_task(struct rq *rq, struct task_struct *p) +{ + if (p == rq->idle || p->policy == SCHED_FIFO) + hrexpiry_clear(rq); + else + hrexpiry_start(rq, US_TO_NS(p->time_slice)); + if (rq->clock - rq->last_tick > HALF_JIFFY_NS) + rq->dither = 0; + else + rq->dither = rq_dither(rq); + + rq->rq_deadline = p->deadline; + rq->rq_prio = p->prio; +#ifdef CONFIG_SMT_NICE + rq->rq_mm = p->mm; + rq->rq_smt_bias = p->smt_bias; +#endif +} + +#ifdef CONFIG_SMT_NICE +static void check_no_siblings(struct rq __maybe_unused *this_rq) {} +static void wake_no_siblings(struct rq __maybe_unused *this_rq) {} +static void (*check_siblings)(struct rq *this_rq) = &check_no_siblings; +static void (*wake_siblings)(struct rq *this_rq) = &wake_no_siblings; + +/* Iterate over smt siblings when we've scheduled a process on cpu and decide + * whether they should continue running or be descheduled. */ +static void check_smt_siblings(struct rq *this_rq) +{ + int other_cpu; + + for_each_cpu(other_cpu, &this_rq->thread_mask) { + struct task_struct *p; + struct rq *rq; + + rq = cpu_rq(other_cpu); + if (rq_idle(rq)) + continue; + p = rq->curr; + if (!smt_schedule(p, this_rq)) + resched_curr(rq); + } +} + +static void wake_smt_siblings(struct rq *this_rq) +{ + int other_cpu; + + for_each_cpu(other_cpu, &this_rq->thread_mask) { + struct rq *rq; + + rq = cpu_rq(other_cpu); + if (rq_idle(rq)) + resched_idle(rq); + } +} +#else +static void check_siblings(struct rq __maybe_unused *this_rq) {} +static void wake_siblings(struct rq __maybe_unused *this_rq) {} +#endif + +/* + * schedule() is the main scheduler function. + * + * The main means of driving the scheduler and thus entering this function are: + * + * 1. Explicit blocking: mutex, semaphore, waitqueue, etc. + * + * 2. TIF_NEED_RESCHED flag is checked on interrupt and userspace return + * paths. For example, see arch/x86/entry_64.S. + * + * To drive preemption between tasks, the scheduler sets the flag in timer + * interrupt handler scheduler_tick(). + * + * 3. Wakeups don't really cause entry into schedule(). They add a + * task to the run-queue and that's it. + * + * Now, if the new task added to the run-queue preempts the current + * task, then the wakeup sets TIF_NEED_RESCHED and schedule() gets + * called on the nearest possible occasion: + * + * - If the kernel is preemptible (CONFIG_PREEMPTION=y): + * + * - in syscall or exception context, at the next outmost + * preempt_enable(). (this might be as soon as the wake_up()'s + * spin_unlock()!) + * + * - in IRQ context, return from interrupt-handler to + * preemptible context + * + * - If the kernel is not preemptible (CONFIG_PREEMPTION is not set) + * then at the next: + * + * - cond_resched() call + * - explicit schedule() call + * - return from syscall or exception to user-space + * - return from interrupt-handler to user-space + * + * WARNING: must be called with preemption disabled! + */ +static void __sched notrace __schedule(bool preempt) +{ + struct task_struct *prev, *next, *idle; + unsigned long *switch_count; + unsigned long prev_state; + bool deactivate = false; + struct rq *rq; + u64 niffies; + int cpu; + + cpu = smp_processor_id(); + rq = cpu_rq(cpu); + prev = rq->curr; + idle = rq->idle; + + schedule_debug(prev, preempt); + + local_irq_disable(); + rcu_note_context_switch(preempt); + + /* + * Make sure that signal_pending_state()->signal_pending() below + * can't be reordered with __set_current_state(TASK_INTERRUPTIBLE) + * done by the caller to avoid the race with signal_wake_up(): + * + * __set_current_state(@state) signal_wake_up() + * schedule() set_tsk_thread_flag(p, TIF_SIGPENDING) + * wake_up_state(p, state) + * LOCK rq->lock LOCK p->pi_state + * smp_mb__after_spinlock() smp_mb__after_spinlock() + * if (signal_pending_state()) if (p->state & @state) + * + * Also, the membarrier system call requires a full memory barrier + * after coming from user-space, before storing to rq->curr. + */ + rq_lock(rq); + smp_mb__after_spinlock(); +#ifdef CONFIG_SMP + if (rq->preempt) { + /* + * Make sure resched_curr hasn't triggered a preemption + * locklessly on a task that has since scheduled away. Spurious + * wakeup of idle is okay though. + */ + if (unlikely(preempt && prev != idle && !test_tsk_need_resched(prev))) { + rq->preempt = NULL; + clear_preempt_need_resched(); + rq_unlock_irq(rq, NULL); + return; + } + rq->preempt = NULL; + } +#endif + + switch_count = &prev->nivcsw; + + /* + * We must load prev->state once (task_struct::state is volatile), such + * that: + * + * - we form a control dependency vs deactivate_task() below. + * - ptrace_{,un}freeze_traced() can change ->state underneath us. + */ + prev_state = prev->state; + if (!preempt && prev_state) { + if (signal_pending_state(prev_state, prev)) { + prev->state = TASK_RUNNING; + } else { + prev->sched_contributes_to_load = + (prev_state & TASK_UNINTERRUPTIBLE) && + !(prev_state & TASK_NOLOAD) && + !(prev->flags & PF_FROPCK); + + if (prev->sched_contributes_to_load) + rq->nr_uninterruptible++; + + /* + * __schedule() ttwu() + * prev_state = prev->state; if (p->on_rq && ...) + * if (prev_state) goto out; + * p->on_rq = 0; smp_acquire__after_ctrl_dep(); + * p->state = TASK_WAKING + * + * Where __schedule() and ttwu() have matching control dependencies. + * + * After this, schedule() must not care about p->state any more. + */ + deactivate = true; + + if (prev->in_iowait) { + atomic_inc(&rq->nr_iowait); + delayacct_blkio_start(); + } + } + switch_count = &prev->nvcsw; + } + + /* + * Store the niffy value here for use by the next task's last_ran + * below to avoid losing niffies due to update_clocks being called + * again after this point. + */ + update_clocks(rq); + niffies = rq->niffies; + update_cpu_clock_switch(rq, prev); + + clear_tsk_need_resched(prev); + clear_preempt_need_resched(); + + if (idle != prev) { + check_deadline(prev, rq); + return_task(prev, rq, cpu, deactivate); + } + + next = earliest_deadline_task(rq, cpu, idle); + if (likely(next->prio != PRIO_LIMIT)) + clear_cpuidle_map(cpu); + else { + set_cpuidle_map(cpu); + update_load_avg(rq, 0); + } + + set_rq_task(rq, next); + next->last_ran = niffies; + + if (likely(prev != next)) { + /* + * Don't reschedule an idle task or deactivated tasks + */ + if (prev == idle) { + inc_nr_running(rq); + if (rt_task(next)) + rq->rt_nr_running++; + } else if (!deactivate) + resched_suitable_idle(prev); + if (unlikely(next == idle)) { + dec_nr_running(rq); + if (rt_task(prev)) + rq->rt_nr_running--; + wake_siblings(rq); + } else + check_siblings(rq); + rq->nr_switches++; + /* + * RCU users of rcu_dereference(rq->curr) may not see + * changes to task_struct made by pick_next_task(). + */ + RCU_INIT_POINTER(rq->curr, next); + /* + * The membarrier system call requires each architecture + * to have a full memory barrier after updating + * rq->curr, before returning to user-space. + * + * Here are the schemes providing that barrier on the + * various architectures: + * - mm ? switch_mm() : mmdrop() for x86, s390, sparc, PowerPC. + * switch_mm() rely on membarrier_arch_switch_mm() on PowerPC. + * - finish_lock_switch() for weakly-ordered + * architectures where spin_unlock is a full barrier, + * - switch_to() for arm64 (weakly-ordered, spin_unlock + * is a RELEASE barrier), + */ + ++*switch_count; + + psi_sched_switch(prev, next, !task_on_rq_queued(prev)); + + trace_sched_switch(preempt, prev, next); + context_switch(rq, prev, next); /* unlocks the rq */ + } else { + check_siblings(rq); + rq_unlock(rq); + local_irq_enable(); + } +} + +void __noreturn do_task_dead(void) +{ + /* Causes final put_task_struct in finish_task_switch(). */ + set_special_state(TASK_DEAD); + + /* Tell freezer to ignore us: */ + current->flags |= PF_NOFREEZE; + __schedule(false); + BUG(); + + /* Avoid "noreturn function does return" - but don't continue if BUG() is a NOP: */ + for (;;) + cpu_relax(); +} + +static inline void sched_submit_work(struct task_struct *tsk) +{ + unsigned int task_flags; + + if (!tsk->state) + return; + + task_flags = tsk->flags; + /* + * If a worker went to sleep, notify and ask workqueue whether + * it wants to wake up a task to maintain concurrency. + * As this function is called inside the schedule() context, + * we disable preemption to avoid it calling schedule() again + * in the possible wakeup of a kworker and because wq_worker_sleeping() + * requires it. + */ + if (task_flags & (PF_WQ_WORKER | PF_IO_WORKER)) { + preempt_disable(); + if (task_flags & PF_WQ_WORKER) + wq_worker_sleeping(tsk); + else + io_wq_worker_sleeping(tsk); + preempt_enable_no_resched(); + } + + if (tsk_is_pi_blocked(tsk)) + return; + + /* + * If we are going to sleep and we have plugged IO queued, + * make sure to submit it to avoid deadlocks. + */ + if (blk_needs_flush_plug(tsk)) + blk_schedule_flush_plug(tsk); +} + +static inline void sched_update_worker(struct task_struct *tsk) +{ + if (tsk->flags & (PF_WQ_WORKER | PF_IO_WORKER)) { + if (tsk->flags & PF_WQ_WORKER) + wq_worker_running(tsk); + else + io_wq_worker_running(tsk); + } +} + +asmlinkage __visible void __sched schedule(void) +{ + struct task_struct *tsk = current; + + sched_submit_work(tsk); + do { + preempt_disable(); + __schedule(false); + sched_preempt_enable_no_resched(); + } while (need_resched()); + sched_update_worker(tsk); +} + +EXPORT_SYMBOL(schedule); + +/* + * synchronize_rcu_tasks() makes sure that no task is stuck in preempted + * state (have scheduled out non-voluntarily) by making sure that all + * tasks have either left the run queue or have gone into user space. + * As idle tasks do not do either, they must not ever be preempted + * (schedule out non-voluntarily). + * + * schedule_idle() is similar to schedule_preempt_disable() except that it + * never enables preemption because it does not call sched_submit_work(). + */ +void __sched schedule_idle(void) +{ + /* + * As this skips calling sched_submit_work(), which the idle task does + * regardless because that function is a nop when the task is in a + * TASK_RUNNING state, make sure this isn't used someplace that the + * current task can be in any other state. Note, idle is always in the + * TASK_RUNNING state. + */ + WARN_ON_ONCE(current->state); + do { + __schedule(false); + } while (need_resched()); +} + +#if defined(CONFIG_CONTEXT_TRACKING) && !defined(CONFIG_HAVE_CONTEXT_TRACKING_OFFSTACK) +asmlinkage __visible void __sched schedule_user(void) +{ + /* + * If we come here after a random call to set_need_resched(), + * or we have been woken up remotely but the IPI has not yet arrived, + * we haven't yet exited the RCU idle mode. Do it here manually until + * we find a better solution. + * + * NB: There are buggy callers of this function. Ideally we + * should warn if prev_state != IN_USER, but that will trigger + * too frequently to make sense yet. + */ + enum ctx_state prev_state = exception_enter(); + schedule(); + exception_exit(prev_state); +} +#endif + +/** + * schedule_preempt_disabled - called with preemption disabled + * + * Returns with preemption disabled. Note: preempt_count must be 1 + */ +void __sched schedule_preempt_disabled(void) +{ + sched_preempt_enable_no_resched(); + schedule(); + preempt_disable(); +} + +static void __sched notrace preempt_schedule_common(void) +{ + do { + /* + * Because the function tracer can trace preempt_count_sub() + * and it also uses preempt_enable/disable_notrace(), if + * NEED_RESCHED is set, the preempt_enable_notrace() called + * by the function tracer will call this function again and + * cause infinite recursion. + * + * Preemption must be disabled here before the function + * tracer can trace. Break up preempt_disable() into two + * calls. One to disable preemption without fear of being + * traced. The other to still record the preemption latency, + * which can also be traced by the function tracer. + */ + preempt_disable_notrace(); + preempt_latency_start(1); + __schedule(true); + preempt_latency_stop(1); + preempt_enable_no_resched_notrace(); + + /* + * Check again in case we missed a preemption opportunity + * between schedule and now. + */ + } while (need_resched()); +} + +#ifdef CONFIG_PREEMPTION +/* + * This is the entry point to schedule() from in-kernel preemption + * off of preempt_enable. + */ +asmlinkage __visible void __sched notrace preempt_schedule(void) +{ + /* + * If there is a non-zero preempt_count or interrupts are disabled, + * we do not want to preempt the current task. Just return.. + */ + if (likely(!preemptible())) + return; + + preempt_schedule_common(); +} +NOKPROBE_SYMBOL(preempt_schedule); +EXPORT_SYMBOL(preempt_schedule); + +#ifdef CONFIG_PREEMPT_DYNAMIC +DEFINE_STATIC_CALL(preempt_schedule, __preempt_schedule_func); +EXPORT_STATIC_CALL_TRAMP(preempt_schedule); +#endif + + +/** + * preempt_schedule_notrace - preempt_schedule called by tracing + * + * The tracing infrastructure uses preempt_enable_notrace to prevent + * recursion and tracing preempt enabling caused by the tracing + * infrastructure itself. But as tracing can happen in areas coming + * from userspace or just about to enter userspace, a preempt enable + * can occur before user_exit() is called. This will cause the scheduler + * to be called when the system is still in usermode. + * + * To prevent this, the preempt_enable_notrace will use this function + * instead of preempt_schedule() to exit user context if needed before + * calling the scheduler. + */ +asmlinkage __visible void __sched notrace preempt_schedule_notrace(void) +{ + enum ctx_state prev_ctx; + + if (likely(!preemptible())) + return; + + do { + /* + * Because the function tracer can trace preempt_count_sub() + * and it also uses preempt_enable/disable_notrace(), if + * NEED_RESCHED is set, the preempt_enable_notrace() called + * by the function tracer will call this function again and + * cause infinite recursion. + * + * Preemption must be disabled here before the function + * tracer can trace. Break up preempt_disable() into two + * calls. One to disable preemption without fear of being + * traced. The other to still record the preemption latency, + * which can also be traced by the function tracer. + */ + preempt_disable_notrace(); + preempt_latency_start(1); + /* + * Needs preempt disabled in case user_exit() is traced + * and the tracer calls preempt_enable_notrace() causing + * an infinite recursion. + */ + prev_ctx = exception_enter(); + __schedule(true); + exception_exit(prev_ctx); + + preempt_latency_stop(1); + preempt_enable_no_resched_notrace(); + } while (need_resched()); +} +EXPORT_SYMBOL_GPL(preempt_schedule_notrace); + +#ifdef CONFIG_PREEMPT_DYNAMIC +DEFINE_STATIC_CALL(preempt_schedule_notrace, __preempt_schedule_notrace_func); +EXPORT_STATIC_CALL_TRAMP(preempt_schedule_notrace); +#endif + +#endif /* CONFIG_PREEMPTION */ + +#ifdef CONFIG_PREEMPT_DYNAMIC + +#include + +/* + * SC:cond_resched + * SC:might_resched + * SC:preempt_schedule + * SC:preempt_schedule_notrace + * SC:irqentry_exit_cond_resched + * + * + * NONE: + * cond_resched <- __cond_resched + * might_resched <- RET0 + * preempt_schedule <- NOP + * preempt_schedule_notrace <- NOP + * irqentry_exit_cond_resched <- NOP + * + * VOLUNTARY: + * cond_resched <- __cond_resched + * might_resched <- __cond_resched + * preempt_schedule <- NOP + * preempt_schedule_notrace <- NOP + * irqentry_exit_cond_resched <- NOP + * + * FULL: + * cond_resched <- RET0 + * might_resched <- RET0 + * preempt_schedule <- preempt_schedule + * preempt_schedule_notrace <- preempt_schedule_notrace + * irqentry_exit_cond_resched <- irqentry_exit_cond_resched + */ + +enum { + preempt_dynamic_none = 0, + preempt_dynamic_voluntary, + preempt_dynamic_full, +}; + +int preempt_dynamic_mode = preempt_dynamic_full; + +int sched_dynamic_mode(const char *str) +{ + if (!strcmp(str, "none")) + return preempt_dynamic_none; + + if (!strcmp(str, "voluntary")) + return preempt_dynamic_voluntary; + + if (!strcmp(str, "full")) + return preempt_dynamic_full; + + return -EINVAL; +} + +void sched_dynamic_update(int mode) +{ + /* + * Avoid {NONE,VOLUNTARY} -> FULL transitions from ever ending up in + * the ZERO state, which is invalid. + */ + static_call_update(cond_resched, __cond_resched); + static_call_update(might_resched, __cond_resched); + static_call_update(preempt_schedule, __preempt_schedule_func); + static_call_update(preempt_schedule_notrace, __preempt_schedule_notrace_func); + static_call_update(irqentry_exit_cond_resched, irqentry_exit_cond_resched); + + switch (mode) { + case preempt_dynamic_none: + static_call_update(cond_resched, __cond_resched); + static_call_update(might_resched, (void *)&__static_call_return0); + static_call_update(preempt_schedule, NULL); + static_call_update(preempt_schedule_notrace, NULL); + static_call_update(irqentry_exit_cond_resched, NULL); + pr_info("Dynamic Preempt: none\n"); + break; + + case preempt_dynamic_voluntary: + static_call_update(cond_resched, __cond_resched); + static_call_update(might_resched, __cond_resched); + static_call_update(preempt_schedule, NULL); + static_call_update(preempt_schedule_notrace, NULL); + static_call_update(irqentry_exit_cond_resched, NULL); + pr_info("Dynamic Preempt: voluntary\n"); + break; + + case preempt_dynamic_full: + static_call_update(cond_resched, (void *)&__static_call_return0); + static_call_update(might_resched, (void *)&__static_call_return0); + static_call_update(preempt_schedule, __preempt_schedule_func); + static_call_update(preempt_schedule_notrace, __preempt_schedule_notrace_func); + static_call_update(irqentry_exit_cond_resched, irqentry_exit_cond_resched); + pr_info("Dynamic Preempt: full\n"); + break; + } + + preempt_dynamic_mode = mode; +} + +static int __init setup_preempt_mode(char *str) +{ + int mode = sched_dynamic_mode(str); + if (mode < 0) { + pr_warn("Dynamic Preempt: unsupported mode: %s\n", str); + return 1; + } + + sched_dynamic_update(mode); + return 0; +} +__setup("preempt=", setup_preempt_mode); + +#endif /* CONFIG_PREEMPT_DYNAMIC */ +/* + * This is the entry point to schedule() from kernel preemption + * off of irq context. + * Note, that this is called and return with irqs disabled. This will + * protect us against recursive calling from irq. + */ +asmlinkage __visible void __sched preempt_schedule_irq(void) +{ + enum ctx_state prev_state; + + /* Catch callers which need to be fixed */ + BUG_ON(preempt_count() || !irqs_disabled()); + + prev_state = exception_enter(); + + do { + preempt_disable(); + local_irq_enable(); + __schedule(true); + local_irq_disable(); + sched_preempt_enable_no_resched(); + } while (need_resched()); + + exception_exit(prev_state); +} + +int default_wake_function(wait_queue_entry_t *curr, unsigned mode, int wake_flags, + void *key) +{ + WARN_ON_ONCE(IS_ENABLED(CONFIG_SCHED_DEBUG) && wake_flags & ~WF_SYNC); + return try_to_wake_up(curr->private, mode, wake_flags); +} +EXPORT_SYMBOL(default_wake_function); + +#ifdef CONFIG_RT_MUTEXES + +static inline int __rt_effective_prio(struct task_struct *pi_task, int prio) +{ + if (pi_task) + prio = min(prio, pi_task->prio); + + return prio; +} + +static inline int rt_effective_prio(struct task_struct *p, int prio) +{ + struct task_struct *pi_task = rt_mutex_get_top_task(p); + + return __rt_effective_prio(pi_task, prio); +} + +/* + * rt_mutex_setprio - set the current priority of a task + * @p: task to boost + * @pi_task: donor task + * + * This function changes the 'effective' priority of a task. It does + * not touch ->normal_prio like __setscheduler(). + * + * Used by the rt_mutex code to implement priority inheritance + * logic. Call site only calls if the priority of the task changed. + */ +void rt_mutex_setprio(struct task_struct *p, struct task_struct *pi_task) +{ + int prio, oldprio; + struct rq *rq; + + /* XXX used to be waiter->prio, not waiter->task->prio */ + prio = __rt_effective_prio(pi_task, p->normal_prio); + + /* + * If nothing changed; bail early. + */ + if (p->pi_top_task == pi_task && prio == p->prio) + return; + + rq = __task_rq_lock(p, NULL); + update_rq_clock(rq); + /* + * Set under pi_lock && rq->lock, such that the value can be used under + * either lock. + * + * Note that there is loads of tricky to make this pointer cache work + * right. rt_mutex_slowunlock()+rt_mutex_postunlock() work together to + * ensure a task is de-boosted (pi_task is set to NULL) before the + * task is allowed to run again (and can exit). This ensures the pointer + * points to a blocked task -- which guarantees the task is present. + */ + p->pi_top_task = pi_task; + + /* + * For FIFO/RR we only need to set prio, if that matches we're done. + */ + if (prio == p->prio) + goto out_unlock; + + /* + * Idle task boosting is a nono in general. There is one + * exception, when PREEMPT_RT and NOHZ is active: + * + * The idle task calls get_next_timer_interrupt() and holds + * the timer wheel base->lock on the CPU and another CPU wants + * to access the timer (probably to cancel it). We can safely + * ignore the boosting request, as the idle CPU runs this code + * with interrupts disabled and will complete the lock + * protected section without being interrupted. So there is no + * real need to boost. + */ + if (unlikely(p == rq->idle)) { + WARN_ON(p != rq->curr); + WARN_ON(p->pi_blocked_on); + goto out_unlock; + } + + trace_sched_pi_setprio(p, pi_task); + oldprio = p->prio; + p->prio = prio; + if (task_running(rq, p)){ + if (prio > oldprio) + resched_task(p); + } else if (task_queued(p)) { + dequeue_task(rq, p, DEQUEUE_SAVE); + enqueue_task(rq, p, ENQUEUE_RESTORE); + if (prio < oldprio) + try_preempt(p, rq); + } +out_unlock: + /* Avoid rq from going away on us: */ + preempt_disable(); + __task_rq_unlock(rq, NULL); + + preempt_enable(); +} +#else +static inline int rt_effective_prio(struct task_struct *p, int prio) +{ + return prio; +} +#endif + +/* + * Adjust the deadline for when the priority is to change, before it's + * changed. + */ +static inline void adjust_deadline(struct task_struct *p, int new_prio) +{ + p->deadline += static_deadline_diff(new_prio) - task_deadline_diff(p); +} + +void set_user_nice(struct task_struct *p, long nice) +{ + int new_static, old_static; + struct rq_flags rf; + struct rq *rq; + + if (task_nice(p) == nice || nice < MIN_NICE || nice > MAX_NICE) + return; + new_static = NICE_TO_PRIO(nice); + /* + * We have to be careful, if called from sys_setpriority(), + * the task might be in the middle of scheduling on another CPU. + */ + rq = task_rq_lock(p, &rf); + update_rq_clock(rq); + + /* + * The RT priorities are set via sched_setscheduler(), but we still + * allow the 'normal' nice value to be set - but as expected + * it won't have any effect on scheduling until the task is + * not SCHED_NORMAL/SCHED_BATCH: + */ + if (has_rt_policy(p)) { + p->static_prio = new_static; + goto out_unlock; + } + + adjust_deadline(p, new_static); + old_static = p->static_prio; + p->static_prio = new_static; + p->prio = effective_prio(p); + + if (task_queued(p)) { + dequeue_task(rq, p, DEQUEUE_SAVE); + enqueue_task(rq, p, ENQUEUE_RESTORE); + if (new_static < old_static) + try_preempt(p, rq); + } else if (task_running(rq, p)) { + set_rq_task(rq, p); + if (old_static < new_static) + resched_task(p); + } +out_unlock: + task_rq_unlock(rq, p, &rf); +} +EXPORT_SYMBOL(set_user_nice); + +/* + * can_nice - check if a task can reduce its nice value + * @p: task + * @nice: nice value + */ +int can_nice(const struct task_struct *p, const int nice) +{ + /* Convert nice value [19,-20] to rlimit style value [1,40] */ + int nice_rlim = nice_to_rlimit(nice); + + return (nice_rlim <= task_rlimit(p, RLIMIT_NICE) || + capable(CAP_SYS_NICE)); +} + +#ifdef __ARCH_WANT_SYS_NICE + +/* + * sys_nice - change the priority of the current process. + * @increment: priority increment + * + * sys_setpriority is a more generic, but much slower function that + * does similar things. + */ +SYSCALL_DEFINE1(nice, int, increment) +{ + long nice, retval; + + /* + * Setpriority might change our priority at the same moment. + * We don't have to worry. Conceptually one call occurs first + * and we have a single winner. + */ + + increment = clamp(increment, -NICE_WIDTH, NICE_WIDTH); + nice = task_nice(current) + increment; + + nice = clamp_val(nice, MIN_NICE, MAX_NICE); + if (increment < 0 && !can_nice(current, nice)) + return -EPERM; + + retval = security_task_setnice(current, nice); + if (retval) + return retval; + + set_user_nice(current, nice); + return 0; +} + +#endif + +/** + * task_prio - return the priority value of a given task. + * @p: the task in question. + * + * Return: The priority value as seen by users in /proc. + * + * sched policy return value kernel prio user prio/nice + * + * normal, batch, [1 ... 41] 101 0/[-20 ... 19] + * idle [42 ... 81] 102 0/[-20 ... 19] + * iso [0 ... 41] 100 0/[-20 ... 19] + * fifo, rr [-2 ... -100] [98 ... 0] [1 ... 99] + */ +int task_prio(const struct task_struct *p) +{ + int delta, prio = p->prio - MAX_RT_PRIO; + + /* rt tasks and iso tasks */ + if (prio <= 0) + goto out; + + /* Convert to ms to avoid overflows */ + delta = NS_TO_MS(p->deadline - task_rq(p)->niffies); + if (unlikely(delta < 0)) + delta = 0; + delta = delta * 40 / ms_longest_deadline_diff(); + if (delta <= 80) + prio += delta; + if (idleprio_task(p)) + prio += 40; +out: + return prio; +} + +#ifdef CONFIG_SMP +static inline bool rt_rq_is_runnable(struct rq *rt_rq) +{ + return rt_rq->rt_nr_running; +} + +/* + * This function computes an effective utilization for the given CPU, to be + * used for frequency selection given the linear relation: f = u * f_max. + * + * The scheduler tracks the following metrics: + * + * cpu_util_{cfs,rt,dl,irq}() + * cpu_bw_dl() + * + * Where the cfs,rt and dl util numbers are tracked with the same metric and + * synchronized windows and are thus directly comparable. + * + * The cfs,rt,dl utilization are the running times measured with rq->clock_task + * which excludes things like IRQ and steal-time. These latter are then accrued + * in the irq utilization. + * + * The DL bandwidth number otoh is not a measured metric but a value computed + * based on the task model parameters and gives the minimal utilization + * required to meet deadlines. + */ +unsigned long effective_cpu_util(int cpu, unsigned long util_cfs, + unsigned long max, enum cpu_util_type type, + struct task_struct *p) +{ + unsigned long dl_util, util, irq; + struct rq *rq = cpu_rq(cpu); + + if (!uclamp_is_used() && + type == FREQUENCY_UTIL && rt_rq_is_runnable(rq)) { + return max; + } + + /* + * Early check to see if IRQ/steal time saturates the CPU, can be + * because of inaccuracies in how we track these -- see + * update_irq_load_avg(). + */ + irq = cpu_util_irq(rq); + if (unlikely(irq >= max)) + return max; + + /* + * Because the time spend on RT/DL tasks is visible as 'lost' time to + * CFS tasks and we use the same metric to track the effective + * utilization (PELT windows are synchronized) we can directly add them + * to obtain the CPU's actual utilization. + * + * CFS and RT utilization can be boosted or capped, depending on + * utilization clamp constraints requested by currently RUNNABLE + * tasks. + * When there are no CFS RUNNABLE tasks, clamps are released and + * frequency will be gracefully reduced with the utilization decay. + */ + util = util_cfs + cpu_util_rt(rq); + if (type == FREQUENCY_UTIL) + util = uclamp_rq_util_with(rq, util, p); + + dl_util = cpu_util_dl(rq); + + /* + * For frequency selection we do not make cpu_util_dl() a permanent part + * of this sum because we want to use cpu_bw_dl() later on, but we need + * to check if the CFS+RT+DL sum is saturated (ie. no idle time) such + * that we select f_max when there is no idle time. + * + * NOTE: numerical errors or stop class might cause us to not quite hit + * saturation when we should -- something for later. + */ + if (util + dl_util >= max) + return max; + + /* + * OTOH, for energy computation we need the estimated running time, so + * include util_dl and ignore dl_bw. + */ + if (type == ENERGY_UTIL) + util += dl_util; + + /* + * There is still idle time; further improve the number by using the + * irq metric. Because IRQ/steal time is hidden from the task clock we + * need to scale the task numbers: + * + * max - irq + * U' = irq + --------- * U + * max + */ + util = scale_irq_capacity(util, irq, max); + util += irq; + + /* + * Bandwidth required by DEADLINE must always be granted while, for + * FAIR and RT, we use blocked utilization of IDLE CPUs as a mechanism + * to gracefully reduce the frequency when no tasks show up for longer + * periods of time. + * + * Ideally we would like to set bw_dl as min/guaranteed freq and util + + * bw_dl as requested freq. However, cpufreq is not yet ready for such + * an interface. So, we only do the latter for now. + */ + if (type == FREQUENCY_UTIL) + util += cpu_bw_dl(rq); + + return min(max, util); +} + +unsigned long sched_cpu_util(int cpu, unsigned long max) +{ + return effective_cpu_util(cpu, cpu_util_cfs(cpu_rq(cpu)), max, + ENERGY_UTIL, NULL); +} +#endif /* CONFIG_SMP */ + +/** + * idle_cpu - is a given CPU idle currently? + * @cpu: the processor in question. + * + * Return: 1 if the CPU is currently idle. 0 otherwise. + */ +int idle_cpu(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + + if (rq->curr != rq->idle) + return 0; + + if (rq->nr_running) + return 0; + +#ifdef CONFIG_SMP + if (rq->ttwu_pending) + return 0; +#endif + + return 1; +} + +/** + * available_idle_cpu - is a given CPU idle for enqueuing work. + * @cpu: the CPU in question. + * + * Return: 1 if the CPU is currently idle. 0 otherwise. + */ +int available_idle_cpu(int cpu) +{ + if (!idle_cpu(cpu)) + return 0; + + if (vcpu_is_preempted(cpu)) + return 0; + + return 1; +} + +/** + * idle_task - return the idle task for a given CPU. + * @cpu: the processor in question. + * + * Return: The idle task for the CPU @cpu. + */ +struct task_struct *idle_task(int cpu) +{ + return cpu_rq(cpu)->idle; +} + +/** + * find_process_by_pid - find a process with a matching PID value. + * @pid: the pid in question. + * + * The task of @pid, if found. %NULL otherwise. + */ +static inline struct task_struct *find_process_by_pid(pid_t pid) +{ + return pid ? find_task_by_vpid(pid) : current; +} + +/* Actually do priority change: must hold rq lock. */ +static void __setscheduler(struct task_struct *p, struct rq *rq, int policy, + int prio, const struct sched_attr *attr, + bool keep_boost) +{ + int oldrtprio, oldprio; + + /* + * If params can't change scheduling class changes aren't allowed + * either. + */ + if (attr->sched_flags & SCHED_FLAG_KEEP_PARAMS) + return; + + p->policy = policy; + oldrtprio = p->rt_priority; + p->rt_priority = prio; + p->normal_prio = normal_prio(p); + oldprio = p->prio; + /* + * Keep a potential priority boosting if called from + * sched_setscheduler(). + */ + p->prio = normal_prio(p); + if (keep_boost) + p->prio = rt_effective_prio(p, p->prio); + + if (task_running(rq, p)) { + set_rq_task(rq, p); + resched_task(p); + } else if (task_queued(p)) { + dequeue_task(rq, p, DEQUEUE_SAVE); + enqueue_task(rq, p, ENQUEUE_RESTORE); + if (p->prio < oldprio || p->rt_priority > oldrtprio) + try_preempt(p, rq); + } +} + +/* + * Check the target process has a UID that matches the current process's + */ +static bool check_same_owner(struct task_struct *p) +{ + const struct cred *cred = current_cred(), *pcred; + bool match; + + rcu_read_lock(); + pcred = __task_cred(p); + match = (uid_eq(cred->euid, pcred->euid) || + uid_eq(cred->euid, pcred->uid)); + rcu_read_unlock(); + return match; +} + +static int __sched_setscheduler(struct task_struct *p, + const struct sched_attr *attr, + bool user, bool pi) +{ + int retval, policy = attr->sched_policy, oldpolicy = -1, priority = attr->sched_priority; + unsigned long rlim_rtprio = 0; + struct rq_flags rf; + int reset_on_fork; + struct rq *rq; + + /* The pi code expects interrupts enabled */ + BUG_ON(pi && in_interrupt()); + + if (is_rt_policy(policy) && !capable(CAP_SYS_NICE)) { + unsigned long lflags; + + if (!lock_task_sighand(p, &lflags)) + return -ESRCH; + rlim_rtprio = task_rlimit(p, RLIMIT_RTPRIO); + unlock_task_sighand(p, &lflags); + if (rlim_rtprio) + goto recheck; + /* + * If the caller requested an RT policy without having the + * necessary rights, we downgrade the policy to SCHED_ISO. + * We also set the parameter to zero to pass the checks. + */ + policy = SCHED_ISO; + priority = 0; + } +recheck: + /* Double check policy once rq lock held */ + if (policy < 0) { + reset_on_fork = p->sched_reset_on_fork; + policy = oldpolicy = p->policy; + } else { + reset_on_fork = !!(policy & SCHED_RESET_ON_FORK); + policy &= ~SCHED_RESET_ON_FORK; + + if (!SCHED_RANGE(policy)) + return -EINVAL; + } + + if (attr->sched_flags & ~(SCHED_FLAG_ALL | SCHED_FLAG_SUGOV)) + return -EINVAL; + + /* + * Valid priorities for SCHED_FIFO and SCHED_RR are + * 1..MAX_RT_PRIO-1, valid priority for SCHED_NORMAL and + * SCHED_BATCH is 0. + */ + if (priority > MAX_RT_PRIO-1) + return -EINVAL; + if (is_rt_policy(policy) != (priority != 0)) + return -EINVAL; + + /* + * Allow unprivileged RT tasks to decrease priority: + */ + if (user && !capable(CAP_SYS_NICE)) { + if (is_rt_policy(policy)) { + unsigned long rlim_rtprio = + task_rlimit(p, RLIMIT_RTPRIO); + + /* Can't set/change the rt policy */ + if (policy != p->policy && !rlim_rtprio) + return -EPERM; + + /* Can't increase priority */ + if (priority > p->rt_priority && + priority > rlim_rtprio) + return -EPERM; + } else { + switch (p->policy) { + /* + * Can only downgrade policies but not back to + * SCHED_NORMAL + */ + case SCHED_ISO: + if (policy == SCHED_ISO) + goto out; + if (policy != SCHED_NORMAL) + return -EPERM; + break; + case SCHED_BATCH: + if (policy == SCHED_BATCH) + goto out; + if (policy != SCHED_IDLEPRIO) + return -EPERM; + break; + case SCHED_IDLEPRIO: + if (policy == SCHED_IDLEPRIO) + goto out; + return -EPERM; + default: + break; + } + } + + /* Can't change other user's priorities */ + if (!check_same_owner(p)) + return -EPERM; + + /* Normal users shall not reset the sched_reset_on_fork flag: */ + if (p->sched_reset_on_fork && !reset_on_fork) + return -EPERM; + } + + if (user) { + retval = security_task_setscheduler(p); + if (retval) + return retval; + } + + if (pi) + cpuset_read_lock(); + + /* + * Make sure no PI-waiters arrive (or leave) while we are + * changing the priority of the task: + * + * To be able to change p->policy safely, the runqueue lock must be + * held. + */ + rq = task_rq_lock(p, &rf); + update_rq_clock(rq); + + /* + * Changing the policy of the stop threads its a very bad idea: + */ + if (p == rq->stop) { + retval = -EINVAL; + goto unlock; + } + + /* + * If not changing anything there's no need to proceed further, + * but store a possible modification of reset_on_fork. + */ + if (unlikely(policy == p->policy && (!is_rt_policy(policy) || + priority == p->rt_priority))) { + p->sched_reset_on_fork = reset_on_fork; + retval = 0; + goto unlock; + } + + /* Re-check policy now with rq lock held */ + if (unlikely(oldpolicy != -1 && oldpolicy != p->policy)) { + policy = oldpolicy = -1; + task_rq_unlock(rq, p, &rf); + if (pi) + cpuset_read_unlock(); + goto recheck; + } + p->sched_reset_on_fork = reset_on_fork; + + __setscheduler(p, rq, policy, priority, attr, pi); + + /* Avoid rq from going away on us: */ + preempt_disable(); + task_rq_unlock(rq, p, &rf); + + if (pi) { + cpuset_read_unlock(); + rt_mutex_adjust_pi(p); + } + preempt_enable(); +out: + return 0; + +unlock: + task_rq_unlock(rq, p, &rf); + if (pi) + cpuset_read_unlock(); + return retval; +} + +static int _sched_setscheduler(struct task_struct *p, int policy, + const struct sched_param *param, bool check) +{ + struct sched_attr attr = { + .sched_policy = policy, + .sched_priority = param->sched_priority, + .sched_nice = PRIO_TO_NICE(p->static_prio), + }; + + return __sched_setscheduler(p, &attr, check, true); +} +/** + * sched_setscheduler - change the scheduling policy and/or RT priority of a thread. + * @p: the task in question. + * @policy: new policy. + * @param: structure containing the new RT priority. + * + * Use sched_set_fifo(), read its comment. + * + * Return: 0 on success. An error code otherwise. + * + * NOTE that the task may be already dead. + */ +int sched_setscheduler(struct task_struct *p, int policy, + const struct sched_param *param) +{ + return _sched_setscheduler(p, policy, param, true); +} + + +int sched_setattr(struct task_struct *p, const struct sched_attr *attr) +{ + return __sched_setscheduler(p, attr, true, true); +} + +int sched_setattr_nocheck(struct task_struct *p, const struct sched_attr *attr) +{ + return __sched_setscheduler(p, attr, false, true); +} + +/** + * sched_setscheduler_nocheck - change the scheduling policy and/or RT priority of a thread from kernelspace. + * @p: the task in question. + * @policy: new policy. + * @param: structure containing the new RT priority. + * + * Just like sched_setscheduler, only don't bother checking if the + * current context has permission. For example, this is needed in + * stop_machine(): we create temporary high priority worker threads, + * but our caller might not have that capability. + * + * Return: 0 on success. An error code otherwise. + */ +int sched_setscheduler_nocheck(struct task_struct *p, int policy, + const struct sched_param *param) +{ + return _sched_setscheduler(p, policy, param, false); +} + +/* + * SCHED_FIFO is a broken scheduler model; that is, it is fundamentally + * incapable of resource management, which is the one thing an OS really should + * be doing. + * + * This is of course the reason it is limited to privileged users only. + * + * Worse still; it is fundamentally impossible to compose static priority + * workloads. You cannot take two correctly working static prio workloads + * and smash them together and still expect them to work. + * + * For this reason 'all' FIFO tasks the kernel creates are basically at: + * + * MAX_RT_PRIO / 2 + * + * The administrator _MUST_ configure the system, the kernel simply doesn't + * know enough information to make a sensible choice. + */ +void sched_set_fifo(struct task_struct *p) +{ + struct sched_param sp = { .sched_priority = MAX_RT_PRIO / 2 }; + WARN_ON_ONCE(sched_setscheduler_nocheck(p, SCHED_FIFO, &sp) != 0); +} +EXPORT_SYMBOL_GPL(sched_set_fifo); + +/* + * For when you don't much care about FIFO, but want to be above SCHED_NORMAL. + */ +void sched_set_fifo_low(struct task_struct *p) +{ + struct sched_param sp = { .sched_priority = 1 }; + WARN_ON_ONCE(sched_setscheduler_nocheck(p, SCHED_FIFO, &sp) != 0); +} +EXPORT_SYMBOL_GPL(sched_set_fifo_low); + +void sched_set_normal(struct task_struct *p, int nice) +{ + struct sched_attr attr = { + .sched_policy = SCHED_NORMAL, + .sched_nice = nice, + }; + WARN_ON_ONCE(sched_setattr_nocheck(p, &attr) != 0); +} +EXPORT_SYMBOL_GPL(sched_set_normal); + +static int +do_sched_setscheduler(pid_t pid, int policy, struct sched_param __user *param) +{ + struct sched_param lparam; + struct task_struct *p; + int retval; + + if (!param || pid < 0) + return -EINVAL; + if (copy_from_user(&lparam, param, sizeof(struct sched_param))) + return -EFAULT; + + rcu_read_lock(); + retval = -ESRCH; + p = find_process_by_pid(pid); + if (likely(p)) + get_task_struct(p); + rcu_read_unlock(); + + if (likely(p)) { + retval = sched_setscheduler(p, policy, &lparam); + put_task_struct(p); + } + + return retval; +} + +/* + * Mimics kernel/events/core.c perf_copy_attr(). + */ +static int sched_copy_attr(struct sched_attr __user *uattr, + struct sched_attr *attr) +{ + u32 size; + int ret; + + /* Zero the full structure, so that a short copy will be nice: */ + memset(attr, 0, sizeof(*attr)); + + ret = get_user(size, &uattr->size); + if (ret) + return ret; + + /* ABI compatibility quirk: */ + if (!size) + size = SCHED_ATTR_SIZE_VER0; + + if (size < SCHED_ATTR_SIZE_VER0 || size > PAGE_SIZE) + goto err_size; + + ret = copy_struct_from_user(attr, sizeof(*attr), uattr, size); + if (ret) { + if (ret == -E2BIG) + goto err_size; + return ret; + } + + /* + * XXX: Do we want to be lenient like existing syscalls; or do we want + * to be strict and return an error on out-of-bounds values? + */ + attr->sched_nice = clamp(attr->sched_nice, -20, 19); + + /* sched/core.c uses zero here but we already know ret is zero */ + return 0; + +err_size: + put_user(sizeof(*attr), &uattr->size); + return -E2BIG; +} + +/* + * sched_setparam() passes in -1 for its policy, to let the functions + * it calls know not to change it. + */ +#define SETPARAM_POLICY -1 + +/** + * sys_sched_setscheduler - set/change the scheduler policy and RT priority + * @pid: the pid in question. + * @policy: new policy. + * @param: structure containing the new RT priority. + * + * Return: 0 on success. An error code otherwise. + */ +SYSCALL_DEFINE3(sched_setscheduler, pid_t, pid, int, policy, struct sched_param __user *, param) +{ + if (policy < 0) + return -EINVAL; + + return do_sched_setscheduler(pid, policy, param); +} + +/** + * sys_sched_setparam - set/change the RT priority of a thread + * @pid: the pid in question. + * @param: structure containing the new RT priority. + * + * Return: 0 on success. An error code otherwise. + */ +SYSCALL_DEFINE2(sched_setparam, pid_t, pid, struct sched_param __user *, param) +{ + return do_sched_setscheduler(pid, SETPARAM_POLICY, param); +} + +/** + * sys_sched_setattr - same as above, but with extended sched_attr + * @pid: the pid in question. + * @uattr: structure containing the extended parameters. + */ +SYSCALL_DEFINE3(sched_setattr, pid_t, pid, struct sched_attr __user *, uattr, + unsigned int, flags) +{ + struct sched_attr attr; + struct task_struct *p; + int retval; + + if (!uattr || pid < 0 || flags) + return -EINVAL; + + retval = sched_copy_attr(uattr, &attr); + if (retval) + return retval; + + if ((int)attr.sched_policy < 0) + return -EINVAL; + if (attr.sched_flags & SCHED_FLAG_KEEP_POLICY) + attr.sched_policy = SETPARAM_POLICY; + + rcu_read_lock(); + retval = -ESRCH; + p = find_process_by_pid(pid); + if (likely(p)) + get_task_struct(p); + rcu_read_unlock(); + + if (likely(p)) { + retval = sched_setattr(p, &attr); + put_task_struct(p); + } + + return retval; +} + +/** + * sys_sched_getscheduler - get the policy (scheduling class) of a thread + * @pid: the pid in question. + * + * Return: On success, the policy of the thread. Otherwise, a negative error + * code. + */ +SYSCALL_DEFINE1(sched_getscheduler, pid_t, pid) +{ + struct task_struct *p; + int retval = -EINVAL; + + if (pid < 0) + goto out_nounlock; + + retval = -ESRCH; + rcu_read_lock(); + p = find_process_by_pid(pid); + if (p) { + retval = security_task_getscheduler(p); + if (!retval) + retval = p->policy; + } + rcu_read_unlock(); + +out_nounlock: + return retval; +} + +/** + * sys_sched_getscheduler - get the RT priority of a thread + * @pid: the pid in question. + * @param: structure containing the RT priority. + * + * Return: On success, 0 and the RT priority is in @param. Otherwise, an error + * code. + */ +SYSCALL_DEFINE2(sched_getparam, pid_t, pid, struct sched_param __user *, param) +{ + struct sched_param lp = { .sched_priority = 0 }; + struct task_struct *p; + int retval = -EINVAL; + + if (!param || pid < 0) + goto out_nounlock; + + rcu_read_lock(); + p = find_process_by_pid(pid); + retval = -ESRCH; + if (!p) + goto out_unlock; + + retval = security_task_getscheduler(p); + if (retval) + goto out_unlock; + + if (has_rt_policy(p)) + lp.sched_priority = p->rt_priority; + rcu_read_unlock(); + + /* + * This one might sleep, we cannot do it with a spinlock held ... + */ + retval = copy_to_user(param, &lp, sizeof(*param)) ? -EFAULT : 0; + +out_nounlock: + return retval; + +out_unlock: + rcu_read_unlock(); + return retval; +} + +/* + * Copy the kernel size attribute structure (which might be larger + * than what user-space knows about) to user-space. + * + * Note that all cases are valid: user-space buffer can be larger or + * smaller than the kernel-space buffer. The usual case is that both + * have the same size. + */ +static int +sched_attr_copy_to_user(struct sched_attr __user *uattr, + struct sched_attr *kattr, + unsigned int usize) +{ + unsigned int ksize = sizeof(*kattr); + + if (!access_ok(uattr, usize)) + return -EFAULT; + + /* + * sched_getattr() ABI forwards and backwards compatibility: + * + * If usize == ksize then we just copy everything to user-space and all is good. + * + * If usize < ksize then we only copy as much as user-space has space for, + * this keeps ABI compatibility as well. We skip the rest. + * + * If usize > ksize then user-space is using a newer version of the ABI, + * which part the kernel doesn't know about. Just ignore it - tooling can + * detect the kernel's knowledge of attributes from the attr->size value + * which is set to ksize in this case. + */ + kattr->size = min(usize, ksize); + + if (copy_to_user(uattr, kattr, kattr->size)) + return -EFAULT; + + return 0; +} + +/** + * sys_sched_getattr - similar to sched_getparam, but with sched_attr + * @pid: the pid in question. + * @uattr: structure containing the extended parameters. + * @usize: sizeof(attr) for fwd/bwd comp. + * @flags: for future extension. + */ +SYSCALL_DEFINE4(sched_getattr, pid_t, pid, struct sched_attr __user *, uattr, + unsigned int, usize, unsigned int, flags) +{ + struct sched_attr kattr = { }; + struct task_struct *p; + int retval; + + if (!uattr || pid < 0 || usize > PAGE_SIZE || + usize < SCHED_ATTR_SIZE_VER0 || flags) + return -EINVAL; + + rcu_read_lock(); + p = find_process_by_pid(pid); + retval = -ESRCH; + if (!p) + goto out_unlock; + + retval = security_task_getscheduler(p); + if (retval) + goto out_unlock; + + kattr.sched_policy = p->policy; + if (rt_task(p)) + kattr.sched_priority = p->rt_priority; + else + kattr.sched_nice = task_nice(p); + + rcu_read_unlock(); + + return sched_attr_copy_to_user(uattr, &kattr, usize); + +out_unlock: + rcu_read_unlock(); + return retval; +} + +long sched_setaffinity(pid_t pid, const struct cpumask *in_mask) +{ + cpumask_var_t cpus_allowed, new_mask; + struct task_struct *p; + int retval; + + rcu_read_lock(); + + p = find_process_by_pid(pid); + if (!p) { + rcu_read_unlock(); + return -ESRCH; + } + + /* Prevent p going away */ + get_task_struct(p); + rcu_read_unlock(); + + if (p->flags & PF_NO_SETAFFINITY) { + retval = -EINVAL; + goto out_put_task; + } + if (!alloc_cpumask_var(&cpus_allowed, GFP_KERNEL)) { + retval = -ENOMEM; + goto out_put_task; + } + if (!alloc_cpumask_var(&new_mask, GFP_KERNEL)) { + retval = -ENOMEM; + goto out_free_cpus_allowed; + } + retval = -EPERM; + if (!check_same_owner(p)) { + rcu_read_lock(); + if (!ns_capable(__task_cred(p)->user_ns, CAP_SYS_NICE)) { + rcu_read_unlock(); + goto out_unlock; + } + rcu_read_unlock(); + } + + retval = security_task_setscheduler(p); + if (retval) + goto out_unlock; + + cpuset_cpus_allowed(p, cpus_allowed); + cpumask_and(new_mask, in_mask, cpus_allowed); +again: + retval = __set_cpus_allowed_ptr(p, new_mask, SCA_CHECK); + + if (!retval) { + cpuset_cpus_allowed(p, cpus_allowed); + if (!cpumask_subset(new_mask, cpus_allowed)) { + /* + * We must have raced with a concurrent cpuset + * update. Just reset the cpus_allowed to the + * cpuset's cpus_allowed + */ + cpumask_copy(new_mask, cpus_allowed); + goto again; + } + } +out_unlock: + free_cpumask_var(new_mask); +out_free_cpus_allowed: + free_cpumask_var(cpus_allowed); +out_put_task: + put_task_struct(p); + return retval; +} + +static int get_user_cpu_mask(unsigned long __user *user_mask_ptr, unsigned len, + cpumask_t *new_mask) +{ + if (len < cpumask_size()) + cpumask_clear(new_mask); + else if (len > cpumask_size()) + len = cpumask_size(); + + return copy_from_user(new_mask, user_mask_ptr, len) ? -EFAULT : 0; +} + + +/** + * sys_sched_setaffinity - set the CPU affinity of a process + * @pid: pid of the process + * @len: length in bytes of the bitmask pointed to by user_mask_ptr + * @user_mask_ptr: user-space pointer to the new CPU mask + * + * Return: 0 on success. An error code otherwise. + */ +SYSCALL_DEFINE3(sched_setaffinity, pid_t, pid, unsigned int, len, + unsigned long __user *, user_mask_ptr) +{ + cpumask_var_t new_mask; + int retval; + + if (!alloc_cpumask_var(&new_mask, GFP_KERNEL)) + return -ENOMEM; + + retval = get_user_cpu_mask(user_mask_ptr, len, new_mask); + if (retval == 0) + retval = sched_setaffinity(pid, new_mask); + free_cpumask_var(new_mask); + return retval; +} + +long sched_getaffinity(pid_t pid, cpumask_t *mask) +{ + struct task_struct *p; + unsigned long flags; + int retval; + + get_online_cpus(); + rcu_read_lock(); + + retval = -ESRCH; + p = find_process_by_pid(pid); + if (!p) + goto out_unlock; + + retval = security_task_getscheduler(p); + if (retval) + goto out_unlock; + + raw_spin_lock_irqsave(&p->pi_lock, flags); + cpumask_and(mask, &p->cpus_mask, cpu_active_mask); + raw_spin_unlock_irqrestore(&p->pi_lock, flags); + +out_unlock: + rcu_read_unlock(); + put_online_cpus(); + + return retval; +} + +/** + * sys_sched_getaffinity - get the CPU affinity of a process + * @pid: pid of the process + * @len: length in bytes of the bitmask pointed to by user_mask_ptr + * @user_mask_ptr: user-space pointer to hold the current CPU mask + * + * Return: 0 on success. An error code otherwise. + */ +SYSCALL_DEFINE3(sched_getaffinity, pid_t, pid, unsigned int, len, + unsigned long __user *, user_mask_ptr) +{ + int ret; + cpumask_var_t mask; + + if ((len * BITS_PER_BYTE) < nr_cpu_ids) + return -EINVAL; + if (len & (sizeof(unsigned long)-1)) + return -EINVAL; + + if (!alloc_cpumask_var(&mask, GFP_KERNEL)) + return -ENOMEM; + + ret = sched_getaffinity(pid, mask); + if (ret == 0) { + unsigned int retlen = min(len, cpumask_size()); + + if (copy_to_user(user_mask_ptr, mask, retlen)) + ret = -EFAULT; + else + ret = retlen; + } + free_cpumask_var(mask); + + return ret; +} + +static void do_sched_yield(void) +{ + struct rq_flags rf; + struct rq *rq; + + if (!sched_yield_type) + return; + + rq = this_rq_lock_irq(&rf); + + if (sched_yield_type > 1) + time_slice_expired(current, rq); + schedstat_inc(rq->yld_count); + + preempt_disable(); + rq_unlock_irq(rq, &rf); + sched_preempt_enable_no_resched(); + + schedule(); +} + +/** + * sys_sched_yield - yield the current processor to other threads. + * + * This function yields the current CPU to other tasks. If there are no + * other threads running on this CPU then this function will return. + * + * Return: 0. + */ +SYSCALL_DEFINE0(sched_yield) +{ + do_sched_yield(); + return 0; +} + +#if !defined(CONFIG_PREEMPTION) || defined(CONFIG_PREEMPT_DYNAMIC) +int __sched __cond_resched(void) +{ + if (should_resched(0)) { + preempt_schedule_common(); + return 1; + } +#ifndef CONFIG_PREEMPT_RCU + rcu_all_qs(); +#endif + return 0; +} +EXPORT_SYMBOL(__cond_resched); +#endif + +#ifdef CONFIG_PREEMPT_DYNAMIC +DEFINE_STATIC_CALL_RET0(cond_resched, __cond_resched); +EXPORT_STATIC_CALL_TRAMP(cond_resched); + +DEFINE_STATIC_CALL_RET0(might_resched, __cond_resched); +EXPORT_STATIC_CALL_TRAMP(might_resched); +#endif + +/* + * __cond_resched_lock() - if a reschedule is pending, drop the given lock, + * call schedule, and on return reacquire the lock. + * + * This works OK both with and without CONFIG_PREEMPTION. We do strange low-level + * operations here to prevent schedule() from being called twice (once via + * spin_unlock(), once by hand). + */ +int __cond_resched_lock(spinlock_t *lock) +{ + int resched = should_resched(PREEMPT_LOCK_OFFSET); + int ret = 0; + + lockdep_assert_held(lock); + + if (spin_needbreak(lock) || resched) { + spin_unlock(lock); + if (resched) + preempt_schedule_common(); + else + cpu_relax(); + ret = 1; + spin_lock(lock); + } + return ret; +} +EXPORT_SYMBOL(__cond_resched_lock); + +int __cond_resched_rwlock_read(rwlock_t *lock) +{ + int resched = should_resched(PREEMPT_LOCK_OFFSET); + int ret = 0; + + lockdep_assert_held_read(lock); + + if (rwlock_needbreak(lock) || resched) { + read_unlock(lock); + if (resched) + preempt_schedule_common(); + else + cpu_relax(); + ret = 1; + read_lock(lock); + } + return ret; +} +EXPORT_SYMBOL(__cond_resched_rwlock_read); + +int __cond_resched_rwlock_write(rwlock_t *lock) +{ + int resched = should_resched(PREEMPT_LOCK_OFFSET); + int ret = 0; + + lockdep_assert_held_write(lock); + + if (rwlock_needbreak(lock) || resched) { + write_unlock(lock); + if (resched) + preempt_schedule_common(); + else + cpu_relax(); + ret = 1; + write_lock(lock); + } + return ret; +} +EXPORT_SYMBOL(__cond_resched_rwlock_write); + +/** + * yield - yield the current processor to other threads. + * + * Do not ever use this function, there's a 99% chance you're doing it wrong. + * + * The scheduler is at all times free to pick the calling task as the most + * eligible task to run, if removing the yield() call from your code breaks + * it, it's already broken. + * + * Typical broken usage is: + * + * while (!event) + * yield(); + * + * where one assumes that yield() will let 'the other' process run that will + * make event true. If the current task is a SCHED_FIFO task that will never + * happen. Never use yield() as a progress guarantee!! + * + * If you want to use yield() to wait for something, use wait_event(). + * If you want to use yield() to be 'nice' for others, use cond_resched(). + * If you still want to use yield(), do not! + */ +void __sched yield(void) +{ + set_current_state(TASK_RUNNING); + do_sched_yield(); +} +EXPORT_SYMBOL(yield); + +/** + * yield_to - yield the current processor to another thread in + * your thread group, or accelerate that thread toward the + * processor it's on. + * @p: target task + * @preempt: whether task preemption is allowed or not + * + * It's the caller's job to ensure that the target task struct + * can't go away on us before we can do any checks. + * + * Return: + * true (>0) if we indeed boosted the target task. + * false (0) if we failed to boost the target. + * -ESRCH if there's no task to yield to. + */ +int __sched yield_to(struct task_struct *p, bool preempt) +{ + struct task_struct *rq_p; + struct rq *rq, *p_rq; + unsigned long flags; + int yielded = 0; + + local_irq_save(flags); + rq = this_rq(); + +again: + p_rq = task_rq(p); + /* + * If we're the only runnable task on the rq and target rq also + * has only one task, there's absolutely no point in yielding. + */ + if (task_running(p_rq, p) || p->state) { + yielded = -ESRCH; + goto out_irq; + } + + double_rq_lock(rq, p_rq); + if (unlikely(task_rq(p) != p_rq)) { + double_rq_unlock(rq, p_rq); + goto again; + } + + yielded = 1; + schedstat_inc(rq->yld_count); + rq_p = rq->curr; + if (p->deadline > rq_p->deadline) + p->deadline = rq_p->deadline; + p->time_slice += rq_p->time_slice; + if (p->time_slice > timeslice()) + p->time_slice = timeslice(); + time_slice_expired(rq_p, rq); + if (preempt && rq != p_rq) + resched_task(p_rq->curr); + double_rq_unlock(rq, p_rq); +out_irq: + local_irq_restore(flags); + + if (yielded > 0) + schedule(); + return yielded; +} +EXPORT_SYMBOL_GPL(yield_to); + +int io_schedule_prepare(void) +{ + int old_iowait = current->in_iowait; + + current->in_iowait = 1; + blk_schedule_flush_plug(current); + + return old_iowait; +} + +void io_schedule_finish(int token) +{ + current->in_iowait = token; +} + +/* + * This task is about to go to sleep on IO. Increment rq->nr_iowait so + * that process accounting knows that this is a task in IO wait state. + * + * But don't do that if it is a deliberate, throttling IO wait (this task + * has set its backing_dev_info: the queue against which it should throttle) + */ + +long __sched io_schedule_timeout(long timeout) +{ + int token; + long ret; + + token = io_schedule_prepare(); + ret = schedule_timeout(timeout); + io_schedule_finish(token); + + return ret; +} +EXPORT_SYMBOL(io_schedule_timeout); + +void __sched io_schedule(void) +{ + int token; + + token = io_schedule_prepare(); + schedule(); + io_schedule_finish(token); +} +EXPORT_SYMBOL(io_schedule); + +/** + * sys_sched_get_priority_max - return maximum RT priority. + * @policy: scheduling class. + * + * Return: On success, this syscall returns the maximum + * rt_priority that can be used by a given scheduling class. + * On failure, a negative error code is returned. + */ +SYSCALL_DEFINE1(sched_get_priority_max, int, policy) +{ + int ret = -EINVAL; + + switch (policy) { + case SCHED_FIFO: + case SCHED_RR: + ret = MAX_RT_PRIO-1; + break; + case SCHED_NORMAL: + case SCHED_BATCH: + case SCHED_ISO: + case SCHED_IDLEPRIO: + ret = 0; + break; + } + return ret; +} + +/** + * sys_sched_get_priority_min - return minimum RT priority. + * @policy: scheduling class. + * + * Return: On success, this syscall returns the minimum + * rt_priority that can be used by a given scheduling class. + * On failure, a negative error code is returned. + */ +SYSCALL_DEFINE1(sched_get_priority_min, int, policy) +{ + int ret = -EINVAL; + + switch (policy) { + case SCHED_FIFO: + case SCHED_RR: + ret = 1; + break; + case SCHED_NORMAL: + case SCHED_BATCH: + case SCHED_ISO: + case SCHED_IDLEPRIO: + ret = 0; + break; + } + return ret; +} + +static int sched_rr_get_interval(pid_t pid, struct timespec64 *t) +{ + struct task_struct *p; + unsigned int time_slice; + struct rq_flags rf; + struct rq *rq; + int retval; + + if (pid < 0) + return -EINVAL; + + retval = -ESRCH; + rcu_read_lock(); + p = find_process_by_pid(pid); + if (!p) + goto out_unlock; + + retval = security_task_getscheduler(p); + if (retval) + goto out_unlock; + + rq = task_rq_lock(p, &rf); + time_slice = p->policy == SCHED_FIFO ? 0 : MS_TO_NS(task_timeslice(p)); + task_rq_unlock(rq, p, &rf); + + rcu_read_unlock(); + *t = ns_to_timespec64(time_slice); + return 0; + +out_unlock: + rcu_read_unlock(); + return retval; +} + +/** + * sys_sched_rr_get_interval - return the default timeslice of a process. + * @pid: pid of the process. + * @interval: userspace pointer to the timeslice value. + * + * this syscall writes the default timeslice value of a given process + * into the user-space timespec buffer. A value of '0' means infinity. + * + * Return: On success, 0 and the timeslice is in @interval. Otherwise, + * an error code. + */ +SYSCALL_DEFINE2(sched_rr_get_interval, pid_t, pid, + struct __kernel_timespec __user *, interval) +{ + struct timespec64 t; + int retval = sched_rr_get_interval(pid, &t); + + if (retval == 0) + retval = put_timespec64(&t, interval); + + return retval; +} + +#ifdef CONFIG_COMPAT_32BIT_TIME +SYSCALL_DEFINE2(sched_rr_get_interval_time32, pid_t, pid, + struct old_timespec32 __user *, interval) +{ + struct timespec64 t; + int retval = sched_rr_get_interval(pid, &t); + + if (retval == 0) + retval = put_old_timespec32(&t, interval); + return retval; +} +#endif + +void sched_show_task(struct task_struct *p) +{ + unsigned long free = 0; + int ppid; + + if (!try_get_task_stack(p)) + return; + + printk(KERN_INFO "%-15.15s %c", p->comm, task_state_to_char(p)); + + if (p->state == TASK_RUNNING) + printk(KERN_CONT " running task "); +#ifdef CONFIG_DEBUG_STACK_USAGE + free = stack_not_used(p); +#endif + ppid = 0; + rcu_read_lock(); + if (pid_alive(p)) + ppid = task_pid_nr(rcu_dereference(p->real_parent)); + rcu_read_unlock(); + pr_cont(" stack:%5lu pid:%5d ppid:%6d flags:0x%08lx\n", + free, task_pid_nr(p), ppid, + (unsigned long)task_thread_info(p)->flags); + + print_worker_info(KERN_INFO, p); + print_stop_info(KERN_INFO, p); + show_stack(p, NULL, KERN_INFO); + put_task_stack(p); +} +EXPORT_SYMBOL_GPL(sched_show_task); + +static inline bool +state_filter_match(unsigned long state_filter, struct task_struct *p) +{ + /* no filter, everything matches */ + if (!state_filter) + return true; + + /* filter, but doesn't match */ + if (!(p->state & state_filter)) + return false; + + /* + * When looking for TASK_UNINTERRUPTIBLE skip TASK_IDLE (allows + * TASK_KILLABLE). + */ + if (state_filter == TASK_UNINTERRUPTIBLE && p->state == TASK_IDLE) + return false; + + return true; +} + +void show_state_filter(unsigned long state_filter) +{ + struct task_struct *g, *p; + + rcu_read_lock(); + for_each_process_thread(g, p) { + /* + * reset the NMI-timeout, listing all files on a slow + * console might take a lot of time: + * Also, reset softlockup watchdogs on all CPUs, because + * another CPU might be blocked waiting for us to process + * an IPI. + */ + touch_nmi_watchdog(); + touch_all_softlockup_watchdogs(); + if (state_filter_match(state_filter, p)) + sched_show_task(p); + } + + rcu_read_unlock(); + /* + * Only show locks if all tasks are dumped: + */ + if (!state_filter) + debug_show_all_locks(); +} + +void dump_cpu_task(int cpu) +{ + pr_info("Task dump for CPU %d:\n", cpu); + sched_show_task(cpu_curr(cpu)); +} + +#ifdef CONFIG_SMP +void set_cpus_allowed_common(struct task_struct *p, const struct cpumask *new_mask, u32 __always_unused flags) +{ + cpumask_copy(&p->cpus_mask, new_mask); + p->nr_cpus_allowed = cpumask_weight(new_mask); +} + +void +__do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) +{ + struct rq *rq = task_rq(p); + + lockdep_assert_held(&p->pi_lock); + + cpumask_copy(&p->cpus_mask, new_mask); + + if (task_queued(p)) { + /* + * Because __kthread_bind() calls this on blocked tasks without + * holding rq->lock. + */ + lockdep_assert_held(rq->lock); + } +} + +/* + * Calling do_set_cpus_allowed from outside the scheduler code should not be + * called on a running or queued task. We should be holding pi_lock. + */ +void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask) +{ + __do_set_cpus_allowed(p, new_mask); + if (needs_other_cpu(p, task_cpu(p))) { + struct rq *rq; + + rq = __task_rq_lock(p, NULL); + set_task_cpu(p, valid_task_cpu(p)); + resched_task(p); + __task_rq_unlock(rq, NULL); + } +} + +void migrate_disable(void) +{ +} +EXPORT_SYMBOL_GPL(migrate_disable); + +void migrate_enable(void) +{ +} +EXPORT_SYMBOL_GPL(migrate_enable); +#endif + +/** + * init_idle - set up an idle thread for a given CPU + * @idle: task in question + * @cpu: cpu the idle task belongs to + * + * NOTE: this function does not set the idle thread's NEED_RESCHED + * flag, to make booting more robust. + */ +void __init init_idle(struct task_struct *idle, int cpu) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long flags; + + /* + * The idle task doesn't need the kthread struct to function, but it + * is dressed up as a per-CPU kthread and thus needs to play the part + * if we want to avoid special-casing it in code that deals with per-CPU + * kthreads. + */ + set_kthread_struct(idle); + + raw_spin_lock_irqsave(&idle->pi_lock, flags); + raw_spin_lock(rq->lock); + idle->last_ran = rq->niffies; + time_slice_expired(idle, rq); + idle->state = TASK_RUNNING; + /* Setting prio to illegal value shouldn't matter when never queued */ + idle->prio = PRIO_LIMIT; + /* + * PF_KTHREAD should already be set at this point; regardless, make it + * look like a proper per-CPU kthread. + */ + idle->flags |= PF_IDLE | PF_KTHREAD | PF_NO_SETAFFINITY; + kthread_set_per_cpu(idle, cpu); + + scs_task_reset(idle); + kasan_unpoison_task_stack(idle); + +#ifdef CONFIG_SMP + /* + * It's possible that init_idle() gets called multiple times on a task, + * in that case do_set_cpus_allowed() will not do the right thing. + * + * And since this is boot we can forgo the serialisation. + */ + set_cpus_allowed_common(idle, cpumask_of(cpu), 0); +#ifdef CONFIG_SMT_NICE + idle->smt_bias = 0; +#endif +#endif + set_rq_task(rq, idle); + + /* Silence PROVE_RCU */ + rcu_read_lock(); + set_task_cpu(idle, cpu); + rcu_read_unlock(); + + rq->idle = idle; + rcu_assign_pointer(rq->curr, idle); + idle->on_rq = TASK_ON_RQ_QUEUED; + raw_spin_unlock(rq->lock); + raw_spin_unlock_irqrestore(&idle->pi_lock, flags); + + /* Set the preempt count _outside_ the spinlocks! */ + init_idle_preempt_count(idle, cpu); + + ftrace_graph_init_idle_task(idle, cpu); + vtime_init_idle(idle, cpu); +#ifdef CONFIG_SMP + sprintf(idle->comm, "%s/%d", INIT_TASK_COMM, cpu); +#endif +} + +int cpuset_cpumask_can_shrink(const struct cpumask __maybe_unused *cur, + const struct cpumask __maybe_unused *trial) +{ + return 1; +} + +int task_can_attach(struct task_struct *p, + const struct cpumask *cs_cpus_allowed) +{ + int ret = 0; + + /* + * Kthreads which disallow setaffinity shouldn't be moved + * to a new cpuset; we don't want to change their CPU + * affinity and isolating such threads by their set of + * allowed nodes is unnecessary. Thus, cpusets are not + * applicable for such threads. This prevents checking for + * success of set_cpus_allowed_ptr() on all attached tasks + * before cpus_mask may be changed. + */ + if (p->flags & PF_NO_SETAFFINITY) + ret = -EINVAL; + + return ret; +} + +void resched_cpu(int cpu) +{ + struct rq *rq = cpu_rq(cpu); + struct rq_flags rf; + + rq_lock_irqsave(rq, &rf); + if (cpu_online(cpu) || cpu == smp_processor_id()) + resched_curr(rq); + rq_unlock_irqrestore(rq, &rf); +} + +#ifdef CONFIG_SMP +#ifdef CONFIG_NO_HZ_COMMON +void select_nohz_load_balancer(int stop_tick) +{ +} + +void set_cpu_sd_state_idle(void) {} +void nohz_balance_enter_idle(int cpu) {} + +/* + * In the semi idle case, use the nearest busy CPU for migrating timers + * from an idle CPU. This is good for power-savings. + * + * We don't do similar optimization for completely idle system, as + * selecting an idle CPU will add more delays to the timers than intended + * (as that CPU's timer base may not be uptodate wrt jiffies etc). + */ +int get_nohz_timer_target(void) +{ + int i, cpu = smp_processor_id(), default_cpu = -1; + struct sched_domain *sd; + + if (housekeeping_cpu(cpu, HK_FLAG_TIMER)) { + if (!idle_cpu(cpu)) + return cpu; + default_cpu = cpu; + } + + rcu_read_lock(); + for_each_domain(cpu, sd) { + for_each_cpu_and(i, sched_domain_span(sd), + housekeeping_cpumask(HK_FLAG_TIMER)) { + if (cpu == i) + continue; + + if (!idle_cpu(i)) { + cpu = i; + goto unlock; + } + } + } + + if (default_cpu == -1) + default_cpu = housekeeping_any_cpu(HK_FLAG_TIMER); + cpu = default_cpu; +unlock: + rcu_read_unlock(); + return cpu; +} + +/* + * When add_timer_on() enqueues a timer into the timer wheel of an + * idle CPU then this timer might expire before the next timer event + * which is scheduled to wake up that CPU. In case of a completely + * idle system the next event might even be infinite time into the + * future. wake_up_idle_cpu() ensures that the CPU is woken up and + * leaves the inner idle loop so the newly added timer is taken into + * account when the CPU goes back to idle and evaluates the timer + * wheel for the next timer event. + */ +void wake_up_idle_cpu(int cpu) +{ + if (cpu == smp_processor_id()) + return; + + if (set_nr_and_not_polling(cpu_rq(cpu)->idle)) + smp_sched_reschedule(cpu); + else + trace_sched_wake_idle_without_ipi(cpu); +} + +static bool wake_up_full_nohz_cpu(int cpu) +{ + /* + * We just need the target to call irq_exit() and re-evaluate + * the next tick. The nohz full kick at least implies that. + * If needed we can still optimize that later with an + * empty IRQ. + */ + if (cpu_is_offline(cpu)) + return true; /* Don't try to wake offline CPUs. */ + if (tick_nohz_full_cpu(cpu)) { + if (cpu != smp_processor_id() || + tick_nohz_tick_stopped()) + tick_nohz_full_kick_cpu(cpu); + return true; + } + + return false; +} + +/* + * Wake up the specified CPU. If the CPU is going offline, it is the + * caller's responsibility to deal with the lost wakeup, for example, + * by hooking into the CPU_DEAD notifier like timers and hrtimers do. + */ +void wake_up_nohz_cpu(int cpu) +{ + if (!wake_up_full_nohz_cpu(cpu)) + wake_up_idle_cpu(cpu); +} +#endif /* CONFIG_NO_HZ_COMMON */ + +/* + * Change a given task's CPU affinity. Migrate the thread to a + * proper CPU and schedule it away if the CPU it's executing on + * is removed from the allowed bitmask. + * + * NOTE: the caller must have a valid reference to the task, the + * task must not exit() & deallocate itself prematurely. The + * call is not atomic; no spinlocks may be held. + */ +static int __set_cpus_allowed_ptr(struct task_struct *p, + const struct cpumask *new_mask, + u32 flags) +{ + const struct cpumask *cpu_valid_mask = cpu_active_mask; + bool queued = false, running_wrong = false, kthread; + unsigned int dest_cpu; + struct rq_flags rf; + struct rq *rq; + int ret = 0; + + rq = task_rq_lock(p, &rf); + update_rq_clock(rq); + + kthread = !!(p->flags & PF_KTHREAD); + if (kthread) { + /* + * Kernel threads are allowed on online && !active CPUs + */ + cpu_valid_mask = cpu_online_mask; + } + + /* + * Must re-check here, to close a race against __kthread_bind(), + * sched_setaffinity() is not guaranteed to observe the flag. + */ + if ((flags & SCA_CHECK) && (p->flags & PF_NO_SETAFFINITY)) { + ret = -EINVAL; + goto out; + } + + if (cpumask_equal(&p->cpus_mask, new_mask)) + goto out; + /* + * Picking a ~random cpu helps in cases where we are changing affinity + * for groups of tasks (ie. cpuset), so that load balancing is not + * immediately required to distribute the tasks within their new mask. + */ + dest_cpu = cpumask_any_and_distribute(cpu_valid_mask, new_mask); + if (dest_cpu >= nr_cpu_ids) { + ret = -EINVAL; + goto out; + } + + queued = task_queued(p); + __do_set_cpus_allowed(p, new_mask); + + if (kthread) { + /* + * For kernel threads that do indeed end up on online && + * !active we want to ensure they are strict per-CPU threads. + */ + WARN_ON(cpumask_intersects(new_mask, cpu_online_mask) && + !cpumask_intersects(new_mask, cpu_active_mask) && + p->nr_cpus_allowed != 1); + } + + /* Can the task run on the task's current CPU? If so, we're done */ + if (cpumask_test_cpu(task_cpu(p), new_mask)) + goto out; + + if (task_running(rq, p)) { + /* Task is running on the wrong cpu now, reschedule it. */ + if (rq == this_rq()) { + set_task_cpu(p, dest_cpu); + set_tsk_need_resched(p); + running_wrong = true; + } else + resched_task(p); + } else { + if (queued) { + /* + * Switch runqueue locks after dequeueing the task + * here while still holding the pi_lock to be holding + * the correct lock for enqueueing. + */ + dequeue_task(rq, p, 0); + rq_unlock(rq); + + rq = cpu_rq(dest_cpu); + rq_lock(rq); + } + set_task_cpu(p, dest_cpu); + if (queued) + enqueue_task(rq, p, 0); + } + if (queued) + try_preempt(p, rq); + if (running_wrong) + preempt_disable(); +out: + task_rq_unlock(rq, p, &rf); + + if (running_wrong) { + __schedule(true); + preempt_enable(); + } + + return ret; +} + +int set_cpus_allowed_ptr(struct task_struct *p, const struct cpumask *new_mask) +{ + return __set_cpus_allowed_ptr(p, new_mask, 0); +} +EXPORT_SYMBOL_GPL(set_cpus_allowed_ptr); + +#ifdef CONFIG_HOTPLUG_CPU +/* + * Run through task list and find tasks affined to the dead cpu, then remove + * that cpu from the list, enable cpu0 and set the zerobound flag. Must hold + * cpu 0 and src_cpu's runqueue locks. We should be holding both rq lock and + * pi_lock to change cpus_mask but it's not going to matter here. + */ +static void bind_zero(int src_cpu) +{ + struct task_struct *p, *t; + struct rq *rq0; + int bound = 0; + + if (src_cpu == 0) + return; + + rq0 = cpu_rq(0); + + do_each_thread(t, p) { + if (cpumask_test_cpu(src_cpu, p->cpus_ptr)) { + bool local = (task_cpu(p) == src_cpu); + struct rq *rq = task_rq(p); + + /* task_running is the cpu stopper thread */ + if (local && task_running(rq, p)) + continue; + atomic_clear_cpu(src_cpu, &p->cpus_mask); + atomic_set_cpu(0, &p->cpus_mask); + p->zerobound = true; + bound++; + if (local) { + bool queued = task_queued(p); + + if (queued) + dequeue_task(rq, p, 0); + set_task_cpu(p, 0); + if (queued) + enqueue_task(rq0, p, 0); + } + } + } while_each_thread(t, p); + + if (bound) { + printk(KERN_INFO "MuQSS removed affinity for %d processes to cpu %d\n", + bound, src_cpu); + } +} + +/* Find processes with the zerobound flag and reenable their affinity for the + * CPU coming alive. */ +static void unbind_zero(int src_cpu) +{ + int unbound = 0, zerobound = 0; + struct task_struct *p, *t; + + if (src_cpu == 0) + return; + + do_each_thread(t, p) { + if (!p->mm) + p->zerobound = false; + if (p->zerobound) { + unbound++; + cpumask_set_cpu(src_cpu, &p->cpus_mask); + /* Once every CPU affinity has been re-enabled, remove + * the zerobound flag */ + if (cpumask_subset(cpu_possible_mask, p->cpus_ptr)) { + p->zerobound = false; + zerobound++; + } + } + } while_each_thread(t, p); + + if (unbound) { + printk(KERN_INFO "MuQSS added affinity for %d processes to cpu %d\n", + unbound, src_cpu); + } + if (zerobound) { + printk(KERN_INFO "MuQSS released forced binding to cpu0 for %d processes\n", + zerobound); + } +} + +/* + * Ensure that the idle task is using init_mm right before its cpu goes + * offline. + */ +void idle_task_exit(void) +{ + struct mm_struct *mm = current->active_mm; + + BUG_ON(cpu_online(smp_processor_id())); + BUG_ON(current != this_rq()->idle); + + if (mm != &init_mm) { + switch_mm(mm, &init_mm, current); + finish_arch_post_lock_switch(); + } + + /* finish_cpu(), as ran on the BP, will clean up the active_mm state */ +} +#else /* CONFIG_HOTPLUG_CPU */ +static void unbind_zero(int src_cpu) {} +#endif /* CONFIG_HOTPLUG_CPU */ + +void sched_set_stop_task(int cpu, struct task_struct *stop) +{ + struct sched_param stop_param = { .sched_priority = STOP_PRIO }; + struct sched_param start_param = { .sched_priority = 0 }; + struct task_struct *old_stop = cpu_rq(cpu)->stop; + + if (stop) { + /* + * Make it appear like a SCHED_FIFO task, its something + * userspace knows about and won't get confused about. + * + * Also, it will make PI more or less work without too + * much confusion -- but then, stop work should not + * rely on PI working anyway. + */ + sched_setscheduler_nocheck(stop, SCHED_FIFO, &stop_param); + } + + cpu_rq(cpu)->stop = stop; + + if (old_stop) { + /* + * Reset it back to a normal scheduling policy so that + * it can die in pieces. + */ + sched_setscheduler_nocheck(old_stop, SCHED_NORMAL, &start_param); + } +} + +#if defined(CONFIG_SCHED_DEBUG) && defined(CONFIG_SYSCTL) + +static struct ctl_table sd_ctl_dir[] = { + { + .procname = "sched_domain", + .mode = 0555, + }, + {} +}; + +static struct ctl_table sd_ctl_root[] = { + { + .procname = "kernel", + .mode = 0555, + .child = sd_ctl_dir, + }, + {} +}; + +static struct ctl_table *sd_alloc_ctl_entry(int n) +{ + struct ctl_table *entry = + kcalloc(n, sizeof(struct ctl_table), GFP_KERNEL); + + return entry; +} + +static void sd_free_ctl_entry(struct ctl_table **tablep) +{ + struct ctl_table *entry; + + /* + * In the intermediate directories, both the child directory and + * procname are dynamically allocated and could fail but the mode + * will always be set. In the lowest directory the names are + * static strings and all have proc handlers. + */ + for (entry = *tablep; entry->mode; entry++) { + if (entry->child) + sd_free_ctl_entry(&entry->child); + if (entry->proc_handler == NULL) + kfree(entry->procname); + } + + kfree(*tablep); + *tablep = NULL; +} + +static void +set_table_entry(struct ctl_table *entry, + const char *procname, void *data, int maxlen, + umode_t mode, proc_handler *proc_handler) +{ + entry->procname = procname; + entry->data = data; + entry->maxlen = maxlen; + entry->mode = mode; + entry->proc_handler = proc_handler; +} + +static struct ctl_table * +sd_alloc_ctl_domain_table(struct sched_domain *sd) +{ + struct ctl_table *table = sd_alloc_ctl_entry(9); + + if (table == NULL) + return NULL; + + set_table_entry(&table[0], "min_interval", &sd->min_interval, sizeof(long), 0644, proc_doulongvec_minmax); + set_table_entry(&table[1], "max_interval", &sd->max_interval, sizeof(long), 0644, proc_doulongvec_minmax); + set_table_entry(&table[2], "busy_factor", &sd->busy_factor, sizeof(int), 0644, proc_dointvec_minmax); + set_table_entry(&table[3], "imbalance_pct", &sd->imbalance_pct, sizeof(int), 0644, proc_dointvec_minmax); + set_table_entry(&table[4], "cache_nice_tries", &sd->cache_nice_tries, sizeof(int), 0644, proc_dointvec_minmax); + set_table_entry(&table[5], "flags", &sd->flags, sizeof(int), 0644, proc_dointvec_minmax); + set_table_entry(&table[6], "max_newidle_lb_cost", &sd->max_newidle_lb_cost, sizeof(long), 0644, proc_doulongvec_minmax); + set_table_entry(&table[7], "name", sd->name, CORENAME_MAX_SIZE, 0444, proc_dostring); + /* &table[8] is terminator */ + + return table; +} + +static struct ctl_table *sd_alloc_ctl_cpu_table(int cpu) +{ + struct ctl_table *entry, *table; + struct sched_domain *sd; + int domain_num = 0, i; + char buf[32]; + + for_each_domain(cpu, sd) + domain_num++; + entry = table = sd_alloc_ctl_entry(domain_num + 1); + if (table == NULL) + return NULL; + + i = 0; + for_each_domain(cpu, sd) { + snprintf(buf, 32, "domain%d", i); + entry->procname = kstrdup(buf, GFP_KERNEL); + entry->mode = 0555; + entry->child = sd_alloc_ctl_domain_table(sd); + entry++; + i++; + } + return table; +} + +static cpumask_var_t sd_sysctl_cpus; +static struct ctl_table_header *sd_sysctl_header; + +void register_sched_domain_sysctl(void) +{ + static struct ctl_table *cpu_entries; + static struct ctl_table **cpu_idx; + char buf[32]; + int i; + + if (!cpu_entries) { + cpu_entries = sd_alloc_ctl_entry(num_possible_cpus() + 1); + if (!cpu_entries) + return; + + WARN_ON(sd_ctl_dir[0].child); + sd_ctl_dir[0].child = cpu_entries; + } + + if (!cpu_idx) { + struct ctl_table *e = cpu_entries; + + cpu_idx = kcalloc(nr_cpu_ids, sizeof(struct ctl_table*), GFP_KERNEL); + if (!cpu_idx) + return; + + /* deal with sparse possible map */ + for_each_possible_cpu(i) { + cpu_idx[i] = e; + e++; + } + } + + if (!cpumask_available(sd_sysctl_cpus)) { + if (!alloc_cpumask_var(&sd_sysctl_cpus, GFP_KERNEL)) + return; + + /* init to possible to not have holes in @cpu_entries */ + cpumask_copy(sd_sysctl_cpus, cpu_possible_mask); + } + + for_each_cpu(i, sd_sysctl_cpus) { + struct ctl_table *e = cpu_idx[i]; + + if (e->child) + sd_free_ctl_entry(&e->child); + + if (!e->procname) { + snprintf(buf, 32, "cpu%d", i); + e->procname = kstrdup(buf, GFP_KERNEL); + } + e->mode = 0555; + e->child = sd_alloc_ctl_cpu_table(i); + + __cpumask_clear_cpu(i, sd_sysctl_cpus); + } + + WARN_ON(sd_sysctl_header); + sd_sysctl_header = register_sysctl_table(sd_ctl_root); +} + +void dirty_sched_domain_sysctl(int cpu) +{ + if (cpumask_available(sd_sysctl_cpus)) + __cpumask_set_cpu(cpu, sd_sysctl_cpus); +} + +/* may be called multiple times per register */ +void unregister_sched_domain_sysctl(void) +{ + unregister_sysctl_table(sd_sysctl_header); + sd_sysctl_header = NULL; +} +#endif /* CONFIG_SYSCTL */ + +void set_rq_online(struct rq *rq) +{ + if (!rq->online) { + cpumask_set_cpu(cpu_of(rq), rq->rd->online); + rq->online = true; + } +} + +void set_rq_offline(struct rq *rq) +{ + if (rq->online) { + int cpu = cpu_of(rq); + + cpumask_clear_cpu(cpu, rq->rd->online); + rq->online = false; + clear_cpuidle_map(cpu); + } +} + +/* + * used to mark begin/end of suspend/resume: + */ +static int num_cpus_frozen; + +/* + * Update cpusets according to cpu_active mask. If cpusets are + * disabled, cpuset_update_active_cpus() becomes a simple wrapper + * around partition_sched_domains(). + * + * If we come here as part of a suspend/resume, don't touch cpusets because we + * want to restore it back to its original state upon resume anyway. + */ +static void cpuset_cpu_active(void) +{ + if (cpuhp_tasks_frozen) { + /* + * num_cpus_frozen tracks how many CPUs are involved in suspend + * resume sequence. As long as this is not the last online + * operation in the resume sequence, just build a single sched + * domain, ignoring cpusets. + */ + partition_sched_domains(1, NULL, NULL); + if (--num_cpus_frozen) + return; + /* + * This is the last CPU online operation. So fall through and + * restore the original sched domains by considering the + * cpuset configurations. + */ + cpuset_force_rebuild(); + } + + cpuset_update_active_cpus(); +} + +static int cpuset_cpu_inactive(unsigned int cpu) +{ + if (!cpuhp_tasks_frozen) { + cpuset_update_active_cpus(); + } else { + num_cpus_frozen++; + partition_sched_domains(1, NULL, NULL); + } + return 0; +} + +int sched_cpu_activate(unsigned int cpu) +{ + struct rq *rq = cpu_rq(cpu); + struct rq_flags rf; + +#ifdef CONFIG_SCHED_SMT + /* + * When going up, increment the number of cores with SMT present. + */ + if (cpumask_weight(cpu_smt_mask(cpu)) == 2) + static_branch_inc_cpuslocked(&sched_smt_present); +#endif + set_cpu_active(cpu, true); + + if (sched_smp_initialized) { + sched_domains_numa_masks_set(cpu); + cpuset_cpu_active(); + } + + /* + * Put the rq online, if not already. This happens: + * + * 1) In the early boot process, because we build the real domains + * after all CPUs have been brought up. + * + * 2) At runtime, if cpuset_cpu_active() fails to rebuild the + * domains. + */ + rq_lock_irqsave(rq, &rf); + if (rq->rd) { + BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span)); + set_rq_online(rq); + } + unbind_zero(cpu); + rq_unlock_irqrestore(rq, &rf); + + return 0; +} + +int sched_cpu_deactivate(unsigned int cpu) +{ + struct rq *rq = cpu_rq(cpu); + struct rq_flags rf; + int ret; + + set_cpu_active(cpu, false); + /* + * We've cleared cpu_active_mask, wait for all preempt-disabled and RCU + * users of this state to go away such that all new such users will + * observe it. + * + * Do sync before park smpboot threads to take care the rcu boost case. + */ + synchronize_rcu(); + + rq_lock_irqsave(rq, &rf); + if (rq->rd) { + update_rq_clock(rq); + BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span)); + set_rq_offline(rq); + } + rq_unlock_irqrestore(rq, &rf); + +#ifdef CONFIG_SCHED_SMT + /* + * When going down, decrement the number of cores with SMT present. + */ + if (cpumask_weight(cpu_smt_mask(cpu)) == 2) + static_branch_dec_cpuslocked(&sched_smt_present); +#endif + + if (!sched_smp_initialized) + return 0; + + ret = cpuset_cpu_inactive(cpu); + if (ret) { + set_cpu_active(cpu, true); + return ret; + } + sched_domains_numa_masks_clear(cpu); + return 0; +} + +int sched_cpu_starting(unsigned int cpu) +{ + sched_tick_start(cpu); + return 0; +} + +#ifdef CONFIG_HOTPLUG_CPU +int sched_cpu_wait_empty(unsigned int __always_unused cpu) +{ + return 0; +} + +int sched_cpu_dying(unsigned int cpu) +{ + struct rq *rq = cpu_rq(cpu); + unsigned long flags; + + /* Handle pending wakeups and then migrate everything off */ + sched_tick_stop(cpu); + + local_irq_save(flags); + double_rq_lock(rq, cpu_rq(0)); + if (rq->rd) { + BUG_ON(!cpumask_test_cpu(cpu, rq->rd->span)); + set_rq_offline(rq); + } + bind_zero(cpu); + double_rq_unlock(rq, cpu_rq(0)); + sched_start_tick(rq, cpu); + hrexpiry_clear(rq); + local_irq_restore(flags); + + return 0; +} +#endif + +#if defined(CONFIG_SCHED_SMT) || defined(CONFIG_SCHED_MC) +/* + * Cheaper version of the below functions in case support for SMT and MC is + * compiled in but CPUs have no siblings. + */ +static bool sole_cpu_idle(struct rq *rq) +{ + return rq_idle(rq); +} +#endif +#ifdef CONFIG_SCHED_SMT +static const cpumask_t *thread_cpumask(int cpu) +{ + return topology_sibling_cpumask(cpu); +} +/* All this CPU's SMT siblings are idle */ +static bool siblings_cpu_idle(struct rq *rq) +{ + return cpumask_subset(&rq->thread_mask, &cpu_idle_map); +} +#endif +#ifdef CONFIG_SCHED_MC +static const cpumask_t *core_cpumask(int cpu) +{ + return topology_core_cpumask(cpu); +} +/* All this CPU's shared cache siblings are idle */ +static bool cache_cpu_idle(struct rq *rq) +{ + return cpumask_subset(&rq->core_mask, &cpu_idle_map); +} +/* MC siblings CPU mask which share the same LLC */ +static const cpumask_t *llc_core_cpumask(int cpu) +{ +#ifdef CONFIG_X86 + return per_cpu(cpu_llc_shared_map, cpu); +#else + return topology_core_cpumask(cpu); +#endif +} +#endif + +enum sched_domain_level { + SD_LV_NONE = 0, + SD_LV_SIBLING, + SD_LV_MC, + SD_LV_BOOK, + SD_LV_CPU, + SD_LV_NODE, + SD_LV_ALLNODES, + SD_LV_MAX +}; + +/* + * Set up the relative cache distance of each online cpu from each + * other in a simple array for quick lookup. Locality is determined + * by the closest sched_domain that CPUs are separated by. CPUs with + * shared cache in SMT and MC are treated as local. Separate CPUs + * (within the same package or physically) within the same node are + * treated as not local. CPUs not even in the same domain (different + * nodes) are treated as very distant. + */ +static void __init select_leaders(void) +{ + struct rq *rq, *other_rq, *leader; + struct sched_domain *sd; + int cpu, other_cpu; +#ifdef CONFIG_SCHED_SMT + bool smt_threads = false; +#endif + + for (cpu = 0; cpu < num_online_cpus(); cpu++) { + rq = cpu_rq(cpu); + leader = NULL; + /* First check if this cpu is in the same node */ + for_each_domain(cpu, sd) { + if (sd->level > SD_LV_MC) + continue; + if (rqshare != RQSHARE_ALL) + leader = NULL; + /* Set locality to local node if not already found lower */ + for_each_cpu(other_cpu, sched_domain_span(sd)) { + if (rqshare >= RQSHARE_SMP) { + other_rq = cpu_rq(other_cpu); + + /* Set the smp_leader to the first CPU */ + if (!leader) + leader = rq; + if (!other_rq->smp_leader) + other_rq->smp_leader = leader; + } + if (rq->cpu_locality[other_cpu] > LOCALITY_SMP) + rq->cpu_locality[other_cpu] = LOCALITY_SMP; + } + } + + /* + * Each runqueue has its own function in case it doesn't have + * siblings of its own allowing mixed topologies. + */ +#ifdef CONFIG_SCHED_MC + leader = NULL; + if (cpumask_weight(core_cpumask(cpu)) > 1) { + cpumask_copy(&rq->core_mask, llc_core_cpumask(cpu)); + cpumask_clear_cpu(cpu, &rq->core_mask); + for_each_cpu(other_cpu, core_cpumask(cpu)) { + if (rqshare == RQSHARE_MC || + (rqshare == RQSHARE_MC_LLC && cpumask_test_cpu(other_cpu, llc_core_cpumask(cpu)))) { + other_rq = cpu_rq(other_cpu); + + /* Set the mc_leader to the first CPU */ + if (!leader) + leader = rq; + if (!other_rq->mc_leader) + other_rq->mc_leader = leader; + } + if (rq->cpu_locality[other_cpu] > LOCALITY_MC) { + /* this is to get LLC into play even in case LLC sharing is not used */ + if (cpumask_test_cpu(other_cpu, llc_core_cpumask(cpu))) + rq->cpu_locality[other_cpu] = LOCALITY_MC_LLC; + else + rq->cpu_locality[other_cpu] = LOCALITY_MC; + } + } + rq->cache_idle = cache_cpu_idle; + } +#endif +#ifdef CONFIG_SCHED_SMT + leader = NULL; + if (cpumask_weight(thread_cpumask(cpu)) > 1) { + cpumask_copy(&rq->thread_mask, thread_cpumask(cpu)); + cpumask_clear_cpu(cpu, &rq->thread_mask); + for_each_cpu(other_cpu, thread_cpumask(cpu)) { + if (rqshare == RQSHARE_SMT) { + other_rq = cpu_rq(other_cpu); + + /* Set the smt_leader to the first CPU */ + if (!leader) + leader = rq; + if (!other_rq->smt_leader) + other_rq->smt_leader = leader; + } + if (rq->cpu_locality[other_cpu] > LOCALITY_SMT) + rq->cpu_locality[other_cpu] = LOCALITY_SMT; + } + rq->siblings_idle = siblings_cpu_idle; + smt_threads = true; + } +#endif + } + +#ifdef CONFIG_SMT_NICE + if (smt_threads) { + check_siblings = &check_smt_siblings; + wake_siblings = &wake_smt_siblings; + smt_schedule = &smt_should_schedule; + } +#endif + + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + for_each_online_cpu(other_cpu) { + printk(KERN_DEBUG "MuQSS locality CPU %d to %d: %d\n", cpu, other_cpu, rq->cpu_locality[other_cpu]); + } + } +} + +/* FIXME freeing locked spinlock */ +static void __init share_and_free_rq(struct rq *leader, struct rq *rq) +{ + WARN_ON(rq->nr_running > 0); + + kfree(rq->node); + kfree(rq->sl); + kfree(rq->lock); + rq->node = leader->node; + rq->sl = leader->sl; + rq->lock = leader->lock; + rq->is_leader = false; + barrier(); + /* To make up for not unlocking the freed runlock */ + preempt_enable(); +} + +static void __init share_rqs(void) +{ + struct rq *rq, *leader; + int cpu; + + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + leader = rq->smp_leader; + + rq_lock(rq); + if (leader && rq != leader) { + printk(KERN_INFO "MuQSS sharing SMP runqueue from CPU %d to CPU %d\n", + leader->cpu, rq->cpu); + share_and_free_rq(leader, rq); + } else + rq_unlock(rq); + } + +#ifdef CONFIG_SCHED_MC + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + leader = rq->mc_leader; + + rq_lock(rq); + if (leader && rq != leader) { + printk(KERN_INFO "MuQSS sharing MC runqueue from CPU %d to CPU %d\n", + leader->cpu, rq->cpu); + share_and_free_rq(leader, rq); + } else + rq_unlock(rq); + } +#endif /* CONFIG_SCHED_MC */ + +#ifdef CONFIG_SCHED_SMT + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + leader = rq->smt_leader; + + rq_lock(rq); + if (leader && rq != leader) { + printk(KERN_INFO "MuQSS sharing SMT runqueue from CPU %d to CPU %d\n", + leader->cpu, rq->cpu); + share_and_free_rq(leader, rq); + } else + rq_unlock(rq); + } +#endif /* CONFIG_SCHED_SMT */ +} + +static void __init setup_rq_orders(void) +{ + int *selected_cpus, *ordered_cpus; + struct rq *rq, *other_rq; + int cpu, other_cpu, i; + + selected_cpus = kmalloc(sizeof(int) * NR_CPUS, GFP_ATOMIC); + ordered_cpus = kmalloc(sizeof(int) * NR_CPUS, GFP_ATOMIC); + + total_runqueues = 0; + for_each_online_cpu(cpu) { + int locality, total_rqs = 0, total_cpus = 0; + + rq = cpu_rq(cpu); + if (rq->is_leader) + total_runqueues++; + + for (locality = LOCALITY_SAME; locality <= LOCALITY_DISTANT; locality++) { + int selected_cpu_cnt, selected_cpu_idx, test_cpu_idx, cpu_idx, best_locality, test_cpu; + int ordered_cpus_idx; + + ordered_cpus_idx = -1; + selected_cpu_cnt = 0; + + for_each_online_cpu(test_cpu) { + if (cpu < num_online_cpus() / 2) + other_cpu = cpu + test_cpu; + else + other_cpu = cpu - test_cpu; + if (other_cpu < 0) + other_cpu += num_online_cpus(); + else + other_cpu %= num_online_cpus(); + /* gather CPUs of the same locality */ + if (rq->cpu_locality[other_cpu] == locality) { + selected_cpus[selected_cpu_cnt] = other_cpu; + selected_cpu_cnt++; + } + } + + /* reserve first CPU as starting point */ + if (selected_cpu_cnt > 0) { + ordered_cpus_idx++; + ordered_cpus[ordered_cpus_idx] = selected_cpus[ordered_cpus_idx]; + selected_cpus[ordered_cpus_idx] = -1; + } + + /* take each CPU and sort it within the same locality based on each inter-CPU localities */ + for (test_cpu_idx = 1; test_cpu_idx < selected_cpu_cnt; test_cpu_idx++) { + /* starting point with worst locality and current CPU */ + best_locality = LOCALITY_DISTANT; + selected_cpu_idx = test_cpu_idx; + + /* try to find the best locality within group */ + for (cpu_idx = 1; cpu_idx < selected_cpu_cnt; cpu_idx++) { + /* if CPU has not been used and locality is better */ + if (selected_cpus[cpu_idx] > -1) { + other_rq = cpu_rq(ordered_cpus[ordered_cpus_idx]); + if (best_locality > other_rq->cpu_locality[selected_cpus[cpu_idx]]) { + /* assign best locality and best CPU idx in array */ + best_locality = other_rq->cpu_locality[selected_cpus[cpu_idx]]; + selected_cpu_idx = cpu_idx; + } + } + } + + /* add our next best CPU to ordered list */ + ordered_cpus_idx++; + ordered_cpus[ordered_cpus_idx] = selected_cpus[selected_cpu_idx]; + /* mark this CPU as used */ + selected_cpus[selected_cpu_idx] = -1; + } + + /* set up RQ and CPU orders */ + for (test_cpu = 0; test_cpu <= ordered_cpus_idx; test_cpu++) { + other_rq = cpu_rq(ordered_cpus[test_cpu]); + /* set up cpu orders */ + rq->cpu_order[total_cpus++] = other_rq; + if (other_rq->is_leader) { + /* set up RQ orders */ + rq->rq_order[total_rqs++] = other_rq; + } + } + } + } + + kfree(selected_cpus); + kfree(ordered_cpus); + +#ifdef CONFIG_X86 + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + for (i = 0; i < total_runqueues; i++) { + printk(KERN_DEBUG "MuQSS CPU %d llc %d RQ order %d RQ %d llc %d\n", cpu, per_cpu(cpu_llc_id, cpu), i, + rq->rq_order[i]->cpu, per_cpu(cpu_llc_id, rq->rq_order[i]->cpu)); + } + } + + for_each_online_cpu(cpu) { + rq = cpu_rq(cpu); + for (i = 0; i < num_online_cpus(); i++) { + printk(KERN_DEBUG "MuQSS CPU %d llc %d CPU order %d RQ %d llc %d\n", cpu, per_cpu(cpu_llc_id, cpu), i, + rq->cpu_order[i]->cpu, per_cpu(cpu_llc_id, rq->cpu_order[i]->cpu)); + } + } +#endif +} + +void __init sched_init_smp(void) +{ + sched_init_numa(); + + /* + * There's no userspace yet to cause hotplug operations; hence all the + * cpu masks are stable and all blatant races in the below code cannot + * happen. + */ + mutex_lock(&sched_domains_mutex); + sched_init_domains(cpu_active_mask); + mutex_unlock(&sched_domains_mutex); + + /* Move init over to a non-isolated CPU */ + if (set_cpus_allowed_ptr(current, housekeeping_cpumask(HK_FLAG_DOMAIN)) < 0) + BUG(); + + local_irq_disable(); + mutex_lock(&sched_domains_mutex); + lock_all_rqs(); + + printk(KERN_INFO "MuQSS possible/present/online CPUs: %d/%d/%d\n", + num_possible_cpus(), num_present_cpus(), num_online_cpus()); + + select_leaders(); + + unlock_all_rqs(); + mutex_unlock(&sched_domains_mutex); + + share_rqs(); + + local_irq_enable(); + + setup_rq_orders(); + + switch (rqshare) { + case RQSHARE_ALL: + /* This should only ever read 1 */ + printk(KERN_INFO "MuQSS runqueue share type ALL total runqueues: %d\n", + total_runqueues); + break; + case RQSHARE_SMP: + printk(KERN_INFO "MuQSS runqueue share type SMP total runqueues: %d\n", + total_runqueues); + break; + case RQSHARE_MC: + printk(KERN_INFO "MuQSS runqueue share type MC total runqueues: %d\n", + total_runqueues); + break; + case RQSHARE_MC_LLC: + printk(KERN_INFO "MuQSS runqueue share type LLC total runqueues: %d\n", + total_runqueues); + break; + case RQSHARE_SMT: + printk(KERN_INFO "MuQSS runqueue share type SMT total runqueues: %d\n", + total_runqueues); + break; + case RQSHARE_NONE: + printk(KERN_INFO "MuQSS runqueue share type NONE total runqueues: %d\n", + total_runqueues); + break; + } + + sched_smp_initialized = true; +} +#else +void __init sched_init_smp(void) +{ + sched_smp_initialized = true; +} +#endif /* CONFIG_SMP */ + +int in_sched_functions(unsigned long addr) +{ + return in_lock_functions(addr) || + (addr >= (unsigned long)__sched_text_start + && addr < (unsigned long)__sched_text_end); +} + +#ifdef CONFIG_CGROUP_SCHED +/* task group related information */ +struct task_group { + struct cgroup_subsys_state css; + + struct rcu_head rcu; + struct list_head list; + + struct task_group *parent; + struct list_head siblings; + struct list_head children; +}; + +/* + * Default task group. + * Every task in system belongs to this group at bootup. + */ +struct task_group root_task_group; +LIST_HEAD(task_groups); + +/* Cacheline aligned slab cache for task_group */ +static struct kmem_cache *task_group_cache __read_mostly; +#endif /* CONFIG_CGROUP_SCHED */ + +void __init sched_init(void) +{ +#ifdef CONFIG_SMP + int cpu_ids; +#endif + int i; + struct rq *rq; + + wait_bit_init(); + + prio_ratios[0] = 128; + for (i = 1 ; i < NICE_WIDTH ; i++) + prio_ratios[i] = prio_ratios[i - 1] * 11 / 10; + + skiplist_node_init(&init_task.node); + +#ifdef CONFIG_SMP + init_defrootdomain(); + cpumask_clear(&cpu_idle_map); +#else + uprq = &per_cpu(runqueues, 0); +#endif + +#ifdef CONFIG_CGROUP_SCHED + task_group_cache = KMEM_CACHE(task_group, 0); + + list_add(&root_task_group.list, &task_groups); + INIT_LIST_HEAD(&root_task_group.children); + INIT_LIST_HEAD(&root_task_group.siblings); +#endif /* CONFIG_CGROUP_SCHED */ + for_each_possible_cpu(i) { + rq = cpu_rq(i); + rq->node = kmalloc(sizeof(skiplist_node), GFP_ATOMIC); + skiplist_init(rq->node); + rq->sl = new_skiplist(rq->node); + rq->lock = kmalloc(sizeof(raw_spinlock_t), GFP_ATOMIC); + raw_spin_lock_init(rq->lock); + rq->nr_running = 0; + rq->nr_uninterruptible = 0; + rq->nr_switches = 0; + rq->clock = rq->old_clock = rq->last_niffy = rq->niffies = 0; + rq->last_jiffy = jiffies; + rq->user_ns = rq->nice_ns = rq->softirq_ns = rq->system_ns = + rq->iowait_ns = rq->idle_ns = 0; + rq->dither = 0; + set_rq_task(rq, &init_task); + rq->iso_ticks = 0; + rq->iso_refractory = false; +#ifdef CONFIG_SMP + rq->is_leader = true; + rq->smp_leader = NULL; +#ifdef CONFIG_SCHED_MC + rq->mc_leader = NULL; +#endif +#ifdef CONFIG_SCHED_SMT + rq->smt_leader = NULL; +#endif + rq->sd = NULL; + rq->rd = NULL; + rq->online = false; + rq->cpu = i; + rq_attach_root(rq, &def_root_domain); +#endif /* CONFIG_SMP */ + init_rq_hrexpiry(rq); + atomic_set(&rq->nr_iowait, 0); + } + +#ifdef CONFIG_SMP + cpu_ids = i; + /* + * Set the base locality for cpu cache distance calculation to + * "distant" (3). Make sure the distance from a CPU to itself is 0. + */ + for_each_possible_cpu(i) { + int j; + + rq = cpu_rq(i); +#ifdef CONFIG_SCHED_SMT + rq->siblings_idle = sole_cpu_idle; +#endif +#ifdef CONFIG_SCHED_MC + rq->cache_idle = sole_cpu_idle; +#endif + rq->cpu_locality = kmalloc(cpu_ids * sizeof(int *), GFP_ATOMIC); + for_each_possible_cpu(j) { + if (i == j) + rq->cpu_locality[j] = LOCALITY_SAME; + else + rq->cpu_locality[j] = LOCALITY_DISTANT; + } + rq->rq_order = kmalloc(cpu_ids * sizeof(struct rq *), GFP_ATOMIC); + rq->cpu_order = kmalloc(cpu_ids * sizeof(struct rq *), GFP_ATOMIC); + rq->rq_order[0] = rq->cpu_order[0] = rq; + for (j = 1; j < cpu_ids; j++) + rq->rq_order[j] = rq->cpu_order[j] = cpu_rq(j); + } +#endif + + /* + * The boot idle thread does lazy MMU switching as well: + */ + mmgrab(&init_mm); + enter_lazy_tlb(&init_mm, current); + + /* + * Make us the idle thread. Technically, schedule() should not be + * called from this thread, however somewhere below it might be, + * but because we are the idle thread, we just pick up running again + * when this runqueue becomes "idle". + */ + init_idle(current, smp_processor_id()); + +#ifdef CONFIG_SMP + idle_thread_set_boot_cpu(); +#endif /* SMP */ + + init_schedstats(); + + psi_init(); +} + +#ifdef CONFIG_DEBUG_ATOMIC_SLEEP +static inline int preempt_count_equals(int preempt_offset) +{ + int nested = preempt_count() + rcu_preempt_depth(); + + return (nested == preempt_offset); +} + +void __might_sleep(const char *file, int line, int preempt_offset) +{ + /* + * Blocking primitives will set (and therefore destroy) current->state, + * since we will exit with TASK_RUNNING make sure we enter with it, + * otherwise we will destroy state. + */ + WARN_ONCE(current->state != TASK_RUNNING && current->task_state_change, + "do not call blocking ops when !TASK_RUNNING; " + "state=%lx set at [<%p>] %pS\n", + current->state, + (void *)current->task_state_change, + (void *)current->task_state_change); + + ___might_sleep(file, line, preempt_offset); +} +EXPORT_SYMBOL(__might_sleep); + +void __cant_sleep(const char *file, int line, int preempt_offset) +{ + static unsigned long prev_jiffy; + + if (irqs_disabled()) + return; + + if (!IS_ENABLED(CONFIG_PREEMPT_COUNT)) + return; + + if (preempt_count() > preempt_offset) + return; + + if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy) + return; + prev_jiffy = jiffies; + + printk(KERN_ERR "BUG: assuming atomic context at %s:%d\n", file, line); + printk(KERN_ERR "in_atomic(): %d, irqs_disabled(): %d, pid: %d, name: %s\n", + in_atomic(), irqs_disabled(), + current->pid, current->comm); + + debug_show_held_locks(current); + dump_stack(); + add_taint(TAINT_WARN, LOCKDEP_STILL_OK); +} +EXPORT_SYMBOL_GPL(__cant_sleep); + +void ___might_sleep(const char *file, int line, int preempt_offset) +{ + /* Ratelimiting timestamp: */ + static unsigned long prev_jiffy; + + unsigned long preempt_disable_ip; + + /* WARN_ON_ONCE() by default, no rate limit required: */ + rcu_sleep_check(); + + if ((preempt_count_equals(preempt_offset) && !irqs_disabled() && + !is_idle_task(current) && !current->non_block_count) || + system_state == SYSTEM_BOOTING || system_state > SYSTEM_RUNNING || + oops_in_progress) + return; + + if (time_before(jiffies, prev_jiffy + HZ) && prev_jiffy) + return; + prev_jiffy = jiffies; + + /* Save this before calling printk(), since that will clobber it: */ + preempt_disable_ip = get_preempt_disable_ip(current); + + printk(KERN_ERR + "BUG: sleeping function called from invalid context at %s:%d\n", + file, line); + printk(KERN_ERR + "in_atomic(): %d, irqs_disabled(): %d, non_block: %d, pid: %d, name: %s\n", + in_atomic(), irqs_disabled(), current->non_block_count, + current->pid, current->comm); + + if (task_stack_end_corrupted(current)) + printk(KERN_EMERG "Thread overran stack, or stack corrupted\n"); + + debug_show_held_locks(current); + if (irqs_disabled()) + print_irqtrace_events(current); + if (IS_ENABLED(CONFIG_DEBUG_PREEMPT) + && !preempt_count_equals(preempt_offset)) { + pr_err("Preemption disabled at:"); + print_ip_sym(KERN_ERR, preempt_disable_ip); + } + dump_stack(); + add_taint(TAINT_WARN, LOCKDEP_STILL_OK); +} +EXPORT_SYMBOL(___might_sleep); +#endif + +#ifdef CONFIG_MAGIC_SYSRQ +static inline void normalise_rt_tasks(void) +{ + struct sched_attr attr = {}; + struct task_struct *g, *p; + struct rq_flags rf; + struct rq *rq; + + read_lock(&tasklist_lock); + for_each_process_thread(g, p) { + /* + * Only normalize user tasks: + */ + if (p->flags & PF_KTHREAD) + continue; + + if (!rt_task(p) && !iso_task(p)) + continue; + + rq = task_rq_lock(p, &rf); + __setscheduler(p, rq, SCHED_NORMAL, 0, &attr, false); + task_rq_unlock(rq, p, &rf); + } + read_unlock(&tasklist_lock); +} + +void normalize_rt_tasks(void) +{ + normalise_rt_tasks(); +} +#endif /* CONFIG_MAGIC_SYSRQ */ + +#if defined(CONFIG_IA64) || defined(CONFIG_KGDB_KDB) +/* + * These functions are only useful for the IA64 MCA handling, or kdb. + * + * They can only be called when the whole system has been + * stopped - every CPU needs to be quiescent, and no scheduling + * activity can take place. Using them for anything else would + * be a serious bug, and as a result, they aren't even visible + * under any other configuration. + */ + +/** + * curr_task - return the current task for a given CPU. + * @cpu: the processor in question. + * + * ONLY VALID WHEN THE WHOLE SYSTEM IS STOPPED! + * + * Return: The current task for @cpu. + */ +struct task_struct *curr_task(int cpu) +{ + return cpu_curr(cpu); +} + +#endif /* defined(CONFIG_IA64) || defined(CONFIG_KGDB_KDB) */ + +#ifdef CONFIG_IA64 +/** + * ia64_set_curr_task - set the current task for a given CPU. + * @cpu: the processor in question. + * @p: the task pointer to set. + * + * Description: This function must only be used when non-maskable interrupts + * are serviced on a separate stack. It allows the architecture to switch the + * notion of the current task on a CPU in a non-blocking manner. This function + * must be called with all CPU's synchronised, and interrupts disabled, the + * and caller must save the original value of the current task (see + * curr_task() above) and restore that value before reenabling interrupts and + * re-starting the system. + * + * ONLY VALID WHEN THE WHOLE SYSTEM IS STOPPED! + */ +void ia64_set_curr_task(int cpu, struct task_struct *p) +{ + cpu_curr(cpu) = p; +} + +#endif + +void init_idle_bootup_task(struct task_struct *idle) +{} + +#ifdef CONFIG_SCHED_DEBUG +__read_mostly bool sched_debug_enabled; + +void proc_sched_show_task(struct task_struct *p, struct pid_namespace *ns, + struct seq_file *m) +{ + seq_printf(m, "%s (%d, #threads: %d)\n", p->comm, task_pid_nr_ns(p, ns), + get_nr_threads(p)); +} + +void proc_sched_set_task(struct task_struct *p) +{} +#endif + +#ifdef CONFIG_CGROUP_SCHED +static void sched_free_group(struct task_group *tg) +{ + kmem_cache_free(task_group_cache, tg); +} + +/* allocate runqueue etc for a new task group */ +struct task_group *sched_create_group(struct task_group *parent) +{ + struct task_group *tg; + + tg = kmem_cache_alloc(task_group_cache, GFP_KERNEL | __GFP_ZERO); + if (!tg) + return ERR_PTR(-ENOMEM); + + return tg; +} + +void sched_online_group(struct task_group *tg, struct task_group *parent) +{ +} + +/* rcu callback to free various structures associated with a task group */ +static void sched_free_group_rcu(struct rcu_head *rhp) +{ + /* Now it should be safe to free those cfs_rqs */ + sched_free_group(container_of(rhp, struct task_group, rcu)); +} + +void sched_destroy_group(struct task_group *tg) +{ + /* Wait for possible concurrent references to cfs_rqs complete */ + call_rcu(&tg->rcu, sched_free_group_rcu); +} + +void sched_offline_group(struct task_group *tg) +{ +} + +static inline struct task_group *css_tg(struct cgroup_subsys_state *css) +{ + return css ? container_of(css, struct task_group, css) : NULL; +} + +static struct cgroup_subsys_state * +cpu_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) +{ + struct task_group *parent = css_tg(parent_css); + struct task_group *tg; + + if (!parent) { + /* This is early initialization for the top cgroup */ + return &root_task_group.css; + } + + tg = sched_create_group(parent); + if (IS_ERR(tg)) + return ERR_PTR(-ENOMEM); + return &tg->css; +} + +/* Expose task group only after completing cgroup initialization */ +static int cpu_cgroup_css_online(struct cgroup_subsys_state *css) +{ + struct task_group *tg = css_tg(css); + struct task_group *parent = css_tg(css->parent); + + if (parent) + sched_online_group(tg, parent); + return 0; +} + +static void cpu_cgroup_css_released(struct cgroup_subsys_state *css) +{ + struct task_group *tg = css_tg(css); + + sched_offline_group(tg); +} + +static void cpu_cgroup_css_free(struct cgroup_subsys_state *css) +{ + struct task_group *tg = css_tg(css); + + /* + * Relies on the RCU grace period between css_released() and this. + */ + sched_free_group(tg); +} + +static void cpu_cgroup_fork(struct task_struct *task) +{ +} + +static int cpu_cgroup_can_attach(struct cgroup_taskset *tset) +{ + return 0; +} + +static void cpu_cgroup_attach(struct cgroup_taskset *tset) +{ +} + +static struct cftype cpu_legacy_files[] = { + { } /* Terminate */ +}; + +static struct cftype cpu_files[] = { + { } /* terminate */ +}; + +static int cpu_extra_stat_show(struct seq_file *sf, + struct cgroup_subsys_state *css) +{ + return 0; +} + +struct cgroup_subsys cpu_cgrp_subsys = { + .css_alloc = cpu_cgroup_css_alloc, + .css_online = cpu_cgroup_css_online, + .css_released = cpu_cgroup_css_released, + .css_free = cpu_cgroup_css_free, + .css_extra_stat_show = cpu_extra_stat_show, + .fork = cpu_cgroup_fork, + .can_attach = cpu_cgroup_can_attach, + .attach = cpu_cgroup_attach, + .legacy_cftypes = cpu_files, + .legacy_cftypes = cpu_legacy_files, + .dfl_cftypes = cpu_files, + .early_init = true, + .threaded = true, +}; +#endif /* CONFIG_CGROUP_SCHED */ + +void call_trace_sched_update_nr_running(struct rq *rq, int count) +{ + trace_sched_update_nr_running_tp(rq, count); +} + +/* CFS Compat */ +#ifdef CONFIG_RCU_TORTURE_TEST +int sysctl_sched_rt_runtime; +#endif diff --git a/kernel/sched/MuQSS.h b/kernel/sched/MuQSS.h new file mode 100644 index 00000000000..62e01b82600 --- /dev/null +++ b/kernel/sched/MuQSS.h @@ -0,0 +1,1084 @@ +/* SPDX-License-Identifier: GPL-2.0 */ +#ifndef MUQSS_SCHED_H +#define MUQSS_SCHED_H + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#ifdef CONFIG_PARAVIRT +#include +#endif + +#include "cpupri.h" + +#include + +#ifdef CONFIG_SCHED_DEBUG +# define SCHED_WARN_ON(x) WARN_ONCE(x, #x) +#else +# define SCHED_WARN_ON(x) ((void)(x)) +#endif + +/* Wake flags. The first three directly map to some SD flag value */ +#define WF_EXEC 0x02 /* Wakeup after exec; maps to SD_BALANCE_EXEC */ +#define WF_FORK 0x04 /* Wakeup after fork; maps to SD_BALANCE_FORK */ +#define WF_TTWU 0x08 /* Wakeup; maps to SD_BALANCE_WAKE */ + +#define WF_SYNC 0x10 /* Waker goes to sleep after wakeup */ +#define WF_MIGRATED 0x20 /* Internal use, task got migrated */ +#define WF_ON_CPU 0x40 /* Wakee is on_cpu */ + +#ifdef CONFIG_SMP +static_assert(WF_EXEC == SD_BALANCE_EXEC); +static_assert(WF_FORK == SD_BALANCE_FORK); +static_assert(WF_TTWU == SD_BALANCE_WAKE); +#endif + +/* task_struct::on_rq states: */ +#define TASK_ON_RQ_QUEUED 1 +#define TASK_ON_RQ_MIGRATING 2 + +extern void call_trace_sched_update_nr_running(struct rq *rq, int count); + +struct rq; + +#ifdef CONFIG_SMP + +static inline bool sched_asym_prefer(int a, int b) +{ + return arch_asym_cpu_priority(a) > arch_asym_cpu_priority(b); +} + +struct perf_domain { + struct em_perf_domain *em_pd; + struct perf_domain *next; + struct rcu_head rcu; +}; + +/* Scheduling group status flags */ +#define SG_OVERLOAD 0x1 /* More than one runnable task on a CPU. */ +#define SG_OVERUTILIZED 0x2 /* One or more CPUs are over-utilized. */ + +/* + * We add the notion of a root-domain which will be used to define per-domain + * variables. Each exclusive cpuset essentially defines an island domain by + * fully partitioning the member cpus from any other cpuset. Whenever a new + * exclusive cpuset is created, we also create and attach a new root-domain + * object. + * + */ +struct root_domain { + atomic_t refcount; + atomic_t rto_count; + struct rcu_head rcu; + cpumask_var_t span; + cpumask_var_t online; + + /* + * Indicate pullable load on at least one CPU, e.g: + * - More than one runnable task + * - Running task is misfit + */ + int overload; + + /* Indicate one or more cpus over-utilized (tipping point) */ + int overutilized; + + /* + * The bit corresponding to a CPU gets set here if such CPU has more + * than one runnable -deadline task (as it is below for RT tasks). + */ + cpumask_var_t dlo_mask; + atomic_t dlo_count; + + /* Replace unused CFS structures with void */ + //struct dl_bw dl_bw; + //struct cpudl cpudl; + void *dl_bw; + void *cpudl; + u64 visit_gen; + + /* + * The "RT overload" flag: it gets set if a CPU has more than + * one runnable RT task. + */ + cpumask_var_t rto_mask; + //struct cpupri cpupri; + void *cpupri; + + unsigned long max_cpu_capacity; + + /* + * NULL-terminated list of performance domains intersecting with the + * CPUs of the rd. Protected by RCU. + */ + struct perf_domain *pd; +}; + +extern void init_defrootdomain(void); +extern int sched_init_domains(const struct cpumask *cpu_map); +extern void rq_attach_root(struct rq *rq, struct root_domain *rd); + +static inline void cpupri_cleanup(void __maybe_unused *cpupri) +{ +} + +static inline void cpudl_cleanup(void __maybe_unused *cpudl) +{ +} + +static inline void init_dl_bw(void __maybe_unused *dl_bw) +{ +} + +static inline int cpudl_init(void __maybe_unused *dl_bw) +{ + return 0; +} + +static inline int cpupri_init(void __maybe_unused *cpupri) +{ + return 0; +} +#endif /* CONFIG_SMP */ + +/* + * This is the main, per-CPU runqueue data structure. + * This data should only be modified by the local cpu. + */ +struct rq { + raw_spinlock_t *lock; + raw_spinlock_t *orig_lock; + + struct task_struct __rcu *curr; + struct task_struct *idle; + struct task_struct *stop; + struct mm_struct *prev_mm; + + unsigned int nr_running; + /* + * This is part of a global counter where only the total sum + * over all CPUs matters. A task can increase this counter on + * one CPU and if it got migrated afterwards it may decrease + * it on another CPU. Always updated under the runqueue lock: + */ + unsigned long nr_uninterruptible; +#ifdef CONFIG_SMP + unsigned int ttwu_pending; +#endif + u64 nr_switches; + + /* Stored data about rq->curr to work outside rq lock */ + u64 rq_deadline; + int rq_prio; + + /* Best queued id for use outside lock */ + u64 best_key; + + unsigned long last_scheduler_tick; /* Last jiffy this RQ ticked */ + unsigned long last_jiffy; /* Last jiffy this RQ updated rq clock */ + u64 niffies; /* Last time this RQ updated rq clock */ + u64 last_niffy; /* Last niffies as updated by local clock */ + u64 last_jiffy_niffies; /* Niffies @ last_jiffy */ + + u64 load_update; /* When we last updated load */ + unsigned long load_avg; /* Rolling load average */ +#ifdef CONFIG_HAVE_SCHED_AVG_IRQ + u64 irq_load_update; /* When we last updated IRQ load */ + unsigned long irq_load_avg; /* Rolling IRQ load average */ +#endif +#ifdef CONFIG_SMT_NICE + struct mm_struct *rq_mm; + int rq_smt_bias; /* Policy/nice level bias across smt siblings */ +#endif + /* Accurate timekeeping data */ + unsigned long user_ns, nice_ns, irq_ns, softirq_ns, system_ns, + iowait_ns, idle_ns; + atomic_t nr_iowait; + +#ifdef CONFIG_MEMBARRIER + int membarrier_state; +#endif + + skiplist_node *node; + skiplist *sl; +#ifdef CONFIG_SMP + struct task_struct *preempt; /* Preempt triggered on this task */ + struct task_struct *preempting; /* Hint only, what task is preempting */ + + int cpu; /* cpu of this runqueue */ + bool online; + + struct root_domain *rd; + struct sched_domain *sd; + + unsigned long cpu_capacity_orig; + + int *cpu_locality; /* CPU relative cache distance */ + struct rq **rq_order; /* Shared RQs ordered by relative cache distance */ + struct rq **cpu_order; /* RQs of discrete CPUs ordered by distance */ + + bool is_leader; + struct rq *smp_leader; /* First physical CPU per node */ +#ifdef CONFIG_SCHED_THERMAL_PRESSURE + struct sched_avg avg_thermal; +#endif /* CONFIG_SCHED_THERMAL_PRESSURE */ +#ifdef CONFIG_SCHED_SMT + struct rq *smt_leader; /* First logical CPU in SMT siblings */ + cpumask_t thread_mask; + bool (*siblings_idle)(struct rq *rq); + /* See if all smt siblings are idle */ +#endif /* CONFIG_SCHED_SMT */ +#ifdef CONFIG_SCHED_MC + struct rq *mc_leader; /* First logical CPU in MC siblings */ + cpumask_t core_mask; + bool (*cache_idle)(struct rq *rq); + /* See if all cache siblings are idle */ +#endif /* CONFIG_SCHED_MC */ +#endif /* CONFIG_SMP */ + +#ifdef CONFIG_IRQ_TIME_ACCOUNTING + u64 prev_irq_time; +#endif /* CONFIG_IRQ_TIME_ACCOUNTING */ +#ifdef CONFIG_PARAVIRT + u64 prev_steal_time; +#endif /* CONFIG_PARAVIRT */ +#ifdef CONFIG_PARAVIRT_TIME_ACCOUNTING + u64 prev_steal_time_rq; +#endif /* CONFIG_PARAVIRT_TIME_ACCOUNTING */ + + u64 clock, old_clock, last_tick; + /* Ensure that all clocks are in the same cache line */ + u64 clock_task ____cacheline_aligned; + int dither; + + int iso_ticks; + bool iso_refractory; + +#ifdef CONFIG_HIGH_RES_TIMERS + struct hrtimer hrexpiry_timer; +#endif + + int rt_nr_running; /* Number real time tasks running */ +#ifdef CONFIG_SCHEDSTATS + + /* latency stats */ + struct sched_info rq_sched_info; + unsigned long long rq_cpu_time; + /* could above be rq->cfs_rq.exec_clock + rq->rt_rq.rt_runtime ? */ + + /* sys_sched_yield() stats */ + unsigned int yld_count; + + /* schedule() stats */ + unsigned int sched_switch; + unsigned int sched_count; + unsigned int sched_goidle; + + /* try_to_wake_up() stats */ + unsigned int ttwu_count; + unsigned int ttwu_local; +#endif /* CONFIG_SCHEDSTATS */ + +#ifdef CONFIG_CPU_IDLE + /* Must be inspected within a rcu lock section */ + struct cpuidle_state *idle_state; +#endif +}; + +static inline u64 __rq_clock_broken(struct rq *rq) +{ + return READ_ONCE(rq->clock); +} + +static inline u64 rq_clock(struct rq *rq) +{ + lockdep_assert_held(rq->lock); + + return rq->clock; +} + +static inline u64 rq_clock_task(struct rq *rq) +{ + lockdep_assert_held(rq->lock); + + return rq->clock_task; +} + +/** + * By default the decay is the default pelt decay period. + * The decay shift can change the decay period in + * multiples of 32. + * Decay shift Decay period(ms) + * 0 32 + * 1 64 + * 2 128 + * 3 256 + * 4 512 + */ +extern int sched_thermal_decay_shift; + +static inline u64 rq_clock_thermal(struct rq *rq) +{ + return rq_clock_task(rq) >> sched_thermal_decay_shift; +} + +struct rq_flags { + unsigned long flags; +}; + +#ifdef CONFIG_SMP +struct rq *cpu_rq(int cpu); +#endif + +#ifndef CONFIG_SMP +extern struct rq *uprq; +#define cpu_rq(cpu) (uprq) +#define this_rq() (uprq) +#define raw_rq() (uprq) +#define task_rq(p) (uprq) +#define cpu_curr(cpu) ((uprq)->curr) +#else /* CONFIG_SMP */ +DECLARE_PER_CPU_SHARED_ALIGNED(struct rq, runqueues); +#define cpu_rq(cpu) (&per_cpu(runqueues, (cpu))) +#define this_rq() this_cpu_ptr(&runqueues) +#define task_rq(p) cpu_rq(task_cpu(p)) +#define cpu_curr(cpu) (cpu_rq(cpu)->curr) +#define raw_rq() raw_cpu_ptr(&runqueues) +#endif /* CONFIG_SMP */ + +static inline int task_current(struct rq *rq, struct task_struct *p) +{ + return rq->curr == p; +} + +static inline int task_running(struct rq *rq, struct task_struct *p) +{ +#ifdef CONFIG_SMP + return p->on_cpu; +#else + return task_current(rq, p); +#endif +} + +static inline int task_on_rq_queued(struct task_struct *p) +{ + return p->on_rq == TASK_ON_RQ_QUEUED; +} + +static inline int task_on_rq_migrating(struct task_struct *p) +{ + return READ_ONCE(p->on_rq) == TASK_ON_RQ_MIGRATING; +} + +static inline void rq_lock(struct rq *rq) + __acquires(rq->lock) +{ + raw_spin_lock(rq->lock); +} + +static inline void rq_unlock(struct rq *rq) + __releases(rq->lock) +{ + raw_spin_unlock(rq->lock); +} + +static inline void rq_lock_irq(struct rq *rq) + __acquires(rq->lock) +{ + raw_spin_lock_irq(rq->lock); +} + +static inline void rq_unlock_irq(struct rq *rq, struct rq_flags __always_unused *rf) + __releases(rq->lock) +{ + raw_spin_unlock_irq(rq->lock); +} + +static inline void rq_lock_irqsave(struct rq *rq, struct rq_flags *rf) + __acquires(rq->lock) +{ + raw_spin_lock_irqsave(rq->lock, rf->flags); +} + +static inline void rq_unlock_irqrestore(struct rq *rq, struct rq_flags *rf) + __releases(rq->lock) +{ + raw_spin_unlock_irqrestore(rq->lock, rf->flags); +} + +static inline struct rq *task_rq_lock(struct task_struct *p, struct rq_flags *rf) + __acquires(p->pi_lock) + __acquires(rq->lock) +{ + struct rq *rq; + + while (42) { + raw_spin_lock_irqsave(&p->pi_lock, rf->flags); + rq = task_rq(p); + raw_spin_lock(rq->lock); + if (likely(rq == task_rq(p))) + break; + raw_spin_unlock(rq->lock); + raw_spin_unlock_irqrestore(&p->pi_lock, rf->flags); + } + return rq; +} + +static inline void task_rq_unlock(struct rq *rq, struct task_struct *p, struct rq_flags *rf) + __releases(rq->lock) + __releases(p->pi_lock) +{ + rq_unlock(rq); + raw_spin_unlock_irqrestore(&p->pi_lock, rf->flags); +} + +static inline struct rq *__task_rq_lock(struct task_struct *p, struct rq_flags __always_unused *rf) + __acquires(rq->lock) +{ + struct rq *rq; + + lockdep_assert_held(&p->pi_lock); + + while (42) { + rq = task_rq(p); + raw_spin_lock(rq->lock); + if (likely(rq == task_rq(p))) + break; + raw_spin_unlock(rq->lock); + } + return rq; +} + +static inline void __task_rq_unlock(struct rq *rq, struct rq_flags __always_unused *rf) +{ + rq_unlock(rq); +} + +static inline struct rq * +this_rq_lock_irq(struct rq_flags *rf) + __acquires(rq->lock) +{ + struct rq *rq; + + local_irq_disable(); + rq = this_rq(); + rq_lock(rq); + return rq; +} + +/* + * {de,en}queue flags: Most not used on MuQSS. + * + * DEQUEUE_SLEEP - task is no longer runnable + * ENQUEUE_WAKEUP - task just became runnable + * + * SAVE/RESTORE - an otherwise spurious dequeue/enqueue, done to ensure tasks + * are in a known state which allows modification. Such pairs + * should preserve as much state as possible. + * + * MOVE - paired with SAVE/RESTORE, explicitly does not preserve the location + * in the runqueue. + * + * ENQUEUE_HEAD - place at front of runqueue (tail if not specified) + * ENQUEUE_REPLENISH - CBS (replenish runtime and postpone deadline) + * ENQUEUE_MIGRATED - the task was migrated during wakeup + * + */ + +#define DEQUEUE_SLEEP 0x01 +#define DEQUEUE_SAVE 0x02 /* matches ENQUEUE_RESTORE */ + +#define ENQUEUE_WAKEUP 0x01 +#define ENQUEUE_RESTORE 0x02 + +#ifdef CONFIG_SMP +#define ENQUEUE_MIGRATED 0x40 +#else +#define ENQUEUE_MIGRATED 0x00 +#endif + +#ifdef CONFIG_NUMA +enum numa_topology_type { + NUMA_DIRECT, + NUMA_GLUELESS_MESH, + NUMA_BACKPLANE, +}; +extern enum numa_topology_type sched_numa_topology_type; +extern int sched_max_numa_distance; +extern bool find_numa_distance(int distance); +extern void sched_init_numa(void); +extern void sched_domains_numa_masks_set(unsigned int cpu); +extern void sched_domains_numa_masks_clear(unsigned int cpu); +extern int sched_numa_find_closest(const struct cpumask *cpus, int cpu); +#else +static inline void sched_init_numa(void) { } +static inline void sched_domains_numa_masks_set(unsigned int cpu) { } +static inline void sched_domains_numa_masks_clear(unsigned int cpu) { } +static inline int sched_numa_find_closest(const struct cpumask *cpus, int cpu) +{ + return nr_cpu_ids; +} +#endif + +extern struct mutex sched_domains_mutex; +extern struct static_key_false sched_schedstats; + +#define rcu_dereference_check_sched_domain(p) \ + rcu_dereference_check((p), \ + lockdep_is_held(&sched_domains_mutex)) + +#define SCA_CHECK 0x01 + +#ifdef CONFIG_SMP + +/* + * The domain tree (rq->sd) is protected by RCU's quiescent state transition. + * See destroy_sched_domains: call_rcu for details. + * + * The domain tree of any CPU may only be accessed from within + * preempt-disabled sections. + */ +#define for_each_domain(cpu, __sd) \ + for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); \ + __sd; __sd = __sd->parent) + +/** + * highest_flag_domain - Return highest sched_domain containing flag. + * @cpu: The cpu whose highest level of sched domain is to + * be returned. + * @flag: The flag to check for the highest sched_domain + * for the given cpu. + * + * Returns the highest sched_domain of a cpu which contains the given flag. + */ +static inline struct sched_domain *highest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd, *hsd = NULL; + + for_each_domain(cpu, sd) { + if (!(sd->flags & flag)) + break; + hsd = sd; + } + + return hsd; +} + +static inline struct sched_domain *lowest_flag_domain(int cpu, int flag) +{ + struct sched_domain *sd; + + for_each_domain(cpu, sd) { + if (sd->flags & flag) + break; + } + + return sd; +} + +DECLARE_PER_CPU(struct sched_domain *, sd_llc); +DECLARE_PER_CPU(int, sd_llc_size); +DECLARE_PER_CPU(int, sd_llc_id); +DECLARE_PER_CPU(struct sched_domain_shared *, sd_llc_shared); +DECLARE_PER_CPU(struct sched_domain *, sd_numa); +DECLARE_PER_CPU(struct sched_domain *, sd_asym_packing); +DECLARE_PER_CPU(struct sched_domain *, sd_asym_cpucapacity); + +struct sched_group_capacity { + atomic_t ref; + /* + * CPU capacity of this group, SCHED_CAPACITY_SCALE being max capacity + * for a single CPU. + */ + unsigned long capacity; + unsigned long min_capacity; /* Min per-CPU capacity in group */ + unsigned long max_capacity; /* Max per-CPU capacity in group */ + unsigned long next_update; + int imbalance; /* XXX unrelated to capacity but shared group state */ + +#ifdef CONFIG_SCHED_DEBUG + int id; +#endif + + unsigned long cpumask[]; /* balance mask */ +}; + +struct sched_group { + struct sched_group *next; /* Must be a circular list */ + atomic_t ref; + + unsigned int group_weight; + struct sched_group_capacity *sgc; + int asym_prefer_cpu; /* cpu of highest priority in group */ + + /* + * The CPUs this group covers. + * + * NOTE: this field is variable length. (Allocated dynamically + * by attaching extra space to the end of the structure, + * depending on how many CPUs the kernel has booted up with) + */ + unsigned long cpumask[0]; +}; + +static inline struct cpumask *sched_group_span(struct sched_group *sg) +{ + return to_cpumask(sg->cpumask); +} + +/* + * See build_balance_mask(). + */ +static inline struct cpumask *group_balance_mask(struct sched_group *sg) +{ + return to_cpumask(sg->sgc->cpumask); +} + +/** + * group_first_cpu - Returns the first cpu in the cpumask of a sched_group. + * @group: The group whose first cpu is to be returned. + */ +static inline unsigned int group_first_cpu(struct sched_group *group) +{ + return cpumask_first(sched_group_span(group)); +} + + +#ifdef CONFIG_SCHED_DEBUG +void update_sched_domain_debugfs(void); +void dirty_sched_domain_sysctl(int cpu); +#else +static inline void update_sched_domain_debugfs(void) +{ +} +static inline void dirty_sched_domain_sysctl(int cpu) +{ +} +#endif + +extern int sched_update_scaling(void); + +extern void flush_smp_call_function_from_idle(void); + +extern void set_cpus_allowed_common(struct task_struct *p, const struct cpumask *new_mask, u32 flags); + +extern void set_rq_online (struct rq *rq); +extern void set_rq_offline(struct rq *rq); +extern bool sched_smp_initialized; + +static inline void update_group_capacity(struct sched_domain *sd, int cpu) +{ +} + +static inline void trigger_load_balance(struct rq *rq) +{ +} + +#else /* CONFIG_SMP */ + +static inline void flush_smp_call_function_from_idle(void) { } + +#endif /* CONFIG_SMP */ + +#ifdef CONFIG_CPU_IDLE +static inline void idle_set_state(struct rq *rq, + struct cpuidle_state *idle_state) +{ + rq->idle_state = idle_state; +} + +static inline struct cpuidle_state *idle_get_state(struct rq *rq) +{ + SCHED_WARN_ON(!rcu_read_lock_held()); + return rq->idle_state; +} +#else +static inline void idle_set_state(struct rq *rq, + struct cpuidle_state *idle_state) +{ +} + +static inline struct cpuidle_state *idle_get_state(struct rq *rq) +{ + return NULL; +} +#endif + +#ifdef CONFIG_SCHED_DEBUG +extern bool sched_debug_enabled; +#endif + +extern void schedule_idle(void); + +#ifdef CONFIG_IRQ_TIME_ACCOUNTING +struct irqtime { + u64 total; + u64 tick_delta; + u64 irq_start_time; + struct u64_stats_sync sync; +}; + +DECLARE_PER_CPU(struct irqtime, cpu_irqtime); + +/* + * Returns the irqtime minus the softirq time computed by ksoftirqd. + * Otherwise ksoftirqd's sum_exec_runtime is substracted its own runtime + * and never move forward. + */ +static inline u64 irq_time_read(int cpu) +{ + struct irqtime *irqtime = &per_cpu(cpu_irqtime, cpu); + unsigned int seq; + u64 total; + + do { + seq = __u64_stats_fetch_begin(&irqtime->sync); + total = irqtime->total; + } while (__u64_stats_fetch_retry(&irqtime->sync, seq)); + + return total; +} +#endif /* CONFIG_IRQ_TIME_ACCOUNTING */ + +static inline bool sched_stop_runnable(struct rq *rq) +{ + return rq->stop && task_on_rq_queued(rq->stop); +} + +#ifdef CONFIG_SMP +static inline int cpu_of(struct rq *rq) +{ + return rq->cpu; +} +#else /* CONFIG_SMP */ +static inline int cpu_of(struct rq *rq) +{ + return 0; +} +#endif + +#ifdef CONFIG_CPU_FREQ +DECLARE_PER_CPU(struct update_util_data *, cpufreq_update_util_data); + +static inline void cpufreq_trigger(struct rq *rq, unsigned int flags) +{ + struct update_util_data *data; + + data = rcu_dereference_sched(*per_cpu_ptr(&cpufreq_update_util_data, + cpu_of(rq))); + + if (data) + data->func(data, rq->niffies, flags); +} +#else +static inline void cpufreq_trigger(struct rq *rq, unsigned int flag) +{ +} +#endif /* CONFIG_CPU_FREQ */ + +static __always_inline +unsigned int uclamp_rq_util_with(struct rq __maybe_unused *rq, unsigned int util, + struct task_struct __maybe_unused *p) +{ + return util; +} + +static inline bool uclamp_is_used(void) +{ + return false; +} + +#ifndef arch_scale_freq_tick +static __always_inline +void arch_scale_freq_tick(void) +{ +} +#endif + +#ifdef arch_scale_freq_capacity +#ifndef arch_scale_freq_invariant +#define arch_scale_freq_invariant() (true) +#endif +#else /* arch_scale_freq_capacity */ +#define arch_scale_freq_invariant() (false) +#endif + +#ifdef CONFIG_64BIT +static inline u64 read_sum_exec_runtime(struct task_struct *t) +{ + return tsk_seruntime(t); +} +#else +static inline u64 read_sum_exec_runtime(struct task_struct *t) +{ + struct rq_flags rf; + u64 ns; + struct rq *rq; + + rq = task_rq_lock(t, &rf); + ns = tsk_seruntime(t); + task_rq_unlock(rq, t, &rf); + + return ns; +} +#endif + +#ifndef arch_scale_freq_capacity +/** + * arch_scale_freq_capacity - get the frequency scale factor of a given CPU. + * @cpu: the CPU in question. + * + * Return: the frequency scale factor normalized against SCHED_CAPACITY_SCALE, i.e. + * + * f_curr + * ------ * SCHED_CAPACITY_SCALE + * f_max + */ +static __always_inline +unsigned long arch_scale_freq_capacity(int cpu) +{ + return SCHED_CAPACITY_SCALE; +} +#endif + +#ifdef CONFIG_NO_HZ_FULL +extern bool sched_can_stop_tick(struct rq *rq); +extern int __init sched_tick_offload_init(void); + +/* + * Tick may be needed by tasks in the runqueue depending on their policy and + * requirements. If tick is needed, lets send the target an IPI to kick it out of + * nohz mode if necessary. + */ +static inline void sched_update_tick_dependency(struct rq *rq) +{ + int cpu = cpu_of(rq); + + if (!tick_nohz_full_cpu(cpu)) + return; + + if (sched_can_stop_tick(rq)) + tick_nohz_dep_clear_cpu(cpu, TICK_DEP_BIT_SCHED); + else + tick_nohz_dep_set_cpu(cpu, TICK_DEP_BIT_SCHED); +} +#else +static inline int sched_tick_offload_init(void) { return 0; } +static inline void sched_update_tick_dependency(struct rq *rq) { } +#endif + +#define SCHED_FLAG_SUGOV 0x10000000 + +#ifdef CONFIG_SMP +/** + * enum cpu_util_type - CPU utilization type + * @FREQUENCY_UTIL: Utilization used to select frequency + * @ENERGY_UTIL: Utilization used during energy calculation + * + * The utilization signals of all scheduling classes (CFS/RT/DL) and IRQ time + * need to be aggregated differently depending on the usage made of them. This + * enum is used within schedutil_freq_util() to differentiate the types of + * enum is used within effective_cpu_util() to differentiate the types of + * utilization expected by the callers, and adjust the aggregation accordingly. + */ +enum cpu_util_type { + FREQUENCY_UTIL, + ENERGY_UTIL, +}; + +unsigned long effective_cpu_util(int cpu, unsigned long util_cfs, + unsigned long max, enum cpu_util_type type, + struct task_struct *p); + +static inline unsigned long cpu_bw_dl(struct rq *rq) +{ + return 0; +} + +static inline unsigned long cpu_util_dl(struct rq *rq) +{ + return 0; +} + +static inline unsigned long cpu_util_cfs(struct rq *rq) +{ + unsigned long ret = READ_ONCE(rq->load_avg); + + if (ret > SCHED_CAPACITY_SCALE) + ret = SCHED_CAPACITY_SCALE; + return ret; +} + +static inline unsigned long cpu_util_rt(struct rq *rq) +{ + unsigned long ret = READ_ONCE(rq->rt_nr_running); + + if (ret > SCHED_CAPACITY_SCALE) + ret = SCHED_CAPACITY_SCALE; + return ret; +} + +#ifdef CONFIG_HAVE_SCHED_AVG_IRQ +static inline unsigned long cpu_util_irq(struct rq *rq) +{ + unsigned long ret = READ_ONCE(rq->irq_load_avg); + + if (ret > SCHED_CAPACITY_SCALE) + ret = SCHED_CAPACITY_SCALE; + return ret; +} + +static inline +unsigned long scale_irq_capacity(unsigned long util, unsigned long irq, unsigned long max) +{ + util *= (max - irq); + util /= max; + + return util; + +} +#else /* CONFIG_HAVE_SCHED_AVG_IRQ */ +static inline unsigned long cpu_util_irq(struct rq *rq) +{ + return 0; +} + +static inline +unsigned long scale_irq_capacity(unsigned long util, unsigned long irq, unsigned long max) +{ + return util; +} +#endif /* CONFIG_HAVE_SCHED_AVG_IRQ */ +#endif /* CONFIG_SMP */ + +#if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL) +#define perf_domain_span(pd) (to_cpumask(((pd)->em_pd->cpus))) + +DECLARE_STATIC_KEY_FALSE(sched_energy_present); + +static inline bool sched_energy_enabled(void) +{ + return static_branch_unlikely(&sched_energy_present); +} + +#else /* ! (CONFIG_ENERGY_MODEL && CONFIG_CPU_FREQ_GOV_SCHEDUTIL) */ + +#define perf_domain_span(pd) NULL +static inline bool sched_energy_enabled(void) { return false; } + +#endif /* CONFIG_ENERGY_MODEL && CONFIG_CPU_FREQ_GOV_SCHEDUTIL */ + +#ifdef CONFIG_MEMBARRIER +/* + * The scheduler provides memory barriers required by membarrier between: + * - prior user-space memory accesses and store to rq->membarrier_state, + * - store to rq->membarrier_state and following user-space memory accesses. + * In the same way it provides those guarantees around store to rq->curr. + */ +static inline void membarrier_switch_mm(struct rq *rq, + struct mm_struct *prev_mm, + struct mm_struct *next_mm) +{ + int membarrier_state; + + if (prev_mm == next_mm) + return; + + membarrier_state = atomic_read(&next_mm->membarrier_state); + if (READ_ONCE(rq->membarrier_state) == membarrier_state) + return; + + WRITE_ONCE(rq->membarrier_state, membarrier_state); +} +#else +static inline void membarrier_switch_mm(struct rq *rq, + struct mm_struct *prev_mm, + struct mm_struct *next_mm) +{ +} +#endif + +#ifdef CONFIG_SMP +static inline bool is_per_cpu_kthread(struct task_struct *p) +{ + if (!(p->flags & PF_KTHREAD)) + return false; + + if (p->nr_cpus_allowed != 1) + return false; + + return true; +} +#endif + +extern void swake_up_all_locked(struct swait_queue_head *q); +extern void __prepare_to_swait(struct swait_queue_head *q, struct swait_queue *wait); + +#ifdef CONFIG_PREEMPT_DYNAMIC +extern int preempt_dynamic_mode; +extern int sched_dynamic_mode(const char *str); +extern void sched_dynamic_update(int mode); +#endif + +/* pelt.h compat CONFIG_SCHED_THERMAL_PRESSURE impossible with MUQSS */ +static inline int +update_thermal_load_avg(u64 now, struct rq *rq, u64 capacity) +{ + return 0; +} + +static inline u64 thermal_load_avg(struct rq *rq) +{ + return 0; +} + +#ifdef CONFIG_RCU_TORTURE_TEST +extern int sysctl_sched_rt_runtime; +#endif + + + +#endif /* MUQSS_SCHED_H */ diff --git a/kernel/sched/core.c b/kernel/sched/core.c index e5858999b54..ec0675c165c 100644 --- a/kernel/sched/core.c +++ b/kernel/sched/core.c @@ -74,7 +74,11 @@ __read_mostly int sysctl_resched_latency_warn_once = 1; * Number of tasks to iterate in a single balance run. * Limited because this is done with IRQs disabled. */ +#ifdef CONFIG_PCK_INTERACTIVE +const_debug unsigned int sysctl_sched_nr_migrate = 128; +#else const_debug unsigned int sysctl_sched_nr_migrate = 32; +#endif /* * period over which we measure -rt task CPU usage in us. @@ -4327,6 +4331,7 @@ context_switch(struct rq *rq, struct task_struct *prev, * finish_task_switch()'s mmdrop(). */ switch_mm_irqs_off(prev->active_mm, next->mm, next); + lru_gen_switch_mm(prev->active_mm, next->mm); if (!prev->mm) { // from kernel /* will mmdrop() in finish_task_switch(). */ @@ -7619,6 +7624,7 @@ void idle_task_exit(void) if (mm != &init_mm) { switch_mm(mm, &init_mm, current); + lru_gen_switch_mm(mm, &init_mm); finish_arch_post_lock_switch(); } diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c index 4f09afd2f32..19d6b1a7ec6 100644 --- a/kernel/sched/cpufreq_schedutil.c +++ b/kernel/sched/cpufreq_schedutil.c @@ -568,7 +568,11 @@ static int sugov_kthread_create(struct sugov_policy *sg_policy) struct task_struct *thread; struct sched_attr attr = { .size = sizeof(struct sched_attr), +#ifdef CONFIG_SCHED_MUQSS + .sched_policy = SCHED_RR, +#else .sched_policy = SCHED_DEADLINE, +#endif .sched_flags = SCHED_FLAG_SUGOV, .sched_nice = 0, .sched_priority = 0, diff --git a/kernel/sched/cpupri.h b/kernel/sched/cpupri.h index d6cba002006..935c7dc48e2 100644 --- a/kernel/sched/cpupri.h +++ b/kernel/sched/cpupri.h @@ -17,6 +17,7 @@ struct cpupri { int *cpu_to_pri; }; +#ifndef CONFIG_SCHED_MUQSS #ifdef CONFIG_SMP int cpupri_find(struct cpupri *cp, struct task_struct *p, struct cpumask *lowest_mask); @@ -27,3 +28,4 @@ void cpupri_set(struct cpupri *cp, int cpu, int pri); int cpupri_init(struct cpupri *cp); void cpupri_cleanup(struct cpupri *cp); #endif +#endif diff --git a/kernel/sched/cputime.c b/kernel/sched/cputime.c index 872e481d509..f8a996d5932 100644 --- a/kernel/sched/cputime.c +++ b/kernel/sched/cputime.c @@ -267,26 +267,6 @@ static inline u64 account_other_time(u64 max) return accounted; } -#ifdef CONFIG_64BIT -static inline u64 read_sum_exec_runtime(struct task_struct *t) -{ - return t->se.sum_exec_runtime; -} -#else -static u64 read_sum_exec_runtime(struct task_struct *t) -{ - u64 ns; - struct rq_flags rf; - struct rq *rq; - - rq = task_rq_lock(t, &rf); - ns = t->se.sum_exec_runtime; - task_rq_unlock(rq, t, &rf); - - return ns; -} -#endif - /* * Accumulate raw cputime values of dead tasks (sig->[us]time) and live * tasks (sum on group iteration) belonging to @tsk's group. @@ -612,7 +592,7 @@ void cputime_adjust(struct task_cputime *curr, struct prev_cputime *prev, void task_cputime_adjusted(struct task_struct *p, u64 *ut, u64 *st) { struct task_cputime cputime = { - .sum_exec_runtime = p->se.sum_exec_runtime, + .sum_exec_runtime = tsk_seruntime(p), }; task_cputime(p, &cputime.utime, &cputime.stime); diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c index f60ef0b4ec3..4096a1dc4e5 100644 --- a/kernel/sched/fair.c +++ b/kernel/sched/fair.c @@ -35,8 +35,13 @@ * * (default: 6ms * (1 + ilog(ncpus)), units: nanoseconds) */ +#ifdef CONFIG_PCK_INTERACTIVE +unsigned int sysctl_sched_latency = 4000000ULL; +static unsigned int normalized_sysctl_sched_latency = 4000000ULL; +#else unsigned int sysctl_sched_latency = 6000000ULL; static unsigned int normalized_sysctl_sched_latency = 6000000ULL; +#endif /* * The initial- and re-scaling of tunables is configurable @@ -56,13 +61,22 @@ unsigned int sysctl_sched_tunable_scaling = SCHED_TUNABLESCALING_LOG; * * (default: 0.75 msec * (1 + ilog(ncpus)), units: nanoseconds) */ +#ifdef CONFIG_PCK_INTERACTIVE +unsigned int sysctl_sched_min_granularity = 400000ULL; +static unsigned int normalized_sysctl_sched_min_granularity = 400000ULL; +#else unsigned int sysctl_sched_min_granularity = 750000ULL; static unsigned int normalized_sysctl_sched_min_granularity = 750000ULL; +#endif /* * This value is kept at sysctl_sched_latency/sysctl_sched_min_granularity */ +#ifdef CONFIG_PCK_INTERACTIVE +static unsigned int sched_nr_latency = 10; +#else static unsigned int sched_nr_latency = 8; +#endif /* * After fork, child runs first. If set to 0 (default) then @@ -79,10 +93,17 @@ unsigned int sysctl_sched_child_runs_first __read_mostly; * * (default: 1 msec * (1 + ilog(ncpus)), units: nanoseconds) */ +#ifdef CONFIG_PCK_INTERACTIVE +unsigned int sysctl_sched_wakeup_granularity = 500000UL; +static unsigned int normalized_sysctl_sched_wakeup_granularity = 500000UL; + +const_debug unsigned int sysctl_sched_migration_cost = 250000UL; +#else unsigned int sysctl_sched_wakeup_granularity = 1000000UL; static unsigned int normalized_sysctl_sched_wakeup_granularity = 1000000UL; const_debug unsigned int sysctl_sched_migration_cost = 500000UL; +#endif int sched_thermal_decay_shift; static int __init setup_sched_thermal_decay_shift(char *str) @@ -133,8 +154,12 @@ int __weak arch_asym_cpu_priority(int cpu) * * (default: 5 msec, units: microseconds) */ +#ifdef CONFIG_PCK_INTERACTIVE +unsigned int sysctl_sched_cfs_bandwidth_slice = 3000UL; +#else unsigned int sysctl_sched_cfs_bandwidth_slice = 5000UL; #endif +#endif static inline void update_load_add(struct load_weight *lw, unsigned long inc) { diff --git a/kernel/sched/idle.c b/kernel/sched/idle.c index 7ca3d3d86c2..46e79fb4bc4 100644 --- a/kernel/sched/idle.c +++ b/kernel/sched/idle.c @@ -265,7 +265,9 @@ static void do_idle(void) /* * Check if we need to update blocked load */ +#ifndef CONFIG_SCHED_MUQSS nohz_run_idle_balance(cpu); +#endif /* * If the arch has a polling bit, we maintain an invariant: @@ -403,6 +405,7 @@ void cpu_startup_entry(enum cpuhp_state state) do_idle(); } +#ifndef CONFIG_SCHED_MUQSS /* * idle-task scheduling class. */ @@ -516,3 +519,4 @@ DEFINE_SCHED_CLASS(idle) = { .switched_to = switched_to_idle, .update_curr = update_curr_idle, }; +#endif /* CONFIG_SCHED_MUQSS */ diff --git a/kernel/sched/sched.h b/kernel/sched/sched.h index 35f7efed75c..fe91fbeeb51 100644 --- a/kernel/sched/sched.h +++ b/kernel/sched/sched.h @@ -2,6 +2,19 @@ /* * Scheduler internal types and methods: */ +#ifdef CONFIG_SCHED_MUQSS +#include "MuQSS.h" + +/* Begin compatibility wrappers for MuQSS/CFS differences */ +#define rq_rt_nr_running(rq) ((rq)->rt_nr_running) +#define rq_h_nr_running(rq) ((rq)->nr_running) + +#else /* CONFIG_SCHED_MUQSS */ + +#define rq_rt_nr_running(rq) ((rq)->rt.rt_nr_running) +#define rq_h_nr_running(rq) ((rq)->cfs.h_nr_running) + + #include #include @@ -990,6 +1003,7 @@ struct rq { struct callback_head *balance_callback; + unsigned char nohz_idle_balance; unsigned char idle_balance; @@ -1561,6 +1575,7 @@ extern int group_balance_cpu(struct sched_group *sg); #ifdef CONFIG_SCHED_DEBUG void update_sched_domain_debugfs(void); void dirty_sched_domain_sysctl(int cpu); + #else static inline void update_sched_domain_debugfs(void) { @@ -1568,6 +1583,7 @@ static inline void update_sched_domain_debugfs(void) static inline void dirty_sched_domain_sysctl(int cpu) { } + #endif extern int sched_update_scaling(void); @@ -2756,3 +2772,24 @@ extern int sched_dynamic_mode(const char *str); extern void sched_dynamic_update(int mode); #endif +/* MuQSS compatibility functions */ +#ifdef CONFIG_64BIT +static inline u64 read_sum_exec_runtime(struct task_struct *t) +{ + return t->se.sum_exec_runtime; +} +#else +static inline u64 read_sum_exec_runtime(struct task_struct *t) +{ + u64 ns; + struct rq_flags rf; + struct rq *rq; + + rq = task_rq_lock(t, &rf); + ns = t->se.sum_exec_runtime; + task_rq_unlock(rq, t, &rf); + + return ns; +} +#endif +#endif /* CONFIG_SCHED_MUQSS */ \ No newline at end of file diff --git a/kernel/sched/topology.c b/kernel/sched/topology.c index 55a0a243e87..210727531bd 100644 --- a/kernel/sched/topology.c +++ b/kernel/sched/topology.c @@ -467,7 +467,11 @@ void rq_attach_root(struct rq *rq, struct root_domain *rd) struct root_domain *old_rd = NULL; unsigned long flags; +#ifdef CONFIG_SCHED_MUQSS + raw_spin_lock_irqsave(rq->lock, flags); +#else raw_spin_lock_irqsave(&rq->lock, flags); +#endif if (rq->rd) { old_rd = rq->rd; @@ -493,7 +497,11 @@ void rq_attach_root(struct rq *rq, struct root_domain *rd) if (cpumask_test_cpu(rq->cpu, cpu_active_mask)) set_rq_online(rq); +#ifdef CONFIG_SCHED_MUQSS + raw_spin_unlock_irqrestore(rq->lock, flags); +#else raw_spin_unlock_irqrestore(&rq->lock, flags); +#endif if (old_rd) call_rcu(&old_rd->rcu, free_rootdomain); diff --git a/kernel/skip_list.c b/kernel/skip_list.c new file mode 100644 index 00000000000..bf5c6e97e13 --- /dev/null +++ b/kernel/skip_list.c @@ -0,0 +1,148 @@ +/* + Copyright (C) 2011,2016 Con Kolivas. + + Code based on example originally by William Pugh. + +Skip Lists are a probabilistic alternative to balanced trees, as +described in the June 1990 issue of CACM and were invented by +William Pugh in 1987. + +A couple of comments about this implementation: +The routine randomLevel has been hard-coded to generate random +levels using p=0.25. It can be easily changed. + +The insertion routine has been implemented so as to use the +dirty hack described in the CACM paper: if a random level is +generated that is more than the current maximum level, the +current maximum level plus one is used instead. + +Levels start at zero and go up to MaxLevel (which is equal to +MaxNumberOfLevels-1). + +The routines defined in this file are: + +init: defines slnode + +new_skiplist: returns a new, empty list + +randomLevel: Returns a random level based on a u64 random seed passed to it. +In MuQSS, the "niffy" time is used for this purpose. + +insert(l,key, value): inserts the binding (key, value) into l. This operation +occurs in O(log n) time. + +delnode(slnode, l, node): deletes any binding of key from the l based on the +actual node value. This operation occurs in O(k) time where k is the +number of levels of the node in question (max 8). The original delete +function occurred in O(log n) time and involved a search. + +MuQSS Notes: In this implementation of skiplists, there are bidirectional +next/prev pointers and the insert function returns a pointer to the actual +node the value is stored. The key here is chosen by the scheduler so as to +sort tasks according to the priority list requirements and is no longer used +by the scheduler after insertion. The scheduler lookup, however, occurs in +O(1) time because it is always the first item in the level 0 linked list. +Since the task struct stores a copy of the node pointer upon skiplist_insert, +it can also remove it much faster than the original implementation with the +aid of prev<->next pointer manipulation and no searching. + +*/ + +#include +#include + +#define MaxNumberOfLevels 8 +#define MaxLevel (MaxNumberOfLevels - 1) + +void skiplist_init(skiplist_node *slnode) +{ + int i; + + slnode->key = 0xFFFFFFFFFFFFFFFF; + slnode->level = 0; + slnode->value = NULL; + for (i = 0; i < MaxNumberOfLevels; i++) + slnode->next[i] = slnode->prev[i] = slnode; +} + +skiplist *new_skiplist(skiplist_node *slnode) +{ + skiplist *l = kzalloc(sizeof(skiplist), GFP_ATOMIC); + + BUG_ON(!l); + l->header = slnode; + return l; +} + +void free_skiplist(skiplist *l) +{ + skiplist_node *p, *q; + + p = l->header; + do { + q = p->next[0]; + p->next[0]->prev[0] = q->prev[0]; + skiplist_node_init(p); + p = q; + } while (p != l->header); + kfree(l); +} + +void skiplist_node_init(skiplist_node *node) +{ + memset(node, 0, sizeof(skiplist_node)); +} + +static inline unsigned int randomLevel(const long unsigned int randseed) +{ + return find_first_bit(&randseed, MaxLevel) / 2; +} + +void skiplist_insert(skiplist *l, skiplist_node *node, keyType key, valueType value, unsigned int randseed) +{ + skiplist_node *update[MaxNumberOfLevels]; + skiplist_node *p, *q; + int k = l->level; + + p = l->header; + do { + while (q = p->next[k], q->key <= key) + p = q; + update[k] = p; + } while (--k >= 0); + + ++l->entries; + k = randomLevel(randseed); + if (k > l->level) { + k = ++l->level; + update[k] = l->header; + } + + node->level = k; + node->key = key; + node->value = value; + do { + p = update[k]; + node->next[k] = p->next[k]; + p->next[k] = node; + node->prev[k] = p; + node->next[k]->prev[k] = node; + } while (--k >= 0); +} + +void skiplist_delete(skiplist *l, skiplist_node *node) +{ + int k, m = node->level; + + for (k = 0; k <= m; k++) { + node->prev[k]->next[k] = node->next[k]; + node->next[k]->prev[k] = node->prev[k]; + } + skiplist_node_init(node); + if (m == l->level) { + while (l->header->next[m] == l->header && l->header->prev[m] == l->header && m > 0) + m--; + l->level = m; + } + l->entries--; +} diff --git a/kernel/sys_ni.c b/kernel/sys_ni.c index 0ea8128468c..af0b1ef09d9 100644 --- a/kernel/sys_ni.c +++ b/kernel/sys_ni.c @@ -151,6 +151,12 @@ COND_SYSCALL_COMPAT(set_robust_list); COND_SYSCALL(get_robust_list); COND_SYSCALL_COMPAT(get_robust_list); +/* kernel/futex2.c */ +COND_SYSCALL(futex_wait); +COND_SYSCALL(futex_wake); +COND_SYSCALL(futex_waitv); +COND_SYSCALL(futex_requeue); + /* kernel/hrtimer.c */ /* kernel/itimer.c */ diff --git a/kernel/sysctl.c b/kernel/sysctl.c index d4a78e08f6d..c17725b75fa 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -103,6 +103,9 @@ #ifdef CONFIG_LOCKUP_DETECTOR #include #endif +#ifdef CONFIG_USER_NS +#include +#endif #if defined(CONFIG_SYSCTL) @@ -120,7 +123,17 @@ static unsigned long long_max = LONG_MAX; static int one_hundred = 100; static int two_hundred = 200; static int one_thousand = 1000; -#ifdef CONFIG_PRINTK +#ifdef CONFIG_SCHED_MUQSS +static int zero = 0; +static int one = 1; +extern int rr_interval; +extern int sched_interactive; +extern int sched_iso_cpu; +extern int sched_yield_type; +extern int hrtimer_granularity_us; +extern int hrtimeout_min_us; +#endif +#if defined(CONFIG_PRINTK) || defined(CONFIG_SCHED_MUQSS) static int ten_thousand = 10000; #endif #ifdef CONFIG_PERF_EVENTS @@ -187,6 +200,10 @@ static enum sysctl_writes_mode sysctl_writes_strict = SYSCTL_WRITES_STRICT; int sysctl_legacy_va_layout; #endif + + + + #ifdef CONFIG_COMPACTION static int min_extfrag_threshold; static int max_extfrag_threshold = 1000; @@ -1729,6 +1746,7 @@ int proc_do_static_key(struct ctl_table *table, int write, } static struct ctl_table kern_table[] = { +#ifndef CONFIG_SCHED_MUQSS { .procname = "sched_child_runs_first", .data = &sysctl_sched_child_runs_first, @@ -1736,6 +1754,7 @@ static struct ctl_table kern_table[] = { .mode = 0644, .proc_handler = proc_dointvec, }, + #ifdef CONFIG_SCHEDSTATS { .procname = "sched_schedstats", @@ -1747,8 +1766,10 @@ static struct ctl_table kern_table[] = { .extra2 = SYSCTL_ONE, }, #endif /* CONFIG_SCHEDSTATS */ + #ifdef CONFIG_NUMA_BALANCING { + .procname = "numa_balancing", .data = NULL, /* filled in by handler */ .maxlen = sizeof(unsigned int), @@ -1758,6 +1779,7 @@ static struct ctl_table kern_table[] = { .extra2 = SYSCTL_ONE, }, #endif /* CONFIG_NUMA_BALANCING */ + { .procname = "sched_rt_period_us", .data = &sysctl_sched_rt_period, @@ -1837,6 +1859,73 @@ static struct ctl_table kern_table[] = { .extra1 = SYSCTL_ONE, }, #endif +#elif defined(CONFIG_SCHED_MUQSS) + { + .procname = "rr_interval", + .data = &rr_interval, + .maxlen = sizeof (int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &one_thousand, + }, + { + .procname = "interactive", + .data = &sched_interactive, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &zero, + .extra2 = &one, + }, + { + .procname = "iso_cpu", + .data = &sched_iso_cpu, + .maxlen = sizeof (int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &zero, + .extra2 = &one_hundred, + }, + { + .procname = "yield_type", + .data = &sched_yield_type, + .maxlen = sizeof (int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &zero, + .extra2 = &two, + }, +#if defined(CONFIG_SMP) && defined(CONFIG_SCHEDSTATS) + { + .procname = "sched_schedstats", + .data = NULL, + .maxlen = sizeof(unsigned int), + .mode = 0644, + .proc_handler = sysctl_schedstats, + .extra1 = SYSCTL_ZERO, + .extra2 = SYSCTL_ONE, + }, +#endif /* CONFIG_SMP && CONFIG_SCHEDSTATS */ + { + .procname = "hrtimer_granularity_us", + .data = &hrtimer_granularity_us, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &ten_thousand, + }, + { + .procname = "hrtimeout_min_us", + .data = &hrtimeout_min_us, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec_minmax, + .extra1 = &one, + .extra2 = &ten_thousand, + }, +#endif /* CONFIG_SCHED_MUQSS */ #if defined(CONFIG_ENERGY_MODEL) && defined(CONFIG_CPU_FREQ_GOV_SCHEDUTIL) { .procname = "sched_energy_aware", @@ -1896,6 +1985,15 @@ static struct ctl_table kern_table[] = { .proc_handler = proc_dointvec, }, #endif +#ifdef CONFIG_USER_NS + { + .procname = "unprivileged_userns_clone", + .data = &unprivileged_userns_clone, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = proc_dointvec, + }, +#endif #ifdef CONFIG_PROC_SYSCTL { .procname = "tainted", @@ -2623,6 +2721,7 @@ static struct ctl_table kern_table[] = { .proc_handler = bpf_unpriv_handler, .extra1 = SYSCTL_ZERO, .extra2 = &two, + }, { .procname = "bpf_stats_enabled", @@ -3086,6 +3185,20 @@ static struct ctl_table vm_table[] = { .extra2 = SYSCTL_ONE, }, #endif + { + .procname = "clean_low_kbytes", + .data = &sysctl_clean_low_kbytes, + .maxlen = sizeof(sysctl_clean_low_kbytes), + .mode = 0644, + .proc_handler = proc_doulongvec_minmax, + }, + { + .procname = "clean_min_kbytes", + .data = &sysctl_clean_min_kbytes, + .maxlen = sizeof(sysctl_clean_min_kbytes), + .mode = 0644, + .proc_handler = proc_doulongvec_minmax, + }, { .procname = "user_reserve_kbytes", .data = &sysctl_user_reserve_kbytes, @@ -3431,4 +3544,4 @@ EXPORT_SYMBOL(proc_dointvec_ms_jiffies); EXPORT_SYMBOL(proc_dostring); EXPORT_SYMBOL(proc_doulongvec_minmax); EXPORT_SYMBOL(proc_doulongvec_ms_jiffies_minmax); -EXPORT_SYMBOL(proc_do_large_bitmap); +EXPORT_SYMBOL(proc_do_large_bitmap); \ No newline at end of file diff --git a/kernel/task_work.c b/kernel/task_work.c index 1698fbe6f0e..081b05acadf 100644 --- a/kernel/task_work.c +++ b/kernel/task_work.c @@ -167,3 +167,4 @@ void task_work_run(void) } while (work); } } +EXPORT_SYMBOL_GPL(task_work_run); diff --git a/kernel/time/Kconfig b/kernel/time/Kconfig index 83e158d016b..99235d57431 100644 --- a/kernel/time/Kconfig +++ b/kernel/time/Kconfig @@ -132,7 +132,7 @@ config CONTEXT_TRACKING config CONTEXT_TRACKING_FORCE bool "Force context tracking" - depends on CONTEXT_TRACKING + depends on CONTEXT_TRACKING && !SCHED_MUQSS default y if !NO_HZ_FULL help The major pre-requirement for full dynticks to work is to diff --git a/kernel/time/clockevents.c b/kernel/time/clockevents.c index f5490222e13..3b3bf431f14 100644 --- a/kernel/time/clockevents.c +++ b/kernel/time/clockevents.c @@ -190,8 +190,14 @@ int clockevents_tick_resume(struct clock_event_device *dev) #ifdef CONFIG_GENERIC_CLOCKEVENTS_MIN_ADJUST +#ifdef CONFIG_SCHED_MUQSS +int __read_mostly hrtimer_granularity_us = 100; +/* Limit min_delta to 100us */ +#define MIN_DELTA_LIMIT (hrtimer_granularity_us * NSEC_PER_USEC) +#else /* Limit min_delta to a jiffie */ #define MIN_DELTA_LIMIT (NSEC_PER_SEC / HZ) +#endif /** * clockevents_increase_min_delta - raise minimum delta of a clock event device diff --git a/kernel/time/hrtimer.c b/kernel/time/hrtimer.c index 4a66725b1d4..2045eb35e3d 100644 --- a/kernel/time/hrtimer.c +++ b/kernel/time/hrtimer.c @@ -2236,3 +2236,115 @@ int __sched schedule_hrtimeout(ktime_t *expires, return schedule_hrtimeout_range(expires, 0, mode); } EXPORT_SYMBOL_GPL(schedule_hrtimeout); + +#ifdef CONFIG_SCHED_MUQSS +/* + * As per schedule_hrtimeout but taskes a millisecond value and returns how + * many milliseconds are left. + */ +long __sched schedule_msec_hrtimeout(long timeout) +{ + struct hrtimer_sleeper t; + int delta, jiffs; + ktime_t expires; + + if (!timeout) { + __set_current_state(TASK_RUNNING); + return 0; + } + + jiffs = msecs_to_jiffies(timeout); + /* + * If regular timer resolution is adequate or hrtimer resolution is not + * (yet) better than Hz, as would occur during startup, use regular + * timers. + */ + if (jiffs > 4 || hrtimer_resolution >= NSEC_PER_SEC / HZ || pm_freezing) + return schedule_timeout(jiffs); + + delta = (timeout % 1000) * NSEC_PER_MSEC; + expires = ktime_set(0, delta); + + hrtimer_init_sleeper_on_stack(&t, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hrtimer_set_expires_range_ns(&t.timer, expires, delta); + + hrtimer_sleeper_start_expires(&t, HRTIMER_MODE_REL); + + if (likely(t.task)) + schedule(); + + hrtimer_cancel(&t.timer); + destroy_hrtimer_on_stack(&t.timer); + + __set_current_state(TASK_RUNNING); + + expires = hrtimer_expires_remaining(&t.timer); + timeout = ktime_to_ms(expires); + return timeout < 0 ? 0 : timeout; +} + +EXPORT_SYMBOL(schedule_msec_hrtimeout); + +#define USECS_PER_SEC 1000000 +extern int hrtimer_granularity_us; + +static inline long schedule_usec_hrtimeout(long timeout) +{ + struct hrtimer_sleeper t; + ktime_t expires; + int delta; + + if (!timeout) { + __set_current_state(TASK_RUNNING); + return 0; + } + + if (hrtimer_resolution >= NSEC_PER_SEC / HZ) + return schedule_timeout(usecs_to_jiffies(timeout)); + + if (timeout < hrtimer_granularity_us) + timeout = hrtimer_granularity_us; + delta = (timeout % USECS_PER_SEC) * NSEC_PER_USEC; + expires = ktime_set(0, delta); + + hrtimer_init_sleeper_on_stack(&t, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + hrtimer_set_expires_range_ns(&t.timer, expires, delta); + + hrtimer_sleeper_start_expires(&t, HRTIMER_MODE_REL); + + if (likely(t.task)) + schedule(); + + hrtimer_cancel(&t.timer); + destroy_hrtimer_on_stack(&t.timer); + + __set_current_state(TASK_RUNNING); + + expires = hrtimer_expires_remaining(&t.timer); + timeout = ktime_to_us(expires); + return timeout < 0 ? 0 : timeout; +} + +int __read_mostly hrtimeout_min_us = 500; + +long __sched schedule_min_hrtimeout(void) +{ + return usecs_to_jiffies(schedule_usec_hrtimeout(hrtimeout_min_us)); +} + +EXPORT_SYMBOL(schedule_min_hrtimeout); + +long __sched schedule_msec_hrtimeout_interruptible(long timeout) +{ + __set_current_state(TASK_INTERRUPTIBLE); + return schedule_msec_hrtimeout(timeout); +} +EXPORT_SYMBOL(schedule_msec_hrtimeout_interruptible); + +long __sched schedule_msec_hrtimeout_uninterruptible(long timeout) +{ + __set_current_state(TASK_UNINTERRUPTIBLE); + return schedule_msec_hrtimeout(timeout); +} +EXPORT_SYMBOL(schedule_msec_hrtimeout_uninterruptible); +#endif /* CONFIG_SCHED_MUQSS */ diff --git a/kernel/time/posix-cpu-timers.c b/kernel/time/posix-cpu-timers.c index aa52fc85dbc..bfdf0bc62bd 100644 --- a/kernel/time/posix-cpu-timers.c +++ b/kernel/time/posix-cpu-timers.c @@ -216,7 +216,7 @@ static void task_sample_cputime(struct task_struct *p, u64 *samples) u64 stime, utime; task_cputime(p, &utime, &stime); - store_samples(samples, stime, utime, p->se.sum_exec_runtime); + store_samples(samples, stime, utime, tsk_seruntime(p)); } static void proc_sample_cputime_atomic(struct task_cputime_atomic *at, @@ -850,7 +850,7 @@ static void check_thread_timers(struct task_struct *tsk, soft = task_rlimit(tsk, RLIMIT_RTTIME); if (soft != RLIM_INFINITY) { /* Task RT timeout is accounted in jiffies. RTTIME is usec */ - unsigned long rttime = tsk->rt.timeout * (USEC_PER_SEC / HZ); + unsigned long rttime = tsk_rttimeout(tsk) * (USEC_PER_SEC / HZ); unsigned long hard = task_rlimit_max(tsk, RLIMIT_RTTIME); /* At the hard limit, send SIGKILL. No further action. */ diff --git a/kernel/time/timer.c b/kernel/time/timer.c index 99b97ccefdb..c2ce4af71b0 100644 --- a/kernel/time/timer.c +++ b/kernel/time/timer.c @@ -44,6 +44,9 @@ #include #include #include +#ifdef CONFIG_SCHED_MUQSS +#include +#endif #include #include @@ -1608,7 +1611,7 @@ static unsigned long __next_timer_interrupt(struct timer_base *base) * Check, if the next hrtimer event is before the next timer wheel * event: */ -static u64 cmp_next_hrtimer_event(u64 basem, u64 expires) +static u64 cmp_next_hrtimer_event(struct timer_base *base, u64 basem, u64 expires) { u64 nextevt = hrtimer_get_next_event(); @@ -1626,6 +1629,11 @@ static u64 cmp_next_hrtimer_event(u64 basem, u64 expires) if (nextevt <= basem) return basem; +#ifdef CONFIG_SCHED_MUQSS + if (nextevt < expires && nextevt - basem <= TICK_NSEC) + base->is_idle = false; +#endif + /* * Round up to the next jiffie. High resolution timers are * off, so the hrtimers are expired in the tick and we need to @@ -1693,7 +1701,7 @@ u64 get_next_timer_interrupt(unsigned long basej, u64 basem) } raw_spin_unlock(&base->lock); - return cmp_next_hrtimer_event(basem, expires); + return cmp_next_hrtimer_event(base, basem, expires); } /** @@ -1887,6 +1895,18 @@ signed long __sched schedule_timeout(signed long timeout) expire = timeout + jiffies; +#if defined(CONFIG_HIGH_RES_TIMERS) && defined(CONFIG_SCHED_MUQSS) + if (timeout == 1 && hrtimer_resolution < NSEC_PER_SEC / HZ) { + /* + * Special case 1 as being a request for the minimum timeout + * and use highres timers to timeout after 1ms to workaround + * the granularity of low Hz tick timers. + */ + if (!schedule_min_hrtimeout()) + return 0; + goto out_timeout; + } +#endif timer.task = current; timer_setup_on_stack(&timer.timer, process_timeout, 0); __mod_timer(&timer.timer, expire, MOD_TIMER_NOTPENDING); @@ -1896,6 +1916,9 @@ signed long __sched schedule_timeout(signed long timeout) /* Remove the timer from the object tracker */ destroy_timer_on_stack(&timer.timer); +#if defined(CONFIG_HIGH_RES_TIMERS) && defined(CONFIG_SCHED_MUQSS) +out_timeout: +#endif timeout = expire - jiffies; out: @@ -2044,6 +2067,18 @@ void msleep(unsigned int msecs) { unsigned long timeout = msecs_to_jiffies(msecs) + 1; +#ifdef CONFIG_SCHED_MUQSS + /* + * Use high resolution timers where the resolution of tick based + * timers is inadequate. + */ + if (timeout < 6 && hrtimer_resolution < NSEC_PER_SEC / HZ && !pm_freezing) { + while (msecs) + msecs = schedule_msec_hrtimeout_uninterruptible(msecs); + return; + } +#endif + while (timeout) timeout = schedule_timeout_uninterruptible(timeout); } @@ -2058,6 +2093,14 @@ unsigned long msleep_interruptible(unsigned int msecs) { unsigned long timeout = msecs_to_jiffies(msecs) + 1; +#ifdef CONFIG_SCHED_MUQSS + if (timeout < 6 && hrtimer_resolution < NSEC_PER_SEC / HZ && !pm_freezing) { + while (msecs && !signal_pending(current)) + msecs = schedule_msec_hrtimeout_interruptible(msecs); + return msecs; + } +#endif + while (timeout && !signal_pending(current)) timeout = schedule_timeout_interruptible(timeout); return jiffies_to_msecs(timeout); diff --git a/kernel/trace/trace_selftest.c b/kernel/trace/trace_selftest.c index adf7ef19400..5c53163120b 100644 --- a/kernel/trace/trace_selftest.c +++ b/kernel/trace/trace_selftest.c @@ -1052,10 +1052,15 @@ static int trace_wakeup_test_thread(void *data) { /* Make this a -deadline thread */ static const struct sched_attr attr = { +#ifdef CONFIG_SCHED_MUQSS + /* No deadline on MuQSS, use RR */ + .sched_policy = SCHED_RR, +#else .sched_policy = SCHED_DEADLINE, .sched_runtime = 100000ULL, .sched_deadline = 10000000ULL, .sched_period = 10000000ULL +#endif }; struct wakeup_test_data *x = data; diff --git a/kernel/user_namespace.c b/kernel/user_namespace.c index 27670ab7a4e..300c72d2a38 100644 --- a/kernel/user_namespace.c +++ b/kernel/user_namespace.c @@ -21,6 +21,13 @@ #include #include +/* sysctl */ +#ifdef CONFIG_USER_NS_UNPRIVILEGED +int unprivileged_userns_clone = 1; +#else +int unprivileged_userns_clone; +#endif + static struct kmem_cache *user_ns_cachep __read_mostly; static DEFINE_MUTEX(userns_state_mutex); diff --git a/lib/Makefile b/lib/Makefile index 2cc359ec1fd..6bfed81d8a7 100644 --- a/lib/Makefile +++ b/lib/Makefile @@ -28,7 +28,7 @@ CFLAGS_string.o += -fno-stack-protector endif lib-y := ctype.o string.o vsprintf.o cmdline.o \ - rbtree.o radix-tree.o timerqueue.o xarray.o \ + rbtree.o radix-tree.o sradix-tree.o timerqueue.o xarray.o \ idr.o extable.o sha1.o irq_regs.o argv_split.o \ flex_proportions.o ratelimit.o show_mem.o \ is_single_threaded.o plist.o decompress.o kobject_uevent.o \ diff --git a/lib/sradix-tree.c b/lib/sradix-tree.c new file mode 100644 index 00000000000..ab21e6309b9 --- /dev/null +++ b/lib/sradix-tree.c @@ -0,0 +1,476 @@ +#include +#include +#include +#include +#include +#include +#include + +static inline int sradix_node_full(struct sradix_tree_root *root, struct sradix_tree_node *node) +{ + return node->fulls == root->stores_size || + (node->height == 1 && node->count == root->stores_size); +} + +/* + * Extend a sradix tree so it can store key @index. + */ +static int sradix_tree_extend(struct sradix_tree_root *root, unsigned long index) +{ + struct sradix_tree_node *node; + unsigned int height; + + if (unlikely(root->rnode == NULL)) { + if (!(node = root->alloc())) + return -ENOMEM; + + node->height = 1; + root->rnode = node; + root->height = 1; + } + + /* Figure out what the height should be. */ + height = root->height; + index >>= root->shift * height; + + while (index) { + index >>= root->shift; + height++; + } + + while (height > root->height) { + unsigned int newheight; + + if (!(node = root->alloc())) + return -ENOMEM; + + /* Increase the height. */ + node->stores[0] = root->rnode; + root->rnode->parent = node; + if (root->extend) + root->extend(node, root->rnode); + + newheight = root->height + 1; + node->height = newheight; + node->count = 1; + if (sradix_node_full(root, root->rnode)) + node->fulls = 1; + + root->rnode = node; + root->height = newheight; + } + + return 0; +} + +/* + * Search the next item from the current node, that is not NULL + * and can satify root->iter(). + */ +void *sradix_tree_next(struct sradix_tree_root *root, + struct sradix_tree_node *node, unsigned long index, + int (*iter)(void *item, unsigned long height)) +{ + unsigned long offset; + void *item; + + if (unlikely(node == NULL)) { + node = root->rnode; + for (offset = 0; offset < root->stores_size; offset++) { + item = node->stores[offset]; + if (item && (!iter || iter(item, node->height))) + break; + } + + if (unlikely(offset >= root->stores_size)) + return NULL; + + if (node->height == 1) + return item; + else + goto go_down; + } + + while (node) { + offset = (index & root->mask) + 1; + for (; offset < root->stores_size; offset++) { + item = node->stores[offset]; + if (item && (!iter || iter(item, node->height))) + break; + } + + if (offset < root->stores_size) + break; + + node = node->parent; + index >>= root->shift; + } + + if (!node) + return NULL; + + while (node->height > 1) { +go_down: + node = item; + for (offset = 0; offset < root->stores_size; offset++) { + item = node->stores[offset]; + if (item && (!iter || iter(item, node->height))) + break; + } + + if (unlikely(offset >= root->stores_size)) + return NULL; + } + + BUG_ON(offset > root->stores_size); + + return item; +} + +/* + * Blindly insert the item to the tree. Typically, we reuse the + * first empty store item. + */ +int sradix_tree_enter(struct sradix_tree_root *root, void **item, int num) +{ + unsigned long index; + unsigned int height; + struct sradix_tree_node *node, *tmp = NULL; + int offset, offset_saved; + void **store = NULL; + int error, i, j, shift; + +go_on: + index = root->min; + + if (root->enter_node && !sradix_node_full(root, root->enter_node)) { + node = root->enter_node; + BUG_ON((index >> (root->shift * root->height))); + } else { + node = root->rnode; + if (node == NULL || (index >> (root->shift * root->height)) + || sradix_node_full(root, node)) { + error = sradix_tree_extend(root, index); + if (error) + return error; + + node = root->rnode; + } + } + + + height = node->height; + shift = (height - 1) * root->shift; + offset = (index >> shift) & root->mask; + while (shift > 0) { + offset_saved = offset; + for (; offset < root->stores_size; offset++) { + store = &node->stores[offset]; + tmp = *store; + + if (!tmp || !sradix_node_full(root, tmp)) + break; + } + BUG_ON(offset >= root->stores_size); + + if (offset != offset_saved) { + index += (offset - offset_saved) << shift; + index &= ~((1UL << shift) - 1); + } + + if (!tmp) { + if (!(tmp = root->alloc())) + return -ENOMEM; + + tmp->height = shift / root->shift; + *store = tmp; + tmp->parent = node; + node->count++; +// if (root->extend) +// root->extend(node, tmp); + } + + node = tmp; + shift -= root->shift; + offset = (index >> shift) & root->mask; + } + + BUG_ON(node->height != 1); + + + store = &node->stores[offset]; + for (i = 0, j = 0; + j < root->stores_size - node->count && + i < root->stores_size - offset && j < num; i++) { + if (!store[i]) { + store[i] = item[j]; + if (root->assign) + root->assign(node, index + i, item[j]); + j++; + } + } + + node->count += j; + root->num += j; + num -= j; + + while (sradix_node_full(root, node)) { + node = node->parent; + if (!node) + break; + + node->fulls++; + } + + if (unlikely(!node)) { + /* All nodes are full */ + root->min = 1 << (root->height * root->shift); + root->enter_node = NULL; + } else { + root->min = index + i - 1; + root->min |= (1UL << (node->height - 1)) - 1; + root->min++; + root->enter_node = node; + } + + if (num) { + item += j; + goto go_on; + } + + return 0; +} + + +/** + * sradix_tree_shrink - shrink height of a sradix tree to minimal + * @root sradix tree root + * + */ +static inline void sradix_tree_shrink(struct sradix_tree_root *root) +{ + /* try to shrink tree height */ + while (root->height > 1) { + struct sradix_tree_node *to_free = root->rnode; + + /* + * The candidate node has more than one child, or its child + * is not at the leftmost store, we cannot shrink. + */ + if (to_free->count != 1 || !to_free->stores[0]) + break; + + root->rnode = to_free->stores[0]; + root->rnode->parent = NULL; + root->height--; + if (unlikely(root->enter_node == to_free)) + root->enter_node = NULL; + root->free(to_free); + } +} + +/* + * Del the item on the known leaf node and index + */ +void sradix_tree_delete_from_leaf(struct sradix_tree_root *root, + struct sradix_tree_node *node, unsigned long index) +{ + unsigned int offset; + struct sradix_tree_node *start, *end; + + BUG_ON(node->height != 1); + + start = node; + while (node && !(--node->count)) + node = node->parent; + + end = node; + if (!node) { + root->rnode = NULL; + root->height = 0; + root->min = 0; + root->num = 0; + root->enter_node = NULL; + } else { + offset = (index >> (root->shift * (node->height - 1))) & root->mask; + if (root->rm) + root->rm(node, offset); + node->stores[offset] = NULL; + root->num--; + if (root->min > index) { + root->min = index; + root->enter_node = node; + } + } + + if (start != end) { + do { + node = start; + start = start->parent; + if (unlikely(root->enter_node == node)) + root->enter_node = end; + root->free(node); + } while (start != end); + + /* + * Note that shrink may free "end", so enter_node still need to + * be checked inside. + */ + sradix_tree_shrink(root); + } else if (node->count == root->stores_size - 1) { + /* It WAS a full leaf node. Update the ancestors */ + node = node->parent; + while (node) { + node->fulls--; + if (node->fulls != root->stores_size - 1) + break; + + node = node->parent; + } + } +} + +void *sradix_tree_lookup(struct sradix_tree_root *root, unsigned long index) +{ + unsigned int height, offset; + struct sradix_tree_node *node; + int shift; + + node = root->rnode; + if (node == NULL || (index >> (root->shift * root->height))) + return NULL; + + height = root->height; + shift = (height - 1) * root->shift; + + do { + offset = (index >> shift) & root->mask; + node = node->stores[offset]; + if (!node) + return NULL; + + shift -= root->shift; + } while (shift >= 0); + + return node; +} + +/* + * Return the item if it exists, otherwise create it in place + * and return the created item. + */ +void *sradix_tree_lookup_create(struct sradix_tree_root *root, + unsigned long index, void *(*item_alloc)(void)) +{ + unsigned int height, offset; + struct sradix_tree_node *node, *tmp; + void *item; + int shift, error; + + if (root->rnode == NULL || (index >> (root->shift * root->height))) { + if (item_alloc) { + error = sradix_tree_extend(root, index); + if (error) + return NULL; + } else { + return NULL; + } + } + + node = root->rnode; + height = root->height; + shift = (height - 1) * root->shift; + + do { + offset = (index >> shift) & root->mask; + if (!node->stores[offset]) { + if (!(tmp = root->alloc())) + return NULL; + + tmp->height = shift / root->shift; + node->stores[offset] = tmp; + tmp->parent = node; + node->count++; + node = tmp; + } else { + node = node->stores[offset]; + } + + shift -= root->shift; + } while (shift > 0); + + BUG_ON(node->height != 1); + offset = index & root->mask; + if (node->stores[offset]) { + return node->stores[offset]; + } else if (item_alloc) { + if (!(item = item_alloc())) + return NULL; + + node->stores[offset] = item; + + /* + * NOTE: we do NOT call root->assign here, since this item is + * newly created by us having no meaning. Caller can call this + * if it's necessary to do so. + */ + + node->count++; + root->num++; + + while (sradix_node_full(root, node)) { + node = node->parent; + if (!node) + break; + + node->fulls++; + } + + if (unlikely(!node)) { + /* All nodes are full */ + root->min = 1 << (root->height * root->shift); + } else { + if (root->min == index) { + root->min |= (1UL << (node->height - 1)) - 1; + root->min++; + root->enter_node = node; + } + } + + return item; + } else { + return NULL; + } + +} + +int sradix_tree_delete(struct sradix_tree_root *root, unsigned long index) +{ + unsigned int height, offset; + struct sradix_tree_node *node; + int shift; + + node = root->rnode; + if (node == NULL || (index >> (root->shift * root->height))) + return -ENOENT; + + height = root->height; + shift = (height - 1) * root->shift; + + do { + offset = (index >> shift) & root->mask; + node = node->stores[offset]; + if (!node) + return -ENOENT; + + shift -= root->shift; + } while (shift > 0); + + offset = index & root->mask; + if (!node->stores[offset]) + return -ENOENT; + + sradix_tree_delete_from_leaf(root, node, index); + + return 0; +} diff --git a/mm/Kconfig b/mm/Kconfig index 02d44e3420f..b352b4068ef 100644 --- a/mm/Kconfig +++ b/mm/Kconfig @@ -121,6 +121,41 @@ config SPARSEMEM_VMEMMAP pfn_to_page and page_to_pfn operations. This is the most efficient option when sufficient kernel resources are available. +config CLEAN_LOW_KBYTES + int "Default value for vm.clean_low_kbytes" + depends on SYSCTL + default "262144" + help + The vm.clean_low_kbytes sysctl knob provides *best-effort* + protection of clean file pages. The clean file pages on the current + node won't be reclaimed under memory pressure when their amount is + below vm.clean_low_kbytes *unless* we threaten to OOM or have + no free swap space or vm.swappiness=0. + + Protection of clean file pages may be used to prevent thrashing and + reducing I/O under low-memory conditions. + + Setting it to a high value may result in a early eviction of anonymous + pages into the swap space by attempting to hold the protected amount of + clean file pages in memory. + +config CLEAN_MIN_KBYTES + int "Default value for vm.clean_min_kbytes" + depends on SYSCTL + default "0" + help + The vm.clean_min_kbytes sysctl knob provides *hard* protection + of clean file pages. The clean file pages on the current node won't be + reclaimed under memory pressure when their amount is below + vm.clean_min_kbytes. + + Hard protection of clean file pages may be used to avoid high latency and + prevent livelock in near-OOM conditions. + + Setting it to a high value may result in a early out-of-memory condition + due to the inability to reclaim the protected amount of clean file pages + when other types of pages cannot be reclaimed. + config HAVE_MEMBLOCK_PHYS_MAP bool @@ -333,6 +368,32 @@ config KSM See Documentation/vm/ksm.rst for more information: KSM is inactive until a program has madvised that an area is MADV_MERGEABLE, and root has set /sys/kernel/mm/ksm/run to 1 (if CONFIG_SYSFS is set). +choice + prompt "Choose UKSM/KSM strategy" + default UKSM + depends on KSM + help + This option allows to select a UKSM/KSM stragety. + +config UKSM + bool "Ultra-KSM for page merging" + depends on KSM + help + UKSM is inspired by the Linux kernel project \u2014 KSM(Kernel Same + page Merging), but with a fundamentally rewritten core algorithm. With + an advanced algorithm, UKSM now can transparently scans all anonymously + mapped user space applications with an significantly improved scan speed + and CPU efficiency. Since KVM is friendly to KSM, KVM can also benefit from + UKSM. Now UKSM has its first stable release and first real world enterprise user. + For more information, please goto its project page. + (github.com/dolohow/uksm) + +config KSM_LEGACY + bool "Legacy KSM implementation" + depends on KSM + help + The legacy KSM implementation from Red Hat. +endchoice config DEFAULT_MMAP_MIN_ADDR int "Low address space to protect from user allocation" @@ -901,4 +962,62 @@ config KMAP_LOCAL # struct io_mapping based helper. Selected by drivers that need them config IO_MAPPING bool + +# the multigenerational lru { +config LRU_GEN + bool "Multigenerational LRU" + depends on MMU + help + A high performance LRU implementation to heavily overcommit workloads + that are not IO bound. See Documentation/vm/multigen_lru.rst for + details. + + Warning: do not enable this option unless you plan to use it because + it introduces a small per-process and per-memcg and per-node memory + overhead. + +config LRU_GEN_ENABLED + bool "Turn on by default" + depends on LRU_GEN + help + The default value of /sys/kernel/mm/lru_gen/enabled is 0. This option + changes it to 1. + + Warning: the default value is the fast path. See + Documentation/static-keys.txt for details. + +config LRU_GEN_STATS + bool "Full stats for debugging" + depends on LRU_GEN + help + This option keeps full stats for each generation, which can be read + from /sys/kernel/debug/lru_gen_full. + + Warning: do not enable this option unless you plan to use it because + it introduces an additional small per-process and per-memcg and + per-node memory overhead. + +config NR_LRU_GENS + int "Max number of generations" + depends on LRU_GEN + range 4 31 + default 7 + help + This will use order_base_2(N+1) spare bits from page flags. + + Warning: do not use numbers larger than necessary because each + generation introduces a small per-node and per-memcg memory overhead. + +config TIERS_PER_GEN + int "Number of tiers per generation" + depends on LRU_GEN + range 2 5 + default 4 + help + This will use N-2 spare bits from page flags. + + Larger values generally offer better protection to active pages under + heavy buffered I/O workloads. +# } + endmenu diff --git a/mm/Makefile b/mm/Makefile index bf71e295e9f..14cda67496d 100644 --- a/mm/Makefile +++ b/mm/Makefile @@ -52,7 +52,7 @@ obj-y := filemap.o mempool.o oom_kill.o fadvise.o \ mm_init.o percpu.o slab_common.o \ compaction.o vmacache.o \ interval_tree.o list_lru.o workingset.o \ - debug.o gup.o mmap_lock.o $(mmu-y) + prfile.o debug.o gup.o mmap_lock.o $(mmu-y) # Give 'page_alloc' its own module-parameter namespace page-alloc-y := page_alloc.o @@ -80,7 +80,8 @@ obj-$(CONFIG_SPARSEMEM) += sparse.o obj-$(CONFIG_SPARSEMEM_VMEMMAP) += sparse-vmemmap.o obj-$(CONFIG_SLOB) += slob.o obj-$(CONFIG_MMU_NOTIFIER) += mmu_notifier.o -obj-$(CONFIG_KSM) += ksm.o +obj-$(CONFIG_KSM_LEGACY) += ksm.o +obj-$(CONFIG_UKSM) += uksm.o obj-$(CONFIG_PAGE_POISONING) += page_poison.o obj-$(CONFIG_SLAB) += slab.o obj-$(CONFIG_SLUB) += slub.o diff --git a/mm/filemap.c b/mm/filemap.c index 66f7e9fdfbc..f9a8ff48e69 100644 --- a/mm/filemap.c +++ b/mm/filemap.c @@ -3240,7 +3240,7 @@ vm_fault_t filemap_page_mkwrite(struct vm_fault *vmf) vm_fault_t ret = VM_FAULT_LOCKED; sb_start_pagefault(mapping->host->i_sb); - file_update_time(vmf->vma->vm_file); + vma_file_update_time(vmf->vma); lock_page(page); if (page->mapping != mapping) { unlock_page(page); diff --git a/mm/huge_memory.c b/mm/huge_memory.c index 9aaf4a8ebee..3dbe908dc38 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -54,7 +54,11 @@ unsigned long transparent_hugepage_flags __read_mostly = #ifdef CONFIG_TRANSPARENT_HUGEPAGE_MADVISE (1<vm_page_prot); entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); page_add_new_anon_rmap(page, vma, haddr, true); - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, true); pgtable_trans_huge_deposit(vma->vm_mm, vmf->pmd, pgtable); set_pmd_at(vma->vm_mm, haddr, vmf->pmd, entry); update_mmu_cache_pmd(vma, vmf->address, vmf->pmd); @@ -2427,7 +2431,8 @@ static void __split_huge_page_tail(struct page *head, int tail, #ifdef CONFIG_64BIT (1L << PG_arch_2) | #endif - (1L << PG_dirty))); + (1L << PG_dirty) | + LRU_GEN_MASK | LRU_USAGE_MASK)); /* ->mapping in first tail page is compound_mapcount */ VM_BUG_ON_PAGE(tail > 2 && page_tail->mapping != TAIL_MAPPING, diff --git a/mm/internal.h b/mm/internal.h index e8fdb531f88..3e68959cdbf 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -197,6 +197,7 @@ extern void prep_compound_page(struct page *page, unsigned int order); extern void post_alloc_hook(struct page *page, unsigned int order, gfp_t gfp_flags); extern int user_min_free_kbytes; +extern atomic_long_t kswapd_waiters; extern void free_unref_page(struct page *page); extern void free_unref_page_list(struct list_head *list); diff --git a/mm/khugepaged.c b/mm/khugepaged.c index d97b20fad6e..6354edc652b 100644 --- a/mm/khugepaged.c +++ b/mm/khugepaged.c @@ -1196,7 +1196,7 @@ static void collapse_huge_page(struct mm_struct *mm, spin_lock(pmd_ptl); BUG_ON(!pmd_none(*pmd)); page_add_new_anon_rmap(new_page, vma, address, true); - lru_cache_add_inactive_or_unevictable(new_page, vma); + lru_cache_add_page_vma(new_page, vma, true); pgtable_trans_huge_deposit(mm, pmd, pgtable); set_pmd_at(mm, address, pmd, _pmd); update_mmu_cache_pmd(vma, address, pmd); diff --git a/mm/ksm.c b/mm/ksm.c index 2f3aaeb34a4..c75372c9d02 100644 --- a/mm/ksm.c +++ b/mm/ksm.c @@ -856,17 +856,6 @@ static int unmerge_ksm_pages(struct vm_area_struct *vma, return err; } -static inline struct stable_node *page_stable_node(struct page *page) -{ - return PageKsm(page) ? page_rmapping(page) : NULL; -} - -static inline void set_page_stable_node(struct page *page, - struct stable_node *stable_node) -{ - page->mapping = (void *)((unsigned long)stable_node | PAGE_MAPPING_KSM); -} - #ifdef CONFIG_SYSFS /* * Only called through the sysfs control interface: diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 946d8481569..17f941495f4 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -4990,6 +4990,7 @@ static void __mem_cgroup_free(struct mem_cgroup *memcg) for_each_node(node) free_mem_cgroup_per_node_info(memcg, node); free_percpu(memcg->vmstats_percpu); + lru_gen_free_mm_list(memcg); kfree(memcg); } @@ -5039,6 +5040,9 @@ static struct mem_cgroup *mem_cgroup_alloc(void) if (alloc_mem_cgroup_per_node_info(memcg, node)) goto fail; + if (lru_gen_alloc_mm_list(memcg)) + goto fail; + if (memcg_wb_domain_init(memcg, GFP_KERNEL)) goto fail; @@ -6000,6 +6004,29 @@ static void mem_cgroup_move_task(void) } #endif +#ifdef CONFIG_LRU_GEN +static void mem_cgroup_attach(struct cgroup_taskset *tset) +{ + struct cgroup_subsys_state *css; + struct task_struct *task = NULL; + + cgroup_taskset_for_each_leader(task, css, tset) + break; + + if (!task) + return; + + task_lock(task); + if (task->mm && task->mm->owner == task) + lru_gen_migrate_mm(task->mm); + task_unlock(task); +} +#else +static void mem_cgroup_attach(struct cgroup_taskset *tset) +{ +} +#endif + static int seq_puts_memcg_tunable(struct seq_file *m, unsigned long value) { if (value == PAGE_COUNTER_MAX) @@ -6341,6 +6368,7 @@ struct cgroup_subsys memory_cgrp_subsys = { .css_reset = mem_cgroup_css_reset, .css_rstat_flush = mem_cgroup_css_rstat_flush, .can_attach = mem_cgroup_can_attach, + .attach = mem_cgroup_attach, .cancel_attach = mem_cgroup_cancel_attach, .post_attach = mem_cgroup_move_task, .dfl_cftypes = memory_files, diff --git a/mm/memory.c b/mm/memory.c index f2293c0df33..f1d9d56518c 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -158,6 +158,25 @@ EXPORT_SYMBOL(zero_pfn); unsigned long highest_memmap_pfn __read_mostly; +#ifdef CONFIG_UKSM +unsigned long uksm_zero_pfn __read_mostly; +EXPORT_SYMBOL_GPL(uksm_zero_pfn); +struct page *empty_uksm_zero_page; + +static int __init setup_uksm_zero_page(void) +{ + empty_uksm_zero_page = alloc_pages(__GFP_ZERO & ~__GFP_MOVABLE, 0); + if (!empty_uksm_zero_page) + panic("Oh boy, that early out of memory?"); + + SetPageReserved(empty_uksm_zero_page); + uksm_zero_pfn = page_to_pfn(empty_uksm_zero_page); + + return 0; +} +core_initcall(setup_uksm_zero_page); +#endif + /* * CONFIG_MMU architectures set up ZERO_PAGE in their paging_init() */ @@ -173,6 +192,7 @@ void mm_trace_rss_stat(struct mm_struct *mm, int member, long count) trace_rss_stat(mm, member, count); } + #if defined(SPLIT_RSS_COUNTING) void sync_mm_rss(struct mm_struct *mm) @@ -841,7 +861,7 @@ copy_present_page(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma copy_user_highpage(new_page, page, addr, src_vma); __SetPageUptodate(new_page); page_add_new_anon_rmap(new_page, dst_vma, addr, false); - lru_cache_add_inactive_or_unevictable(new_page, dst_vma); + lru_cache_add_page_vma(new_page, dst_vma, false); rss[mm_counter(new_page)]++; /* All done, just insert the new page copy in the child */ @@ -880,6 +900,11 @@ copy_present_pte(struct vm_area_struct *dst_vma, struct vm_area_struct *src_vma, get_page(page); page_dup_rmap(page, false); rss[mm_counter(page)]++; + + /* Should return NULL in vm_normal_page() */ + uksm_bugon_zeropage(pte); + } else { + uksm_map_zero_page(pte); } /* @@ -1255,8 +1280,10 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, ptent = ptep_get_and_clear_full(mm, addr, pte, tlb->fullmm); tlb_remove_tlb_entry(tlb, pte, addr); - if (unlikely(!page)) + if (unlikely(!page)) { + uksm_unmap_zero_page(ptent); continue; + } if (!PageAnon(page)) { if (pte_dirty(ptent)) { @@ -2658,6 +2685,7 @@ static inline bool cow_user_page(struct page *dst, struct page *src, if (likely(src)) { copy_user_highpage(dst, src, addr, vma); + uksm_cow_page(vma, src); return true; } @@ -2904,6 +2932,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) vmf->address); if (!new_page) goto oom; + uksm_cow_pte(vma, vmf->orig_pte); } else { new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, vmf->address); @@ -2946,7 +2975,9 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) mm_counter_file(old_page)); inc_mm_counter_fast(mm, MM_ANONPAGES); } + uksm_bugon_zeropage(vmf->orig_pte); } else { + uksm_unmap_zero_page(vmf->orig_pte); inc_mm_counter_fast(mm, MM_ANONPAGES); } flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte)); @@ -2963,7 +2994,7 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) */ ptep_clear_flush_notify(vma, vmf->address, vmf->pte); page_add_new_anon_rmap(new_page, vma, vmf->address, false); - lru_cache_add_inactive_or_unevictable(new_page, vma); + lru_cache_add_page_vma(new_page, vma, true); /* * We call the notify macro here because, when using secondary * mmu page tables (such as kvm shadow page tables), we want the @@ -3522,7 +3553,7 @@ vm_fault_t do_swap_page(struct vm_fault *vmf) /* ksm created a completely new copy */ if (unlikely(page != swapcache && swapcache)) { page_add_new_anon_rmap(page, vma, vmf->address, false); - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, true); } else { do_page_add_anon_rmap(page, vma, vmf->address, exclusive); } @@ -3669,7 +3700,7 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES); page_add_new_anon_rmap(page, vma, vmf->address, false); - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, true); setpte: set_pte_at(vma->vm_mm, vmf->address, vmf->pte, entry); @@ -3839,7 +3870,7 @@ void do_set_pte(struct vm_fault *vmf, struct page *page, unsigned long addr) if (write && !(vma->vm_flags & VM_SHARED)) { inc_mm_counter_fast(vma->vm_mm, MM_ANONPAGES); page_add_new_anon_rmap(page, vma, addr, false); - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, true); } else { inc_mm_counter_fast(vma->vm_mm, mm_counter_file(page)); page_add_file_rmap(page, false); diff --git a/mm/migrate.c b/mm/migrate.c index 047209d6602..2bd61823f11 100644 --- a/mm/migrate.c +++ b/mm/migrate.c @@ -2968,7 +2968,7 @@ static void migrate_vma_insert_page(struct migrate_vma *migrate, inc_mm_counter(mm, MM_ANONPAGES); page_add_new_anon_rmap(page, vma, addr, false); if (!is_zone_device_page(page)) - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, false); get_page(page); if (flush) { diff --git a/mm/mm_init.c b/mm/mm_init.c index 9ddaf0e1b0a..ef0deadb90a 100644 --- a/mm/mm_init.c +++ b/mm/mm_init.c @@ -65,14 +65,16 @@ void __init mminit_verify_pageflags_layout(void) shift = 8 * sizeof(unsigned long); width = shift - SECTIONS_WIDTH - NODES_WIDTH - ZONES_WIDTH - - LAST_CPUPID_SHIFT - KASAN_TAG_WIDTH; + - LAST_CPUPID_SHIFT - KASAN_TAG_WIDTH - LRU_GEN_WIDTH - LRU_USAGE_WIDTH; mminit_dprintk(MMINIT_TRACE, "pageflags_layout_widths", - "Section %d Node %d Zone %d Lastcpupid %d Kasantag %d Flags %d\n", + "Section %d Node %d Zone %d Lastcpupid %d Kasantag %d Gen %d Tier %d Flags %d\n", SECTIONS_WIDTH, NODES_WIDTH, ZONES_WIDTH, LAST_CPUPID_WIDTH, KASAN_TAG_WIDTH, + LRU_GEN_WIDTH, + LRU_USAGE_WIDTH, NR_PAGEFLAGS); mminit_dprintk(MMINIT_TRACE, "pageflags_layout_shifts", "Section %d Node %d Zone %d Lastcpupid %d Kasantag %d\n", diff --git a/mm/mmap.c b/mm/mmap.c index 0584e540246..cb70acbc1bd 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -46,6 +46,7 @@ #include #include #include +#include #include #include @@ -185,8 +186,9 @@ static struct vm_area_struct *remove_vma(struct vm_area_struct *vma) if (vma->vm_ops && vma->vm_ops->close) vma->vm_ops->close(vma); if (vma->vm_file) - fput(vma->vm_file); + vma_fput(vma); mpol_put(vma_policy(vma)); + uksm_remove_vma(vma); vm_area_free(vma); return next; } @@ -754,9 +756,16 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, long adjust_next = 0; int remove_next = 0; +/* + * to avoid deadlock, ksm_remove_vma must be done before any spin_lock is + * acquired + */ + uksm_remove_vma(vma); + if (next && !insert) { struct vm_area_struct *exporter = NULL, *importer = NULL; + uksm_remove_vma(next); if (end >= next->vm_end) { /* * vma expands, overlapping all the next, and @@ -887,6 +896,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, end_changed = true; } vma->vm_pgoff = pgoff; + if (adjust_next) { next->vm_start += adjust_next; next->vm_pgoff += adjust_next >> PAGE_SHIFT; @@ -955,7 +965,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, if (remove_next) { if (file) { uprobe_munmap(next, next->vm_start, next->vm_end); - fput(file); + vma_fput(vma); } if (next->anon_vma) anon_vma_merge(vma, next); @@ -991,6 +1001,7 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, if (remove_next == 2) { remove_next = 1; end = next->vm_end; + uksm_remove_vma(next); goto again; } else if (next) @@ -1017,10 +1028,14 @@ int __vma_adjust(struct vm_area_struct *vma, unsigned long start, */ VM_WARN_ON(mm->highest_vm_end != vm_end_gap(vma)); } + } else { + if (next && !insert) + uksm_vma_add_new(next); } if (insert && file) uprobe_mmap(insert); + uksm_vma_add_new(vma); validate_mm(mm); return 0; @@ -1476,6 +1491,9 @@ unsigned long do_mmap(struct file *file, unsigned long addr, vm_flags = calc_vm_prot_bits(prot, pkey) | calc_vm_flag_bits(flags) | mm->def_flags | VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC; + /* If uksm is enabled, we add VM_MERGEABLE to new VMAs. */ + uksm_vm_flags_mod(&vm_flags); + if (flags & MAP_LOCKED) if (!can_do_mlock()) return -EPERM; @@ -1871,6 +1889,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, allow_write_access(file); } file = vma->vm_file; + uksm_vma_add_new(vma); out: perf_event_mmap(vma); @@ -1901,7 +1920,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, return addr; unmap_and_free_vma: - fput(vma->vm_file); + vma_fput(vma); vma->vm_file = NULL; /* Undo any partial mapping done by a device driver. */ @@ -1913,6 +1932,7 @@ unsigned long mmap_region(struct file *file, unsigned long addr, if (vm_flags & VM_DENYWRITE) allow_write_access(file); free_vma: + uksm_remove_vma(vma); vm_area_free(vma); unacct_error: if (charged) @@ -2761,7 +2781,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma, goto out_free_mpol; if (new->vm_file) - get_file(new->vm_file); + vma_get_file(new); if (new->vm_ops && new->vm_ops->open) new->vm_ops->open(new); @@ -2772,6 +2792,8 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma, else err = vma_adjust(vma, vma->vm_start, addr, vma->vm_pgoff, new); + uksm_vma_add_new(new); + /* Success. */ if (!err) return 0; @@ -2780,7 +2802,7 @@ int __split_vma(struct mm_struct *mm, struct vm_area_struct *vma, if (new->vm_ops && new->vm_ops->close) new->vm_ops->close(new); if (new->vm_file) - fput(new->vm_file); + vma_fput(new); unlink_anon_vmas(new); out_free_mpol: mpol_put(vma_policy(new)); @@ -2973,7 +2995,7 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size, struct vm_area_struct *vma; unsigned long populate = 0; unsigned long ret = -EINVAL; - struct file *file; + struct file *file, *prfile; pr_warn_once("%s (%d) uses deprecated remap_file_pages() syscall. See Documentation/vm/remap_file_pages.rst.\n", current->comm, current->pid); @@ -3032,10 +3054,27 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size, if (vma->vm_flags & VM_LOCKED) flags |= MAP_LOCKED; - file = get_file(vma->vm_file); + vma_get_file(vma); + file = vma->vm_file; + prfile = vma->vm_prfile; ret = do_mmap(vma->vm_file, start, size, prot, flags, pgoff, &populate, NULL); + if (!IS_ERR_VALUE(ret) && file && prfile) { + struct vm_area_struct *new_vma; + + new_vma = find_vma(mm, ret); + if (!new_vma->vm_prfile) + new_vma->vm_prfile = prfile; + if (new_vma != vma) + get_file(prfile); + } + /* + * two fput()s instead of vma_fput(vma), + * coz vma may not be available anymore. + */ fput(file); + if (prfile) + fput(prfile); out: mmap_write_unlock(mm); if (populate) @@ -3063,6 +3102,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla if ((flags & (~VM_EXEC)) != 0) return -EINVAL; flags |= VM_DATA_DEFAULT_FLAGS | VM_ACCOUNT | mm->def_flags; + uksm_vm_flags_mod(&flags); mapped_addr = get_unmapped_area(NULL, addr, len, 0, MAP_FIXED); if (IS_ERR_VALUE(mapped_addr)) @@ -3108,6 +3148,7 @@ static int do_brk_flags(unsigned long addr, unsigned long len, unsigned long fla vma->vm_flags = flags; vma->vm_page_prot = vm_get_page_prot(flags); vma_link(mm, vma, prev, rb_link, rb_parent); + uksm_vma_add_new(vma); out: perf_event_mmap(vma); mm->total_vm += len >> PAGE_SHIFT; @@ -3185,6 +3226,12 @@ void exit_mmap(struct mm_struct *mm) mmap_write_unlock(mm); } + /* + * Taking write lock on mmap does not harm others, + * but it's crucial for uksm to avoid races. + */ + mmap_write_lock(mm); + if (mm->locked_vm) { vma = mm->mmap; while (vma) { @@ -3220,6 +3267,11 @@ void exit_mmap(struct mm_struct *mm) cond_resched(); } vm_unacct_memory(nr_accounted); + + mm->mmap = NULL; + mm->mm_rb = RB_ROOT; + vmacache_invalidate(mm); + mmap_write_unlock(mm); } /* Insert vm structure into process list sorted by address @@ -3322,11 +3374,12 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap, if (anon_vma_clone(new_vma, vma)) goto out_free_mempol; if (new_vma->vm_file) - get_file(new_vma->vm_file); + vma_get_file(new_vma); if (new_vma->vm_ops && new_vma->vm_ops->open) new_vma->vm_ops->open(new_vma); vma_link(mm, new_vma, prev, rb_link, rb_parent); *need_rmap_locks = false; + uksm_vma_add_new(new_vma); } return new_vma; @@ -3491,6 +3544,7 @@ static struct vm_area_struct *__install_special_mapping( vm_stat_account(mm, vma->vm_flags, len >> PAGE_SHIFT); perf_event_mmap(vma); + uksm_vma_add_new(vma); return vma; diff --git a/mm/mmzone.c b/mm/mmzone.c index eb89d6e018e..2ec0d779342 100644 --- a/mm/mmzone.c +++ b/mm/mmzone.c @@ -81,6 +81,8 @@ void lruvec_init(struct lruvec *lruvec) for_each_lru(lru) INIT_LIST_HEAD(&lruvec->lists[lru]); + + lru_gen_init_lruvec(lruvec); } #if defined(CONFIG_NUMA_BALANCING) && !defined(LAST_CPUPID_NOT_IN_PAGE_FLAGS) diff --git a/mm/nommu.c b/mm/nommu.c index 85a3a68dffb..a2bee44172c 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -523,7 +523,7 @@ static void __put_nommu_region(struct vm_region *region) up_write(&nommu_region_sem); if (region->vm_file) - fput(region->vm_file); + vmr_fput(region); /* IO memory and memory shared directly out of the pagecache * from ramfs/tmpfs mustn't be released here */ @@ -655,7 +655,7 @@ static void delete_vma(struct mm_struct *mm, struct vm_area_struct *vma) if (vma->vm_ops && vma->vm_ops->close) vma->vm_ops->close(vma); if (vma->vm_file) - fput(vma->vm_file); + vma_fput(vma); put_nommu_region(vma->vm_region); vm_area_free(vma); } @@ -1178,7 +1178,7 @@ unsigned long do_mmap(struct file *file, goto error_just_free; } } - fput(region->vm_file); + vmr_fput(region); kmem_cache_free(vm_region_jar, region); region = pregion; result = start; @@ -1255,10 +1255,10 @@ unsigned long do_mmap(struct file *file, up_write(&nommu_region_sem); error: if (region->vm_file) - fput(region->vm_file); + vmr_fput(region); kmem_cache_free(vm_region_jar, region); if (vma->vm_file) - fput(vma->vm_file); + vma_fput(vma); vm_area_free(vma); return ret; diff --git a/mm/page_alloc.c b/mm/page_alloc.c index de392908bf6..65172b65de4 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -118,6 +118,8 @@ typedef int __bitwise fpi_t; */ #define FPI_SKIP_KASAN_POISON ((__force fpi_t)BIT(2)) +atomic_long_t kswapd_waiters = ATOMIC_LONG_INIT(0); + /* prevent >1 _updater_ of zone percpu pageset ->high and ->batch fields */ static DEFINE_MUTEX(pcp_batch_high_lock); #define MIN_PERCPU_PAGELIST_FRACTION (8) @@ -331,20 +333,7 @@ compound_page_dtor * const compound_page_dtors[NR_COMPOUND_DTORS] = { int min_free_kbytes = 1024; int user_min_free_kbytes = -1; -#ifdef CONFIG_DISCONTIGMEM -/* - * DiscontigMem defines memory ranges as separate pg_data_t even if the ranges - * are not on separate NUMA nodes. Functionally this works but with - * watermark_boost_factor, it can reclaim prematurely as the ranges can be - * quite small. By default, do not boost watermarks on discontigmem as in - * many cases very high-order allocations like THP are likely to be - * unsupported and the premature reclaim offsets the advantage of long-term - * fragmentation avoidance. - */ int watermark_boost_factor __read_mostly; -#else -int watermark_boost_factor __read_mostly = 15000; -#endif int watermark_scale_factor = 10; static unsigned long nr_kernel_pages __initdata; @@ -4722,6 +4711,7 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, int no_progress_loops; unsigned int cpuset_mems_cookie; int reserve_flags; + bool woke_kswapd = false; /* * We also sanity check to catch abuse of atomic reserves being used by @@ -4755,8 +4745,13 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, if (!ac->preferred_zoneref->zone) goto nopage; - if (alloc_flags & ALLOC_KSWAPD) + if (alloc_flags & ALLOC_KSWAPD) { + if (!woke_kswapd) { + atomic_long_inc(&kswapd_waiters); + woke_kswapd = true; + } wake_all_kswapds(order, gfp_mask, ac); + } /* * The adjusted alloc_flags might result in immediate success, so try @@ -4961,9 +4956,12 @@ __alloc_pages_slowpath(gfp_t gfp_mask, unsigned int order, goto retry; } fail: - warn_alloc(gfp_mask, ac->nodemask, - "page allocation failure: order:%u", order); got_pg: + if (woke_kswapd) + atomic_long_dec(&kswapd_waiters); + if (!page) + warn_alloc(gfp_mask, ac->nodemask, + "page allocation failure: order:%u", order); return page; } diff --git a/mm/prfile.c b/mm/prfile.c new file mode 100644 index 00000000000..00d51187c32 --- /dev/null +++ b/mm/prfile.c @@ -0,0 +1,86 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * Mainly for aufs which mmap(2) different file and wants to print different + * path in /proc/PID/maps. + * Call these functions via macros defined in linux/mm.h. + * + * See Documentation/filesystems/aufs/design/06mmap.txt + * + * Copyright (c) 2014-2020 Junjro R. Okajima + * Copyright (c) 2014 Ian Campbell + */ + +#include +#include +#include + +/* #define PRFILE_TRACE */ +static inline void prfile_trace(struct file *f, struct file *pr, + const char func[], int line, const char func2[]) +{ +#ifdef PRFILE_TRACE + if (pr) + pr_info("%s:%d: %s, %pD2\n", func, line, func2, f); +#endif +} + +void vma_do_file_update_time(struct vm_area_struct *vma, const char func[], + int line) +{ + struct file *f = vma->vm_file, *pr = vma->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + file_update_time(f); + if (f && pr) + file_update_time(pr); +} + +struct file *vma_do_pr_or_file(struct vm_area_struct *vma, const char func[], + int line) +{ + struct file *f = vma->vm_file, *pr = vma->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + return (f && pr) ? pr : f; +} + +void vma_do_get_file(struct vm_area_struct *vma, const char func[], int line) +{ + struct file *f = vma->vm_file, *pr = vma->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + get_file(f); + if (f && pr) + get_file(pr); +} + +void vma_do_fput(struct vm_area_struct *vma, const char func[], int line) +{ + struct file *f = vma->vm_file, *pr = vma->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + fput(f); + if (f && pr) + fput(pr); +} + +#ifndef CONFIG_MMU +struct file *vmr_do_pr_or_file(struct vm_region *region, const char func[], + int line) +{ + struct file *f = region->vm_file, *pr = region->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + return (f && pr) ? pr : f; +} + +void vmr_do_fput(struct vm_region *region, const char func[], int line) +{ + struct file *f = region->vm_file, *pr = region->vm_prfile; + + prfile_trace(f, pr, func, line, __func__); + fput(f); + if (f && pr) + fput(pr); +} +#endif /* !CONFIG_MMU */ diff --git a/mm/rmap.c b/mm/rmap.c index e05c300048e..1a33e394f51 100644 --- a/mm/rmap.c +++ b/mm/rmap.c @@ -72,6 +72,7 @@ #include #include #include +#include #include @@ -789,6 +790,11 @@ static bool page_referenced_one(struct page *page, struct vm_area_struct *vma, } if (pvmw.pte) { + /* the multigenerational lru exploits the spatial locality */ + if (lru_gen_enabled() && pte_young(*pvmw.pte)) { + lru_gen_scan_around(&pvmw); + referenced++; + } if (ptep_clear_flush_young_notify(vma, address, pvmw.pte)) { /* diff --git a/mm/swap.c b/mm/swap.c index dfb48cf9c2c..96ce95eeb2c 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -433,6 +433,8 @@ void mark_page_accessed(struct page *page) * this list is never rotated or maintained, so marking an * evictable page accessed has no effect. */ + } else if (lru_gen_enabled()) { + page_inc_usage(page); } else if (!PageActive(page)) { /* * If the page is on the LRU, queue it for activation via @@ -478,15 +480,14 @@ void lru_cache_add(struct page *page) EXPORT_SYMBOL(lru_cache_add); /** - * lru_cache_add_inactive_or_unevictable + * lru_cache_add_page_vma * @page: the page to be added to LRU * @vma: vma in which page is mapped for determining reclaimability * - * Place @page on the inactive or unevictable LRU list, depending on its - * evictability. + * Place @page on an LRU list, depending on its evictability. */ -void lru_cache_add_inactive_or_unevictable(struct page *page, - struct vm_area_struct *vma) +void lru_cache_add_page_vma(struct page *page, struct vm_area_struct *vma, + bool faulting) { bool unevictable; @@ -503,6 +504,11 @@ void lru_cache_add_inactive_or_unevictable(struct page *page, __mod_zone_page_state(page_zone(page), NR_MLOCK, nr_pages); count_vm_events(UNEVICTABLE_PGMLOCKED, nr_pages); } + + /* tell the multigenerational lru that the page is being faulted in */ + if (lru_gen_enabled() && !unevictable && faulting) + SetPageActive(page); + lru_cache_add(page); } @@ -529,7 +535,7 @@ void lru_cache_add_inactive_or_unevictable(struct page *page, */ static void lru_deactivate_file_fn(struct page *page, struct lruvec *lruvec) { - bool active = PageActive(page); + bool active = PageActive(page) || lru_gen_enabled(); int nr_pages = thp_nr_pages(page); if (PageUnevictable(page)) @@ -569,7 +575,7 @@ static void lru_deactivate_file_fn(struct page *page, struct lruvec *lruvec) static void lru_deactivate_fn(struct page *page, struct lruvec *lruvec) { - if (PageActive(page) && !PageUnevictable(page)) { + if (!PageUnevictable(page) && (PageActive(page) || lru_gen_enabled())) { int nr_pages = thp_nr_pages(page); del_page_from_lru_list(page, lruvec); @@ -684,7 +690,7 @@ void deactivate_file_page(struct page *page) */ void deactivate_page(struct page *page) { - if (PageLRU(page) && PageActive(page) && !PageUnevictable(page)) { + if (PageLRU(page) && !PageUnevictable(page) && (PageActive(page) || lru_gen_enabled())) { struct pagevec *pvec; local_lock(&lru_pvecs.lock); diff --git a/mm/swapfile.c b/mm/swapfile.c index 996afa8131c..8b5ca15df12 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -1936,7 +1936,7 @@ static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd, page_add_anon_rmap(page, vma, addr, false); } else { /* ksm created a completely new copy */ page_add_new_anon_rmap(page, vma, addr, false); - lru_cache_add_inactive_or_unevictable(page, vma); + lru_cache_add_page_vma(page, vma, false); } swap_free(entry); out: @@ -2702,6 +2702,8 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile) err = 0; atomic_inc(&proc_poll_event); wake_up_interruptible(&proc_poll_wait); + /* stop tracking anon if the multigenerational lru is turned off */ + lru_gen_set_state(false, false, true); out_dput: filp_close(victim, NULL); @@ -3348,6 +3350,8 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags) mutex_unlock(&swapon_mutex); atomic_inc(&proc_poll_event); wake_up_interruptible(&proc_poll_wait); + /* start tracking anon if the multigenerational lru is turned on */ + lru_gen_set_state(true, false, true); error = 0; goto out; diff --git a/mm/uksm.c b/mm/uksm.c new file mode 100644 index 00000000000..e4732c00be6 --- /dev/null +++ b/mm/uksm.c @@ -0,0 +1,5614 @@ +/* + * Ultra KSM. Copyright (C) 2011-2012 Nai Xia + * + * This is an improvement upon KSM. Some basic data structures and routines + * are borrowed from ksm.c . + * + * Its new features: + * 1. Full system scan: + * It automatically scans all user processes' anonymous VMAs. Kernel-user + * interaction to submit a memory area to KSM is no longer needed. + * + * 2. Rich area detection: + * It automatically detects rich areas containing abundant duplicated + * pages based. Rich areas are given a full scan speed. Poor areas are + * sampled at a reasonable speed with very low CPU consumption. + * + * 3. Ultra Per-page scan speed improvement: + * A new hash algorithm is proposed. As a result, on a machine with + * Core(TM)2 Quad Q9300 CPU in 32-bit mode and 800MHZ DDR2 main memory, it + * can scan memory areas that does not contain duplicated pages at speed of + * 627MB/sec ~ 2445MB/sec and can merge duplicated areas at speed of + * 477MB/sec ~ 923MB/sec. + * + * 4. Thrashing area avoidance: + * Thrashing area(an VMA that has frequent Ksm page break-out) can be + * filtered out. My benchmark shows it's more efficient than KSM's per-page + * hash value based volatile page detection. + * + * + * 5. Misc changes upon KSM: + * * It has a fully x86-opitmized memcmp dedicated for 4-byte-aligned page + * comparison. It's much faster than default C version on x86. + * * rmap_item now has an struct *page member to loosely cache a + * address-->page mapping, which reduces too much time-costly + * follow_page(). + * * The VMA creation/exit procedures are hooked to let the Ultra KSM know. + * * try_to_merge_two_pages() now can revert a pte if it fails. No break_ + * ksm is needed for this case. + * + * 6. Full Zero Page consideration(contributed by Figo Zhang) + * Now uksmd consider full zero pages as special pages and merge them to an + * special unswappable uksm zero page. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include +#include "internal.h" + +#ifdef CONFIG_X86 +#undef memcmp + +#ifdef CONFIG_X86_32 +#define memcmp memcmpx86_32 +/* + * Compare 4-byte-aligned address s1 and s2, with length n + */ +int memcmpx86_32(void *s1, void *s2, size_t n) +{ + size_t num = n / 4; + register int res; + + __asm__ __volatile__ + ( + "testl %3,%3\n\t" + "repe; cmpsd\n\t" + "je 1f\n\t" + "sbbl %0,%0\n\t" + "orl $1,%0\n" + "1:" + : "=&a" (res), "+&S" (s1), "+&D" (s2), "+&c" (num) + : "0" (0) + : "cc"); + + return res; +} + +/* + * Check the page is all zero ? + */ +static int is_full_zero(const void *s1, size_t len) +{ + unsigned char same; + + len /= 4; + + __asm__ __volatile__ + ("repe; scasl;" + "sete %0" + : "=qm" (same), "+D" (s1), "+c" (len) + : "a" (0) + : "cc"); + + return same; +} + + +#elif defined(CONFIG_X86_64) +#define memcmp memcmpx86_64 +/* + * Compare 8-byte-aligned address s1 and s2, with length n + */ +int memcmpx86_64(void *s1, void *s2, size_t n) +{ + size_t num = n / 8; + register int res; + + __asm__ __volatile__ + ( + "testq %q3,%q3\n\t" + "repe; cmpsq\n\t" + "je 1f\n\t" + "sbbq %q0,%q0\n\t" + "orq $1,%q0\n" + "1:" + : "=&a" (res), "+&S" (s1), "+&D" (s2), "+&c" (num) + : "0" (0) + : "cc"); + + return res; +} + +static int is_full_zero(const void *s1, size_t len) +{ + unsigned char same; + + len /= 8; + + __asm__ __volatile__ + ("repe; scasq;" + "sete %0" + : "=qm" (same), "+D" (s1), "+c" (len) + : "a" (0) + : "cc"); + + return same; +} + +#endif +#else +static int is_full_zero(const void *s1, size_t len) +{ + unsigned long *src = s1; + int i; + + len /= sizeof(*src); + + for (i = 0; i < len; i++) { + if (src[i]) + return 0; + } + + return 1; +} +#endif + +#define UKSM_RUNG_ROUND_FINISHED (1 << 0) +#define TIME_RATIO_SCALE 10000 + +#define SLOT_TREE_NODE_SHIFT 8 +#define SLOT_TREE_NODE_STORE_SIZE (1UL << SLOT_TREE_NODE_SHIFT) +struct slot_tree_node { + unsigned long size; + struct sradix_tree_node snode; + void *stores[SLOT_TREE_NODE_STORE_SIZE]; +}; + +static struct kmem_cache *slot_tree_node_cachep; + +static struct sradix_tree_node *slot_tree_node_alloc(void) +{ + struct slot_tree_node *p; + + p = kmem_cache_zalloc(slot_tree_node_cachep, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (!p) + return NULL; + + return &p->snode; +} + +static void slot_tree_node_free(struct sradix_tree_node *node) +{ + struct slot_tree_node *p; + + p = container_of(node, struct slot_tree_node, snode); + kmem_cache_free(slot_tree_node_cachep, p); +} + +static void slot_tree_node_extend(struct sradix_tree_node *parent, + struct sradix_tree_node *child) +{ + struct slot_tree_node *p, *c; + + p = container_of(parent, struct slot_tree_node, snode); + c = container_of(child, struct slot_tree_node, snode); + + p->size += c->size; +} + +void slot_tree_node_assign(struct sradix_tree_node *node, + unsigned int index, void *item) +{ + struct vma_slot *slot = item; + struct slot_tree_node *cur; + + slot->snode = node; + slot->sindex = index; + + while (node) { + cur = container_of(node, struct slot_tree_node, snode); + cur->size += slot->pages; + node = node->parent; + } +} + +void slot_tree_node_rm(struct sradix_tree_node *node, unsigned int offset) +{ + struct vma_slot *slot; + struct slot_tree_node *cur; + unsigned long pages; + + if (node->height == 1) { + slot = node->stores[offset]; + pages = slot->pages; + } else { + cur = container_of(node->stores[offset], + struct slot_tree_node, snode); + pages = cur->size; + } + + while (node) { + cur = container_of(node, struct slot_tree_node, snode); + cur->size -= pages; + node = node->parent; + } +} + +unsigned long slot_iter_index; +int slot_iter(void *item, unsigned long height) +{ + struct slot_tree_node *node; + struct vma_slot *slot; + + if (height == 1) { + slot = item; + if (slot_iter_index < slot->pages) { + /*in this one*/ + return 1; + } else { + slot_iter_index -= slot->pages; + return 0; + } + + } else { + node = container_of(item, struct slot_tree_node, snode); + if (slot_iter_index < node->size) { + /*in this one*/ + return 1; + } else { + slot_iter_index -= node->size; + return 0; + } + } +} + + +static inline void slot_tree_init_root(struct sradix_tree_root *root) +{ + init_sradix_tree_root(root, SLOT_TREE_NODE_SHIFT); + root->alloc = slot_tree_node_alloc; + root->free = slot_tree_node_free; + root->extend = slot_tree_node_extend; + root->assign = slot_tree_node_assign; + root->rm = slot_tree_node_rm; +} + +void slot_tree_init(void) +{ + slot_tree_node_cachep = kmem_cache_create("slot_tree_node", + sizeof(struct slot_tree_node), 0, + SLAB_PANIC | SLAB_RECLAIM_ACCOUNT, + NULL); +} + + +/* Each rung of this ladder is a list of VMAs having a same scan ratio */ +struct scan_rung { + //struct list_head scanned_list; + struct sradix_tree_root vma_root; + struct sradix_tree_root vma_root2; + + struct vma_slot *current_scan; + unsigned long current_offset; + + /* + * The initial value for current_offset, it should loop over + * [0~ step - 1] to let all slot have its chance to be scanned. + */ + unsigned long offset_init; + unsigned long step; /* dynamic step for current_offset */ + unsigned int flags; + unsigned long pages_to_scan; + //unsigned long fully_scanned_slots; + /* + * a little bit tricky - if cpu_time_ratio > 0, then the value is the + * the cpu time ratio it can spend in rung_i for every scan + * period. if < 0, then it is the cpu time ratio relative to the + * max cpu percentage user specified. Both in unit of + * 1/TIME_RATIO_SCALE + */ + int cpu_ratio; + + /* + * How long it will take for all slots in this rung to be fully + * scanned? If it's zero, we don't care about the cover time: + * it's fully scanned. + */ + unsigned int cover_msecs; + //unsigned long vma_num; + //unsigned long pages; /* Sum of all slot's pages in rung */ +}; + +/** + * node of either the stable or unstale rbtree + * + */ +struct tree_node { + struct rb_node node; /* link in the main (un)stable rbtree */ + struct rb_root sub_root; /* rb_root for sublevel collision rbtree */ + u32 hash; + unsigned long count; /* TODO: merged with sub_root */ + struct list_head all_list; /* all tree nodes in stable/unstable tree */ +}; + +/** + * struct stable_node - node of the stable rbtree + * @node: rb node of this ksm page in the stable tree + * @hlist: hlist head of rmap_items using this ksm page + * @kpfn: page frame number of this ksm page + */ +struct stable_node { + struct rb_node node; /* link in sub-rbtree */ + struct tree_node *tree_node; /* it's tree node root in stable tree, NULL if it's in hell list */ + struct hlist_head hlist; + unsigned long kpfn; + u32 hash_max; /* if ==0 then it's not been calculated yet */ + struct list_head all_list; /* in a list for all stable nodes */ +}; + +/** + * struct node_vma - group rmap_items linked in a same stable + * node together. + */ +struct node_vma { + union { + struct vma_slot *slot; + unsigned long key; /* slot is used as key sorted on hlist */ + }; + struct hlist_node hlist; + struct hlist_head rmap_hlist; + struct stable_node *head; +}; + +/** + * struct rmap_item - reverse mapping item for virtual addresses + * @rmap_list: next rmap_item in mm_slot's singly-linked rmap_list + * @anon_vma: pointer to anon_vma for this mm,address, when in stable tree + * @mm: the memory structure this rmap_item is pointing into + * @address: the virtual address this rmap_item tracks (+ flags in low bits) + * @node: rb node of this rmap_item in the unstable tree + * @head: pointer to stable_node heading this list in the stable tree + * @hlist: link into hlist of rmap_items hanging off that stable_node + */ +struct rmap_item { + struct vma_slot *slot; + struct page *page; + unsigned long address; /* + low bits used for flags below */ + unsigned long hash_round; + unsigned long entry_index; + union { + struct {/* when in unstable tree */ + struct rb_node node; + struct tree_node *tree_node; + u32 hash_max; + }; + struct { /* when in stable tree */ + struct node_vma *head; + struct hlist_node hlist; + struct anon_vma *anon_vma; + }; + }; +} __aligned(4); + +struct rmap_list_entry { + union { + struct rmap_item *item; + unsigned long addr; + }; + /* lowest bit is used for is_addr tag */ +} __aligned(4); /* 4 aligned to fit in to pages*/ + + +/* Basic data structure definition ends */ + + +/* + * Flags for rmap_item to judge if it's listed in the stable/unstable tree. + * The flags use the low bits of rmap_item.address + */ +#define UNSTABLE_FLAG 0x1 +#define STABLE_FLAG 0x2 +#define get_rmap_addr(x) ((x)->address & PAGE_MASK) + +/* + * rmap_list_entry helpers + */ +#define IS_ADDR_FLAG 1 +#define is_addr(ptr) ((unsigned long)(ptr) & IS_ADDR_FLAG) +#define set_is_addr(ptr) ((ptr) |= IS_ADDR_FLAG) +#define get_clean_addr(ptr) (((ptr) & ~(__typeof__(ptr))IS_ADDR_FLAG)) + + +/* + * High speed caches for frequently allocated and freed structs + */ +static struct kmem_cache *rmap_item_cache; +static struct kmem_cache *stable_node_cache; +static struct kmem_cache *node_vma_cache; +static struct kmem_cache *vma_slot_cache; +static struct kmem_cache *tree_node_cache; +#define UKSM_KMEM_CACHE(__struct, __flags) kmem_cache_create("uksm_"#__struct,\ + sizeof(struct __struct), __alignof__(struct __struct),\ + (__flags), NULL) + +/* Array of all scan_rung, uksm_scan_ladder[0] having the minimum scan ratio */ +#define SCAN_LADDER_SIZE 4 +static struct scan_rung uksm_scan_ladder[SCAN_LADDER_SIZE]; + +/* The evaluation rounds uksmd has finished */ +static unsigned long long uksm_eval_round = 1; + +/* + * we add 1 to this var when we consider we should rebuild the whole + * unstable tree. + */ +static unsigned long uksm_hash_round = 1; + +/* + * How many times the whole memory is scanned. + */ +static unsigned long long fully_scanned_round = 1; + +/* The total number of virtual pages of all vma slots */ +static u64 uksm_pages_total; + +/* The number of pages has been scanned since the start up */ +static u64 uksm_pages_scanned; + +static u64 scanned_virtual_pages; + +/* The number of pages has been scanned since last encode_benefit call */ +static u64 uksm_pages_scanned_last; + +/* If the scanned number is tooo large, we encode it here */ +static u64 pages_scanned_stored; + +static unsigned long pages_scanned_base; + +/* The number of nodes in the stable tree */ +static unsigned long uksm_pages_shared; + +/* The number of page slots additionally sharing those nodes */ +static unsigned long uksm_pages_sharing; + +/* The number of nodes in the unstable tree */ +static unsigned long uksm_pages_unshared; + +/* + * Milliseconds ksmd should sleep between scans, + * >= 100ms to be consistent with + * scan_time_to_sleep_msec() + */ +static unsigned int uksm_sleep_jiffies; + +/* The real value for the uksmd next sleep */ +static unsigned int uksm_sleep_real; + +/* Saved value for user input uksm_sleep_jiffies when it's enlarged */ +static unsigned int uksm_sleep_saved; + +/* Max percentage of cpu utilization ksmd can take to scan in one batch */ +static unsigned int uksm_max_cpu_percentage; + +static int uksm_cpu_governor; + +static char *uksm_cpu_governor_str[4] = { "full", "medium", "low", "quiet" }; + +struct uksm_cpu_preset_s { + int cpu_ratio[SCAN_LADDER_SIZE]; + unsigned int cover_msecs[SCAN_LADDER_SIZE]; + unsigned int max_cpu; /* percentage */ +}; + +struct uksm_cpu_preset_s uksm_cpu_preset[4] = { + { {20, 40, -2500, -10000}, {1000, 500, 200, 50}, 95}, + { {20, 30, -2500, -10000}, {1000, 500, 400, 100}, 50}, + { {10, 20, -5000, -10000}, {1500, 1000, 1000, 250}, 20}, + { {10, 20, 40, 75}, {2000, 1000, 1000, 1000}, 1}, +}; + +/* The default value for uksm_ema_page_time if it's not initialized */ +#define UKSM_PAGE_TIME_DEFAULT 500 + +/*cost to scan one page by expotional moving average in nsecs */ +static unsigned long uksm_ema_page_time = UKSM_PAGE_TIME_DEFAULT; + +/* The expotional moving average alpha weight, in percentage. */ +#define EMA_ALPHA 20 + +/* + * The threshold used to filter out thrashing areas, + * If it == 0, filtering is disabled, otherwise it's the percentage up-bound + * of the thrashing ratio of all areas. Any area with a bigger thrashing ratio + * will be considered as having a zero duplication ratio. + */ +static unsigned int uksm_thrash_threshold = 50; + +/* How much dedup ratio is considered to be abundant*/ +static unsigned int uksm_abundant_threshold = 10; + +/* All slots having merged pages in this eval round. */ +struct list_head vma_slot_dedup = LIST_HEAD_INIT(vma_slot_dedup); + +/* How many times the ksmd has slept since startup */ +static unsigned long long uksm_sleep_times; + +#define UKSM_RUN_STOP 0 +#define UKSM_RUN_MERGE 1 +static unsigned int uksm_run = 1; + +static DECLARE_WAIT_QUEUE_HEAD(uksm_thread_wait); +static DEFINE_MUTEX(uksm_thread_mutex); + +/* + * List vma_slot_new is for newly created vma_slot waiting to be added by + * ksmd. If one cannot be added(e.g. due to it's too small), it's moved to + * vma_slot_noadd. vma_slot_del is the list for vma_slot whose corresponding + * VMA has been removed/freed. + */ +struct list_head vma_slot_new = LIST_HEAD_INIT(vma_slot_new); +struct list_head vma_slot_noadd = LIST_HEAD_INIT(vma_slot_noadd); +struct list_head vma_slot_del = LIST_HEAD_INIT(vma_slot_del); +static DEFINE_SPINLOCK(vma_slot_list_lock); + +/* The unstable tree heads */ +static struct rb_root root_unstable_tree = RB_ROOT; + +/* + * All tree_nodes are in a list to be freed at once when unstable tree is + * freed after each scan round. + */ +static struct list_head unstable_tree_node_list = + LIST_HEAD_INIT(unstable_tree_node_list); + +/* List contains all stable nodes */ +static struct list_head stable_node_list = LIST_HEAD_INIT(stable_node_list); + +/* + * When the hash strength is changed, the stable tree must be delta_hashed and + * re-structured. We use two set of below structs to speed up the + * re-structuring of stable tree. + */ +static struct list_head +stable_tree_node_list[2] = {LIST_HEAD_INIT(stable_tree_node_list[0]), + LIST_HEAD_INIT(stable_tree_node_list[1])}; + +static struct list_head *stable_tree_node_listp = &stable_tree_node_list[0]; +static struct rb_root root_stable_tree[2] = {RB_ROOT, RB_ROOT}; +static struct rb_root *root_stable_treep = &root_stable_tree[0]; +static unsigned long stable_tree_index; + +/* The hash strength needed to hash a full page */ +#define HASH_STRENGTH_FULL (PAGE_SIZE / sizeof(u32)) + +/* The hash strength needed for loop-back hashing */ +#define HASH_STRENGTH_MAX (HASH_STRENGTH_FULL + 10) + +/* The random offsets in a page */ +static u32 *random_nums; + +/* The hash strength */ +static unsigned long hash_strength = HASH_STRENGTH_FULL >> 4; + +/* The delta value each time the hash strength increases or decreases */ +static unsigned long hash_strength_delta; +#define HASH_STRENGTH_DELTA_MAX 5 + +/* The time we have saved due to random_sample_hash */ +static u64 rshash_pos; + +/* The time we have wasted due to hash collision */ +static u64 rshash_neg; + +struct uksm_benefit { + u64 pos; + u64 neg; + u64 scanned; + unsigned long base; +} benefit; + +/* + * The relative cost of memcmp, compared to 1 time unit of random sample + * hash, this value is tested when ksm module is initialized + */ +static unsigned long memcmp_cost; + +static unsigned long rshash_neg_cont_zero; +static unsigned long rshash_cont_obscure; + +/* The possible states of hash strength adjustment heuristic */ +enum rshash_states { + RSHASH_STILL, + RSHASH_TRYUP, + RSHASH_TRYDOWN, + RSHASH_NEW, + RSHASH_PRE_STILL, +}; + +/* The possible direction we are about to adjust hash strength */ +enum rshash_direct { + GO_UP, + GO_DOWN, + OBSCURE, + STILL, +}; + +/* random sampling hash state machine */ +static struct { + enum rshash_states state; + enum rshash_direct pre_direct; + u8 below_count; + /* Keep a lookup window of size 5, iff above_count/below_count > 3 + * in this window we stop trying. + */ + u8 lookup_window_index; + u64 stable_benefit; + unsigned long turn_point_down; + unsigned long turn_benefit_down; + unsigned long turn_point_up; + unsigned long turn_benefit_up; + unsigned long stable_point; +} rshash_state; + +/*zero page hash table, hash_strength [0 ~ HASH_STRENGTH_MAX]*/ +static u32 *zero_hash_table; + +static inline struct node_vma *alloc_node_vma(void) +{ + struct node_vma *node_vma; + + node_vma = kmem_cache_zalloc(node_vma_cache, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (node_vma) { + INIT_HLIST_HEAD(&node_vma->rmap_hlist); + INIT_HLIST_NODE(&node_vma->hlist); + } + return node_vma; +} + +static inline void free_node_vma(struct node_vma *node_vma) +{ + kmem_cache_free(node_vma_cache, node_vma); +} + + +static inline struct vma_slot *alloc_vma_slot(void) +{ + struct vma_slot *slot; + + /* + * In case ksm is not initialized by now. + * Oops, we need to consider the call site of uksm_init() in the future. + */ + if (!vma_slot_cache) + return NULL; + + slot = kmem_cache_zalloc(vma_slot_cache, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (slot) { + INIT_LIST_HEAD(&slot->slot_list); + INIT_LIST_HEAD(&slot->dedup_list); + slot->flags |= UKSM_SLOT_NEED_RERAND; + } + return slot; +} + +static inline void free_vma_slot(struct vma_slot *vma_slot) +{ + kmem_cache_free(vma_slot_cache, vma_slot); +} + + + +static inline struct rmap_item *alloc_rmap_item(void) +{ + struct rmap_item *rmap_item; + + rmap_item = kmem_cache_zalloc(rmap_item_cache, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (rmap_item) { + /* bug on lowest bit is not clear for flag use */ + BUG_ON(is_addr(rmap_item)); + } + return rmap_item; +} + +static inline void free_rmap_item(struct rmap_item *rmap_item) +{ + rmap_item->slot = NULL; /* debug safety */ + kmem_cache_free(rmap_item_cache, rmap_item); +} + +static inline struct stable_node *alloc_stable_node(void) +{ + struct stable_node *node; + + node = kmem_cache_alloc(stable_node_cache, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (!node) + return NULL; + + INIT_HLIST_HEAD(&node->hlist); + list_add(&node->all_list, &stable_node_list); + return node; +} + +static inline void free_stable_node(struct stable_node *stable_node) +{ + list_del(&stable_node->all_list); + kmem_cache_free(stable_node_cache, stable_node); +} + +static inline struct tree_node *alloc_tree_node(struct list_head *list) +{ + struct tree_node *node; + + node = kmem_cache_zalloc(tree_node_cache, GFP_KERNEL | + __GFP_NORETRY | __GFP_NOWARN); + if (!node) + return NULL; + + list_add(&node->all_list, list); + return node; +} + +static inline void free_tree_node(struct tree_node *node) +{ + list_del(&node->all_list); + kmem_cache_free(tree_node_cache, node); +} + +static void uksm_drop_anon_vma(struct rmap_item *rmap_item) +{ + struct anon_vma *anon_vma = rmap_item->anon_vma; + + put_anon_vma(anon_vma); +} + + +/** + * Remove a stable node from stable_tree, may unlink from its tree_node and + * may remove its parent tree_node if no other stable node is pending. + * + * @stable_node The node need to be removed + * @unlink_rb Will this node be unlinked from the rbtree? + * @remove_tree_ node Will its tree_node be removed if empty? + */ +static void remove_node_from_stable_tree(struct stable_node *stable_node, + int unlink_rb, int remove_tree_node) +{ + struct node_vma *node_vma; + struct rmap_item *rmap_item; + struct hlist_node *n; + + if (!hlist_empty(&stable_node->hlist)) { + hlist_for_each_entry_safe(node_vma, n, + &stable_node->hlist, hlist) { + hlist_for_each_entry(rmap_item, &node_vma->rmap_hlist, hlist) { + uksm_pages_sharing--; + + uksm_drop_anon_vma(rmap_item); + rmap_item->address &= PAGE_MASK; + } + free_node_vma(node_vma); + cond_resched(); + } + + /* the last one is counted as shared */ + uksm_pages_shared--; + uksm_pages_sharing++; + } + + if (stable_node->tree_node && unlink_rb) { + rb_erase(&stable_node->node, + &stable_node->tree_node->sub_root); + + if (RB_EMPTY_ROOT(&stable_node->tree_node->sub_root) && + remove_tree_node) { + rb_erase(&stable_node->tree_node->node, + root_stable_treep); + free_tree_node(stable_node->tree_node); + } else { + stable_node->tree_node->count--; + } + } + + free_stable_node(stable_node); +} + + +/* + * get_uksm_page: checks if the page indicated by the stable node + * is still its ksm page, despite having held no reference to it. + * In which case we can trust the content of the page, and it + * returns the gotten page; but if the page has now been zapped, + * remove the stale node from the stable tree and return NULL. + * + * You would expect the stable_node to hold a reference to the ksm page. + * But if it increments the page's count, swapping out has to wait for + * ksmd to come around again before it can free the page, which may take + * seconds or even minutes: much too unresponsive. So instead we use a + * "keyhole reference": access to the ksm page from the stable node peeps + * out through its keyhole to see if that page still holds the right key, + * pointing back to this stable node. This relies on freeing a PageAnon + * page to reset its page->mapping to NULL, and relies on no other use of + * a page to put something that might look like our key in page->mapping. + * + * include/linux/pagemap.h page_cache_get_speculative() is a good reference, + * but this is different - made simpler by uksm_thread_mutex being held, but + * interesting for assuming that no other use of the struct page could ever + * put our expected_mapping into page->mapping (or a field of the union which + * coincides with page->mapping). The RCU calls are not for KSM at all, but + * to keep the page_count protocol described with page_cache_get_speculative. + * + * Note: it is possible that get_uksm_page() will return NULL one moment, + * then page the next, if the page is in between page_freeze_refs() and + * page_unfreeze_refs(): this shouldn't be a problem anywhere, the page + * is on its way to being freed; but it is an anomaly to bear in mind. + * + * @unlink_rb: if the removal of this node will firstly unlink from + * its rbtree. stable_node_reinsert will prevent this when restructuring the + * node from its old tree. + * + * @remove_tree_node: if this is the last one of its tree_node, will the + * tree_node be freed ? If we are inserting stable node, this tree_node may + * be reused, so don't free it. + */ +static struct page *get_uksm_page(struct stable_node *stable_node, + int unlink_rb, int remove_tree_node) +{ + struct page *page; + void *expected_mapping; + unsigned long kpfn; + + expected_mapping = (void *)((unsigned long)stable_node | + PAGE_MAPPING_KSM); +again: + kpfn = READ_ONCE(stable_node->kpfn); + page = pfn_to_page(kpfn); + + /* + * page is computed from kpfn, so on most architectures reading + * page->mapping is naturally ordered after reading node->kpfn, + * but on Alpha we need to be more careful. + */ + smp_rmb(); + + if (READ_ONCE(page->mapping) != expected_mapping) + goto stale; + + /* + * We cannot do anything with the page while its refcount is 0. + * Usually 0 means free, or tail of a higher-order page: in which + * case this node is no longer referenced, and should be freed; + * however, it might mean that the page is under page_freeze_refs(). + * The __remove_mapping() case is easy, again the node is now stale; + * but if page is swapcache in migrate_page_move_mapping(), it might + * still be our page, in which case it's essential to keep the node. + */ + while (!get_page_unless_zero(page)) { + /* + * Another check for page->mapping != expected_mapping would + * work here too. We have chosen the !PageSwapCache test to + * optimize the common case, when the page is or is about to + * be freed: PageSwapCache is cleared (under spin_lock_irq) + * in the freeze_refs section of __remove_mapping(); but Anon + * page->mapping reset to NULL later, in free_pages_prepare(). + */ + if (!PageSwapCache(page)) + goto stale; + cpu_relax(); + } + + if (READ_ONCE(page->mapping) != expected_mapping) { + put_page(page); + goto stale; + } + + lock_page(page); + if (READ_ONCE(page->mapping) != expected_mapping) { + unlock_page(page); + put_page(page); + goto stale; + } + unlock_page(page); + return page; +stale: + /* + * We come here from above when page->mapping or !PageSwapCache + * suggests that the node is stale; but it might be under migration. + * We need smp_rmb(), matching the smp_wmb() in ksm_migrate_page(), + * before checking whether node->kpfn has been changed. + */ + smp_rmb(); + if (stable_node->kpfn != kpfn) + goto again; + + remove_node_from_stable_tree(stable_node, unlink_rb, remove_tree_node); + + return NULL; +} + +/* + * Removing rmap_item from stable or unstable tree. + * This function will clean the information from the stable/unstable tree. + */ +static inline void remove_rmap_item_from_tree(struct rmap_item *rmap_item) +{ + if (rmap_item->address & STABLE_FLAG) { + struct stable_node *stable_node; + struct node_vma *node_vma; + struct page *page; + + node_vma = rmap_item->head; + stable_node = node_vma->head; + page = get_uksm_page(stable_node, 1, 1); + if (!page) + goto out; + + /* + * page lock is needed because it's racing with + * try_to_unmap_ksm(), etc. + */ + lock_page(page); + hlist_del(&rmap_item->hlist); + + if (hlist_empty(&node_vma->rmap_hlist)) { + hlist_del(&node_vma->hlist); + free_node_vma(node_vma); + } + unlock_page(page); + + put_page(page); + if (hlist_empty(&stable_node->hlist)) { + /* do NOT call remove_node_from_stable_tree() here, + * it's possible for a forked rmap_item not in + * stable tree while the in-tree rmap_items were + * deleted. + */ + uksm_pages_shared--; + } else + uksm_pages_sharing--; + + + uksm_drop_anon_vma(rmap_item); + } else if (rmap_item->address & UNSTABLE_FLAG) { + if (rmap_item->hash_round == uksm_hash_round) { + + rb_erase(&rmap_item->node, + &rmap_item->tree_node->sub_root); + if (RB_EMPTY_ROOT(&rmap_item->tree_node->sub_root)) { + rb_erase(&rmap_item->tree_node->node, + &root_unstable_tree); + + free_tree_node(rmap_item->tree_node); + } else + rmap_item->tree_node->count--; + } + uksm_pages_unshared--; + } + + rmap_item->address &= PAGE_MASK; + rmap_item->hash_max = 0; + +out: + cond_resched(); /* we're called from many long loops */ +} + +static inline int slot_in_uksm(struct vma_slot *slot) +{ + return list_empty(&slot->slot_list); +} + +/* + * Test if the mm is exiting + */ +static inline bool uksm_test_exit(struct mm_struct *mm) +{ + return atomic_read(&mm->mm_users) == 0; +} + +static inline unsigned long vma_pool_size(struct vma_slot *slot) +{ + return round_up(sizeof(struct rmap_list_entry) * slot->pages, + PAGE_SIZE) >> PAGE_SHIFT; +} + +#define CAN_OVERFLOW_U64(x, delta) (U64_MAX - (x) < (delta)) + +/* must be done with sem locked */ +static int slot_pool_alloc(struct vma_slot *slot) +{ + unsigned long pool_size; + + if (slot->rmap_list_pool) + return 0; + + pool_size = vma_pool_size(slot); + slot->rmap_list_pool = kcalloc(pool_size, sizeof(struct page *), + GFP_KERNEL); + if (!slot->rmap_list_pool) + return -ENOMEM; + + slot->pool_counts = kcalloc(pool_size, sizeof(unsigned int), + GFP_KERNEL); + if (!slot->pool_counts) { + kfree(slot->rmap_list_pool); + return -ENOMEM; + } + + slot->pool_size = pool_size; + BUG_ON(CAN_OVERFLOW_U64(uksm_pages_total, slot->pages)); + slot->flags |= UKSM_SLOT_IN_UKSM; + uksm_pages_total += slot->pages; + + return 0; +} + +/* + * Called after vma is unlinked from its mm + */ +void uksm_remove_vma(struct vm_area_struct *vma) +{ + struct vma_slot *slot; + + if (!vma->uksm_vma_slot) + return; + + spin_lock(&vma_slot_list_lock); + slot = vma->uksm_vma_slot; + if (!slot) + goto out; + + if (slot_in_uksm(slot)) { + /** + * This slot has been added by ksmd, so move to the del list + * waiting ksmd to free it. + */ + list_add_tail(&slot->slot_list, &vma_slot_del); + } else { + /** + * It's still on new list. It's ok to free slot directly. + */ + list_del(&slot->slot_list); + free_vma_slot(slot); + } +out: + vma->uksm_vma_slot = NULL; + spin_unlock(&vma_slot_list_lock); +} + +/** + * Need to do two things: + * 1. check if slot was moved to del list + * 2. make sure the mmap_sem is manipulated under valid vma. + * + * My concern here is that in some cases, this may make + * vma_slot_list_lock() waiters to serialized further by some + * sem->wait_lock, can this really be expensive? + * + * + * @return + * 0: if successfully locked mmap_sem + * -ENOENT: this slot was moved to del list + * -EBUSY: vma lock failed + */ +static int try_down_read_slot_mmap_sem(struct vma_slot *slot) +{ + struct vm_area_struct *vma; + struct mm_struct *mm; + struct rw_semaphore *sem; + + spin_lock(&vma_slot_list_lock); + + /* the slot_list was removed and inited from new list, when it enters + * uksm_list. If now it's not empty, then it must be moved to del list + */ + if (!slot_in_uksm(slot)) { + spin_unlock(&vma_slot_list_lock); + return -ENOENT; + } + + BUG_ON(slot->pages != vma_pages(slot->vma)); + /* Ok, vma still valid */ + vma = slot->vma; + mm = vma->vm_mm; + sem = &mm->mmap_lock; + + if (uksm_test_exit(mm)) { + spin_unlock(&vma_slot_list_lock); + return -ENOENT; + } + + if (down_read_trylock(sem)) { + spin_unlock(&vma_slot_list_lock); + if (slot_pool_alloc(slot)) { + uksm_remove_vma(vma); + up_read(sem); + return -ENOENT; + } + return 0; + } + + spin_unlock(&vma_slot_list_lock); + return -EBUSY; +} + +static inline unsigned long +vma_page_address(struct page *page, struct vm_area_struct *vma) +{ + pgoff_t pgoff = page->index; + unsigned long address; + + address = vma->vm_start + ((pgoff - vma->vm_pgoff) << PAGE_SHIFT); + if (unlikely(address < vma->vm_start || address >= vma->vm_end)) { + /* page should be within @vma mapping range */ + return -EFAULT; + } + return address; +} + + +/* return 0 on success with the item's mmap_sem locked */ +static inline int get_mergeable_page_lock_mmap(struct rmap_item *item) +{ + struct mm_struct *mm; + struct vma_slot *slot = item->slot; + int err = -EINVAL; + + struct page *page; + + /* + * try_down_read_slot_mmap_sem() returns non-zero if the slot + * has been removed by uksm_remove_vma(). + */ + if (try_down_read_slot_mmap_sem(slot)) + return -EBUSY; + + mm = slot->vma->vm_mm; + + if (uksm_test_exit(mm)) + goto failout_up; + + page = item->page; + rcu_read_lock(); + if (!get_page_unless_zero(page)) { + rcu_read_unlock(); + goto failout_up; + } + + /* No need to consider huge page here. */ + if (item->slot->vma->anon_vma != page_anon_vma(page) || + vma_page_address(page, item->slot->vma) != get_rmap_addr(item)) { + /* + * TODO: + * should we release this item becase of its stale page + * mapping? + */ + put_page(page); + rcu_read_unlock(); + goto failout_up; + } + rcu_read_unlock(); + return 0; + +failout_up: + mmap_read_unlock(mm); + return err; +} + +/* + * What kind of VMA is considered ? + */ +static inline int vma_can_enter(struct vm_area_struct *vma) +{ + return uksm_flags_can_scan(vma->vm_flags); +} + +/* + * Called whenever a fresh new vma is created A new vma_slot. + * is created and inserted into a global list Must be called. + * after vma is inserted to its mm. + */ +void uksm_vma_add_new(struct vm_area_struct *vma) +{ + struct vma_slot *slot; + + if (!vma_can_enter(vma)) { + vma->uksm_vma_slot = NULL; + return; + } + + slot = alloc_vma_slot(); + if (!slot) { + vma->uksm_vma_slot = NULL; + return; + } + + vma->uksm_vma_slot = slot; + vma->vm_flags |= VM_MERGEABLE; + slot->vma = vma; + slot->mm = vma->vm_mm; + slot->ctime_j = jiffies; + slot->pages = vma_pages(vma); + spin_lock(&vma_slot_list_lock); + list_add_tail(&slot->slot_list, &vma_slot_new); + spin_unlock(&vma_slot_list_lock); +} + +/* 32/3 < they < 32/2 */ +#define shiftl 8 +#define shiftr 12 + +#define HASH_FROM_TO(from, to) \ +for (index = from; index < to; index++) { \ + pos = random_nums[index]; \ + hash += key[pos]; \ + hash += (hash << shiftl); \ + hash ^= (hash >> shiftr); \ +} + + +#define HASH_FROM_DOWN_TO(from, to) \ +for (index = from - 1; index >= to; index--) { \ + hash ^= (hash >> shiftr); \ + hash ^= (hash >> (shiftr*2)); \ + hash -= (hash << shiftl); \ + hash += (hash << (shiftl*2)); \ + pos = random_nums[index]; \ + hash -= key[pos]; \ +} + +/* + * The main random sample hash function. + */ +static u32 random_sample_hash(void *addr, u32 hash_strength) +{ + u32 hash = 0xdeadbeef; + int index, pos, loop = hash_strength; + u32 *key = (u32 *)addr; + + if (loop > HASH_STRENGTH_FULL) + loop = HASH_STRENGTH_FULL; + + HASH_FROM_TO(0, loop); + + if (hash_strength > HASH_STRENGTH_FULL) { + loop = hash_strength - HASH_STRENGTH_FULL; + HASH_FROM_TO(0, loop); + } + + return hash; +} + + +/** + * It's used when hash strength is adjusted + * + * @addr The page's virtual address + * @from The original hash strength + * @to The hash strength changed to + * @hash The hash value generated with "from" hash value + * + * return the hash value + */ +static u32 delta_hash(void *addr, int from, int to, u32 hash) +{ + u32 *key = (u32 *)addr; + int index, pos; /* make sure they are int type */ + + if (to > from) { + if (from >= HASH_STRENGTH_FULL) { + from -= HASH_STRENGTH_FULL; + to -= HASH_STRENGTH_FULL; + HASH_FROM_TO(from, to); + } else if (to <= HASH_STRENGTH_FULL) { + HASH_FROM_TO(from, to); + } else { + HASH_FROM_TO(from, HASH_STRENGTH_FULL); + HASH_FROM_TO(0, to - HASH_STRENGTH_FULL); + } + } else { + if (from <= HASH_STRENGTH_FULL) { + HASH_FROM_DOWN_TO(from, to); + } else if (to >= HASH_STRENGTH_FULL) { + from -= HASH_STRENGTH_FULL; + to -= HASH_STRENGTH_FULL; + HASH_FROM_DOWN_TO(from, to); + } else { + HASH_FROM_DOWN_TO(from - HASH_STRENGTH_FULL, 0); + HASH_FROM_DOWN_TO(HASH_STRENGTH_FULL, to); + } + } + + return hash; +} + +/** + * + * Called when: rshash_pos or rshash_neg is about to overflow or a scan round + * has finished. + * + * return 0 if no page has been scanned since last call, 1 otherwise. + */ +static inline int encode_benefit(void) +{ + u64 scanned_delta, pos_delta, neg_delta; + unsigned long base = benefit.base; + + scanned_delta = uksm_pages_scanned - uksm_pages_scanned_last; + + if (!scanned_delta) + return 0; + + scanned_delta >>= base; + pos_delta = rshash_pos >> base; + neg_delta = rshash_neg >> base; + + if (CAN_OVERFLOW_U64(benefit.pos, pos_delta) || + CAN_OVERFLOW_U64(benefit.neg, neg_delta) || + CAN_OVERFLOW_U64(benefit.scanned, scanned_delta)) { + benefit.scanned >>= 1; + benefit.neg >>= 1; + benefit.pos >>= 1; + benefit.base++; + scanned_delta >>= 1; + pos_delta >>= 1; + neg_delta >>= 1; + } + + benefit.pos += pos_delta; + benefit.neg += neg_delta; + benefit.scanned += scanned_delta; + + BUG_ON(!benefit.scanned); + + rshash_pos = rshash_neg = 0; + uksm_pages_scanned_last = uksm_pages_scanned; + + return 1; +} + +static inline void reset_benefit(void) +{ + benefit.pos = 0; + benefit.neg = 0; + benefit.base = 0; + benefit.scanned = 0; +} + +static inline void inc_rshash_pos(unsigned long delta) +{ + if (CAN_OVERFLOW_U64(rshash_pos, delta)) + encode_benefit(); + + rshash_pos += delta; +} + +static inline void inc_rshash_neg(unsigned long delta) +{ + if (CAN_OVERFLOW_U64(rshash_neg, delta)) + encode_benefit(); + + rshash_neg += delta; +} + + +static inline u32 page_hash(struct page *page, unsigned long hash_strength, + int cost_accounting) +{ + u32 val; + unsigned long delta; + + void *addr = kmap_atomic(page); + + val = random_sample_hash(addr, hash_strength); + kunmap_atomic(addr); + + if (cost_accounting) { + if (hash_strength < HASH_STRENGTH_FULL) + delta = HASH_STRENGTH_FULL - hash_strength; + else + delta = 0; + + inc_rshash_pos(delta); + } + + return val; +} + +static int memcmp_pages_with_cost(struct page *page1, struct page *page2, + int cost_accounting) +{ + char *addr1, *addr2; + int ret; + + addr1 = kmap_atomic(page1); + addr2 = kmap_atomic(page2); + ret = memcmp(addr1, addr2, PAGE_SIZE); + kunmap_atomic(addr2); + kunmap_atomic(addr1); + + if (cost_accounting) + inc_rshash_neg(memcmp_cost); + + return ret; +} + +static inline int pages_identical_with_cost(struct page *page1, struct page *page2) +{ + return !memcmp_pages_with_cost(page1, page2, 0); +} + +static inline int is_page_full_zero(struct page *page) +{ + char *addr; + int ret; + + addr = kmap_atomic(page); + ret = is_full_zero(addr, PAGE_SIZE); + kunmap_atomic(addr); + + return ret; +} + +static int write_protect_page(struct vm_area_struct *vma, struct page *page, + pte_t *orig_pte, pte_t *old_pte) +{ + struct mm_struct *mm = vma->vm_mm; + struct page_vma_mapped_walk pvmw = { + .page = page, + .vma = vma, + }; + struct mmu_notifier_range range; + int swapped; + int err = -EFAULT; + + pvmw.address = page_address_in_vma(page, vma); + if (pvmw.address == -EFAULT) + goto out; + + BUG_ON(PageTransCompound(page)); + + mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, mm, pvmw.address, + pvmw.address + PAGE_SIZE); + mmu_notifier_invalidate_range_start(&range); + + if (!page_vma_mapped_walk(&pvmw)) + goto out_mn; + if (WARN_ONCE(!pvmw.pte, "Unexpected PMD mapping?")) + goto out_unlock; + + if (old_pte) + *old_pte = *pvmw.pte; + + if (pte_write(*pvmw.pte) || pte_dirty(*pvmw.pte) || + (pte_protnone(*pvmw.pte) && pte_savedwrite(*pvmw.pte)) || mm_tlb_flush_pending(mm)) { + pte_t entry; + + swapped = PageSwapCache(page); + flush_cache_page(vma, pvmw.address, page_to_pfn(page)); + /* + * Ok this is tricky, when get_user_pages_fast() run it doesn't + * take any lock, therefore the check that we are going to make + * with the pagecount against the mapcount is racey and + * O_DIRECT can happen right after the check. + * So we clear the pte and flush the tlb before the check + * this assure us that no O_DIRECT can happen after the check + * or in the middle of the check. + */ + entry = ptep_clear_flush_notify(vma, pvmw.address, pvmw.pte); + /* + * Check that no O_DIRECT or similar I/O is in progress on the + * page + */ + if (page_mapcount(page) + 1 + swapped != page_count(page)) { + set_pte_at(mm, pvmw.address, pvmw.pte, entry); + goto out_unlock; + } + if (pte_dirty(entry)) + set_page_dirty(page); + + if (pte_protnone(entry)) + entry = pte_mkclean(pte_clear_savedwrite(entry)); + else + entry = pte_mkclean(pte_wrprotect(entry)); + + set_pte_at_notify(mm, pvmw.address, pvmw.pte, entry); + } + *orig_pte = *pvmw.pte; + err = 0; + +out_unlock: + page_vma_mapped_walk_done(&pvmw); +out_mn: + mmu_notifier_invalidate_range_end(&range); +out: + return err; +} + +#define MERGE_ERR_PGERR 1 /* the page is invalid cannot continue */ +#define MERGE_ERR_COLLI 2 /* there is a collision */ +#define MERGE_ERR_COLLI_MAX 3 /* collision at the max hash strength */ +#define MERGE_ERR_CHANGED 4 /* the page has changed since last hash */ + + +/** + * replace_page - replace page in vma by new ksm page + * @vma: vma that holds the pte pointing to page + * @page: the page we are replacing by kpage + * @kpage: the ksm page we replace page by + * @orig_pte: the original value of the pte + * + * Returns 0 on success, MERGE_ERR_PGERR on failure. + */ +static int replace_page(struct vm_area_struct *vma, struct page *page, + struct page *kpage, pte_t orig_pte) +{ + struct mm_struct *mm = vma->vm_mm; + struct mmu_notifier_range range; + pgd_t *pgd; + p4d_t *p4d; + pud_t *pud; + pmd_t *pmd; + pte_t *ptep; + spinlock_t *ptl; + pte_t entry; + + unsigned long addr; + int err = MERGE_ERR_PGERR; + + addr = page_address_in_vma(page, vma); + if (addr == -EFAULT) + goto out; + + pgd = pgd_offset(mm, addr); + if (!pgd_present(*pgd)) + goto out; + + p4d = p4d_offset(pgd, addr); + pud = pud_offset(p4d, addr); + if (!pud_present(*pud)) + goto out; + + pmd = pmd_offset(pud, addr); + BUG_ON(pmd_trans_huge(*pmd)); + if (!pmd_present(*pmd)) + goto out; + + mmu_notifier_range_init(&range, MMU_NOTIFY_CLEAR, 0, vma, mm, addr, + addr + PAGE_SIZE); + mmu_notifier_invalidate_range_start(&range); + + ptep = pte_offset_map_lock(mm, pmd, addr, &ptl); + if (!pte_same(*ptep, orig_pte)) { + pte_unmap_unlock(ptep, ptl); + goto out_mn; + } + + flush_cache_page(vma, addr, pte_pfn(*ptep)); + ptep_clear_flush_notify(vma, addr, ptep); + entry = mk_pte(kpage, vma->vm_page_prot); + + /* special treatment is needed for zero_page */ + if ((page_to_pfn(kpage) == uksm_zero_pfn) || + (page_to_pfn(kpage) == zero_pfn)) { + entry = pte_mkspecial(entry); + dec_mm_counter(mm, MM_ANONPAGES); + inc_zone_page_state(page, NR_UKSM_ZERO_PAGES); + } else { + get_page(kpage); + page_add_anon_rmap(kpage, vma, addr, false); + } + + set_pte_at_notify(mm, addr, ptep, entry); + + page_remove_rmap(page, false); + if (!page_mapped(page)) + try_to_free_swap(page); + put_page(page); + + pte_unmap_unlock(ptep, ptl); + err = 0; +out_mn: + mmu_notifier_invalidate_range_end(&range); +out: + return err; +} + + +/** + * Fully hash a page with HASH_STRENGTH_MAX return a non-zero hash value. The + * zero hash value at HASH_STRENGTH_MAX is used to indicated that its + * hash_max member has not been calculated. + * + * @page The page needs to be hashed + * @hash_old The hash value calculated with current hash strength + * + * return the new hash value calculated at HASH_STRENGTH_MAX + */ +static inline u32 page_hash_max(struct page *page, u32 hash_old) +{ + u32 hash_max = 0; + void *addr; + + addr = kmap_atomic(page); + hash_max = delta_hash(addr, hash_strength, + HASH_STRENGTH_MAX, hash_old); + + kunmap_atomic(addr); + + if (!hash_max) + hash_max = 1; + + inc_rshash_neg(HASH_STRENGTH_MAX - hash_strength); + return hash_max; +} + +/* + * We compare the hash again, to ensure that it is really a hash collision + * instead of being caused by page write. + */ +static inline int check_collision(struct rmap_item *rmap_item, + u32 hash) +{ + int err; + struct page *page = rmap_item->page; + + /* if this rmap_item has already been hash_maxed, then the collision + * must appears in the second-level rbtree search. In this case we check + * if its hash_max value has been changed. Otherwise, the collision + * happens in the first-level rbtree search, so we check against it's + * current hash value. + */ + if (rmap_item->hash_max) { + inc_rshash_neg(memcmp_cost); + inc_rshash_neg(HASH_STRENGTH_MAX - hash_strength); + + if (rmap_item->hash_max == page_hash_max(page, hash)) + err = MERGE_ERR_COLLI; + else + err = MERGE_ERR_CHANGED; + } else { + inc_rshash_neg(memcmp_cost + hash_strength); + + if (page_hash(page, hash_strength, 0) == hash) + err = MERGE_ERR_COLLI; + else + err = MERGE_ERR_CHANGED; + } + + return err; +} + +/** + * Try to merge a rmap_item.page with a kpage in stable node. kpage must + * already be a ksm page. + * + * @return 0 if the pages were merged, -EFAULT otherwise. + */ +static int try_to_merge_with_uksm_page(struct rmap_item *rmap_item, + struct page *kpage, u32 hash) +{ + struct vm_area_struct *vma = rmap_item->slot->vma; + struct mm_struct *mm = vma->vm_mm; + pte_t orig_pte = __pte(0); + int err = MERGE_ERR_PGERR; + struct page *page; + + if (uksm_test_exit(mm)) + goto out; + + page = rmap_item->page; + + if (page == kpage) { /* ksm page forked */ + err = 0; + goto out; + } + + /* + * We need the page lock to read a stable PageSwapCache in + * write_protect_page(). We use trylock_page() instead of + * lock_page() because we don't want to wait here - we + * prefer to continue scanning and merging different pages, + * then come back to this page when it is unlocked. + */ + if (!trylock_page(page)) + goto out; + + if (!PageAnon(page) || !PageKsm(kpage)) + goto out_unlock; + + if (PageTransCompound(page)) { + err = split_huge_page(page); + if (err) + goto out_unlock; + } + + /* + * If this anonymous page is mapped only here, its pte may need + * to be write-protected. If it's mapped elsewhere, all of its + * ptes are necessarily already write-protected. But in either + * case, we need to lock and check page_count is not raised. + */ + if (write_protect_page(vma, page, &orig_pte, NULL) == 0) { + if (pages_identical_with_cost(page, kpage)) + err = replace_page(vma, page, kpage, orig_pte); + else + err = check_collision(rmap_item, hash); + } + + if ((vma->vm_flags & VM_LOCKED) && kpage && !err) { + munlock_vma_page(page); + if (!PageMlocked(kpage)) { + unlock_page(page); + lock_page(kpage); + mlock_vma_page(kpage); + page = kpage; /* for final unlock */ + } + } + +out_unlock: + unlock_page(page); +out: + return err; +} + + + +/** + * If two pages fail to merge in try_to_merge_two_pages, then we have a chance + * to restore a page mapping that has been changed in try_to_merge_two_pages. + * + * @return 0 on success. + */ +static int restore_uksm_page_pte(struct vm_area_struct *vma, unsigned long addr, + pte_t orig_pte, pte_t wprt_pte) +{ + struct mm_struct *mm = vma->vm_mm; + pgd_t *pgd; + p4d_t *p4d; + pud_t *pud; + pmd_t *pmd; + pte_t *ptep; + spinlock_t *ptl; + + int err = -EFAULT; + + pgd = pgd_offset(mm, addr); + if (!pgd_present(*pgd)) + goto out; + + p4d = p4d_offset(pgd, addr); + pud = pud_offset(p4d, addr); + if (!pud_present(*pud)) + goto out; + + pmd = pmd_offset(pud, addr); + if (!pmd_present(*pmd)) + goto out; + + ptep = pte_offset_map_lock(mm, pmd, addr, &ptl); + if (!pte_same(*ptep, wprt_pte)) { + /* already copied, let it be */ + pte_unmap_unlock(ptep, ptl); + goto out; + } + + /* + * Good boy, still here. When we still get the ksm page, it does not + * return to the free page pool, there is no way that a pte was changed + * to other page and gets back to this page. And remind that ksm page + * do not reuse in do_wp_page(). So it's safe to restore the original + * pte. + */ + flush_cache_page(vma, addr, pte_pfn(*ptep)); + ptep_clear_flush_notify(vma, addr, ptep); + set_pte_at_notify(mm, addr, ptep, orig_pte); + + pte_unmap_unlock(ptep, ptl); + err = 0; +out: + return err; +} + +/** + * try_to_merge_two_pages() - take two identical pages and prepare + * them to be merged into one page(rmap_item->page) + * + * @return 0 if we successfully merged two identical pages into + * one ksm page. MERGE_ERR_COLLI if it's only a hash collision + * search in rbtree. MERGE_ERR_CHANGED if rmap_item has been + * changed since it's hashed. MERGE_ERR_PGERR otherwise. + * + */ +static int try_to_merge_two_pages(struct rmap_item *rmap_item, + struct rmap_item *tree_rmap_item, + u32 hash) +{ + pte_t orig_pte1 = __pte(0), orig_pte2 = __pte(0); + pte_t wprt_pte1 = __pte(0), wprt_pte2 = __pte(0); + struct vm_area_struct *vma1 = rmap_item->slot->vma; + struct vm_area_struct *vma2 = tree_rmap_item->slot->vma; + struct page *page = rmap_item->page; + struct page *tree_page = tree_rmap_item->page; + int err = MERGE_ERR_PGERR; + struct address_space *saved_mapping; + + + if (rmap_item->page == tree_rmap_item->page) + goto out; + + if (!trylock_page(page)) + goto out; + + if (!PageAnon(page)) + goto out_unlock; + + if (PageTransCompound(page)) { + err = split_huge_page(page); + if (err) + goto out_unlock; + } + + if (write_protect_page(vma1, page, &wprt_pte1, &orig_pte1) != 0) { + unlock_page(page); + goto out; + } + + /* + * While we hold page lock, upgrade page from + * PageAnon+anon_vma to PageKsm+NULL stable_node: + * stable_tree_insert() will update stable_node. + */ + saved_mapping = page->mapping; + set_page_stable_node(page, NULL); + mark_page_accessed(page); + if (!PageDirty(page)) + SetPageDirty(page); + + unlock_page(page); + + if (!trylock_page(tree_page)) + goto restore_out; + + if (!PageAnon(tree_page)) { + unlock_page(tree_page); + goto restore_out; + } + + if (PageTransCompound(tree_page)) { + err = split_huge_page(tree_page); + if (err) { + unlock_page(tree_page); + goto restore_out; + } + } + + if (write_protect_page(vma2, tree_page, &wprt_pte2, &orig_pte2) != 0) { + unlock_page(tree_page); + goto restore_out; + } + + if (pages_identical_with_cost(page, tree_page)) { + err = replace_page(vma2, tree_page, page, wprt_pte2); + if (err) { + unlock_page(tree_page); + goto restore_out; + } + + if ((vma2->vm_flags & VM_LOCKED)) { + munlock_vma_page(tree_page); + if (!PageMlocked(page)) { + unlock_page(tree_page); + lock_page(page); + mlock_vma_page(page); + tree_page = page; /* for final unlock */ + } + } + + unlock_page(tree_page); + + goto out; /* success */ + + } else { + if (tree_rmap_item->hash_max && + tree_rmap_item->hash_max == rmap_item->hash_max) { + err = MERGE_ERR_COLLI_MAX; + } else if (page_hash(page, hash_strength, 0) == + page_hash(tree_page, hash_strength, 0)) { + inc_rshash_neg(memcmp_cost + hash_strength * 2); + err = MERGE_ERR_COLLI; + } else { + err = MERGE_ERR_CHANGED; + } + + unlock_page(tree_page); + } + +restore_out: + lock_page(page); + if (!restore_uksm_page_pte(vma1, get_rmap_addr(rmap_item), + orig_pte1, wprt_pte1)) + page->mapping = saved_mapping; + +out_unlock: + unlock_page(page); +out: + return err; +} + +static inline int hash_cmp(u32 new_val, u32 node_val) +{ + if (new_val > node_val) + return 1; + else if (new_val < node_val) + return -1; + else + return 0; +} + +static inline u32 rmap_item_hash_max(struct rmap_item *item, u32 hash) +{ + u32 hash_max = item->hash_max; + + if (!hash_max) { + hash_max = page_hash_max(item->page, hash); + + item->hash_max = hash_max; + } + + return hash_max; +} + + + +/** + * stable_tree_search() - search the stable tree for a page + * + * @item: the rmap_item we are comparing with + * @hash: the hash value of this item->page already calculated + * + * @return the page we have found, NULL otherwise. The page returned has + * been gotten. + */ +static struct page *stable_tree_search(struct rmap_item *item, u32 hash) +{ + struct rb_node *node = root_stable_treep->rb_node; + struct tree_node *tree_node; + unsigned long hash_max; + struct page *page = item->page; + struct stable_node *stable_node; + + stable_node = page_stable_node(page); + if (stable_node) { + /* ksm page forked, that is + * if (PageKsm(page) && !in_stable_tree(rmap_item)) + * it's actually gotten once outside. + */ + get_page(page); + return page; + } + + while (node) { + int cmp; + + tree_node = rb_entry(node, struct tree_node, node); + + cmp = hash_cmp(hash, tree_node->hash); + + if (cmp < 0) + node = node->rb_left; + else if (cmp > 0) + node = node->rb_right; + else + break; + } + + if (!node) + return NULL; + + if (tree_node->count == 1) { + stable_node = rb_entry(tree_node->sub_root.rb_node, + struct stable_node, node); + BUG_ON(!stable_node); + + goto get_page_out; + } + + /* + * ok, we have to search the second + * level subtree, hash the page to a + * full strength. + */ + node = tree_node->sub_root.rb_node; + BUG_ON(!node); + hash_max = rmap_item_hash_max(item, hash); + + while (node) { + int cmp; + + stable_node = rb_entry(node, struct stable_node, node); + + cmp = hash_cmp(hash_max, stable_node->hash_max); + + if (cmp < 0) + node = node->rb_left; + else if (cmp > 0) + node = node->rb_right; + else + goto get_page_out; + } + + return NULL; + +get_page_out: + page = get_uksm_page(stable_node, 1, 1); + return page; +} + +static int try_merge_rmap_item(struct rmap_item *item, + struct page *kpage, + struct page *tree_page) +{ + struct vm_area_struct *vma = item->slot->vma; + struct page_vma_mapped_walk pvmw = { + .page = kpage, + .vma = vma, + }; + + pvmw.address = get_rmap_addr(item); + if (!page_vma_mapped_walk(&pvmw)) + return 0; + + if (pte_write(*pvmw.pte)) { + /* has changed, abort! */ + page_vma_mapped_walk_done(&pvmw); + return 0; + } + + get_page(tree_page); + page_add_anon_rmap(tree_page, vma, pvmw.address, false); + + flush_cache_page(vma, pvmw.address, page_to_pfn(kpage)); + ptep_clear_flush_notify(vma, pvmw.address, pvmw.pte); + set_pte_at_notify(vma->vm_mm, pvmw.address, pvmw.pte, + mk_pte(tree_page, vma->vm_page_prot)); + + page_remove_rmap(kpage, false); + put_page(kpage); + + page_vma_mapped_walk_done(&pvmw); + + return 1; +} + +/** + * try_to_merge_with_stable_page() - when two rmap_items need to be inserted + * into stable tree, the page was found to be identical to a stable ksm page, + * this is the last chance we can merge them into one. + * + * @item1: the rmap_item holding the page which we wanted to insert + * into stable tree. + * @item2: the other rmap_item we found when unstable tree search + * @oldpage: the page currently mapped by the two rmap_items + * @tree_page: the page we found identical in stable tree node + * @success1: return if item1 is successfully merged + * @success2: return if item2 is successfully merged + */ +static void try_merge_with_stable(struct rmap_item *item1, + struct rmap_item *item2, + struct page **kpage, + struct page *tree_page, + int *success1, int *success2) +{ + struct vm_area_struct *vma1 = item1->slot->vma; + struct vm_area_struct *vma2 = item2->slot->vma; + *success1 = 0; + *success2 = 0; + + if (unlikely(*kpage == tree_page)) { + /* I don't think this can really happen */ + pr_warn("UKSM: unexpected condition detected in " + "%s -- *kpage == tree_page !\n", __func__); + *success1 = 1; + *success2 = 1; + return; + } + + if (!PageAnon(*kpage) || !PageKsm(*kpage)) + goto failed; + + if (!trylock_page(tree_page)) + goto failed; + + /* If the oldpage is still ksm and still pointed + * to in the right place, and still write protected, + * we are confident it's not changed, no need to + * memcmp anymore. + * be ware, we cannot take nested pte locks, + * deadlock risk. + */ + if (!try_merge_rmap_item(item1, *kpage, tree_page)) + goto unlock_failed; + + /* ok, then vma2, remind that pte1 already set */ + if (!try_merge_rmap_item(item2, *kpage, tree_page)) + goto success_1; + + *success2 = 1; +success_1: + *success1 = 1; + + + if ((*success1 && vma1->vm_flags & VM_LOCKED) || + (*success2 && vma2->vm_flags & VM_LOCKED)) { + munlock_vma_page(*kpage); + if (!PageMlocked(tree_page)) + mlock_vma_page(tree_page); + } + + /* + * We do not need oldpage any more in the caller, so can break the lock + * now. + */ + unlock_page(*kpage); + *kpage = tree_page; /* Get unlocked outside. */ + return; + +unlock_failed: + unlock_page(tree_page); +failed: + return; +} + +static inline void stable_node_hash_max(struct stable_node *node, + struct page *page, u32 hash) +{ + u32 hash_max = node->hash_max; + + if (!hash_max) { + hash_max = page_hash_max(page, hash); + node->hash_max = hash_max; + } +} + +static inline +struct stable_node *new_stable_node(struct tree_node *tree_node, + struct page *kpage, u32 hash_max) +{ + struct stable_node *new_stable_node; + + new_stable_node = alloc_stable_node(); + if (!new_stable_node) + return NULL; + + new_stable_node->kpfn = page_to_pfn(kpage); + new_stable_node->hash_max = hash_max; + new_stable_node->tree_node = tree_node; + set_page_stable_node(kpage, new_stable_node); + + return new_stable_node; +} + +static inline +struct stable_node *first_level_insert(struct tree_node *tree_node, + struct rmap_item *rmap_item, + struct rmap_item *tree_rmap_item, + struct page **kpage, u32 hash, + int *success1, int *success2) +{ + int cmp; + struct page *tree_page; + u32 hash_max = 0; + struct stable_node *stable_node, *new_snode; + struct rb_node *parent = NULL, **new; + + /* this tree node contains no sub-tree yet */ + stable_node = rb_entry(tree_node->sub_root.rb_node, + struct stable_node, node); + + tree_page = get_uksm_page(stable_node, 1, 0); + if (tree_page) { + cmp = memcmp_pages_with_cost(*kpage, tree_page, 1); + if (!cmp) { + try_merge_with_stable(rmap_item, tree_rmap_item, kpage, + tree_page, success1, success2); + put_page(tree_page); + if (!*success1 && !*success2) + goto failed; + + return stable_node; + + } else { + /* + * collision in first level try to create a subtree. + * A new node need to be created. + */ + put_page(tree_page); + + stable_node_hash_max(stable_node, tree_page, + tree_node->hash); + hash_max = rmap_item_hash_max(rmap_item, hash); + cmp = hash_cmp(hash_max, stable_node->hash_max); + + parent = &stable_node->node; + if (cmp < 0) + new = &parent->rb_left; + else if (cmp > 0) + new = &parent->rb_right; + else + goto failed; + } + + } else { + /* the only stable_node deleted, we reuse its tree_node. + */ + parent = NULL; + new = &tree_node->sub_root.rb_node; + } + + new_snode = new_stable_node(tree_node, *kpage, hash_max); + if (!new_snode) + goto failed; + + rb_link_node(&new_snode->node, parent, new); + rb_insert_color(&new_snode->node, &tree_node->sub_root); + tree_node->count++; + *success1 = *success2 = 1; + + return new_snode; + +failed: + return NULL; +} + +static inline +struct stable_node *stable_subtree_insert(struct tree_node *tree_node, + struct rmap_item *rmap_item, + struct rmap_item *tree_rmap_item, + struct page **kpage, u32 hash, + int *success1, int *success2) +{ + struct page *tree_page; + u32 hash_max; + struct stable_node *stable_node, *new_snode; + struct rb_node *parent, **new; + +research: + parent = NULL; + new = &tree_node->sub_root.rb_node; + BUG_ON(!*new); + hash_max = rmap_item_hash_max(rmap_item, hash); + while (*new) { + int cmp; + + stable_node = rb_entry(*new, struct stable_node, node); + + cmp = hash_cmp(hash_max, stable_node->hash_max); + + if (cmp < 0) { + parent = *new; + new = &parent->rb_left; + } else if (cmp > 0) { + parent = *new; + new = &parent->rb_right; + } else { + tree_page = get_uksm_page(stable_node, 1, 0); + if (tree_page) { + cmp = memcmp_pages_with_cost(*kpage, tree_page, 1); + if (!cmp) { + try_merge_with_stable(rmap_item, + tree_rmap_item, kpage, + tree_page, success1, success2); + + put_page(tree_page); + if (!*success1 && !*success2) + goto failed; + /* + * successfully merged with a stable + * node + */ + return stable_node; + } else { + put_page(tree_page); + goto failed; + } + } else { + /* + * stable node may be deleted, + * and subtree maybe + * restructed, cannot + * continue, research it. + */ + if (tree_node->count) { + goto research; + } else { + /* reuse the tree node*/ + parent = NULL; + new = &tree_node->sub_root.rb_node; + } + } + } + } + + new_snode = new_stable_node(tree_node, *kpage, hash_max); + if (!new_snode) + goto failed; + + rb_link_node(&new_snode->node, parent, new); + rb_insert_color(&new_snode->node, &tree_node->sub_root); + tree_node->count++; + *success1 = *success2 = 1; + + return new_snode; + +failed: + return NULL; +} + + +/** + * stable_tree_insert() - try to insert a merged page in unstable tree to + * the stable tree + * + * @kpage: the page need to be inserted + * @hash: the current hash of this page + * @rmap_item: the rmap_item being scanned + * @tree_rmap_item: the rmap_item found on unstable tree + * @success1: return if rmap_item is merged + * @success2: return if tree_rmap_item is merged + * + * @return the stable_node on stable tree if at least one + * rmap_item is inserted into stable tree, NULL + * otherwise. + */ +static struct stable_node * +stable_tree_insert(struct page **kpage, u32 hash, + struct rmap_item *rmap_item, + struct rmap_item *tree_rmap_item, + int *success1, int *success2) +{ + struct rb_node **new = &root_stable_treep->rb_node; + struct rb_node *parent = NULL; + struct stable_node *stable_node; + struct tree_node *tree_node; + u32 hash_max = 0; + + *success1 = *success2 = 0; + + while (*new) { + int cmp; + + tree_node = rb_entry(*new, struct tree_node, node); + + cmp = hash_cmp(hash, tree_node->hash); + + if (cmp < 0) { + parent = *new; + new = &parent->rb_left; + } else if (cmp > 0) { + parent = *new; + new = &parent->rb_right; + } else + break; + } + + if (*new) { + if (tree_node->count == 1) { + stable_node = first_level_insert(tree_node, rmap_item, + tree_rmap_item, kpage, + hash, success1, success2); + } else { + stable_node = stable_subtree_insert(tree_node, + rmap_item, tree_rmap_item, kpage, + hash, success1, success2); + } + } else { + + /* no tree node found */ + tree_node = alloc_tree_node(stable_tree_node_listp); + if (!tree_node) { + stable_node = NULL; + goto out; + } + + stable_node = new_stable_node(tree_node, *kpage, hash_max); + if (!stable_node) { + free_tree_node(tree_node); + goto out; + } + + tree_node->hash = hash; + rb_link_node(&tree_node->node, parent, new); + rb_insert_color(&tree_node->node, root_stable_treep); + parent = NULL; + new = &tree_node->sub_root.rb_node; + + rb_link_node(&stable_node->node, parent, new); + rb_insert_color(&stable_node->node, &tree_node->sub_root); + tree_node->count++; + *success1 = *success2 = 1; + } + +out: + return stable_node; +} + + +/** + * get_tree_rmap_item_page() - try to get the page and lock the mmap_sem + * + * @return 0 on success, -EBUSY if unable to lock the mmap_sem, + * -EINVAL if the page mapping has been changed. + */ +static inline int get_tree_rmap_item_page(struct rmap_item *tree_rmap_item) +{ + int err; + + err = get_mergeable_page_lock_mmap(tree_rmap_item); + + if (err == -EINVAL) { + /* its page map has been changed, remove it */ + remove_rmap_item_from_tree(tree_rmap_item); + } + + /* The page is gotten and mmap_sem is locked now. */ + return err; +} + + +/** + * unstable_tree_search_insert() - search an unstable tree rmap_item with the + * same hash value. Get its page and trylock the mmap_sem + */ +static inline +struct rmap_item *unstable_tree_search_insert(struct rmap_item *rmap_item, + u32 hash) + +{ + struct rb_node **new = &root_unstable_tree.rb_node; + struct rb_node *parent = NULL; + struct tree_node *tree_node; + u32 hash_max; + struct rmap_item *tree_rmap_item; + + while (*new) { + int cmp; + + tree_node = rb_entry(*new, struct tree_node, node); + + cmp = hash_cmp(hash, tree_node->hash); + + if (cmp < 0) { + parent = *new; + new = &parent->rb_left; + } else if (cmp > 0) { + parent = *new; + new = &parent->rb_right; + } else + break; + } + + if (*new) { + /* got the tree_node */ + if (tree_node->count == 1) { + tree_rmap_item = rb_entry(tree_node->sub_root.rb_node, + struct rmap_item, node); + BUG_ON(!tree_rmap_item); + + goto get_page_out; + } + + /* well, search the collision subtree */ + new = &tree_node->sub_root.rb_node; + BUG_ON(!*new); + hash_max = rmap_item_hash_max(rmap_item, hash); + + while (*new) { + int cmp; + + tree_rmap_item = rb_entry(*new, struct rmap_item, + node); + + cmp = hash_cmp(hash_max, tree_rmap_item->hash_max); + parent = *new; + if (cmp < 0) + new = &parent->rb_left; + else if (cmp > 0) + new = &parent->rb_right; + else + goto get_page_out; + } + } else { + /* alloc a new tree_node */ + tree_node = alloc_tree_node(&unstable_tree_node_list); + if (!tree_node) + return NULL; + + tree_node->hash = hash; + rb_link_node(&tree_node->node, parent, new); + rb_insert_color(&tree_node->node, &root_unstable_tree); + parent = NULL; + new = &tree_node->sub_root.rb_node; + } + + /* did not found even in sub-tree */ + rmap_item->tree_node = tree_node; + rmap_item->address |= UNSTABLE_FLAG; + rmap_item->hash_round = uksm_hash_round; + rb_link_node(&rmap_item->node, parent, new); + rb_insert_color(&rmap_item->node, &tree_node->sub_root); + + uksm_pages_unshared++; + return NULL; + +get_page_out: + if (tree_rmap_item->page == rmap_item->page) + return NULL; + + if (get_tree_rmap_item_page(tree_rmap_item)) + return NULL; + + return tree_rmap_item; +} + +static void hold_anon_vma(struct rmap_item *rmap_item, + struct anon_vma *anon_vma) +{ + rmap_item->anon_vma = anon_vma; + get_anon_vma(anon_vma); +} + + +/** + * stable_tree_append() - append a rmap_item to a stable node. Deduplication + * ratio statistics is done in this function. + * + */ +static void stable_tree_append(struct rmap_item *rmap_item, + struct stable_node *stable_node, int logdedup) +{ + struct node_vma *node_vma = NULL, *new_node_vma, *node_vma_cont = NULL; + unsigned long key = (unsigned long)rmap_item->slot; + unsigned long factor = rmap_item->slot->rung->step; + + BUG_ON(!stable_node); + rmap_item->address |= STABLE_FLAG; + + if (hlist_empty(&stable_node->hlist)) { + uksm_pages_shared++; + goto node_vma_new; + } else { + uksm_pages_sharing++; + } + + hlist_for_each_entry(node_vma, &stable_node->hlist, hlist) { + if (node_vma->key >= key) + break; + + if (logdedup) { + node_vma->slot->pages_bemerged += factor; + if (list_empty(&node_vma->slot->dedup_list)) + list_add(&node_vma->slot->dedup_list, + &vma_slot_dedup); + } + } + + if (node_vma) { + if (node_vma->key == key) { + node_vma_cont = hlist_entry_safe(node_vma->hlist.next, struct node_vma, hlist); + goto node_vma_ok; + } else if (node_vma->key > key) { + node_vma_cont = node_vma; + } + } + +node_vma_new: + /* no same vma already in node, alloc a new node_vma */ + new_node_vma = alloc_node_vma(); + BUG_ON(!new_node_vma); + new_node_vma->head = stable_node; + new_node_vma->slot = rmap_item->slot; + + if (!node_vma) { + hlist_add_head(&new_node_vma->hlist, &stable_node->hlist); + } else if (node_vma->key != key) { + if (node_vma->key < key) + hlist_add_behind(&new_node_vma->hlist, &node_vma->hlist); + else { + hlist_add_before(&new_node_vma->hlist, + &node_vma->hlist); + } + + } + node_vma = new_node_vma; + +node_vma_ok: /* ok, ready to add to the list */ + rmap_item->head = node_vma; + hlist_add_head(&rmap_item->hlist, &node_vma->rmap_hlist); + hold_anon_vma(rmap_item, rmap_item->slot->vma->anon_vma); + if (logdedup) { + rmap_item->slot->pages_merged++; + if (node_vma_cont) { + node_vma = node_vma_cont; + hlist_for_each_entry_continue(node_vma, hlist) { + node_vma->slot->pages_bemerged += factor; + if (list_empty(&node_vma->slot->dedup_list)) + list_add(&node_vma->slot->dedup_list, + &vma_slot_dedup); + } + } + } +} + +/* + * We use break_ksm to break COW on a ksm page: it's a stripped down + * + * if (get_user_pages(addr, 1, 1, 1, &page, NULL) == 1) + * put_page(page); + * + * but taking great care only to touch a ksm page, in a VM_MERGEABLE vma, + * in case the application has unmapped and remapped mm,addr meanwhile. + * Could a ksm page appear anywhere else? Actually yes, in a VM_PFNMAP + * mmap of /dev/mem or /dev/kmem, where we would not want to touch it. + */ +static int break_ksm(struct vm_area_struct *vma, unsigned long addr) +{ + struct page *page; + int ret = 0; + + do { + cond_resched(); + page = follow_page(vma, addr, FOLL_GET | FOLL_MIGRATION | FOLL_REMOTE); + if (IS_ERR_OR_NULL(page)) + break; + if (PageKsm(page)) { + ret = handle_mm_fault(vma, addr, + FAULT_FLAG_WRITE | FAULT_FLAG_REMOTE, + NULL); + } else + ret = VM_FAULT_WRITE; + put_page(page); + } while (!(ret & (VM_FAULT_WRITE | VM_FAULT_SIGBUS | VM_FAULT_SIGSEGV | VM_FAULT_OOM))); + /* + * We must loop because handle_mm_fault() may back out if there's + * any difficulty e.g. if pte accessed bit gets updated concurrently. + * + * VM_FAULT_WRITE is what we have been hoping for: it indicates that + * COW has been broken, even if the vma does not permit VM_WRITE; + * but note that a concurrent fault might break PageKsm for us. + * + * VM_FAULT_SIGBUS could occur if we race with truncation of the + * backing file, which also invalidates anonymous pages: that's + * okay, that truncation will have unmapped the PageKsm for us. + * + * VM_FAULT_OOM: at the time of writing (late July 2009), setting + * aside mem_cgroup limits, VM_FAULT_OOM would only be set if the + * current task has TIF_MEMDIE set, and will be OOM killed on return + * to user; and ksmd, having no mm, would never be chosen for that. + * + * But if the mm is in a limited mem_cgroup, then the fault may fail + * with VM_FAULT_OOM even if the current task is not TIF_MEMDIE; and + * even ksmd can fail in this way - though it's usually breaking ksm + * just to undo a merge it made a moment before, so unlikely to oom. + * + * That's a pity: we might therefore have more kernel pages allocated + * than we're counting as nodes in the stable tree; but uksm_do_scan + * will retry to break_cow on each pass, so should recover the page + * in due course. The important thing is to not let VM_MERGEABLE + * be cleared while any such pages might remain in the area. + */ + return (ret & VM_FAULT_OOM) ? -ENOMEM : 0; +} + +static void break_cow(struct rmap_item *rmap_item) +{ + struct vm_area_struct *vma = rmap_item->slot->vma; + struct mm_struct *mm = vma->vm_mm; + unsigned long addr = get_rmap_addr(rmap_item); + + if (uksm_test_exit(mm)) + goto out; + + break_ksm(vma, addr); +out: + return; +} + +/* + * Though it's very tempting to unmerge in_stable_tree(rmap_item)s rather + * than check every pte of a given vma, the locking doesn't quite work for + * that - an rmap_item is assigned to the stable tree after inserting ksm + * page and upping mmap_sem. Nor does it fit with the way we skip dup'ing + * rmap_items from parent to child at fork time (so as not to waste time + * if exit comes before the next scan reaches it). + * + * Similarly, although we'd like to remove rmap_items (so updating counts + * and freeing memory) when unmerging an area, it's easier to leave that + * to the next pass of ksmd - consider, for example, how ksmd might be + * in cmp_and_merge_page on one of the rmap_items we would be removing. + */ +inline int unmerge_uksm_pages(struct vm_area_struct *vma, + unsigned long start, unsigned long end) +{ + unsigned long addr; + int err = 0; + + for (addr = start; addr < end && !err; addr += PAGE_SIZE) { + if (uksm_test_exit(vma->vm_mm)) + break; + if (signal_pending(current)) + err = -ERESTARTSYS; + else + err = break_ksm(vma, addr); + } + return err; +} + +static inline void inc_uksm_pages_scanned(void) +{ + u64 delta; + + + if (uksm_pages_scanned == U64_MAX) { + encode_benefit(); + + delta = uksm_pages_scanned >> pages_scanned_base; + + if (CAN_OVERFLOW_U64(pages_scanned_stored, delta)) { + pages_scanned_stored >>= 1; + delta >>= 1; + pages_scanned_base++; + } + + pages_scanned_stored += delta; + + uksm_pages_scanned = uksm_pages_scanned_last = 0; + } + + uksm_pages_scanned++; +} + +static inline int find_zero_page_hash(int strength, u32 hash) +{ + return (zero_hash_table[strength] == hash); +} + +static +int cmp_and_merge_zero_page(struct vm_area_struct *vma, struct page *page) +{ + struct page *zero_page = empty_uksm_zero_page; + struct mm_struct *mm = vma->vm_mm; + pte_t orig_pte = __pte(0); + int err = -EFAULT; + + if (uksm_test_exit(mm)) + goto out; + + if (!trylock_page(page)) + goto out; + + if (!PageAnon(page)) + goto out_unlock; + + if (PageTransCompound(page)) { + err = split_huge_page(page); + if (err) + goto out_unlock; + } + + if (write_protect_page(vma, page, &orig_pte, 0) == 0) { + if (is_page_full_zero(page)) + err = replace_page(vma, page, zero_page, orig_pte); + } + +out_unlock: + unlock_page(page); +out: + return err; +} + +/* + * cmp_and_merge_page() - first see if page can be merged into the stable + * tree; if not, compare hash to previous and if it's the same, see if page + * can be inserted into the unstable tree, or merged with a page already there + * and both transferred to the stable tree. + * + * @page: the page that we are searching identical page to. + * @rmap_item: the reverse mapping into the virtual address of this page + */ +static void cmp_and_merge_page(struct rmap_item *rmap_item, u32 hash) +{ + struct rmap_item *tree_rmap_item; + struct page *page; + struct page *kpage = NULL; + u32 hash_max; + int err; + unsigned int success1, success2; + struct stable_node *snode; + int cmp; + struct rb_node *parent = NULL, **new; + + remove_rmap_item_from_tree(rmap_item); + page = rmap_item->page; + + /* We first start with searching the page inside the stable tree */ + kpage = stable_tree_search(rmap_item, hash); + if (kpage) { + err = try_to_merge_with_uksm_page(rmap_item, kpage, + hash); + if (!err) { + /* + * The page was successfully merged, add + * its rmap_item to the stable tree. + * page lock is needed because it's + * racing with try_to_unmap_ksm(), etc. + */ + lock_page(kpage); + snode = page_stable_node(kpage); + stable_tree_append(rmap_item, snode, 1); + unlock_page(kpage); + put_page(kpage); + return; /* success */ + } + put_page(kpage); + + /* + * if it's a collision and it has been search in sub-rbtree + * (hash_max != 0), we want to abort, because if it is + * successfully merged in unstable tree, the collision trends to + * happen again. + */ + if (err == MERGE_ERR_COLLI && rmap_item->hash_max) + return; + } + + tree_rmap_item = + unstable_tree_search_insert(rmap_item, hash); + if (tree_rmap_item) { + err = try_to_merge_two_pages(rmap_item, tree_rmap_item, hash); + /* + * As soon as we merge this page, we want to remove the + * rmap_item of the page we have merged with from the unstable + * tree, and insert it instead as new node in the stable tree. + */ + if (!err) { + kpage = page; + remove_rmap_item_from_tree(tree_rmap_item); + lock_page(kpage); + snode = stable_tree_insert(&kpage, hash, + rmap_item, tree_rmap_item, + &success1, &success2); + + /* + * Do not log dedup for tree item, it's not counted as + * scanned in this round. + */ + if (success2) + stable_tree_append(tree_rmap_item, snode, 0); + + /* + * The order of these two stable append is important: + * we are scanning rmap_item. + */ + if (success1) + stable_tree_append(rmap_item, snode, 1); + + /* + * The original kpage may be unlocked inside + * stable_tree_insert() already. This page + * should be unlocked before doing + * break_cow(). + */ + unlock_page(kpage); + + if (!success1) + break_cow(rmap_item); + + if (!success2) + break_cow(tree_rmap_item); + + } else if (err == MERGE_ERR_COLLI) { + BUG_ON(tree_rmap_item->tree_node->count > 1); + + rmap_item_hash_max(tree_rmap_item, + tree_rmap_item->tree_node->hash); + + hash_max = rmap_item_hash_max(rmap_item, hash); + cmp = hash_cmp(hash_max, tree_rmap_item->hash_max); + parent = &tree_rmap_item->node; + if (cmp < 0) + new = &parent->rb_left; + else if (cmp > 0) + new = &parent->rb_right; + else + goto put_up_out; + + rmap_item->tree_node = tree_rmap_item->tree_node; + rmap_item->address |= UNSTABLE_FLAG; + rmap_item->hash_round = uksm_hash_round; + rb_link_node(&rmap_item->node, parent, new); + rb_insert_color(&rmap_item->node, + &tree_rmap_item->tree_node->sub_root); + rmap_item->tree_node->count++; + } else { + /* + * either one of the page has changed or they collide + * at the max hash, we consider them as ill items. + */ + remove_rmap_item_from_tree(tree_rmap_item); + } +put_up_out: + put_page(tree_rmap_item->page); + mmap_read_unlock(tree_rmap_item->slot->vma->vm_mm); + } +} + + + + +static inline unsigned long get_pool_index(struct vma_slot *slot, + unsigned long index) +{ + unsigned long pool_index; + + pool_index = (sizeof(struct rmap_list_entry *) * index) >> PAGE_SHIFT; + if (pool_index >= slot->pool_size) + BUG(); + return pool_index; +} + +static inline unsigned long index_page_offset(unsigned long index) +{ + return offset_in_page(sizeof(struct rmap_list_entry *) * index); +} + +static inline +struct rmap_list_entry *get_rmap_list_entry(struct vma_slot *slot, + unsigned long index, int need_alloc) +{ + unsigned long pool_index; + struct page *page; + void *addr; + + + pool_index = get_pool_index(slot, index); + if (!slot->rmap_list_pool[pool_index]) { + if (!need_alloc) + return NULL; + + page = alloc_page(GFP_KERNEL | __GFP_ZERO | __GFP_NOWARN); + if (!page) + return NULL; + + slot->rmap_list_pool[pool_index] = page; + } + + addr = kmap(slot->rmap_list_pool[pool_index]); + addr += index_page_offset(index); + + return addr; +} + +static inline void put_rmap_list_entry(struct vma_slot *slot, + unsigned long index) +{ + unsigned long pool_index; + + pool_index = get_pool_index(slot, index); + BUG_ON(!slot->rmap_list_pool[pool_index]); + kunmap(slot->rmap_list_pool[pool_index]); +} + +static inline int entry_is_new(struct rmap_list_entry *entry) +{ + return !entry->item; +} + +static inline unsigned long get_index_orig_addr(struct vma_slot *slot, + unsigned long index) +{ + return slot->vma->vm_start + (index << PAGE_SHIFT); +} + +static inline unsigned long get_entry_address(struct rmap_list_entry *entry) +{ + unsigned long addr; + + if (is_addr(entry->addr)) + addr = get_clean_addr(entry->addr); + else if (entry->item) + addr = get_rmap_addr(entry->item); + else + BUG(); + + return addr; +} + +static inline struct rmap_item *get_entry_item(struct rmap_list_entry *entry) +{ + if (is_addr(entry->addr)) + return NULL; + + return entry->item; +} + +static inline void inc_rmap_list_pool_count(struct vma_slot *slot, + unsigned long index) +{ + unsigned long pool_index; + + pool_index = get_pool_index(slot, index); + BUG_ON(!slot->rmap_list_pool[pool_index]); + slot->pool_counts[pool_index]++; +} + +static inline void dec_rmap_list_pool_count(struct vma_slot *slot, + unsigned long index) +{ + unsigned long pool_index; + + pool_index = get_pool_index(slot, index); + BUG_ON(!slot->rmap_list_pool[pool_index]); + BUG_ON(!slot->pool_counts[pool_index]); + slot->pool_counts[pool_index]--; +} + +static inline int entry_has_rmap(struct rmap_list_entry *entry) +{ + return !is_addr(entry->addr) && entry->item; +} + +static inline void swap_entries(struct rmap_list_entry *entry1, + unsigned long index1, + struct rmap_list_entry *entry2, + unsigned long index2) +{ + struct rmap_list_entry tmp; + + /* swapping two new entries is meaningless */ + BUG_ON(entry_is_new(entry1) && entry_is_new(entry2)); + + tmp = *entry1; + *entry1 = *entry2; + *entry2 = tmp; + + if (entry_has_rmap(entry1)) + entry1->item->entry_index = index1; + + if (entry_has_rmap(entry2)) + entry2->item->entry_index = index2; + + if (entry_has_rmap(entry1) && !entry_has_rmap(entry2)) { + inc_rmap_list_pool_count(entry1->item->slot, index1); + dec_rmap_list_pool_count(entry1->item->slot, index2); + } else if (!entry_has_rmap(entry1) && entry_has_rmap(entry2)) { + inc_rmap_list_pool_count(entry2->item->slot, index2); + dec_rmap_list_pool_count(entry2->item->slot, index1); + } +} + +static inline void free_entry_item(struct rmap_list_entry *entry) +{ + unsigned long index; + struct rmap_item *item; + + if (!is_addr(entry->addr)) { + BUG_ON(!entry->item); + item = entry->item; + entry->addr = get_rmap_addr(item); + set_is_addr(entry->addr); + index = item->entry_index; + remove_rmap_item_from_tree(item); + dec_rmap_list_pool_count(item->slot, index); + free_rmap_item(item); + } +} + +static inline int pool_entry_boundary(unsigned long index) +{ + unsigned long linear_addr; + + linear_addr = sizeof(struct rmap_list_entry *) * index; + return index && !offset_in_page(linear_addr); +} + +static inline void try_free_last_pool(struct vma_slot *slot, + unsigned long index) +{ + unsigned long pool_index; + + pool_index = get_pool_index(slot, index); + if (slot->rmap_list_pool[pool_index] && + !slot->pool_counts[pool_index]) { + __free_page(slot->rmap_list_pool[pool_index]); + slot->rmap_list_pool[pool_index] = NULL; + slot->flags |= UKSM_SLOT_NEED_SORT; + } + +} + +static inline unsigned long vma_item_index(struct vm_area_struct *vma, + struct rmap_item *item) +{ + return (get_rmap_addr(item) - vma->vm_start) >> PAGE_SHIFT; +} + +static int within_same_pool(struct vma_slot *slot, + unsigned long i, unsigned long j) +{ + unsigned long pool_i, pool_j; + + pool_i = get_pool_index(slot, i); + pool_j = get_pool_index(slot, j); + + return (pool_i == pool_j); +} + +static void sort_rmap_entry_list(struct vma_slot *slot) +{ + unsigned long i, j; + struct rmap_list_entry *entry, *swap_entry; + + entry = get_rmap_list_entry(slot, 0, 0); + for (i = 0; i < slot->pages; ) { + + if (!entry) + goto skip_whole_pool; + + if (entry_is_new(entry)) + goto next_entry; + + if (is_addr(entry->addr)) { + entry->addr = 0; + goto next_entry; + } + + j = vma_item_index(slot->vma, entry->item); + if (j == i) + goto next_entry; + + if (within_same_pool(slot, i, j)) + swap_entry = entry + j - i; + else + swap_entry = get_rmap_list_entry(slot, j, 1); + + swap_entries(entry, i, swap_entry, j); + if (!within_same_pool(slot, i, j)) + put_rmap_list_entry(slot, j); + continue; + +skip_whole_pool: + i += PAGE_SIZE / sizeof(*entry); + if (i < slot->pages) + entry = get_rmap_list_entry(slot, i, 0); + continue; + +next_entry: + if (i >= slot->pages - 1 || + !within_same_pool(slot, i, i + 1)) { + put_rmap_list_entry(slot, i); + if (i + 1 < slot->pages) + entry = get_rmap_list_entry(slot, i + 1, 0); + } else + entry++; + i++; + continue; + } + + /* free empty pool entries which contain no rmap_item */ + /* CAN be simplied to based on only pool_counts when bug freed !!!!! */ + for (i = 0; i < slot->pool_size; i++) { + unsigned char has_rmap; + void *addr; + + if (!slot->rmap_list_pool[i]) + continue; + + has_rmap = 0; + addr = kmap(slot->rmap_list_pool[i]); + BUG_ON(!addr); + for (j = 0; j < PAGE_SIZE / sizeof(*entry); j++) { + entry = (struct rmap_list_entry *)addr + j; + if (is_addr(entry->addr)) + continue; + if (!entry->item) + continue; + has_rmap = 1; + } + kunmap(slot->rmap_list_pool[i]); + if (!has_rmap) { + BUG_ON(slot->pool_counts[i]); + __free_page(slot->rmap_list_pool[i]); + slot->rmap_list_pool[i] = NULL; + } + } + + slot->flags &= ~UKSM_SLOT_NEED_SORT; +} + +/* + * vma_fully_scanned() - if all the pages in this slot have been scanned. + */ +static inline int vma_fully_scanned(struct vma_slot *slot) +{ + return slot->pages_scanned == slot->pages; +} + +/** + * get_next_rmap_item() - Get the next rmap_item in a vma_slot according to + * its random permutation. This function is embedded with the random + * permutation index management code. + */ +static struct rmap_item *get_next_rmap_item(struct vma_slot *slot, u32 *hash) +{ + unsigned long rand_range, addr, swap_index, scan_index; + struct rmap_item *item = NULL; + struct rmap_list_entry *scan_entry, *swap_entry = NULL; + struct page *page; + + scan_index = swap_index = slot->pages_scanned % slot->pages; + + if (pool_entry_boundary(scan_index)) + try_free_last_pool(slot, scan_index - 1); + + if (vma_fully_scanned(slot)) { + if (slot->flags & UKSM_SLOT_NEED_SORT) + slot->flags |= UKSM_SLOT_NEED_RERAND; + else + slot->flags &= ~UKSM_SLOT_NEED_RERAND; + if (slot->flags & UKSM_SLOT_NEED_SORT) + sort_rmap_entry_list(slot); + } + + scan_entry = get_rmap_list_entry(slot, scan_index, 1); + if (!scan_entry) + return NULL; + + if (entry_is_new(scan_entry)) { + scan_entry->addr = get_index_orig_addr(slot, scan_index); + set_is_addr(scan_entry->addr); + } + + if (slot->flags & UKSM_SLOT_NEED_RERAND) { + rand_range = slot->pages - scan_index; + BUG_ON(!rand_range); + swap_index = scan_index + (prandom_u32() % rand_range); + } + + if (swap_index != scan_index) { + swap_entry = get_rmap_list_entry(slot, swap_index, 1); + + if (!swap_entry) + return NULL; + + if (entry_is_new(swap_entry)) { + swap_entry->addr = get_index_orig_addr(slot, + swap_index); + set_is_addr(swap_entry->addr); + } + swap_entries(scan_entry, scan_index, swap_entry, swap_index); + } + + addr = get_entry_address(scan_entry); + item = get_entry_item(scan_entry); + BUG_ON(addr > slot->vma->vm_end || addr < slot->vma->vm_start); + + page = follow_page(slot->vma, addr, FOLL_GET); + if (IS_ERR_OR_NULL(page)) + goto nopage; + + if (!PageAnon(page)) + goto putpage; + + /*check is zero_page pfn or uksm_zero_page*/ + if ((page_to_pfn(page) == zero_pfn) + || (page_to_pfn(page) == uksm_zero_pfn)) + goto putpage; + + flush_anon_page(slot->vma, page, addr); + flush_dcache_page(page); + + + *hash = page_hash(page, hash_strength, 1); + inc_uksm_pages_scanned(); + /*if the page content all zero, re-map to zero-page*/ + if (find_zero_page_hash(hash_strength, *hash)) { + if (!cmp_and_merge_zero_page(slot->vma, page)) { + slot->pages_merged++; + + /* For full-zero pages, no need to create rmap item */ + goto putpage; + } else { + inc_rshash_neg(memcmp_cost / 2); + } + } + + if (!item) { + item = alloc_rmap_item(); + if (item) { + /* It has already been zeroed */ + item->slot = slot; + item->address = addr; + item->entry_index = scan_index; + scan_entry->item = item; + inc_rmap_list_pool_count(slot, scan_index); + } else + goto putpage; + } + + BUG_ON(item->slot != slot); + /* the page may have changed */ + item->page = page; + put_rmap_list_entry(slot, scan_index); + if (swap_entry) + put_rmap_list_entry(slot, swap_index); + return item; + +putpage: + put_page(page); + page = NULL; +nopage: + /* no page, store addr back and free rmap_item if possible */ + free_entry_item(scan_entry); + put_rmap_list_entry(slot, scan_index); + if (swap_entry) + put_rmap_list_entry(slot, swap_index); + return NULL; +} + +static inline int in_stable_tree(struct rmap_item *rmap_item) +{ + return rmap_item->address & STABLE_FLAG; +} + +/** + * scan_vma_one_page() - scan the next page in a vma_slot. Called with + * mmap_sem locked. + */ +static noinline void scan_vma_one_page(struct vma_slot *slot) +{ + u32 hash; + struct mm_struct *mm; + struct rmap_item *rmap_item = NULL; + struct vm_area_struct *vma = slot->vma; + + mm = vma->vm_mm; + BUG_ON(!mm); + BUG_ON(!slot); + + rmap_item = get_next_rmap_item(slot, &hash); + if (!rmap_item) + goto out1; + + if (PageKsm(rmap_item->page) && in_stable_tree(rmap_item)) + goto out2; + + cmp_and_merge_page(rmap_item, hash); +out2: + put_page(rmap_item->page); +out1: + slot->pages_scanned++; + slot->this_sampled++; + if (slot->fully_scanned_round != fully_scanned_round) + scanned_virtual_pages++; + + if (vma_fully_scanned(slot)) + slot->fully_scanned_round = fully_scanned_round; +} + +static inline unsigned long rung_get_pages(struct scan_rung *rung) +{ + struct slot_tree_node *node; + + if (!rung->vma_root.rnode) + return 0; + + node = container_of(rung->vma_root.rnode, struct slot_tree_node, snode); + + return node->size; +} + +#define RUNG_SAMPLED_MIN 3 + +static inline +void uksm_calc_rung_step(struct scan_rung *rung, + unsigned long page_time, unsigned long ratio) +{ + unsigned long sampled, pages; + + /* will be fully scanned ? */ + if (!rung->cover_msecs) { + rung->step = 1; + return; + } + + sampled = rung->cover_msecs * (NSEC_PER_MSEC / TIME_RATIO_SCALE) + * ratio / page_time; + + /* + * Before we finsish a scan round and expensive per-round jobs, + * we need to have a chance to estimate the per page time. So + * the sampled number can not be too small. + */ + if (sampled < RUNG_SAMPLED_MIN) + sampled = RUNG_SAMPLED_MIN; + + pages = rung_get_pages(rung); + if (likely(pages > sampled)) + rung->step = pages / sampled; + else + rung->step = 1; +} + +static inline int step_need_recalc(struct scan_rung *rung) +{ + unsigned long pages, stepmax; + + pages = rung_get_pages(rung); + stepmax = pages / RUNG_SAMPLED_MIN; + + return pages && (rung->step > pages || + (stepmax && rung->step > stepmax)); +} + +static inline +void reset_current_scan(struct scan_rung *rung, int finished, int step_recalc) +{ + struct vma_slot *slot; + + if (finished) + rung->flags |= UKSM_RUNG_ROUND_FINISHED; + + if (step_recalc || step_need_recalc(rung)) { + uksm_calc_rung_step(rung, uksm_ema_page_time, rung->cpu_ratio); + BUG_ON(step_need_recalc(rung)); + } + + slot_iter_index = prandom_u32() % rung->step; + BUG_ON(!rung->vma_root.rnode); + slot = sradix_tree_next(&rung->vma_root, NULL, 0, slot_iter); + BUG_ON(!slot); + + rung->current_scan = slot; + rung->current_offset = slot_iter_index; +} + +static inline struct sradix_tree_root *slot_get_root(struct vma_slot *slot) +{ + return &slot->rung->vma_root; +} + +/* + * return if resetted. + */ +static int advance_current_scan(struct scan_rung *rung) +{ + unsigned short n; + struct vma_slot *slot, *next = NULL; + + BUG_ON(!rung->vma_root.num); + + slot = rung->current_scan; + n = (slot->pages - rung->current_offset) % rung->step; + slot_iter_index = rung->step - n; + next = sradix_tree_next(&rung->vma_root, slot->snode, + slot->sindex, slot_iter); + + if (next) { + rung->current_offset = slot_iter_index; + rung->current_scan = next; + return 0; + } else { + reset_current_scan(rung, 1, 0); + return 1; + } +} + +static inline void rung_rm_slot(struct vma_slot *slot) +{ + struct scan_rung *rung = slot->rung; + struct sradix_tree_root *root; + + if (rung->current_scan == slot) + advance_current_scan(rung); + + root = slot_get_root(slot); + sradix_tree_delete_from_leaf(root, slot->snode, slot->sindex); + slot->snode = NULL; + if (step_need_recalc(rung)) { + uksm_calc_rung_step(rung, uksm_ema_page_time, rung->cpu_ratio); + BUG_ON(step_need_recalc(rung)); + } + + /* In case advance_current_scan loop back to this slot again */ + if (rung->vma_root.num && rung->current_scan == slot) + reset_current_scan(slot->rung, 1, 0); +} + +static inline void rung_add_new_slots(struct scan_rung *rung, + struct vma_slot **slots, unsigned long num) +{ + int err; + struct vma_slot *slot; + unsigned long i; + struct sradix_tree_root *root = &rung->vma_root; + + err = sradix_tree_enter(root, (void **)slots, num); + BUG_ON(err); + + for (i = 0; i < num; i++) { + slot = slots[i]; + slot->rung = rung; + BUG_ON(vma_fully_scanned(slot)); + } + + if (rung->vma_root.num == num) + reset_current_scan(rung, 0, 1); +} + +static inline int rung_add_one_slot(struct scan_rung *rung, + struct vma_slot *slot) +{ + int err; + + err = sradix_tree_enter(&rung->vma_root, (void **)&slot, 1); + if (err) + return err; + + slot->rung = rung; + if (rung->vma_root.num == 1) + reset_current_scan(rung, 0, 1); + + return 0; +} + +/* + * Return true if the slot is deleted from its rung. + */ +static inline int vma_rung_enter(struct vma_slot *slot, struct scan_rung *rung) +{ + struct scan_rung *old_rung = slot->rung; + int err; + + if (old_rung == rung) + return 0; + + rung_rm_slot(slot); + err = rung_add_one_slot(rung, slot); + if (err) { + err = rung_add_one_slot(old_rung, slot); + WARN_ON(err); /* OOPS, badly OOM, we lost this slot */ + } + + return 1; +} + +static inline int vma_rung_up(struct vma_slot *slot) +{ + struct scan_rung *rung; + + rung = slot->rung; + if (slot->rung != &uksm_scan_ladder[SCAN_LADDER_SIZE-1]) + rung++; + + return vma_rung_enter(slot, rung); +} + +static inline int vma_rung_down(struct vma_slot *slot) +{ + struct scan_rung *rung; + + rung = slot->rung; + if (slot->rung != &uksm_scan_ladder[0]) + rung--; + + return vma_rung_enter(slot, rung); +} + +/** + * cal_dedup_ratio() - Calculate the deduplication ratio for this slot. + */ +static unsigned long cal_dedup_ratio(struct vma_slot *slot) +{ + unsigned long ret; + unsigned long pages; + + pages = slot->this_sampled; + if (!pages) + return 0; + + BUG_ON(slot->pages_scanned == slot->last_scanned); + + ret = slot->pages_merged; + + /* Thrashing area filtering */ + if (ret && uksm_thrash_threshold) { + if (slot->pages_cowed * 100 / slot->pages_merged + > uksm_thrash_threshold) { + ret = 0; + } else { + ret = slot->pages_merged - slot->pages_cowed; + } + } + + return ret * 100 / pages; +} + +/** + * cal_dedup_ratio() - Calculate the deduplication ratio for this slot. + */ +static unsigned long cal_dedup_ratio_old(struct vma_slot *slot) +{ + unsigned long ret; + unsigned long pages; + + pages = slot->pages; + if (!pages) + return 0; + + ret = slot->pages_bemerged; + + /* Thrashing area filtering */ + if (ret && uksm_thrash_threshold) { + if (slot->pages_cowed * 100 / slot->pages_bemerged + > uksm_thrash_threshold) { + ret = 0; + } else { + ret = slot->pages_bemerged - slot->pages_cowed; + } + } + + return ret * 100 / pages; +} + +/** + * stable_node_reinsert() - When the hash_strength has been adjusted, the + * stable tree need to be restructured, this is the function re-inserting the + * stable node. + */ +static inline void stable_node_reinsert(struct stable_node *new_node, + struct page *page, + struct rb_root *root_treep, + struct list_head *tree_node_listp, + u32 hash) +{ + struct rb_node **new = &root_treep->rb_node; + struct rb_node *parent = NULL; + struct stable_node *stable_node; + struct tree_node *tree_node; + struct page *tree_page; + int cmp; + + while (*new) { + int cmp; + + tree_node = rb_entry(*new, struct tree_node, node); + + cmp = hash_cmp(hash, tree_node->hash); + + if (cmp < 0) { + parent = *new; + new = &parent->rb_left; + } else if (cmp > 0) { + parent = *new; + new = &parent->rb_right; + } else + break; + } + + if (*new) { + /* find a stable tree node with same first level hash value */ + stable_node_hash_max(new_node, page, hash); + if (tree_node->count == 1) { + stable_node = rb_entry(tree_node->sub_root.rb_node, + struct stable_node, node); + tree_page = get_uksm_page(stable_node, 1, 0); + if (tree_page) { + stable_node_hash_max(stable_node, + tree_page, hash); + put_page(tree_page); + + /* prepare for stable node insertion */ + + cmp = hash_cmp(new_node->hash_max, + stable_node->hash_max); + parent = &stable_node->node; + if (cmp < 0) + new = &parent->rb_left; + else if (cmp > 0) + new = &parent->rb_right; + else + goto failed; + + goto add_node; + } else { + /* the only stable_node deleted, the tree node + * was not deleted. + */ + goto tree_node_reuse; + } + } + + /* well, search the collision subtree */ + new = &tree_node->sub_root.rb_node; + parent = NULL; + BUG_ON(!*new); + while (*new) { + int cmp; + + stable_node = rb_entry(*new, struct stable_node, node); + + cmp = hash_cmp(new_node->hash_max, + stable_node->hash_max); + + if (cmp < 0) { + parent = *new; + new = &parent->rb_left; + } else if (cmp > 0) { + parent = *new; + new = &parent->rb_right; + } else { + /* oh, no, still a collision */ + goto failed; + } + } + + goto add_node; + } + + /* no tree node found */ + tree_node = alloc_tree_node(tree_node_listp); + if (!tree_node) { + pr_err("UKSM: memory allocation error!\n"); + goto failed; + } else { + tree_node->hash = hash; + rb_link_node(&tree_node->node, parent, new); + rb_insert_color(&tree_node->node, root_treep); + +tree_node_reuse: + /* prepare for stable node insertion */ + parent = NULL; + new = &tree_node->sub_root.rb_node; + } + +add_node: + rb_link_node(&new_node->node, parent, new); + rb_insert_color(&new_node->node, &tree_node->sub_root); + new_node->tree_node = tree_node; + tree_node->count++; + return; + +failed: + /* This can only happen when two nodes have collided + * in two levels. + */ + new_node->tree_node = NULL; + return; +} + +static inline void free_all_tree_nodes(struct list_head *list) +{ + struct tree_node *node, *tmp; + + list_for_each_entry_safe(node, tmp, list, all_list) { + free_tree_node(node); + } +} + +/** + * stable_tree_delta_hash() - Delta hash the stable tree from previous hash + * strength to the current hash_strength. It re-structures the hole tree. + */ +static inline void stable_tree_delta_hash(u32 prev_hash_strength) +{ + struct stable_node *node, *tmp; + struct rb_root *root_new_treep; + struct list_head *new_tree_node_listp; + + stable_tree_index = (stable_tree_index + 1) % 2; + root_new_treep = &root_stable_tree[stable_tree_index]; + new_tree_node_listp = &stable_tree_node_list[stable_tree_index]; + *root_new_treep = RB_ROOT; + BUG_ON(!list_empty(new_tree_node_listp)); + + /* + * we need to be safe, the node could be removed by get_uksm_page() + */ + list_for_each_entry_safe(node, tmp, &stable_node_list, all_list) { + void *addr; + struct page *node_page; + u32 hash; + + /* + * We are completely re-structuring the stable nodes to a new + * stable tree. We don't want to touch the old tree unlinks and + * old tree_nodes. The old tree_nodes will be freed at once. + */ + node_page = get_uksm_page(node, 0, 0); + if (!node_page) + continue; + + if (node->tree_node) { + hash = node->tree_node->hash; + + addr = kmap_atomic(node_page); + + hash = delta_hash(addr, prev_hash_strength, + hash_strength, hash); + kunmap_atomic(addr); + } else { + /* + *it was not inserted to rbtree due to collision in last + *round scan. + */ + hash = page_hash(node_page, hash_strength, 0); + } + + stable_node_reinsert(node, node_page, root_new_treep, + new_tree_node_listp, hash); + put_page(node_page); + } + + root_stable_treep = root_new_treep; + free_all_tree_nodes(stable_tree_node_listp); + BUG_ON(!list_empty(stable_tree_node_listp)); + stable_tree_node_listp = new_tree_node_listp; +} + +static inline void inc_hash_strength(unsigned long delta) +{ + hash_strength += 1 << delta; + if (hash_strength > HASH_STRENGTH_MAX) + hash_strength = HASH_STRENGTH_MAX; +} + +static inline void dec_hash_strength(unsigned long delta) +{ + unsigned long change = 1 << delta; + + if (hash_strength <= change + 1) + hash_strength = 1; + else + hash_strength -= change; +} + +static inline void inc_hash_strength_delta(void) +{ + hash_strength_delta++; + if (hash_strength_delta > HASH_STRENGTH_DELTA_MAX) + hash_strength_delta = HASH_STRENGTH_DELTA_MAX; +} + +static inline unsigned long get_current_neg_ratio(void) +{ + u64 pos = benefit.pos; + u64 neg = benefit.neg; + + if (!neg) + return 0; + + if (!pos || neg > pos) + return 100; + + if (neg > div64_u64(U64_MAX, 100)) + pos = div64_u64(pos, 100); + else + neg *= 100; + + return div64_u64(neg, pos); +} + +static inline unsigned long get_current_benefit(void) +{ + u64 pos = benefit.pos; + u64 neg = benefit.neg; + u64 scanned = benefit.scanned; + + if (neg > pos) + return 0; + + return div64_u64((pos - neg), scanned); +} + +static inline int judge_rshash_direction(void) +{ + u64 current_neg_ratio, stable_benefit; + u64 current_benefit, delta = 0; + int ret = STILL; + + /* + * Try to probe a value after the boot, and in case the system + * are still for a long time. + */ + if ((fully_scanned_round & 0xFFULL) == 10) { + ret = OBSCURE; + goto out; + } + + current_neg_ratio = get_current_neg_ratio(); + + if (current_neg_ratio == 0) { + rshash_neg_cont_zero++; + if (rshash_neg_cont_zero > 2) + return GO_DOWN; + else + return STILL; + } + rshash_neg_cont_zero = 0; + + if (current_neg_ratio > 90) { + ret = GO_UP; + goto out; + } + + current_benefit = get_current_benefit(); + stable_benefit = rshash_state.stable_benefit; + + if (!stable_benefit) { + ret = OBSCURE; + goto out; + } + + if (current_benefit > stable_benefit) + delta = current_benefit - stable_benefit; + else if (current_benefit < stable_benefit) + delta = stable_benefit - current_benefit; + + delta = div64_u64(100 * delta, stable_benefit); + + if (delta > 50) { + rshash_cont_obscure++; + if (rshash_cont_obscure > 2) + return OBSCURE; + else + return STILL; + } + +out: + rshash_cont_obscure = 0; + return ret; +} + +/** + * rshash_adjust() - The main function to control the random sampling state + * machine for hash strength adapting. + * + * return true if hash_strength has changed. + */ +static inline int rshash_adjust(void) +{ + unsigned long prev_hash_strength = hash_strength; + + if (!encode_benefit()) + return 0; + + switch (rshash_state.state) { + case RSHASH_STILL: + switch (judge_rshash_direction()) { + case GO_UP: + if (rshash_state.pre_direct == GO_DOWN) + hash_strength_delta = 0; + + inc_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + rshash_state.stable_benefit = get_current_benefit(); + rshash_state.pre_direct = GO_UP; + break; + + case GO_DOWN: + if (rshash_state.pre_direct == GO_UP) + hash_strength_delta = 0; + + dec_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + rshash_state.stable_benefit = get_current_benefit(); + rshash_state.pre_direct = GO_DOWN; + break; + + case OBSCURE: + rshash_state.stable_point = hash_strength; + rshash_state.turn_point_down = hash_strength; + rshash_state.turn_point_up = hash_strength; + rshash_state.turn_benefit_down = get_current_benefit(); + rshash_state.turn_benefit_up = get_current_benefit(); + rshash_state.lookup_window_index = 0; + rshash_state.state = RSHASH_TRYDOWN; + dec_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + break; + + case STILL: + break; + default: + BUG(); + } + break; + + case RSHASH_TRYDOWN: + if (rshash_state.lookup_window_index++ % 5 == 0) + rshash_state.below_count = 0; + + if (get_current_benefit() < rshash_state.stable_benefit) + rshash_state.below_count++; + else if (get_current_benefit() > + rshash_state.turn_benefit_down) { + rshash_state.turn_point_down = hash_strength; + rshash_state.turn_benefit_down = get_current_benefit(); + } + + if (rshash_state.below_count >= 3 || + judge_rshash_direction() == GO_UP || + hash_strength == 1) { + hash_strength = rshash_state.stable_point; + hash_strength_delta = 0; + inc_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + rshash_state.lookup_window_index = 0; + rshash_state.state = RSHASH_TRYUP; + hash_strength_delta = 0; + } else { + dec_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + } + break; + + case RSHASH_TRYUP: + if (rshash_state.lookup_window_index++ % 5 == 0) + rshash_state.below_count = 0; + + if (get_current_benefit() < rshash_state.turn_benefit_down) + rshash_state.below_count++; + else if (get_current_benefit() > rshash_state.turn_benefit_up) { + rshash_state.turn_point_up = hash_strength; + rshash_state.turn_benefit_up = get_current_benefit(); + } + + if (rshash_state.below_count >= 3 || + judge_rshash_direction() == GO_DOWN || + hash_strength == HASH_STRENGTH_MAX) { + hash_strength = rshash_state.turn_benefit_up > + rshash_state.turn_benefit_down ? + rshash_state.turn_point_up : + rshash_state.turn_point_down; + + rshash_state.state = RSHASH_PRE_STILL; + } else { + inc_hash_strength(hash_strength_delta); + inc_hash_strength_delta(); + } + + break; + + case RSHASH_NEW: + case RSHASH_PRE_STILL: + rshash_state.stable_benefit = get_current_benefit(); + rshash_state.state = RSHASH_STILL; + hash_strength_delta = 0; + break; + default: + BUG(); + } + + /* rshash_neg = rshash_pos = 0; */ + reset_benefit(); + + if (prev_hash_strength != hash_strength) + stable_tree_delta_hash(prev_hash_strength); + + return prev_hash_strength != hash_strength; +} + +/** + * round_update_ladder() - The main function to do update of all the + * adjustments whenever a scan round is finished. + */ +static noinline void round_update_ladder(void) +{ + int i; + unsigned long dedup; + struct vma_slot *slot, *tmp_slot; + + for (i = 0; i < SCAN_LADDER_SIZE; i++) + uksm_scan_ladder[i].flags &= ~UKSM_RUNG_ROUND_FINISHED; + + list_for_each_entry_safe(slot, tmp_slot, &vma_slot_dedup, dedup_list) { + + /* slot may be rung_rm_slot() when mm exits */ + if (slot->snode) { + dedup = cal_dedup_ratio_old(slot); + if (dedup && dedup >= uksm_abundant_threshold) + vma_rung_up(slot); + } + + slot->pages_bemerged = 0; + slot->pages_cowed = 0; + + list_del_init(&slot->dedup_list); + } +} + +static void uksm_del_vma_slot(struct vma_slot *slot) +{ + int i, j; + struct rmap_list_entry *entry; + + if (slot->snode) { + /* + * In case it just failed when entering the rung, it's not + * necessary. + */ + rung_rm_slot(slot); + } + + if (!list_empty(&slot->dedup_list)) + list_del(&slot->dedup_list); + + if (!slot->rmap_list_pool || !slot->pool_counts) { + /* In case it OOMed in uksm_vma_enter() */ + goto out; + } + + for (i = 0; i < slot->pool_size; i++) { + void *addr; + + if (!slot->rmap_list_pool[i]) + continue; + + addr = kmap(slot->rmap_list_pool[i]); + for (j = 0; j < PAGE_SIZE / sizeof(*entry); j++) { + entry = (struct rmap_list_entry *)addr + j; + if (is_addr(entry->addr)) + continue; + if (!entry->item) + continue; + + remove_rmap_item_from_tree(entry->item); + free_rmap_item(entry->item); + slot->pool_counts[i]--; + } + BUG_ON(slot->pool_counts[i]); + kunmap(slot->rmap_list_pool[i]); + __free_page(slot->rmap_list_pool[i]); + } + kfree(slot->rmap_list_pool); + kfree(slot->pool_counts); + +out: + slot->rung = NULL; + if (slot->flags & UKSM_SLOT_IN_UKSM) { + BUG_ON(uksm_pages_total < slot->pages); + uksm_pages_total -= slot->pages; + } + + if (slot->fully_scanned_round == fully_scanned_round) + scanned_virtual_pages -= slot->pages; + else + scanned_virtual_pages -= slot->pages_scanned; + free_vma_slot(slot); +} + + +#define SPIN_LOCK_PERIOD 32 +static struct vma_slot *cleanup_slots[SPIN_LOCK_PERIOD]; +static inline void cleanup_vma_slots(void) +{ + struct vma_slot *slot; + int i; + + i = 0; + spin_lock(&vma_slot_list_lock); + while (!list_empty(&vma_slot_del)) { + slot = list_entry(vma_slot_del.next, + struct vma_slot, slot_list); + list_del(&slot->slot_list); + cleanup_slots[i++] = slot; + if (i == SPIN_LOCK_PERIOD) { + spin_unlock(&vma_slot_list_lock); + while (--i >= 0) + uksm_del_vma_slot(cleanup_slots[i]); + i = 0; + spin_lock(&vma_slot_list_lock); + } + } + spin_unlock(&vma_slot_list_lock); + + while (--i >= 0) + uksm_del_vma_slot(cleanup_slots[i]); +} + +/* + * Expotional moving average formula + */ +static inline unsigned long ema(unsigned long curr, unsigned long last_ema) +{ + /* + * For a very high burst, even the ema cannot work well, a false very + * high per-page time estimation can result in feedback in very high + * overhead of context switch and rung update -- this will then lead + * to higher per-paper time, this may not converge. + * + * Instead, we try to approach this value in a binary manner. + */ + if (curr > last_ema * 10) + return last_ema * 2; + + return (EMA_ALPHA * curr + (100 - EMA_ALPHA) * last_ema) / 100; +} + +/* + * convert cpu ratio in 1/TIME_RATIO_SCALE configured by user to + * nanoseconds based on current uksm_sleep_jiffies. + */ +static inline unsigned long cpu_ratio_to_nsec(unsigned int ratio) +{ + return NSEC_PER_USEC * jiffies_to_usecs(uksm_sleep_jiffies) / + (TIME_RATIO_SCALE - ratio) * ratio; +} + + +static inline unsigned long rung_real_ratio(int cpu_time_ratio) +{ + unsigned long ret; + + BUG_ON(!cpu_time_ratio); + + if (cpu_time_ratio > 0) + ret = cpu_time_ratio; + else + ret = (unsigned long)(-cpu_time_ratio) * + uksm_max_cpu_percentage / 100UL; + + return ret ? ret : 1; +} + +static noinline void uksm_calc_scan_pages(void) +{ + struct scan_rung *ladder = uksm_scan_ladder; + unsigned long sleep_usecs, nsecs; + unsigned long ratio; + int i; + unsigned long per_page; + + if (uksm_ema_page_time > 100000 || + (((unsigned long) uksm_eval_round & (256UL - 1)) == 0UL)) + uksm_ema_page_time = UKSM_PAGE_TIME_DEFAULT; + + per_page = uksm_ema_page_time; + BUG_ON(!per_page); + + /* + * For every 8 eval round, we try to probe a uksm_sleep_jiffies value + * based on saved user input. + */ + if (((unsigned long) uksm_eval_round & (8UL - 1)) == 0UL) + uksm_sleep_jiffies = uksm_sleep_saved; + + /* We require a rung scan at least 1 page in a period. */ + nsecs = per_page; + ratio = rung_real_ratio(ladder[0].cpu_ratio); + if (cpu_ratio_to_nsec(ratio) < nsecs) { + sleep_usecs = nsecs * (TIME_RATIO_SCALE - ratio) / ratio + / NSEC_PER_USEC; + uksm_sleep_jiffies = usecs_to_jiffies(sleep_usecs) + 1; + } + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + ratio = rung_real_ratio(ladder[i].cpu_ratio); + ladder[i].pages_to_scan = cpu_ratio_to_nsec(ratio) / + per_page; + BUG_ON(!ladder[i].pages_to_scan); + uksm_calc_rung_step(&ladder[i], per_page, ratio); + } +} + +/* + * From the scan time of this round (ns) to next expected min sleep time + * (ms), be careful of the possible overflows. ratio is taken from + * rung_real_ratio() + */ +static inline +unsigned int scan_time_to_sleep(unsigned long long scan_time, unsigned long ratio) +{ + scan_time >>= 20; /* to msec level now */ + BUG_ON(scan_time > (ULONG_MAX / TIME_RATIO_SCALE)); + + return (unsigned int) ((unsigned long) scan_time * + (TIME_RATIO_SCALE - ratio) / ratio); +} + +#define __round_mask(x, y) ((__typeof__(x))((y)-1)) +#define round_up(x, y) ((((x)-1) | __round_mask(x, y))+1) + +static void uksm_vma_enter(struct vma_slot **slots, unsigned long num) +{ + struct scan_rung *rung; + + rung = &uksm_scan_ladder[0]; + rung_add_new_slots(rung, slots, num); +} + +static struct vma_slot *batch_slots[SLOT_TREE_NODE_STORE_SIZE]; + +static void uksm_enter_all_slots(void) +{ + struct vma_slot *slot; + unsigned long index; + struct list_head empty_vma_list; + int i; + + i = 0; + index = 0; + INIT_LIST_HEAD(&empty_vma_list); + + spin_lock(&vma_slot_list_lock); + while (!list_empty(&vma_slot_new)) { + slot = list_entry(vma_slot_new.next, + struct vma_slot, slot_list); + + if (!slot->vma->anon_vma) { + list_move(&slot->slot_list, &empty_vma_list); + } else if (vma_can_enter(slot->vma)) { + batch_slots[index++] = slot; + list_del_init(&slot->slot_list); + } else { + list_move(&slot->slot_list, &vma_slot_noadd); + } + + if (++i == SPIN_LOCK_PERIOD || + (index && !(index % SLOT_TREE_NODE_STORE_SIZE))) { + spin_unlock(&vma_slot_list_lock); + + if (index && !(index % SLOT_TREE_NODE_STORE_SIZE)) { + uksm_vma_enter(batch_slots, index); + index = 0; + } + i = 0; + cond_resched(); + spin_lock(&vma_slot_list_lock); + } + } + + list_splice(&empty_vma_list, &vma_slot_new); + + spin_unlock(&vma_slot_list_lock); + + if (index) + uksm_vma_enter(batch_slots, index); + +} + +static inline int rung_round_finished(struct scan_rung *rung) +{ + return rung->flags & UKSM_RUNG_ROUND_FINISHED; +} + +static inline void judge_slot(struct vma_slot *slot) +{ + struct scan_rung *rung = slot->rung; + unsigned long dedup; + int deleted; + + dedup = cal_dedup_ratio(slot); + if (vma_fully_scanned(slot) && uksm_thrash_threshold) + deleted = vma_rung_enter(slot, &uksm_scan_ladder[0]); + else if (dedup && dedup >= uksm_abundant_threshold) + deleted = vma_rung_up(slot); + else + deleted = vma_rung_down(slot); + + slot->pages_merged = 0; + slot->pages_cowed = 0; + slot->this_sampled = 0; + + if (vma_fully_scanned(slot)) + slot->pages_scanned = 0; + + slot->last_scanned = slot->pages_scanned; + + /* If its deleted in above, then rung was already advanced. */ + if (!deleted) + advance_current_scan(rung); +} + + +static inline int hash_round_finished(void) +{ + if (scanned_virtual_pages > (uksm_pages_total >> 2)) { + scanned_virtual_pages = 0; + if (uksm_pages_scanned) + fully_scanned_round++; + + return 1; + } else { + return 0; + } +} + +#define UKSM_MMSEM_BATCH 5 +#define BUSY_RETRY 100 + +/** + * uksm_do_scan() - the main worker function. + */ +static noinline void uksm_do_scan(void) +{ + struct vma_slot *slot, *iter; + struct mm_struct *busy_mm; + unsigned char round_finished, all_rungs_emtpy; + int i, err, mmsem_batch; + unsigned long pcost; + long long delta_exec; + unsigned long vpages, max_cpu_ratio; + unsigned long long start_time, end_time, scan_time; + unsigned int expected_jiffies; + + might_sleep(); + + vpages = 0; + + start_time = task_sched_runtime(current); + max_cpu_ratio = 0; + mmsem_batch = 0; + + for (i = 0; i < SCAN_LADDER_SIZE;) { + struct scan_rung *rung = &uksm_scan_ladder[i]; + unsigned long ratio; + int busy_retry; + + if (!rung->pages_to_scan) { + i++; + continue; + } + + if (!rung->vma_root.num) { + rung->pages_to_scan = 0; + i++; + continue; + } + + ratio = rung_real_ratio(rung->cpu_ratio); + if (ratio > max_cpu_ratio) + max_cpu_ratio = ratio; + + busy_retry = BUSY_RETRY; + /* + * Do not consider rung_round_finished() here, just used up the + * rung->pages_to_scan quota. + */ + while (rung->pages_to_scan && rung->vma_root.num && + likely(!freezing(current))) { + int reset = 0; + + slot = rung->current_scan; + + BUG_ON(vma_fully_scanned(slot)); + + if (mmsem_batch) + err = 0; + else + err = try_down_read_slot_mmap_sem(slot); + + if (err == -ENOENT) { +rm_slot: + rung_rm_slot(slot); + continue; + } + + busy_mm = slot->mm; + + if (err == -EBUSY) { + /* skip other vmas on the same mm */ + do { + reset = advance_current_scan(rung); + iter = rung->current_scan; + busy_retry--; + if (iter->vma->vm_mm != busy_mm || + !busy_retry || reset) + break; + } while (1); + + if (iter->vma->vm_mm != busy_mm) { + continue; + } else { + /* scan round finsished */ + break; + } + } + + BUG_ON(!vma_can_enter(slot->vma)); + if (uksm_test_exit(slot->vma->vm_mm)) { + mmsem_batch = 0; + mmap_read_unlock(slot->vma->vm_mm); + goto rm_slot; + } + + if (mmsem_batch) + mmsem_batch--; + else + mmsem_batch = UKSM_MMSEM_BATCH; + + /* Ok, we have take the mmap_sem, ready to scan */ + scan_vma_one_page(slot); + rung->pages_to_scan--; + vpages++; + + if (rung->current_offset + rung->step > slot->pages - 1 + || vma_fully_scanned(slot)) { + mmap_read_unlock(slot->vma->vm_mm); + judge_slot(slot); + mmsem_batch = 0; + } else { + rung->current_offset += rung->step; + if (!mmsem_batch) + mmap_read_unlock(slot->vma->vm_mm); + } + + busy_retry = BUSY_RETRY; + cond_resched(); + } + + if (mmsem_batch) { + mmap_read_unlock(slot->vma->vm_mm); + mmsem_batch = 0; + } + + if (freezing(current)) + break; + + cond_resched(); + } + end_time = task_sched_runtime(current); + delta_exec = end_time - start_time; + + if (freezing(current)) + return; + + cleanup_vma_slots(); + uksm_enter_all_slots(); + + round_finished = 1; + all_rungs_emtpy = 1; + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + struct scan_rung *rung = &uksm_scan_ladder[i]; + + if (rung->vma_root.num) { + all_rungs_emtpy = 0; + if (!rung_round_finished(rung)) + round_finished = 0; + } + } + + if (all_rungs_emtpy) + round_finished = 0; + + if (round_finished) { + round_update_ladder(); + uksm_eval_round++; + + if (hash_round_finished() && rshash_adjust()) { + /* Reset the unstable root iff hash strength changed */ + uksm_hash_round++; + root_unstable_tree = RB_ROOT; + free_all_tree_nodes(&unstable_tree_node_list); + } + + /* + * A number of pages can hang around indefinitely on per-cpu + * pagevecs, raised page count preventing write_protect_page + * from merging them. Though it doesn't really matter much, + * it is puzzling to see some stuck in pages_volatile until + * other activity jostles them out, and they also prevented + * LTP's KSM test from succeeding deterministically; so drain + * them here (here rather than on entry to uksm_do_scan(), + * so we don't IPI too often when pages_to_scan is set low). + */ + lru_add_drain_all(); + } + + + if (vpages && delta_exec > 0) { + pcost = (unsigned long) delta_exec / vpages; + if (likely(uksm_ema_page_time)) + uksm_ema_page_time = ema(pcost, uksm_ema_page_time); + else + uksm_ema_page_time = pcost; + } + + uksm_calc_scan_pages(); + uksm_sleep_real = uksm_sleep_jiffies; + /* in case of radical cpu bursts, apply the upper bound */ + end_time = task_sched_runtime(current); + if (max_cpu_ratio && end_time > start_time) { + scan_time = end_time - start_time; + expected_jiffies = msecs_to_jiffies( + scan_time_to_sleep(scan_time, max_cpu_ratio)); + + if (expected_jiffies > uksm_sleep_real) + uksm_sleep_real = expected_jiffies; + + /* We have a 1 second up bound for responsiveness. */ + if (jiffies_to_msecs(uksm_sleep_real) > MSEC_PER_SEC) + uksm_sleep_real = msecs_to_jiffies(1000); + } + + return; +} + +static int ksmd_should_run(void) +{ + return uksm_run & UKSM_RUN_MERGE; +} + +static int uksm_scan_thread(void *nothing) +{ + set_freezable(); + set_user_nice(current, 5); + + while (!kthread_should_stop()) { + mutex_lock(&uksm_thread_mutex); + if (ksmd_should_run()) + uksm_do_scan(); + mutex_unlock(&uksm_thread_mutex); + + try_to_freeze(); + + if (ksmd_should_run()) { + schedule_timeout_interruptible(uksm_sleep_real); + uksm_sleep_times++; + } else { + wait_event_freezable(uksm_thread_wait, + ksmd_should_run() || kthread_should_stop()); + } + } + return 0; +} + +void rmap_walk_ksm(struct page *page, struct rmap_walk_control *rwc) +{ + struct stable_node *stable_node; + struct node_vma *node_vma; + struct rmap_item *rmap_item; + int search_new_forks = 0; + unsigned long address; + + VM_BUG_ON_PAGE(!PageKsm(page), page); + VM_BUG_ON_PAGE(!PageLocked(page), page); + + stable_node = page_stable_node(page); + if (!stable_node) + return; +again: + hlist_for_each_entry(node_vma, &stable_node->hlist, hlist) { + hlist_for_each_entry(rmap_item, &node_vma->rmap_hlist, hlist) { + struct anon_vma *anon_vma = rmap_item->anon_vma; + struct anon_vma_chain *vmac; + struct vm_area_struct *vma; + + cond_resched(); + anon_vma_lock_read(anon_vma); + anon_vma_interval_tree_foreach(vmac, &anon_vma->rb_root, + 0, ULONG_MAX) { + cond_resched(); + vma = vmac->vma; + address = get_rmap_addr(rmap_item); + + if (address < vma->vm_start || + address >= vma->vm_end) + continue; + + if ((rmap_item->slot->vma == vma) == + search_new_forks) + continue; + + if (rwc->invalid_vma && rwc->invalid_vma(vma, rwc->arg)) + continue; + + if (!rwc->rmap_one(page, vma, address, rwc->arg)) { + anon_vma_unlock_read(anon_vma); + return; + } + + if (rwc->done && rwc->done(page)) { + anon_vma_unlock_read(anon_vma); + return; + } + } + anon_vma_unlock_read(anon_vma); + } + } + if (!search_new_forks++) + goto again; +} + +#ifdef CONFIG_MIGRATION +/* Common ksm interface but may be specific to uksm */ +void ksm_migrate_page(struct page *newpage, struct page *oldpage) +{ + struct stable_node *stable_node; + + VM_BUG_ON_PAGE(!PageLocked(oldpage), oldpage); + VM_BUG_ON_PAGE(!PageLocked(newpage), newpage); + VM_BUG_ON(newpage->mapping != oldpage->mapping); + + stable_node = page_stable_node(newpage); + if (stable_node) { + VM_BUG_ON(stable_node->kpfn != page_to_pfn(oldpage)); + stable_node->kpfn = page_to_pfn(newpage); + /* + * newpage->mapping was set in advance; now we need smp_wmb() + * to make sure that the new stable_node->kpfn is visible + * to get_ksm_page() before it can see that oldpage->mapping + * has gone stale (or that PageSwapCache has been cleared). + */ + smp_wmb(); + set_page_stable_node(oldpage, NULL); + } +} +#endif /* CONFIG_MIGRATION */ + +#ifdef CONFIG_MEMORY_HOTREMOVE +static struct stable_node *uksm_check_stable_tree(unsigned long start_pfn, + unsigned long end_pfn) +{ + struct rb_node *node; + + for (node = rb_first(root_stable_treep); node; node = rb_next(node)) { + struct stable_node *stable_node; + + stable_node = rb_entry(node, struct stable_node, node); + if (stable_node->kpfn >= start_pfn && + stable_node->kpfn < end_pfn) + return stable_node; + } + return NULL; +} + +static int uksm_memory_callback(struct notifier_block *self, + unsigned long action, void *arg) +{ + struct memory_notify *mn = arg; + struct stable_node *stable_node; + + switch (action) { + case MEM_GOING_OFFLINE: + /* + * Keep it very simple for now: just lock out ksmd and + * MADV_UNMERGEABLE while any memory is going offline. + * mutex_lock_nested() is necessary because lockdep was alarmed + * that here we take uksm_thread_mutex inside notifier chain + * mutex, and later take notifier chain mutex inside + * uksm_thread_mutex to unlock it. But that's safe because both + * are inside mem_hotplug_mutex. + */ + mutex_lock_nested(&uksm_thread_mutex, SINGLE_DEPTH_NESTING); + break; + + case MEM_OFFLINE: + /* + * Most of the work is done by page migration; but there might + * be a few stable_nodes left over, still pointing to struct + * pages which have been offlined: prune those from the tree. + */ + while ((stable_node = uksm_check_stable_tree(mn->start_pfn, + mn->start_pfn + mn->nr_pages)) != NULL) + remove_node_from_stable_tree(stable_node, 1, 1); + /* fallthrough */ + + case MEM_CANCEL_OFFLINE: + mutex_unlock(&uksm_thread_mutex); + break; + } + return NOTIFY_OK; +} +#endif /* CONFIG_MEMORY_HOTREMOVE */ + +#ifdef CONFIG_SYSFS +/* + * This all compiles without CONFIG_SYSFS, but is a waste of space. + */ + +#define UKSM_ATTR_RO(_name) \ + static struct kobj_attribute _name##_attr = __ATTR_RO(_name) +#define UKSM_ATTR(_name) \ + static struct kobj_attribute _name##_attr = \ + __ATTR(_name, 0644, _name##_show, _name##_store) + +static ssize_t max_cpu_percentage_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%u\n", uksm_max_cpu_percentage); +} + +static ssize_t max_cpu_percentage_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + unsigned long max_cpu_percentage; + int err; + + err = kstrtoul(buf, 10, &max_cpu_percentage); + if (err || max_cpu_percentage > 100) + return -EINVAL; + + if (max_cpu_percentage == 100) + max_cpu_percentage = 99; + else if (max_cpu_percentage < 10) + max_cpu_percentage = 10; + + uksm_max_cpu_percentage = max_cpu_percentage; + + return count; +} +UKSM_ATTR(max_cpu_percentage); + +static ssize_t sleep_millisecs_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%u\n", jiffies_to_msecs(uksm_sleep_jiffies)); +} + +static ssize_t sleep_millisecs_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + unsigned long msecs; + int err; + + err = kstrtoul(buf, 10, &msecs); + if (err || msecs > MSEC_PER_SEC) + return -EINVAL; + + uksm_sleep_jiffies = msecs_to_jiffies(msecs); + uksm_sleep_saved = uksm_sleep_jiffies; + + return count; +} +UKSM_ATTR(sleep_millisecs); + + +static ssize_t cpu_governor_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + int n = sizeof(uksm_cpu_governor_str) / sizeof(char *); + int i; + + buf[0] = '\0'; + for (i = 0; i < n ; i++) { + if (uksm_cpu_governor == i) + strcat(buf, "["); + + strcat(buf, uksm_cpu_governor_str[i]); + + if (uksm_cpu_governor == i) + strcat(buf, "]"); + + strcat(buf, " "); + } + strcat(buf, "\n"); + + return strlen(buf); +} + +static inline void init_performance_values(void) +{ + int i; + struct scan_rung *rung; + struct uksm_cpu_preset_s *preset = uksm_cpu_preset + uksm_cpu_governor; + + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = uksm_scan_ladder + i; + rung->cpu_ratio = preset->cpu_ratio[i]; + rung->cover_msecs = preset->cover_msecs[i]; + } + + uksm_max_cpu_percentage = preset->max_cpu; +} + +static ssize_t cpu_governor_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int n = sizeof(uksm_cpu_governor_str) / sizeof(char *); + + for (n--; n >= 0 ; n--) { + if (!strncmp(buf, uksm_cpu_governor_str[n], + strlen(uksm_cpu_governor_str[n]))) + break; + } + + if (n < 0) + return -EINVAL; + else + uksm_cpu_governor = n; + + init_performance_values(); + + return count; +} +UKSM_ATTR(cpu_governor); + +static ssize_t run_show(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%u\n", uksm_run); +} + +static ssize_t run_store(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int err; + unsigned long flags; + + err = kstrtoul(buf, 10, &flags); + if (err || flags > UINT_MAX) + return -EINVAL; + if (flags > UKSM_RUN_MERGE) + return -EINVAL; + + mutex_lock(&uksm_thread_mutex); + if (uksm_run != flags) + uksm_run = flags; + mutex_unlock(&uksm_thread_mutex); + + if (flags & UKSM_RUN_MERGE) + wake_up_interruptible(&uksm_thread_wait); + + return count; +} +UKSM_ATTR(run); + +static ssize_t abundant_threshold_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%u\n", uksm_abundant_threshold); +} + +static ssize_t abundant_threshold_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int err; + unsigned long flags; + + err = kstrtoul(buf, 10, &flags); + if (err || flags > 99) + return -EINVAL; + + uksm_abundant_threshold = flags; + + return count; +} +UKSM_ATTR(abundant_threshold); + +static ssize_t thrash_threshold_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%u\n", uksm_thrash_threshold); +} + +static ssize_t thrash_threshold_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int err; + unsigned long flags; + + err = kstrtoul(buf, 10, &flags); + if (err || flags > 99) + return -EINVAL; + + uksm_thrash_threshold = flags; + + return count; +} +UKSM_ATTR(thrash_threshold); + +static ssize_t cpu_ratios_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + int i, size; + struct scan_rung *rung; + char *p = buf; + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = &uksm_scan_ladder[i]; + + if (rung->cpu_ratio > 0) + size = sprintf(p, "%d ", rung->cpu_ratio); + else + size = sprintf(p, "MAX/%d ", + TIME_RATIO_SCALE / -rung->cpu_ratio); + + p += size; + } + + *p++ = '\n'; + *p = '\0'; + + return p - buf; +} + +static ssize_t cpu_ratios_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int i, cpuratios[SCAN_LADDER_SIZE], err; + unsigned long value; + struct scan_rung *rung; + char *p, *end = NULL; + + p = kzalloc(count, GFP_KERNEL); + if (!p) + return -ENOMEM; + + memcpy(p, buf, count); + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + if (i != SCAN_LADDER_SIZE - 1) { + end = strchr(p, ' '); + if (!end) + return -EINVAL; + + *end = '\0'; + } + + if (strstr(p, "MAX/")) { + p = strchr(p, '/') + 1; + err = kstrtoul(p, 10, &value); + if (err || value > TIME_RATIO_SCALE || !value) + return -EINVAL; + + cpuratios[i] = -(int) (TIME_RATIO_SCALE / value); + } else { + err = kstrtoul(p, 10, &value); + if (err || value > TIME_RATIO_SCALE || !value) + return -EINVAL; + + cpuratios[i] = value; + } + + p = end + 1; + } + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = &uksm_scan_ladder[i]; + + rung->cpu_ratio = cpuratios[i]; + } + + return count; +} +UKSM_ATTR(cpu_ratios); + +static ssize_t eval_intervals_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + int i, size; + struct scan_rung *rung; + char *p = buf; + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = &uksm_scan_ladder[i]; + size = sprintf(p, "%u ", rung->cover_msecs); + p += size; + } + + *p++ = '\n'; + *p = '\0'; + + return p - buf; +} + +static ssize_t eval_intervals_store(struct kobject *kobj, + struct kobj_attribute *attr, + const char *buf, size_t count) +{ + int i, err; + unsigned long values[SCAN_LADDER_SIZE]; + struct scan_rung *rung; + char *p, *end = NULL; + ssize_t ret = count; + + p = kzalloc(count + 2, GFP_KERNEL); + if (!p) + return -ENOMEM; + + memcpy(p, buf, count); + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + if (i != SCAN_LADDER_SIZE - 1) { + end = strchr(p, ' '); + if (!end) { + ret = -EINVAL; + goto out; + } + + *end = '\0'; + } + + err = kstrtoul(p, 10, &values[i]); + if (err) { + ret = -EINVAL; + goto out; + } + + p = end + 1; + } + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = &uksm_scan_ladder[i]; + + rung->cover_msecs = values[i]; + } + +out: + kfree(p); + return ret; +} +UKSM_ATTR(eval_intervals); + +static ssize_t ema_per_page_time_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", uksm_ema_page_time); +} +UKSM_ATTR_RO(ema_per_page_time); + +static ssize_t pages_shared_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", uksm_pages_shared); +} +UKSM_ATTR_RO(pages_shared); + +static ssize_t pages_sharing_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", uksm_pages_sharing); +} +UKSM_ATTR_RO(pages_sharing); + +static ssize_t pages_unshared_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", uksm_pages_unshared); +} +UKSM_ATTR_RO(pages_unshared); + +static ssize_t full_scans_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%llu\n", fully_scanned_round); +} +UKSM_ATTR_RO(full_scans); + +static ssize_t pages_scanned_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + unsigned long base = 0; + u64 delta, ret; + + if (pages_scanned_stored) { + base = pages_scanned_base; + ret = pages_scanned_stored; + delta = uksm_pages_scanned >> base; + if (CAN_OVERFLOW_U64(ret, delta)) { + ret >>= 1; + delta >>= 1; + base++; + ret += delta; + } + } else { + ret = uksm_pages_scanned; + } + + while (ret > ULONG_MAX) { + ret >>= 1; + base++; + } + + if (base) + return sprintf(buf, "%lu * 2^%lu\n", (unsigned long)ret, base); + else + return sprintf(buf, "%lu\n", (unsigned long)ret); +} +UKSM_ATTR_RO(pages_scanned); + +static ssize_t hash_strength_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%lu\n", hash_strength); +} +UKSM_ATTR_RO(hash_strength); + +static ssize_t sleep_times_show(struct kobject *kobj, + struct kobj_attribute *attr, char *buf) +{ + return sprintf(buf, "%llu\n", uksm_sleep_times); +} +UKSM_ATTR_RO(sleep_times); + + +static struct attribute *uksm_attrs[] = { + &max_cpu_percentage_attr.attr, + &sleep_millisecs_attr.attr, + &cpu_governor_attr.attr, + &run_attr.attr, + &ema_per_page_time_attr.attr, + &pages_shared_attr.attr, + &pages_sharing_attr.attr, + &pages_unshared_attr.attr, + &full_scans_attr.attr, + &pages_scanned_attr.attr, + &hash_strength_attr.attr, + &sleep_times_attr.attr, + &thrash_threshold_attr.attr, + &abundant_threshold_attr.attr, + &cpu_ratios_attr.attr, + &eval_intervals_attr.attr, + NULL, +}; + +static struct attribute_group uksm_attr_group = { + .attrs = uksm_attrs, + .name = "uksm", +}; +#endif /* CONFIG_SYSFS */ + +static inline void init_scan_ladder(void) +{ + int i; + struct scan_rung *rung; + + for (i = 0; i < SCAN_LADDER_SIZE; i++) { + rung = uksm_scan_ladder + i; + slot_tree_init_root(&rung->vma_root); + } + + init_performance_values(); + uksm_calc_scan_pages(); +} + +static inline int cal_positive_negative_costs(void) +{ + struct page *p1, *p2; + unsigned char *addr1, *addr2; + unsigned long i, time_start, hash_cost; + unsigned long loopnum = 0; + + /*IMPORTANT: volatile is needed to prevent over-optimization by gcc. */ + volatile u32 hash; + volatile int ret; + + p1 = alloc_page(GFP_KERNEL); + if (!p1) + return -ENOMEM; + + p2 = alloc_page(GFP_KERNEL); + if (!p2) + return -ENOMEM; + + addr1 = kmap_atomic(p1); + addr2 = kmap_atomic(p2); + memset(addr1, prandom_u32(), PAGE_SIZE); + memcpy(addr2, addr1, PAGE_SIZE); + + /* make sure that the two pages differ in last byte */ + addr2[PAGE_SIZE-1] = ~addr2[PAGE_SIZE-1]; + kunmap_atomic(addr2); + kunmap_atomic(addr1); + + time_start = jiffies; + while (jiffies - time_start < 100) { + for (i = 0; i < 100; i++) + hash = page_hash(p1, HASH_STRENGTH_FULL, 0); + loopnum += 100; + } + hash_cost = (jiffies - time_start); + + time_start = jiffies; + for (i = 0; i < loopnum; i++) + ret = pages_identical_with_cost(p1, p2); + memcmp_cost = HASH_STRENGTH_FULL * (jiffies - time_start); + memcmp_cost /= hash_cost; + pr_info("UKSM: relative memcmp_cost = %lu " + "hash=%u cmp_ret=%d.\n", + memcmp_cost, hash, ret); + + __free_page(p1); + __free_page(p2); + return 0; +} + +static int init_zeropage_hash_table(void) +{ + struct page *page; + char *addr; + int i; + + page = alloc_page(GFP_KERNEL); + if (!page) + return -ENOMEM; + + addr = kmap_atomic(page); + memset(addr, 0, PAGE_SIZE); + kunmap_atomic(addr); + + zero_hash_table = kmalloc_array(HASH_STRENGTH_MAX, sizeof(u32), + GFP_KERNEL); + if (!zero_hash_table) + return -ENOMEM; + + for (i = 0; i < HASH_STRENGTH_MAX; i++) + zero_hash_table[i] = page_hash(page, i, 0); + + __free_page(page); + + return 0; +} + +static inline int init_random_sampling(void) +{ + unsigned long i; + + random_nums = kmalloc(PAGE_SIZE, GFP_KERNEL); + if (!random_nums) + return -ENOMEM; + + for (i = 0; i < HASH_STRENGTH_FULL; i++) + random_nums[i] = i; + + for (i = 0; i < HASH_STRENGTH_FULL; i++) { + unsigned long rand_range, swap_index, tmp; + + rand_range = HASH_STRENGTH_FULL - i; + swap_index = i + prandom_u32() % rand_range; + tmp = random_nums[i]; + random_nums[i] = random_nums[swap_index]; + random_nums[swap_index] = tmp; + } + + rshash_state.state = RSHASH_NEW; + rshash_state.below_count = 0; + rshash_state.lookup_window_index = 0; + + return cal_positive_negative_costs(); +} + +static int __init uksm_slab_init(void) +{ + rmap_item_cache = UKSM_KMEM_CACHE(rmap_item, 0); + if (!rmap_item_cache) + goto out; + + stable_node_cache = UKSM_KMEM_CACHE(stable_node, 0); + if (!stable_node_cache) + goto out_free1; + + node_vma_cache = UKSM_KMEM_CACHE(node_vma, 0); + if (!node_vma_cache) + goto out_free2; + + vma_slot_cache = UKSM_KMEM_CACHE(vma_slot, 0); + if (!vma_slot_cache) + goto out_free3; + + tree_node_cache = UKSM_KMEM_CACHE(tree_node, 0); + if (!tree_node_cache) + goto out_free4; + + return 0; + +out_free4: + kmem_cache_destroy(vma_slot_cache); +out_free3: + kmem_cache_destroy(node_vma_cache); +out_free2: + kmem_cache_destroy(stable_node_cache); +out_free1: + kmem_cache_destroy(rmap_item_cache); +out: + return -ENOMEM; +} + +static void __init uksm_slab_free(void) +{ + kmem_cache_destroy(stable_node_cache); + kmem_cache_destroy(rmap_item_cache); + kmem_cache_destroy(node_vma_cache); + kmem_cache_destroy(vma_slot_cache); + kmem_cache_destroy(tree_node_cache); +} + +/* Common interface to ksm, different to it. */ +int ksm_madvise(struct vm_area_struct *vma, unsigned long start, + unsigned long end, int advice, unsigned long *vm_flags) +{ + int err; + + switch (advice) { + case MADV_MERGEABLE: + return 0; /* just ignore the advice */ + + case MADV_UNMERGEABLE: + if (!(*vm_flags & VM_MERGEABLE) || !uksm_flags_can_scan(*vm_flags)) + return 0; /* just ignore the advice */ + + if (vma->anon_vma) { + err = unmerge_uksm_pages(vma, start, end); + if (err) + return err; + } + + uksm_remove_vma(vma); + *vm_flags &= ~VM_MERGEABLE; + break; + } + + return 0; +} + +/* Common interface to ksm, actually the same. */ +struct page *ksm_might_need_to_copy(struct page *page, + struct vm_area_struct *vma, unsigned long address) +{ + struct anon_vma *anon_vma = page_anon_vma(page); + struct page *new_page; + + if (PageKsm(page)) { + if (page_stable_node(page)) + return page; /* no need to copy it */ + } else if (!anon_vma) { + return page; /* no need to copy it */ + } else if (anon_vma->root == vma->anon_vma->root && + page->index == linear_page_index(vma, address)) { + return page; /* still no need to copy it */ + } + if (!PageUptodate(page)) + return page; /* let do_swap_page report the error */ + + new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address); + if (new_page) { + copy_user_highpage(new_page, page, address, vma); + + SetPageDirty(new_page); + __SetPageUptodate(new_page); + __SetPageLocked(new_page); + } + + return new_page; +} + +/* Copied from mm/ksm.c and required from 5.1 */ +bool reuse_ksm_page(struct page *page, + struct vm_area_struct *vma, + unsigned long address) +{ +#ifdef CONFIG_DEBUG_VM + if (WARN_ON(is_zero_pfn(page_to_pfn(page))) || + WARN_ON(!page_mapped(page)) || + WARN_ON(!PageLocked(page))) { + dump_page(page, "reuse_ksm_page"); + return false; + } +#endif + + if (PageSwapCache(page) || !page_stable_node(page)) + return false; + /* Prohibit parallel get_ksm_page() */ + if (!page_ref_freeze(page, 1)) + return false; + + page_move_anon_rmap(page, vma); + page->index = linear_page_index(vma, address); + page_ref_unfreeze(page, 1); + + return true; +} + +static int __init uksm_init(void) +{ + struct task_struct *uksm_thread; + int err; + + uksm_sleep_jiffies = msecs_to_jiffies(100); + uksm_sleep_saved = uksm_sleep_jiffies; + + slot_tree_init(); + init_scan_ladder(); + + + err = init_random_sampling(); + if (err) + goto out_free2; + + err = uksm_slab_init(); + if (err) + goto out_free1; + + err = init_zeropage_hash_table(); + if (err) + goto out_free0; + + uksm_thread = kthread_run(uksm_scan_thread, NULL, "uksmd"); + if (IS_ERR(uksm_thread)) { + pr_err("uksm: creating kthread failed\n"); + err = PTR_ERR(uksm_thread); + goto out_free; + } + +#ifdef CONFIG_SYSFS + err = sysfs_create_group(mm_kobj, &uksm_attr_group); + if (err) { + pr_err("uksm: register sysfs failed\n"); + kthread_stop(uksm_thread); + goto out_free; + } +#else + uksm_run = UKSM_RUN_MERGE; /* no way for user to start it */ + +#endif /* CONFIG_SYSFS */ + +#ifdef CONFIG_MEMORY_HOTREMOVE + /* + * Choose a high priority since the callback takes uksm_thread_mutex: + * later callbacks could only be taking locks which nest within that. + */ + hotplug_memory_notifier(uksm_memory_callback, 100); +#endif + return 0; + +out_free: + kfree(zero_hash_table); +out_free0: + uksm_slab_free(); +out_free1: + kfree(random_nums); +out_free2: + kfree(uksm_scan_ladder); + return err; +} + +#ifdef MODULE +subsys_initcall(ksm_init); +#else +late_initcall(uksm_init); +#endif + diff --git a/mm/userfaultfd.c b/mm/userfaultfd.c index 63a73e164d5..747a2d7eb5b 100644 --- a/mm/userfaultfd.c +++ b/mm/userfaultfd.c @@ -123,7 +123,7 @@ static int mcopy_atomic_pte(struct mm_struct *dst_mm, inc_mm_counter(dst_mm, MM_ANONPAGES); page_add_new_anon_rmap(page, dst_vma, dst_addr, false); - lru_cache_add_inactive_or_unevictable(page, dst_vma); + lru_cache_add_page_vma(page, dst_vma, true); set_pte_at(dst_mm, dst_addr, dst_pte, _dst_pte); diff --git a/mm/vmscan.c b/mm/vmscan.c index 5199b9696ba..6f0f83686a5 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -49,6 +49,11 @@ #include #include #include +#include +#include +#include +#include +#include #include #include @@ -118,6 +123,19 @@ struct scan_control { /* The file pages on the current node are dangerously low */ unsigned int file_is_tiny:1; + /* + * The clean file pages on the current node won't be reclaimed when + * their amount is below vm.clean_low_kbytes *unless* we threaten + * to OOM or have no free swap space or vm.swappiness=0. + */ + unsigned int clean_below_low:1; + + /* + * The clean file pages on the current node won't be reclaimed when + * their amount is below vm.clean_min_kbytes. + */ + unsigned int clean_below_min:1; + /* Allocation order */ s8 order; @@ -164,6 +182,17 @@ struct scan_control { #define prefetchw_prev_lru_page(_page, _base, _field) do { } while (0) #endif +#if CONFIG_CLEAN_LOW_KBYTES < 0 +#error "CONFIG_CLEAN_LOW_KBYTES must be >= 0" +#endif + +#if CONFIG_CLEAN_MIN_KBYTES < 0 +#error "CONFIG_CLEAN_MIN_KBYTES must be >= 0" +#endif + +unsigned long sysctl_clean_low_kbytes __read_mostly = CONFIG_CLEAN_LOW_KBYTES; +unsigned long sysctl_clean_min_kbytes __read_mostly = CONFIG_CLEAN_MIN_KBYTES; + /* * From 0 .. 200. Higher means more swappy. */ @@ -1093,9 +1122,11 @@ static int __remove_mapping(struct address_space *mapping, struct page *page, if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; - mem_cgroup_swapout(page, swap); + + /* get a shadow entry before page_memcg() is cleared */ if (reclaimed && !mapping_exiting(mapping)) shadow = workingset_eviction(page, target_memcg); + mem_cgroup_swapout(page, swap); __delete_from_swap_cache(page, swap, shadow); xa_unlock_irqrestore(&mapping->i_pages, flags); put_swap_page(page, swap); @@ -1306,6 +1337,11 @@ static unsigned int shrink_page_list(struct list_head *page_list, if (!sc->may_unmap && page_mapped(page)) goto keep_locked; + /* in case the page was found accessed by lru_gen_scan_around() */ + if (lru_gen_enabled() && !ignore_references && + page_mapped(page) && PageReferenced(page)) + goto keep_locked; + may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); @@ -2421,6 +2457,106 @@ enum scan_balance { SCAN_FILE, }; +static void prepare_scan_count(pg_data_t *pgdat, struct scan_control *sc) +{ + unsigned long file; + struct lruvec *target_lruvec; + + if (lru_gen_enabled()) + return; + + target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat); + + /* + * Determine the scan balance between anon and file LRUs. + */ + spin_lock_irq(&target_lruvec->lru_lock); + sc->anon_cost = target_lruvec->anon_cost; + sc->file_cost = target_lruvec->file_cost; + spin_unlock_irq(&target_lruvec->lru_lock); + + /* + * Target desirable inactive:active list ratios for the anon + * and file LRU lists. + */ + if (!sc->force_deactivate) { + unsigned long refaults; + + refaults = lruvec_page_state(target_lruvec, + WORKINGSET_ACTIVATE_ANON); + if (refaults != target_lruvec->refaults[0] || + inactive_is_low(target_lruvec, LRU_INACTIVE_ANON)) + sc->may_deactivate |= DEACTIVATE_ANON; + else + sc->may_deactivate &= ~DEACTIVATE_ANON; + + /* + * When refaults are being observed, it means a new + * workingset is being established. Deactivate to get + * rid of any stale active pages quickly. + */ + refaults = lruvec_page_state(target_lruvec, + WORKINGSET_ACTIVATE_FILE); + if (refaults != target_lruvec->refaults[1] || + inactive_is_low(target_lruvec, LRU_INACTIVE_FILE)) + sc->may_deactivate |= DEACTIVATE_FILE; + else + sc->may_deactivate &= ~DEACTIVATE_FILE; + } else + sc->may_deactivate = DEACTIVATE_ANON | DEACTIVATE_FILE; + + /* + * If we have plenty of inactive file pages that aren't + * thrashing, try to reclaim those first before touching + * anonymous pages. + */ + file = lruvec_page_state(target_lruvec, NR_INACTIVE_FILE); + if (file >> sc->priority && !(sc->may_deactivate & DEACTIVATE_FILE)) + sc->cache_trim_mode = 1; + else + sc->cache_trim_mode = 0; + + /* + * Prevent the reclaimer from falling into the cache trap: as + * cache pages start out inactive, every cache fault will tip + * the scan balance towards the file LRU. And as the file LRU + * shrinks, so does the window for rotation from references. + * This means we have a runaway feedback loop where a tiny + * thrashing file LRU becomes infinitely more attractive than + * anon pages. Try to detect this based on file LRU size. + */ + if (!cgroup_reclaim(sc)) { + unsigned long total_high_wmark = 0; + unsigned long free, anon; + int z; + + free = sum_zone_node_page_state(pgdat->node_id, NR_FREE_PAGES); + file = node_page_state(pgdat, NR_ACTIVE_FILE) + + node_page_state(pgdat, NR_INACTIVE_FILE); + + for (z = 0; z < MAX_NR_ZONES; z++) { + struct zone *zone = &pgdat->node_zones[z]; + + if (!managed_zone(zone)) + continue; + + total_high_wmark += high_wmark_pages(zone); + } + + /* + * Consider anon: if that's low too, this isn't a + * runaway file reclaim problem, but rather just + * extreme pressure. Reclaim as per usual then. + */ + anon = node_page_state(pgdat, NR_INACTIVE_ANON); + + sc->file_is_tiny = + file + free <= total_high_wmark && + !(sc->may_deactivate & DEACTIVATE_ANON) && + anon >> sc->priority; + } +} + /* * Determine how aggressively the anon and file LRU lists should be * scanned. The relative value of each set of LRU lists is determined @@ -2478,6 +2614,16 @@ static void get_scan_count(struct lruvec *lruvec, struct scan_control *sc, goto out; } + /* + * Force-scan anon if clean file pages is under vm.clean_min_kbytes + * or vm.clean_low_kbytes (unless the swappiness setting + * disagrees with swapping). + */ + if ((sc->clean_below_low || sc->clean_below_min) && swappiness) { + scan_balance = SCAN_ANON; + goto out; + } + /* * If there is enough inactive page cache, we do not reclaim * anything from the anonymous working right now. @@ -2614,131 +2760,2631 @@ static void get_scan_count(struct lruvec *lruvec, struct scan_control *sc, BUG(); } + /* + * Don't reclaim clean file pages when their amount is below + * vm.clean_min_kbytes. + */ + if (file && sc->clean_below_min) + scan = 0; + nr[lru] = scan; } } -static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc) +#ifdef CONFIG_LRU_GEN + +/* + * After pages are faulted in, the aging must scan them twice before the + * eviction can consider them. The first scan clears the accessed bit set during + * initial faults. And the second scan makes sure they haven't been used since + * the first scan. + */ +#define MIN_NR_GENS 2 + +#define MAX_BATCH_SIZE 8192 + +/****************************************************************************** + * shorthand helpers + ******************************************************************************/ + +#define DEFINE_MAX_SEQ() \ + unsigned long max_seq = READ_ONCE(lruvec->evictable.max_seq) + +#define DEFINE_MIN_SEQ() \ + unsigned long min_seq[ANON_AND_FILE] = { \ + READ_ONCE(lruvec->evictable.min_seq[0]), \ + READ_ONCE(lruvec->evictable.min_seq[1]), \ + } + +#define for_each_type_zone(type, zone) \ + for ((type) = 0; (type) < ANON_AND_FILE; (type)++) \ + for ((zone) = 0; (zone) < MAX_NR_ZONES; (zone)++) + +#define for_each_gen_type_zone(gen, type, zone) \ + for ((gen) = 0; (gen) < MAX_NR_GENS; (gen)++) \ + for ((type) = 0; (type) < ANON_AND_FILE; (type)++) \ + for ((zone) = 0; (zone) < MAX_NR_ZONES; (zone)++) + +static int page_lru_gen(struct page *page) { - unsigned long nr[NR_LRU_LISTS]; - unsigned long targets[NR_LRU_LISTS]; - unsigned long nr_to_scan; - enum lru_list lru; - unsigned long nr_reclaimed = 0; - unsigned long nr_to_reclaim = sc->nr_to_reclaim; - struct blk_plug plug; - bool scan_adjusted; + return ((page->flags & LRU_GEN_MASK) >> LRU_GEN_PGOFF) - 1; +} - get_scan_count(lruvec, sc, nr); +static int get_nr_gens(struct lruvec *lruvec, int type) +{ + return lruvec->evictable.max_seq - lruvec->evictable.min_seq[type] + 1; +} - /* Record the original scan target for proportional adjustments later */ - memcpy(targets, nr, sizeof(nr)); +static int min_nr_gens(unsigned long max_seq, unsigned long *min_seq, int swappiness) +{ + return max_seq - max(min_seq[!swappiness], min_seq[1]) + 1; +} - /* - * Global reclaiming within direct reclaim at DEF_PRIORITY is a normal - * event that can occur when there is little memory pressure e.g. - * multiple streaming readers/writers. Hence, we do not abort scanning - * when the requested number of pages are reclaimed when scanning at - * DEF_PRIORITY on the assumption that the fact we are direct - * reclaiming implies that kswapd is not keeping up and it is best to - * do a batch of work at once. For memcg reclaim one check is made to - * abort proportional reclaim if either the file or anon lru has already - * dropped to zero at the first pass. - */ - scan_adjusted = (!cgroup_reclaim(sc) && !current_is_kswapd() && - sc->priority == DEF_PRIORITY); +static int max_nr_gens(unsigned long max_seq, unsigned long *min_seq, int swappiness) +{ + return max_seq - min(min_seq[!swappiness], min_seq[1]) + 1; +} - blk_start_plug(&plug); - while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || - nr[LRU_INACTIVE_FILE]) { - unsigned long nr_anon, nr_file, percentage; - unsigned long nr_scanned; +static bool __maybe_unused seq_is_valid(struct lruvec *lruvec) +{ + lockdep_assert_held(&lruvec->lru_lock); - for_each_evictable_lru(lru) { - if (nr[lru]) { - nr_to_scan = min(nr[lru], SWAP_CLUSTER_MAX); - nr[lru] -= nr_to_scan; + return get_nr_gens(lruvec, 0) >= MIN_NR_GENS && + get_nr_gens(lruvec, 0) <= MAX_NR_GENS && + get_nr_gens(lruvec, 1) >= MIN_NR_GENS && + get_nr_gens(lruvec, 1) <= MAX_NR_GENS; +} - nr_reclaimed += shrink_list(lru, nr_to_scan, - lruvec, sc); - } - } +/****************************************************************************** + * refault feedback loop + ******************************************************************************/ - cond_resched(); +/* + * A feedback loop modeled after the PID controller. Currently supports the + * proportional (P) and the integral (I) terms; the derivative (D) term can be + * added if necessary. The setpoint (SP) is the desired position; the process + * variable (PV) is the measured position. The error is the difference between + * the SP and the PV. A positive error results in a positive control output + * correction, which, in our case, is to allow eviction. + * + * The P term is the current refault rate refaulted/(evicted+activated), which + * has a weight of 1. The I term is the arithmetic mean of the last N refault + * rates, weighted by geometric series 1/2, 1/4, ..., 1/(1<evictable; + int hist = hist_from_seq_or_gen(lrugen->min_seq[type]); + + pos->refaulted = lrugen->avg_refaulted[type][tier] + + atomic_long_read(&lrugen->refaulted[hist][type][tier]); + pos->total = lrugen->avg_total[type][tier] + + atomic_long_read(&lrugen->evicted[hist][type][tier]); + if (tier) + pos->total += lrugen->activated[hist][type][tier - 1]; + pos->gain = gain; +} - /* - * For kswapd and memcg, reclaim at least the number of pages - * requested. Ensure that the anon and file LRUs are scanned - * proportionally what was requested by get_scan_count(). We - * stop reclaiming one LRU and reduce the amount scanning - * proportional to the original scan target. - */ - nr_file = nr[LRU_INACTIVE_FILE] + nr[LRU_ACTIVE_FILE]; - nr_anon = nr[LRU_INACTIVE_ANON] + nr[LRU_ACTIVE_ANON]; +static void reset_controller_pos(struct lruvec *lruvec, int gen, int type) +{ + int tier; + int hist = hist_from_seq_or_gen(gen); + struct lrugen *lrugen = &lruvec->evictable; + bool carryover = gen == lru_gen_from_seq(lrugen->min_seq[type]); - /* - * It's just vindictive to attack the larger once the smaller - * has gone to zero. And given the way we stop scanning the - * smaller below, this makes sure that we only make one nudge - * towards proportionality once we've got nr_to_reclaim. - */ - if (!nr_file || !nr_anon) - break; + if (!carryover && NR_STAT_GENS == 1) + return; - if (nr_file > nr_anon) { - unsigned long scan_target = targets[LRU_INACTIVE_ANON] + - targets[LRU_ACTIVE_ANON] + 1; - lru = LRU_BASE; - percentage = nr_anon * 100 / scan_target; - } else { - unsigned long scan_target = targets[LRU_INACTIVE_FILE] + - targets[LRU_ACTIVE_FILE] + 1; - lru = LRU_FILE; - percentage = nr_file * 100 / scan_target; - } + for (tier = 0; tier < MAX_NR_TIERS; tier++) { + if (carryover) { + unsigned long sum; - /* Stop scanning the smaller of the LRU */ - nr[lru] = 0; - nr[lru + LRU_ACTIVE] = 0; + sum = lrugen->avg_refaulted[type][tier] + + atomic_long_read(&lrugen->refaulted[hist][type][tier]); + WRITE_ONCE(lrugen->avg_refaulted[type][tier], sum / 2); - /* - * Recalculate the other LRU scan count based on its original - * scan target and the percentage scanning already complete - */ - lru = (lru == LRU_FILE) ? LRU_BASE : LRU_FILE; - nr_scanned = targets[lru] - nr[lru]; - nr[lru] = targets[lru] * (100 - percentage) / 100; - nr[lru] -= min(nr[lru], nr_scanned); + sum = lrugen->avg_total[type][tier] + + atomic_long_read(&lrugen->evicted[hist][type][tier]); + if (tier) + sum += lrugen->activated[hist][type][tier - 1]; + WRITE_ONCE(lrugen->avg_total[type][tier], sum / 2); - lru += LRU_ACTIVE; - nr_scanned = targets[lru] - nr[lru]; - nr[lru] = targets[lru] * (100 - percentage) / 100; - nr[lru] -= min(nr[lru], nr_scanned); + if (NR_STAT_GENS > 1) + continue; + } - scan_adjusted = true; + atomic_long_set(&lrugen->refaulted[hist][type][tier], 0); + atomic_long_set(&lrugen->evicted[hist][type][tier], 0); + if (tier) + WRITE_ONCE(lrugen->activated[hist][type][tier - 1], 0); } - blk_finish_plug(&plug); - sc->nr_reclaimed += nr_reclaimed; +} +static bool positive_ctrl_err(struct controller_pos *sp, struct controller_pos *pv) +{ /* - * Even if we did not try to evict anon pages at all, we want to - * rebalance the anon lru active/inactive ratio. + * Allow eviction if the PV has a limited number of refaulted pages or a + * lower refault rate than the SP. */ - if (total_swap_pages && inactive_is_low(lruvec, LRU_INACTIVE_ANON)) - shrink_active_list(SWAP_CLUSTER_MAX, lruvec, - sc, LRU_ACTIVE_ANON); + return pv->refaulted < SWAP_CLUSTER_MAX || + pv->refaulted * max(sp->total, 1UL) * sp->gain <= + sp->refaulted * max(pv->total, 1UL) * pv->gain; } -/* Use reclaim/compaction for costly allocs or under memory pressure */ -static bool in_reclaim_compaction(struct scan_control *sc) -{ - if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && - (sc->order > PAGE_ALLOC_COSTLY_ORDER || - sc->priority < DEF_PRIORITY - 2)) - return true; +/****************************************************************************** + * mm_struct list + ******************************************************************************/ + +enum { + MM_SCHED_ACTIVE, /* running processes */ + MM_SCHED_INACTIVE, /* sleeping processes */ + MM_LOCK_CONTENTION, /* lock contentions */ + MM_VMA_INTERVAL, /* VMAs within the range of each PUD/PMD/PTE */ + MM_LEAF_OTHER_NODE, /* entries not from the node under reclaim */ + MM_LEAF_OTHER_MEMCG, /* entries not from the memcg under reclaim */ + MM_LEAF_OLD, /* old entries */ + MM_LEAF_YOUNG, /* young entries */ + MM_LEAF_DIRTY, /* dirty entries */ + MM_LEAF_HOLE, /* non-present entries */ + MM_NONLEAF_OLD, /* old non-leaf PMD entries */ + MM_NONLEAF_YOUNG, /* young non-leaf PMD entries */ + NR_MM_STATS +}; + +/* mnemonic codes for the stats above */ +#define MM_STAT_CODES "aicvnmoydhlu" + +struct lru_gen_mm_list { + /* the head of a global or per-memcg mm_struct list */ + struct list_head head; + /* protects the list */ + spinlock_t lock; + struct { + /* set to max_seq after each round of walk */ + unsigned long cur_seq; + /* the next mm on the list to walk */ + struct list_head *iter; + /* to wait for the last worker to finish */ + struct wait_queue_head wait; + /* the number of concurrent workers */ + int nr_workers; + /* stats for debugging */ + unsigned long stats[NR_STAT_GENS][NR_MM_STATS]; + } nodes[0]; +}; + +static struct lru_gen_mm_list *global_mm_list; + +static struct lru_gen_mm_list *alloc_mm_list(void) +{ + int nid; + struct lru_gen_mm_list *mm_list; + + mm_list = kzalloc(struct_size(mm_list, nodes, nr_node_ids), GFP_KERNEL); + if (!mm_list) + return NULL; + + INIT_LIST_HEAD(&mm_list->head); + spin_lock_init(&mm_list->lock); + + for_each_node(nid) { + mm_list->nodes[nid].cur_seq = MIN_NR_GENS; + mm_list->nodes[nid].iter = &mm_list->head; + init_waitqueue_head(&mm_list->nodes[nid].wait); + } + + return mm_list; +} + +static struct lru_gen_mm_list *get_mm_list(struct mem_cgroup *memcg) +{ +#ifdef CONFIG_MEMCG + if (!mem_cgroup_disabled()) + return memcg ? memcg->mm_list : root_mem_cgroup->mm_list; +#endif + VM_BUG_ON(memcg); + + return global_mm_list; +} + +void lru_gen_init_mm(struct mm_struct *mm) +{ + INIT_LIST_HEAD(&mm->lrugen.list); +#ifdef CONFIG_MEMCG + mm->lrugen.memcg = NULL; +#endif +#ifndef CONFIG_ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH + atomic_set(&mm->lrugen.nr_cpus, 0); +#endif + nodes_clear(mm->lrugen.nodes); +} + +void lru_gen_add_mm(struct mm_struct *mm) +{ + struct mem_cgroup *memcg = get_mem_cgroup_from_mm(mm); + struct lru_gen_mm_list *mm_list = get_mm_list(memcg); + + VM_BUG_ON_MM(!list_empty(&mm->lrugen.list), mm); +#ifdef CONFIG_MEMCG + VM_BUG_ON_MM(mm->lrugen.memcg, mm); + WRITE_ONCE(mm->lrugen.memcg, memcg); +#endif + spin_lock(&mm_list->lock); + list_add_tail(&mm->lrugen.list, &mm_list->head); + spin_unlock(&mm_list->lock); +} + +void lru_gen_del_mm(struct mm_struct *mm) +{ + int nid; +#ifdef CONFIG_MEMCG + struct lru_gen_mm_list *mm_list = get_mm_list(mm->lrugen.memcg); +#else + struct lru_gen_mm_list *mm_list = get_mm_list(NULL); +#endif + + spin_lock(&mm_list->lock); + + for_each_node(nid) { + if (mm_list->nodes[nid].iter != &mm->lrugen.list) + continue; + + mm_list->nodes[nid].iter = mm_list->nodes[nid].iter->next; + if (mm_list->nodes[nid].iter == &mm_list->head) + WRITE_ONCE(mm_list->nodes[nid].cur_seq, + mm_list->nodes[nid].cur_seq + 1); + } + + list_del_init(&mm->lrugen.list); + + spin_unlock(&mm_list->lock); + +#ifdef CONFIG_MEMCG + mem_cgroup_put(mm->lrugen.memcg); + WRITE_ONCE(mm->lrugen.memcg, NULL); +#endif +} + +#ifdef CONFIG_MEMCG +int lru_gen_alloc_mm_list(struct mem_cgroup *memcg) +{ + if (mem_cgroup_disabled()) + return 0; + + memcg->mm_list = alloc_mm_list(); + + return memcg->mm_list ? 0 : -ENOMEM; +} + +void lru_gen_free_mm_list(struct mem_cgroup *memcg) +{ + kfree(memcg->mm_list); + memcg->mm_list = NULL; +} + +void lru_gen_migrate_mm(struct mm_struct *mm) +{ + struct mem_cgroup *memcg; + + lockdep_assert_held(&mm->owner->alloc_lock); + + if (mem_cgroup_disabled()) + return; + + rcu_read_lock(); + memcg = mem_cgroup_from_task(mm->owner); + rcu_read_unlock(); + if (memcg == mm->lrugen.memcg) + return; + + VM_BUG_ON_MM(!mm->lrugen.memcg, mm); + VM_BUG_ON_MM(list_empty(&mm->lrugen.list), mm); + + lru_gen_del_mm(mm); + lru_gen_add_mm(mm); +} + +static bool mm_has_migrated(struct mm_struct *mm, struct mem_cgroup *memcg) +{ + return READ_ONCE(mm->lrugen.memcg) != memcg; +} +#else +static bool mm_has_migrated(struct mm_struct *mm, struct mem_cgroup *memcg) +{ + return false; +} +#endif + +struct mm_walk_args { + struct mem_cgroup *memcg; + unsigned long max_seq; + unsigned long start_pfn; + unsigned long end_pfn; + unsigned long next_addr; + int node_id; + int swappiness; + int batch_size; + int nr_pages[MAX_NR_GENS][ANON_AND_FILE][MAX_NR_ZONES]; + int mm_stats[NR_MM_STATS]; + unsigned long bitmap[0]; +}; + +static int size_of_mm_walk_args(void) +{ + int size = sizeof(struct mm_walk_args); + + if (IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE) || + IS_ENABLED(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG)) + size += sizeof(unsigned long) * BITS_TO_LONGS(PTRS_PER_PMD); + + return size; +} + +static void reset_mm_stats(struct lru_gen_mm_list *mm_list, bool last, + struct mm_walk_args *args) +{ + int i; + int nid = args->node_id; + int hist = hist_from_seq_or_gen(args->max_seq); + + lockdep_assert_held(&mm_list->lock); + + for (i = 0; i < NR_MM_STATS; i++) { + WRITE_ONCE(mm_list->nodes[nid].stats[hist][i], + mm_list->nodes[nid].stats[hist][i] + args->mm_stats[i]); + args->mm_stats[i] = 0; + } + + if (!last || NR_STAT_GENS == 1) + return; + + hist = hist_from_seq_or_gen(args->max_seq + 1); + for (i = 0; i < NR_MM_STATS; i++) + WRITE_ONCE(mm_list->nodes[nid].stats[hist][i], 0); +} + +static bool should_skip_mm(struct mm_struct *mm, struct mm_walk_args *args) +{ + int type; + unsigned long size = 0; + + if (!lru_gen_mm_is_active(mm) && !node_isset(args->node_id, mm->lrugen.nodes)) + return true; + + if (mm_is_oom_victim(mm)) + return true; + + for (type = !args->swappiness; type < ANON_AND_FILE; type++) { + size += type ? get_mm_counter(mm, MM_FILEPAGES) : + get_mm_counter(mm, MM_ANONPAGES) + + get_mm_counter(mm, MM_SHMEMPAGES); + } + + /* leave the legwork to the rmap if mappings are too sparse */ + if (size < max(SWAP_CLUSTER_MAX, mm_pgtables_bytes(mm) / PAGE_SIZE)) + return true; + + return !mmget_not_zero(mm); +} + +/* To support multiple workers that concurrently walk an mm_struct list. */ +static bool get_next_mm(struct mm_walk_args *args, struct mm_struct **iter) +{ + bool last = true; + struct mm_struct *mm = NULL; + int nid = args->node_id; + struct lru_gen_mm_list *mm_list = get_mm_list(args->memcg); + + if (*iter) + mmput_async(*iter); + else if (args->max_seq <= READ_ONCE(mm_list->nodes[nid].cur_seq)) + return false; + + spin_lock(&mm_list->lock); + + VM_BUG_ON(args->max_seq > mm_list->nodes[nid].cur_seq + 1); + VM_BUG_ON(*iter && args->max_seq < mm_list->nodes[nid].cur_seq); + VM_BUG_ON(*iter && !mm_list->nodes[nid].nr_workers); + + if (args->max_seq <= mm_list->nodes[nid].cur_seq) { + last = *iter; + goto done; + } + + if (mm_list->nodes[nid].iter == &mm_list->head) { + VM_BUG_ON(*iter || mm_list->nodes[nid].nr_workers); + mm_list->nodes[nid].iter = mm_list->nodes[nid].iter->next; + } + + while (!mm && mm_list->nodes[nid].iter != &mm_list->head) { + mm = list_entry(mm_list->nodes[nid].iter, struct mm_struct, lrugen.list); + mm_list->nodes[nid].iter = mm_list->nodes[nid].iter->next; + if (should_skip_mm(mm, args)) + mm = NULL; + + args->mm_stats[mm ? MM_SCHED_ACTIVE : MM_SCHED_INACTIVE]++; + } + + if (mm_list->nodes[nid].iter == &mm_list->head) + WRITE_ONCE(mm_list->nodes[nid].cur_seq, + mm_list->nodes[nid].cur_seq + 1); +done: + if (*iter && !mm) + mm_list->nodes[nid].nr_workers--; + if (!*iter && mm) + mm_list->nodes[nid].nr_workers++; + + last = last && !mm_list->nodes[nid].nr_workers && + mm_list->nodes[nid].iter == &mm_list->head; + + reset_mm_stats(mm_list, last, args); + + spin_unlock(&mm_list->lock); + + *iter = mm; + if (mm) + node_clear(nid, mm->lrugen.nodes); + + return last; +} + +/****************************************************************************** + * the aging + ******************************************************************************/ + +static void update_batch_size(struct page *page, int old_gen, int new_gen, + struct mm_walk_args *args) +{ + int type = page_is_file_lru(page); + int zone = page_zonenum(page); + int delta = thp_nr_pages(page); + + VM_BUG_ON(old_gen >= MAX_NR_GENS); + VM_BUG_ON(new_gen >= MAX_NR_GENS); + + args->batch_size++; + + args->nr_pages[old_gen][type][zone] -= delta; + args->nr_pages[new_gen][type][zone] += delta; +} + +static void reset_batch_size(struct lruvec *lruvec, struct mm_walk_args *args) +{ + int gen, type, zone; + struct lrugen *lrugen = &lruvec->evictable; + + if (!args->batch_size) + return; + + args->batch_size = 0; + + spin_lock_irq(&lruvec->lru_lock); + + for_each_gen_type_zone(gen, type, zone) { + enum lru_list lru = type * LRU_FILE; + int total = args->nr_pages[gen][type][zone]; + + if (!total) + continue; + + args->nr_pages[gen][type][zone] = 0; + WRITE_ONCE(lrugen->sizes[gen][type][zone], + lrugen->sizes[gen][type][zone] + total); + + if (lru_gen_is_active(lruvec, gen)) + lru += LRU_ACTIVE; + update_lru_size(lruvec, lru, zone, total); + } + + spin_unlock_irq(&lruvec->lru_lock); +} + +static int page_update_gen(struct page *page, int new_gen) +{ + int old_gen; + unsigned long old_flags, new_flags; + + VM_BUG_ON(new_gen >= MAX_NR_GENS); + + do { + old_flags = READ_ONCE(page->flags); + + old_gen = ((old_flags & LRU_GEN_MASK) >> LRU_GEN_PGOFF) - 1; + if (old_gen < 0) { + new_flags = old_flags | BIT(PG_referenced); + continue; + } + + new_flags = (old_flags & ~(LRU_GEN_MASK | LRU_USAGE_MASK | LRU_TIER_FLAGS)) | + ((new_gen + 1UL) << LRU_GEN_PGOFF); + } while (new_flags != old_flags && + cmpxchg(&page->flags, old_flags, new_flags) != old_flags); + + return old_gen; +} + +static int should_skip_vma(unsigned long start, unsigned long end, struct mm_walk *walk) +{ + struct address_space *mapping; + struct vm_area_struct *vma = walk->vma; + struct mm_walk_args *args = walk->private; + + if (!vma_is_accessible(vma) || is_vm_hugetlb_page(vma) || + (vma->vm_flags & (VM_LOCKED | VM_SPECIAL))) + return true; + + if (vma_is_anonymous(vma)) + return !args->swappiness; + + if (WARN_ON_ONCE(!vma->vm_file || !vma->vm_file->f_mapping)) + return true; + + mapping = vma->vm_file->f_mapping; + if (!mapping->a_ops->writepage) + return true; + + return (shmem_mapping(mapping) && !args->swappiness) || mapping_unevictable(mapping); +} + +/* + * Some userspace memory allocators create many single-page VMAs. So instead of + * returning back to the PGD table for each of such VMAs, we finish at least an + * entire PMD table and therefore avoid many zigzags. This optimizes page table + * walks for workloads that have large numbers of tiny VMAs. + * + * We scan PMD tables in two passes. The first pass reaches to PTE tables and + * doesn't take the PMD lock. The second pass clears the accessed bit on PMD + * entries and needs to take the PMD lock. The second pass is only done on the + * PMD entries that first pass has found the accessed bit is set, namely + * 1) leaf entries mapping huge pages from the node under reclaim, and + * 2) non-leaf entries whose leaf entries only map pages from the node under + * reclaim, when CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG=y. + */ +static bool get_next_vma(struct mm_walk *walk, unsigned long mask, unsigned long size, + unsigned long *start, unsigned long *end) +{ + unsigned long next = round_up(*end, size); + struct mm_walk_args *args = walk->private; + + VM_BUG_ON(mask & size); + VM_BUG_ON(*start >= *end); + VM_BUG_ON((next & mask) != (*start & mask)); + + while (walk->vma) { + if (next >= walk->vma->vm_end) { + walk->vma = walk->vma->vm_next; + continue; + } + + if ((next & mask) != (walk->vma->vm_start & mask)) + return false; + + if (should_skip_vma(walk->vma->vm_start, walk->vma->vm_end, walk)) { + walk->vma = walk->vma->vm_next; + continue; + } + + *start = max(next, walk->vma->vm_start); + next = (next | ~mask) + 1; + /* rounded-up boundaries can wrap to 0 */ + *end = next && next < walk->vma->vm_end ? next : walk->vma->vm_end; + + args->mm_stats[MM_VMA_INTERVAL]++; + + return true; + } + + return false; +} + +static bool walk_pte_range(pmd_t *pmd, unsigned long start, unsigned long end, + struct mm_walk *walk) +{ + int i; + pte_t *pte; + spinlock_t *ptl; + unsigned long addr; + int remote = 0; + struct mm_walk_args *args = walk->private; + int old_gen, new_gen = lru_gen_from_seq(args->max_seq); + + VM_BUG_ON(pmd_leaf(*pmd)); + + pte = pte_offset_map_lock(walk->mm, pmd, start & PMD_MASK, &ptl); + arch_enter_lazy_mmu_mode(); +restart: + for (i = pte_index(start), addr = start; addr != end; i++, addr += PAGE_SIZE) { + struct page *page; + unsigned long pfn = pte_pfn(pte[i]); + + if (!pte_present(pte[i]) || is_zero_pfn(pfn)) { + args->mm_stats[MM_LEAF_HOLE]++; + continue; + } + + if (WARN_ON_ONCE(pte_devmap(pte[i]) || pte_special(pte[i]))) + continue; + + if (!pte_young(pte[i])) { + args->mm_stats[MM_LEAF_OLD]++; + continue; + } + + VM_BUG_ON(!pfn_valid(pfn)); + if (pfn < args->start_pfn || pfn >= args->end_pfn) { + args->mm_stats[MM_LEAF_OTHER_NODE]++; + remote++; + continue; + } + + page = compound_head(pfn_to_page(pfn)); + if (page_to_nid(page) != args->node_id) { + args->mm_stats[MM_LEAF_OTHER_NODE]++; + remote++; + continue; + } + + if (page_memcg_rcu(page) != args->memcg) { + args->mm_stats[MM_LEAF_OTHER_MEMCG]++; + continue; + } + + VM_BUG_ON(addr < walk->vma->vm_start || addr >= walk->vma->vm_end); + if (!ptep_test_and_clear_young(walk->vma, addr, pte + i)) + continue; + + if (pte_dirty(pte[i]) && !PageDirty(page) && + !(PageAnon(page) && PageSwapBacked(page) && !PageSwapCache(page))) { + set_page_dirty(page); + args->mm_stats[MM_LEAF_DIRTY]++; + } + + old_gen = page_update_gen(page, new_gen); + if (old_gen >= 0 && old_gen != new_gen) + update_batch_size(page, old_gen, new_gen, args); + args->mm_stats[MM_LEAF_YOUNG]++; + } + + if (i < PTRS_PER_PTE && get_next_vma(walk, PMD_MASK, PAGE_SIZE, &start, &end)) + goto restart; + + arch_leave_lazy_mmu_mode(); + pte_unmap_unlock(pte, ptl); + + return IS_ENABLED(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG) && !remote; +} + +#if defined(CONFIG_TRANSPARENT_HUGEPAGE) || defined(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG) +static void __walk_pmd_range(pud_t *pud, unsigned long start, + struct vm_area_struct *vma, struct mm_walk *walk) +{ + int i; + pmd_t *pmd; + spinlock_t *ptl; + struct mm_walk_args *args = walk->private; + int old_gen, new_gen = lru_gen_from_seq(args->max_seq); + + VM_BUG_ON(pud_leaf(*pud)); + + start &= PUD_MASK; + pmd = pmd_offset(pud, start); + ptl = pmd_lock(walk->mm, pmd); + arch_enter_lazy_mmu_mode(); + + for_each_set_bit(i, args->bitmap, PTRS_PER_PMD) { + struct page *page; + unsigned long pfn = pmd_pfn(pmd[i]); + unsigned long addr = start + i * PMD_SIZE; + + if (!pmd_present(pmd[i]) || is_huge_zero_pmd(pmd[i])) { + args->mm_stats[MM_LEAF_HOLE]++; + continue; + } + + if (WARN_ON_ONCE(pmd_devmap(pmd[i]))) + continue; + + if (!pmd_young(pmd[i])) { + args->mm_stats[MM_LEAF_OLD]++; + continue; + } + + if (!pmd_trans_huge(pmd[i])) { + if (IS_ENABLED(CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG) && + pmdp_test_and_clear_young(vma, addr, pmd + i)) + args->mm_stats[MM_NONLEAF_YOUNG]++; + continue; + } + + VM_BUG_ON(!pfn_valid(pfn)); + if (pfn < args->start_pfn || pfn >= args->end_pfn) { + args->mm_stats[MM_LEAF_OTHER_NODE]++; + continue; + } + + page = pfn_to_page(pfn); + VM_BUG_ON_PAGE(PageTail(page), page); + if (page_to_nid(page) != args->node_id) { + args->mm_stats[MM_LEAF_OTHER_NODE]++; + continue; + } + + if (page_memcg_rcu(page) != args->memcg) { + args->mm_stats[MM_LEAF_OTHER_MEMCG]++; + continue; + } + + VM_BUG_ON(addr < vma->vm_start || addr >= vma->vm_end); + if (!pmdp_test_and_clear_young(vma, addr, pmd + i)) + continue; + + if (pmd_dirty(pmd[i]) && !PageDirty(page) && + !(PageAnon(page) && PageSwapBacked(page) && !PageSwapCache(page))) { + set_page_dirty(page); + args->mm_stats[MM_LEAF_DIRTY]++; + } + + old_gen = page_update_gen(page, new_gen); + if (old_gen >= 0 && old_gen != new_gen) + update_batch_size(page, old_gen, new_gen, args); + args->mm_stats[MM_LEAF_YOUNG]++; + } + + arch_leave_lazy_mmu_mode(); + spin_unlock(ptl); + + bitmap_zero(args->bitmap, PTRS_PER_PMD); +} +#else +static void __walk_pmd_range(pud_t *pud, unsigned long start, + struct vm_area_struct *vma, struct mm_walk *walk) +{ +} +#endif + +static void walk_pmd_range(pud_t *pud, unsigned long start, unsigned long end, + struct mm_walk *walk) +{ + int i; + pmd_t *pmd; + unsigned long next; + unsigned long addr; + struct vm_area_struct *vma; + int leaf = 0; + int nonleaf = 0; + struct mm_walk_args *args = walk->private; + + VM_BUG_ON(pud_leaf(*pud)); + + pmd = pmd_offset(pud, start & PUD_MASK); +restart: + vma = walk->vma; + for (i = pmd_index(start), addr = start; addr != end; i++, addr = next) { + pmd_t val = pmd_read_atomic(pmd + i); + + /* for pmd_read_atomic() */ + barrier(); + + next = pmd_addr_end(addr, end); + + if (!pmd_present(val)) { + args->mm_stats[MM_LEAF_HOLE]++; + continue; + } + +#ifdef CONFIG_TRANSPARENT_HUGEPAGE + if (pmd_trans_huge(val)) { + unsigned long pfn = pmd_pfn(val); + + if (is_huge_zero_pmd(val)) { + args->mm_stats[MM_LEAF_HOLE]++; + continue; + } + + if (!pmd_young(val)) { + args->mm_stats[MM_LEAF_OLD]++; + continue; + } + + if (pfn < args->start_pfn || pfn >= args->end_pfn) { + args->mm_stats[MM_LEAF_OTHER_NODE]++; + continue; + } + + __set_bit(i, args->bitmap); + leaf++; + continue; + } +#endif + +#ifdef CONFIG_HAVE_ARCH_PARENT_PMD_YOUNG + if (!pmd_young(val)) { + args->mm_stats[MM_NONLEAF_OLD]++; + continue; + } +#endif + if (walk_pte_range(&val, addr, next, walk)) { + __set_bit(i, args->bitmap); + nonleaf++; + } + } + + if (leaf) { + __walk_pmd_range(pud, start, vma, walk); + leaf = nonleaf = 0; + } + + if (i < PTRS_PER_PMD && get_next_vma(walk, PUD_MASK, PMD_SIZE, &start, &end)) + goto restart; + + if (nonleaf) + __walk_pmd_range(pud, start, vma, walk); +} + +static int walk_pud_range(p4d_t *p4d, unsigned long start, unsigned long end, + struct mm_walk *walk) +{ + int i; + pud_t *pud; + unsigned long addr; + unsigned long next; + struct mm_walk_args *args = walk->private; + + VM_BUG_ON(p4d_leaf(*p4d)); + + pud = pud_offset(p4d, start & P4D_MASK); +restart: + for (i = pud_index(start), addr = start; addr != end; i++, addr = next) { + pud_t val = READ_ONCE(pud[i]); + + next = pud_addr_end(addr, end); + + if (!pud_present(val) || WARN_ON_ONCE(pud_leaf(val))) + continue; + + walk_pmd_range(&val, addr, next, walk); + + if (args->batch_size >= MAX_BATCH_SIZE) { + end = (addr | ~PUD_MASK) + 1; + goto done; + } + } + + if (i < PTRS_PER_PUD && get_next_vma(walk, P4D_MASK, PUD_SIZE, &start, &end)) + goto restart; + + end = round_up(end, P4D_SIZE); +done: + /* rounded-up boundaries can wrap to 0 */ + args->next_addr = end && walk->vma ? max(end, walk->vma->vm_start) : 0; + + return -EAGAIN; +} + +static void walk_mm(struct mm_walk_args *args, struct mm_struct *mm) +{ + static const struct mm_walk_ops mm_walk_ops = { + .test_walk = should_skip_vma, + .p4d_entry = walk_pud_range, + }; + + int err; + struct mem_cgroup *memcg = args->memcg; + struct lruvec *lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(args->node_id)); + + args->next_addr = FIRST_USER_ADDRESS; + + do { + unsigned long start = args->next_addr; + unsigned long end = mm->highest_vm_end; + + err = -EBUSY; + + preempt_disable(); + rcu_read_lock(); + +#ifdef CONFIG_MEMCG + if (memcg && atomic_read(&memcg->moving_account)) { + args->mm_stats[MM_LOCK_CONTENTION]++; + goto contended; + } +#endif + if (!mmap_read_trylock(mm)) { + args->mm_stats[MM_LOCK_CONTENTION]++; + goto contended; + } + + err = walk_page_range(mm, start, end, &mm_walk_ops, args); + + mmap_read_unlock(mm); + + reset_batch_size(lruvec, args); +contended: + rcu_read_unlock(); + preempt_enable(); + + cond_resched(); + } while (err == -EAGAIN && args->next_addr && + !mm_is_oom_victim(mm) && !mm_has_migrated(mm, memcg)); +} + +static void page_inc_gen(struct page *page, struct lruvec *lruvec, bool front) +{ + int old_gen, new_gen; + unsigned long old_flags, new_flags; + int type = page_is_file_lru(page); + int zone = page_zonenum(page); + struct lrugen *lrugen = &lruvec->evictable; + + old_gen = lru_gen_from_seq(lrugen->min_seq[type]); + + do { + old_flags = READ_ONCE(page->flags); + + /* in case the aging has updated old_gen */ + new_gen = ((old_flags & LRU_GEN_MASK) >> LRU_GEN_PGOFF) - 1; + VM_BUG_ON_PAGE(new_gen < 0, page); + if (new_gen >= 0 && new_gen != old_gen) + goto sort; + + new_gen = (old_gen + 1) % MAX_NR_GENS; + + new_flags = (old_flags & ~(LRU_GEN_MASK | LRU_USAGE_MASK | LRU_TIER_FLAGS)) | + ((new_gen + 1UL) << LRU_GEN_PGOFF); + /* mark the page for reclaim if it's pending writeback */ + if (front) + new_flags |= BIT(PG_reclaim); + } while (cmpxchg(&page->flags, old_flags, new_flags) != old_flags); + + lru_gen_update_size(page, lruvec, old_gen, new_gen); +sort: + if (front) + list_move(&page->lru, &lrugen->lists[new_gen][type][zone]); + else + list_move_tail(&page->lru, &lrugen->lists[new_gen][type][zone]); +} + +static bool try_inc_min_seq(struct lruvec *lruvec, int type) +{ + int gen, zone; + bool success = false; + struct lrugen *lrugen = &lruvec->evictable; + + VM_BUG_ON(!seq_is_valid(lruvec)); + + while (get_nr_gens(lruvec, type) > MIN_NR_GENS) { + gen = lru_gen_from_seq(lrugen->min_seq[type]); + + for (zone = 0; zone < MAX_NR_ZONES; zone++) { + if (!list_empty(&lrugen->lists[gen][type][zone])) + return success; + } + + reset_controller_pos(lruvec, gen, type); + WRITE_ONCE(lrugen->min_seq[type], lrugen->min_seq[type] + 1); + + success = true; + } + + return success; +} + +static bool inc_min_seq(struct lruvec *lruvec, int type) +{ + int gen, zone; + int batch_size = 0; + struct lrugen *lrugen = &lruvec->evictable; + + VM_BUG_ON(!seq_is_valid(lruvec)); + + if (get_nr_gens(lruvec, type) != MAX_NR_GENS) + return true; + + gen = lru_gen_from_seq(lrugen->min_seq[type]); + + for (zone = 0; zone < MAX_NR_ZONES; zone++) { + struct list_head *head = &lrugen->lists[gen][type][zone]; + + while (!list_empty(head)) { + struct page *page = lru_to_page(head); + + VM_BUG_ON_PAGE(PageTail(page), page); + VM_BUG_ON_PAGE(PageUnevictable(page), page); + VM_BUG_ON_PAGE(PageActive(page), page); + VM_BUG_ON_PAGE(page_is_file_lru(page) != type, page); + VM_BUG_ON_PAGE(page_zonenum(page) != zone, page); + + prefetchw_prev_lru_page(page, head, flags); + + page_inc_gen(page, lruvec, false); + + if (++batch_size == MAX_BATCH_SIZE) + return false; + } + + VM_BUG_ON(lrugen->sizes[gen][type][zone]); + } + + reset_controller_pos(lruvec, gen, type); + WRITE_ONCE(lrugen->min_seq[type], lrugen->min_seq[type] + 1); + + return true; +} + +static void inc_max_seq(struct lruvec *lruvec) +{ + int gen, type, zone; + struct lrugen *lrugen = &lruvec->evictable; + + spin_lock_irq(&lruvec->lru_lock); + + VM_BUG_ON(!seq_is_valid(lruvec)); + + for (type = 0; type < ANON_AND_FILE; type++) { + if (try_inc_min_seq(lruvec, type)) + continue; + + while (!inc_min_seq(lruvec, type)) { + spin_unlock_irq(&lruvec->lru_lock); + cond_resched(); + spin_lock_irq(&lruvec->lru_lock); + } + } + + gen = lru_gen_from_seq(lrugen->max_seq - 1); + for_each_type_zone(type, zone) { + enum lru_list lru = type * LRU_FILE; + long total = lrugen->sizes[gen][type][zone]; + + if (!total) + continue; + + WARN_ON_ONCE(total != (int)total); + + update_lru_size(lruvec, lru, zone, total); + update_lru_size(lruvec, lru + LRU_ACTIVE, zone, -total); + } + + gen = lru_gen_from_seq(lrugen->max_seq + 1); + for_each_type_zone(type, zone) { + VM_BUG_ON(lrugen->sizes[gen][type][zone]); + VM_BUG_ON(!list_empty(&lrugen->lists[gen][type][zone])); + } + + for (type = 0; type < ANON_AND_FILE; type++) + reset_controller_pos(lruvec, gen, type); + + WRITE_ONCE(lrugen->timestamps[gen], jiffies); + /* make sure all preceding modifications appear first */ + smp_store_release(&lrugen->max_seq, lrugen->max_seq + 1); + + spin_unlock_irq(&lruvec->lru_lock); +} + +/* Main function used by the foreground, the background and the user-triggered aging. */ +static bool walk_mm_list(struct lruvec *lruvec, unsigned long max_seq, + struct scan_control *sc, int swappiness, struct mm_walk_args *args) +{ + bool last; + bool alloc = !args; + struct mm_struct *mm = NULL; + struct lrugen *lrugen = &lruvec->evictable; + struct pglist_data *pgdat = lruvec_pgdat(lruvec); + int nid = pgdat->node_id; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + struct lru_gen_mm_list *mm_list = get_mm_list(memcg); + + VM_BUG_ON(max_seq > READ_ONCE(lrugen->max_seq)); + + if (alloc) { + args = kvzalloc_node(size_of_mm_walk_args(), GFP_KERNEL, nid); + if (WARN_ON_ONCE(!args)) + return false; + } + + args->memcg = memcg; + args->max_seq = max_seq; + args->start_pfn = pgdat->node_start_pfn; + args->end_pfn = pgdat_end_pfn(pgdat); + args->node_id = nid; + args->swappiness = swappiness; + + do { + last = get_next_mm(args, &mm); + if (mm) + walk_mm(args, mm); + + cond_resched(); + } while (mm); + + if (alloc) + kvfree(args); + + if (!last) { + /* the foreground aging prefers not to wait */ + if (!current_is_kswapd() && sc->priority < DEF_PRIORITY - 2) + wait_event_killable(mm_list->nodes[nid].wait, + max_seq < READ_ONCE(lrugen->max_seq)); + + return max_seq < READ_ONCE(lrugen->max_seq); + } + + VM_BUG_ON(max_seq != READ_ONCE(lrugen->max_seq)); + + inc_max_seq(lruvec); + + /* order against inc_max_seq() */ + smp_mb(); + /* either we see any waiters or they will see updated max_seq */ + if (waitqueue_active(&mm_list->nodes[nid].wait)) + wake_up_all(&mm_list->nodes[nid].wait); + + wakeup_flusher_threads(WB_REASON_VMSCAN); + + return true; +} + +void lru_gen_scan_around(struct page_vma_mapped_walk *pvmw) +{ + int i; + pte_t *pte; + int old_gen, new_gen; + unsigned long start; + unsigned long end; + unsigned long addr; + struct lruvec *lruvec; + struct mem_cgroup *memcg; + struct pglist_data *pgdat = page_pgdat(pvmw->page); + unsigned long bitmap[BITS_TO_LONGS(SWAP_CLUSTER_MAX * 2)] = {}; + + lockdep_assert_held(pvmw->ptl); + VM_BUG_ON_PAGE(PageTail(pvmw->page), pvmw->page); + + start = max(pvmw->address & PMD_MASK, pvmw->vma->vm_start); + end = pmd_addr_end(pvmw->address, pvmw->vma->vm_end); + + if (end - start > SWAP_CLUSTER_MAX * 2 * PAGE_SIZE) { + if (pvmw->address - start < SWAP_CLUSTER_MAX * PAGE_SIZE) + end = start + SWAP_CLUSTER_MAX * 2 * PAGE_SIZE; + else if (end - pvmw->address < SWAP_CLUSTER_MAX * PAGE_SIZE) + start = end - SWAP_CLUSTER_MAX * 2 * PAGE_SIZE; + else { + start = pvmw->address - SWAP_CLUSTER_MAX * PAGE_SIZE; + end = pvmw->address + SWAP_CLUSTER_MAX * PAGE_SIZE; + } + } + + pte = pvmw->pte - (pvmw->address - start) / PAGE_SIZE; + + arch_enter_lazy_mmu_mode(); + + lock_page_memcg(pvmw->page); + lruvec = lock_page_lruvec_irq(pvmw->page); + + memcg = page_memcg(pvmw->page); + new_gen = lru_gen_from_seq(lruvec->evictable.max_seq); + + for (i = 0, addr = start; addr != end; i++, addr += PAGE_SIZE) { + struct page *page; + unsigned long pfn = pte_pfn(pte[i]); + + if (!pte_present(pte[i]) || is_zero_pfn(pfn)) + continue; + + if (WARN_ON_ONCE(pte_devmap(pte[i]) || pte_special(pte[i]))) + continue; + + if (!pte_young(pte[i])) + continue; + + VM_BUG_ON(!pfn_valid(pfn)); + if (pfn < pgdat->node_start_pfn || pfn >= pgdat_end_pfn(pgdat)) + continue; + + page = compound_head(pfn_to_page(pfn)); + if (page_to_nid(page) != pgdat->node_id) + continue; + + if (page_memcg_rcu(page) != memcg) + continue; + + VM_BUG_ON(addr < pvmw->vma->vm_start || addr >= pvmw->vma->vm_end); + if (!ptep_test_and_clear_young(pvmw->vma, addr, pte + i)) + continue; + + if (pte_dirty(pte[i]) && !PageDirty(page) && + !(PageAnon(page) && PageSwapBacked(page) && !PageSwapCache(page))) + __set_bit(i, bitmap); + + old_gen = page_update_gen(page, new_gen); + if (old_gen >= 0 && old_gen != new_gen) + lru_gen_update_size(page, lruvec, old_gen, new_gen); + } + + unlock_page_lruvec_irq(lruvec); + unlock_page_memcg(pvmw->page); + + arch_leave_lazy_mmu_mode(); + + for_each_set_bit(i, bitmap, SWAP_CLUSTER_MAX * 2) + set_page_dirty(pte_page(pte[i])); +} + +/****************************************************************************** + * the eviction + ******************************************************************************/ + +static bool should_skip_page(struct page *page, struct scan_control *sc) +{ + if (!sc->may_unmap && page_mapped(page)) + return true; + + if (!(sc->may_writepage && (sc->gfp_mask & __GFP_IO)) && + (PageDirty(page) || (PageAnon(page) && !PageSwapCache(page)))) + return true; + + if (!get_page_unless_zero(page)) + return true; + + if (!TestClearPageLRU(page)) { + put_page(page); + return true; + } + + return false; +} + +static bool sort_page(struct page *page, struct lruvec *lruvec, int tier_to_isolate) +{ + bool success; + int gen = page_lru_gen(page); + int type = page_is_file_lru(page); + int zone = page_zonenum(page); + int tier = lru_tier_from_usage(page_tier_usage(page)); + struct lrugen *lrugen = &lruvec->evictable; + + VM_BUG_ON_PAGE(gen == -1, page); + VM_BUG_ON_PAGE(tier_to_isolate < 0, page); + + /* a lazy-free page that has been written into? */ + if (type && PageDirty(page) && PageAnon(page)) { + success = lru_gen_deletion(page, lruvec); + VM_BUG_ON_PAGE(!success, page); + SetPageSwapBacked(page); + add_page_to_lru_list_tail(page, lruvec); + return true; + } + + /* page_update_gen() has updated the gen #? */ + if (gen != lru_gen_from_seq(lrugen->min_seq[type])) { + list_move(&page->lru, &lrugen->lists[gen][type][zone]); + return true; + } + + /* activate this page if its tier has a higher refault rate */ + if (tier_to_isolate < tier) { + int hist = hist_from_seq_or_gen(gen); + + page_inc_gen(page, lruvec, false); + WRITE_ONCE(lrugen->activated[hist][type][tier - 1], + lrugen->activated[hist][type][tier - 1] + thp_nr_pages(page)); + inc_lruvec_state(lruvec, WORKINGSET_ACTIVATE_BASE + type); + return true; + } + + /* mark this page for reclaim if it's pending writeback */ + if (PageWriteback(page) || (type && PageDirty(page))) { + page_inc_gen(page, lruvec, true); + return true; + } + + return false; +} + +static void isolate_page(struct page *page, struct lruvec *lruvec) +{ + bool success; + + success = lru_gen_deletion(page, lruvec); + VM_BUG_ON_PAGE(!success, page); + + if (PageActive(page)) { + ClearPageActive(page); + /* make sure shrink_page_list() rejects this page */ + SetPageReferenced(page); + return; + } + + /* make sure shrink_page_list() doesn't try to write this page */ + ClearPageReclaim(page); + /* make sure shrink_page_list() doesn't reject this page */ + ClearPageReferenced(page); +} + +static int scan_pages(struct lruvec *lruvec, struct scan_control *sc, long *nr_to_scan, + int type, int tier, struct list_head *list) +{ + bool success; + int gen, zone; + enum vm_event_item item; + int sorted = 0; + int scanned = 0; + int isolated = 0; + int batch_size = 0; + struct lrugen *lrugen = &lruvec->evictable; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + + VM_BUG_ON(!list_empty(list)); + + if (get_nr_gens(lruvec, type) == MIN_NR_GENS) + return -ENOENT; + + gen = lru_gen_from_seq(lrugen->min_seq[type]); + + for (zone = sc->reclaim_idx; zone >= 0; zone--) { + LIST_HEAD(moved); + int skipped = 0; + struct list_head *head = &lrugen->lists[gen][type][zone]; + + while (!list_empty(head)) { + struct page *page = lru_to_page(head); + int delta = thp_nr_pages(page); + + VM_BUG_ON_PAGE(PageTail(page), page); + VM_BUG_ON_PAGE(PageUnevictable(page), page); + VM_BUG_ON_PAGE(PageActive(page), page); + VM_BUG_ON_PAGE(page_is_file_lru(page) != type, page); + VM_BUG_ON_PAGE(page_zonenum(page) != zone, page); + + prefetchw_prev_lru_page(page, head, flags); + + scanned += delta; + + if (sort_page(page, lruvec, tier)) + sorted += delta; + else if (should_skip_page(page, sc)) { + list_move(&page->lru, &moved); + skipped += delta; + } else { + isolate_page(page, lruvec); + list_add(&page->lru, list); + isolated += delta; + } + + if (isolated >= SWAP_CLUSTER_MAX || ++batch_size == MAX_BATCH_SIZE) + break; + } + + list_splice(&moved, head); + __count_zid_vm_events(PGSCAN_SKIP, zone, skipped); + + if (isolated >= SWAP_CLUSTER_MAX || batch_size == MAX_BATCH_SIZE) + break; + } + + success = try_inc_min_seq(lruvec, type); + + item = current_is_kswapd() ? PGSCAN_KSWAPD : PGSCAN_DIRECT; + if (!cgroup_reclaim(sc)) { + __count_vm_events(item, scanned); + __count_vm_events(PGREFILL, sorted); + } + __count_memcg_events(memcg, item, scanned); + __count_memcg_events(memcg, PGREFILL, sorted); + __count_vm_events(PGSCAN_ANON + type, scanned); + + *nr_to_scan -= scanned; + + if (*nr_to_scan <= 0 || success || isolated) + return isolated; + /* + * We may have trouble finding eligible pages due to reclaim_idx, + * may_unmap and may_writepage. The following check makes sure we won't + * be stuck if we aren't making enough progress. + */ + return batch_size == MAX_BATCH_SIZE && sorted >= scanned / 2 ? 0 : -ENOENT; +} + +static int get_tier_to_isolate(struct lruvec *lruvec, int type) +{ + int tier; + struct controller_pos sp, pv; + + /* + * Ideally we don't want to evict upper tiers that have higher refault + * rates. However, we need to leave a margin for the fluctuations in + * refault rates. So we use a larger gain factor to make sure upper + * tiers are indeed more active. We choose 2 because the lowest upper + * tier would have twice of the refault rate of the base tier, according + * to their numbers of accesses. + */ + read_controller_pos(&sp, lruvec, type, 0, 1); + for (tier = 1; tier < MAX_NR_TIERS; tier++) { + read_controller_pos(&pv, lruvec, type, tier, 2); + if (!positive_ctrl_err(&sp, &pv)) + break; + } + + return tier - 1; +} + +static int get_type_to_scan(struct lruvec *lruvec, int swappiness, int *tier_to_isolate) +{ + int type, tier; + struct controller_pos sp, pv; + int gain[ANON_AND_FILE] = { swappiness, 200 - swappiness }; + + /* + * Compare the refault rates between the base tiers of anon and file to + * determine which type to evict. Also need to compare the refault rates + * of the upper tiers of the selected type with that of the base tier of + * the other type to determine which tier of the selected type to evict. + */ + read_controller_pos(&sp, lruvec, 0, 0, gain[0]); + read_controller_pos(&pv, lruvec, 1, 0, gain[1]); + type = positive_ctrl_err(&sp, &pv); + + read_controller_pos(&sp, lruvec, !type, 0, gain[!type]); + for (tier = 1; tier < MAX_NR_TIERS; tier++) { + read_controller_pos(&pv, lruvec, type, tier, gain[type]); + if (!positive_ctrl_err(&sp, &pv)) + break; + } + + *tier_to_isolate = tier - 1; + + return type; +} + +static int isolate_pages(struct lruvec *lruvec, struct scan_control *sc, int swappiness, + long *nr_to_scan, int *type_to_scan, struct list_head *list) +{ + int i; + int type; + int isolated; + int tier = -1; + DEFINE_MAX_SEQ(); + DEFINE_MIN_SEQ(); + + VM_BUG_ON(!seq_is_valid(lruvec)); + + if (max_nr_gens(max_seq, min_seq, swappiness) == MIN_NR_GENS) + return 0; + /* + * Try to select a type based on generations and swappiness, and if that + * fails, fall back to get_type_to_scan(). When anon and file are both + * available from the same generation, swappiness 200 is interpreted as + * anon first and swappiness 1 is interpreted as file first. + */ + type = !swappiness || min_seq[0] > min_seq[1] || + (min_seq[0] == min_seq[1] && swappiness != 200 && + (swappiness == 1 || get_type_to_scan(lruvec, swappiness, &tier))); + + if (tier == -1) + tier = get_tier_to_isolate(lruvec, type); + + for (i = !swappiness; i < ANON_AND_FILE; i++) { + isolated = scan_pages(lruvec, sc, nr_to_scan, type, tier, list); + if (isolated >= 0) + break; + + type = !type; + tier = get_tier_to_isolate(lruvec, type); + } + + if (isolated < 0) + isolated = *nr_to_scan = 0; + + *type_to_scan = type; + + return isolated; +} + +/* Main function used by the foreground, the background and the user-triggered eviction. */ +static bool evict_pages(struct lruvec *lruvec, struct scan_control *sc, int swappiness, + long *nr_to_scan) +{ + int type; + int isolated; + int reclaimed; + LIST_HEAD(list); + struct page *page; + enum vm_event_item item; + struct reclaim_stat stat; + struct pglist_data *pgdat = lruvec_pgdat(lruvec); + + spin_lock_irq(&lruvec->lru_lock); + + isolated = isolate_pages(lruvec, sc, swappiness, nr_to_scan, &type, &list); + VM_BUG_ON(list_empty(&list) == !!isolated); + + if (isolated) + __mod_node_page_state(pgdat, NR_ISOLATED_ANON + type, isolated); + + spin_unlock_irq(&lruvec->lru_lock); + + if (!isolated) + goto done; + + reclaimed = shrink_page_list(&list, pgdat, sc, &stat, false); + /* + * We need to prevent rejected pages from being added back to the same + * lists they were isolated from. Otherwise we may risk looping on them + * forever. We use PageActive() or !PageReferenced() && PageWorkingset() + * to tell lru_gen_addition() not to add them to the oldest generation. + */ + list_for_each_entry(page, &list, lru) { + if (PageMlocked(page)) + continue; + + if (page_mapped(page) && PageReferenced(page)) + SetPageActive(page); + else { + ClearPageActive(page); + SetPageWorkingset(page); + } + ClearPageReferenced(page); + } + + spin_lock_irq(&lruvec->lru_lock); + + move_pages_to_lru(lruvec, &list); + + __mod_node_page_state(pgdat, NR_ISOLATED_ANON + type, -isolated); + + item = current_is_kswapd() ? PGSTEAL_KSWAPD : PGSTEAL_DIRECT; + if (!cgroup_reclaim(sc)) + __count_vm_events(item, reclaimed); + __count_memcg_events(lruvec_memcg(lruvec), item, reclaimed); + __count_vm_events(PGSTEAL_ANON + type, reclaimed); + + spin_unlock_irq(&lruvec->lru_lock); + + mem_cgroup_uncharge_list(&list); + free_unref_page_list(&list); + + sc->nr_reclaimed += reclaimed; +done: + return *nr_to_scan > 0 && sc->nr_reclaimed < sc->nr_to_reclaim; +} + +/****************************************************************************** + * page reclaim + ******************************************************************************/ + +static int get_swappiness(struct lruvec *lruvec) +{ + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + int swappiness = mem_cgroup_get_nr_swap_pages(memcg) >= (long)SWAP_CLUSTER_MAX ? + mem_cgroup_swappiness(memcg) : 0; + + VM_BUG_ON(swappiness > 200U); + + return swappiness; +} + +static DEFINE_RATELIMIT_STATE(lru_gen_min_ttl, 0, 1); + +static unsigned long get_nr_to_scan(struct lruvec *lruvec, struct scan_control *sc, + int swappiness) +{ + int gen, type, zone; + int nr_gens; + long nr_to_scan = 0; + struct lrugen *lrugen = &lruvec->evictable; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + DEFINE_MAX_SEQ(); + DEFINE_MIN_SEQ(); + + for (type = !swappiness; type < ANON_AND_FILE; type++) { + unsigned long seq; + + for (seq = min_seq[type]; seq <= max_seq; seq++) { + gen = lru_gen_from_seq(seq); + + for (zone = 0; zone <= sc->reclaim_idx; zone++) + nr_to_scan += READ_ONCE(lrugen->sizes[gen][type][zone]); + } + } + + + if (nr_to_scan <= 0) + return 0; + + nr_gens = max_nr_gens(max_seq, min_seq, swappiness); + + if (current_is_kswapd()) { + gen = lru_gen_from_seq(max_seq - nr_gens + 1); + if (time_is_before_eq_jiffies(READ_ONCE(lrugen->timestamps[gen]) + + READ_ONCE(lru_gen_min_ttl.interval))) + sc->file_is_tiny = 0; + + /* leave the work to age_lru_gens() */ + if (nr_gens == MIN_NR_GENS) + return 0; + + if (nr_to_scan >= sc->nr_to_reclaim) + sc->force_deactivate = 0; + } + + nr_to_scan = max(nr_to_scan >> sc->priority, (long)!mem_cgroup_online(memcg)); + if (!nr_to_scan || nr_gens > MIN_NR_GENS) + return nr_to_scan; + + /* move onto other memcgs if we haven't tried them all yet */ + if (memcg && !sc->force_deactivate) { + sc->skipped_deactivate = 1; + return 0; + } + + return walk_mm_list(lruvec, max_seq, sc, swappiness, NULL) ? nr_to_scan : 0; +} + +static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc) +{ + struct blk_plug plug; + unsigned long scanned = 0; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + + lru_add_drain(); + + blk_start_plug(&plug); + + while (true) { + long nr_to_scan; + int swappiness = sc->may_swap ? get_swappiness(lruvec) : 0; + + nr_to_scan = get_nr_to_scan(lruvec, sc, swappiness) - scanned; + if (nr_to_scan <= 0) + break; + + scanned += nr_to_scan; + + if (!evict_pages(lruvec, sc, swappiness, &nr_to_scan)) + break; + + scanned -= nr_to_scan; + + if (mem_cgroup_below_min(memcg) || + (mem_cgroup_below_low(memcg) && !sc->memcg_low_reclaim)) + break; + + cond_resched(); + } + + blk_finish_plug(&plug); +} + +/****************************************************************************** + * the background aging + ******************************************************************************/ + +static int lru_gen_spread = MIN_NR_GENS; + +static void try_walk_mm_list(struct lruvec *lruvec, struct scan_control *sc) +{ + int gen, type, zone; + long old_and_young[2] = {}; + int spread = READ_ONCE(lru_gen_spread); + int swappiness = get_swappiness(lruvec); + struct lrugen *lrugen = &lruvec->evictable; + struct pglist_data *pgdat = lruvec_pgdat(lruvec); + DEFINE_MAX_SEQ(); + DEFINE_MIN_SEQ(); + + lru_add_drain(); + + for (type = !swappiness; type < ANON_AND_FILE; type++) { + unsigned long seq; + + for (seq = min_seq[type]; seq <= max_seq; seq++) { + gen = lru_gen_from_seq(seq); + + for (zone = 0; zone < MAX_NR_ZONES; zone++) + old_and_young[seq == max_seq] += + READ_ONCE(lrugen->sizes[gen][type][zone]); + } + } + + old_and_young[0] = max(old_and_young[0], 0L); + old_and_young[1] = max(old_and_young[1], 0L); + + /* try to spread pages out across spread+1 generations */ + if (old_and_young[0] >= old_and_young[1] * spread && + min_nr_gens(max_seq, min_seq, swappiness) > max(spread, MIN_NR_GENS)) + return; + + walk_mm_list(lruvec, max_seq, sc, swappiness, pgdat->mm_walk_args); +} + +static void lru_gen_age_node(struct pglist_data *pgdat, struct scan_control *sc) +{ + struct mem_cgroup *memcg; + + VM_BUG_ON(!current_is_kswapd()); + + if (sc->file_is_tiny && READ_ONCE(lru_gen_min_ttl.interval) && + __ratelimit(&lru_gen_min_ttl)) { + struct oom_control oc = { + .gfp_mask = sc->gfp_mask, + .order = sc->order, + }; + + if (mutex_trylock(&oom_lock)) { + out_of_memory(&oc); + mutex_unlock(&oom_lock); + } + } + + sc->file_is_tiny = 1; + + if (!mem_cgroup_disabled() && !sc->force_deactivate) { + /* we may clear this later in get_nr_to_scan() */ + sc->force_deactivate = 1; + return; + } + + sc->force_deactivate = 0; + + memcg = mem_cgroup_iter(NULL, NULL, NULL); + do { + struct lruvec *lruvec = mem_cgroup_lruvec(memcg, pgdat); + + if (!mem_cgroup_below_min(memcg) && + (!mem_cgroup_below_low(memcg) || sc->memcg_low_reclaim)) + try_walk_mm_list(lruvec, sc); + + cond_resched(); + } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL))); +} + +/****************************************************************************** + * state change + ******************************************************************************/ + +#ifdef CONFIG_LRU_GEN_ENABLED +DEFINE_STATIC_KEY_TRUE(lru_gen_static_key); +#else +DEFINE_STATIC_KEY_FALSE(lru_gen_static_key); +#endif + +static DEFINE_MUTEX(lru_gen_state_mutex); +static int lru_gen_nr_swapfiles __read_mostly; + +static bool __maybe_unused state_is_valid(struct lruvec *lruvec) +{ + int gen, type, zone; + enum lru_list lru; + struct lrugen *lrugen = &lruvec->evictable; + + for_each_evictable_lru(lru) { + type = is_file_lru(lru); + + if (lrugen->enabled[type] && !list_empty(&lruvec->lists[lru])) + return false; + } + + for_each_gen_type_zone(gen, type, zone) { + if (!lrugen->enabled[type] && !list_empty(&lrugen->lists[gen][type][zone])) + return false; + + VM_WARN_ON_ONCE(!lrugen->enabled[type] && lrugen->sizes[gen][type][zone]); + } + + return true; +} + +static bool fill_lru_gen_lists(struct lruvec *lruvec) +{ + enum lru_list lru; + int batch_size = 0; + + for_each_evictable_lru(lru) { + int type = is_file_lru(lru); + bool active = is_active_lru(lru); + struct list_head *head = &lruvec->lists[lru]; + + if (!lruvec->evictable.enabled[type]) + continue; + + while (!list_empty(head)) { + bool success; + struct page *page = lru_to_page(head); + + VM_BUG_ON_PAGE(PageTail(page), page); + VM_BUG_ON_PAGE(PageUnevictable(page), page); + VM_BUG_ON_PAGE(PageActive(page) != active, page); + VM_BUG_ON_PAGE(page_lru_gen(page) != -1, page); + VM_BUG_ON_PAGE(page_is_file_lru(page) != type, page); + + prefetchw_prev_lru_page(page, head, flags); + + del_page_from_lru_list(page, lruvec); + success = lru_gen_addition(page, lruvec, true); + VM_BUG_ON(!success); + + if (++batch_size == MAX_BATCH_SIZE) + return false; + } + } + + return true; +} + +static bool drain_lru_gen_lists(struct lruvec *lruvec) +{ + int gen, type, zone; + int batch_size = 0; + + for_each_gen_type_zone(gen, type, zone) { + struct list_head *head = &lruvec->evictable.lists[gen][type][zone]; + + if (lruvec->evictable.enabled[type]) + continue; + + while (!list_empty(head)) { + bool success; + struct page *page = lru_to_page(head); + + VM_BUG_ON_PAGE(PageTail(page), page); + VM_BUG_ON_PAGE(PageUnevictable(page), page); + VM_BUG_ON_PAGE(PageActive(page), page); + VM_BUG_ON_PAGE(page_is_file_lru(page) != type, page); + VM_BUG_ON_PAGE(page_zonenum(page) != zone, page); + + prefetchw_prev_lru_page(page, head, flags); + + success = lru_gen_deletion(page, lruvec); + VM_BUG_ON(!success); + add_page_to_lru_list(page, lruvec); + + if (++batch_size == MAX_BATCH_SIZE) + return false; + } + } + + return true; +} + +/* + * For file page tracking, we enable/disable it according to the main switch. + * For anon page tracking, we only enabled it when the main switch is on and + * there is at least one swapfile; we disable it when there are no swapfiles + * regardless of the value of the main switch. Otherwise, we will eventually + * reach the max size of the sliding window and have to call inc_min_seq(), + * which brings an unnecessary overhead. + */ +void lru_gen_set_state(bool enable, bool main, bool swap) +{ + struct mem_cgroup *memcg; + + mem_hotplug_begin(); + mutex_lock(&lru_gen_state_mutex); + cgroup_lock(); + + main = main && enable != lru_gen_enabled(); + swap = swap && !(enable ? lru_gen_nr_swapfiles++ : --lru_gen_nr_swapfiles); + swap = swap && lru_gen_enabled(); + if (!main && !swap) + goto unlock; + + if (main) { + if (enable) + static_branch_enable(&lru_gen_static_key); + else + static_branch_disable(&lru_gen_static_key); + } + + memcg = mem_cgroup_iter(NULL, NULL, NULL); + do { + int nid; + + for_each_node_state(nid, N_MEMORY) { + struct lruvec *lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(nid)); + struct lrugen *lrugen = &lruvec->evictable; + + spin_lock_irq(&lruvec->lru_lock); + + VM_BUG_ON(!seq_is_valid(lruvec)); + VM_BUG_ON(!state_is_valid(lruvec)); + + WRITE_ONCE(lrugen->enabled[0], lru_gen_enabled() && lru_gen_nr_swapfiles); + WRITE_ONCE(lrugen->enabled[1], lru_gen_enabled()); + + while (!(enable ? fill_lru_gen_lists(lruvec) : + drain_lru_gen_lists(lruvec))) { + spin_unlock_irq(&lruvec->lru_lock); + cond_resched(); + spin_lock_irq(&lruvec->lru_lock); + } + + spin_unlock_irq(&lruvec->lru_lock); + } + + cond_resched(); + } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL))); +unlock: + cgroup_unlock(); + mutex_unlock(&lru_gen_state_mutex); + mem_hotplug_done(); +} + +static int __meminit __maybe_unused lru_gen_online_mem(struct notifier_block *self, + unsigned long action, void *arg) +{ + struct mem_cgroup *memcg; + struct memory_notify *mnb = arg; + int nid = mnb->status_change_nid; + + if (action != MEM_GOING_ONLINE || nid == NUMA_NO_NODE) + return NOTIFY_DONE; + + mutex_lock(&lru_gen_state_mutex); + cgroup_lock(); + + memcg = mem_cgroup_iter(NULL, NULL, NULL); + do { + struct lruvec *lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(nid)); + struct lrugen *lrugen = &lruvec->evictable; + + VM_BUG_ON(!seq_is_valid(lruvec)); + VM_BUG_ON(!state_is_valid(lruvec)); + + WRITE_ONCE(lrugen->enabled[0], lru_gen_enabled() && lru_gen_nr_swapfiles); + WRITE_ONCE(lrugen->enabled[1], lru_gen_enabled()); + } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL))); + + cgroup_unlock(); + mutex_unlock(&lru_gen_state_mutex); + + return NOTIFY_DONE; +} + +static void lru_gen_start_kswapd(int nid) +{ + struct pglist_data *pgdat = NODE_DATA(nid); + + pgdat->mm_walk_args = kvzalloc_node(size_of_mm_walk_args(), GFP_KERNEL, nid); + WARN_ON_ONCE(!pgdat->mm_walk_args); +} + +static void lru_gen_stop_kswapd(int nid) +{ + struct pglist_data *pgdat = NODE_DATA(nid); + + kvfree(pgdat->mm_walk_args); +} + +/****************************************************************************** + * sysfs interface + ******************************************************************************/ + +static ssize_t show_lru_gen_spread(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%d\n", READ_ONCE(lru_gen_spread)); +} + +static ssize_t store_lru_gen_spread(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t len) +{ + int spread; + + if (kstrtoint(buf, 10, &spread) || spread >= MAX_NR_GENS) + return -EINVAL; + + WRITE_ONCE(lru_gen_spread, spread); + + return len; +} + +static struct kobj_attribute lru_gen_spread_attr = __ATTR( + spread, 0644, show_lru_gen_spread, store_lru_gen_spread +); + +static ssize_t show_lru_gen_min_ttl(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return sprintf(buf, "%u\n", jiffies_to_msecs(READ_ONCE(lru_gen_min_ttl.interval))); +} + +static ssize_t store_lru_gen_min_ttl(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t len) +{ + unsigned int msecs; + + if (kstrtouint(buf, 10, &msecs)) + return -EINVAL; + + WRITE_ONCE(lru_gen_min_ttl.interval, msecs_to_jiffies(msecs)); + + return len; +} + +static struct kobj_attribute lru_gen_min_ttl_attr = __ATTR( + min_ttl_ms, 0644, show_lru_gen_min_ttl, store_lru_gen_min_ttl +); + +static ssize_t show_lru_gen_enabled(struct kobject *kobj, struct kobj_attribute *attr, + char *buf) +{ + return snprintf(buf, PAGE_SIZE, "%d\n", lru_gen_enabled()); +} + +static ssize_t store_lru_gen_enabled(struct kobject *kobj, struct kobj_attribute *attr, + const char *buf, size_t len) +{ + int enable; + + if (kstrtoint(buf, 10, &enable)) + return -EINVAL; + + lru_gen_set_state(enable, true, false); + + return len; +} + +static struct kobj_attribute lru_gen_enabled_attr = __ATTR( + enabled, 0644, show_lru_gen_enabled, store_lru_gen_enabled +); + +static struct attribute *lru_gen_attrs[] = { + &lru_gen_spread_attr.attr, + &lru_gen_min_ttl_attr.attr, + &lru_gen_enabled_attr.attr, + NULL +}; + +static struct attribute_group lru_gen_attr_group = { + .name = "lru_gen", + .attrs = lru_gen_attrs, +}; + +/****************************************************************************** + * debugfs interface + ******************************************************************************/ + +static void *lru_gen_seq_start(struct seq_file *m, loff_t *pos) +{ + struct mem_cgroup *memcg; + loff_t nr_to_skip = *pos; + + m->private = kzalloc(PATH_MAX, GFP_KERNEL); + if (!m->private) + return ERR_PTR(-ENOMEM); + + memcg = mem_cgroup_iter(NULL, NULL, NULL); + do { + int nid; + + for_each_node_state(nid, N_MEMORY) { + if (!nr_to_skip--) + return mem_cgroup_lruvec(memcg, NODE_DATA(nid)); + } + } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL))); + + return NULL; +} + +static void lru_gen_seq_stop(struct seq_file *m, void *v) +{ + if (!IS_ERR_OR_NULL(v)) + mem_cgroup_iter_break(NULL, lruvec_memcg(v)); + + kfree(m->private); + m->private = NULL; +} + +static void *lru_gen_seq_next(struct seq_file *m, void *v, loff_t *pos) +{ + int nid = lruvec_pgdat(v)->node_id; + struct mem_cgroup *memcg = lruvec_memcg(v); + + ++*pos; + + nid = next_memory_node(nid); + if (nid == MAX_NUMNODES) { + memcg = mem_cgroup_iter(NULL, memcg, NULL); + if (!memcg) + return NULL; + + nid = first_memory_node; + } + + return mem_cgroup_lruvec(memcg, NODE_DATA(nid)); +} + +static void lru_gen_seq_show_full(struct seq_file *m, struct lruvec *lruvec, + unsigned long max_seq, unsigned long *min_seq, + unsigned long seq) +{ + int i; + int type, tier; + int hist = hist_from_seq_or_gen(seq); + struct lrugen *lrugen = &lruvec->evictable; + int nid = lruvec_pgdat(lruvec)->node_id; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + struct lru_gen_mm_list *mm_list = get_mm_list(memcg); + + for (tier = 0; tier < MAX_NR_TIERS; tier++) { + seq_printf(m, " %10d", tier); + for (type = 0; type < ANON_AND_FILE; type++) { + unsigned long n[3] = {}; + + if (seq == max_seq) { + n[0] = READ_ONCE(lrugen->avg_refaulted[type][tier]); + n[1] = READ_ONCE(lrugen->avg_total[type][tier]); + + seq_printf(m, " %10luR %10luT %10lu ", n[0], n[1], n[2]); + } else if (seq == min_seq[type] || NR_STAT_GENS > 1) { + n[0] = atomic_long_read(&lrugen->refaulted[hist][type][tier]); + n[1] = atomic_long_read(&lrugen->evicted[hist][type][tier]); + if (tier) + n[2] = READ_ONCE(lrugen->activated[hist][type][tier - 1]); + + seq_printf(m, " %10lur %10lue %10lua", n[0], n[1], n[2]); + } else + seq_puts(m, " 0 0 0 "); + } + seq_putc(m, '\n'); + } + + seq_puts(m, " "); + for (i = 0; i < NR_MM_STATS; i++) { + if (seq == max_seq && NR_STAT_GENS == 1) + seq_printf(m, " %10lu%c", READ_ONCE(mm_list->nodes[nid].stats[hist][i]), + toupper(MM_STAT_CODES[i])); + else if (seq != max_seq && NR_STAT_GENS > 1) + seq_printf(m, " %10lu%c", READ_ONCE(mm_list->nodes[nid].stats[hist][i]), + MM_STAT_CODES[i]); + else + seq_puts(m, " 0 "); + } + seq_putc(m, '\n'); +} + +static int lru_gen_seq_show(struct seq_file *m, void *v) +{ + unsigned long seq; + bool full = !debugfs_real_fops(m->file)->write; + struct lruvec *lruvec = v; + struct lrugen *lrugen = &lruvec->evictable; + int nid = lruvec_pgdat(lruvec)->node_id; + struct mem_cgroup *memcg = lruvec_memcg(lruvec); + DEFINE_MAX_SEQ(); + DEFINE_MIN_SEQ(); + + if (nid == first_memory_node) { +#ifdef CONFIG_MEMCG + if (memcg) + cgroup_path(memcg->css.cgroup, m->private, PATH_MAX); +#endif + seq_printf(m, "memcg %5hu %s\n", mem_cgroup_id(memcg), (char *)m->private); + } + + seq_printf(m, " node %5d\n", nid); + + seq = full ? (max_seq < MAX_NR_GENS ? 0 : max_seq - MAX_NR_GENS + 1) : + min(min_seq[0], min_seq[1]); + + for (; seq <= max_seq; seq++) { + int gen, type, zone; + unsigned int msecs; + + gen = lru_gen_from_seq(seq); + msecs = jiffies_to_msecs(jiffies - READ_ONCE(lrugen->timestamps[gen])); + + seq_printf(m, " %10lu %10u", seq, msecs); + + for (type = 0; type < ANON_AND_FILE; type++) { + long size = 0; + + if (seq < min_seq[type]) { + seq_puts(m, " -0 "); + continue; + } + + for (zone = 0; zone < MAX_NR_ZONES; zone++) + size += READ_ONCE(lrugen->sizes[gen][type][zone]); + + seq_printf(m, " %10lu ", max(size, 0L)); + } + + seq_putc(m, '\n'); + + if (full) + lru_gen_seq_show_full(m, lruvec, max_seq, min_seq, seq); + } + + return 0; +} + +static const struct seq_operations lru_gen_seq_ops = { + .start = lru_gen_seq_start, + .stop = lru_gen_seq_stop, + .next = lru_gen_seq_next, + .show = lru_gen_seq_show, +}; + +static int advance_max_seq(struct lruvec *lruvec, unsigned long seq, int swappiness) +{ + struct scan_control sc = { + .target_mem_cgroup = lruvec_memcg(lruvec), + }; + DEFINE_MAX_SEQ(); + + if (seq == max_seq) + walk_mm_list(lruvec, max_seq, &sc, swappiness, NULL); + + return seq > max_seq ? -EINVAL : 0; +} + +static int advance_min_seq(struct lruvec *lruvec, unsigned long seq, int swappiness, + unsigned long nr_to_reclaim) +{ + struct blk_plug plug; + int err = -EINTR; + long nr_to_scan = LONG_MAX; + struct scan_control sc = { + .nr_to_reclaim = nr_to_reclaim, + .target_mem_cgroup = lruvec_memcg(lruvec), + .may_writepage = 1, + .may_unmap = 1, + .may_swap = 1, + .reclaim_idx = MAX_NR_ZONES - 1, + .gfp_mask = GFP_KERNEL, + }; + DEFINE_MAX_SEQ(); + + if (seq >= max_seq - 1) + return -EINVAL; + + blk_start_plug(&plug); + + while (!signal_pending(current)) { + DEFINE_MIN_SEQ(); + + if (seq < min(min_seq[!swappiness], min_seq[swappiness < 200]) || + !evict_pages(lruvec, &sc, swappiness, &nr_to_scan)) { + err = 0; + break; + } + + cond_resched(); + } + + blk_finish_plug(&plug); + + return err; +} + +static int advance_seq(char cmd, int memcg_id, int nid, unsigned long seq, + int swappiness, unsigned long nr_to_reclaim) +{ + struct lruvec *lruvec; + int err = -EINVAL; + struct mem_cgroup *memcg = NULL; + + if (!mem_cgroup_disabled()) { + rcu_read_lock(); + memcg = mem_cgroup_from_id(memcg_id); +#ifdef CONFIG_MEMCG + if (memcg && !css_tryget(&memcg->css)) + memcg = NULL; +#endif + rcu_read_unlock(); + + if (!memcg) + goto done; + } + if (memcg_id != mem_cgroup_id(memcg)) + goto done; + + if (nid < 0 || nid >= MAX_NUMNODES || !node_state(nid, N_MEMORY)) + goto done; + + lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(nid)); + + if (swappiness == -1) + swappiness = get_swappiness(lruvec); + else if (swappiness > 200U) + goto done; + + switch (cmd) { + case '+': + err = advance_max_seq(lruvec, seq, swappiness); + break; + case '-': + err = advance_min_seq(lruvec, seq, swappiness, nr_to_reclaim); + break; + } +done: + mem_cgroup_put(memcg); + + return err; +} + +static ssize_t lru_gen_seq_write(struct file *file, const char __user *src, + size_t len, loff_t *pos) +{ + void *buf; + char *cur, *next; + int err = 0; + + buf = kvmalloc(len + 1, GFP_USER); + if (!buf) + return -ENOMEM; + + if (copy_from_user(buf, src, len)) { + kvfree(buf); + return -EFAULT; + } + + next = buf; + next[len] = '\0'; + + while ((cur = strsep(&next, ",;\n"))) { + int n; + int end; + char cmd; + unsigned int memcg_id; + unsigned int nid; + unsigned long seq; + unsigned int swappiness = -1; + unsigned long nr_to_reclaim = -1; + + cur = skip_spaces(cur); + if (!*cur) + continue; + + n = sscanf(cur, "%c %u %u %lu %n %u %n %lu %n", &cmd, &memcg_id, &nid, + &seq, &end, &swappiness, &end, &nr_to_reclaim, &end); + if (n < 4 || cur[end]) { + err = -EINVAL; + break; + } + + err = advance_seq(cmd, memcg_id, nid, seq, swappiness, nr_to_reclaim); + if (err) + break; + } + + kvfree(buf); + + return err ? : len; +} + +static int lru_gen_seq_open(struct inode *inode, struct file *file) +{ + return seq_open(file, &lru_gen_seq_ops); +} + +static const struct file_operations lru_gen_rw_fops = { + .open = lru_gen_seq_open, + .read = seq_read, + .write = lru_gen_seq_write, + .llseek = seq_lseek, + .release = seq_release, +}; + +static const struct file_operations lru_gen_ro_fops = { + .open = lru_gen_seq_open, + .read = seq_read, + .llseek = seq_lseek, + .release = seq_release, +}; + +/****************************************************************************** + * initialization + ******************************************************************************/ + +void lru_gen_init_lruvec(struct lruvec *lruvec) +{ + int i; + int gen, type, zone; + struct lrugen *lrugen = &lruvec->evictable; + + lrugen->max_seq = MIN_NR_GENS + 1; + lrugen->enabled[0] = lru_gen_enabled() && lru_gen_nr_swapfiles; + lrugen->enabled[1] = lru_gen_enabled(); + + for (i = 0; i <= MIN_NR_GENS + 1; i++) + lrugen->timestamps[i] = jiffies; + + for_each_gen_type_zone(gen, type, zone) + INIT_LIST_HEAD(&lrugen->lists[gen][type][zone]); +} + +static int __init init_lru_gen(void) +{ + BUILD_BUG_ON(MIN_NR_GENS + 1 >= MAX_NR_GENS); + BUILD_BUG_ON(BIT(LRU_GEN_WIDTH) <= MAX_NR_GENS); + BUILD_BUG_ON(sizeof(MM_STAT_CODES) != NR_MM_STATS + 1); + + VM_BUG_ON(PMD_SIZE / PAGE_SIZE != PTRS_PER_PTE); + VM_BUG_ON(PUD_SIZE / PMD_SIZE != PTRS_PER_PMD); + VM_BUG_ON(P4D_SIZE / PUD_SIZE != PTRS_PER_PUD); + + if (mem_cgroup_disabled()) { + global_mm_list = alloc_mm_list(); + if (WARN_ON_ONCE(!global_mm_list)) + return -ENOMEM; + } + + if (hotplug_memory_notifier(lru_gen_online_mem, 0)) + pr_err("lru_gen: failed to subscribe hotplug notifications\n"); + + if (sysfs_create_group(mm_kobj, &lru_gen_attr_group)) + pr_err("lru_gen: failed to create sysfs group\n"); + + debugfs_create_file("lru_gen", 0644, NULL, NULL, &lru_gen_rw_fops); + debugfs_create_file("lru_gen_full", 0444, NULL, NULL, &lru_gen_ro_fops); + + return 0; +}; +/* + * We want to run as early as possible because debug code may call mm_alloc() + * and mmput(). Out only dependency mm_kobj is initialized one stage earlier. + */ +arch_initcall(init_lru_gen); + +#else /* CONFIG_LRU_GEN */ + +static void lru_gen_shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc) +{ +} + +static void lru_gen_age_node(struct pglist_data *pgdat, struct scan_control *sc) +{ +} + +static void lru_gen_start_kswapd(int nid) +{ +} + +static void lru_gen_stop_kswapd(int nid) +{ +} + +#endif /* CONFIG_LRU_GEN */ + +static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc) +{ + unsigned long nr[NR_LRU_LISTS]; + unsigned long targets[NR_LRU_LISTS]; + unsigned long nr_to_scan; + enum lru_list lru; + unsigned long nr_reclaimed = 0; + unsigned long nr_to_reclaim = sc->nr_to_reclaim; + struct blk_plug plug; + bool scan_adjusted; + + if (lru_gen_enabled()) { + lru_gen_shrink_lruvec(lruvec, sc); + return; + } + + get_scan_count(lruvec, sc, nr); + + /* Record the original scan target for proportional adjustments later */ + memcpy(targets, nr, sizeof(nr)); + + /* + * Global reclaiming within direct reclaim at DEF_PRIORITY is a normal + * event that can occur when there is little memory pressure e.g. + * multiple streaming readers/writers. Hence, we do not abort scanning + * when the requested number of pages are reclaimed when scanning at + * DEF_PRIORITY on the assumption that the fact we are direct + * reclaiming implies that kswapd is not keeping up and it is best to + * do a batch of work at once. For memcg reclaim one check is made to + * abort proportional reclaim if either the file or anon lru has already + * dropped to zero at the first pass. + */ + scan_adjusted = (!cgroup_reclaim(sc) && !current_is_kswapd() && + sc->priority == DEF_PRIORITY); + + blk_start_plug(&plug); + while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || + nr[LRU_INACTIVE_FILE]) { + unsigned long nr_anon, nr_file, percentage; + unsigned long nr_scanned; + + for_each_evictable_lru(lru) { + if (nr[lru]) { + nr_to_scan = min(nr[lru], SWAP_CLUSTER_MAX); + nr[lru] -= nr_to_scan; + + nr_reclaimed += shrink_list(lru, nr_to_scan, + lruvec, sc); + } + } + + cond_resched(); + + if (nr_reclaimed < nr_to_reclaim || scan_adjusted) + continue; + + /* + * For kswapd and memcg, reclaim at least the number of pages + * requested. Ensure that the anon and file LRUs are scanned + * proportionally what was requested by get_scan_count(). We + * stop reclaiming one LRU and reduce the amount scanning + * proportional to the original scan target. + */ + nr_file = nr[LRU_INACTIVE_FILE] + nr[LRU_ACTIVE_FILE]; + nr_anon = nr[LRU_INACTIVE_ANON] + nr[LRU_ACTIVE_ANON]; + + /* + * It's just vindictive to attack the larger once the smaller + * has gone to zero. And given the way we stop scanning the + * smaller below, this makes sure that we only make one nudge + * towards proportionality once we've got nr_to_reclaim. + */ + if (!nr_file || !nr_anon) + break; + + if (nr_file > nr_anon) { + unsigned long scan_target = targets[LRU_INACTIVE_ANON] + + targets[LRU_ACTIVE_ANON] + 1; + lru = LRU_BASE; + percentage = nr_anon * 100 / scan_target; + } else { + unsigned long scan_target = targets[LRU_INACTIVE_FILE] + + targets[LRU_ACTIVE_FILE] + 1; + lru = LRU_FILE; + percentage = nr_file * 100 / scan_target; + } + + /* Stop scanning the smaller of the LRU */ + nr[lru] = 0; + nr[lru + LRU_ACTIVE] = 0; + + /* + * Recalculate the other LRU scan count based on its original + * scan target and the percentage scanning already complete + */ + lru = (lru == LRU_FILE) ? LRU_BASE : LRU_FILE; + nr_scanned = targets[lru] - nr[lru]; + nr[lru] = targets[lru] * (100 - percentage) / 100; + nr[lru] -= min(nr[lru], nr_scanned); + + lru += LRU_ACTIVE; + nr_scanned = targets[lru] - nr[lru]; + nr[lru] = targets[lru] * (100 - percentage) / 100; + nr[lru] -= min(nr[lru], nr_scanned); + + scan_adjusted = true; + } + blk_finish_plug(&plug); + sc->nr_reclaimed += nr_reclaimed; + + /* + * Even if we did not try to evict anon pages at all, we want to + * rebalance the anon lru active/inactive ratio. + */ + if (total_swap_pages && inactive_is_low(lruvec, LRU_INACTIVE_ANON)) + shrink_active_list(SWAP_CLUSTER_MAX, lruvec, + sc, LRU_ACTIVE_ANON); +} + +/* Use reclaim/compaction for costly allocs or under memory pressure */ +static bool in_reclaim_compaction(struct scan_control *sc) +{ + if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && + (sc->order > PAGE_ALLOC_COSTLY_ORDER || + sc->priority < DEF_PRIORITY - 2)) + return true; return false; } @@ -2866,7 +5512,6 @@ static void shrink_node(pg_data_t *pgdat, struct scan_control *sc) unsigned long nr_reclaimed, nr_scanned; struct lruvec *target_lruvec; bool reclaimable = false; - unsigned long file; target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat); @@ -2876,92 +5521,34 @@ static void shrink_node(pg_data_t *pgdat, struct scan_control *sc) nr_reclaimed = sc->nr_reclaimed; nr_scanned = sc->nr_scanned; - /* - * Determine the scan balance between anon and file LRUs. - */ - spin_lock_irq(&target_lruvec->lru_lock); - sc->anon_cost = target_lruvec->anon_cost; - sc->file_cost = target_lruvec->file_cost; - spin_unlock_irq(&target_lruvec->lru_lock); + prepare_scan_count(pgdat, sc); /* - * Target desirable inactive:active list ratios for the anon - * and file LRU lists. + * Check the number of clean file pages to protect them from + * reclaiming if their amount is below the specified. */ - if (!sc->force_deactivate) { - unsigned long refaults; - - refaults = lruvec_page_state(target_lruvec, - WORKINGSET_ACTIVATE_ANON); - if (refaults != target_lruvec->refaults[0] || - inactive_is_low(target_lruvec, LRU_INACTIVE_ANON)) - sc->may_deactivate |= DEACTIVATE_ANON; - else - sc->may_deactivate &= ~DEACTIVATE_ANON; - + if (sysctl_clean_low_kbytes || sysctl_clean_min_kbytes) { + unsigned long reclaimable_file, dirty, clean; + + reclaimable_file = + node_page_state(pgdat, NR_ACTIVE_FILE) + + node_page_state(pgdat, NR_INACTIVE_FILE) + + node_page_state(pgdat, NR_ISOLATED_FILE); + dirty = node_page_state(pgdat, NR_FILE_DIRTY); /* - * When refaults are being observed, it means a new - * workingset is being established. Deactivate to get - * rid of any stale active pages quickly. + * node_page_state() sum can go out of sync since + * all the values are not read at once. */ - refaults = lruvec_page_state(target_lruvec, - WORKINGSET_ACTIVATE_FILE); - if (refaults != target_lruvec->refaults[1] || - inactive_is_low(target_lruvec, LRU_INACTIVE_FILE)) - sc->may_deactivate |= DEACTIVATE_FILE; + if (likely(reclaimable_file > dirty)) + clean = (reclaimable_file - dirty) << (PAGE_SHIFT - 10); else - sc->may_deactivate &= ~DEACTIVATE_FILE; - } else - sc->may_deactivate = DEACTIVATE_ANON | DEACTIVATE_FILE; - - /* - * If we have plenty of inactive file pages that aren't - * thrashing, try to reclaim those first before touching - * anonymous pages. - */ - file = lruvec_page_state(target_lruvec, NR_INACTIVE_FILE); - if (file >> sc->priority && !(sc->may_deactivate & DEACTIVATE_FILE)) - sc->cache_trim_mode = 1; - else - sc->cache_trim_mode = 0; - - /* - * Prevent the reclaimer from falling into the cache trap: as - * cache pages start out inactive, every cache fault will tip - * the scan balance towards the file LRU. And as the file LRU - * shrinks, so does the window for rotation from references. - * This means we have a runaway feedback loop where a tiny - * thrashing file LRU becomes infinitely more attractive than - * anon pages. Try to detect this based on file LRU size. - */ - if (!cgroup_reclaim(sc)) { - unsigned long total_high_wmark = 0; - unsigned long free, anon; - int z; - - free = sum_zone_node_page_state(pgdat->node_id, NR_FREE_PAGES); - file = node_page_state(pgdat, NR_ACTIVE_FILE) + - node_page_state(pgdat, NR_INACTIVE_FILE); - - for (z = 0; z < MAX_NR_ZONES; z++) { - struct zone *zone = &pgdat->node_zones[z]; - if (!managed_zone(zone)) - continue; - - total_high_wmark += high_wmark_pages(zone); - } - - /* - * Consider anon: if that's low too, this isn't a - * runaway file reclaim problem, but rather just - * extreme pressure. Reclaim as per usual then. - */ - anon = node_page_state(pgdat, NR_INACTIVE_ANON); + clean = 0; - sc->file_is_tiny = - file + free <= total_high_wmark && - !(sc->may_deactivate & DEACTIVATE_ANON) && - anon >> sc->priority; + sc->clean_below_low = clean < sysctl_clean_low_kbytes; + sc->clean_below_min = clean < sysctl_clean_min_kbytes; + } else { + sc->clean_below_low = false; + sc->clean_below_min = false; } shrink_node_memcgs(pgdat, sc); @@ -3182,6 +5769,9 @@ static void snapshot_refaults(struct mem_cgroup *target_memcg, pg_data_t *pgdat) struct lruvec *target_lruvec; unsigned long refaults; + if (lru_gen_enabled()) + return; + target_lruvec = mem_cgroup_lruvec(target_memcg, pgdat); refaults = lruvec_page_state(target_lruvec, WORKINGSET_ACTIVATE_ANON); target_lruvec->refaults[0] = refaults; @@ -3556,6 +6146,11 @@ static void age_active_anon(struct pglist_data *pgdat, struct mem_cgroup *memcg; struct lruvec *lruvec; + if (lru_gen_enabled()) { + lru_gen_age_node(pgdat, sc); + return; + } + if (!total_swap_pages) return; @@ -3881,7 +6476,8 @@ static int balance_pgdat(pg_data_t *pgdat, int order, int highest_zoneidx) __fs_reclaim_release(); ret = try_to_freeze(); __fs_reclaim_acquire(); - if (ret || kthread_should_stop()) + if (ret || kthread_should_stop() || + !atomic_long_read(&kswapd_waiters)) break; /* @@ -4236,6 +6832,8 @@ int kswapd_run(int nid) if (pgdat->kswapd) return 0; + lru_gen_start_kswapd(nid); + pgdat->kswapd = kthread_run(kswapd, pgdat, "kswapd%d", nid); if (IS_ERR(pgdat->kswapd)) { /* failure at boot is fatal */ @@ -4258,6 +6856,7 @@ void kswapd_stop(int nid) if (kswapd) { kthread_stop(kswapd); NODE_DATA(nid)->kswapd = NULL; + lru_gen_stop_kswapd(nid); } } diff --git a/mm/vmstat.c b/mm/vmstat.c index cccee36b289..4c516987da8 100644 --- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -1231,6 +1231,9 @@ const char * const vmstat_text[] = { "nr_swapcached", #endif +#ifdef CONFIG_UKSM + "nr_uksm_zero_pages", +#endif /* enum writeback_stat_item counters */ "nr_dirty_threshold", "nr_dirty_background_threshold", diff --git a/mm/workingset.c b/mm/workingset.c index b7cdeca5a76..3f3f03d51ea 100644 --- a/mm/workingset.c +++ b/mm/workingset.c @@ -168,9 +168,9 @@ * refault distance will immediately activate the refaulting page. */ -#define EVICTION_SHIFT ((BITS_PER_LONG - BITS_PER_XA_VALUE) + \ - 1 + NODES_SHIFT + MEM_CGROUP_ID_SHIFT) -#define EVICTION_MASK (~0UL >> EVICTION_SHIFT) +#define EVICTION_SHIFT (BITS_PER_XA_VALUE - MEM_CGROUP_ID_SHIFT - NODES_SHIFT) +#define EVICTION_MASK (BIT(EVICTION_SHIFT) - 1) +#define WORKINGSET_WIDTH 1 /* * Eviction timestamps need to be able to cover the full range of @@ -182,38 +182,129 @@ */ static unsigned int bucket_order __read_mostly; -static void *pack_shadow(int memcgid, pg_data_t *pgdat, unsigned long eviction, - bool workingset) +static void *pack_shadow(int memcg_id, struct pglist_data *pgdat, unsigned long val) { - eviction >>= bucket_order; - eviction &= EVICTION_MASK; - eviction = (eviction << MEM_CGROUP_ID_SHIFT) | memcgid; - eviction = (eviction << NODES_SHIFT) | pgdat->node_id; - eviction = (eviction << 1) | workingset; + val = (val << MEM_CGROUP_ID_SHIFT) | memcg_id; + val = (val << NODES_SHIFT) | pgdat->node_id; - return xa_mk_value(eviction); + return xa_mk_value(val); } -static void unpack_shadow(void *shadow, int *memcgidp, pg_data_t **pgdat, - unsigned long *evictionp, bool *workingsetp) +static unsigned long unpack_shadow(void *shadow, int *memcg_id, struct pglist_data **pgdat) { - unsigned long entry = xa_to_value(shadow); - int memcgid, nid; - bool workingset; + unsigned long val = xa_to_value(shadow); + + *pgdat = NODE_DATA(val & (BIT(NODES_SHIFT) - 1)); + val >>= NODES_SHIFT; + *memcg_id = val & (BIT(MEM_CGROUP_ID_SHIFT) - 1); + + return val >> MEM_CGROUP_ID_SHIFT; +} + +#ifdef CONFIG_LRU_GEN + +#if LRU_GEN_SHIFT + LRU_USAGE_SHIFT >= EVICTION_SHIFT +#error "Please try smaller NODES_SHIFT, NR_LRU_GENS and TIERS_PER_GEN configurations" +#endif + +static void page_set_usage(struct page *page, int usage) +{ + unsigned long old_flags, new_flags; + + VM_BUG_ON(usage > BIT(LRU_USAGE_WIDTH)); + + if (!usage) + return; + + do { + old_flags = READ_ONCE(page->flags); + new_flags = (old_flags & ~LRU_USAGE_MASK) | LRU_TIER_FLAGS | + ((usage - 1UL) << LRU_USAGE_PGOFF); + } while (new_flags != old_flags && + cmpxchg(&page->flags, old_flags, new_flags) != old_flags); +} + +/* Return a token to be stored in the shadow entry of a page being evicted. */ +static void *lru_gen_eviction(struct page *page) +{ + int hist, tier; + unsigned long token; + unsigned long min_seq; + struct lruvec *lruvec; + struct lrugen *lrugen; + int type = page_is_file_lru(page); + int usage = page_tier_usage(page); + struct mem_cgroup *memcg = page_memcg(page); + struct pglist_data *pgdat = page_pgdat(page); + + lruvec = mem_cgroup_lruvec(memcg, pgdat); + lrugen = &lruvec->evictable; + min_seq = READ_ONCE(lrugen->min_seq[type]); + token = (min_seq << LRU_USAGE_SHIFT) | usage; + + hist = hist_from_seq_or_gen(min_seq); + tier = lru_tier_from_usage(usage); + atomic_long_add(thp_nr_pages(page), &lrugen->evicted[hist][type][tier]); + + return pack_shadow(mem_cgroup_id(memcg), pgdat, token); +} - workingset = entry & 1; - entry >>= 1; - nid = entry & ((1UL << NODES_SHIFT) - 1); - entry >>= NODES_SHIFT; - memcgid = entry & ((1UL << MEM_CGROUP_ID_SHIFT) - 1); - entry >>= MEM_CGROUP_ID_SHIFT; - - *memcgidp = memcgid; - *pgdat = NODE_DATA(nid); - *evictionp = entry << bucket_order; - *workingsetp = workingset; +/* Account a refaulted page based on the token stored in its shadow entry. */ +static void lru_gen_refault(struct page *page, void *shadow) +{ + int hist, tier, usage; + int memcg_id; + unsigned long token; + unsigned long min_seq; + struct lruvec *lruvec; + struct lrugen *lrugen; + struct pglist_data *pgdat; + struct mem_cgroup *memcg; + int type = page_is_file_lru(page); + + token = unpack_shadow(shadow, &memcg_id, &pgdat); + if (page_pgdat(page) != pgdat) + return; + + rcu_read_lock(); + memcg = page_memcg_rcu(page); + if (mem_cgroup_id(memcg) != memcg_id) + goto unlock; + + usage = token & (BIT(LRU_USAGE_SHIFT) - 1); + token >>= LRU_USAGE_SHIFT; + + lruvec = mem_cgroup_lruvec(memcg, pgdat); + lrugen = &lruvec->evictable; + min_seq = READ_ONCE(lrugen->min_seq[type]); + if (token != (min_seq & (EVICTION_MASK >> LRU_USAGE_SHIFT))) + goto unlock; + + page_set_usage(page, usage); + + hist = hist_from_seq_or_gen(min_seq); + tier = lru_tier_from_usage(usage); + atomic_long_add(thp_nr_pages(page), &lrugen->refaulted[hist][type][tier]); + inc_lruvec_state(lruvec, WORKINGSET_REFAULT_BASE + type); + if (tier) + inc_lruvec_state(lruvec, WORKINGSET_RESTORE_BASE + type); +unlock: + rcu_read_unlock(); } +#else /* CONFIG_LRU_GEN */ + +static void *lru_gen_eviction(struct page *page) +{ + return NULL; +} + +static void lru_gen_refault(struct page *page, void *shadow) +{ +} + +#endif /* CONFIG_LRU_GEN */ + /** * workingset_age_nonresident - age non-resident entries as LRU ages * @lruvec: the lruvec that was aged @@ -262,12 +353,17 @@ void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg) VM_BUG_ON_PAGE(page_count(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page); + if (lru_gen_enabled()) + return lru_gen_eviction(page); + lruvec = mem_cgroup_lruvec(target_memcg, pgdat); /* XXX: target_memcg can be NULL, go through lruvec */ memcgid = mem_cgroup_id(lruvec_memcg(lruvec)); eviction = atomic_long_read(&lruvec->nonresident_age); + eviction >>= bucket_order; + eviction = (eviction << WORKINGSET_WIDTH) | PageWorkingset(page); workingset_age_nonresident(lruvec, thp_nr_pages(page)); - return pack_shadow(memcgid, pgdat, eviction, PageWorkingset(page)); + return pack_shadow(memcgid, pgdat, eviction); } /** @@ -294,7 +390,12 @@ void workingset_refault(struct page *page, void *shadow) bool workingset; int memcgid; - unpack_shadow(shadow, &memcgid, &pgdat, &eviction, &workingset); + if (lru_gen_enabled()) { + lru_gen_refault(page, shadow); + return; + } + + eviction = unpack_shadow(shadow, &memcgid, &pgdat); rcu_read_lock(); /* @@ -318,6 +419,8 @@ void workingset_refault(struct page *page, void *shadow) goto out; eviction_lruvec = mem_cgroup_lruvec(eviction_memcg, pgdat); refault = atomic_long_read(&eviction_lruvec->nonresident_age); + workingset = eviction & (BIT(WORKINGSET_WIDTH) - 1); + eviction = (eviction >> WORKINGSET_WIDTH) << bucket_order; /* * Calculate the refault distance @@ -335,7 +438,7 @@ void workingset_refault(struct page *page, void *shadow) * longest time, so the occasional inappropriate activation * leading to pressure on the active list is not a problem. */ - refault_distance = (refault - eviction) & EVICTION_MASK; + refault_distance = (refault - eviction) & (EVICTION_MASK >> WORKINGSET_WIDTH); /* * The activation decision for this page is made at the level @@ -593,7 +696,7 @@ static int __init workingset_init(void) unsigned int max_order; int ret; - BUILD_BUG_ON(BITS_PER_LONG < EVICTION_SHIFT); + BUILD_BUG_ON(EVICTION_SHIFT < WORKINGSET_WIDTH); /* * Calculate the eviction bucket size to cover the longest * actionable refault distance, which is currently half of @@ -601,7 +704,7 @@ static int __init workingset_init(void) * some more pages at runtime, so keep working with up to * double the initial memory by using totalram_pages as-is. */ - timestamp_bits = BITS_PER_LONG - EVICTION_SHIFT; + timestamp_bits = EVICTION_SHIFT - WORKINGSET_WIDTH; max_order = fls_long(totalram_pages() - 1); if (max_order > timestamp_bits) bucket_order = max_order - timestamp_bits; diff --git a/net/core/pktgen.c b/net/core/pktgen.c index 3fba429f1f5..9a3a9a6eb83 100644 --- a/net/core/pktgen.c +++ b/net/core/pktgen.c @@ -1894,7 +1894,7 @@ static void pktgen_mark_device(const struct pktgen_net *pn, const char *ifname) mutex_unlock(&pktgen_thread_lock); pr_debug("%s: waiting for %s to disappear....\n", __func__, ifname); - schedule_timeout_interruptible(msecs_to_jiffies(msec_per_try)); + schedule_msec_hrtimeout_interruptible((msec_per_try)); mutex_lock(&pktgen_thread_lock); if (++i >= max_tries) { diff --git a/net/ipv4/Kconfig b/net/ipv4/Kconfig index 87983e70f03..a833a7a67ce 100644 --- a/net/ipv4/Kconfig +++ b/net/ipv4/Kconfig @@ -669,6 +669,24 @@ config TCP_CONG_BBR AQM schemes that do not provide a delay signal. It requires the fq ("Fair Queue") pacing packet scheduler. +config TCP_CONG_BBR2 + tristate "BBR2 TCP" + default n + help + + BBR2 TCP congestion control is a model-based congestion control + algorithm that aims to maximize network utilization, keep queues and + retransmit rates low, and to be able to coexist with Reno/CUBIC in + common scenarios. It builds an explicit model of the network path. It + tolerates a targeted degree of random packet loss and delay that are + unrelated to congestion. It can operate over LAN, WAN, cellular, wifi, + or cable modem links, and can use DCTCP-L4S-style ECN signals. It can + coexist with flows that use loss-based congestion control, and can + operate with shallow buffers, deep buffers, bufferbloat, policers, or + AQM schemes that do not provide a delay signal. It requires pacing, + using either TCP internal pacing or the fq ("Fair Queue") pacing packet + scheduler. + choice prompt "Default TCP congestion control" default DEFAULT_CUBIC @@ -706,6 +724,9 @@ choice config DEFAULT_BBR bool "BBR" if TCP_CONG_BBR=y + config DEFAULT_BBR2 + bool "BBR2" if TCP_CONG_BBR2=y + config DEFAULT_RENO bool "Reno" endchoice @@ -730,6 +751,7 @@ config DEFAULT_TCP_CONG default "dctcp" if DEFAULT_DCTCP default "cdg" if DEFAULT_CDG default "bbr" if DEFAULT_BBR + default "bbr2" if DEFAULT_BBR2 default "cubic" config TCP_MD5SIG diff --git a/net/ipv4/Makefile b/net/ipv4/Makefile index bbdd9c44f14..8dee1547d82 100644 --- a/net/ipv4/Makefile +++ b/net/ipv4/Makefile @@ -46,6 +46,7 @@ obj-$(CONFIG_INET_TCP_DIAG) += tcp_diag.o obj-$(CONFIG_INET_UDP_DIAG) += udp_diag.o obj-$(CONFIG_INET_RAW_DIAG) += raw_diag.o obj-$(CONFIG_TCP_CONG_BBR) += tcp_bbr.o +obj-$(CONFIG_TCP_CONG_BBR2) += tcp_bbr2.o obj-$(CONFIG_TCP_CONG_BIC) += tcp_bic.o obj-$(CONFIG_TCP_CONG_CDG) += tcp_cdg.o obj-$(CONFIG_TCP_CONG_CUBIC) += tcp_cubic.o diff --git a/net/ipv4/bpf_tcp_ca.c b/net/ipv4/bpf_tcp_ca.c index 9e41eff4a68..f26b0fcbb9f 100644 --- a/net/ipv4/bpf_tcp_ca.c +++ b/net/ipv4/bpf_tcp_ca.c @@ -17,7 +17,7 @@ static u32 optional_ops[] = { offsetof(struct tcp_congestion_ops, cwnd_event), offsetof(struct tcp_congestion_ops, in_ack_event), offsetof(struct tcp_congestion_ops, pkts_acked), - offsetof(struct tcp_congestion_ops, min_tso_segs), + offsetof(struct tcp_congestion_ops, tso_segs), offsetof(struct tcp_congestion_ops, sndbuf_expand), offsetof(struct tcp_congestion_ops, cong_control), }; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 4c9a5b99bf7..ef6f09bba33 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -3049,6 +3049,7 @@ int tcp_disconnect(struct sock *sk, int flags) tp->rx_opt.dsack = 0; tp->rx_opt.num_sacks = 0; tp->rcv_ooopack = 0; + tp->fast_ack_mode = 0; /* Clean up fastopen related fields */ diff --git a/net/ipv4/tcp_bbr.c b/net/ipv4/tcp_bbr.c index 6ea3dc2e421..8ef512fefe2 100644 --- a/net/ipv4/tcp_bbr.c +++ b/net/ipv4/tcp_bbr.c @@ -292,26 +292,40 @@ static void bbr_set_pacing_rate(struct sock *sk, u32 bw, int gain) sk->sk_pacing_rate = rate; } -/* override sysctl_tcp_min_tso_segs */ static u32 bbr_min_tso_segs(struct sock *sk) { return sk->sk_pacing_rate < (bbr_min_tso_rate >> 3) ? 1 : 2; } +/* Return the number of segments BBR would like in a TSO/GSO skb, given + * a particular max gso size as a constraint. + */ +static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now, + u32 gso_max_size) +{ + u32 segs; + u64 bytes; + + /* Budget a TSO/GSO burst size allowance based on bw (pacing_rate). */ + bytes = sk->sk_pacing_rate >> sk->sk_pacing_shift; + + bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER); + segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk)); + return segs; +} + +/* Custom tcp_tso_autosize() for BBR, used at transmit time to cap skb size. */ +static u32 bbr_tso_segs(struct sock *sk, unsigned int mss_now) +{ + return bbr_tso_segs_generic(sk, mss_now, sk->sk_gso_max_size); +} + +/* Like bbr_tso_segs(), using mss_cache, ignoring driver's sk_gso_max_size. */ static u32 bbr_tso_segs_goal(struct sock *sk) { struct tcp_sock *tp = tcp_sk(sk); - u32 segs, bytes; - - /* Sort of tcp_tso_autosize() but ignoring - * driver provided sk_gso_max_size. - */ - bytes = min_t(unsigned long, - sk->sk_pacing_rate >> READ_ONCE(sk->sk_pacing_shift), - GSO_MAX_SIZE - 1 - MAX_TCP_HEADER); - segs = max_t(u32, bytes / tp->mss_cache, bbr_min_tso_segs(sk)); - return min(segs, 0x7FU); + return bbr_tso_segs_generic(sk, tp->mss_cache, GSO_MAX_SIZE); } /* Save "last known good" cwnd so we can restore it after losses or PROBE_RTT */ @@ -1147,7 +1161,7 @@ static struct tcp_congestion_ops tcp_bbr_cong_ops __read_mostly = { .undo_cwnd = bbr_undo_cwnd, .cwnd_event = bbr_cwnd_event, .ssthresh = bbr_ssthresh, - .min_tso_segs = bbr_min_tso_segs, + .tso_segs = bbr_tso_segs, .get_info = bbr_get_info, .set_state = bbr_set_state, }; diff --git a/net/ipv4/tcp_bbr2.c b/net/ipv4/tcp_bbr2.c new file mode 100644 index 00000000000..5510adc92bb --- /dev/null +++ b/net/ipv4/tcp_bbr2.c @@ -0,0 +1,2671 @@ +/* BBR (Bottleneck Bandwidth and RTT) congestion control, v2 + * + * BBRv2 is a model-based congestion control algorithm that aims for low + * queues, low loss, and (bounded) Reno/CUBIC coexistence. To maintain a model + * of the network path, it uses measurements of bandwidth and RTT, as well as + * (if they occur) packet loss and/or DCTCP/L4S-style ECN signals. Note that + * although it can use ECN or loss signals explicitly, it does not require + * either; it can bound its in-flight data based on its estimate of the BDP. + * + * The model has both higher and lower bounds for the operating range: + * lo: bw_lo, inflight_lo: conservative short-term lower bound + * hi: bw_hi, inflight_hi: robust long-term upper bound + * The bandwidth-probing time scale is (a) extended dynamically based on + * estimated BDP to improve coexistence with Reno/CUBIC; (b) bounded by + * an interactive wall-clock time-scale to be more scalable and responsive + * than Reno and CUBIC. + * + * Here is a state transition diagram for BBR: + * + * | + * V + * +---> STARTUP ----+ + * | | | + * | V | + * | DRAIN ----+ + * | | | + * | V | + * +---> PROBE_BW ----+ + * | ^ | | + * | | | | + * | +----+ | + * | | + * +---- PROBE_RTT <--+ + * + * A BBR flow starts in STARTUP, and ramps up its sending rate quickly. + * When it estimates the pipe is full, it enters DRAIN to drain the queue. + * In steady state a BBR flow only uses PROBE_BW and PROBE_RTT. + * A long-lived BBR flow spends the vast majority of its time remaining + * (repeatedly) in PROBE_BW, fully probing and utilizing the pipe's bandwidth + * in a fair manner, with a small, bounded queue. *If* a flow has been + * continuously sending for the entire min_rtt window, and hasn't seen an RTT + * sample that matches or decreases its min_rtt estimate for 10 seconds, then + * it briefly enters PROBE_RTT to cut inflight to a minimum value to re-probe + * the path's two-way propagation delay (min_rtt). When exiting PROBE_RTT, if + * we estimated that we reached the full bw of the pipe then we enter PROBE_BW; + * otherwise we enter STARTUP to try to fill the pipe. + * + * BBR is described in detail in: + * "BBR: Congestion-Based Congestion Control", + * Neal Cardwell, Yuchung Cheng, C. Stephen Gunn, Soheil Hassas Yeganeh, + * Van Jacobson. ACM Queue, Vol. 14 No. 5, September-October 2016. + * + * There is a public e-mail list for discussing BBR development and testing: + * https://groups.google.com/forum/#!forum/bbr-dev + * + * NOTE: BBR might be used with the fq qdisc ("man tc-fq") with pacing enabled, + * otherwise TCP stack falls back to an internal pacing using one high + * resolution timer per TCP socket and may use more resources. + */ +#include +#include +#include +#include +#include + +#include "tcp_dctcp.h" + +/* Scale factor for rate in pkt/uSec unit to avoid truncation in bandwidth + * estimation. The rate unit ~= (1500 bytes / 1 usec / 2^24) ~= 715 bps. + * This handles bandwidths from 0.06pps (715bps) to 256Mpps (3Tbps) in a u32. + * Since the minimum window is >=4 packets, the lower bound isn't + * an issue. The upper bound isn't an issue with existing technologies. + */ +#define BW_SCALE 24 +#define BW_UNIT (1 << BW_SCALE) + +#define BBR_SCALE 8 /* scaling factor for fractions in BBR (e.g. gains) */ +#define BBR_UNIT (1 << BBR_SCALE) + +#define FLAG_DEBUG_VERBOSE 0x1 /* Verbose debugging messages */ +#define FLAG_DEBUG_LOOPBACK 0x2 /* Do NOT skip loopback addr */ + +#define CYCLE_LEN 8 /* number of phases in a pacing gain cycle */ + +/* BBR has the following modes for deciding how fast to send: */ +enum bbr_mode { + BBR_STARTUP, /* ramp up sending rate rapidly to fill pipe */ + BBR_DRAIN, /* drain any queue created during startup */ + BBR_PROBE_BW, /* discover, share bw: pace around estimated bw */ + BBR_PROBE_RTT, /* cut inflight to min to probe min_rtt */ +}; + +/* How does the incoming ACK stream relate to our bandwidth probing? */ +enum bbr_ack_phase { + BBR_ACKS_INIT, /* not probing; not getting probe feedback */ + BBR_ACKS_REFILLING, /* sending at est. bw to fill pipe */ + BBR_ACKS_PROBE_STARTING, /* inflight rising to probe bw */ + BBR_ACKS_PROBE_FEEDBACK, /* getting feedback from bw probing */ + BBR_ACKS_PROBE_STOPPING, /* stopped probing; still getting feedback */ +}; + +/* BBR congestion control block */ +struct bbr { + u32 min_rtt_us; /* min RTT in min_rtt_win_sec window */ + u32 min_rtt_stamp; /* timestamp of min_rtt_us */ + u32 probe_rtt_done_stamp; /* end time for BBR_PROBE_RTT mode */ + u32 probe_rtt_min_us; /* min RTT in bbr_probe_rtt_win_ms window */ + u32 probe_rtt_min_stamp; /* timestamp of probe_rtt_min_us*/ + u32 next_rtt_delivered; /* scb->tx.delivered at end of round */ + u32 prior_rcv_nxt; /* tp->rcv_nxt when CE state last changed */ + u64 cycle_mstamp; /* time of this cycle phase start */ + u32 mode:3, /* current bbr_mode in state machine */ + prev_ca_state:3, /* CA state on previous ACK */ + packet_conservation:1, /* use packet conservation? */ + round_start:1, /* start of packet-timed tx->ack round? */ + ce_state:1, /* If most recent data has CE bit set */ + bw_probe_up_rounds:5, /* cwnd-limited rounds in PROBE_UP */ + try_fast_path:1, /* can we take fast path? */ + unused2:11, + idle_restart:1, /* restarting after idle? */ + probe_rtt_round_done:1, /* a BBR_PROBE_RTT round at 4 pkts? */ + cycle_idx:3, /* current index in pacing_gain cycle array */ + has_seen_rtt:1; /* have we seen an RTT sample yet? */ + u32 pacing_gain:11, /* current gain for setting pacing rate */ + cwnd_gain:11, /* current gain for setting cwnd */ + full_bw_reached:1, /* reached full bw in Startup? */ + full_bw_cnt:2, /* number of rounds without large bw gains */ + init_cwnd:7; /* initial cwnd */ + u32 prior_cwnd; /* prior cwnd upon entering loss recovery */ + u32 full_bw; /* recent bw, to estimate if pipe is full */ + + /* For tracking ACK aggregation: */ + u64 ack_epoch_mstamp; /* start of ACK sampling epoch */ + u16 extra_acked[2]; /* max excess data ACKed in epoch */ + u32 ack_epoch_acked:20, /* packets (S)ACKed in sampling epoch */ + extra_acked_win_rtts:5, /* age of extra_acked, in round trips */ + extra_acked_win_idx:1, /* current index in extra_acked array */ + /* BBR v2 state: */ + unused1:2, + startup_ecn_rounds:2, /* consecutive hi ECN STARTUP rounds */ + loss_in_cycle:1, /* packet loss in this cycle? */ + ecn_in_cycle:1; /* ECN in this cycle? */ + u32 loss_round_delivered; /* scb->tx.delivered ending loss round */ + u32 undo_bw_lo; /* bw_lo before latest losses */ + u32 undo_inflight_lo; /* inflight_lo before latest losses */ + u32 undo_inflight_hi; /* inflight_hi before latest losses */ + u32 bw_latest; /* max delivered bw in last round trip */ + u32 bw_lo; /* lower bound on sending bandwidth */ + u32 bw_hi[2]; /* upper bound of sending bandwidth range*/ + u32 inflight_latest; /* max delivered data in last round trip */ + u32 inflight_lo; /* lower bound of inflight data range */ + u32 inflight_hi; /* upper bound of inflight data range */ + u32 bw_probe_up_cnt; /* packets delivered per inflight_hi incr */ + u32 bw_probe_up_acks; /* packets (S)ACKed since inflight_hi incr */ + u32 probe_wait_us; /* PROBE_DOWN until next clock-driven probe */ + u32 ecn_eligible:1, /* sender can use ECN (RTT, handshake)? */ + ecn_alpha:9, /* EWMA delivered_ce/delivered; 0..256 */ + bw_probe_samples:1, /* rate samples reflect bw probing? */ + prev_probe_too_high:1, /* did last PROBE_UP go too high? */ + stopped_risky_probe:1, /* last PROBE_UP stopped due to risk? */ + rounds_since_probe:8, /* packet-timed rounds since probed bw */ + loss_round_start:1, /* loss_round_delivered round trip? */ + loss_in_round:1, /* loss marked in this round trip? */ + ecn_in_round:1, /* ECN marked in this round trip? */ + ack_phase:3, /* bbr_ack_phase: meaning of ACKs */ + loss_events_in_round:4,/* losses in STARTUP round */ + initialized:1; /* has bbr_init() been called? */ + u32 alpha_last_delivered; /* tp->delivered at alpha update */ + u32 alpha_last_delivered_ce; /* tp->delivered_ce at alpha update */ + + /* Params configurable using setsockopt. Refer to correspoding + * module param for detailed description of params. + */ + struct bbr_params { + u32 high_gain:11, /* max allowed value: 2047 */ + drain_gain:10, /* max allowed value: 1023 */ + cwnd_gain:11; /* max allowed value: 2047 */ + u32 cwnd_min_target:4, /* max allowed value: 15 */ + min_rtt_win_sec:5, /* max allowed value: 31 */ + probe_rtt_mode_ms:9, /* max allowed value: 511 */ + full_bw_cnt:3, /* max allowed value: 7 */ + cwnd_tso_budget:1, /* allowed values: {0, 1} */ + unused3:6, + drain_to_target:1, /* boolean */ + precise_ece_ack:1, /* boolean */ + extra_acked_in_startup:1, /* allowed values: {0, 1} */ + fast_path:1; /* boolean */ + u32 full_bw_thresh:10, /* max allowed value: 1023 */ + startup_cwnd_gain:11, /* max allowed value: 2047 */ + bw_probe_pif_gain:9, /* max allowed value: 511 */ + usage_based_cwnd:1, /* boolean */ + unused2:1; + u16 probe_rtt_win_ms:14, /* max allowed value: 16383 */ + refill_add_inc:2; /* max allowed value: 3 */ + u16 extra_acked_gain:11, /* max allowed value: 2047 */ + extra_acked_win_rtts:5; /* max allowed value: 31*/ + u16 pacing_gain[CYCLE_LEN]; /* max allowed value: 1023 */ + /* Mostly BBR v2 parameters below here: */ + u32 ecn_alpha_gain:8, /* max allowed value: 255 */ + ecn_factor:8, /* max allowed value: 255 */ + ecn_thresh:8, /* max allowed value: 255 */ + beta:8; /* max allowed value: 255 */ + u32 ecn_max_rtt_us:19, /* max allowed value: 524287 */ + bw_probe_reno_gain:9, /* max allowed value: 511 */ + full_loss_cnt:4; /* max allowed value: 15 */ + u32 probe_rtt_cwnd_gain:8, /* max allowed value: 255 */ + inflight_headroom:8, /* max allowed value: 255 */ + loss_thresh:8, /* max allowed value: 255 */ + bw_probe_max_rounds:8; /* max allowed value: 255 */ + u32 bw_probe_rand_rounds:4, /* max allowed value: 15 */ + bw_probe_base_us:26, /* usecs: 0..2^26-1 (67 secs) */ + full_ecn_cnt:2; /* max allowed value: 3 */ + u32 bw_probe_rand_us:26, /* usecs: 0..2^26-1 (67 secs) */ + undo:1, /* boolean */ + tso_rtt_shift:4, /* max allowed value: 15 */ + unused5:1; + u32 ecn_reprobe_gain:9, /* max allowed value: 511 */ + unused1:14, + ecn_alpha_init:9; /* max allowed value: 256 */ + } params; + + struct { + u32 snd_isn; /* Initial sequence number */ + u32 rs_bw; /* last valid rate sample bw */ + u32 target_cwnd; /* target cwnd, based on BDP */ + u8 undo:1, /* Undo even happened but not yet logged */ + unused:7; + char event; /* single-letter event debug codes */ + u16 unused2; + } debug; +}; + +struct bbr_context { + u32 sample_bw; + u32 target_cwnd; + u32 log:1; +}; + +/* Window length of min_rtt filter (in sec). Max allowed value is 31 (0x1F) */ +static u32 bbr_min_rtt_win_sec = 10; +/* Minimum time (in ms) spent at bbr_cwnd_min_target in BBR_PROBE_RTT mode. + * Max allowed value is 511 (0x1FF). + */ +static u32 bbr_probe_rtt_mode_ms = 200; +/* Window length of probe_rtt_min_us filter (in ms), and consequently the + * typical interval between PROBE_RTT mode entries. + * Note that bbr_probe_rtt_win_ms must be <= bbr_min_rtt_win_sec * MSEC_PER_SEC + */ +static u32 bbr_probe_rtt_win_ms = 5000; +/* Skip TSO below the following bandwidth (bits/sec): */ +static int bbr_min_tso_rate = 1200000; + +/* Use min_rtt to help adapt TSO burst size, with smaller min_rtt resulting + * in bigger TSO bursts. By default we cut the RTT-based allowance in half + * for every 2^9 usec (aka 512 us) of RTT, so that the RTT-based allowance + * is below 1500 bytes after 6 * ~500 usec = 3ms. + */ +static u32 bbr_tso_rtt_shift = 9; /* halve allowance per 2^9 usecs, 512us */ + +/* Select cwnd TSO budget approach: + * 0: padding + * 1: flooring + */ +static uint bbr_cwnd_tso_budget = 1; + +/* Pace at ~1% below estimated bw, on average, to reduce queue at bottleneck. + * In order to help drive the network toward lower queues and low latency while + * maintaining high utilization, the average pacing rate aims to be slightly + * lower than the estimated bandwidth. This is an important aspect of the + * design. + */ +static const int bbr_pacing_margin_percent = 1; + +/* We use a high_gain value of 2/ln(2) because it's the smallest pacing gain + * that will allow a smoothly increasing pacing rate that will double each RTT + * and send the same number of packets per RTT that an un-paced, slow-starting + * Reno or CUBIC flow would. Max allowed value is 2047 (0x7FF). + */ +static int bbr_high_gain = BBR_UNIT * 2885 / 1000 + 1; +/* The gain for deriving startup cwnd. Max allowed value is 2047 (0x7FF). */ +static int bbr_startup_cwnd_gain = BBR_UNIT * 2885 / 1000 + 1; +/* The pacing gain of 1/high_gain in BBR_DRAIN is calculated to typically drain + * the queue created in BBR_STARTUP in a single round. Max allowed value + * is 1023 (0x3FF). + */ +static int bbr_drain_gain = BBR_UNIT * 1000 / 2885; +/* The gain for deriving steady-state cwnd tolerates delayed/stretched ACKs. + * Max allowed value is 2047 (0x7FF). + */ +static int bbr_cwnd_gain = BBR_UNIT * 2; +/* The pacing_gain values for the PROBE_BW gain cycle, to discover/share bw. + * Max allowed value for each element is 1023 (0x3FF). + */ +enum bbr_pacing_gain_phase { + BBR_BW_PROBE_UP = 0, /* push up inflight to probe for bw/vol */ + BBR_BW_PROBE_DOWN = 1, /* drain excess inflight from the queue */ + BBR_BW_PROBE_CRUISE = 2, /* use pipe, w/ headroom in queue/pipe */ + BBR_BW_PROBE_REFILL = 3, /* v2: refill the pipe again to 100% */ +}; +static int bbr_pacing_gain[] = { + BBR_UNIT * 5 / 4, /* probe for more available bw */ + BBR_UNIT * 3 / 4, /* drain queue and/or yield bw to other flows */ + BBR_UNIT, BBR_UNIT, BBR_UNIT, /* cruise at 1.0*bw to utilize pipe, */ + BBR_UNIT, BBR_UNIT, BBR_UNIT /* without creating excess queue... */ +}; + +/* Try to keep at least this many packets in flight, if things go smoothly. For + * smooth functioning, a sliding window protocol ACKing every other packet + * needs at least 4 packets in flight. Max allowed value is 15 (0xF). + */ +static u32 bbr_cwnd_min_target = 4; + +/* Cwnd to BDP proportion in PROBE_RTT mode scaled by BBR_UNIT. Default: 50%. + * Use 0 to disable. Max allowed value is 255. + */ +static u32 bbr_probe_rtt_cwnd_gain = BBR_UNIT * 1 / 2; + +/* To estimate if BBR_STARTUP mode (i.e. high_gain) has filled pipe... */ +/* If bw has increased significantly (1.25x), there may be more bw available. + * Max allowed value is 1023 (0x3FF). + */ +static u32 bbr_full_bw_thresh = BBR_UNIT * 5 / 4; +/* But after 3 rounds w/o significant bw growth, estimate pipe is full. + * Max allowed value is 7 (0x7). + */ +static u32 bbr_full_bw_cnt = 3; + +static u32 bbr_flags; /* Debugging related stuff */ + +/* Whether to debug using printk. + */ +static bool bbr_debug_with_printk; + +/* Whether to debug using ftrace event tcp:tcp_bbr_event. + * Ignored when bbr_debug_with_printk is set. + */ +static bool bbr_debug_ftrace; + +/* Experiment: each cycle, try to hold sub-unity gain until inflight <= BDP. */ +static bool bbr_drain_to_target = true; /* default: enabled */ + +/* Experiment: Flags to control BBR with ECN behavior. + */ +static bool bbr_precise_ece_ack = true; /* default: enabled */ + +/* The max rwin scaling shift factor is 14 (RFC 1323), so the max sane rwin is + * (2^(16+14) B)/(1024 B/packet) = 1M packets. + */ +static u32 bbr_cwnd_warn_val = 1U << 20; + +static u16 bbr_debug_port_mask; + +/* BBR module parameters. These are module parameters only in Google prod. + * Upstream these are intentionally not module parameters. + */ +static int bbr_pacing_gain_size = CYCLE_LEN; + +/* Gain factor for adding extra_acked to target cwnd: */ +static int bbr_extra_acked_gain = 256; + +/* Window length of extra_acked window. Max allowed val is 31. */ +static u32 bbr_extra_acked_win_rtts = 5; + +/* Max allowed val for ack_epoch_acked, after which sampling epoch is reset */ +static u32 bbr_ack_epoch_acked_reset_thresh = 1U << 20; + +/* Time period for clamping cwnd increment due to ack aggregation */ +static u32 bbr_extra_acked_max_us = 100 * 1000; + +/* Use extra acked in startup ? + * 0: disabled + * 1: use latest extra_acked value from 1-2 rtt in startup + */ +static int bbr_extra_acked_in_startup = 1; /* default: enabled */ + +/* Experiment: don't grow cwnd beyond twice of what we just probed. */ +static bool bbr_usage_based_cwnd; /* default: disabled */ + +/* For lab testing, researchers can enable BBRv2 ECN support with this flag, + * when they know that any ECN marks that the connections experience will be + * DCTCP/L4S-style ECN marks, rather than RFC3168 ECN marks. + * TODO(ncardwell): Production use of the BBRv2 ECN functionality depends on + * negotiation or configuration that is outside the scope of the BBRv2 + * alpha release. + */ +static bool bbr_ecn_enable = false; + +module_param_named(min_tso_rate, bbr_min_tso_rate, int, 0644); +module_param_named(tso_rtt_shift, bbr_tso_rtt_shift, int, 0644); +module_param_named(high_gain, bbr_high_gain, int, 0644); +module_param_named(drain_gain, bbr_drain_gain, int, 0644); +module_param_named(startup_cwnd_gain, bbr_startup_cwnd_gain, int, 0644); +module_param_named(cwnd_gain, bbr_cwnd_gain, int, 0644); +module_param_array_named(pacing_gain, bbr_pacing_gain, int, + &bbr_pacing_gain_size, 0644); +module_param_named(cwnd_min_target, bbr_cwnd_min_target, uint, 0644); +module_param_named(probe_rtt_cwnd_gain, + bbr_probe_rtt_cwnd_gain, uint, 0664); +module_param_named(cwnd_warn_val, bbr_cwnd_warn_val, uint, 0664); +module_param_named(debug_port_mask, bbr_debug_port_mask, ushort, 0644); +module_param_named(flags, bbr_flags, uint, 0644); +module_param_named(debug_ftrace, bbr_debug_ftrace, bool, 0644); +module_param_named(debug_with_printk, bbr_debug_with_printk, bool, 0644); +module_param_named(min_rtt_win_sec, bbr_min_rtt_win_sec, uint, 0644); +module_param_named(probe_rtt_mode_ms, bbr_probe_rtt_mode_ms, uint, 0644); +module_param_named(probe_rtt_win_ms, bbr_probe_rtt_win_ms, uint, 0644); +module_param_named(full_bw_thresh, bbr_full_bw_thresh, uint, 0644); +module_param_named(full_bw_cnt, bbr_full_bw_cnt, uint, 0644); +module_param_named(cwnd_tso_bduget, bbr_cwnd_tso_budget, uint, 0664); +module_param_named(extra_acked_gain, bbr_extra_acked_gain, int, 0664); +module_param_named(extra_acked_win_rtts, + bbr_extra_acked_win_rtts, uint, 0664); +module_param_named(extra_acked_max_us, + bbr_extra_acked_max_us, uint, 0664); +module_param_named(ack_epoch_acked_reset_thresh, + bbr_ack_epoch_acked_reset_thresh, uint, 0664); +module_param_named(drain_to_target, bbr_drain_to_target, bool, 0664); +module_param_named(precise_ece_ack, bbr_precise_ece_ack, bool, 0664); +module_param_named(extra_acked_in_startup, + bbr_extra_acked_in_startup, int, 0664); +module_param_named(usage_based_cwnd, bbr_usage_based_cwnd, bool, 0664); +module_param_named(ecn_enable, bbr_ecn_enable, bool, 0664); + +static void bbr2_exit_probe_rtt(struct sock *sk); +static void bbr2_reset_congestion_signals(struct sock *sk); + +static void bbr_check_probe_rtt_done(struct sock *sk); + +/* Do we estimate that STARTUP filled the pipe? */ +static bool bbr_full_bw_reached(const struct sock *sk) +{ + const struct bbr *bbr = inet_csk_ca(sk); + + return bbr->full_bw_reached; +} + +/* Return the windowed max recent bandwidth sample, in pkts/uS << BW_SCALE. */ +static u32 bbr_max_bw(const struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + return max(bbr->bw_hi[0], bbr->bw_hi[1]); +} + +/* Return the estimated bandwidth of the path, in pkts/uS << BW_SCALE. */ +static u32 bbr_bw(const struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + return min(bbr_max_bw(sk), bbr->bw_lo); +} + +/* Return maximum extra acked in past k-2k round trips, + * where k = bbr_extra_acked_win_rtts. + */ +static u16 bbr_extra_acked(const struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + return max(bbr->extra_acked[0], bbr->extra_acked[1]); +} + +/* Return rate in bytes per second, optionally with a gain. + * The order here is chosen carefully to avoid overflow of u64. This should + * work for input rates of up to 2.9Tbit/sec and gain of 2.89x. + */ +static u64 bbr_rate_bytes_per_sec(struct sock *sk, u64 rate, int gain, + int margin) +{ + unsigned int mss = tcp_sk(sk)->mss_cache; + + rate *= mss; + rate *= gain; + rate >>= BBR_SCALE; + rate *= USEC_PER_SEC / 100 * (100 - margin); + rate >>= BW_SCALE; + rate = max(rate, 1ULL); + return rate; +} + +static u64 bbr_bw_bytes_per_sec(struct sock *sk, u64 rate) +{ + return bbr_rate_bytes_per_sec(sk, rate, BBR_UNIT, 0); +} + +static u64 bbr_rate_kbps(struct sock *sk, u64 rate) +{ + rate = bbr_bw_bytes_per_sec(sk, rate); + rate *= 8; + do_div(rate, 1000); + return rate; +} + +static u32 bbr_tso_segs_goal(struct sock *sk); +static void bbr_debug(struct sock *sk, u32 acked, + const struct rate_sample *rs, struct bbr_context *ctx) +{ + static const char ca_states[] = { + [TCP_CA_Open] = 'O', + [TCP_CA_Disorder] = 'D', + [TCP_CA_CWR] = 'C', + [TCP_CA_Recovery] = 'R', + [TCP_CA_Loss] = 'L', + }; + static const char mode[] = { + 'G', /* Growing - BBR_STARTUP */ + 'D', /* Drain - BBR_DRAIN */ + 'W', /* Window - BBR_PROBE_BW */ + 'M', /* Min RTT - BBR_PROBE_RTT */ + }; + static const char ack_phase[] = { /* bbr_ack_phase strings */ + 'I', /* BBR_ACKS_INIT - 'Init' */ + 'R', /* BBR_ACKS_REFILLING - 'Refilling' */ + 'B', /* BBR_ACKS_PROBE_STARTING - 'Before' */ + 'F', /* BBR_ACKS_PROBE_FEEDBACK - 'Feedback' */ + 'A', /* BBR_ACKS_PROBE_STOPPING - 'After' */ + }; + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + const u32 una = tp->snd_una - bbr->debug.snd_isn; + const u32 fack = tcp_highest_sack_seq(tp); + const u16 dport = ntohs(inet_sk(sk)->inet_dport); + bool is_port_match = (bbr_debug_port_mask && + ((dport & bbr_debug_port_mask) == 0)); + char debugmsg[320]; + + if (sk->sk_state == TCP_SYN_SENT) + return; /* no bbr_init() yet if SYN retransmit -> CA_Loss */ + + if (!tp->snd_cwnd || tp->snd_cwnd > bbr_cwnd_warn_val) { + char addr[INET6_ADDRSTRLEN + 10] = { 0 }; + + if (sk->sk_family == AF_INET) + snprintf(addr, sizeof(addr), "%pI4:%u", + &inet_sk(sk)->inet_daddr, dport); + else if (sk->sk_family == AF_INET6) + snprintf(addr, sizeof(addr), "%pI6:%u", + &sk->sk_v6_daddr, dport); + + WARN_ONCE(1, + "BBR %s cwnd alert: %u " + "snd_una: %u ca: %d pacing_gain: %u cwnd_gain: %u " + "bw: %u rtt: %u min_rtt: %u " + "acked: %u tso_segs: %u " + "bw: %d %ld %d pif: %u\n", + addr, tp->snd_cwnd, + una, inet_csk(sk)->icsk_ca_state, + bbr->pacing_gain, bbr->cwnd_gain, + bbr_max_bw(sk), (tp->srtt_us >> 3), bbr->min_rtt_us, + acked, bbr_tso_segs_goal(sk), + rs->delivered, rs->interval_us, rs->is_retrans, + tcp_packets_in_flight(tp)); + } + + if (likely(!bbr_debug_with_printk && !bbr_debug_ftrace)) + return; + + if (!sock_flag(sk, SOCK_DBG) && !is_port_match) + return; + + if (!ctx->log && !tp->app_limited && !(bbr_flags & FLAG_DEBUG_VERBOSE)) + return; + + if (ipv4_is_loopback(inet_sk(sk)->inet_daddr) && + !(bbr_flags & FLAG_DEBUG_LOOPBACK)) + return; + + snprintf(debugmsg, sizeof(debugmsg) - 1, + "BBR %pI4:%-5u %5u,%03u:%-7u %c " + "%c %2u br %2u cr %2d rtt %5ld d %2d i %5ld mrtt %d %cbw %llu " + "bw %llu lb %llu ib %llu qb %llu " + "a %u if %2u %c %c dl %u l %u al %u # %u t %u %c %c " + "lr %d er %d ea %d bwl %lld il %d ih %d c %d " + "v %d %c %u %c %s\n", + &inet_sk(sk)->inet_daddr, dport, + una / 1000, una % 1000, fack - tp->snd_una, + ca_states[inet_csk(sk)->icsk_ca_state], + bbr->debug.undo ? '@' : mode[bbr->mode], + tp->snd_cwnd, + bbr_extra_acked(sk), /* br (legacy): extra_acked */ + rs->tx_in_flight, /* cr (legacy): tx_inflight */ + rs->rtt_us, + rs->delivered, + rs->interval_us, + bbr->min_rtt_us, + rs->is_app_limited ? '_' : 'l', + bbr_rate_kbps(sk, ctx->sample_bw), /* lbw: latest sample bw */ + bbr_rate_kbps(sk, bbr_max_bw(sk)), /* bw: max bw */ + 0ULL, /* lb: [obsolete] */ + 0ULL, /* ib: [obsolete] */ + (u64)sk->sk_pacing_rate * 8 / 1000, + acked, + tcp_packets_in_flight(tp), + rs->is_ack_delayed ? 'd' : '.', + bbr->round_start ? '*' : '.', + tp->delivered, tp->lost, + tp->app_limited, + 0, /* #: [obsolete] */ + ctx->target_cwnd, + tp->reord_seen ? 'r' : '.', /* r: reordering seen? */ + ca_states[bbr->prev_ca_state], + (rs->lost + rs->delivered) > 0 ? + (1000 * rs->lost / + (rs->lost + rs->delivered)) : 0, /* lr: loss rate x1000 */ + (rs->delivered) > 0 ? + (1000 * rs->delivered_ce / + (rs->delivered)) : 0, /* er: ECN rate x1000 */ + 1000 * bbr->ecn_alpha >> BBR_SCALE, /* ea: ECN alpha x1000 */ + bbr->bw_lo == ~0U ? + -1 : (s64)bbr_rate_kbps(sk, bbr->bw_lo), /* bwl */ + bbr->inflight_lo, /* il */ + bbr->inflight_hi, /* ih */ + bbr->bw_probe_up_cnt, /* c */ + 2, /* v: version */ + bbr->debug.event, + bbr->cycle_idx, + ack_phase[bbr->ack_phase], + bbr->bw_probe_samples ? "Y" : "N"); + debugmsg[sizeof(debugmsg) - 1] = 0; + + /* printk takes a higher precedence. */ + if (bbr_debug_with_printk) + printk(KERN_DEBUG "%s", debugmsg); + + if (unlikely(bbr->debug.undo)) + bbr->debug.undo = 0; +} + +/* Convert a BBR bw and gain factor to a pacing rate in bytes per second. */ +static unsigned long bbr_bw_to_pacing_rate(struct sock *sk, u32 bw, int gain) +{ + u64 rate = bw; + + rate = bbr_rate_bytes_per_sec(sk, rate, gain, + bbr_pacing_margin_percent); + rate = min_t(u64, rate, sk->sk_max_pacing_rate); + return rate; +} + +/* Initialize pacing rate to: high_gain * init_cwnd / RTT. */ +static void bbr_init_pacing_rate_from_rtt(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u64 bw; + u32 rtt_us; + + if (tp->srtt_us) { /* any RTT sample yet? */ + rtt_us = max(tp->srtt_us >> 3, 1U); + bbr->has_seen_rtt = 1; + } else { /* no RTT sample yet */ + rtt_us = USEC_PER_MSEC; /* use nominal default RTT */ + } + bw = (u64)tp->snd_cwnd * BW_UNIT; + do_div(bw, rtt_us); + sk->sk_pacing_rate = bbr_bw_to_pacing_rate(sk, bw, bbr->params.high_gain); +} + +/* Pace using current bw estimate and a gain factor. */ +static void bbr_set_pacing_rate(struct sock *sk, u32 bw, int gain) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + unsigned long rate = bbr_bw_to_pacing_rate(sk, bw, gain); + + if (unlikely(!bbr->has_seen_rtt && tp->srtt_us)) + bbr_init_pacing_rate_from_rtt(sk); + if (bbr_full_bw_reached(sk) || rate > sk->sk_pacing_rate) + sk->sk_pacing_rate = rate; +} + +static u32 bbr_min_tso_segs(struct sock *sk) +{ + return sk->sk_pacing_rate < (bbr_min_tso_rate >> 3) ? 1 : 2; +} + +/* Return the number of segments BBR would like in a TSO/GSO skb, given + * a particular max gso size as a constraint. + */ +static u32 bbr_tso_segs_generic(struct sock *sk, unsigned int mss_now, + u32 gso_max_size) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 segs, r; + u64 bytes; + + /* Budget a TSO/GSO burst size allowance based on bw (pacing_rate). */ + bytes = sk->sk_pacing_rate >> sk->sk_pacing_shift; + + /* Budget a TSO/GSO burst size allowance based on min_rtt. For every + * K = 2^tso_rtt_shift microseconds of min_rtt, halve the burst. + * The min_rtt-based burst allowance is: 64 KBytes / 2^(min_rtt/K) + */ + if (bbr->params.tso_rtt_shift) { + r = bbr->min_rtt_us >> bbr->params.tso_rtt_shift; + if (r < BITS_PER_TYPE(u32)) /* prevent undefined behavior */ + bytes += GSO_MAX_SIZE >> r; + } + + bytes = min_t(u32, bytes, gso_max_size - 1 - MAX_TCP_HEADER); + segs = max_t(u32, bytes / mss_now, bbr_min_tso_segs(sk)); + return segs; +} + +/* Custom tcp_tso_autosize() for BBR, used at transmit time to cap skb size. */ +static u32 bbr_tso_segs(struct sock *sk, unsigned int mss_now) +{ + return bbr_tso_segs_generic(sk, mss_now, sk->sk_gso_max_size); +} + +/* Like bbr_tso_segs(), using mss_cache, ignoring driver's sk_gso_max_size. */ +static u32 bbr_tso_segs_goal(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + + return bbr_tso_segs_generic(sk, tp->mss_cache, GSO_MAX_SIZE); +} + +/* Save "last known good" cwnd so we can restore it after losses or PROBE_RTT */ +static void bbr_save_cwnd(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr->prev_ca_state < TCP_CA_Recovery && bbr->mode != BBR_PROBE_RTT) + bbr->prior_cwnd = tp->snd_cwnd; /* this cwnd is good enough */ + else /* loss recovery or BBR_PROBE_RTT have temporarily cut cwnd */ + bbr->prior_cwnd = max(bbr->prior_cwnd, tp->snd_cwnd); +} + +static void bbr_cwnd_event(struct sock *sk, enum tcp_ca_event event) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + if (event == CA_EVENT_TX_START && tp->app_limited) { + bbr->idle_restart = 1; + bbr->ack_epoch_mstamp = tp->tcp_mstamp; + bbr->ack_epoch_acked = 0; + /* Avoid pointless buffer overflows: pace at est. bw if we don't + * need more speed (we're restarting from idle and app-limited). + */ + if (bbr->mode == BBR_PROBE_BW) + bbr_set_pacing_rate(sk, bbr_bw(sk), BBR_UNIT); + else if (bbr->mode == BBR_PROBE_RTT) + bbr_check_probe_rtt_done(sk); + } else if ((event == CA_EVENT_ECN_IS_CE || + event == CA_EVENT_ECN_NO_CE) && + bbr_ecn_enable && + bbr->params.precise_ece_ack) { + u32 state = bbr->ce_state; + dctcp_ece_ack_update(sk, event, &bbr->prior_rcv_nxt, &state); + bbr->ce_state = state; + if (tp->fast_ack_mode == 2 && event == CA_EVENT_ECN_IS_CE) + tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS); + } +} + +/* Calculate bdp based on min RTT and the estimated bottleneck bandwidth: + * + * bdp = ceil(bw * min_rtt * gain) + * + * The key factor, gain, controls the amount of queue. While a small gain + * builds a smaller queue, it becomes more vulnerable to noise in RTT + * measurements (e.g., delayed ACKs or other ACK compression effects). This + * noise may cause BBR to under-estimate the rate. + */ +static u32 bbr_bdp(struct sock *sk, u32 bw, int gain) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 bdp; + u64 w; + + /* If we've never had a valid RTT sample, cap cwnd at the initial + * default. This should only happen when the connection is not using TCP + * timestamps and has retransmitted all of the SYN/SYNACK/data packets + * ACKed so far. In this case, an RTO can cut cwnd to 1, in which + * case we need to slow-start up toward something safe: initial cwnd. + */ + if (unlikely(bbr->min_rtt_us == ~0U)) /* no valid RTT samples yet? */ + return bbr->init_cwnd; /* be safe: cap at initial cwnd */ + + w = (u64)bw * bbr->min_rtt_us; + + /* Apply a gain to the given value, remove the BW_SCALE shift, and + * round the value up to avoid a negative feedback loop. + */ + bdp = (((w * gain) >> BBR_SCALE) + BW_UNIT - 1) / BW_UNIT; + + return bdp; +} + +/* To achieve full performance in high-speed paths, we budget enough cwnd to + * fit full-sized skbs in-flight on both end hosts to fully utilize the path: + * - one skb in sending host Qdisc, + * - one skb in sending host TSO/GSO engine + * - one skb being received by receiver host LRO/GRO/delayed-ACK engine + * Don't worry, at low rates (bbr_min_tso_rate) this won't bloat cwnd because + * in such cases tso_segs_goal is 1. The minimum cwnd is 4 packets, + * which allows 2 outstanding 2-packet sequences, to try to keep pipe + * full even with ACK-every-other-packet delayed ACKs. + */ +static u32 bbr_quantization_budget(struct sock *sk, u32 cwnd) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 tso_segs_goal; + + tso_segs_goal = 3 * bbr_tso_segs_goal(sk); + + /* Allow enough full-sized skbs in flight to utilize end systems. */ + if (bbr->params.cwnd_tso_budget == 1) { + cwnd = max_t(u32, cwnd, tso_segs_goal); + cwnd = max_t(u32, cwnd, bbr->params.cwnd_min_target); + } else { + cwnd += tso_segs_goal; + cwnd = (cwnd + 1) & ~1U; + } + /* Ensure gain cycling gets inflight above BDP even for small BDPs. */ + if (bbr->mode == BBR_PROBE_BW && bbr->cycle_idx == BBR_BW_PROBE_UP) + cwnd += 2; + + return cwnd; +} + +/* Find inflight based on min RTT and the estimated bottleneck bandwidth. */ +static u32 bbr_inflight(struct sock *sk, u32 bw, int gain) +{ + u32 inflight; + + inflight = bbr_bdp(sk, bw, gain); + inflight = bbr_quantization_budget(sk, inflight); + + return inflight; +} + +/* With pacing at lower layers, there's often less data "in the network" than + * "in flight". With TSQ and departure time pacing at lower layers (e.g. fq), + * we often have several skbs queued in the pacing layer with a pre-scheduled + * earliest departure time (EDT). BBR adapts its pacing rate based on the + * inflight level that it estimates has already been "baked in" by previous + * departure time decisions. We calculate a rough estimate of the number of our + * packets that might be in the network at the earliest departure time for the + * next skb scheduled: + * in_network_at_edt = inflight_at_edt - (EDT - now) * bw + * If we're increasing inflight, then we want to know if the transmit of the + * EDT skb will push inflight above the target, so inflight_at_edt includes + * bbr_tso_segs_goal() from the skb departing at EDT. If decreasing inflight, + * then estimate if inflight will sink too low just before the EDT transmit. + */ +static u32 bbr_packets_in_net_at_edt(struct sock *sk, u32 inflight_now) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u64 now_ns, edt_ns, interval_us; + u32 interval_delivered, inflight_at_edt; + + now_ns = tp->tcp_clock_cache; + edt_ns = max(tp->tcp_wstamp_ns, now_ns); + interval_us = div_u64(edt_ns - now_ns, NSEC_PER_USEC); + interval_delivered = (u64)bbr_bw(sk) * interval_us >> BW_SCALE; + inflight_at_edt = inflight_now; + if (bbr->pacing_gain > BBR_UNIT) /* increasing inflight */ + inflight_at_edt += bbr_tso_segs_goal(sk); /* include EDT skb */ + if (interval_delivered >= inflight_at_edt) + return 0; + return inflight_at_edt - interval_delivered; +} + +/* Find the cwnd increment based on estimate of ack aggregation */ +static u32 bbr_ack_aggregation_cwnd(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 max_aggr_cwnd, aggr_cwnd = 0; + + if (bbr->params.extra_acked_gain && + (bbr_full_bw_reached(sk) || bbr->params.extra_acked_in_startup)) { + max_aggr_cwnd = ((u64)bbr_bw(sk) * bbr_extra_acked_max_us) + / BW_UNIT; + aggr_cwnd = (bbr->params.extra_acked_gain * bbr_extra_acked(sk)) + >> BBR_SCALE; + aggr_cwnd = min(aggr_cwnd, max_aggr_cwnd); + } + + return aggr_cwnd; +} + +/* Returns the cwnd for PROBE_RTT mode. */ +static u32 bbr_probe_rtt_cwnd(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr->params.probe_rtt_cwnd_gain == 0) + return bbr->params.cwnd_min_target; + return max_t(u32, bbr->params.cwnd_min_target, + bbr_bdp(sk, bbr_bw(sk), bbr->params.probe_rtt_cwnd_gain)); +} + +/* Slow-start up toward target cwnd (if bw estimate is growing, or packet loss + * has drawn us down below target), or snap down to target if we're above it. + */ +static void bbr_set_cwnd(struct sock *sk, const struct rate_sample *rs, + u32 acked, u32 bw, int gain, u32 cwnd, + struct bbr_context *ctx) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u32 target_cwnd = 0, prev_cwnd = tp->snd_cwnd, max_probe; + + if (!acked) + goto done; /* no packet fully ACKed; just apply caps */ + + target_cwnd = bbr_bdp(sk, bw, gain); + + /* Increment the cwnd to account for excess ACKed data that seems + * due to aggregation (of data and/or ACKs) visible in the ACK stream. + */ + target_cwnd += bbr_ack_aggregation_cwnd(sk); + target_cwnd = bbr_quantization_budget(sk, target_cwnd); + + /* If we're below target cwnd, slow start cwnd toward target cwnd. */ + bbr->debug.target_cwnd = target_cwnd; + + /* Update cwnd and enable fast path if cwnd reaches target_cwnd. */ + bbr->try_fast_path = 0; + if (bbr_full_bw_reached(sk)) { /* only cut cwnd if we filled the pipe */ + cwnd += acked; + if (cwnd >= target_cwnd) { + cwnd = target_cwnd; + bbr->try_fast_path = 1; + } + } else if (cwnd < target_cwnd || cwnd < 2 * bbr->init_cwnd) { + cwnd += acked; + } else { + bbr->try_fast_path = 1; + } + + /* When growing cwnd, don't grow beyond twice what we just probed. */ + if (bbr->params.usage_based_cwnd) { + max_probe = max(2 * tp->max_packets_out, tp->snd_cwnd); + cwnd = min(cwnd, max_probe); + } + + cwnd = max_t(u32, cwnd, bbr->params.cwnd_min_target); +done: + tp->snd_cwnd = min(cwnd, tp->snd_cwnd_clamp); /* apply global cap */ + if (bbr->mode == BBR_PROBE_RTT) /* drain queue, refresh min_rtt */ + tp->snd_cwnd = min_t(u32, tp->snd_cwnd, bbr_probe_rtt_cwnd(sk)); + + ctx->target_cwnd = target_cwnd; + ctx->log = (tp->snd_cwnd != prev_cwnd); +} + +/* See if we have reached next round trip */ +static void bbr_update_round_start(struct sock *sk, + const struct rate_sample *rs, struct bbr_context *ctx) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr->round_start = 0; + + /* See if we've reached the next RTT */ + if (rs->interval_us > 0 && + !before(rs->prior_delivered, bbr->next_rtt_delivered)) { + bbr->next_rtt_delivered = tp->delivered; + bbr->round_start = 1; + } +} + +/* Calculate the bandwidth based on how fast packets are delivered */ +static void bbr_calculate_bw_sample(struct sock *sk, + const struct rate_sample *rs, struct bbr_context *ctx) +{ + struct bbr *bbr = inet_csk_ca(sk); + u64 bw = 0; + + /* Divide delivered by the interval to find a (lower bound) bottleneck + * bandwidth sample. Delivered is in packets and interval_us in uS and + * ratio will be <<1 for most connections. So delivered is first scaled. + * Round up to allow growth at low rates, even with integer division. + */ + if (rs->interval_us > 0) { + if (WARN_ONCE(rs->delivered < 0, + "negative delivered: %d interval_us: %ld\n", + rs->delivered, rs->interval_us)) + return; + + bw = DIV_ROUND_UP_ULL((u64)rs->delivered * BW_UNIT, rs->interval_us); + } + + ctx->sample_bw = bw; + bbr->debug.rs_bw = bw; +} + +/* Estimates the windowed max degree of ack aggregation. + * This is used to provision extra in-flight data to keep sending during + * inter-ACK silences. + * + * Degree of ack aggregation is estimated as extra data acked beyond expected. + * + * max_extra_acked = "maximum recent excess data ACKed beyond max_bw * interval" + * cwnd += max_extra_acked + * + * Max extra_acked is clamped by cwnd and bw * bbr_extra_acked_max_us (100 ms). + * Max filter is an approximate sliding window of 5-10 (packet timed) round + * trips for non-startup phase, and 1-2 round trips for startup. + */ +static void bbr_update_ack_aggregation(struct sock *sk, + const struct rate_sample *rs) +{ + u32 epoch_us, expected_acked, extra_acked; + struct bbr *bbr = inet_csk_ca(sk); + struct tcp_sock *tp = tcp_sk(sk); + u32 extra_acked_win_rtts_thresh = bbr->params.extra_acked_win_rtts; + + if (!bbr->params.extra_acked_gain || rs->acked_sacked <= 0 || + rs->delivered < 0 || rs->interval_us <= 0) + return; + + if (bbr->round_start) { + bbr->extra_acked_win_rtts = min(0x1F, + bbr->extra_acked_win_rtts + 1); + if (bbr->params.extra_acked_in_startup && + !bbr_full_bw_reached(sk)) + extra_acked_win_rtts_thresh = 1; + if (bbr->extra_acked_win_rtts >= + extra_acked_win_rtts_thresh) { + bbr->extra_acked_win_rtts = 0; + bbr->extra_acked_win_idx = bbr->extra_acked_win_idx ? + 0 : 1; + bbr->extra_acked[bbr->extra_acked_win_idx] = 0; + } + } + + /* Compute how many packets we expected to be delivered over epoch. */ + epoch_us = tcp_stamp_us_delta(tp->delivered_mstamp, + bbr->ack_epoch_mstamp); + expected_acked = ((u64)bbr_bw(sk) * epoch_us) / BW_UNIT; + + /* Reset the aggregation epoch if ACK rate is below expected rate or + * significantly large no. of ack received since epoch (potentially + * quite old epoch). + */ + if (bbr->ack_epoch_acked <= expected_acked || + (bbr->ack_epoch_acked + rs->acked_sacked >= + bbr_ack_epoch_acked_reset_thresh)) { + bbr->ack_epoch_acked = 0; + bbr->ack_epoch_mstamp = tp->delivered_mstamp; + expected_acked = 0; + } + + /* Compute excess data delivered, beyond what was expected. */ + bbr->ack_epoch_acked = min_t(u32, 0xFFFFF, + bbr->ack_epoch_acked + rs->acked_sacked); + extra_acked = bbr->ack_epoch_acked - expected_acked; + extra_acked = min(extra_acked, tp->snd_cwnd); + if (extra_acked > bbr->extra_acked[bbr->extra_acked_win_idx]) + bbr->extra_acked[bbr->extra_acked_win_idx] = extra_acked; +} + +/* Estimate when the pipe is full, using the change in delivery rate: BBR + * estimates that STARTUP filled the pipe if the estimated bw hasn't changed by + * at least bbr_full_bw_thresh (25%) after bbr_full_bw_cnt (3) non-app-limited + * rounds. Why 3 rounds: 1: rwin autotuning grows the rwin, 2: we fill the + * higher rwin, 3: we get higher delivery rate samples. Or transient + * cross-traffic or radio noise can go away. CUBIC Hystart shares a similar + * design goal, but uses delay and inter-ACK spacing instead of bandwidth. + */ +static void bbr_check_full_bw_reached(struct sock *sk, + const struct rate_sample *rs) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 bw_thresh; + + if (bbr_full_bw_reached(sk) || !bbr->round_start || rs->is_app_limited) + return; + + bw_thresh = (u64)bbr->full_bw * bbr->params.full_bw_thresh >> BBR_SCALE; + if (bbr_max_bw(sk) >= bw_thresh) { + bbr->full_bw = bbr_max_bw(sk); + bbr->full_bw_cnt = 0; + return; + } + ++bbr->full_bw_cnt; + bbr->full_bw_reached = bbr->full_bw_cnt >= bbr->params.full_bw_cnt; +} + +/* If pipe is probably full, drain the queue and then enter steady-state. */ +static bool bbr_check_drain(struct sock *sk, const struct rate_sample *rs, + struct bbr_context *ctx) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr->mode == BBR_STARTUP && bbr_full_bw_reached(sk)) { + bbr->mode = BBR_DRAIN; /* drain queue we created */ + tcp_sk(sk)->snd_ssthresh = + bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT); + bbr2_reset_congestion_signals(sk); + } /* fall through to check if in-flight is already small: */ + if (bbr->mode == BBR_DRAIN && + bbr_packets_in_net_at_edt(sk, tcp_packets_in_flight(tcp_sk(sk))) <= + bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT)) + return true; /* exiting DRAIN now */ + return false; +} + +static void bbr_check_probe_rtt_done(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + if (!(bbr->probe_rtt_done_stamp && + after(tcp_jiffies32, bbr->probe_rtt_done_stamp))) + return; + + bbr->probe_rtt_min_stamp = tcp_jiffies32; /* schedule next PROBE_RTT */ + tp->snd_cwnd = max(tp->snd_cwnd, bbr->prior_cwnd); + bbr2_exit_probe_rtt(sk); +} + +/* The goal of PROBE_RTT mode is to have BBR flows cooperatively and + * periodically drain the bottleneck queue, to converge to measure the true + * min_rtt (unloaded propagation delay). This allows the flows to keep queues + * small (reducing queuing delay and packet loss) and achieve fairness among + * BBR flows. + * + * The min_rtt filter window is 10 seconds. When the min_rtt estimate expires, + * we enter PROBE_RTT mode and cap the cwnd at bbr_cwnd_min_target=4 packets. + * After at least bbr_probe_rtt_mode_ms=200ms and at least one packet-timed + * round trip elapsed with that flight size <= 4, we leave PROBE_RTT mode and + * re-enter the previous mode. BBR uses 200ms to approximately bound the + * performance penalty of PROBE_RTT's cwnd capping to roughly 2% (200ms/10s). + * + * Note that flows need only pay 2% if they are busy sending over the last 10 + * seconds. Interactive applications (e.g., Web, RPCs, video chunks) often have + * natural silences or low-rate periods within 10 seconds where the rate is low + * enough for long enough to drain its queue in the bottleneck. We pick up + * these min RTT measurements opportunistically with our min_rtt filter. :-) + */ +static void bbr_update_min_rtt(struct sock *sk, const struct rate_sample *rs) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + bool probe_rtt_expired, min_rtt_expired; + u32 expire; + + /* Track min RTT in probe_rtt_win_ms to time next PROBE_RTT state. */ + expire = bbr->probe_rtt_min_stamp + + msecs_to_jiffies(bbr->params.probe_rtt_win_ms); + probe_rtt_expired = after(tcp_jiffies32, expire); + if (rs->rtt_us >= 0 && + (rs->rtt_us <= bbr->probe_rtt_min_us || + (probe_rtt_expired && !rs->is_ack_delayed))) { + bbr->probe_rtt_min_us = rs->rtt_us; + bbr->probe_rtt_min_stamp = tcp_jiffies32; + } + /* Track min RTT seen in the min_rtt_win_sec filter window: */ + expire = bbr->min_rtt_stamp + bbr->params.min_rtt_win_sec * HZ; + min_rtt_expired = after(tcp_jiffies32, expire); + if (bbr->probe_rtt_min_us <= bbr->min_rtt_us || + min_rtt_expired) { + bbr->min_rtt_us = bbr->probe_rtt_min_us; + bbr->min_rtt_stamp = bbr->probe_rtt_min_stamp; + } + + if (bbr->params.probe_rtt_mode_ms > 0 && probe_rtt_expired && + !bbr->idle_restart && bbr->mode != BBR_PROBE_RTT) { + bbr->mode = BBR_PROBE_RTT; /* dip, drain queue */ + bbr_save_cwnd(sk); /* note cwnd so we can restore it */ + bbr->probe_rtt_done_stamp = 0; + bbr->ack_phase = BBR_ACKS_PROBE_STOPPING; + bbr->next_rtt_delivered = tp->delivered; + } + + if (bbr->mode == BBR_PROBE_RTT) { + /* Ignore low rate samples during this mode. */ + tp->app_limited = + (tp->delivered + tcp_packets_in_flight(tp)) ? : 1; + /* Maintain min packets in flight for max(200 ms, 1 round). */ + if (!bbr->probe_rtt_done_stamp && + tcp_packets_in_flight(tp) <= bbr_probe_rtt_cwnd(sk)) { + bbr->probe_rtt_done_stamp = tcp_jiffies32 + + msecs_to_jiffies(bbr->params.probe_rtt_mode_ms); + bbr->probe_rtt_round_done = 0; + bbr->next_rtt_delivered = tp->delivered; + } else if (bbr->probe_rtt_done_stamp) { + if (bbr->round_start) + bbr->probe_rtt_round_done = 1; + if (bbr->probe_rtt_round_done) + bbr_check_probe_rtt_done(sk); + } + } + /* Restart after idle ends only once we process a new S/ACK for data */ + if (rs->delivered > 0) + bbr->idle_restart = 0; +} + +static void bbr_update_gains(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + switch (bbr->mode) { + case BBR_STARTUP: + bbr->pacing_gain = bbr->params.high_gain; + bbr->cwnd_gain = bbr->params.startup_cwnd_gain; + break; + case BBR_DRAIN: + bbr->pacing_gain = bbr->params.drain_gain; /* slow, to drain */ + bbr->cwnd_gain = bbr->params.startup_cwnd_gain; /* keep cwnd */ + break; + case BBR_PROBE_BW: + bbr->pacing_gain = bbr->params.pacing_gain[bbr->cycle_idx]; + bbr->cwnd_gain = bbr->params.cwnd_gain; + break; + case BBR_PROBE_RTT: + bbr->pacing_gain = BBR_UNIT; + bbr->cwnd_gain = BBR_UNIT; + break; + default: + WARN_ONCE(1, "BBR bad mode: %u\n", bbr->mode); + break; + } +} + +static void bbr_init(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + int i; + + WARN_ON_ONCE(tp->snd_cwnd >= bbr_cwnd_warn_val); + + bbr->initialized = 1; + bbr->params.high_gain = min(0x7FF, bbr_high_gain); + bbr->params.drain_gain = min(0x3FF, bbr_drain_gain); + bbr->params.startup_cwnd_gain = min(0x7FF, bbr_startup_cwnd_gain); + bbr->params.cwnd_gain = min(0x7FF, bbr_cwnd_gain); + bbr->params.cwnd_tso_budget = min(0x1U, bbr_cwnd_tso_budget); + bbr->params.cwnd_min_target = min(0xFU, bbr_cwnd_min_target); + bbr->params.min_rtt_win_sec = min(0x1FU, bbr_min_rtt_win_sec); + bbr->params.probe_rtt_mode_ms = min(0x1FFU, bbr_probe_rtt_mode_ms); + bbr->params.full_bw_cnt = min(0x7U, bbr_full_bw_cnt); + bbr->params.full_bw_thresh = min(0x3FFU, bbr_full_bw_thresh); + bbr->params.extra_acked_gain = min(0x7FF, bbr_extra_acked_gain); + bbr->params.extra_acked_win_rtts = min(0x1FU, bbr_extra_acked_win_rtts); + bbr->params.drain_to_target = bbr_drain_to_target ? 1 : 0; + bbr->params.precise_ece_ack = bbr_precise_ece_ack ? 1 : 0; + bbr->params.extra_acked_in_startup = bbr_extra_acked_in_startup ? 1 : 0; + bbr->params.probe_rtt_cwnd_gain = min(0xFFU, bbr_probe_rtt_cwnd_gain); + bbr->params.probe_rtt_win_ms = + min(0x3FFFU, + min_t(u32, bbr_probe_rtt_win_ms, + bbr->params.min_rtt_win_sec * MSEC_PER_SEC)); + for (i = 0; i < CYCLE_LEN; i++) + bbr->params.pacing_gain[i] = min(0x3FF, bbr_pacing_gain[i]); + bbr->params.usage_based_cwnd = bbr_usage_based_cwnd ? 1 : 0; + bbr->params.tso_rtt_shift = min(0xFU, bbr_tso_rtt_shift); + + bbr->debug.snd_isn = tp->snd_una; + bbr->debug.target_cwnd = 0; + bbr->debug.undo = 0; + + bbr->init_cwnd = min(0x7FU, tp->snd_cwnd); + bbr->prior_cwnd = tp->prior_cwnd; + tp->snd_ssthresh = TCP_INFINITE_SSTHRESH; + bbr->next_rtt_delivered = 0; + bbr->prev_ca_state = TCP_CA_Open; + bbr->packet_conservation = 0; + + bbr->probe_rtt_done_stamp = 0; + bbr->probe_rtt_round_done = 0; + bbr->probe_rtt_min_us = tcp_min_rtt(tp); + bbr->probe_rtt_min_stamp = tcp_jiffies32; + bbr->min_rtt_us = tcp_min_rtt(tp); + bbr->min_rtt_stamp = tcp_jiffies32; + + bbr->has_seen_rtt = 0; + bbr_init_pacing_rate_from_rtt(sk); + + bbr->round_start = 0; + bbr->idle_restart = 0; + bbr->full_bw_reached = 0; + bbr->full_bw = 0; + bbr->full_bw_cnt = 0; + bbr->cycle_mstamp = 0; + bbr->cycle_idx = 0; + bbr->mode = BBR_STARTUP; + bbr->debug.rs_bw = 0; + + bbr->ack_epoch_mstamp = tp->tcp_mstamp; + bbr->ack_epoch_acked = 0; + bbr->extra_acked_win_rtts = 0; + bbr->extra_acked_win_idx = 0; + bbr->extra_acked[0] = 0; + bbr->extra_acked[1] = 0; + + bbr->ce_state = 0; + bbr->prior_rcv_nxt = tp->rcv_nxt; + bbr->try_fast_path = 0; + + cmpxchg(&sk->sk_pacing_status, SK_PACING_NONE, SK_PACING_NEEDED); +} + +static u32 bbr_sndbuf_expand(struct sock *sk) +{ + /* Provision 3 * cwnd since BBR may slow-start even during recovery. */ + return 3; +} + +/* __________________________________________________________________________ + * + * Functions new to BBR v2 ("bbr") congestion control are below here. + * __________________________________________________________________________ + */ + +/* Incorporate a new bw sample into the current window of our max filter. */ +static void bbr2_take_bw_hi_sample(struct sock *sk, u32 bw) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr->bw_hi[1] = max(bw, bbr->bw_hi[1]); +} + +/* Keep max of last 1-2 cycles. Each PROBE_BW cycle, flip filter window. */ +static void bbr2_advance_bw_hi_filter(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (!bbr->bw_hi[1]) + return; /* no samples in this window; remember old window */ + bbr->bw_hi[0] = bbr->bw_hi[1]; + bbr->bw_hi[1] = 0; +} + +/* How much do we want in flight? Our BDP, unless congestion cut cwnd. */ +static u32 bbr2_target_inflight(struct sock *sk) +{ + u32 bdp = bbr_inflight(sk, bbr_bw(sk), BBR_UNIT); + + return min(bdp, tcp_sk(sk)->snd_cwnd); +} + +static bool bbr2_is_probing_bandwidth(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + return (bbr->mode == BBR_STARTUP) || + (bbr->mode == BBR_PROBE_BW && + (bbr->cycle_idx == BBR_BW_PROBE_REFILL || + bbr->cycle_idx == BBR_BW_PROBE_UP)); +} + +/* Has the given amount of time elapsed since we marked the phase start? */ +static bool bbr2_has_elapsed_in_phase(const struct sock *sk, u32 interval_us) +{ + const struct tcp_sock *tp = tcp_sk(sk); + const struct bbr *bbr = inet_csk_ca(sk); + + return tcp_stamp_us_delta(tp->tcp_mstamp, + bbr->cycle_mstamp + interval_us) > 0; +} + +static void bbr2_handle_queue_too_high_in_startup(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr->full_bw_reached = 1; + bbr->inflight_hi = bbr_inflight(sk, bbr_max_bw(sk), BBR_UNIT); +} + +/* Exit STARTUP upon N consecutive rounds with ECN mark rate > ecn_thresh. */ +static void bbr2_check_ecn_too_high_in_startup(struct sock *sk, u32 ce_ratio) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr_full_bw_reached(sk) || !bbr->ecn_eligible || + !bbr->params.full_ecn_cnt || !bbr->params.ecn_thresh) + return; + + if (ce_ratio >= bbr->params.ecn_thresh) + bbr->startup_ecn_rounds++; + else + bbr->startup_ecn_rounds = 0; + + if (bbr->startup_ecn_rounds >= bbr->params.full_ecn_cnt) { + bbr->debug.event = 'E'; /* ECN caused STARTUP exit */ + bbr2_handle_queue_too_high_in_startup(sk); + return; + } +} + +static void bbr2_update_ecn_alpha(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + s32 delivered, delivered_ce; + u64 alpha, ce_ratio; + u32 gain; + + if (bbr->params.ecn_factor == 0) + return; + + delivered = tp->delivered - bbr->alpha_last_delivered; + delivered_ce = tp->delivered_ce - bbr->alpha_last_delivered_ce; + + if (delivered == 0 || /* avoid divide by zero */ + WARN_ON_ONCE(delivered < 0 || delivered_ce < 0)) /* backwards? */ + return; + + /* See if we should use ECN sender logic for this connection. */ + if (!bbr->ecn_eligible && bbr_ecn_enable && + (bbr->min_rtt_us <= bbr->params.ecn_max_rtt_us || + !bbr->params.ecn_max_rtt_us)) + bbr->ecn_eligible = 1; + + ce_ratio = (u64)delivered_ce << BBR_SCALE; + do_div(ce_ratio, delivered); + gain = bbr->params.ecn_alpha_gain; + alpha = ((BBR_UNIT - gain) * bbr->ecn_alpha) >> BBR_SCALE; + alpha += (gain * ce_ratio) >> BBR_SCALE; + bbr->ecn_alpha = min_t(u32, alpha, BBR_UNIT); + + bbr->alpha_last_delivered = tp->delivered; + bbr->alpha_last_delivered_ce = tp->delivered_ce; + + bbr2_check_ecn_too_high_in_startup(sk, ce_ratio); +} + +/* Each round trip of BBR_BW_PROBE_UP, double volume of probing data. */ +static void bbr2_raise_inflight_hi_slope(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u32 growth_this_round, cnt; + + /* Calculate "slope": packets S/Acked per inflight_hi increment. */ + growth_this_round = 1 << bbr->bw_probe_up_rounds; + bbr->bw_probe_up_rounds = min(bbr->bw_probe_up_rounds + 1, 30); + cnt = tp->snd_cwnd / growth_this_round; + cnt = max(cnt, 1U); + bbr->bw_probe_up_cnt = cnt; + bbr->debug.event = 'G'; /* Grow inflight_hi slope */ +} + +/* In BBR_BW_PROBE_UP, not seeing high loss/ECN/queue, so raise inflight_hi. */ +static void bbr2_probe_inflight_hi_upward(struct sock *sk, + const struct rate_sample *rs) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u32 delta; + + if (!tp->is_cwnd_limited || tp->snd_cwnd < bbr->inflight_hi) { + bbr->bw_probe_up_acks = 0; /* don't accmulate unused credits */ + return; /* not fully using inflight_hi, so don't grow it */ + } + + /* For each bw_probe_up_cnt packets ACKed, increase inflight_hi by 1. */ + bbr->bw_probe_up_acks += rs->acked_sacked; + if (bbr->bw_probe_up_acks >= bbr->bw_probe_up_cnt) { + delta = bbr->bw_probe_up_acks / bbr->bw_probe_up_cnt; + bbr->bw_probe_up_acks -= delta * bbr->bw_probe_up_cnt; + bbr->inflight_hi += delta; + bbr->debug.event = 'I'; /* Increment inflight_hi */ + } + + if (bbr->round_start) + bbr2_raise_inflight_hi_slope(sk); +} + +/* Does loss/ECN rate for this sample say inflight is "too high"? + * This is used by both the bbr_check_loss_too_high_in_startup() function, + * which can be used in either v1 or v2, and the PROBE_UP phase of v2, which + * uses it to notice when loss/ECN rates suggest inflight is too high. + */ +static bool bbr2_is_inflight_too_high(const struct sock *sk, + const struct rate_sample *rs) +{ + const struct bbr *bbr = inet_csk_ca(sk); + u32 loss_thresh, ecn_thresh; + + if (rs->lost > 0 && rs->tx_in_flight) { + loss_thresh = (u64)rs->tx_in_flight * bbr->params.loss_thresh >> + BBR_SCALE; + if (rs->lost > loss_thresh) + return true; + } + + if (rs->delivered_ce > 0 && rs->delivered > 0 && + bbr->ecn_eligible && bbr->params.ecn_thresh) { + ecn_thresh = (u64)rs->delivered * bbr->params.ecn_thresh >> + BBR_SCALE; + if (rs->delivered_ce >= ecn_thresh) + return true; + } + + return false; +} + +/* Calculate the tx_in_flight level that corresponded to excessive loss. + * We find "lost_prefix" segs of the skb where loss rate went too high, + * by solving for "lost_prefix" in the following equation: + * lost / inflight >= loss_thresh + * (lost_prev + lost_prefix) / (inflight_prev + lost_prefix) >= loss_thresh + * Then we take that equation, convert it to fixed point, and + * round up to the nearest packet. + */ +static u32 bbr2_inflight_hi_from_lost_skb(const struct sock *sk, + const struct rate_sample *rs, + const struct sk_buff *skb) +{ + const struct bbr *bbr = inet_csk_ca(sk); + u32 loss_thresh = bbr->params.loss_thresh; + u32 pcount, divisor, inflight_hi; + s32 inflight_prev, lost_prev; + u64 loss_budget, lost_prefix; + + pcount = tcp_skb_pcount(skb); + + /* How much data was in flight before this skb? */ + inflight_prev = rs->tx_in_flight - pcount; + if (WARN_ONCE(inflight_prev < 0, + "tx_in_flight: %u pcount: %u reneg: %u", + rs->tx_in_flight, pcount, tcp_sk(sk)->is_sack_reneg)) + return ~0U; + + /* How much inflight data was marked lost before this skb? */ + lost_prev = rs->lost - pcount; + if (WARN_ON_ONCE(lost_prev < 0)) + return ~0U; + + /* At what prefix of this lost skb did losss rate exceed loss_thresh? */ + loss_budget = (u64)inflight_prev * loss_thresh + BBR_UNIT - 1; + loss_budget >>= BBR_SCALE; + if (lost_prev >= loss_budget) { + lost_prefix = 0; /* previous losses crossed loss_thresh */ + } else { + lost_prefix = loss_budget - lost_prev; + lost_prefix <<= BBR_SCALE; + divisor = BBR_UNIT - loss_thresh; + if (WARN_ON_ONCE(!divisor)) /* loss_thresh is 8 bits */ + return ~0U; + do_div(lost_prefix, divisor); + } + + inflight_hi = inflight_prev + lost_prefix; + return inflight_hi; +} + +/* If loss/ECN rates during probing indicated we may have overfilled a + * buffer, return an operating point that tries to leave unutilized headroom in + * the path for other flows, for fairness convergence and lower RTTs and loss. + */ +static u32 bbr2_inflight_with_headroom(const struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 headroom, headroom_fraction; + + if (bbr->inflight_hi == ~0U) + return ~0U; + + headroom_fraction = bbr->params.inflight_headroom; + headroom = ((u64)bbr->inflight_hi * headroom_fraction) >> BBR_SCALE; + headroom = max(headroom, 1U); + return max_t(s32, bbr->inflight_hi - headroom, + bbr->params.cwnd_min_target); +} + +/* Bound cwnd to a sensible level, based on our current probing state + * machine phase and model of a good inflight level (inflight_lo, inflight_hi). + */ +static void bbr2_bound_cwnd_for_inflight_model(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u32 cap; + + /* tcp_rcv_synsent_state_process() currently calls tcp_ack() + * and thus cong_control() without first initializing us(!). + */ + if (!bbr->initialized) + return; + + cap = ~0U; + if (bbr->mode == BBR_PROBE_BW && + bbr->cycle_idx != BBR_BW_PROBE_CRUISE) { + /* Probe to see if more packets fit in the path. */ + cap = bbr->inflight_hi; + } else { + if (bbr->mode == BBR_PROBE_RTT || + (bbr->mode == BBR_PROBE_BW && + bbr->cycle_idx == BBR_BW_PROBE_CRUISE)) + cap = bbr2_inflight_with_headroom(sk); + } + /* Adapt to any loss/ECN since our last bw probe. */ + cap = min(cap, bbr->inflight_lo); + + cap = max_t(u32, cap, bbr->params.cwnd_min_target); + tp->snd_cwnd = min(cap, tp->snd_cwnd); +} + +/* Estimate a short-term lower bound on the capacity available now, based + * on measurements of the current delivery process and recent history. When we + * are seeing loss/ECN at times when we are not probing bw, then conservatively + * move toward flow balance by multiplicatively cutting our short-term + * estimated safe rate and volume of data (bw_lo and inflight_lo). We use a + * multiplicative decrease in order to converge to a lower capacity in time + * logarithmic in the magnitude of the decrease. + * + * However, we do not cut our short-term estimates lower than the current rate + * and volume of delivered data from this round trip, since from the current + * delivery process we can estimate the measured capacity available now. + * + * Anything faster than that approach would knowingly risk high loss, which can + * cause low bw for Reno/CUBIC and high loss recovery latency for + * request/response flows using any congestion control. + */ +static void bbr2_adapt_lower_bounds(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u32 ecn_cut, ecn_inflight_lo, beta; + + /* We only use lower-bound estimates when not probing bw. + * When probing we need to push inflight higher to probe bw. + */ + if (bbr2_is_probing_bandwidth(sk)) + return; + + /* ECN response. */ + if (bbr->ecn_in_round && bbr->ecn_eligible && bbr->params.ecn_factor) { + /* Reduce inflight to (1 - alpha*ecn_factor). */ + ecn_cut = (BBR_UNIT - + ((bbr->ecn_alpha * bbr->params.ecn_factor) >> + BBR_SCALE)); + if (bbr->inflight_lo == ~0U) + bbr->inflight_lo = tp->snd_cwnd; + ecn_inflight_lo = (u64)bbr->inflight_lo * ecn_cut >> BBR_SCALE; + } else { + ecn_inflight_lo = ~0U; + } + + /* Loss response. */ + if (bbr->loss_in_round) { + /* Reduce bw and inflight to (1 - beta). */ + if (bbr->bw_lo == ~0U) + bbr->bw_lo = bbr_max_bw(sk); + if (bbr->inflight_lo == ~0U) + bbr->inflight_lo = tp->snd_cwnd; + beta = bbr->params.beta; + bbr->bw_lo = + max_t(u32, bbr->bw_latest, + (u64)bbr->bw_lo * + (BBR_UNIT - beta) >> BBR_SCALE); + bbr->inflight_lo = + max_t(u32, bbr->inflight_latest, + (u64)bbr->inflight_lo * + (BBR_UNIT - beta) >> BBR_SCALE); + } + + /* Adjust to the lower of the levels implied by loss or ECN. */ + bbr->inflight_lo = min(bbr->inflight_lo, ecn_inflight_lo); +} + +/* Reset any short-term lower-bound adaptation to congestion, so that we can + * push our inflight up. + */ +static void bbr2_reset_lower_bounds(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr->bw_lo = ~0U; + bbr->inflight_lo = ~0U; +} + +/* After bw probing (STARTUP/PROBE_UP), reset signals before entering a state + * machine phase where we adapt our lower bound based on congestion signals. + */ +static void bbr2_reset_congestion_signals(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr->loss_in_round = 0; + bbr->ecn_in_round = 0; + bbr->loss_in_cycle = 0; + bbr->ecn_in_cycle = 0; + bbr->bw_latest = 0; + bbr->inflight_latest = 0; +} + +/* Update (most of) our congestion signals: track the recent rate and volume of + * delivered data, presence of loss, and EWMA degree of ECN marking. + */ +static void bbr2_update_congestion_signals( + struct sock *sk, const struct rate_sample *rs, struct bbr_context *ctx) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + u64 bw; + + bbr->loss_round_start = 0; + if (rs->interval_us <= 0 || !rs->acked_sacked) + return; /* Not a valid observation */ + bw = ctx->sample_bw; + + if (!rs->is_app_limited || bw >= bbr_max_bw(sk)) + bbr2_take_bw_hi_sample(sk, bw); + + bbr->loss_in_round |= (rs->losses > 0); + + /* Update rate and volume of delivered data from latest round trip: */ + bbr->bw_latest = max_t(u32, bbr->bw_latest, ctx->sample_bw); + bbr->inflight_latest = max_t(u32, bbr->inflight_latest, rs->delivered); + + if (before(rs->prior_delivered, bbr->loss_round_delivered)) + return; /* skip the per-round-trip updates */ + /* Now do per-round-trip updates. */ + bbr->loss_round_delivered = tp->delivered; /* mark round trip */ + bbr->loss_round_start = 1; + bbr2_adapt_lower_bounds(sk); + + /* Update windowed "latest" (single-round-trip) filters. */ + bbr->loss_in_round = 0; + bbr->ecn_in_round = 0; + bbr->bw_latest = ctx->sample_bw; + bbr->inflight_latest = rs->delivered; +} + +/* Bandwidth probing can cause loss. To help coexistence with loss-based + * congestion control we spread out our probing in a Reno-conscious way. Due to + * the shape of the Reno sawtooth, the time required between loss epochs for an + * idealized Reno flow is a number of round trips that is the BDP of that + * flow. We count packet-timed round trips directly, since measured RTT can + * vary widely, and Reno is driven by packet-timed round trips. + */ +static bool bbr2_is_reno_coexistence_probe_time(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 inflight, rounds, reno_gain, reno_rounds; + + /* Random loss can shave some small percentage off of our inflight + * in each round. To survive this, flows need robust periodic probes. + */ + rounds = bbr->params.bw_probe_max_rounds; + + reno_gain = bbr->params.bw_probe_reno_gain; + if (reno_gain) { + inflight = bbr2_target_inflight(sk); + reno_rounds = ((u64)inflight * reno_gain) >> BBR_SCALE; + rounds = min(rounds, reno_rounds); + } + return bbr->rounds_since_probe >= rounds; +} + +/* How long do we want to wait before probing for bandwidth (and risking + * loss)? We randomize the wait, for better mixing and fairness convergence. + * + * We bound the Reno-coexistence inter-bw-probe time to be 62-63 round trips. + * This is calculated to allow fairness with a 25Mbps, 30ms Reno flow, + * (eg 4K video to a broadband user): + * BDP = 25Mbps * .030sec /(1514bytes) = 61.9 packets + * + * We bound the BBR-native inter-bw-probe wall clock time to be: + * (a) higher than 2 sec: to try to avoid causing loss for a long enough time + * to allow Reno at 30ms to get 4K video bw, the inter-bw-probe time must + * be at least: 25Mbps * .030sec / (1514bytes) * 0.030sec = 1.9secs + * (b) lower than 3 sec: to ensure flows can start probing in a reasonable + * amount of time to discover unutilized bw on human-scale interactive + * time-scales (e.g. perhaps traffic from a web page download that we + * were competing with is now complete). + */ +static void bbr2_pick_probe_wait(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + /* Decide the random round-trip bound for wait until probe: */ + bbr->rounds_since_probe = + prandom_u32_max(bbr->params.bw_probe_rand_rounds); + /* Decide the random wall clock bound for wait until probe: */ + bbr->probe_wait_us = bbr->params.bw_probe_base_us + + prandom_u32_max(bbr->params.bw_probe_rand_us); +} + +static void bbr2_set_cycle_idx(struct sock *sk, int cycle_idx) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr->cycle_idx = cycle_idx; + /* New phase, so need to update cwnd and pacing rate. */ + bbr->try_fast_path = 0; +} + +/* Send at estimated bw to fill the pipe, but not queue. We need this phase + * before PROBE_UP, because as soon as we send faster than the available bw + * we will start building a queue, and if the buffer is shallow we can cause + * loss. If we do not fill the pipe before we cause this loss, our bw_hi and + * inflight_hi estimates will underestimate. + */ +static void bbr2_start_bw_probe_refill(struct sock *sk, u32 bw_probe_up_rounds) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr2_reset_lower_bounds(sk); + if (bbr->inflight_hi != ~0U) + bbr->inflight_hi += bbr->params.refill_add_inc; + bbr->bw_probe_up_rounds = bw_probe_up_rounds; + bbr->bw_probe_up_acks = 0; + bbr->stopped_risky_probe = 0; + bbr->ack_phase = BBR_ACKS_REFILLING; + bbr->next_rtt_delivered = tp->delivered; + bbr2_set_cycle_idx(sk, BBR_BW_PROBE_REFILL); +} + +/* Now probe max deliverable data rate and volume. */ +static void bbr2_start_bw_probe_up(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr->ack_phase = BBR_ACKS_PROBE_STARTING; + bbr->next_rtt_delivered = tp->delivered; + bbr->cycle_mstamp = tp->tcp_mstamp; + bbr2_set_cycle_idx(sk, BBR_BW_PROBE_UP); + bbr2_raise_inflight_hi_slope(sk); +} + +/* Start a new PROBE_BW probing cycle of some wall clock length. Pick a wall + * clock time at which to probe beyond an inflight that we think to be + * safe. This will knowingly risk packet loss, so we want to do this rarely, to + * keep packet loss rates low. Also start a round-trip counter, to probe faster + * if we estimate a Reno flow at our BDP would probe faster. + */ +static void bbr2_start_bw_probe_down(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr2_reset_congestion_signals(sk); + bbr->bw_probe_up_cnt = ~0U; /* not growing inflight_hi any more */ + bbr2_pick_probe_wait(sk); + bbr->cycle_mstamp = tp->tcp_mstamp; /* start wall clock */ + bbr->ack_phase = BBR_ACKS_PROBE_STOPPING; + bbr->next_rtt_delivered = tp->delivered; + bbr2_set_cycle_idx(sk, BBR_BW_PROBE_DOWN); +} + +/* Cruise: maintain what we estimate to be a neutral, conservative + * operating point, without attempting to probe up for bandwidth or down for + * RTT, and only reducing inflight in response to loss/ECN signals. + */ +static void bbr2_start_bw_probe_cruise(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr->inflight_lo != ~0U) + bbr->inflight_lo = min(bbr->inflight_lo, bbr->inflight_hi); + + bbr2_set_cycle_idx(sk, BBR_BW_PROBE_CRUISE); +} + +/* Loss and/or ECN rate is too high while probing. + * Adapt (once per bw probe) by cutting inflight_hi and then restarting cycle. + */ +static void bbr2_handle_inflight_too_high(struct sock *sk, + const struct rate_sample *rs) +{ + struct bbr *bbr = inet_csk_ca(sk); + const u32 beta = bbr->params.beta; + + bbr->prev_probe_too_high = 1; + bbr->bw_probe_samples = 0; /* only react once per probe */ + bbr->debug.event = 'L'; /* Loss/ECN too high */ + /* If we are app-limited then we are not robustly + * probing the max volume of inflight data we think + * might be safe (analogous to how app-limited bw + * samples are not known to be robustly probing bw). + */ + if (!rs->is_app_limited) + bbr->inflight_hi = max_t(u32, rs->tx_in_flight, + (u64)bbr2_target_inflight(sk) * + (BBR_UNIT - beta) >> BBR_SCALE); + if (bbr->mode == BBR_PROBE_BW && bbr->cycle_idx == BBR_BW_PROBE_UP) + bbr2_start_bw_probe_down(sk); +} + +/* If we're seeing bw and loss samples reflecting our bw probing, adapt + * using the signals we see. If loss or ECN mark rate gets too high, then adapt + * inflight_hi downward. If we're able to push inflight higher without such + * signals, push higher: adapt inflight_hi upward. + */ +static bool bbr2_adapt_upper_bounds(struct sock *sk, + const struct rate_sample *rs) +{ + struct bbr *bbr = inet_csk_ca(sk); + + /* Track when we'll see bw/loss samples resulting from our bw probes. */ + if (bbr->ack_phase == BBR_ACKS_PROBE_STARTING && bbr->round_start) + bbr->ack_phase = BBR_ACKS_PROBE_FEEDBACK; + if (bbr->ack_phase == BBR_ACKS_PROBE_STOPPING && bbr->round_start) { + /* End of samples from bw probing phase. */ + bbr->bw_probe_samples = 0; + bbr->ack_phase = BBR_ACKS_INIT; + /* At this point in the cycle, our current bw sample is also + * our best recent chance at finding the highest available bw + * for this flow. So now is the best time to forget the bw + * samples from the previous cycle, by advancing the window. + */ + if (bbr->mode == BBR_PROBE_BW && !rs->is_app_limited) + bbr2_advance_bw_hi_filter(sk); + /* If we had an inflight_hi, then probed and pushed inflight all + * the way up to hit that inflight_hi without seeing any + * high loss/ECN in all the resulting ACKs from that probing, + * then probe up again, this time letting inflight persist at + * inflight_hi for a round trip, then accelerating beyond. + */ + if (bbr->mode == BBR_PROBE_BW && + bbr->stopped_risky_probe && !bbr->prev_probe_too_high) { + bbr->debug.event = 'R'; /* reprobe */ + bbr2_start_bw_probe_refill(sk, 0); + return true; /* yes, decided state transition */ + } + } + + if (bbr2_is_inflight_too_high(sk, rs)) { + if (bbr->bw_probe_samples) /* sample is from bw probing? */ + bbr2_handle_inflight_too_high(sk, rs); + } else { + /* Loss/ECN rate is declared safe. Adjust upper bound upward. */ + if (bbr->inflight_hi == ~0U) /* no excess queue signals yet? */ + return false; + + /* To be resilient to random loss, we must raise inflight_hi + * if we observe in any phase that a higher level is safe. + */ + if (rs->tx_in_flight > bbr->inflight_hi) { + bbr->inflight_hi = rs->tx_in_flight; + bbr->debug.event = 'U'; /* raise up inflight_hi */ + } + + if (bbr->mode == BBR_PROBE_BW && + bbr->cycle_idx == BBR_BW_PROBE_UP) + bbr2_probe_inflight_hi_upward(sk, rs); + } + + return false; +} + +/* Check if it's time to probe for bandwidth now, and if so, kick it off. */ +static bool bbr2_check_time_to_probe_bw(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 n; + + /* If we seem to be at an operating point where we are not seeing loss + * but we are seeing ECN marks, then when the ECN marks cease we reprobe + * quickly (in case a burst of cross-traffic has ceased and freed up bw, + * or in case we are sharing with multiplicatively probing traffic). + */ + if (bbr->params.ecn_reprobe_gain && bbr->ecn_eligible && + bbr->ecn_in_cycle && !bbr->loss_in_cycle && + inet_csk(sk)->icsk_ca_state == TCP_CA_Open) { + bbr->debug.event = 'A'; /* *A*ll clear to probe *A*gain */ + /* Calculate n so that when bbr2_raise_inflight_hi_slope() + * computes growth_this_round as 2^n it will be roughly the + * desired volume of data (inflight_hi*ecn_reprobe_gain). + */ + n = ilog2((((u64)bbr->inflight_hi * + bbr->params.ecn_reprobe_gain) >> BBR_SCALE)); + bbr2_start_bw_probe_refill(sk, n); + return true; + } + + if (bbr2_has_elapsed_in_phase(sk, bbr->probe_wait_us) || + bbr2_is_reno_coexistence_probe_time(sk)) { + bbr2_start_bw_probe_refill(sk, 0); + return true; + } + return false; +} + +/* Is it time to transition from PROBE_DOWN to PROBE_CRUISE? */ +static bool bbr2_check_time_to_cruise(struct sock *sk, u32 inflight, u32 bw) +{ + struct bbr *bbr = inet_csk_ca(sk); + bool is_under_bdp, is_long_enough; + + /* Always need to pull inflight down to leave headroom in queue. */ + if (inflight > bbr2_inflight_with_headroom(sk)) + return false; + + is_under_bdp = inflight <= bbr_inflight(sk, bw, BBR_UNIT); + if (bbr->params.drain_to_target) + return is_under_bdp; + + is_long_enough = bbr2_has_elapsed_in_phase(sk, bbr->min_rtt_us); + return is_under_bdp || is_long_enough; +} + +/* PROBE_BW state machine: cruise, refill, probe for bw, or drain? */ +static void bbr2_update_cycle_phase(struct sock *sk, + const struct rate_sample *rs) +{ + struct bbr *bbr = inet_csk_ca(sk); + bool is_risky = false, is_queuing = false; + u32 inflight, bw; + + if (!bbr_full_bw_reached(sk)) + return; + + /* In DRAIN, PROBE_BW, or PROBE_RTT, adjust upper bounds. */ + if (bbr2_adapt_upper_bounds(sk, rs)) + return; /* already decided state transition */ + + if (bbr->mode != BBR_PROBE_BW) + return; + + inflight = bbr_packets_in_net_at_edt(sk, rs->prior_in_flight); + bw = bbr_max_bw(sk); + + switch (bbr->cycle_idx) { + /* First we spend most of our time cruising with a pacing_gain of 1.0, + * which paces at the estimated bw, to try to fully use the pipe + * without building queue. If we encounter loss/ECN marks, we adapt + * by slowing down. + */ + case BBR_BW_PROBE_CRUISE: + if (bbr2_check_time_to_probe_bw(sk)) + return; /* already decided state transition */ + break; + + /* After cruising, when it's time to probe, we first "refill": we send + * at the estimated bw to fill the pipe, before probing higher and + * knowingly risking overflowing the bottleneck buffer (causing loss). + */ + case BBR_BW_PROBE_REFILL: + if (bbr->round_start) { + /* After one full round trip of sending in REFILL, we + * start to see bw samples reflecting our REFILL, which + * may be putting too much data in flight. + */ + bbr->bw_probe_samples = 1; + bbr2_start_bw_probe_up(sk); + } + break; + + /* After we refill the pipe, we probe by using a pacing_gain > 1.0, to + * probe for bw. If we have not seen loss/ECN, we try to raise inflight + * to at least pacing_gain*BDP; note that this may take more than + * min_rtt if min_rtt is small (e.g. on a LAN). + * + * We terminate PROBE_UP bandwidth probing upon any of the following: + * + * (1) We've pushed inflight up to hit the inflight_hi target set in the + * most recent previous bw probe phase. Thus we want to start + * draining the queue immediately because it's very likely the most + * recently sent packets will fill the queue and cause drops. + * (checked here) + * (2) We have probed for at least 1*min_rtt_us, and the + * estimated queue is high enough (inflight > 1.25 * estimated_bdp). + * (checked here) + * (3) Loss filter says loss rate is "too high". + * (checked in bbr_is_inflight_too_high()) + * (4) ECN filter says ECN mark rate is "too high". + * (checked in bbr_is_inflight_too_high()) + */ + case BBR_BW_PROBE_UP: + if (bbr->prev_probe_too_high && + inflight >= bbr->inflight_hi) { + bbr->stopped_risky_probe = 1; + is_risky = true; + bbr->debug.event = 'D'; /* D for danger */ + } else if (bbr2_has_elapsed_in_phase(sk, bbr->min_rtt_us) && + inflight >= + bbr_inflight(sk, bw, + bbr->params.bw_probe_pif_gain)) { + is_queuing = true; + bbr->debug.event = 'Q'; /* building Queue */ + } + if (is_risky || is_queuing) { + bbr->prev_probe_too_high = 0; /* no loss/ECN (yet) */ + bbr2_start_bw_probe_down(sk); /* restart w/ down */ + } + break; + + /* After probing in PROBE_UP, we have usually accumulated some data in + * the bottleneck buffer (if bw probing didn't find more bw). We next + * enter PROBE_DOWN to try to drain any excess data from the queue. To + * do this, we use a pacing_gain < 1.0. We hold this pacing gain until + * our inflight is less then that target cruising point, which is the + * minimum of (a) the amount needed to leave headroom, and (b) the + * estimated BDP. Once inflight falls to match the target, we estimate + * the queue is drained; persisting would underutilize the pipe. + */ + case BBR_BW_PROBE_DOWN: + if (bbr2_check_time_to_probe_bw(sk)) + return; /* already decided state transition */ + if (bbr2_check_time_to_cruise(sk, inflight, bw)) + bbr2_start_bw_probe_cruise(sk); + break; + + default: + WARN_ONCE(1, "BBR invalid cycle index %u\n", bbr->cycle_idx); + } +} + +/* Exiting PROBE_RTT, so return to bandwidth probing in STARTUP or PROBE_BW. */ +static void bbr2_exit_probe_rtt(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr2_reset_lower_bounds(sk); + if (bbr_full_bw_reached(sk)) { + bbr->mode = BBR_PROBE_BW; + /* Raising inflight after PROBE_RTT may cause loss, so reset + * the PROBE_BW clock and schedule the next bandwidth probe for + * a friendly and randomized future point in time. + */ + bbr2_start_bw_probe_down(sk); + /* Since we are exiting PROBE_RTT, we know inflight is + * below our estimated BDP, so it is reasonable to cruise. + */ + bbr2_start_bw_probe_cruise(sk); + } else { + bbr->mode = BBR_STARTUP; + } +} + +/* Exit STARTUP based on loss rate > 1% and loss gaps in round >= N. Wait until + * the end of the round in recovery to get a good estimate of how many packets + * have been lost, and how many we need to drain with a low pacing rate. + */ +static void bbr2_check_loss_too_high_in_startup(struct sock *sk, + const struct rate_sample *rs) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr_full_bw_reached(sk)) + return; + + /* For STARTUP exit, check the loss rate at the end of each round trip + * of Recovery episodes in STARTUP. We check the loss rate at the end + * of the round trip to filter out noisy/low loss and have a better + * sense of inflight (extent of loss), so we can drain more accurately. + */ + if (rs->losses && bbr->loss_events_in_round < 0xf) + bbr->loss_events_in_round++; /* update saturating counter */ + if (bbr->params.full_loss_cnt && bbr->loss_round_start && + inet_csk(sk)->icsk_ca_state == TCP_CA_Recovery && + bbr->loss_events_in_round >= bbr->params.full_loss_cnt && + bbr2_is_inflight_too_high(sk, rs)) { + bbr->debug.event = 'P'; /* Packet loss caused STARTUP exit */ + bbr2_handle_queue_too_high_in_startup(sk); + return; + } + if (bbr->loss_round_start) + bbr->loss_events_in_round = 0; +} + +/* If we are done draining, advance into steady state operation in PROBE_BW. */ +static void bbr2_check_drain(struct sock *sk, const struct rate_sample *rs, + struct bbr_context *ctx) +{ + struct bbr *bbr = inet_csk_ca(sk); + + if (bbr_check_drain(sk, rs, ctx)) { + bbr->mode = BBR_PROBE_BW; + bbr2_start_bw_probe_down(sk); + } +} + +static void bbr2_update_model(struct sock *sk, const struct rate_sample *rs, + struct bbr_context *ctx) +{ + bbr2_update_congestion_signals(sk, rs, ctx); + bbr_update_ack_aggregation(sk, rs); + bbr2_check_loss_too_high_in_startup(sk, rs); + bbr_check_full_bw_reached(sk, rs); + bbr2_check_drain(sk, rs, ctx); + bbr2_update_cycle_phase(sk, rs); + bbr_update_min_rtt(sk, rs); +} + +/* Fast path for app-limited case. + * + * On each ack, we execute bbr state machine, which primarily consists of: + * 1) update model based on new rate sample, and + * 2) update control based on updated model or state change. + * + * There are certain workload/scenarios, e.g. app-limited case, where + * either we can skip updating model or we can skip update of both model + * as well as control. This provides signifcant softirq cpu savings for + * processing incoming acks. + * + * In case of app-limited, if there is no congestion (loss/ecn) and + * if observed bw sample is less than current estimated bw, then we can + * skip some of the computation in bbr state processing: + * + * - if there is no rtt/mode/phase change: In this case, since all the + * parameters of the network model are constant, we can skip model + * as well control update. + * + * - else we can skip rest of the model update. But we still need to + * update the control to account for the new rtt/mode/phase. + * + * Returns whether we can take fast path or not. + */ +static bool bbr2_fast_path(struct sock *sk, bool *update_model, + const struct rate_sample *rs, struct bbr_context *ctx) +{ + struct bbr *bbr = inet_csk_ca(sk); + u32 prev_min_rtt_us, prev_mode; + + if (bbr->params.fast_path && bbr->try_fast_path && + rs->is_app_limited && ctx->sample_bw < bbr_max_bw(sk) && + !bbr->loss_in_round && !bbr->ecn_in_round) { + prev_mode = bbr->mode; + prev_min_rtt_us = bbr->min_rtt_us; + bbr2_check_drain(sk, rs, ctx); + bbr2_update_cycle_phase(sk, rs); + bbr_update_min_rtt(sk, rs); + + if (bbr->mode == prev_mode && + bbr->min_rtt_us == prev_min_rtt_us && + bbr->try_fast_path) + return true; + + /* Skip model update, but control still needs to be updated */ + *update_model = false; + } + return false; +} + +static void bbr2_main(struct sock *sk, const struct rate_sample *rs) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + struct bbr_context ctx = { 0 }; + bool update_model = true; + u32 bw; + + bbr->debug.event = '.'; /* init to default NOP (no event yet) */ + + bbr_update_round_start(sk, rs, &ctx); + if (bbr->round_start) { + bbr->rounds_since_probe = + min_t(s32, bbr->rounds_since_probe + 1, 0xFF); + bbr2_update_ecn_alpha(sk); + } + + bbr->ecn_in_round |= rs->is_ece; + bbr_calculate_bw_sample(sk, rs, &ctx); + + if (bbr2_fast_path(sk, &update_model, rs, &ctx)) + goto out; + + if (update_model) + bbr2_update_model(sk, rs, &ctx); + + bbr_update_gains(sk); + bw = bbr_bw(sk); + bbr_set_pacing_rate(sk, bw, bbr->pacing_gain); + bbr_set_cwnd(sk, rs, rs->acked_sacked, bw, bbr->cwnd_gain, + tp->snd_cwnd, &ctx); + bbr2_bound_cwnd_for_inflight_model(sk); + +out: + bbr->prev_ca_state = inet_csk(sk)->icsk_ca_state; + bbr->loss_in_cycle |= rs->lost > 0; + bbr->ecn_in_cycle |= rs->delivered_ce > 0; + + bbr_debug(sk, rs->acked_sacked, rs, &ctx); +} + +/* Module parameters that are settable by TCP_CONGESTION_PARAMS are declared + * down here, so that the algorithm functions that use the parameters must use + * the per-socket parameters; if they accidentally use the global version + * then there will be a compile error. + * TODO(ncardwell): move all per-socket parameters down to this section. + */ + +/* On losses, scale down inflight and pacing rate by beta scaled by BBR_SCALE. + * No loss response when 0. Max allwed value is 255. + */ +static u32 bbr_beta = BBR_UNIT * 30 / 100; + +/* Gain factor for ECN mark ratio samples, scaled by BBR_SCALE. + * Max allowed value is 255. + */ +static u32 bbr_ecn_alpha_gain = BBR_UNIT * 1 / 16; /* 1/16 = 6.25% */ + +/* The initial value for the ecn_alpha state variable. Default and max + * BBR_UNIT (256), representing 1.0. This allows a flow to respond quickly + * to congestion if the bottleneck is congested when the flow starts up. + */ +static u32 bbr_ecn_alpha_init = BBR_UNIT; /* 1.0, to respond quickly */ + +/* On ECN, cut inflight_lo to (1 - ecn_factor * ecn_alpha) scaled by BBR_SCALE. + * No ECN based bounding when 0. Max allwed value is 255. + */ +static u32 bbr_ecn_factor = BBR_UNIT * 1 / 3; /* 1/3 = 33% */ + +/* Estimate bw probing has gone too far if CE ratio exceeds this threshold. + * Scaled by BBR_SCALE. Disabled when 0. Max allowed is 255. + */ +static u32 bbr_ecn_thresh = BBR_UNIT * 1 / 2; /* 1/2 = 50% */ + +/* Max RTT (in usec) at which to use sender-side ECN logic. + * Disabled when 0 (ECN allowed at any RTT). + * Max allowed for the parameter is 524287 (0x7ffff) us, ~524 ms. + */ +static u32 bbr_ecn_max_rtt_us = 5000; + +/* If non-zero, if in a cycle with no losses but some ECN marks, after ECN + * clears then use a multiplicative increase to quickly reprobe bw by + * starting inflight probing at the given multiple of inflight_hi. + * Default for this experimental knob is 0 (disabled). + * Planned value for experiments: BBR_UNIT * 1 / 2 = 128, representing 0.5. + */ +static u32 bbr_ecn_reprobe_gain; + +/* Estimate bw probing has gone too far if loss rate exceeds this level. */ +static u32 bbr_loss_thresh = BBR_UNIT * 2 / 100; /* 2% loss */ + +/* Exit STARTUP if number of loss marking events in a Recovery round is >= N, + * and loss rate is higher than bbr_loss_thresh. + * Disabled if 0. Max allowed value is 15 (0xF). + */ +static u32 bbr_full_loss_cnt = 8; + +/* Exit STARTUP if number of round trips with ECN mark rate above ecn_thresh + * meets this count. Max allowed value is 3. + */ +static u32 bbr_full_ecn_cnt = 2; + +/* Fraction of unutilized headroom to try to leave in path upon high loss. */ +static u32 bbr_inflight_headroom = BBR_UNIT * 15 / 100; + +/* Multiplier to get target inflight (as multiple of BDP) for PROBE_UP phase. + * Default is 1.25x, as in BBR v1. Max allowed is 511. + */ +static u32 bbr_bw_probe_pif_gain = BBR_UNIT * 5 / 4; + +/* Multiplier to get Reno-style probe epoch duration as: k * BDP round trips. + * If zero, disables this BBR v2 Reno-style BDP-scaled coexistence mechanism. + * Max allowed is 511. + */ +static u32 bbr_bw_probe_reno_gain = BBR_UNIT; + +/* Max number of packet-timed rounds to wait before probing for bandwidth. If + * we want to tolerate 1% random loss per round, and not have this cut our + * inflight too much, we must probe for bw periodically on roughly this scale. + * If low, limits Reno/CUBIC coexistence; if high, limits loss tolerance. + * We aim to be fair with Reno/CUBIC up to a BDP of at least: + * BDP = 25Mbps * .030sec /(1514bytes) = 61.9 packets + */ +static u32 bbr_bw_probe_max_rounds = 63; + +/* Max amount of randomness to inject in round counting for Reno-coexistence. + * Max value is 15. + */ +static u32 bbr_bw_probe_rand_rounds = 2; + +/* Use BBR-native probe time scale starting at this many usec. + * We aim to be fair with Reno/CUBIC up to an inter-loss time epoch of at least: + * BDP*RTT = 25Mbps * .030sec /(1514bytes) * 0.030sec = 1.9 secs + */ +static u32 bbr_bw_probe_base_us = 2 * USEC_PER_SEC; /* 2 secs */ + +/* Use BBR-native probes spread over this many usec: */ +static u32 bbr_bw_probe_rand_us = 1 * USEC_PER_SEC; /* 1 secs */ + +/* Undo the model changes made in loss recovery if recovery was spurious? */ +static bool bbr_undo = true; + +/* Use fast path if app-limited, no loss/ECN, and target cwnd was reached? */ +static bool bbr_fast_path = true; /* default: enabled */ + +/* Use fast ack mode ? */ +static int bbr_fast_ack_mode = 1; /* default: rwnd check off */ + +/* How much to additively increase inflight_hi when entering REFILL? */ +static u32 bbr_refill_add_inc; /* default: disabled */ + +module_param_named(beta, bbr_beta, uint, 0644); +module_param_named(ecn_alpha_gain, bbr_ecn_alpha_gain, uint, 0644); +module_param_named(ecn_alpha_init, bbr_ecn_alpha_init, uint, 0644); +module_param_named(ecn_factor, bbr_ecn_factor, uint, 0644); +module_param_named(ecn_thresh, bbr_ecn_thresh, uint, 0644); +module_param_named(ecn_max_rtt_us, bbr_ecn_max_rtt_us, uint, 0644); +module_param_named(ecn_reprobe_gain, bbr_ecn_reprobe_gain, uint, 0644); +module_param_named(loss_thresh, bbr_loss_thresh, uint, 0664); +module_param_named(full_loss_cnt, bbr_full_loss_cnt, uint, 0664); +module_param_named(full_ecn_cnt, bbr_full_ecn_cnt, uint, 0664); +module_param_named(inflight_headroom, bbr_inflight_headroom, uint, 0664); +module_param_named(bw_probe_pif_gain, bbr_bw_probe_pif_gain, uint, 0664); +module_param_named(bw_probe_reno_gain, bbr_bw_probe_reno_gain, uint, 0664); +module_param_named(bw_probe_max_rounds, bbr_bw_probe_max_rounds, uint, 0664); +module_param_named(bw_probe_rand_rounds, bbr_bw_probe_rand_rounds, uint, 0664); +module_param_named(bw_probe_base_us, bbr_bw_probe_base_us, uint, 0664); +module_param_named(bw_probe_rand_us, bbr_bw_probe_rand_us, uint, 0664); +module_param_named(undo, bbr_undo, bool, 0664); +module_param_named(fast_path, bbr_fast_path, bool, 0664); +module_param_named(fast_ack_mode, bbr_fast_ack_mode, uint, 0664); +module_param_named(refill_add_inc, bbr_refill_add_inc, uint, 0664); + +static void bbr2_init(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr_init(sk); /* run shared init code for v1 and v2 */ + + /* BBR v2 parameters: */ + bbr->params.beta = min_t(u32, 0xFFU, bbr_beta); + bbr->params.ecn_alpha_gain = min_t(u32, 0xFFU, bbr_ecn_alpha_gain); + bbr->params.ecn_alpha_init = min_t(u32, BBR_UNIT, bbr_ecn_alpha_init); + bbr->params.ecn_factor = min_t(u32, 0xFFU, bbr_ecn_factor); + bbr->params.ecn_thresh = min_t(u32, 0xFFU, bbr_ecn_thresh); + bbr->params.ecn_max_rtt_us = min_t(u32, 0x7ffffU, bbr_ecn_max_rtt_us); + bbr->params.ecn_reprobe_gain = min_t(u32, 0x1FF, bbr_ecn_reprobe_gain); + bbr->params.loss_thresh = min_t(u32, 0xFFU, bbr_loss_thresh); + bbr->params.full_loss_cnt = min_t(u32, 0xFU, bbr_full_loss_cnt); + bbr->params.full_ecn_cnt = min_t(u32, 0x3U, bbr_full_ecn_cnt); + bbr->params.inflight_headroom = + min_t(u32, 0xFFU, bbr_inflight_headroom); + bbr->params.bw_probe_pif_gain = + min_t(u32, 0x1FFU, bbr_bw_probe_pif_gain); + bbr->params.bw_probe_reno_gain = + min_t(u32, 0x1FFU, bbr_bw_probe_reno_gain); + bbr->params.bw_probe_max_rounds = + min_t(u32, 0xFFU, bbr_bw_probe_max_rounds); + bbr->params.bw_probe_rand_rounds = + min_t(u32, 0xFU, bbr_bw_probe_rand_rounds); + bbr->params.bw_probe_base_us = + min_t(u32, (1 << 26) - 1, bbr_bw_probe_base_us); + bbr->params.bw_probe_rand_us = + min_t(u32, (1 << 26) - 1, bbr_bw_probe_rand_us); + bbr->params.undo = bbr_undo; + bbr->params.fast_path = bbr_fast_path ? 1 : 0; + bbr->params.refill_add_inc = min_t(u32, 0x3U, bbr_refill_add_inc); + + /* BBR v2 state: */ + bbr->initialized = 1; + /* Start sampling ECN mark rate after first full flight is ACKed: */ + bbr->loss_round_delivered = tp->delivered + 1; + bbr->loss_round_start = 0; + bbr->undo_bw_lo = 0; + bbr->undo_inflight_lo = 0; + bbr->undo_inflight_hi = 0; + bbr->loss_events_in_round = 0; + bbr->startup_ecn_rounds = 0; + bbr2_reset_congestion_signals(sk); + bbr->bw_lo = ~0U; + bbr->bw_hi[0] = 0; + bbr->bw_hi[1] = 0; + bbr->inflight_lo = ~0U; + bbr->inflight_hi = ~0U; + bbr->bw_probe_up_cnt = ~0U; + bbr->bw_probe_up_acks = 0; + bbr->bw_probe_up_rounds = 0; + bbr->probe_wait_us = 0; + bbr->stopped_risky_probe = 0; + bbr->ack_phase = BBR_ACKS_INIT; + bbr->rounds_since_probe = 0; + bbr->bw_probe_samples = 0; + bbr->prev_probe_too_high = 0; + bbr->ecn_eligible = 0; + bbr->ecn_alpha = bbr->params.ecn_alpha_init; + bbr->alpha_last_delivered = 0; + bbr->alpha_last_delivered_ce = 0; + + tp->fast_ack_mode = min_t(u32, 0x2U, bbr_fast_ack_mode); +} + +/* Core TCP stack informs us that the given skb was just marked lost. */ +static void bbr2_skb_marked_lost(struct sock *sk, const struct sk_buff *skb) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + struct tcp_skb_cb *scb = TCP_SKB_CB(skb); + struct rate_sample rs; + + /* Capture "current" data over the full round trip of loss, + * to have a better chance to see the full capacity of the path. + */ + if (!bbr->loss_in_round) /* first loss in this round trip? */ + bbr->loss_round_delivered = tp->delivered; /* set round trip */ + bbr->loss_in_round = 1; + bbr->loss_in_cycle = 1; + + if (!bbr->bw_probe_samples) + return; /* not an skb sent while probing for bandwidth */ + if (unlikely(!scb->tx.delivered_mstamp)) + return; /* skb was SACKed, reneged, marked lost; ignore it */ + /* We are probing for bandwidth. Construct a rate sample that + * estimates what happened in the flight leading up to this lost skb, + * then see if the loss rate went too high, and if so at which packet. + */ + memset(&rs, 0, sizeof(rs)); + rs.tx_in_flight = scb->tx.in_flight; + rs.lost = tp->lost - scb->tx.lost; + rs.is_app_limited = scb->tx.is_app_limited; + if (bbr2_is_inflight_too_high(sk, &rs)) { + rs.tx_in_flight = bbr2_inflight_hi_from_lost_skb(sk, &rs, skb); + bbr2_handle_inflight_too_high(sk, &rs); + } +} + +/* Revert short-term model if current loss recovery event was spurious. */ +static u32 bbr2_undo_cwnd(struct sock *sk) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + bbr->debug.undo = 1; + bbr->full_bw = 0; /* spurious slow-down; reset full pipe detection */ + bbr->full_bw_cnt = 0; + bbr->loss_in_round = 0; + + if (!bbr->params.undo) + return tp->snd_cwnd; + + /* Revert to cwnd and other state saved before loss episode. */ + bbr->bw_lo = max(bbr->bw_lo, bbr->undo_bw_lo); + bbr->inflight_lo = max(bbr->inflight_lo, bbr->undo_inflight_lo); + bbr->inflight_hi = max(bbr->inflight_hi, bbr->undo_inflight_hi); + return bbr->prior_cwnd; +} + +/* Entering loss recovery, so save state for when we undo recovery. */ +static u32 bbr2_ssthresh(struct sock *sk) +{ + struct bbr *bbr = inet_csk_ca(sk); + + bbr_save_cwnd(sk); + /* For undo, save state that adapts based on loss signal. */ + bbr->undo_bw_lo = bbr->bw_lo; + bbr->undo_inflight_lo = bbr->inflight_lo; + bbr->undo_inflight_hi = bbr->inflight_hi; + return tcp_sk(sk)->snd_ssthresh; +} + +static enum tcp_bbr2_phase bbr2_get_phase(struct bbr *bbr) +{ + switch (bbr->mode) { + case BBR_STARTUP: + return BBR2_PHASE_STARTUP; + case BBR_DRAIN: + return BBR2_PHASE_DRAIN; + case BBR_PROBE_BW: + break; + case BBR_PROBE_RTT: + return BBR2_PHASE_PROBE_RTT; + default: + return BBR2_PHASE_INVALID; + } + switch (bbr->cycle_idx) { + case BBR_BW_PROBE_UP: + return BBR2_PHASE_PROBE_BW_UP; + case BBR_BW_PROBE_DOWN: + return BBR2_PHASE_PROBE_BW_DOWN; + case BBR_BW_PROBE_CRUISE: + return BBR2_PHASE_PROBE_BW_CRUISE; + case BBR_BW_PROBE_REFILL: + return BBR2_PHASE_PROBE_BW_REFILL; + default: + return BBR2_PHASE_INVALID; + } +} + +static size_t bbr2_get_info(struct sock *sk, u32 ext, int *attr, + union tcp_cc_info *info) +{ + if (ext & (1 << (INET_DIAG_BBRINFO - 1)) || + ext & (1 << (INET_DIAG_VEGASINFO - 1))) { + struct bbr *bbr = inet_csk_ca(sk); + u64 bw = bbr_bw_bytes_per_sec(sk, bbr_bw(sk)); + u64 bw_hi = bbr_bw_bytes_per_sec(sk, bbr_max_bw(sk)); + u64 bw_lo = bbr->bw_lo == ~0U ? + ~0ULL : bbr_bw_bytes_per_sec(sk, bbr->bw_lo); + + memset(&info->bbr2, 0, sizeof(info->bbr2)); + info->bbr2.bbr_bw_lsb = (u32)bw; + info->bbr2.bbr_bw_msb = (u32)(bw >> 32); + info->bbr2.bbr_min_rtt = bbr->min_rtt_us; + info->bbr2.bbr_pacing_gain = bbr->pacing_gain; + info->bbr2.bbr_cwnd_gain = bbr->cwnd_gain; + info->bbr2.bbr_bw_hi_lsb = (u32)bw_hi; + info->bbr2.bbr_bw_hi_msb = (u32)(bw_hi >> 32); + info->bbr2.bbr_bw_lo_lsb = (u32)bw_lo; + info->bbr2.bbr_bw_lo_msb = (u32)(bw_lo >> 32); + info->bbr2.bbr_mode = bbr->mode; + info->bbr2.bbr_phase = (__u8)bbr2_get_phase(bbr); + info->bbr2.bbr_version = (__u8)2; + info->bbr2.bbr_inflight_lo = bbr->inflight_lo; + info->bbr2.bbr_inflight_hi = bbr->inflight_hi; + info->bbr2.bbr_extra_acked = bbr_extra_acked(sk); + *attr = INET_DIAG_BBRINFO; + return sizeof(info->bbr2); + } + return 0; +} + +static void bbr2_set_state(struct sock *sk, u8 new_state) +{ + struct tcp_sock *tp = tcp_sk(sk); + struct bbr *bbr = inet_csk_ca(sk); + + if (new_state == TCP_CA_Loss) { + struct rate_sample rs = { .losses = 1 }; + struct bbr_context ctx = { 0 }; + + bbr->prev_ca_state = TCP_CA_Loss; + bbr->full_bw = 0; + if (!bbr2_is_probing_bandwidth(sk) && bbr->inflight_lo == ~0U) { + /* bbr_adapt_lower_bounds() needs cwnd before + * we suffered an RTO, to update inflight_lo: + */ + bbr->inflight_lo = + max(tp->snd_cwnd, bbr->prior_cwnd); + } + bbr_debug(sk, 0, &rs, &ctx); + } else if (bbr->prev_ca_state == TCP_CA_Loss && + new_state != TCP_CA_Loss) { + tp->snd_cwnd = max(tp->snd_cwnd, bbr->prior_cwnd); + bbr->try_fast_path = 0; /* bound cwnd using latest model */ + } +} + +static struct tcp_congestion_ops tcp_bbr2_cong_ops __read_mostly = { + .flags = TCP_CONG_NON_RESTRICTED | TCP_CONG_WANTS_CE_EVENTS, + .name = "bbr2", + .owner = THIS_MODULE, + .init = bbr2_init, + .cong_control = bbr2_main, + .sndbuf_expand = bbr_sndbuf_expand, + .skb_marked_lost = bbr2_skb_marked_lost, + .undo_cwnd = bbr2_undo_cwnd, + .cwnd_event = bbr_cwnd_event, + .ssthresh = bbr2_ssthresh, + .tso_segs = bbr_tso_segs, + .get_info = bbr2_get_info, + .set_state = bbr2_set_state, +}; + +static int __init bbr_register(void) +{ + BUILD_BUG_ON(sizeof(struct bbr) > ICSK_CA_PRIV_SIZE); + return tcp_register_congestion_control(&tcp_bbr2_cong_ops); +} + +static void __exit bbr_unregister(void) +{ + tcp_unregister_congestion_control(&tcp_bbr2_cong_ops); +} + +module_init(bbr_register); +module_exit(bbr_unregister); + +MODULE_AUTHOR("Van Jacobson "); +MODULE_AUTHOR("Neal Cardwell "); +MODULE_AUTHOR("Yuchung Cheng "); +MODULE_AUTHOR("Soheil Hassas Yeganeh "); +MODULE_AUTHOR("Priyaranjan Jha "); +MODULE_AUTHOR("Yousuk Seung "); +MODULE_AUTHOR("Kevin Yang "); +MODULE_AUTHOR("Arjun Roy "); + +MODULE_LICENSE("Dual BSD/GPL"); +MODULE_DESCRIPTION("TCP BBR (Bottleneck Bandwidth and RTT)"); diff --git a/net/ipv4/tcp_cong.c b/net/ipv4/tcp_cong.c index db5831e6c13..153ed9010c0 100644 --- a/net/ipv4/tcp_cong.c +++ b/net/ipv4/tcp_cong.c @@ -179,6 +179,7 @@ void tcp_init_congestion_control(struct sock *sk) struct inet_connection_sock *icsk = inet_csk(sk); tcp_sk(sk)->prior_ssthresh = 0; + tcp_sk(sk)->fast_ack_mode = 0; if (icsk->icsk_ca_ops->init) icsk->icsk_ca_ops->init(sk); if (tcp_ca_needs_ecn(sk)) diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 0f1b4bfddfd..a365f183bcc 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -348,7 +348,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb) tcp_enter_quickack_mode(sk, 2); break; case INET_ECN_CE: - if (tcp_ca_needs_ecn(sk)) + if (tcp_ca_wants_ce_events(sk)) tcp_ca_event(sk, CA_EVENT_ECN_IS_CE); if (!(tp->ecn_flags & TCP_ECN_DEMAND_CWR)) { @@ -359,7 +359,7 @@ static void __tcp_ecn_check_ce(struct sock *sk, const struct sk_buff *skb) tp->ecn_flags |= TCP_ECN_SEEN; break; default: - if (tcp_ca_needs_ecn(sk)) + if (tcp_ca_wants_ce_events(sk)) tcp_ca_event(sk, CA_EVENT_ECN_NO_CE); tp->ecn_flags |= TCP_ECN_SEEN; break; @@ -1039,7 +1039,12 @@ static void tcp_verify_retransmit_hint(struct tcp_sock *tp, struct sk_buff *skb) */ static void tcp_notify_skb_loss_event(struct tcp_sock *tp, const struct sk_buff *skb) { + struct sock *sk = (struct sock *)tp; + const struct tcp_congestion_ops *ca_ops = inet_csk(sk)->icsk_ca_ops; + tp->lost += tcp_skb_pcount(skb); + if (ca_ops->skb_marked_lost) + ca_ops->skb_marked_lost(sk, skb); } void tcp_mark_skb_lost(struct sock *sk, struct sk_buff *skb) @@ -1420,6 +1425,17 @@ static bool tcp_shifted_skb(struct sock *sk, struct sk_buff *prev, WARN_ON_ONCE(tcp_skb_pcount(skb) < pcount); tcp_skb_pcount_add(skb, -pcount); + /* Adjust tx.in_flight as pcount is shifted from skb to prev. */ + if (WARN_ONCE(TCP_SKB_CB(skb)->tx.in_flight < pcount, + "prev in_flight: %u skb in_flight: %u pcount: %u", + TCP_SKB_CB(prev)->tx.in_flight, + TCP_SKB_CB(skb)->tx.in_flight, + pcount)) + TCP_SKB_CB(skb)->tx.in_flight = 0; + else + TCP_SKB_CB(skb)->tx.in_flight -= pcount; + TCP_SKB_CB(prev)->tx.in_flight += pcount; + /* When we're adding to gso_segs == 1, gso_size will be zero, * in theory this shouldn't be necessary but as long as DSACK * code can come after this skb later on it's better to keep @@ -3189,7 +3205,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb, long seq_rtt_us = -1L; long ca_rtt_us = -1L; u32 pkts_acked = 0; - u32 last_in_flight = 0; bool rtt_update; int flag = 0; @@ -3225,7 +3240,6 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb, if (!first_ackt) first_ackt = last_ackt; - last_in_flight = TCP_SKB_CB(skb)->tx.in_flight; if (before(start_seq, reord)) reord = start_seq; if (!after(scb->end_seq, tp->high_seq)) @@ -3291,8 +3305,8 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb, seq_rtt_us = tcp_stamp_us_delta(tp->tcp_mstamp, first_ackt); ca_rtt_us = tcp_stamp_us_delta(tp->tcp_mstamp, last_ackt); - if (pkts_acked == 1 && last_in_flight < tp->mss_cache && - last_in_flight && !prior_sacked && fully_acked && + if (pkts_acked == 1 && fully_acked && !prior_sacked && + (tp->snd_una - prior_snd_una) < tp->mss_cache && sack->rate->prior_delivered + 1 == tp->delivered && !(flag & (FLAG_CA_ALERT | FLAG_SYN_ACKED))) { /* Conservatively mark a delayed ACK. It's typically @@ -3349,9 +3363,10 @@ static int tcp_clean_rtx_queue(struct sock *sk, const struct sk_buff *ack_skb, if (icsk->icsk_ca_ops->pkts_acked) { struct ack_sample sample = { .pkts_acked = pkts_acked, - .rtt_us = sack->rate->rtt_us, - .in_flight = last_in_flight }; + .rtt_us = sack->rate->rtt_us }; + sample.in_flight = tp->mss_cache * + (tp->delivered - sack->rate->prior_delivered); icsk->icsk_ca_ops->pkts_acked(sk, &sample); } @@ -3749,6 +3764,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) prior_fack = tcp_is_sack(tp) ? tcp_highest_sack_seq(tp) : tp->snd_una; rs.prior_in_flight = tcp_packets_in_flight(tp); + tcp_rate_check_app_limited(sk); /* ts_recent update must be made after we are sure that the packet * is in window. @@ -3846,6 +3862,7 @@ static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag) delivered = tcp_newly_delivered(sk, delivered, flag); lost = tp->lost - lost; /* freshly marked lost */ rs.is_ack_delayed = !!(flag & FLAG_ACK_MAYBE_DELAYED); + rs.is_ece = !!(flag & FLAG_ECE); tcp_rate_gen(sk, delivered, lost, is_sack_reneg, sack_state.rate); tcp_cong_control(sk, ack, delivered, flag, sack_state.rate); tcp_xmit_recovery(sk, rexmit); @@ -5414,13 +5431,14 @@ static void __tcp_ack_snd_check(struct sock *sk, int ofo_possible) /* More than one full frame received... */ if (((tp->rcv_nxt - tp->rcv_wup) > inet_csk(sk)->icsk_ack.rcv_mss && + (tp->fast_ack_mode == 1 || /* ... and right edge of window advances far enough. * (tcp_recvmsg() will send ACK otherwise). * If application uses SO_RCVLOWAT, we want send ack now if * we have not received enough bytes to satisfy the condition. */ - (tp->rcv_nxt - tp->copied_seq < sk->sk_rcvlowat || - __tcp_select_window(sk) >= tp->rcv_wnd)) || + (tp->rcv_nxt - tp->copied_seq < sk->sk_rcvlowat || + __tcp_select_window(sk) >= tp->rcv_wnd))) || /* We ACK each frame or... */ tcp_in_quickack_mode(sk) || /* Protocol state mandates a one-time immediate ACK */ diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 29553fce850..9d87dc854c8 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1256,8 +1256,6 @@ static int __tcp_transmit_skb(struct sock *sk, struct sk_buff *skb, tp->tcp_wstamp_ns = max(tp->tcp_wstamp_ns, tp->tcp_clock_cache); skb->skb_mstamp_ns = tp->tcp_wstamp_ns; if (clone_it) { - TCP_SKB_CB(skb)->tx.in_flight = TCP_SKB_CB(skb)->end_seq - - tp->snd_una; oskb = skb; tcp_skb_tsorted_save(oskb) { @@ -1536,7 +1534,7 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue, { struct tcp_sock *tp = tcp_sk(sk); struct sk_buff *buff; - int nsize, old_factor; + int nsize, old_factor, inflight_prev; long limit; int nlen; u8 flags; @@ -1615,6 +1613,15 @@ int tcp_fragment(struct sock *sk, enum tcp_queue tcp_queue, if (diff) tcp_adjust_pcount(sk, skb, diff); + + /* Set buff tx.in_flight as if buff were sent by itself. */ + inflight_prev = TCP_SKB_CB(skb)->tx.in_flight - old_factor; + if (WARN_ONCE(inflight_prev < 0, + "inconsistent: tx.in_flight: %u old_factor: %d", + TCP_SKB_CB(skb)->tx.in_flight, old_factor)) + inflight_prev = 0; + TCP_SKB_CB(buff)->tx.in_flight = inflight_prev + + tcp_skb_pcount(buff); } /* Link BUFF into the send queue. */ @@ -1983,13 +1990,12 @@ static u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now, static u32 tcp_tso_segs(struct sock *sk, unsigned int mss_now) { const struct tcp_congestion_ops *ca_ops = inet_csk(sk)->icsk_ca_ops; - u32 min_tso, tso_segs; - - min_tso = ca_ops->min_tso_segs ? - ca_ops->min_tso_segs(sk) : - sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs; + u32 tso_segs; - tso_segs = tcp_tso_autosize(sk, mss_now, min_tso); + tso_segs = ca_ops->tso_segs ? + ca_ops->tso_segs(sk, mss_now) : + tcp_tso_autosize(sk, mss_now, + sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs); return min_t(u32, tso_segs, sk->sk_gso_max_segs); } @@ -2629,6 +2635,7 @@ static bool tcp_write_xmit(struct sock *sk, unsigned int mss_now, int nonagle, skb->skb_mstamp_ns = tp->tcp_wstamp_ns = tp->tcp_clock_cache; list_move_tail(&skb->tcp_tsorted_anchor, &tp->tsorted_sent_queue); tcp_init_tso_segs(skb, mss_now); + tcp_set_tx_in_flight(sk, skb); goto repair; /* Skip network transmission */ } diff --git a/net/ipv4/tcp_rate.c b/net/ipv4/tcp_rate.c index 0de69356596..796fa6e5310 100644 --- a/net/ipv4/tcp_rate.c +++ b/net/ipv4/tcp_rate.c @@ -34,6 +34,24 @@ * ready to send in the write queue. */ +void tcp_set_tx_in_flight(struct sock *sk, struct sk_buff *skb) +{ + struct tcp_sock *tp = tcp_sk(sk); + u32 in_flight; + + /* Check, sanitize, and record packets in flight after skb was sent. */ + in_flight = tcp_packets_in_flight(tp) + tcp_skb_pcount(skb); + if (WARN_ONCE(in_flight > TCPCB_IN_FLIGHT_MAX, + "insane in_flight %u cc %s mss %u " + "cwnd %u pif %u %u %u %u\n", + in_flight, inet_csk(sk)->icsk_ca_ops->name, + tp->mss_cache, tp->snd_cwnd, + tp->packets_out, tp->retrans_out, + tp->sacked_out, tp->lost_out)) + in_flight = TCPCB_IN_FLIGHT_MAX; + TCP_SKB_CB(skb)->tx.in_flight = in_flight; +} + /* Snapshot the current delivery information in the skb, to generate * a rate sample later when the skb is (s)acked in tcp_rate_skb_delivered(). */ @@ -65,7 +83,10 @@ void tcp_rate_skb_sent(struct sock *sk, struct sk_buff *skb) TCP_SKB_CB(skb)->tx.first_tx_mstamp = tp->first_tx_mstamp; TCP_SKB_CB(skb)->tx.delivered_mstamp = tp->delivered_mstamp; TCP_SKB_CB(skb)->tx.delivered = tp->delivered; + TCP_SKB_CB(skb)->tx.delivered_ce = tp->delivered_ce; + TCP_SKB_CB(skb)->tx.lost = tp->lost; TCP_SKB_CB(skb)->tx.is_app_limited = tp->app_limited ? 1 : 0; + tcp_set_tx_in_flight(sk, skb); } /* When an skb is sacked or acked, we fill in the rate sample with the (prior) @@ -86,16 +107,20 @@ void tcp_rate_skb_delivered(struct sock *sk, struct sk_buff *skb, if (!rs->prior_delivered || after(scb->tx.delivered, rs->prior_delivered)) { + rs->prior_lost = scb->tx.lost; + rs->prior_delivered_ce = scb->tx.delivered_ce; rs->prior_delivered = scb->tx.delivered; rs->prior_mstamp = scb->tx.delivered_mstamp; rs->is_app_limited = scb->tx.is_app_limited; rs->is_retrans = scb->sacked & TCPCB_RETRANS; + rs->tx_in_flight = scb->tx.in_flight; /* Record send time of most recently ACKed packet: */ tp->first_tx_mstamp = tcp_skb_timestamp_us(skb); /* Find the duration of the "send phase" of this window: */ - rs->interval_us = tcp_stamp_us_delta(tp->first_tx_mstamp, - scb->tx.first_tx_mstamp); + rs->interval_us = tcp_stamp32_us_delta( + tp->first_tx_mstamp, + scb->tx.first_tx_mstamp); } /* Mark off the skb delivered once it's sacked to avoid being @@ -137,6 +162,11 @@ void tcp_rate_gen(struct sock *sk, u32 delivered, u32 lost, return; } rs->delivered = tp->delivered - rs->prior_delivered; + rs->lost = tp->lost - rs->prior_lost; + + rs->delivered_ce = tp->delivered_ce - rs->prior_delivered_ce; + /* delivered_ce occupies less than 32 bits in the skb control block */ + rs->delivered_ce &= TCPCB_DELIVERED_CE_MASK; /* Model sending data and receiving ACKs as separate pipeline phases * for a window. Usually the ACK phase is longer, but with ACK @@ -144,7 +174,7 @@ void tcp_rate_gen(struct sock *sk, u32 delivered, u32 lost, * longer phase. */ snd_us = rs->interval_us; /* send phase */ - ack_us = tcp_stamp_us_delta(tp->tcp_mstamp, + ack_us = tcp_stamp32_us_delta(tp->tcp_mstamp, rs->prior_mstamp); /* ack phase */ rs->interval_us = max(snd_us, ack_us); diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 4ef08079ccf..b5b24caa8ba 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -607,6 +607,7 @@ void tcp_write_timer_handler(struct sock *sk) goto out; } + tcp_rate_check_app_limited(sk); tcp_mstamp_refresh(tcp_sk(sk)); event = icsk->icsk_pending; diff --git a/scripts/headers_install.sh b/scripts/headers_install.sh index dd554bd436c..75030ad939a 100755 --- a/scripts/headers_install.sh +++ b/scripts/headers_install.sh @@ -89,6 +89,7 @@ include/uapi/linux/atmdev.h:CONFIG_COMPAT include/uapi/linux/eventpoll.h:CONFIG_PM_SLEEP include/uapi/linux/hw_breakpoint.h:CONFIG_HAVE_MIXED_BREAKPOINTS_REGS include/uapi/linux/pktcdvd.h:CONFIG_CDROM_PKTCDVD_WCACHE +include/uapi/linux/sched.h:CONFIG_SCHED_MUQSS " for c in $configs diff --git a/scripts/mkcompile_h b/scripts/mkcompile_h index a72b154de7b..fd1d53b7ae3 100755 --- a/scripts/mkcompile_h +++ b/scripts/mkcompile_h @@ -41,8 +41,8 @@ else fi UTS_VERSION="#$VERSION" -CONFIG_FLAGS="" -if [ -n "$SMP" ] ; then CONFIG_FLAGS="SMP"; fi +CONFIG_FLAGS="PCK" +if [ -n "$SMP" ] ; then CONFIG_FLAGS="$CONFIG_FLAGS SMP"; fi if [ -n "$PREEMPT" ] ; then CONFIG_FLAGS="$CONFIG_FLAGS PREEMPT"; fi if [ -n "$PREEMPT_RT" ] ; then CONFIG_FLAGS="$CONFIG_FLAGS PREEMPT_RT"; fi diff --git a/security/security.c b/security/security.c index b38155b2de8..386c2741886 100644 --- a/security/security.c +++ b/security/security.c @@ -1146,6 +1146,7 @@ int security_path_rmdir(const struct path *dir, struct dentry *dentry) return 0; return call_int_hook(path_rmdir, 0, dir, dentry); } +EXPORT_SYMBOL_GPL(security_path_rmdir); int security_path_unlink(const struct path *dir, struct dentry *dentry) { @@ -1162,6 +1163,7 @@ int security_path_symlink(const struct path *dir, struct dentry *dentry, return 0; return call_int_hook(path_symlink, 0, dir, dentry, old_name); } +EXPORT_SYMBOL_GPL(security_path_symlink); int security_path_link(struct dentry *old_dentry, const struct path *new_dir, struct dentry *new_dentry) @@ -1170,6 +1172,7 @@ int security_path_link(struct dentry *old_dentry, const struct path *new_dir, return 0; return call_int_hook(path_link, 0, old_dentry, new_dir, new_dentry); } +EXPORT_SYMBOL_GPL(security_path_link); int security_path_rename(const struct path *old_dir, struct dentry *old_dentry, const struct path *new_dir, struct dentry *new_dentry, @@ -1197,6 +1200,7 @@ int security_path_truncate(const struct path *path) return 0; return call_int_hook(path_truncate, 0, path); } +EXPORT_SYMBOL_GPL(security_path_truncate); int security_path_chmod(const struct path *path, umode_t mode) { @@ -1204,6 +1208,7 @@ int security_path_chmod(const struct path *path, umode_t mode) return 0; return call_int_hook(path_chmod, 0, path, mode); } +EXPORT_SYMBOL_GPL(security_path_chmod); int security_path_chown(const struct path *path, kuid_t uid, kgid_t gid) { @@ -1211,6 +1216,7 @@ int security_path_chown(const struct path *path, kuid_t uid, kgid_t gid) return 0; return call_int_hook(path_chown, 0, path, uid, gid); } +EXPORT_SYMBOL_GPL(security_path_chown); int security_path_chroot(const struct path *path) { @@ -1311,6 +1317,7 @@ int security_inode_permission(struct inode *inode, int mask) return 0; return call_int_hook(inode_permission, 0, inode, mask); } +EXPORT_SYMBOL_GPL(security_inode_permission); int security_inode_setattr(struct dentry *dentry, struct iattr *attr) { @@ -1508,6 +1515,7 @@ int security_file_permission(struct file *file, int mask) return fsnotify_perm(file, mask); } +EXPORT_SYMBOL_GPL(security_file_permission); int security_file_alloc(struct file *file) { diff --git a/sound/pci/maestro3.c b/sound/pci/maestro3.c index a0c91290324..8792e2a5c61 100644 --- a/sound/pci/maestro3.c +++ b/sound/pci/maestro3.c @@ -1989,7 +1989,7 @@ static void snd_m3_ac97_reset(struct snd_m3 *chip) outw(0, io + GPIO_DATA); outw(dir | GPO_PRIMARY_AC97, io + GPIO_DIRECTION); - schedule_timeout_uninterruptible(msecs_to_jiffies(delay1)); + schedule_msec_hrtimeout_uninterruptible((delay1)); outw(GPO_PRIMARY_AC97, io + GPIO_DATA); udelay(5); @@ -1997,7 +1997,7 @@ static void snd_m3_ac97_reset(struct snd_m3 *chip) outw(IO_SRAM_ENABLE | SERIAL_AC_LINK_ENABLE, io + RING_BUS_CTRL_A); outw(~0, io + GPIO_MASK); - schedule_timeout_uninterruptible(msecs_to_jiffies(delay2)); + schedule_msec_hrtimeout_uninterruptible((delay2)); if (! snd_m3_try_read_vendor(chip)) break; diff --git a/sound/soc/codecs/rt5631.c b/sound/soc/codecs/rt5631.c index 38356ea2bd6..8a8b4db5a8d 100644 --- a/sound/soc/codecs/rt5631.c +++ b/sound/soc/codecs/rt5631.c @@ -417,7 +417,7 @@ static void onebit_depop_mute_stage(struct snd_soc_component *component, int ena hp_zc = snd_soc_component_read(component, RT5631_INT_ST_IRQ_CTRL_2); snd_soc_component_write(component, RT5631_INT_ST_IRQ_CTRL_2, hp_zc & 0xf7ff); if (enable) { - schedule_timeout_uninterruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_uninterruptible((10)); /* config one-bit depop parameter */ rt5631_write_index(component, RT5631_SPK_INTL_CTRL, 0x307f); snd_soc_component_update_bits(component, RT5631_HP_OUT_VOL, @@ -529,7 +529,7 @@ static void depop_seq_mute_stage(struct snd_soc_component *component, int enable hp_zc = snd_soc_component_read(component, RT5631_INT_ST_IRQ_CTRL_2); snd_soc_component_write(component, RT5631_INT_ST_IRQ_CTRL_2, hp_zc & 0xf7ff); if (enable) { - schedule_timeout_uninterruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_uninterruptible((10)); /* config depop sequence parameter */ rt5631_write_index(component, RT5631_SPK_INTL_CTRL, 0x302f); diff --git a/sound/soc/codecs/wm8350.c b/sound/soc/codecs/wm8350.c index 15d42ce3b21..897fced9589 100644 --- a/sound/soc/codecs/wm8350.c +++ b/sound/soc/codecs/wm8350.c @@ -234,10 +234,10 @@ static void wm8350_pga_work(struct work_struct *work) out2->ramp == WM8350_RAMP_UP) { /* delay is longer over 0dB as increases are larger */ if (i >= WM8350_OUTn_0dB) - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (2)); else - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (1)); } else udelay(50); /* doesn't matter if we delay longer */ @@ -1121,7 +1121,7 @@ static int wm8350_set_bias_level(struct snd_soc_component *component, (platform->dis_out4 << 6)); /* wait for discharge */ - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (platform-> cap_discharge_msecs)); @@ -1137,7 +1137,7 @@ static int wm8350_set_bias_level(struct snd_soc_component *component, WM8350_VBUFEN); /* wait for vmid */ - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (platform-> vmid_charge_msecs)); @@ -1188,7 +1188,7 @@ static int wm8350_set_bias_level(struct snd_soc_component *component, wm8350_reg_write(wm8350, WM8350_POWER_MGMT_1, pm1); /* wait */ - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (platform-> vmid_discharge_msecs)); @@ -1206,7 +1206,7 @@ static int wm8350_set_bias_level(struct snd_soc_component *component, pm1 | WM8350_OUTPUT_DRAIN_EN); /* wait */ - schedule_timeout_interruptible(msecs_to_jiffies + schedule_msec_hrtimeout_interruptible( (platform->drain_msecs)); pm1 &= ~WM8350_BIASEN; diff --git a/sound/soc/codecs/wm8900.c b/sound/soc/codecs/wm8900.c index a9a6d766a17..45bf31de628 100644 --- a/sound/soc/codecs/wm8900.c +++ b/sound/soc/codecs/wm8900.c @@ -1104,7 +1104,7 @@ static int wm8900_set_bias_level(struct snd_soc_component *component, /* Need to let things settle before stopping the clock * to ensure that restart works, see "Stopping the * master clock" in the datasheet. */ - schedule_timeout_interruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_interruptible(1); snd_soc_component_write(component, WM8900_REG_POWER2, WM8900_REG_POWER2_SYSCLK_ENA); break; diff --git a/sound/soc/codecs/wm9713.c b/sound/soc/codecs/wm9713.c index e0ce32dd4a8..eb91c0282aa 100644 --- a/sound/soc/codecs/wm9713.c +++ b/sound/soc/codecs/wm9713.c @@ -199,7 +199,7 @@ static int wm9713_voice_shutdown(struct snd_soc_dapm_widget *w, /* Gracefully shut down the voice interface. */ snd_soc_component_update_bits(component, AC97_HANDSET_RATE, 0x0f00, 0x0200); - schedule_timeout_interruptible(msecs_to_jiffies(1)); + schedule_msec_hrtimeout_interruptible(1); snd_soc_component_update_bits(component, AC97_HANDSET_RATE, 0x0f00, 0x0f00); snd_soc_component_update_bits(component, AC97_EXTENDED_MID, 0x1000, 0x1000); @@ -868,7 +868,7 @@ static int wm9713_set_pll(struct snd_soc_component *component, wm9713->pll_in = freq_in; /* wait 10ms AC97 link frames for the link to stabilise */ - schedule_timeout_interruptible(msecs_to_jiffies(10)); + schedule_msec_hrtimeout_interruptible((10)); return 0; } diff --git a/sound/soc/soc-dapm.c b/sound/soc/soc-dapm.c index 91bf939d523..1bf7118d96a 100644 --- a/sound/soc/soc-dapm.c +++ b/sound/soc/soc-dapm.c @@ -154,7 +154,7 @@ static void dapm_assert_locked(struct snd_soc_dapm_context *dapm) static void pop_wait(u32 pop_time) { if (pop_time) - schedule_timeout_uninterruptible(msecs_to_jiffies(pop_time)); + schedule_msec_hrtimeout_uninterruptible((pop_time)); } __printf(3, 4) diff --git a/sound/usb/card.h b/sound/usb/card.h index a741e7da83a..866319a5744 100644 --- a/sound/usb/card.h +++ b/sound/usb/card.h @@ -54,6 +54,7 @@ struct snd_urb_ctx { struct snd_usb_endpoint *ep; int index; /* index for urb array */ int packets; /* number of packets per urb */ + int queued; /* queued data bytes by this urb */ int packet_size[MAX_PACKS_HS]; /* size of packets for next submission */ struct list_head ready_list; }; @@ -157,9 +158,12 @@ struct snd_usb_substream { unsigned int stream_offset_adj; /* Bytes to drop from beginning of stream (for non-compliant devices) */ unsigned int running: 1; /* running status */ + unsigned int period_elapsed_pending; /* delay period handling */ + unsigned int buffer_bytes; /* buffer size in bytes */ + unsigned int inflight_bytes; /* in-flight data bytes on buffer (for playback) */ unsigned int hwptr_done; /* processed byte position in the buffer */ - unsigned int transfer_done; /* processed frames since last period update */ + unsigned int transfer_done; /* processed frames since last period update */ unsigned int frame_limit; /* limits number of packets in URB */ /* data and sync endpoints for this stream */ @@ -174,8 +178,7 @@ struct snd_usb_substream { struct list_head fmt_list; /* format list */ spinlock_t lock; - int last_frame_number; /* stored frame number */ - int last_delay; /* stored delay */ + unsigned int last_frame_number; /* stored frame number */ struct { int marker; diff --git a/sound/usb/endpoint.c b/sound/usb/endpoint.c index 014c4386282..da649211bff 100644 --- a/sound/usb/endpoint.c +++ b/sound/usb/endpoint.c @@ -275,6 +275,7 @@ static void prepare_silent_urb(struct snd_usb_endpoint *ep, urb->number_of_packets = ctx->packets; urb->transfer_buffer_length = offs * ep->stride + ctx->packets * extra; + ctx->queued = 0; } /* diff --git a/sound/usb/line6/pcm.c b/sound/usb/line6/pcm.c index fdbdfb7bce9..fa8e8faf3eb 100644 --- a/sound/usb/line6/pcm.c +++ b/sound/usb/line6/pcm.c @@ -127,7 +127,7 @@ static void line6_wait_clear_audio_urbs(struct snd_line6_pcm *line6pcm, if (!alive) break; set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(1); + schedule_min_hrtimeout(); } while (--timeout > 0); if (alive) dev_err(line6pcm->line6->ifcdev, diff --git a/sound/usb/pcm.c b/sound/usb/pcm.c index e5311b6bb3f..4e5031a6806 100644 --- a/sound/usb/pcm.c +++ b/sound/usb/pcm.c @@ -29,15 +29,21 @@ #define SUBSTREAM_FLAG_SYNC_EP_STARTED 1 /* return the estimated delay based on USB frame counters */ -snd_pcm_uframes_t snd_usb_pcm_delay(struct snd_usb_substream *subs, - unsigned int rate) +static snd_pcm_uframes_t snd_usb_pcm_delay(struct snd_usb_substream *subs, + struct snd_pcm_runtime *runtime) { - int current_frame_number; - int frame_diff; + unsigned int current_frame_number; + unsigned int frame_diff; int est_delay; + int queued; - if (!subs->last_delay) - return 0; /* short path */ + if (subs->direction == SNDRV_PCM_STREAM_PLAYBACK) { + queued = bytes_to_frames(runtime, subs->inflight_bytes); + if (!queued) + return 0; + } else if (!subs->running) { + return 0; + } current_frame_number = usb_get_current_frame_number(subs->dev); /* @@ -49,14 +55,14 @@ snd_pcm_uframes_t snd_usb_pcm_delay(struct snd_usb_substream *subs, /* Approximation based on number of samples per USB frame (ms), some truncation for 44.1 but the estimate is good enough */ - est_delay = frame_diff * rate / 1000; - if (subs->direction == SNDRV_PCM_STREAM_PLAYBACK) - est_delay = subs->last_delay - est_delay; - else - est_delay = subs->last_delay + est_delay; + est_delay = frame_diff * runtime->rate / 1000; + + if (subs->direction == SNDRV_PCM_STREAM_PLAYBACK) { + est_delay = queued - est_delay; + if (est_delay < 0) + est_delay = 0; + } - if (est_delay < 0) - est_delay = 0; return est_delay; } @@ -65,17 +71,17 @@ snd_pcm_uframes_t snd_usb_pcm_delay(struct snd_usb_substream *subs, */ static snd_pcm_uframes_t snd_usb_pcm_pointer(struct snd_pcm_substream *substream) { - struct snd_usb_substream *subs = substream->runtime->private_data; + struct snd_pcm_runtime *runtime = substream->runtime; + struct snd_usb_substream *subs = runtime->private_data; unsigned int hwptr_done; if (atomic_read(&subs->stream->chip->shutdown)) return SNDRV_PCM_POS_XRUN; spin_lock(&subs->lock); hwptr_done = subs->hwptr_done; - substream->runtime->delay = snd_usb_pcm_delay(subs, - substream->runtime->rate); + runtime->delay = snd_usb_pcm_delay(subs, runtime); spin_unlock(&subs->lock); - return hwptr_done / (substream->runtime->frame_bits >> 3); + return bytes_to_frames(runtime, hwptr_done); } /* @@ -600,17 +606,14 @@ static int snd_usb_pcm_prepare(struct snd_pcm_substream *substream) goto unlock; /* reset the pointer */ + subs->buffer_bytes = frames_to_bytes(runtime, runtime->buffer_size); + subs->inflight_bytes = 0; subs->hwptr_done = 0; subs->transfer_done = 0; - subs->last_delay = 0; subs->last_frame_number = 0; + subs->period_elapsed_pending = 0; runtime->delay = 0; - /* for playback, submit the URBs now; otherwise, the first hwptr_done - * updates for all URBs would happen at the same time when starting */ - if (subs->direction == SNDRV_PCM_STREAM_PLAYBACK) - ret = start_endpoints(subs); - unlock: snd_usb_unlock_shutdown(chip); return ret; @@ -1147,28 +1150,23 @@ static void retire_capture_urb(struct snd_usb_substream *subs, spin_lock_irqsave(&subs->lock, flags); oldptr = subs->hwptr_done; subs->hwptr_done += bytes; - if (subs->hwptr_done >= runtime->buffer_size * stride) - subs->hwptr_done -= runtime->buffer_size * stride; + if (subs->hwptr_done >= subs->buffer_bytes) + subs->hwptr_done -= subs->buffer_bytes; frames = (bytes + (oldptr % stride)) / stride; subs->transfer_done += frames; if (subs->transfer_done >= runtime->period_size) { subs->transfer_done -= runtime->period_size; period_elapsed = 1; } - /* capture delay is by construction limited to one URB, - * reset delays here - */ - runtime->delay = subs->last_delay = 0; /* realign last_frame_number */ subs->last_frame_number = current_frame_number; - subs->last_frame_number &= 0xFF; /* keep 8 LSBs */ spin_unlock_irqrestore(&subs->lock, flags); /* copy a data chunk */ - if (oldptr + bytes > runtime->buffer_size * stride) { - unsigned int bytes1 = - runtime->buffer_size * stride - oldptr; + if (oldptr + bytes > subs->buffer_bytes) { + unsigned int bytes1 = subs->buffer_bytes - oldptr; + memcpy(runtime->dma_area + oldptr, cp, bytes1); memcpy(runtime->dma_area, cp + bytes1, bytes - bytes1); } else { @@ -1180,17 +1178,29 @@ static void retire_capture_urb(struct snd_usb_substream *subs, snd_pcm_period_elapsed(subs->pcm_substream); } +static void urb_ctx_queue_advance(struct snd_usb_substream *subs, + struct urb *urb, unsigned int bytes) +{ + struct snd_urb_ctx *ctx = urb->context; + + ctx->queued += bytes; + subs->inflight_bytes += bytes; + subs->hwptr_done += bytes; + if (subs->hwptr_done >= subs->buffer_bytes) + subs->hwptr_done -= subs->buffer_bytes; +} + static inline void fill_playback_urb_dsd_dop(struct snd_usb_substream *subs, struct urb *urb, unsigned int bytes) { struct snd_pcm_runtime *runtime = subs->pcm_substream->runtime; - unsigned int stride = runtime->frame_bits >> 3; unsigned int dst_idx = 0; unsigned int src_idx = subs->hwptr_done; - unsigned int wrap = runtime->buffer_size * stride; + unsigned int wrap = subs->buffer_bytes; u8 *dst = urb->transfer_buffer; u8 *src = runtime->dma_area; u8 marker[] = { 0x05, 0xfa }; + unsigned int queued = 0; /* * The DSP DOP format defines a way to transport DSD samples over @@ -1229,12 +1239,29 @@ static inline void fill_playback_urb_dsd_dop(struct snd_usb_substream *subs, dst[dst_idx++] = bitrev8(src[idx]); else dst[dst_idx++] = src[idx]; - - subs->hwptr_done++; + queued++; } } - if (subs->hwptr_done >= runtime->buffer_size * stride) - subs->hwptr_done -= runtime->buffer_size * stride; + + urb_ctx_queue_advance(subs, urb, queued); +} + +/* copy bit-reversed bytes onto transfer buffer */ +static void fill_playback_urb_dsd_bitrev(struct snd_usb_substream *subs, + struct urb *urb, unsigned int bytes) +{ + struct snd_pcm_runtime *runtime = subs->pcm_substream->runtime; + const u8 *src = runtime->dma_area; + u8 *buf = urb->transfer_buffer; + int i, ofs = subs->hwptr_done; + + for (i = 0; i < bytes; i++) { + *buf++ = bitrev8(src[ofs]); + if (++ofs >= subs->buffer_bytes) + ofs = 0; + } + + urb_ctx_queue_advance(subs, urb, bytes); } static void copy_to_urb(struct snd_usb_substream *subs, struct urb *urb, @@ -1242,10 +1269,10 @@ static void copy_to_urb(struct snd_usb_substream *subs, struct urb *urb, { struct snd_pcm_runtime *runtime = subs->pcm_substream->runtime; - if (subs->hwptr_done + bytes > runtime->buffer_size * stride) { + if (subs->hwptr_done + bytes > subs->buffer_bytes) { /* err, the transferred area goes over buffer boundary. */ - unsigned int bytes1 = - runtime->buffer_size * stride - subs->hwptr_done; + unsigned int bytes1 = subs->buffer_bytes - subs->hwptr_done; + memcpy(urb->transfer_buffer + offset, runtime->dma_area + subs->hwptr_done, bytes1); memcpy(urb->transfer_buffer + offset + bytes1, @@ -1254,9 +1281,8 @@ static void copy_to_urb(struct snd_usb_substream *subs, struct urb *urb, memcpy(urb->transfer_buffer + offset, runtime->dma_area + subs->hwptr_done, bytes); } - subs->hwptr_done += bytes; - if (subs->hwptr_done >= runtime->buffer_size * stride) - subs->hwptr_done -= runtime->buffer_size * stride; + + urb_ctx_queue_advance(subs, urb, bytes); } static unsigned int copy_to_urb_quirk(struct snd_usb_substream *subs, @@ -1295,17 +1321,18 @@ static void prepare_playback_urb(struct snd_usb_substream *subs, int i, stride, period_elapsed = 0; unsigned long flags; - stride = runtime->frame_bits >> 3; + stride = ep->stride; frames = 0; + ctx->queued = 0; urb->number_of_packets = 0; spin_lock_irqsave(&subs->lock, flags); subs->frame_limit += ep->max_urb_frames; for (i = 0; i < ctx->packets; i++) { counts = snd_usb_endpoint_next_packet_size(ep, ctx, i); /* set up descriptor */ - urb->iso_frame_desc[i].offset = frames * ep->stride; - urb->iso_frame_desc[i].length = counts * ep->stride; + urb->iso_frame_desc[i].offset = frames * stride; + urb->iso_frame_desc[i].length = counts * stride; frames += counts; urb->number_of_packets++; subs->transfer_done += counts; @@ -1320,14 +1347,14 @@ static void prepare_playback_urb(struct snd_usb_substream *subs, frames -= subs->transfer_done; counts -= subs->transfer_done; urb->iso_frame_desc[i].length = - counts * ep->stride; + counts * stride; subs->transfer_done = 0; } i++; if (i < ctx->packets) { /* add a transfer delimiter */ urb->iso_frame_desc[i].offset = - frames * ep->stride; + frames * stride; urb->iso_frame_desc[i].length = 0; urb->number_of_packets++; } @@ -1340,24 +1367,14 @@ static void prepare_playback_urb(struct snd_usb_substream *subs, !snd_usb_endpoint_implicit_feedback_sink(ep)) break; } - bytes = frames * ep->stride; + bytes = frames * stride; if (unlikely(ep->cur_format == SNDRV_PCM_FORMAT_DSD_U16_LE && subs->cur_audiofmt->dsd_dop)) { fill_playback_urb_dsd_dop(subs, urb, bytes); } else if (unlikely(ep->cur_format == SNDRV_PCM_FORMAT_DSD_U8 && subs->cur_audiofmt->dsd_bitrev)) { - /* bit-reverse the bytes */ - u8 *buf = urb->transfer_buffer; - for (i = 0; i < bytes; i++) { - int idx = (subs->hwptr_done + i) - % (runtime->buffer_size * stride); - buf[i] = bitrev8(runtime->dma_area[idx]); - } - - subs->hwptr_done += bytes; - if (subs->hwptr_done >= runtime->buffer_size * stride) - subs->hwptr_done -= runtime->buffer_size * stride; + fill_playback_urb_dsd_bitrev(subs, urb, bytes); } else { /* usual PCM */ if (!subs->tx_length_quirk) @@ -1367,14 +1384,7 @@ static void prepare_playback_urb(struct snd_usb_substream *subs, /* bytes is now amount of outgoing data */ } - /* update delay with exact number of samples queued */ - runtime->delay = subs->last_delay; - runtime->delay += frames; - subs->last_delay = runtime->delay; - - /* realign last_frame_number */ subs->last_frame_number = usb_get_current_frame_number(subs->dev); - subs->last_frame_number &= 0xFF; /* keep 8 LSBs */ if (subs->trigger_tstamp_pending_update) { /* this is the first actual URB submitted, @@ -1384,6 +1394,10 @@ static void prepare_playback_urb(struct snd_usb_substream *subs, subs->trigger_tstamp_pending_update = false; } + if (period_elapsed && !subs->running) { + subs->period_elapsed_pending = 1; + period_elapsed = 0; + } spin_unlock_irqrestore(&subs->lock, flags); urb->transfer_buffer_length = bytes; if (period_elapsed) @@ -1398,55 +1412,32 @@ static void retire_playback_urb(struct snd_usb_substream *subs, struct urb *urb) { unsigned long flags; - struct snd_pcm_runtime *runtime = subs->pcm_substream->runtime; - struct snd_usb_endpoint *ep = subs->data_endpoint; - int processed = urb->transfer_buffer_length / ep->stride; - int est_delay; - - /* ignore the delay accounting when processed=0 is given, i.e. - * silent payloads are processed before handling the actual data - */ - if (!processed) - return; + struct snd_urb_ctx *ctx = urb->context; + bool period_elapsed = false; spin_lock_irqsave(&subs->lock, flags); - if (!subs->last_delay) - goto out; /* short path */ - - est_delay = snd_usb_pcm_delay(subs, runtime->rate); - /* update delay with exact number of samples played */ - if (processed > subs->last_delay) - subs->last_delay = 0; - else - subs->last_delay -= processed; - runtime->delay = subs->last_delay; - - /* - * Report when delay estimate is off by more than 2ms. - * The error should be lower than 2ms since the estimate relies - * on two reads of a counter updated every ms. - */ - if (abs(est_delay - subs->last_delay) * 1000 > runtime->rate * 2) - dev_dbg_ratelimited(&subs->dev->dev, - "delay: estimated %d, actual %d\n", - est_delay, subs->last_delay); - - if (!subs->running) { - /* update last_frame_number for delay counting here since - * prepare_playback_urb won't be called during pause - */ - subs->last_frame_number = - usb_get_current_frame_number(subs->dev) & 0xff; + if (ctx->queued) { + if (subs->inflight_bytes >= ctx->queued) + subs->inflight_bytes -= ctx->queued; + else + subs->inflight_bytes = 0; } - out: + subs->last_frame_number = usb_get_current_frame_number(subs->dev); + if (subs->running) { + period_elapsed = subs->period_elapsed_pending; + subs->period_elapsed_pending = 0; + } spin_unlock_irqrestore(&subs->lock, flags); + if (period_elapsed) + snd_pcm_period_elapsed(subs->pcm_substream); } static int snd_usb_substream_playback_trigger(struct snd_pcm_substream *substream, int cmd) { struct snd_usb_substream *subs = substream->runtime->private_data; + int err; switch (cmd) { case SNDRV_PCM_TRIGGER_START: @@ -1457,6 +1448,14 @@ static int snd_usb_substream_playback_trigger(struct snd_pcm_substream *substrea prepare_playback_urb, retire_playback_urb, subs); + if (cmd == SNDRV_PCM_TRIGGER_START) { + err = start_endpoints(subs); + if (err < 0) { + snd_usb_endpoint_set_callback(subs->data_endpoint, + NULL, NULL, NULL); + return err; + } + } subs->running = 1; dev_dbg(&subs->dev->dev, "%d:%d Start Playback PCM\n", subs->cur_audiofmt->iface, @@ -1504,6 +1503,7 @@ static int snd_usb_substream_capture_trigger(struct snd_pcm_substream *substream snd_usb_endpoint_set_callback(subs->data_endpoint, NULL, retire_capture_urb, subs); + subs->last_frame_number = usb_get_current_frame_number(subs->dev); subs->running = 1; dev_dbg(&subs->dev->dev, "%d:%d Start Capture PCM\n", subs->cur_audiofmt->iface, diff --git a/sound/usb/pcm.h b/sound/usb/pcm.h index 06c586467d3..493a4e34d78 100644 --- a/sound/usb/pcm.h +++ b/sound/usb/pcm.h @@ -2,9 +2,6 @@ #ifndef __USBAUDIO_PCM_H #define __USBAUDIO_PCM_H -snd_pcm_uframes_t snd_usb_pcm_delay(struct snd_usb_substream *subs, - unsigned int rate); - void snd_usb_set_pcm_ops(struct snd_pcm *pcm, int stream); int snd_usb_pcm_suspend(struct snd_usb_stream *as); int snd_usb_pcm_resume(struct snd_usb_stream *as); diff --git a/tools/include/uapi/asm-generic/unistd.h b/tools/include/uapi/asm-generic/unistd.h index d2a942086fc..de966a1016d 100644 --- a/tools/include/uapi/asm-generic/unistd.h +++ b/tools/include/uapi/asm-generic/unistd.h @@ -872,8 +872,17 @@ __SYSCALL(__NR_landlock_add_rule, sys_landlock_add_rule) #define __NR_landlock_restrict_self 446 __SYSCALL(__NR_landlock_restrict_self, sys_landlock_restrict_self) +#define __NR_futex_wait 443 +__SYSCALL(__NR_futex_wait, sys_futex_wait) + +#define __NR_futex_wake 444 +__SYSCALL(__NR_futex_wake, sys_futex_wake) + +#define __NR_futex_waitv 445 +__SC_COMP(__NR_futex_waitv, sys_futex_waitv, compat_sys_futex_waitv) + #undef __NR_syscalls -#define __NR_syscalls 447 +#define __NR_syscalls 450 /* * 32 bit systems traditionally used different diff --git a/tools/perf/arch/x86/entry/syscalls/syscall_64.tbl b/tools/perf/arch/x86/entry/syscalls/syscall_64.tbl index ce18119ea0d..faa5a3442e4 100644 --- a/tools/perf/arch/x86/entry/syscalls/syscall_64.tbl +++ b/tools/perf/arch/x86/entry/syscalls/syscall_64.tbl @@ -368,6 +368,9 @@ 444 common landlock_create_ruleset sys_landlock_create_ruleset 445 common landlock_add_rule sys_landlock_add_rule 446 common landlock_restrict_self sys_landlock_restrict_self +447 common futex_wait sys_futex_wait +448 common futex_wake sys_futex_wake +449 common futex_waitv sys_futex_waitv # # Due to a historical design error, certain syscalls are numbered differently diff --git a/tools/testing/selftests/futex/functional/.gitignore b/tools/testing/selftests/futex/functional/.gitignore index 0efcd494daa..af7557e821d 100644 --- a/tools/testing/selftests/futex/functional/.gitignore +++ b/tools/testing/selftests/futex/functional/.gitignore @@ -6,3 +6,6 @@ futex_wait_private_mapped_file futex_wait_timeout futex_wait_uninitialized_heap futex_wait_wouldblock +futex2_wait +futex2_waitv +futex2_requeue diff --git a/tools/testing/selftests/futex/functional/Makefile b/tools/testing/selftests/futex/functional/Makefile index 23207829ec7..3ccb9ea58dd 100644 --- a/tools/testing/selftests/futex/functional/Makefile +++ b/tools/testing/selftests/futex/functional/Makefile @@ -1,10 +1,11 @@ # SPDX-License-Identifier: GPL-2.0 -INCLUDES := -I../include -I../../ +INCLUDES := -I../include -I../../ -I../../../../../usr/include/ CFLAGS := $(CFLAGS) -g -O2 -Wall -D_GNU_SOURCE -pthread $(INCLUDES) LDLIBS := -lpthread -lrt HEADERS := \ ../include/futextest.h \ + ../include/futex2test.h \ ../include/atomic.h \ ../include/logging.h TEST_GEN_FILES := \ @@ -14,7 +15,10 @@ TEST_GEN_FILES := \ futex_requeue_pi_signal_restart \ futex_requeue_pi_mismatched_ops \ futex_wait_uninitialized_heap \ - futex_wait_private_mapped_file + futex_wait_private_mapped_file \ + futex2_wait \ + futex2_waitv \ + futex2_requeue TEST_PROGS := run.sh diff --git a/tools/testing/selftests/futex/functional/futex2_requeue.c b/tools/testing/selftests/futex/functional/futex2_requeue.c new file mode 100644 index 00000000000..05629c2257d --- /dev/null +++ b/tools/testing/selftests/futex/functional/futex2_requeue.c @@ -0,0 +1,164 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/****************************************************************************** + * + * Copyright Collabora Ltd., 2021 + * + * DESCRIPTION + * Test requeue mechanism of futex2, using 32bit sized futexes. + * + * AUTHOR + * André Almeida + * + * HISTORY + * 2021-Feb-5: Initial version by André + * + *****************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "futex2test.h" +#include "logging.h" + +#define TEST_NAME "futex2-wait" +#define timeout_ns 30000000 +#define WAKE_WAIT_US 10000 +volatile futex_t *f1; + +void usage(char *prog) +{ + printf("Usage: %s\n", prog); + printf(" -c Use color\n"); + printf(" -h Display this help message\n"); + printf(" -v L Verbosity level: %d=QUIET %d=CRITICAL %d=INFO\n", + VQUIET, VCRITICAL, VINFO); +} + +void *waiterfn(void *arg) +{ + struct timespec64 to64; + + /* setting absolute timeout for futex2 */ + if (gettime64(CLOCK_MONOTONIC, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_nsec += timeout_ns; + + if (to64.tv_nsec >= 1000000000) { + to64.tv_sec++; + to64.tv_nsec -= 1000000000; + } + + if (futex2_wait(f1, *f1, FUTEX_32, &to64)) + printf("waiter failed errno %d\n", errno); + + return NULL; +} + +int main(int argc, char *argv[]) +{ + pthread_t waiter[10]; + int res, ret = RET_PASS; + int c, i; + volatile futex_t _f1 = 0; + volatile futex_t f2 = 0; + struct futex_requeue r1, r2; + + f1 = &_f1; + + r1.flags = FUTEX_32; + r2.flags = FUTEX_32; + + r1.uaddr = f1; + r2.uaddr = &f2; + + while ((c = getopt(argc, argv, "cht:v:")) != -1) { + switch (c) { + case 'c': + log_color(1); + break; + case 'h': + usage(basename(argv[0])); + exit(0); + case 'v': + log_verbosity(atoi(optarg)); + break; + default: + usage(basename(argv[0])); + exit(1); + } + } + + ksft_print_header(); + ksft_set_plan(2); + ksft_print_msg("%s: Test FUTEX2_REQUEUE\n", + basename(argv[0])); + + /* + * Requeue a waiter from f1 to f2, and wake f2. + */ + if (pthread_create(&waiter[0], NULL, waiterfn, NULL)) + error("pthread_create failed\n", errno); + + usleep(WAKE_WAIT_US); + + res = futex2_requeue(&r1, &r2, 0, 1, 0, 0); + if (res != 1) { + ksft_test_result_fail("futex2_requeue private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } + + + info("Calling private futex2_wake on f2: %u @ %p with val=%u\n", f2, &f2, f2); + res = futex2_wake(&f2, 1, FUTEX_32); + if (res != 1) { + ksft_test_result_fail("futex2_requeue private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_requeue simple succeeds\n"); + } + + + /* + * Create 10 waiters at f1. At futex_requeue, wake 3 and requeue 7. + * At futex_wake, wake INT_MAX (should be exaclty 7). + */ + for (i = 0; i < 10; i++) { + if (pthread_create(&waiter[i], NULL, waiterfn, NULL)) + error("pthread_create failed\n", errno); + } + + usleep(WAKE_WAIT_US); + + res = futex2_requeue(&r1, &r2, 3, 7, 0, 0); + if (res != 10) { + ksft_test_result_fail("futex2_requeue private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } + + res = futex2_wake(&f2, INT_MAX, FUTEX_32); + if (res != 7) { + ksft_test_result_fail("futex2_requeue private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_requeue succeeds\n"); + } + + ksft_print_cnts(); + return ret; +} diff --git a/tools/testing/selftests/futex/functional/futex2_wait.c b/tools/testing/selftests/futex/functional/futex2_wait.c new file mode 100644 index 00000000000..752a26b33bf --- /dev/null +++ b/tools/testing/selftests/futex/functional/futex2_wait.c @@ -0,0 +1,209 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/****************************************************************************** + * + * Copyright Collabora Ltd., 2021 + * + * DESCRIPTION + * Test wait/wake mechanism of futex2, using 32bit sized futexes. + * + * AUTHOR + * André Almeida + * + * HISTORY + * 2021-Feb-5: Initial version by André + * + *****************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "futex2test.h" +#include "logging.h" + +#define TEST_NAME "futex2-wait" +#define timeout_ns 30000000 +#define WAKE_WAIT_US 10000 +#define SHM_PATH "futex2_shm_file" +futex_t *f1; + +void usage(char *prog) +{ + printf("Usage: %s\n", prog); + printf(" -c Use color\n"); + printf(" -h Display this help message\n"); + printf(" -v L Verbosity level: %d=QUIET %d=CRITICAL %d=INFO\n", + VQUIET, VCRITICAL, VINFO); +} + +void *waiterfn(void *arg) +{ + struct timespec64 to64; + unsigned int flags = 0; + + if (arg) + flags = *((unsigned int *) arg); + + /* setting absolute timeout for futex2 */ + if (gettime64(CLOCK_MONOTONIC, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_nsec += timeout_ns; + + if (to64.tv_nsec >= 1000000000) { + to64.tv_sec++; + to64.tv_nsec -= 1000000000; + } + + if (futex2_wait(f1, *f1, FUTEX_32 | flags, &to64)) + printf("waiter failed errno %d\n", errno); + + return NULL; +} + +void *waitershm(void *arg) +{ + futex2_wait(arg, 0, FUTEX_32 | FUTEX_SHARED_FLAG, NULL); + + return NULL; +} + +int main(int argc, char *argv[]) +{ + pthread_t waiter; + unsigned int flags = FUTEX_SHARED_FLAG; + int res, ret = RET_PASS; + int c; + futex_t f_private = 0; + + f1 = &f_private; + + while ((c = getopt(argc, argv, "cht:v:")) != -1) { + switch (c) { + case 'c': + log_color(1); + break; + case 'h': + usage(basename(argv[0])); + exit(0); + case 'v': + log_verbosity(atoi(optarg)); + break; + default: + usage(basename(argv[0])); + exit(1); + } + } + + ksft_print_header(); + ksft_set_plan(3); + ksft_print_msg("%s: Test FUTEX2_WAIT\n", + basename(argv[0])); + + /* Testing a private futex */ + info("Calling private futex2_wait on f1: %u @ %p with val=%u\n", *f1, f1, *f1); + + if (pthread_create(&waiter, NULL, waiterfn, NULL)) + error("pthread_create failed\n", errno); + + usleep(WAKE_WAIT_US); + + info("Calling private futex2_wake on f1: %u @ %p with val=%u\n", *f1, f1, *f1); + res = futex2_wake(f1, 1, FUTEX_32); + if (res != 1) { + ksft_test_result_fail("futex2_wake private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wake private succeeds\n"); + } + + int shm_id = shmget(IPC_PRIVATE, 4096, IPC_CREAT | 0666); + + if (shm_id < 0) { + perror("shmget"); + exit(1); + } + + /* Testing an anon page shared memory */ + unsigned int *shared_data = shmat(shm_id, NULL, 0); + + *shared_data = 0; + f1 = shared_data; + + info("Calling shared futex2_wait on f1: %u @ %p with val=%u\n", *f1, f1, *f1); + + if (pthread_create(&waiter, NULL, waiterfn, &flags)) + error("pthread_create failed\n", errno); + + usleep(WAKE_WAIT_US); + + info("Calling shared futex2_wake on f1: %u @ %p with val=%u\n", *f1, f1, *f1); + res = futex2_wake(f1, 1, FUTEX_32 | FUTEX_SHARED_FLAG); + if (res != 1) { + ksft_test_result_fail("futex2_wake shared (shmget) returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wake shared (shmget) succeeds\n"); + } + + shmdt(shared_data); + + /* Testing a file backed shared memory */ + void *shm; + int fd, pid; + + f_private = 0; + + fd = open(SHM_PATH, O_RDWR | O_CREAT, S_IRUSR | S_IWUSR); + if (fd < 0) { + perror("open"); + exit(1); + } + + res = ftruncate(fd, sizeof(f_private)); + if (res) { + perror("ftruncate"); + exit(1); + } + + shm = mmap(NULL, sizeof(f_private), PROT_READ | PROT_WRITE, MAP_SHARED, fd, 0); + if (shm == MAP_FAILED) { + perror("mmap"); + exit(1); + } + + memcpy(shm, &f_private, sizeof(f_private)); + + pthread_create(&waiter, NULL, waitershm, shm); + + usleep(WAKE_WAIT_US); + + res = futex2_wake(shm, 1, FUTEX_32 | FUTEX_SHARED_FLAG); + if (res != 1) { + ksft_test_result_fail("futex2_wake shared (mmap) returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wake shared (mmap) succeeds\n"); + } + + munmap(shm, sizeof(f_private)); + + remove(SHM_PATH); + + ksft_print_cnts(); + return ret; +} diff --git a/tools/testing/selftests/futex/functional/futex2_waitv.c b/tools/testing/selftests/futex/functional/futex2_waitv.c new file mode 100644 index 00000000000..8ba74f1cbd5 --- /dev/null +++ b/tools/testing/selftests/futex/functional/futex2_waitv.c @@ -0,0 +1,157 @@ +// SPDX-License-Identifier: GPL-2.0-or-later +/****************************************************************************** + * + * Copyright Collabora Ltd., 2021 + * + * DESCRIPTION + * Test waitv/wake mechanism of futex2, using 32bit sized futexes. + * + * AUTHOR + * André Almeida + * + * HISTORY + * 2021-Feb-5: Initial version by André + * + *****************************************************************************/ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include "futex2test.h" +#include "logging.h" + +#define TEST_NAME "futex2-wait" +#define timeout_ns 1000000000 +#define WAKE_WAIT_US 10000 +#define NR_FUTEXES 30 +struct futex_waitv waitv[NR_FUTEXES]; +u_int32_t futexes[NR_FUTEXES] = {0}; + +void usage(char *prog) +{ + printf("Usage: %s\n", prog); + printf(" -c Use color\n"); + printf(" -h Display this help message\n"); + printf(" -v L Verbosity level: %d=QUIET %d=CRITICAL %d=INFO\n", + VQUIET, VCRITICAL, VINFO); +} + +void *waiterfn(void *arg) +{ + struct timespec64 to64; + int res; + + /* setting absolute timeout for futex2 */ + if (gettime64(CLOCK_MONOTONIC, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_sec++; + + res = futex2_waitv(waitv, NR_FUTEXES, 0, &to64); + if (res < 0) { + ksft_test_result_fail("futex2_waitv private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + } else if (res != NR_FUTEXES - 1) { + ksft_test_result_fail("futex2_waitv private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + } + + return NULL; +} + +int main(int argc, char *argv[]) +{ + pthread_t waiter; + int res, ret = RET_PASS; + int c, i; + + while ((c = getopt(argc, argv, "cht:v:")) != -1) { + switch (c) { + case 'c': + log_color(1); + break; + case 'h': + usage(basename(argv[0])); + exit(0); + case 'v': + log_verbosity(atoi(optarg)); + break; + default: + usage(basename(argv[0])); + exit(1); + } + } + + ksft_print_header(); + ksft_set_plan(2); + ksft_print_msg("%s: Test FUTEX2_WAITV\n", + basename(argv[0])); + + for (i = 0; i < NR_FUTEXES; i++) { + waitv[i].uaddr = &futexes[i]; + waitv[i].flags = FUTEX_32; + waitv[i].val = 0; + } + + /* Private waitv */ + if (pthread_create(&waiter, NULL, waiterfn, NULL)) + error("pthread_create failed\n", errno); + + usleep(WAKE_WAIT_US); + + res = futex2_wake(waitv[NR_FUTEXES - 1].uaddr, 1, FUTEX_32); + if (res != 1) { + ksft_test_result_fail("futex2_waitv private returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_waitv private succeeds\n"); + } + + /* Shared waitv */ + for (i = 0; i < NR_FUTEXES; i++) { + int shm_id = shmget(IPC_PRIVATE, 4096, IPC_CREAT | 0666); + + if (shm_id < 0) { + perror("shmget"); + exit(1); + } + + unsigned int *shared_data = shmat(shm_id, NULL, 0); + + *shared_data = 0; + waitv[i].uaddr = shared_data; + waitv[i].flags = FUTEX_32 | FUTEX_SHARED_FLAG; + waitv[i].val = 0; + } + + if (pthread_create(&waiter, NULL, waiterfn, NULL)) + error("pthread_create failed\n", errno); + + usleep(WAKE_WAIT_US); + + res = futex2_wake(waitv[NR_FUTEXES - 1].uaddr, 1, FUTEX_32 | FUTEX_SHARED_FLAG); + if (res != 1) { + ksft_test_result_fail("futex2_waitv shared returned: %d %s\n", + res ? errno : res, + res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_waitv shared succeeds\n"); + } + + for (i = 0; i < NR_FUTEXES; i++) + shmdt(waitv[i].uaddr); + + ksft_print_cnts(); + return ret; +} diff --git a/tools/testing/selftests/futex/functional/futex_wait_timeout.c b/tools/testing/selftests/futex/functional/futex_wait_timeout.c index ee55e6d389a..4569bf303b0 100644 --- a/tools/testing/selftests/futex/functional/futex_wait_timeout.c +++ b/tools/testing/selftests/futex/functional/futex_wait_timeout.c @@ -11,6 +11,7 @@ * * HISTORY * 2009-Nov-6: Initial version by Darren Hart + * 2021-Feb-5: Add futex2 test by André * *****************************************************************************/ @@ -20,7 +21,7 @@ #include #include #include -#include "futextest.h" +#include "futex2test.h" #include "logging.h" #define TEST_NAME "futex-wait-timeout" @@ -40,7 +41,8 @@ void usage(char *prog) int main(int argc, char *argv[]) { futex_t f1 = FUTEX_INITIALIZER; - struct timespec to; + struct timespec to = {.tv_sec = 0, .tv_nsec = timeout_ns}; + struct timespec64 to64; int res, ret = RET_PASS; int c; @@ -65,22 +67,60 @@ int main(int argc, char *argv[]) } ksft_print_header(); - ksft_set_plan(1); + ksft_set_plan(3); ksft_print_msg("%s: Block on a futex and wait for timeout\n", basename(argv[0])); ksft_print_msg("\tArguments: timeout=%ldns\n", timeout_ns); - /* initialize timeout */ - to.tv_sec = 0; - to.tv_nsec = timeout_ns; - info("Calling futex_wait on f1: %u @ %p\n", f1, &f1); res = futex_wait(&f1, f1, &to, FUTEX_PRIVATE_FLAG); if (!res || errno != ETIMEDOUT) { - fail("futex_wait returned %d\n", ret < 0 ? errno : ret); + ksft_test_result_fail("futex_wait returned %d\n", ret < 0 ? errno : ret); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex_wait timeout succeeds\n"); + } + + /* setting absolute monotonic timeout for futex2 */ + if (gettime64(CLOCK_MONOTONIC, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_nsec += timeout_ns; + + if (to64.tv_nsec >= 1000000000) { + to64.tv_sec++; + to64.tv_nsec -= 1000000000; + } + + info("Calling futex2_wait on f1: %u @ %p\n", f1, &f1); + res = futex2_wait(&f1, f1, FUTEX_32, &to64); + if (!res || errno != ETIMEDOUT) { + ksft_test_result_fail("futex2_wait monotonic returned %d\n", ret < 0 ? errno : ret); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wait monotonic timeout succeeds\n"); + } + + /* setting absolute realtime timeout for futex2 */ + if (gettime64(CLOCK_REALTIME, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_nsec += timeout_ns; + + if (to64.tv_nsec >= 1000000000) { + to64.tv_sec++; + to64.tv_nsec -= 1000000000; + } + + info("Calling futex2_wait on f1: %u @ %p\n", f1, &f1); + res = futex2_wait(&f1, f1, FUTEX_32 | FUTEX_CLOCK_REALTIME, &to64); + if (!res || errno != ETIMEDOUT) { + ksft_test_result_fail("futex2_wait realtime returned %d\n", ret < 0 ? errno : ret); ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wait realtime timeout succeeds\n"); } - print_result(TEST_NAME, ret); + ksft_print_cnts(); return ret; } diff --git a/tools/testing/selftests/futex/functional/futex_wait_wouldblock.c b/tools/testing/selftests/futex/functional/futex_wait_wouldblock.c index 0ae390ff816..b1d463ebb33 100644 --- a/tools/testing/selftests/futex/functional/futex_wait_wouldblock.c +++ b/tools/testing/selftests/futex/functional/futex_wait_wouldblock.c @@ -12,6 +12,7 @@ * * HISTORY * 2009-Nov-14: Initial version by Gowrishankar + * 2021-Feb-5: Add futex2 test by André * *****************************************************************************/ @@ -21,7 +22,7 @@ #include #include #include -#include "futextest.h" +#include "futex2test.h" #include "logging.h" #define TEST_NAME "futex-wait-wouldblock" @@ -39,6 +40,7 @@ void usage(char *prog) int main(int argc, char *argv[]) { struct timespec to = {.tv_sec = 0, .tv_nsec = timeout_ns}; + struct timespec64 to64; futex_t f1 = FUTEX_INITIALIZER; int res, ret = RET_PASS; int c; @@ -61,18 +63,41 @@ int main(int argc, char *argv[]) } ksft_print_header(); - ksft_set_plan(1); + ksft_set_plan(2); ksft_print_msg("%s: Test the unexpected futex value in FUTEX_WAIT\n", basename(argv[0])); info("Calling futex_wait on f1: %u @ %p with val=%u\n", f1, &f1, f1+1); res = futex_wait(&f1, f1+1, &to, FUTEX_PRIVATE_FLAG); if (!res || errno != EWOULDBLOCK) { - fail("futex_wait returned: %d %s\n", + ksft_test_result_fail("futex_wait returned: %d %s\n", res ? errno : res, res ? strerror(errno) : ""); ret = RET_FAIL; + } else { + ksft_test_result_pass("futex_wait wouldblock succeeds\n"); } - print_result(TEST_NAME, ret); + /* setting absolute timeout for futex2 */ + if (gettime64(CLOCK_MONOTONIC, &to64)) + error("gettime64 failed\n", errno); + + to64.tv_nsec += timeout_ns; + + if (to64.tv_nsec >= 1000000000) { + to64.tv_sec++; + to64.tv_nsec -= 1000000000; + } + + info("Calling futex2_wait on f1: %u @ %p with val=%u\n", f1, &f1, f1+1); + res = futex2_wait(&f1, f1+1, FUTEX_32, &to64); + if (!res || errno != EWOULDBLOCK) { + ksft_test_result_fail("futex2_wait returned: %d %s\n", + res ? errno : res, res ? strerror(errno) : ""); + ret = RET_FAIL; + } else { + ksft_test_result_pass("futex2_wait wouldblock succeeds\n"); + } + + ksft_print_cnts(); return ret; } diff --git a/tools/testing/selftests/futex/functional/run.sh b/tools/testing/selftests/futex/functional/run.sh index 1acb6ace168..18b3883d723 100755 --- a/tools/testing/selftests/futex/functional/run.sh +++ b/tools/testing/selftests/futex/functional/run.sh @@ -73,3 +73,9 @@ echo echo ./futex_wait_uninitialized_heap $COLOR ./futex_wait_private_mapped_file $COLOR + +echo +./futex2_wait $COLOR + +echo +./futex2_waitv $COLOR diff --git a/tools/testing/selftests/futex/include/futex2test.h b/tools/testing/selftests/futex/include/futex2test.h new file mode 100644 index 00000000000..faa4027ce5b --- /dev/null +++ b/tools/testing/selftests/futex/include/futex2test.h @@ -0,0 +1,121 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/****************************************************************************** + * + * Copyright Collabora Ltd., 2021 + * + * DESCRIPTION + * Futex2 library addons for old futex library + * + * AUTHOR + * André Almeida + * + * HISTORY + * 2021-Feb-5: Initial version by André + * + *****************************************************************************/ +#include "futextest.h" +#include + +#define NSEC_PER_SEC 1000000000L + +#ifndef FUTEX_8 +# define FUTEX_8 0 +#endif +#ifndef FUTEX_16 +# define FUTEX_16 1 +#endif +#ifndef FUTEX_32 +# define FUTEX_32 2 +#endif + +#ifndef FUTEX_SHARED_FLAG +#define FUTEX_SHARED_FLAG 8 +#endif + +#ifndef FUTEX_WAITV_MAX +#define FUTEX_WAITV_MAX 128 +struct futex_waitv { + void *uaddr; + unsigned int val; + unsigned int flags; +}; +#endif + +/* + * - Y2038 section for 32-bit applications - + * + * Remove this when glibc is ready for y2038. Then, always compile with + * `-DTIME_BITS=64` or `-D__USE_TIME_BITS64`. glibc will provide both + * timespec64 and clock_gettime64 so we won't need to define here. + */ +#if defined(__i386__) || __TIMESIZE == 32 +# define NR_gettime __NR_clock_gettime64 +#else +# define NR_gettime __NR_clock_gettime +#endif + +struct timespec64 { + long long tv_sec; /* seconds */ + long long tv_nsec; /* nanoseconds */ +}; + +int gettime64(clock_t clockid, struct timespec64 *tv) +{ + return syscall(NR_gettime, clockid, tv); +} +/* + * - End of Y2038 section - + */ + +/** + * futex2_wait - If (*uaddr == val), wait at uaddr until timo + * @uaddr: User address to wait on + * @val: Expected value at uaddr, return if is not equal + * @flags: Operation flags + * @timo: Optional timeout for operation + */ +static inline int futex2_wait(volatile void *uaddr, unsigned long val, + unsigned long flags, struct timespec64 *timo) +{ + return syscall(__NR_futex_wait, uaddr, val, flags, timo); +} + +/** + * futex2_wake - Wake a number of waiters at uaddr + * @uaddr: Address to wake + * @nr: Number of waiters to wake + * @flags: Operation flags + */ +static inline int futex2_wake(volatile void *uaddr, unsigned int nr, unsigned long flags) +{ + return syscall(__NR_futex_wake, uaddr, nr, flags); +} + +/** + * futex2_waitv - Wait at multiple futexes, wake on any + * @waiters: Array of waiters + * @nr_waiters: Length of waiters array + * @flags: Operation flags + * @timo: Optional timeout for operation + */ +static inline int futex2_waitv(volatile struct futex_waitv *waiters, unsigned long nr_waiters, + unsigned long flags, struct timespec64 *timo) +{ + return syscall(__NR_futex_waitv, waiters, nr_waiters, flags, timo); +} + +/** + * futex2_requeue - Wake futexes at uaddr1 and requeue from uaddr1 to uaddr2 + * @uaddr1: Original address to wake and requeue from + * @uaddr2: Address to requeue to + * @nr_wake: Number of futexes to wake at uaddr1 before requeuing + * @nr_requeue: Number of futexes to requeue from uaddr1 to uaddr2 + * @cmpval: If (uaddr1->uaddr != cmpval), return immediatally + * @flgas: Operation flags + */ +static inline int futex2_requeue(struct futex_requeue *uaddr1, struct futex_requeue *uaddr2, + unsigned int nr_wake, unsigned int nr_requeue, + unsigned int cmpval, unsigned long flags) +{ + return syscall(__NR_futex_requeue, uaddr1, uaddr2, nr_wake, nr_requeue, cmpval, flags); +}