sfrench/cifs-2.6.git
2 years agoMerge branch 'for-linus' into test
Jens Axboe [Wed, 7 Feb 2018 22:54:25 +0000 (15:54 -0700)]
Merge branch 'for-linus' into test

* for-linus:
  block, bfq: add requeue-request hook
  bcache: fix for data collapse after re-attaching an attached device
  bcache: return attach error when no cache set exist
  bcache: set writeback_rate_update_seconds in range [1, 60] seconds
  bcache: fix for allocator and register thread race
  bcache: set error_limit correctly
  bcache: properly set task state in bch_writeback_thread()
  bcache: fix high CPU occupancy during journal
  bcache: add journal statistic
  block: Add should_fail_bio() for bpf error injection
  blk-wbt: account flush requests correctly

2 years agoMerge branch 'master' into test
Jens Axboe [Wed, 7 Feb 2018 22:54:20 +0000 (15:54 -0700)]
Merge branch 'master' into test

* master: (1190 commits)
  ASoC: stm32: add of dependency for stm32 drivers
  ASoC: mt8173-rt5650: fix child-node lookup
  ASoC: dapm: fix debugfs read using path->connected
  platform/x86: samsung-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  platform/x86: ideapad-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  platform/x86: dell-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  seq_file: Introduce DEFINE_SHOW_ATTRIBUTE() helper macro
  Documentation/sysctl/user.txt: fix typo
  MAINTAINERS: update ARM/QUALCOMM SUPPORT patterns
  MAINTAINERS: update various PALM patterns
  MAINTAINERS: update "ARM/OXNAS platform support" patterns
  MAINTAINERS: update Cortina/Gemini patterns
  MAINTAINERS: remove ARM/CLKDEV SUPPORT file pattern
  MAINTAINERS: remove ANDROID ION pattern
  mm: docs: add blank lines to silence sphinx "Unexpected indentation" errors
  mm: docs: fix parameter names mismatch
  mm: docs: fixup punctuation
  pipe: read buffer limits atomically
  pipe: simplify round_pipe_size()
  pipe: reject F_SETPIPE_SZ with size over UINT_MAX
  ...

2 years agoMerge tag 'modules-for-v4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/jeyu...
Linus Torvalds [Wed, 7 Feb 2018 22:29:34 +0000 (14:29 -0800)]
Merge tag 'modules-for-v4.16' of git://git./linux/kernel/git/jeyu/linux

Pull modules updates from Jessica Yu:
 "Minor code cleanups and MAINTAINERS update"

* tag 'modules-for-v4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/jeyu/linux:
  modpost: Remove trailing semicolon
  ftrace/module: Move ftrace_release_mod() to ddebug_cleanup label
  MAINTAINERS: Remove from module & paravirt maintenance

2 years agoMerge tag 'iversion-v4.16-2' of git://git.kernel.org/pub/scm/linux/kernel/git/jlayton...
Linus Torvalds [Wed, 7 Feb 2018 22:25:22 +0000 (14:25 -0800)]
Merge tag 'iversion-v4.16-2' of git://git./linux/kernel/git/jlayton/linux

Pull inode->i_version cleanup from Jeff Layton:
 "Goffredo went ahead and sent a patch to rename this function, and
  reverse its sense, as we discussed last week.

  The patch is very straightforward and I figure it's probably best to
  go ahead and merge this to get the API as settled as possible"

* tag 'iversion-v4.16-2' of git://git.kernel.org/pub/scm/linux/kernel/git/jlayton/linux:
  iversion: Rename make inode_cmp_iversion{+raw} to inode_eq_iversion{+raw}

2 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs
Linus Torvalds [Wed, 7 Feb 2018 22:23:06 +0000 (14:23 -0800)]
Merge branch 'for_linus' of git://git./linux/kernel/git/jack/linux-fs

Pull UDF and ext2 fixlets from Jan Kara:
 "A UDF fix and an ext2 cleanup"

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs:
  ext2: drop unneeded newline
  udf: Sanitize nanoseconds for time stamps

2 years agoblock, bfq: add requeue-request hook
Paolo Valente [Wed, 7 Feb 2018 21:19:20 +0000 (22:19 +0100)]
block, bfq: add requeue-request hook

Commit 'a6a252e64914 ("blk-mq-sched: decide how to handle flush rq via
RQF_FLUSH_SEQ")' makes all non-flush re-prepared requests for a device
be re-inserted into the active I/O scheduler for that device. As a
consequence, I/O schedulers may get the same request inserted again,
even several times, without a finish_request invoked on that request
before each re-insertion.

This fact is the cause of the failure reported in [1]. For an I/O
scheduler, every re-insertion of the same re-prepared request is
equivalent to the insertion of a new request. For schedulers like
mq-deadline or kyber, this fact causes no harm. In contrast, it
confuses a stateful scheduler like BFQ, which keeps state for an I/O
request, until the finish_request hook is invoked on the request. In
particular, BFQ may get stuck, waiting forever for the number of
request dispatches, of the same request, to be balanced by an equal
number of request completions (while there will be one completion for
that request). In this state, BFQ may refuse to serve I/O requests
from other bfq_queues. The hang reported in [1] then follows.

However, the above re-prepared requests undergo a requeue, thus the
requeue_request hook of the active elevator is invoked for these
requests, if set. This commit then addresses the above issue by
properly implementing the hook requeue_request in BFQ.

[1] https://marc.info/?l=linux-block&m=151211117608676

Reported-by: Ivan Kozik <ivan@ludios.org>
Reported-by: Alban Browaeys <alban.browaeys@gmail.com>
Tested-by: Mike Galbraith <efault@gmx.de>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Signed-off-by: Serena Ziviani <ziviani.serena@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agoMerge tag 'regulator-fix-v4.16-suspend' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 7 Feb 2018 21:17:07 +0000 (13:17 -0800)]
Merge tag 'regulator-fix-v4.16-suspend' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fix from Mark Brown:
 "Fix suspend to idle.

  Testing on mainline after the initial regulator pull request went in
  identified a regression for suspend to idle due to it calling the
  suspend operations with states that it wasn't realized could happen,
  this patch fixes the problem"

* tag 'regulator-fix-v4.16-suspend' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: Fix suspend to idle

2 years agoMerge tag 'fbdev-v4.16' of git://github.com/bzolnier/linux
Linus Torvalds [Wed, 7 Feb 2018 21:10:43 +0000 (13:10 -0800)]
Merge tag 'fbdev-v4.16' of git://github.com/bzolnier/linux

Pull fbdev updates from Bartlomiej Zolnierkiewicz:
 "There is nothing really major here:

   - fix display-timings lookup in the Device Tree in atmel_lcdfb driver
     (Johan Hovold)

   - fix video mode and line_length to be set correctly in vfb driver
     (Pieter "PoroCYon" Sluys)

   - fix returning nonsensical values to the user-space on GIO_FONTX
     ioctl when using dummy console (Nicolas Pitre)

   - add missing license tag to mmpfb driver (Arnd Bergmann)

   - convert radeonfb and pxa3xx_gcu drivers to use ktime_get[_ts64]()
     instead of the deprecated do_gettimeofday() (Arnd Bergmann)

   - switch udlfb driver from using the pr_*() logging functions to the
     dev_*() ones + related cleanups (Ladislav Michl)

   - use __raw I/O accessors also on arm64 (Ji Zhang)

   - fix Kconfig help text for intelfb driver (Randy Dunlap)

   - do not duplicate features data in omapfb driver (Ladislav Michl)

   - misc cleanups (Colin Ian King, Markus Elfring, Rasmus Villemoes,
     Vasyl Gomonovych, Himanshu Jha, Michael Trimarchi)"

* tag 'fbdev-v4.16' of git://github.com/bzolnier/linux: (25 commits)
  video: udlfb: Switch from the pr_*() to the dev_*() logging functions
  video: udlfb: Constify read only data
  video: fbdev/mmp: add MODULE_LICENSE
  console/dummy: leave .con_font_get set to NULL
  fbdev: mxsfb: use framebuffer_alloc in the correct way
  video: udlfb: Do not name private data 'dev'
  video: udlfb: Remove noisy warnings
  video: udlfb: Remove redundant gdev variable
  video: udlfb: Remove unnecessary local variable
  fbdev: auo_k190x: Use zeroing memory allocator instead of allocator/memset
  vfb: fix video mode and line_length being set when loaded
  fbdev: arm64 use __raw I/O memory api
  omapfb: dss: Do not duplicate features data
  video: fbdev: omap2: Use PTR_ERR_OR_ZERO()
  fbdev: au1200fb: delete duplicate header contents
  fbdev: pxa3xx: use ktime_get_ts64 for time stamps
  fbdev: radeon: use ktime_get() for HZ calibration
  video: smscufx: Improve a size determination in two functions
  video: udlfb: Delete an unnecessary return statement in two functions
  video: udlfb: Improve a size determination in dlfb_alloc_urb_list()
  ...

2 years agoMerge tag 'platform-drivers-x86-v4.16-2' of git://git.infradead.org/linux-platform...
Linus Torvalds [Wed, 7 Feb 2018 20:47:23 +0000 (12:47 -0800)]
Merge tag 'platform-drivers-x86-v4.16-2' of git://git.infradead.org/linux-platform-drivers-x86

Pull more x86 platform-drivers updates from Andy Shevchenko:
 "The DEFINE_SHOW_ATTRIBUTE() macro was defined privately in three
  locations and is useful for new and old users to avoid a lot of code
  duplication.

  Move the macro to seq_file.h.

  Along with above, clean up three drivers to use that macro.

  This, due to dependencies, was sent separately since affected changes
  weren't upstream originally yet. The rationale of doing this now is to
  allow use of new macro in v4.17 cycle in a conflictless manner"

* tag 'platform-drivers-x86-v4.16-2' of git://git.infradead.org/linux-platform-drivers-x86:
  platform/x86: samsung-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  platform/x86: ideapad-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  platform/x86: dell-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
  seq_file: Introduce DEFINE_SHOW_ATTRIBUTE() helper macro

2 years agoMerge tag 'asoc-v4.16-5' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie...
Linus Torvalds [Wed, 7 Feb 2018 20:11:09 +0000 (12:11 -0800)]
Merge tag 'asoc-v4.16-5' of git://git./linux/kernel/git/broonie/sound

Pull more ASoC updates from Mark Brown:
 "With the merge window having been delayed for another week here's
  another batch of updates that came in during that week.

  There's a few important fixes in here, mainly a fix for I/O on a
  number of devices caused by some of the component rework and a fix for
  a potential issue if more than one component in a link provides
  compressed operations. The I/O fixes are particularly important as the
  problem causes a power regression on a number of OMAP platforms"

* tag 'asoc-v4.16-5' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound: (22 commits)
  ASoC: stm32: add of dependency for stm32 drivers
  ASoC: mt8173-rt5650: fix child-node lookup
  ASoC: dapm: fix debugfs read using path->connected
  ASoC: compress: Fixup error messages
  ASoC: compress: Remove some extraneous blank lines
  ASoC: compress: Correct handling of copy callback
  ASoC: Intel: kbl: Enable mclk and ssp sclk early
  ASoC: Intel: Skylake: Add extended I2S config blob support in Clock driver
  ASoC: Intel: Skylake: Add ssp clock driver
  ASoC: Fix twl4030 and 6040 regression by adding back read and write
  ASoC: sun8i-codec: Add ADC support for a33
  ASoC: rockchip: Use dummy_dai for rt5514 dsp dailink
  ASoC: soc-pcm: rename .pmdown_time to .use_pmdown_time for Component
  ASoC: ak4613: call dummy write for PW_MGMT1/3 when Playback
  ASoC: soc-pcm: don't call flush_delayed_work() many times in soc_pcm_private_free()
  ASoC: soc-core: snd_soc_rtdcom_lookup() cares component driver name
  ASoC: sam9x5_wm8731: Drop 'ASoC' prefix from error messages
  ASoC: sam9g20_wm8731: use dev_*() logging functions
  ASoC: max98373 Changed SPDX header in C++ comments style
  ASoC: dmic: Fix check of return value from read of 'num-channels'
  ...

2 years agoMerge tag 'linux-watchdog-4.16-rc1' of git://www.linux-watchdog.org/linux-watchdog
Linus Torvalds [Wed, 7 Feb 2018 19:54:34 +0000 (11:54 -0800)]
Merge tag 'linux-watchdog-4.16-rc1' of git://linux-watchdog.org/linux-watchdog

Pull watchdog updates from Wim Van Sebroeck:

 - new watchdog device drivers for Realtek RTD1295 and Spreadtrum SC9860
   platform

 - add support for the following devices: jz4780 SoC, AST25xx series SoC
   and r8a77970 SoC

 - convert to watchdog framework: i6300esb_wdt, xen_wdt and sp5100_tco

 - several fixes for watchdog core

 - remove at32ap700x and obsolete documentation

 - gpio: Convert to use GPIO descriptors

 - rename gemini into FTWDT010 as this IP block is generc from Faraday
   Technology

 - various clean-ups and small bugfixes

 - add Guenter Roeck as co-maintainer

 - change maintainers e-mail address

* tag 'linux-watchdog-4.16-rc1' of git://www.linux-watchdog.org/linux-watchdog: (74 commits)
  documentation: watchdog: remove documentation of w83697hf_wdt/w83697ug_wdt
  documentation: watchdog: remove documentation for ixp2000
  documentation: watchdog: remove documentation of at32ap700x_wdt
  watchdog: remove at32ap700x_wdt
  watchdog: sp5100_tco: Add support for recent FCH versions
  watchdog: sp5100-tco: Abort if watchdog is disabled by hardware
  watchdog: sp5100_tco: Use bit operations
  watchdog: sp5100_tco: Convert to use watchdog subsystem
  watchdog: sp5100_tco: Clean up function and variable names
  watchdog: sp5100_tco: Use dev_ print functions where possible
  watchdog: sp5100_tco: Match PCI device early
  watchdog: sp5100_tco: Clean up sp5100_tco_setupdevice
  watchdog: sp5100_tco: Use standard error codes
  watchdog: sp5100_tco: Use request_muxed_region where possible
  watchdog: sp5100_tco: Fix watchdog disable bit
  watchdog: sp5100_tco: Always use SP5100_IO_PM_{INDEX_REG,DATA_REG}
  watchdog: core: make sure the watchdog_worker is not deferred
  watchdog: mt7621: switch to using managed devm_watchdog_register_device()
  watchdog: mt7621: set WDOG_HW_RUNNING bit when appropriate
  watchdog: imx2_wdt: restore previous timeout after suspend+resume
  ...

2 years agobcache: fix for data collapse after re-attaching an attached device
Tang Junhui [Wed, 7 Feb 2018 19:41:46 +0000 (11:41 -0800)]
bcache: fix for data collapse after re-attaching an attached device

back-end device sdm has already attached a cache_set with ID
f67ebe1f-f8bc-4d73-bfe5-9dc88607f119, then try to attach with
another cache set, and it returns with an error:
[root]# cd /sys/block/sdm/bcache
[root]# echo 5ccd0a63-148e-48b8-afa2-aca9cbd6279f > attach
-bash: echo: write error: Invalid argument

After that, execute a command to modify the label of bcache
device:
[root]# echo data_disk1 > label

Then we reboot the system, when the system power on, the back-end
device can not attach to cache_set, a messages show in the log:
Feb  5 12:05:52 ceph152 kernel: [922385.508498] bcache:
bch_cached_dev_attach() couldn't find uuid for sdm in set

In sysfs_attach(), dc->sb.set_uuid was assigned to the value
which input through sysfs, no matter whether it is success
or not in bch_cached_dev_attach(). For example, If the back-end
device has already attached to an cache set, bch_cached_dev_attach()
would fail, but dc->sb.set_uuid was changed. Then modify the
label of bcache device, it will call bch_write_bdev_super(),
which would write the dc->sb.set_uuid to the super block, so we
record a wrong cache set ID in the super block, after the system
reboot, the cache set couldn't find the uuid of the back-end
device, so the bcache device couldn't exist and use any more.

In this patch, we don't assigned cache set ID to dc->sb.set_uuid
in sysfs_attach() directly, but input it into bch_cached_dev_attach(),
and assigned dc->sb.set_uuid to the cache set ID after the back-end
device attached to the cache set successful.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: return attach error when no cache set exist
Tang Junhui [Wed, 7 Feb 2018 19:41:45 +0000 (11:41 -0800)]
bcache: return attach error when no cache set exist

I attach a back-end device to a cache set, and the cache set is not
registered yet, this back-end device did not attach successfully, and no
error returned:
[root]# echo 87859280-fec6-4bcc-20df7ca8f86b > /sys/block/sde/bcache/attach
[root]#

In sysfs_attach(), the return value "v" is initialized to "size" in
the beginning, and if no cache set exist in bch_cache_sets, the "v" value
would not change any more, and return to sysfs, sysfs regard it as success
since the "size" is a positive number.

This patch fixes this issue by assigning "v" with "-ENOENT" in the
initialization.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: set writeback_rate_update_seconds in range [1, 60] seconds
Coly Li [Wed, 7 Feb 2018 19:41:44 +0000 (11:41 -0800)]
bcache: set writeback_rate_update_seconds in range [1, 60] seconds

dc->writeback_rate_update_seconds can be set via sysfs and its value can
be set to [1, ULONG_MAX].  It does not make sense to set such a large
value, 60 seconds is long enough value considering the default 5 seconds
works well for long time.

Because dc->writeback_rate_update is a special delayed work, it re-arms
itself inside the delayed work routine update_writeback_rate(). When
stopping it by cancel_delayed_work_sync(), there should be a timeout to
wait and make sure the re-armed delayed work is stopped too. A small max
value of dc->writeback_rate_update_seconds is also helpful to decide a
reasonable small timeout.

This patch limits sysfs interface to set dc->writeback_rate_update_seconds
in range of [1, 60] seconds, and replaces the hand-coded number by macros.

Changelog:
v2: fix a rebase typo in v4, which is pointed out by Michael Lyle.
v1: initial version.

Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: fix for allocator and register thread race
Tang Junhui [Wed, 7 Feb 2018 19:41:43 +0000 (11:41 -0800)]
bcache: fix for allocator and register thread race

After long time running of random small IO writing,
I reboot the machine, and after the machine power on,
I found bcache got stuck, the stack is:
[root@ceph153 ~]# cat /proc/2510/task/*/stack
[<ffffffffa06b2455>] closure_sync+0x25/0x90 [bcache]
[<ffffffffa06b6be8>] bch_journal+0x118/0x2b0 [bcache]
[<ffffffffa06b6dc7>] bch_journal_meta+0x47/0x70 [bcache]
[<ffffffffa06be8f7>] bch_prio_write+0x237/0x340 [bcache]
[<ffffffffa06a8018>] bch_allocator_thread+0x3c8/0x3d0 [bcache]
[<ffffffff810a631f>] kthread+0xcf/0xe0
[<ffffffff8164c318>] ret_from_fork+0x58/0x90
[<ffffffffffffffff>] 0xffffffffffffffff
[root@ceph153 ~]# cat /proc/2038/task/*/stack
[<ffffffffa06b1abd>] __bch_btree_map_nodes+0x12d/0x150 [bcache]
[<ffffffffa06b1bd1>] bch_btree_insert+0xf1/0x170 [bcache]
[<ffffffffa06b637f>] bch_journal_replay+0x13f/0x230 [bcache]
[<ffffffffa06c75fe>] run_cache_set+0x79a/0x7c2 [bcache]
[<ffffffffa06c0cf8>] register_bcache+0xd48/0x1310 [bcache]
[<ffffffff812f702f>] kobj_attr_store+0xf/0x20
[<ffffffff8125b216>] sysfs_write_file+0xc6/0x140
[<ffffffff811dfbfd>] vfs_write+0xbd/0x1e0
[<ffffffff811e069f>] SyS_write+0x7f/0xe0
[<ffffffff8164c3c9>] system_call_fastpath+0x16/0x1
The stack shows the register thread and allocator thread
were getting stuck when registering cache device.

I reboot the machine several times, the issue always
exsit in this machine.

I debug the code, and found the call trace as bellow:
register_bcache()
   ==>run_cache_set()
      ==>bch_journal_replay()
         ==>bch_btree_insert()
            ==>__bch_btree_map_nodes()
               ==>btree_insert_fn()
                  ==>btree_split() //node need split
                     ==>btree_check_reserve()
In btree_check_reserve(), It will check if there is enough buckets
of RESERVE_BTREE type, since allocator thread did not work yet, so
no buckets of RESERVE_BTREE type allocated, so the register thread
waits on c->btree_cache_wait, and goes to sleep.

Then the allocator thread initialized, the call trace is bellow:
bch_allocator_thread()
==>bch_prio_write()
   ==>bch_journal_meta()
      ==>bch_journal()
         ==>journal_wait_for_write()
In journal_wait_for_write(), It will check if journal is full by
journal_full(), but the long time random small IO writing
causes the exhaustion of journal buckets(journal.blocks_free=0),
In order to release the journal buckets,
the allocator calls btree_flush_write() to flush keys to
btree nodes, and waits on c->journal.wait until btree nodes writing
over or there has already some journal buckets space, then the
allocator thread goes to sleep. but in btree_flush_write(), since
bch_journal_replay() is not finished, so no btree nodes have journal
(condition "if (btree_current_write(b)->journal)" never satisfied),
so we got no btree node to flush, no journal bucket released,
and allocator sleep all the times.

Through the above analysis, we can see that:
1) Register thread wait for allocator thread to allocate buckets of
   RESERVE_BTREE type;
2) Alloctor thread wait for register thread to replay journal, so it
   can flush btree nodes and get journal bucket.
   then they are all got stuck by waiting for each other.

Hua Rui provided a patch for me, by allocating some buckets of
RESERVE_BTREE type in advance, so the register thread can get bucket
when btree node splitting and no need to waiting for the allocator
thread. I tested it, it has effect, and register thread run a step
forward, but finally are still got stuck, the reason is only 8 bucket
of RESERVE_BTREE type were allocated, and in bch_journal_replay(),
after 2 btree nodes splitting, only 4 bucket of RESERVE_BTREE type left,
then btree_check_reserve() is not satisfied anymore, so it goes to sleep
again, and in the same time, alloctor thread did not flush enough btree
nodes to release a journal bucket, so they all got stuck again.

So we need to allocate more buckets of RESERVE_BTREE type in advance,
but how much is enough?  By experience and test, I think it should be
as much as journal buckets. Then I modify the code as this patch,
and test in the machine, and it works.

This patch modified base on Hua Rui’s patch, and allocate more buckets
of RESERVE_BTREE type in advance to avoid register thread and allocate
thread going to wait for each other.

[patch v2] ca->sb.njournal_buckets would be 0 in the first time after
cache creation, and no journal exists, so just 8 btree buckets is OK.

Signed-off-by: Hua Rui <huarui.dev@gmail.com>
Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: set error_limit correctly
Coly Li [Wed, 7 Feb 2018 19:41:42 +0000 (11:41 -0800)]
bcache: set error_limit correctly

Struct cache uses io_errors for two purposes,
- Error decay: when cache set error_decay is set, io_errors is used to
  generate a small piece of delay when I/O error happens.
- I/O errors counter: in order to generate big enough value for error
  decay, I/O errors counter value is stored by left shifting 20 bits (a.k.a
  IO_ERROR_SHIFT).

In function bch_count_io_errors(), if I/O errors counter reaches cache set
error limit, bch_cache_set_error() will be called to retire the whold cache
set. But current code is problematic when checking the error limit, see the
following code piece from bch_count_io_errors(),

 90     if (error) {
 91             char buf[BDEVNAME_SIZE];
 92             unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
 93                                                 &ca->io_errors);
 94             errors >>= IO_ERROR_SHIFT;
 95
 96             if (errors < ca->set->error_limit)
 97                     pr_err("%s: IO error on %s, recovering",
 98                            bdevname(ca->bdev, buf), m);
 99             else
100                     bch_cache_set_error(ca->set,
101                                         "%s: too many IO errors %s",
102                                         bdevname(ca->bdev, buf), m);
103     }

At line 94, errors is right shifting IO_ERROR_SHIFT bits, now it is real
errors counter to compare at line 96. But ca->set->error_limit is initia-
lized with an amplified value in bch_cache_set_alloc(),
1545         c->error_limit  = 8 << IO_ERROR_SHIFT;

It means by default, in bch_count_io_errors(), before 8<<20 errors happened
bch_cache_set_error() won't be called to retire the problematic cache
device. If the average request size is 64KB, it means bcache won't handle
failed device until 512GB data is requested. This is too large to be an I/O
threashold. So I believe the correct error limit should be much less.

This patch sets default cache set error limit to 8, then in
bch_count_io_errors() when errors counter reaches 8 (if it is default
value), function bch_cache_set_error() will be called to retire the whole
cache set. This patch also removes bits shifting when store or show
io_error_limit value via sysfs interface.

Nowadays most of SSDs handle internal flash failure automatically by LBA
address re-indirect mapping. If an I/O error can be observed by upper layer
code, it will be a notable error because that SSD can not re-indirect
map the problematic LBA address to an available flash block. This situation
indicates the whole SSD will be failed very soon. Therefore setting 8 as
the default io error limit value makes sense, it is enough for most of
cache devices.

Changelog:
v2: add reviewed-by from Hannes.
v1: initial version for review.

Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Cc: Junhui Tang <tang.junhui@zte.com.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: properly set task state in bch_writeback_thread()
Coly Li [Wed, 7 Feb 2018 19:41:41 +0000 (11:41 -0800)]
bcache: properly set task state in bch_writeback_thread()

Kernel thread routine bch_writeback_thread() has the following code block,

447         down_write(&dc->writeback_lock);
448~450     if (check conditions) {
451                 up_write(&dc->writeback_lock);
452                 set_current_state(TASK_INTERRUPTIBLE);
453
454                 if (kthread_should_stop())
455                         return 0;
456
457                 schedule();
458                 continue;
459         }

If condition check is true, its task state is set to TASK_INTERRUPTIBLE
and call schedule() to wait for others to wake up it.

There are 2 issues in current code,
1, Task state is set to TASK_INTERRUPTIBLE after the condition checks, if
   another process changes the condition and call wake_up_process(dc->
   writeback_thread), then at line 452 task state is set back to
   TASK_INTERRUPTIBLE, the writeback kernel thread will lose a chance to be
   waken up.
2, At line 454 if kthread_should_stop() is true, writeback kernel thread
   will return to kernel/kthread.c:kthread() with TASK_INTERRUPTIBLE and
   call do_exit(). It is not good to enter do_exit() with task state
   TASK_INTERRUPTIBLE, in following code path might_sleep() is called and a
   warning message is reported by __might_sleep(): "WARNING: do not call
   blocking ops when !TASK_RUNNING; state=1 set at [xxxx]".

For the first issue, task state should be set before condition checks.
Ineed because dc->writeback_lock is required when modifying all the
conditions, calling set_current_state() inside code block where dc->
writeback_lock is hold is safe. But this is quite implicit, so I still move
set_current_state() before all the condition checks.

For the second issue, frankley speaking it does not hurt when kernel thread
exits with TASK_INTERRUPTIBLE state, but this warning message scares users,
makes them feel there might be something risky with bcache and hurt their
data.  Setting task state to TASK_RUNNING before returning fixes this
problem.

In alloc.c:allocator_wait(), there is also a similar issue, and is also
fixed in this patch.

Changelog:
v3: merge two similar fixes into one patch
v2: fix the race issue in v1 patch.
v1: initial buggy fix.

Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Cc: Michael Lyle <mlyle@lyle.org>
Cc: Junhui Tang <tang.junhui@zte.com.cn>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: fix high CPU occupancy during journal
Tang Junhui [Wed, 7 Feb 2018 19:41:40 +0000 (11:41 -0800)]
bcache: fix high CPU occupancy during journal

After long time small writing I/O running, we found the occupancy of CPU
is very high and I/O performance has been reduced by about half:

[root@ceph151 internal]# top
top - 15:51:05 up 1 day,2:43,  4 users,  load average: 16.89, 15.15, 16.53
Tasks: 2063 total,   4 running, 2059 sleeping,   0 stopped,   0 zombie
%Cpu(s):4.3 us, 17.1 sy 0.0 ni, 66.1 id, 12.0 wa,  0.0 hi,  0.5 si,  0.0 st
KiB Mem : 65450044 total, 24586420 free, 38909008 used,  1954616 buff/cache
KiB Swap: 65667068 total, 65667068 free,        0 used. 25136812 avail Mem

  PID USER PR NI    VIRT    RES    SHR S %CPU %MEM     TIME+ COMMAND
 2023 root 20  0       0      0      0 S 55.1  0.0   0:04.42 kworker/11:191
14126 root 20  0       0      0      0 S 42.9  0.0   0:08.72 kworker/10:3
 9292 root 20  0       0      0      0 S 30.4  0.0   1:10.99 kworker/6:1
 8553 ceph 20  0 4242492 1.805g  18804 S 30.0  2.9 410:07.04 ceph-osd
12287 root 20  0       0      0      0 S 26.7  0.0   0:28.13 kworker/7:85
31019 root 20  0       0      0      0 S 26.1  0.0   1:30.79 kworker/22:1
 1787 root 20  0       0      0      0 R 25.7  0.0   5:18.45 kworker/8:7
32169 root 20  0       0      0      0 S 14.5  0.0   1:01.92 kworker/23:1
21476 root 20  0       0      0      0 S 13.9  0.0   0:05.09 kworker/1:54
 2204 root 20  0       0      0      0 S 12.5  0.0   1:25.17 kworker/9:10
16994 root 20  0       0      0      0 S 12.2  0.0   0:06.27 kworker/5:106
15714 root 20  0       0      0      0 R 10.9  0.0   0:01.85 kworker/19:2
 9661 ceph 20  0 4246876 1.731g  18800 S 10.6  2.8 403:00.80 ceph-osd
11460 ceph 20  0 4164692 2.206g  18876 S 10.6  3.5 360:27.19 ceph-osd
 9960 root 20  0       0      0      0 S 10.2  0.0   0:02.75 kworker/2:139
11699 ceph 20  0 4169244 1.920g  18920 S 10.2  3.1 355:23.67 ceph-osd
 6843 ceph 20  0 4197632 1.810g  18900 S  9.6  2.9 380:08.30 ceph-osd

The kernel work consumed a lot of CPU, and I found they are running journal
work, The journal is reclaiming source and flush btree node with surprising
frequency.

Through further analysis, we found that in btree_flush_write(), we try to
get a btree node with the smallest fifo idex to flush by traverse all the
btree nodein c->bucket_hash, after we getting it, since no locker protects
it, this btree node may have been written to cache device by other works,
and if this occurred, we retry to traverse in c->bucket_hash and get
another btree node. When the problem occurrd, the retry times is very high,
and we consume a lot of CPU in looking for a appropriate btree node.

In this patch, we try to record 128 btree nodes with the smallest fifo idex
in heap, and pop one by one when we need to flush btree node. It greatly
reduces the time for the loop to find the appropriate BTREE node, and also
reduce the occupancy of CPU.

[note by mpl: this triggers a checkpatch error because of adjacent,
pre-existing style violations]

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agobcache: add journal statistic
Tang Junhui [Wed, 7 Feb 2018 19:41:39 +0000 (11:41 -0800)]
bcache: add journal statistic

Sometimes, Journal takes up a lot of CPU, we need statistics
to know what's the journal is doing. So this patch provide
some journal statistics:
1) reclaim: how many times the journal try to reclaim resource,
   usually the journal bucket or/and the pin are exhausted.
2) flush_write: how many times the journal try to flush btree node
   to cache device, usually the journal bucket are exhausted.
3) retry_flush_write: how many times the journal retry to flush
   the next btree node, usually the previous tree node have been
   flushed by other thread.
we show these statistic by sysfs interface. Through these statistics
We can totally see the status of journal module when the CPU is too
high.

Signed-off-by: Tang Junhui <tang.junhui@zte.com.cn>
Reviewed-by: Michael Lyle <mlyle@lyle.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
2 years agoMerge tag 'riscv-for-linus-4.16-merge_window' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Wed, 7 Feb 2018 19:33:08 +0000 (11:33 -0800)]
Merge tag 'riscv-for-linus-4.16-merge_window' of git://git./linux/kernel/git/palmer/riscv-linux

Pull RISC-V updates from Palmer Dabbelt:
 "This contains the fixes we'd like to target for the 4.16 merge window.
  It's not as much as I was originally hoping to do but between glibc,
  the chip, and FOSDEM there just wasn't enough time to get everything
  put together. As such, this merge window is essentially just going to
  be small changes. This includes mostly cleanups:

   - A build fix failure to the audit test cases.

     RISC-V doesn't have renameat because the generic syscall ABI moved
     to renameat2 by the time of our port. The syscall audit test cases
     don't understand this, so I added a trivial fix. This went through
     mailing list review during the 4.15 merge window, but nobody has
     picked it up so I think it's best to just do this here.

   - The removal of our command-line argument processing code. The
     "mem_end" stuff was broken and the rest duplicated generic device
     tree code. The generic code was already being called.

   - Some unused/redundant code has been removed, including
     __ARCH_HAVE_MMU, current_pgdir, and the initialization of
     init_mm.pgd.

   - SUM is disabled upon taking a trap, which means that user memory is
     protected during traps taking inside copy_{to,from}_user().

   - The sptbr CSR has been renamed to satp in C code. We haven't
     changed the assembly code in order to maintain compatibility with
     binutils 2.29, which doesn't understand the new name.

  Additionally, we're adding some new features:

   - Basic ftrace support, thanks to Alan Kao!

   - Support for ZONE_DMA32.

     This is necessary for all the normal reasons, but also to deal with
     a deficiency in the Xilinx PCIe controller we're using on our
     FPGA-based systems. While the ZONE_DMA32 addition should be
     sufficient for most uses, it doesn't complete the fix for the
     Xilinx controller.

   - TLB shootdowns now only target the harts where they're necessary,
     instead of applying to all harts in the system.

  These patches have all been sitting on our linux-next branch for a
  while now. Due to time constraints this is all I feel comfortable
  submitting during the 4.16 merge window, hopefully we'll do better
  next time!"

[ Note to self: "harts" is RISC-V speak for "hardware threads".  I had
  to look that up.    - Linus ]

* tag 'riscv-for-linus-4.16-merge_window' of git://git.kernel.org/pub/scm/linux/kernel/git/palmer/riscv-linux:
  riscv: inline set_pgdir into its only caller
  riscv: rename sptbr to satp
  riscv: don't read back satp in paging_init
  riscv: remove the unused current_pgdir function
  riscv: add ZONE_DMA32
  RISC-V: Limit the scope of TLB shootdowns
  riscv: disable SUM in the exception handler
  riscv: remove redundant unlikely()
  riscv: remove unused __ARCH_HAVE_MMU define
  riscv/ftrace: Add basic support
  RISC-V: Remove mem_end command line processing
  RISC-V: Remove duplicate command-line parsing logic
  audit: Avoid build failures on systems without renameat

2 years agoMerge tag 'mips_fixes_4.16_1' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan...
Linus Torvalds [Wed, 7 Feb 2018 19:31:05 +0000 (11:31 -0800)]
Merge tag 'mips_fixes_4.16_1' of git://git./linux/kernel/git/jhogan/mips

Pull MIPS fixes from James Hogan:
 "A couple of MIPS fixes for 4.16-rc1, including an important regression
  in 4.15 and a rather more longstanding corner case build fix.

  These are separate from the main pull request as one of the bugs fixed
  was only recently introduced in v4.15-rc8.

   - Fix CPS regression on older binutils due to MIPS_ISA_LEVEL_RAW fix
     (4.15)

   - Fix allmodconfig + CONFIG_MACH_TX49XX=y builds due to incorrect use
     of IS_ENABLED() (2.6.28)"

* tag 'mips_fixes_4.16_1' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips:
  MIPS: TXx9: use IS_BUILTIN() for CONFIG_LEDS_CLASS
  MIPS: CPS: Fix MIPS_ISA_LEVEL_RAW fallout

2 years agoMerge tag 'mips_4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips
Linus Torvalds [Wed, 7 Feb 2018 19:22:44 +0000 (11:22 -0800)]
Merge tag 'mips_4.16' of git://git./linux/kernel/git/jhogan/mips

Pull MIPS updates from James Hogan:
 "These are the main MIPS changes for 4.16.

  Rough overview:

   (1) Basic support for the Ingenic JZ4770 based GCW Zero open-source
       handheld video game console

   (2) Support for the Ranchu board (used by Android emulator)

   (3) Various cleanups and misc improvements

  More detailed summary:

  Fixes:
   - Fix generic platform's USB_*HCI_BIG_ENDIAN selects (4.9)
   - Fix vmlinuz default build when ZBOOT selected
   - Fix clean up of vmlinuz targets
   - Fix command line duplication (in preparation for Ingenic JZ4770)

  Miscellaneous:
   - Allow Processor ID reads to be to be optimised away by the compiler
     (improves performance when running in guest)
   - Push ARCH_MIGHT_HAVE_PC_SERIO/PARPORT down to platform level to
     disable on generic platform with Ranchu board support
   - Add helpers for assembler macro instructions for older assemblers
   - Use assembler macro instructions to support VZ, XPA & MSA
     operations on older assemblers, removing C wrapper duplication
   - Various improvements to VZ & XPA assembly wrappers
   - Add drivers/platform/mips/ to MIPS MAINTAINERS entry

  Minor cleanups:
   - Misc FPU emulation cleanups (removal of unnecessary include, moving
     macros to common header, checkpatch and sparse fixes)
   - Remove duplicate assignment of core in play_dead()
   - Remove duplication in watchpoint handling
   - Remove mips_dma_mapping_error() stub
   - Use NULL instead of 0 in prepare_ftrace_return()
   - Use proper kernel-doc Return keyword for
     __compute_return_epc_for_insn()
   - Remove duplicate semicolon in csum_fold()

  Platform support:

  Broadcom:
   - Enable ZBOOT on BCM47xx

  Generic platform:
   - Add Ranchu board support, used by Android emulator
   - Fix machine compatible string matching for Ranchu
   - Support GIC in EIC mode

  Ingenic platforms:
   - Add DT, defconfig and other support for JZ4770 SoC and GCW Zero
   - Support dynamnic machine types (i.e. JZ4740 / JZ4770 / JZ4780)
   - Add Ingenic JZ4770 CGU clocks
   - General Ingenic clk changes to prepare for JZ4770 SoC support
   - Use common command line handling code
   - Add DT vendor prefix to GCW (Game Consoles Worldwide)

  Loongson:
   - Add MAINTAINERS entry for Loongson2 and Loongson3 platforms
   - Drop 32-bit support for Loongson 2E/2F devices
   - Fix build failures due to multiple use of 'MEM_RESERVED'"

* tag 'mips_4.16' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips: (53 commits)
  MIPS: Malta: Sanitize mouse and keyboard configuration.
  MIPS: Update defconfigs after previous patch.
  MIPS: Push ARCH_MIGHT_HAVE_PC_SERIO down to platform level
  MIPS: Push ARCH_MIGHT_HAVE_PC_PARPORT down to platform level
  MIPS: SMP-CPS: Remove duplicate assignment of core in play_dead
  MIPS: Generic: Support GIC in EIC mode
  MIPS: generic: Fix Makefile alignment
  MIPS: generic: Fix ranchu_of_match[] termination
  MIPS: generic: Fix machine compatible matching
  MIPS: Loongson fix name confict - MEM_RESERVED
  MIPS: bcm47xx: enable ZBOOT support
  MIPS: Fix trailing semicolon
  MIPS: Watch: Avoid duplication of bits in mips_read_watch_registers
  MIPS: Watch: Avoid duplication of bits in mips_install_watch_registers.
  MIPS: MSA: Update helpers to use new asm macros
  MIPS: XPA: Standardise readx/writex accessors
  MIPS: XPA: Allow use of $0 (zero) to MTHC0
  MIPS: XPA: Use XPA instructions in assembly
  MIPS: VZ: Pass GC0 register names in $n format
  MIPS: VZ: Update helpers to use new asm macros
  ...

2 years agoMerge tag 'docs-4.16-2' of git://git.lwn.net/linux
Linus Torvalds [Wed, 7 Feb 2018 17:42:59 +0000 (09:42 -0800)]
Merge tag 'docs-4.16-2' of git://git.lwn.net/linux

Pull more documentation updates from Jonathan Corbet:
 "A few late-arriving fixes, along with Konstantin's PGP document that
  had no reason to wait another cycle"

* tag 'docs-4.16-2' of git://git.lwn.net/linux:
  Documentation/process: tweak pgp maintainer guide
  Documentation/admin-guide: fixes for thunderbolt.rst
  Documentation: mips: Update AU1xxx_IDE Kconfig dependencies
  Fix broken link in Documentation/process/kernel-docs.rst
  Documentation/process: kernel maintainer PGP guide

2 years agoMerge remote-tracking branches 'asoc/topic/sam9x5_wm8731', 'asoc/topic/sgtl5000'...
Mark Brown [Wed, 7 Feb 2018 11:25:47 +0000 (11:25 +0000)]
Merge remote-tracking branches 'asoc/topic/sam9x5_wm8731', 'asoc/topic/sgtl5000' and 'asoc/topic/sun8i-codec' into asoc-next

2 years agoMerge remote-tracking branches 'asoc/topic/max98373', 'asoc/topic/mtk', 'asoc/topic...
Mark Brown [Wed, 7 Feb 2018 11:25:44 +0000 (11:25 +0000)]
Merge remote-tracking branches 'asoc/topic/max98373', 'asoc/topic/mtk', 'asoc/topic/pcm', 'asoc/topic/rockchip' and 'asoc/topic/sam9g20_wm8731' into asoc-next

2 years agoMerge remote-tracking branches 'asoc/topic/ak4613', 'asoc/topic/core', 'asoc/topic...
Mark Brown [Wed, 7 Feb 2018 11:25:41 +0000 (11:25 +0000)]
Merge remote-tracking branches 'asoc/topic/ak4613', 'asoc/topic/core', 'asoc/topic/dmic' and 'asoc/topic/intel' into asoc-next

2 years agoMerge remote-tracking branch 'asoc/topic/compress' into asoc-next
Mark Brown [Wed, 7 Feb 2018 11:24:21 +0000 (11:24 +0000)]
Merge remote-tracking branch 'asoc/topic/compress' into asoc-next

2 years agoMerge remote-tracking branch 'asoc/fix/twl-breakage' into asoc-next
Mark Brown [Wed, 7 Feb 2018 11:23:48 +0000 (11:23 +0000)]
Merge remote-tracking branch 'asoc/fix/twl-breakage' into asoc-next

2 years agoMerge remote-tracking branches 'asoc/fix/compress', 'asoc/fix/core', 'asoc/fix/dapm...
Mark Brown [Wed, 7 Feb 2018 11:21:02 +0000 (11:21 +0000)]
Merge remote-tracking branches 'asoc/fix/compress', 'asoc/fix/core', 'asoc/fix/dapm', 'asoc/fix/mtk' and 'asoc/fix/stm' into asoc-next

2 years agoASoC: stm32: add of dependency for stm32 drivers
Olivier Moysan [Mon, 5 Feb 2018 10:49:45 +0000 (11:49 +0100)]
ASoC: stm32: add of dependency for stm32 drivers

Add of dependency for STM32 ASoC drivers.
DFSDM of dependency is already inherited
from STM32_DFSDM_ADC dependency.

Signed-off-by: olivier moysan <olivier.moysan@st.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
2 years agoASoC: mt8173-rt5650: fix child-node lookup
Johan Hovold [Wed, 31 Jan 2018 05:59:21 +0000 (16:59 +1100)]
ASoC: mt8173-rt5650: fix child-node lookup

This driver used the wrong OF-helper when looking up the optional
capture-codec child node during probe. Instead of searching just
children of the sound node, a tree-wide depth-first search starting at
the unrelated platform node was done. Not only could this end up
matching an unrelated node or no node at all; the platform node could
also be prematurely freed since of_find_node_by_name() drops a reference
to its first argument. This particular pattern has been observed leading
to crashes after probe deferrals in other drivers.

Fix this by dropping the broken call to of_find_node_by_name() and
keeping only the second, correct lookup using of_get_child_by_name()
while taking care not to bail out if the optional node is missing.

Note that this also addresses two capture-codec node-reference leaks
(one for each of the original helper calls).

Compile tested only.

Fixes: d349caeb0510 ("ASoC: mediatek: Add second I2S on mt8173-rt5650 machine driver")
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Mark Brown <broonie@kernel.org>
2 years agoASoC: dapm: fix debugfs read using path->connected
KaiChieh Chuang [Mon, 5 Feb 2018 05:00:00 +0000 (13:00 +0800)]
ASoC: dapm: fix debugfs read using path->connected

This fix a bug in dapm_widget_power_read_file(),
where it may sent opposite order of source/sink widget
into the p->connected().

for example,
static int connected_check(source, sink);
{"w_sink", NULL, "w_source", connected_check}

the dapm_widget_power_read_file() will query p->connected()
in following case
p->conneted("w_source", "w_sink")
p->conneted("w_sink", "w_source")
we should avoid the last case, since it's the wrong order (source/sink)
as declared in snd_soc_dapm_route.

Signed-off-by: KaiChieh Chuang <kaichieh.chuang@mediatek.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
2 years agoplatform/x86: samsung-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
Andy Shevchenko [Mon, 22 Jan 2018 16:05:46 +0000 (18:05 +0200)]
platform/x86: samsung-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro

...instead of open coding file operations followed by custom ->open()
callbacks per each attribute.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agoplatform/x86: ideapad-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
Andy Shevchenko [Mon, 22 Jan 2018 16:05:45 +0000 (18:05 +0200)]
platform/x86: ideapad-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro

...instead of open coding file operations followed by custom ->open()
callbacks per each attribute.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agoplatform/x86: dell-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro
Andy Shevchenko [Mon, 22 Jan 2018 16:05:44 +0000 (18:05 +0200)]
platform/x86: dell-laptop: Re-use DEFINE_SHOW_ATTRIBUTE() macro

...instead of open coding file operations followed by custom ->open()
callbacks per each attribute.

Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agoseq_file: Introduce DEFINE_SHOW_ATTRIBUTE() helper macro
Andy Shevchenko [Mon, 22 Jan 2018 16:05:43 +0000 (18:05 +0200)]
seq_file: Introduce DEFINE_SHOW_ATTRIBUTE() helper macro

The DEFINE_SHOW_ATTRIBUTE() helper macro would be useful for current
users, which are many of them, and for new comers to decrease code
duplication.

Acked-by: Lee Jones <lee.jones@linaro.org>
Acked-by: Darren Hart (VMware) <dvhart@infradead.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
2 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Wed, 7 Feb 2018 06:15:42 +0000 (22:15 -0800)]
Merge branch 'akpm' (patches from Andrew)

Merge misc updates from Andrew Morton:

 - kasan updates

 - procfs

 - lib/bitmap updates

 - other lib/ updates

 - checkpatch tweaks

 - rapidio

 - ubsan

 - pipe fixes and cleanups

 - lots of other misc bits

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (114 commits)
  Documentation/sysctl/user.txt: fix typo
  MAINTAINERS: update ARM/QUALCOMM SUPPORT patterns
  MAINTAINERS: update various PALM patterns
  MAINTAINERS: update "ARM/OXNAS platform support" patterns
  MAINTAINERS: update Cortina/Gemini patterns
  MAINTAINERS: remove ARM/CLKDEV SUPPORT file pattern
  MAINTAINERS: remove ANDROID ION pattern
  mm: docs: add blank lines to silence sphinx "Unexpected indentation" errors
  mm: docs: fix parameter names mismatch
  mm: docs: fixup punctuation
  pipe: read buffer limits atomically
  pipe: simplify round_pipe_size()
  pipe: reject F_SETPIPE_SZ with size over UINT_MAX
  pipe: fix off-by-one error when checking buffer limits
  pipe: actually allow root to exceed the pipe buffer limits
  pipe, sysctl: remove pipe_proc_fn()
  pipe, sysctl: drop 'min' parameter from pipe-max-size converter
  kasan: rework Kconfig settings
  crash_dump: is_kdump_kernel can be boolean
  kernel/mutex: mutex_is_locked can be boolean
  ...

2 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 7 Feb 2018 03:57:31 +0000 (19:57 -0800)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler updates from Ingo Molnar:

 - membarrier updates (Mathieu Desnoyers)

 - SMP balancing optimizations (Mel Gorman)

 - stats update optimizations (Peter Zijlstra)

 - RT scheduler race fixes (Steven Rostedt)

 - misc fixes and updates

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/fair: Use a recently used CPU as an idle candidate and the basis for SIS
  sched/fair: Do not migrate if the prev_cpu is idle
  sched/fair: Restructure wake_affine*() to return a CPU id
  sched/fair: Remove unnecessary parameters from wake_affine_idle()
  sched/rt: Make update_curr_rt() more accurate
  sched/rt: Up the root domain ref count when passing it around via IPIs
  sched/rt: Use container_of() to get root domain in rto_push_irq_work_func()
  sched/core: Optimize update_stats_*()
  sched/core: Optimize ttwu_stat()
  membarrier/selftest: Test private expedited sync core command
  membarrier/arm64: Provide core serializing command
  membarrier/x86: Provide core serializing command
  membarrier: Provide core serializing command, *_SYNC_CORE
  lockin/x86: Implement sync_core_before_usermode()
  locking: Introduce sync_core_before_usermode()
  membarrier/selftest: Test global expedited command
  membarrier: Provide GLOBAL_EXPEDITED command
  membarrier: Document scheduler barrier requirements
  powerpc, membarrier: Skip memory barrier in switch_mm()
  membarrier/selftest: Test private expedited command

2 years agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 7 Feb 2018 03:56:00 +0000 (19:56 -0800)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf fixes from Ingo Molnar:
 "Tooling fixes, plus add missing interval sampling to certain x86 PEBS
  events"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf tools: Add trace/beauty/generated/ into .gitignore
  perf trace: Fix call-graph output
  x86/events/intel/ds: Add PERF_SAMPLE_PERIOD into PEBS_FREERUNNING_FLAGS
  perf record: Fix period option handling
  perf evsel: Fix period/freq terms setup
  tools headers: Synchoronize x86 features UAPI headers
  tools headers: Synchronize uapi/linux/sched.h
  tools headers: Sync {tools/,}arch/powerpc/include/uapi/asm/kvm.h
  tooling headers: Synchronize updated s390 kvm UAPI headers
  tools headers: Synchronize sound/asound.h

2 years agoMerge branch 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 7 Feb 2018 03:28:19 +0000 (19:28 -0800)]
Merge branch 'locking-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull locking fixlets from Ingo Molnar:
 "An endianness fix and a jump labels branch hint update"

* 'locking-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  locking/qrwlock: include asm/byteorder.h as needed
  jump_label: Add branch hints to static_branch_{un,}likely()

2 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Wed, 7 Feb 2018 03:00:42 +0000 (19:00 -0800)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) Fix error path in netdevsim, from Jakub Kicinski.

 2) Default values listed in tcp_wmem and tcp_rmem documentation were
    inaccurate, from Tonghao Zhang.

 3) Fix route leaks in SCTP, both for ipv4 and ipv6. From Alexey Kodanev
    and Tommi Rantala.

 4) Fix "MASK < Y" meant to be "MASK << Y" in xgbe driver, from Wolfram
    Sang.

 5) Use after free in u32_destroy_key(), from Paolo Abeni.

 6) Fix two TX issues in be2net driver, from Suredh Reddy.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (25 commits)
  be2net: Handle transmit completion errors in Lancer
  be2net: Fix HW stall issue in Lancer
  RDS: IB: Fix null pointer issue
  nfp: fix kdoc warnings on nested structures
  sample/bpf: fix erspan metadata
  net: erspan: fix erspan config overwrite
  net: erspan: fix metadata extraction
  cls_u32: fix use after free in u32_destroy_key()
  net: amd-xgbe: fix comparison to bitshift when dealing with a mask
  net: phy: Handle not having GPIO enabled in the kernel
  ibmvnic: fix empty firmware version and errors cleanup
  sctp: fix dst refcnt leak in sctp_v4_get_dst
  sctp: fix dst refcnt leak in sctp_v6_get_dst()
  dwc-xlgmac: remove Jie Deng as co-maintainer
  doc: Change the min default value of tcp_wmem/tcp_rmem.
  samples/bpf: use bpf_set_link_xdp_fd
  libbpf: add missing SPDX-License-Identifier
  libbpf: add error reporting in XDP
  libbpf: add function to setup XDP
  tools: add netlink.h and if_link.h in tools uapi
  ...

2 years agoDocumentation/sysctl/user.txt: fix typo
Kangmin Park [Tue, 6 Feb 2018 23:42:44 +0000 (15:42 -0800)]
Documentation/sysctl/user.txt: fix typo

Fix 'documetation' to 'documentation'

Link: http://lkml.kernel.org/r/CAKW4uUxRPZz59aWAX8ytaCB5=Qh6d_CvAnO7rYq-6NRAnQJbDA@mail.gmail.com
Signed-off-by: Kangmin Park <l4stpr0gr4m@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Jiri Kosina <trivial@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: update ARM/QUALCOMM SUPPORT patterns
Joe Perches [Tue, 6 Feb 2018 23:42:41 +0000 (15:42 -0800)]
MAINTAINERS: update ARM/QUALCOMM SUPPORT patterns

Commit 321737416c72d ("tty: serial: msm: Move header file into driver")
removed the .h file, update the patterns.

Link: http://lkml.kernel.org/r/2b7478bc4c35ab3ac6b06b4edd3b645a8c34a4a2.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Stephen Boyd <sboyd@codeaurora.org>
Cc: Andy Gross <andy.gross@linaro.org>
Cc: David Brown <david.brown@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: update various PALM patterns
Joe Perches [Tue, 6 Feb 2018 23:42:37 +0000 (15:42 -0800)]
MAINTAINERS: update various PALM patterns

Commit 4c25c5d2985c ("ARM: pxa: make more mach/*.h files local") moved
the files around, update the patterns.

Link: http://lkml.kernel.org/r/a291f6f61e378a1f35e266fe4c5f646b9feeaa6a.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Marek Vasut <marek.vasut@gmail.com>
Cc: Tomas Cech <sleep_walker@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: update "ARM/OXNAS platform support" patterns
Joe Perches [Tue, 6 Feb 2018 23:42:33 +0000 (15:42 -0800)]
MAINTAINERS: update "ARM/OXNAS platform support" patterns

Commit 9e6c62b05c1b ("ARM: dts: rename oxnas dts files") renamed the
files, update the patterns.

[akpm@linux-foundation.org: crunch into a single globbed term, per Arnd]
Link: http://lkml.kernel.org/r/b39d779e143b3c0a4e7dff827346e509447e3e8e.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Daniel Golle <daniel@makrotopia.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: update Cortina/Gemini patterns
Joe Perches [Tue, 6 Feb 2018 23:42:30 +0000 (15:42 -0800)]
MAINTAINERS: update Cortina/Gemini patterns

Commit 4d5ae32f5e1e ("net: ethernet: Add a driver for Gemini gigabit
ethernet") added invalid patterns.  Fix it.

Link: http://lkml.kernel.org/r/65b104609e0071d0fbe0dcce3a8e6138a4cf8c25.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Hans Ulli Kroll <ulli.kroll@googlemail.com>
Cc: David S. Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: remove ARM/CLKDEV SUPPORT file pattern
Joe Perches [Tue, 6 Feb 2018 23:42:26 +0000 (15:42 -0800)]
MAINTAINERS: remove ARM/CLKDEV SUPPORT file pattern

Commit 34d2f4d3a4d6 ("ARM: Use generic clkdev.h header") removed the
file, remove the pattern.

Link: http://lkml.kernel.org/r/41bfff9449a5894b94f583983b6c6cb46f4cd821.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Russell King <linux@armlinux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoMAINTAINERS: remove ANDROID ION pattern
Joe Perches [Tue, 6 Feb 2018 23:42:23 +0000 (15:42 -0800)]
MAINTAINERS: remove ANDROID ION pattern

The file drivers/staging/android/uapi/ion_test.h was removed by commit
9828282e33a0 ("staging: android: ion: Remove old platform support")

Remove the pattern.

Link: http://lkml.kernel.org/r/182debec22002c9a1de44e79a7441288942b205c.1517147485.git.joe@perches.com
Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Laura Abbott <labbott@redhat.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: docs: add blank lines to silence sphinx "Unexpected indentation" errors
Mike Rapoport [Tue, 6 Feb 2018 23:42:19 +0000 (15:42 -0800)]
mm: docs: add blank lines to silence sphinx "Unexpected indentation" errors

Link: http://lkml.kernel.org/r/1516700871-22279-4-git-send-email-rppt@linux.vnet.ibm.com
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: docs: fix parameter names mismatch
Mike Rapoport [Tue, 6 Feb 2018 23:42:16 +0000 (15:42 -0800)]
mm: docs: fix parameter names mismatch

There are several places where parameter descriptions do no match the
actual code.  Fix it.

Link: http://lkml.kernel.org/r/1516700871-22279-3-git-send-email-rppt@linux.vnet.ibm.com
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: docs: fixup punctuation
Mike Rapoport [Tue, 6 Feb 2018 23:42:13 +0000 (15:42 -0800)]
mm: docs: fixup punctuation

so that kernel-doc will properly recognize the parameter and function
descriptions.

Link: http://lkml.kernel.org/r/1516700871-22279-2-git-send-email-rppt@linux.vnet.ibm.com
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe: read buffer limits atomically
Eric Biggers [Tue, 6 Feb 2018 23:42:08 +0000 (15:42 -0800)]
pipe: read buffer limits atomically

The pipe buffer limits are accessed without any locking, and may be
changed at any time by the sysctl handlers.  In theory this could cause
problems for expressions like the following:

    pipe_user_pages_hard && user_bufs > pipe_user_pages_hard

...  since the assembly code might reference the 'pipe_user_pages_hard'
memory location multiple times, and if the admin removes the limit by
setting it to 0, there is a very brief window where processes could
incorrectly observe the limit to be exceeded.

Fix this by loading the limits with READ_ONCE() prior to use.

Link: http://lkml.kernel.org/r/20180111052902.14409-8-ebiggers3@gmail.com
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Willy Tarreau <w@1wt.eu>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe: simplify round_pipe_size()
Eric Biggers [Tue, 6 Feb 2018 23:42:05 +0000 (15:42 -0800)]
pipe: simplify round_pipe_size()

round_pipe_size() calculates the number of pages the requested size
corresponds to, then rounds the page count up to the next power of 2.

However, it also rounds everything < PAGE_SIZE up to PAGE_SIZE.
Therefore, there's no need to actually translate the size into a page
count; we just need to round the size up to the next power of 2.

We do need to verify the size isn't greater than (1 << 31), since on
32-bit systems roundup_pow_of_two() would be undefined in that case.  But
that can just be combined with the UINT_MAX check which we need anyway
now.

Finally, update pipe_set_size() to not redundantly check the return value
of round_pipe_size() for the "invalid size" case twice.

Link: http://lkml.kernel.org/r/20180111052902.14409-7-ebiggers3@gmail.com
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe: reject F_SETPIPE_SZ with size over UINT_MAX
Eric Biggers [Tue, 6 Feb 2018 23:42:00 +0000 (15:42 -0800)]
pipe: reject F_SETPIPE_SZ with size over UINT_MAX

A pipe's size is represented as an 'unsigned int'.  As expected, writing a
value greater than UINT_MAX to /proc/sys/fs/pipe-max-size fails with
EINVAL.  However, the F_SETPIPE_SZ fcntl silently truncates such values to
32 bits, rather than failing with EINVAL as expected.  (It *does* fail
with EINVAL for values above (1 << 31) but <= UINT_MAX.)

Fix this by moving the check against UINT_MAX into round_pipe_size() which
is called in both cases.

Link: http://lkml.kernel.org/r/20180111052902.14409-6-ebiggers3@gmail.com
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe: fix off-by-one error when checking buffer limits
Eric Biggers [Tue, 6 Feb 2018 23:41:56 +0000 (15:41 -0800)]
pipe: fix off-by-one error when checking buffer limits

With pipe-user-pages-hard set to 'N', users were actually only allowed up
to 'N - 1' buffers; and likewise for pipe-user-pages-soft.

Fix this to allow up to 'N' buffers, as would be expected.

Link: http://lkml.kernel.org/r/20180111052902.14409-5-ebiggers3@gmail.com
Fixes: b0b91d18e2e9 ("pipe: fix limit checking in pipe_set_size()")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Willy Tarreau <w@1wt.eu>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe: actually allow root to exceed the pipe buffer limits
Eric Biggers [Tue, 6 Feb 2018 23:41:53 +0000 (15:41 -0800)]
pipe: actually allow root to exceed the pipe buffer limits

pipe-user-pages-hard and pipe-user-pages-soft are only supposed to apply
to unprivileged users, as documented in both Documentation/sysctl/fs.txt
and the pipe(7) man page.

However, the capabilities are actually only checked when increasing a
pipe's size using F_SETPIPE_SZ, not when creating a new pipe.  Therefore,
if pipe-user-pages-hard has been set, the root user can run into it and be
unable to create pipes.  Similarly, if pipe-user-pages-soft has been set,
the root user can run into it and have their pipes limited to 1 page each.

Fix this by allowing the privileged override in both cases.

Link: http://lkml.kernel.org/r/20180111052902.14409-4-ebiggers3@gmail.com
Fixes: 759c01142a5d ("pipe: limit the per-user amount of pages allocated in pipes")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe, sysctl: remove pipe_proc_fn()
Eric Biggers [Tue, 6 Feb 2018 23:41:49 +0000 (15:41 -0800)]
pipe, sysctl: remove pipe_proc_fn()

pipe_proc_fn() is no longer needed, as it only calls through to
proc_dopipe_max_size().  Just put proc_dopipe_max_size() in the ctl_table
entry directly, and remove the unneeded EXPORT_SYMBOL() and the ENOSYS
stub for it.

(The reason the ENOSYS stub isn't needed is that the pipe-max-size
ctl_table entry is located directly in 'kern_table' rather than being
registered separately.  Therefore, the entry is already only defined when
the kernel is built with sysctl support.)

Link: http://lkml.kernel.org/r/20180111052902.14409-3-ebiggers3@gmail.com
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopipe, sysctl: drop 'min' parameter from pipe-max-size converter
Eric Biggers [Tue, 6 Feb 2018 23:41:45 +0000 (15:41 -0800)]
pipe, sysctl: drop 'min' parameter from pipe-max-size converter

Patch series "pipe: buffer limits fixes and cleanups", v2.

This series simplifies the sysctl handler for pipe-max-size and fixes
another set of bugs related to the pipe buffer limits:

- The root user wasn't allowed to exceed the limits when creating new
  pipes.

- There was an off-by-one error when checking the limits, so a limit of
  N was actually treated as N - 1.

- F_SETPIPE_SZ accepted values over UINT_MAX.

- Reading the pipe buffer limits could be racy.

This patch (of 7):

Before validating the given value against pipe_min_size,
do_proc_dopipe_max_size_conv() calls round_pipe_size(), which rounds the
value up to pipe_min_size.  Therefore, the second check against
pipe_min_size is redundant.  Remove it.

Link: http://lkml.kernel.org/r/20180111052902.14409-2-ebiggers3@gmail.com
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Acked-by: Joe Lawrence <joe.lawrence@redhat.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Mikulas Patocka <mpatocka@redhat.com>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokasan: rework Kconfig settings
Arnd Bergmann [Tue, 6 Feb 2018 23:41:41 +0000 (15:41 -0800)]
kasan: rework Kconfig settings

We get a lot of very large stack frames using gcc-7.0.1 with the default
-fsanitize-address-use-after-scope --param asan-stack=1 options, which can
easily cause an overflow of the kernel stack, e.g.

  drivers/gpu/drm/i915/gvt/handlers.c:2434:1: warning: the frame size of 46176 bytes is larger than 3072 bytes
  drivers/net/wireless/ralink/rt2x00/rt2800lib.c:5650:1: warning: the frame size of 23632 bytes is larger than 3072 bytes
  lib/atomic64_test.c:250:1: warning: the frame size of 11200 bytes is larger than 3072 bytes
  drivers/gpu/drm/i915/gvt/handlers.c:2621:1: warning: the frame size of 9208 bytes is larger than 3072 bytes
  drivers/media/dvb-frontends/stv090x.c:3431:1: warning: the frame size of 6816 bytes is larger than 3072 bytes
  fs/fscache/stats.c:287:1: warning: the frame size of 6536 bytes is larger than 3072 bytes

To reduce this risk, -fsanitize-address-use-after-scope is now split out
into a separate CONFIG_KASAN_EXTRA Kconfig option, leading to stack
frames that are smaller than 2 kilobytes most of the time on x86_64.  An
earlier version of this patch also prevented combining KASAN_EXTRA with
KASAN_INLINE, but that is no longer necessary with gcc-7.0.1.

All patches to get the frame size below 2048 bytes with CONFIG_KASAN=y
and CONFIG_KASAN_EXTRA=n have been merged by maintainers now, so we can
bring back that default now.  KASAN_EXTRA=y still causes lots of
warnings but now defaults to !COMPILE_TEST to disable it in
allmodconfig, and it remains disabled in all other defconfigs since it
is a new option.  I arbitrarily raise the warning limit for KASAN_EXTRA
to 3072 to reduce the noise, but an allmodconfig kernel still has around
50 warnings on gcc-7.

I experimented a bit more with smaller stack frames and have another
follow-up series that reduces the warning limit for 64-bit architectures
to 1280 bytes (without CONFIG_KASAN).

With earlier versions of this patch series, I also had patches to address
the warnings we get with KASAN and/or KASAN_EXTRA, using a
"noinline_if_stackbloat" annotation.

That annotation now got replaced with a gcc-8 bugfix (see
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=81715) and a workaround for
older compilers, which means that KASAN_EXTRA is now just as bad as
before and will lead to an instant stack overflow in a few extreme
cases.

This reverts parts of commit 3f181b4d8652 ("lib/Kconfig.debug: disable
-Wframe-larger-than warnings with KASAN=y").  Two patches in linux-next
should be merged first to avoid introducing warnings in an allmodconfig
build:
  3cd890dbe2a4 ("media: dvb-frontends: fix i2c access helpers for KASAN")
  16c3ada89cff ("media: r820t: fix r820t_write_reg for KASAN")

Do we really need to backport this?

I think we do: without this patch, enabling KASAN will lead to
unavoidable kernel stack overflow in certain device drivers when built
with gcc-7 or higher on linux-4.10+ or any version that contains a
backport of commit c5caf21ab0cf8.  Most people are probably still on
older compilers, but it will get worse over time as they upgrade their
distros.

The warnings we get on kernels older than this should all be for code
that uses dangerously large stack frames, though most of them do not
cause an actual stack overflow by themselves.The asan-stack option was
added in linux-4.0, and commit 3f181b4d8652 ("lib/Kconfig.debug:
disable -Wframe-larger-than warnings with KASAN=y") effectively turned
off the warning for allmodconfig kernels, so I would like to see this
fix backported to any kernels later than 4.0.

I have done dozens of fixes for individual functions with stack frames
larger than 2048 bytes with asan-stack, and I plan to make sure that
all those fixes make it into the stable kernels as well (most are
already there).

Part of the complication here is that asan-stack (from 4.0) was
originally assumed to always require much larger stacks, but that
turned out to be a combination of multiple gcc bugs that we have now
worked around and fixed, but sanitize-address-use-after-scope (from
v4.10) has a much higher inherent stack usage and also suffers from at
least three other problems that we have analyzed but not yet fixed
upstream, each of them makes the stack usage more severe than it should
be.

Link: http://lkml.kernel.org/r/20171221134744.2295529-1-arnd@arndb.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Andrey Konovalov <andreyknvl@google.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agocrash_dump: is_kdump_kernel can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:38 +0000 (15:41 -0800)]
crash_dump: is_kdump_kernel can be boolean

Make is_kdump_kernel return bool due to this particular function only
using either one or zero as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513308799-19232-8-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/mutex: mutex_is_locked can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:35 +0000 (15:41 -0800)]
kernel/mutex: mutex_is_locked can be boolean

Make mutex_is_locked return bool due to this particular function only
using either one or zero as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-7-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/module: module_is_live can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:31 +0000 (15:41 -0800)]
kernel/module: module_is_live can be boolean

Make module_is_live return bool due to this particular function only using
either one or zero as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-6-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/resource: iomem_is_exclusive can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:28 +0000 (15:41 -0800)]
kernel/resource: iomem_is_exclusive can be boolean

Make iomem_is_exclusive return bool due to this particular function only
using either one or zero as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-5-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/cpuset: current_cpuset_is_being_rebound can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:24 +0000 (15:41 -0800)]
kernel/cpuset: current_cpuset_is_being_rebound can be boolean

Make current_cpuset_is_being_rebound return bool due to this particular
function only using either one or zero as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-4-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/lockref: __lockref_is_dead can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:21 +0000 (15:41 -0800)]
lib/lockref: __lockref_is_dead can be boolean

Make __lockref_is_dead return bool due to this function only using either
true or false as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-3-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm/memblock: memblock_is_map/region_memory can be boolean
Yaowei Bai [Tue, 6 Feb 2018 23:41:18 +0000 (15:41 -0800)]
mm/memblock: memblock_is_map/region_memory can be boolean

Make memblock_is_map/region_memory return bool due to these two
functions only using either true or false as its return value.

No functional change.

Link: http://lkml.kernel.org/r/1513266622-15860-2-git-send-email-baiyaowei@cmss.chinamobile.com
Signed-off-by: Yaowei Bai <baiyaowei@cmss.chinamobile.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agogenirq: remove unneeded kallsyms include
Sergey Senozhatsky [Tue, 6 Feb 2018 23:41:13 +0000 (15:41 -0800)]
genirq: remove unneeded kallsyms include

The file was converted from print_symbol() to %pf some time ago in
commit ef26f20cd117 ("genirq: Print threaded handler in spurious debug
output").  kallsyms does not seem to be needed anymore.

Link: http://lkml.kernel.org/r/20171208025616.16267-10-sergey.senozhatsky@gmail.com
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agohrtimer: remove unneeded kallsyms include
Sergey Senozhatsky [Tue, 6 Feb 2018 23:41:10 +0000 (15:41 -0800)]
hrtimer: remove unneeded kallsyms include

hrtimer does not seem to use any of kallsyms functions/defines.

Link: http://lkml.kernel.org/r/20171208025616.16267-9-sergey.senozhatsky@gmail.com
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm: remove unneeded kallsyms include
Sergey Senozhatsky [Tue, 6 Feb 2018 23:41:06 +0000 (15:41 -0800)]
mm: remove unneeded kallsyms include

The file was converted from print_symbol() to %pSR a while ago in commit
071361d3473e ("mm: Convert print_symbol to %pSR").  kallsyms does not
seem to be needed anymore.

Link: http://lkml.kernel.org/r/20171208025616.16267-3-sergey.senozhatsky@gmail.com
Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agomm/userfaultfd.c: remove duplicate include
Pravin Shedge [Tue, 6 Feb 2018 23:41:03 +0000 (15:41 -0800)]
mm/userfaultfd.c: remove duplicate include

These duplicate includes have been found with scripts/checkincludes.pl but
they have been removed manually to avoid removing false positives.

Link: http://lkml.kernel.org/r/1512580957-6071-1-git-send-email-pravin.shedge4linux@gmail.com
Signed-off-by: Pravin Shedge <pravin.shedge4linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agovfs: remove might_sleep() from clear_inode()
Shakeel Butt [Tue, 6 Feb 2018 23:40:59 +0000 (15:40 -0800)]
vfs: remove might_sleep() from clear_inode()

Commit 7994e6f72543 ("vfs: Move waiting for inode writeback from
end_writeback() to evict_inode()") removed inode_sync_wait() from
end_writeback() and commit dbd5768f87ff ("vfs: Rename end_writeback() to
clear_inode()") renamed end_writeback() to clear_inode().

After these patches there is no sleeping operation in clear_inode().
So, remove might_sleep() from it.

Link: http://lkml.kernel.org/r/20171108004354.40308-1-shakeelb@google.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Greg Thelen <gthelen@google.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoarch/score/kernel/setup.c: combine two seq_printf() calls into one call in show_cpuinfo()
Markus Elfring [Tue, 6 Feb 2018 23:40:56 +0000 (15:40 -0800)]
arch/score/kernel/setup.c: combine two seq_printf() calls into one call in show_cpuinfo()

Some data were printed into a sequence by two separate function calls.
Print the same data by a single function call instead.

This issue was detected by using the Coccinelle software.

Link: http://lkml.kernel.org/r/ddcfff3a-9502-6ce0-b08a-365eb55ce958@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Cc: Chen Liqin <liqin.linux@gmail.com>
Cc: Lennox Wu <lennox.wu@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoipc/mqueue.c: have RT tasks queue in by priority in wq_add()
Jonathan Haws [Tue, 6 Feb 2018 23:40:52 +0000 (15:40 -0800)]
ipc/mqueue.c: have RT tasks queue in by priority in wq_add()

Previous behavior added tasks to the work queue using the static_prio
value instead of the dynamic priority value in prio.  This caused RT tasks
to be added to the work queue in a FIFO manner rather than by priority.
Normal tasks were handled by priority.

This fix utilizes the dynamic priority of the task to ensure that both RT
and normal tasks are added to the work queue in priority order.  Utilizing
the dynamic priority (prio) rather than the base priority (normal_prio)
was chosen to ensure that if a task had a boosted priority when it was
added to the work queue, it would be woken sooner to to ensure that it
releases any other locks it may be holding in a more timely manner.  It is
understood that the task could have a lower priority when it wakes than
when it was added to the queue in this (unlikely) case.

Link: http://lkml.kernel.org/r/1513006652-7014-1-git-send-email-jhaws@sdl.usu.edu
Signed-off-by: Jonathan Haws <jhaws@sdl.usu.edu>
Reviewed-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Reviewed-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoipc: fix ipc data structures inconsistency
Philippe Mikoyan [Tue, 6 Feb 2018 23:40:49 +0000 (15:40 -0800)]
ipc: fix ipc data structures inconsistency

As described in the title, this patch fixes <ipc>id_ds inconsistency when
<ipc>ctl_stat executes concurrently with some ds-changing function, e.g.
shmat, msgsnd or whatever.

For instance, if shmctl(IPC_STAT) is running concurrently
with shmat, following data structure can be returned:
{... shm_lpid = 0, shm_nattch = 1, ...}

Link: http://lkml.kernel.org/r/20171202153456.6514-1-philippe.mikoyan@skat.systems
Signed-off-by: Philippe Mikoyan <philippe.mikoyan@skat.systems>
Reviewed-by: Davidlohr Bueso <dbueso@suse.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/ubsan: remove returns-nonnull-attribute checks
Andrey Ryabinin [Tue, 6 Feb 2018 23:40:45 +0000 (15:40 -0800)]
lib/ubsan: remove returns-nonnull-attribute checks

Similarly to type mismatch checks, new GCC 8.x and Clang also changed for
ABI for returns_nonnull checks.  While we can update our code to conform
the new ABI it's more reasonable to just remove it.  Because it's just
dead code, we don't have any single user of returns_nonnull attribute in
the whole kernel.

And AFAIU the advantage that this attribute could bring would be mitigated
by -fno-delete-null-pointer-checks cflag that we use to build the kernel.
So it's unlikely we will have a lot of returns_nonnull attribute in
future.

So let's just remove the code, it has no use.

[aryabinin@virtuozzo.com: fix warning]
Link: http://lkml.kernel.org/r/20180122165711.11510-1-aryabinin@virtuozzo.com
Link: http://lkml.kernel.org/r/20180119152853.16806-2-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Sodagudi Prasad <psodagud@codeaurora.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/ubsan: add type mismatch handler for new GCC/Clang
Andrey Ryabinin [Tue, 6 Feb 2018 23:40:42 +0000 (15:40 -0800)]
lib/ubsan: add type mismatch handler for new GCC/Clang

UBSAN=y fails to build with new GCC/clang:

    arch/x86/kernel/head64.o: In function `sanitize_boot_params':
    arch/x86/include/asm/bootparam_utils.h:37: undefined reference to `__ubsan_handle_type_mismatch_v1'

because Clang and GCC 8 slightly changed ABI for 'type mismatch' errors.
Compiler now uses new __ubsan_handle_type_mismatch_v1() function with
slightly modified 'struct type_mismatch_data'.

Let's add new 'struct type_mismatch_data_common' which is independent from
compiler's layout of 'struct type_mismatch_data'.  And make
__ubsan_handle_type_mismatch[_v1]() functions transform compiler-dependent
type mismatch data to our internal representation.  This way, we can
support both old and new compilers with minimal amount of change.

Link: http://lkml.kernel.org/r/20180119152853.16806-1-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Reported-by: Sodagudi Prasad <psodagud@codeaurora.org>
Cc: <stable@vger.kernel.org> [4.5+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/ubsan.c: s/missaligned/misaligned/
Andrew Morton [Tue, 6 Feb 2018 23:40:38 +0000 (15:40 -0800)]
lib/ubsan.c: s/missaligned/misaligned/

A vist from the spelling fairy.

Cc: David Laight <David.Laight@ACULAB.COM>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agobuild_bug.h: remove BUILD_BUG_ON_NULL()
Masahiro Yamada [Tue, 6 Feb 2018 23:40:35 +0000 (15:40 -0800)]
build_bug.h: remove BUILD_BUG_ON_NULL()

This macro is only used by net/ipv6/mcast.c, but there is no reason
why it must be BUILD_BUG_ON_NULL().

Replace it with BUILD_BUG_ON_ZERO(), and remove BUILD_BUG_ON_NULL()
definition from <linux/build_bug.h>.

Link: http://lkml.kernel.org/r/1515121833-3174-3-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agoinclude/linux/genl_magic_func.h: remove own BUILD_BUG_ON*() defines
Masahiro Yamada [Tue, 6 Feb 2018 23:40:31 +0000 (15:40 -0800)]
include/linux/genl_magic_func.h: remove own BUILD_BUG_ON*() defines

Do not duplicate BUILD_BUG_ON*.  Use ones from <linux/build_bug.h>.

Link: http://lkml.kernel.org/r/1515121833-3174-2-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Ian Abbott <abbotti@mev.co.uk>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Hideaki YOSHIFUJI <yoshfuji@linux-ipv6.org>
Cc: Alexey Kuznetsov <kuznet@ms2.inr.ac.ru>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokcov: detect double association with a single task
Dmitry Vyukov [Tue, 6 Feb 2018 23:40:28 +0000 (15:40 -0800)]
kcov: detect double association with a single task

Currently KCOV_ENABLE does not check if the current task is already
associated with another kcov descriptor.  As the result it is possible
to associate a single task with more than one kcov descriptor, which
later leads to a memory leak of the old descriptor.  This relation is
really meant to be one-to-one (task has only one back link).

Extend validation to detect such misuse.

Link: http://lkml.kernel.org/r/20180122082520.15716-1-dvyukov@google.com
Fixes: 5c9a8750a640 ("kernel: add kcov code coverage")
Signed-off-by: Dmitry Vyukov <dvyukov@google.com>
Reported-by: Shankara Pailoor <sp3485@columbia.edu>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/relay.c: revert "kernel/relay.c: fix potential memory leak"
Eric Biggers [Tue, 6 Feb 2018 23:40:24 +0000 (15:40 -0800)]
kernel/relay.c: revert "kernel/relay.c: fix potential memory leak"

This reverts commit ba62bafe942b ("kernel/relay.c: fix potential memory leak").

This commit introduced a double free bug, because 'chan' is already
freed by the line:

    kref_put(&chan->kref, relay_destroy_channel);

This bug was found by syzkaller, using the BLKTRACESETUP ioctl.

Link: http://lkml.kernel.org/r/20180127004759.101823-1-ebiggers3@gmail.com
Fixes: ba62bafe942b ("kernel/relay.c: fix potential memory leak")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Zhouyi Zhou <yizhouzhou@ict.ac.cn>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: <stable@vger.kernel.org> [4.7+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopps: parport: use timespec64 instead of timespec
Arnd Bergmann [Tue, 6 Feb 2018 23:40:21 +0000 (15:40 -0800)]
pps: parport: use timespec64 instead of timespec

getnstimeofday() is deprecated, so I'm converting this to use
ktime_get_real_ts64() as a safe replacement.  I considered using
ktime_get_real() instead, but since the algorithm here depends on the
exact timing, I decided to introduce fewer changes and leave the code
that determines the nanoseconds since the last seconds wrap untouched.

It's not entirely clear to me whether we should also change the time
base to CLOCK_BOOTTIME or CLOCK_TAI.  With boottime, we would be
independent of changes due to settimeofday() and only see the speed
adjustment from the upstream clock source, with the downside of having
the signal be at an arbirary offset from the start of the UTC second
signal.  With CLOCK_TAI, we would use the same offset from the UTC
second as before and still suffer from settimeofday() adjustments, but
would be less confused during leap seconds.

Both boottime and tai only offer usable (i.e.  avoiding ktime_t to
timespec64 conversion) interfaces for ktime_t though, so either way,
changing it wouldn't take significantly more work.  CLOCK_MONOTONIC
could be used with ktime_get_ts64(), but would lose synchronization
across a suspend/resume cycle, which seems worse.

Link: http://lkml.kernel.org/r/20180116171451.3095620-1-arnd@arndb.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Rodolfo Giometti <giometti@enneenne.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agopids: introduce find_get_task_by_vpid() helper
Mike Rapoport [Tue, 6 Feb 2018 23:40:17 +0000 (15:40 -0800)]
pids: introduce find_get_task_by_vpid() helper

There are several functions that do find_task_by_vpid() followed by
get_task_struct().  We can use a helper function instead.

Link: http://lkml.kernel.org/r/1509602027-11337-1-git-send-email-rppt@linux.vnet.ibm.com
Signed-off-by: Mike Rapoport <rppt@linux.vnet.ibm.com>
Acked-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agodrivers/rapidio/devices/tsi721_dma.c: adjust six checks for null pointers
Markus Elfring [Tue, 6 Feb 2018 23:40:13 +0000 (15:40 -0800)]
drivers/rapidio/devices/tsi721_dma.c: adjust six checks for null pointers

checkpatch pointed out the following:

  Comparison to NULL could be written !...

Thus fix the affected source code places.

Link: http://lkml.kernel.org/r/c3e79a1a-891e-cb62-990f-bd99839311b9@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agodrivers/rapidio/devices/tsi721_dma.c: delete an unnecessary variable initialisation...
Markus Elfring [Tue, 6 Feb 2018 23:40:10 +0000 (15:40 -0800)]
drivers/rapidio/devices/tsi721_dma.c: delete an unnecessary variable initialisation in tsi721_alloc_chan_resources()

The local variable "desc" will eventually be set to an appropriate pointer
a bit later. Thus omit the explicit initialisation at the beginning.

Link: http://lkml.kernel.org/r/728d8e20-4ae9-d661-d932-2d99ce67e71f@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agodrivers/rapidio/devices/tsi721_dma.c: delete an error message for a failed memory...
Markus Elfring [Tue, 6 Feb 2018 23:40:06 +0000 (15:40 -0800)]
drivers/rapidio/devices/tsi721_dma.c: delete an error message for a failed memory allocation in tsi721_alloc_chan_resources()

Omit an extra message for a memory allocation failure in this function.

This issue was detected by using the Coccinelle software.

Link: http://lkml.kernel.org/r/a96fcaf8-ea24-bcac-0214-273620349d42@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: move 12 EXPORT_SYMBOL_GPL() calls to function implementations
Markus Elfring [Tue, 6 Feb 2018 23:40:01 +0000 (15:40 -0800)]
rapidio: move 12 EXPORT_SYMBOL_GPL() calls to function implementations

checkpatch pointed information out like the following.

  WARNING: EXPORT_SYMBOL(foo); should immediately follow its function/variable

Thus fix the affected source code places.

Link: http://lkml.kernel.org/r/5f51f606-ece8-7bff-bb86-81d182c49b98@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: return an error code only as a constant in two functions
Markus Elfring [Tue, 6 Feb 2018 23:39:58 +0000 (15:39 -0800)]
rapidio: return an error code only as a constant in two functions

* Return an error code without storing it in an intermediate variable.

* Delete the label "out" and local variable "rc" which became unnecessary
  with this refactoring.

Link: http://lkml.kernel.org/r/8db441e1-c227-64e7-4747-095da63d7a2e@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: delete an unnecessary variable initialisation in three functions
Markus Elfring [Tue, 6 Feb 2018 23:39:55 +0000 (15:39 -0800)]
rapidio: delete an unnecessary variable initialisation in three functions

The local variable "rc" will be set to an appropriate value a bit later.
Thus omit the explicit initialisation at the beginning.

Link: http://lkml.kernel.org/r/ed729e8c-815c-8b56-a48f-1be579d0cc2b@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: improve a size determination in five functions
Markus Elfring [Tue, 6 Feb 2018 23:39:51 +0000 (15:39 -0800)]
rapidio: improve a size determination in five functions

Replace the specification of data structures by pointer dereferences as
the parameter for the operator "sizeof" to make the corresponding size
determination a bit safer according to the Linux coding style
convention.

This issue was detected by using the Coccinelle software.

Link: http://lkml.kernel.org/r/495f571c-fb4d-b1d5-a6e5-494f2c537a8d@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: adjust five function calls together with a variable assignment
Markus Elfring [Tue, 6 Feb 2018 23:39:48 +0000 (15:39 -0800)]
rapidio: adjust five function calls together with a variable assignment

checkpatch pointed information out like the following.

  ERROR: do not use assignment in if condition

Thus fix the affected source code places.

Link: http://lkml.kernel.org/r/6f5631ed-ec3e-503d-850a-ab09b5736dbe@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: adjust 12 checks for null pointers
Markus Elfring [Tue, 6 Feb 2018 23:39:44 +0000 (15:39 -0800)]
rapidio: adjust 12 checks for null pointers

checkpatch pointed information out like the following.

  Comparison to NULL could be written ...

Thus fix the affected source code places.

Link: http://lkml.kernel.org/r/739f9f1c-3ebe-c21b-c5e2-31976cfbc0ee@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agorapidio: delete an error message for a failed memory allocation in rio_init_mports()
Markus Elfring [Tue, 6 Feb 2018 23:39:41 +0000 (15:39 -0800)]
rapidio: delete an error message for a failed memory allocation in rio_init_mports()

Patch series "RapidIO: Adjustments for some function implementations".

This patch (of 7):

Omit an extra message for a memory allocation failure in this function.

This issue was detected by using the Coccinelle software.

Link: http://lkml.kernel.org/r/dfd6887b-2521-cbf2-af5b-e70b5fa4c33c@users.sourceforge.net
Signed-off-by: Markus Elfring <elfring@users.sourceforge.net>
Acked-by: Alexandre Bounine <alexandre.bounine@idt.com>
Cc: Matt Porter <mporter@kernel.crashing.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agocpumask: make cpumask_size() return "unsigned int"
Alexey Dobriyan [Tue, 6 Feb 2018 23:39:37 +0000 (15:39 -0800)]
cpumask: make cpumask_size() return "unsigned int"

CPUmasks are never big enough to warrant 64-bit code.

Space savings:

add/remove: 0/0 grow/shrink: 1/4 up/down: 3/-17 (-14)
Function                                     old     new   delta
sched_init_numa                             1530    1533      +3
compat_sys_sched_setaffinity                 160     159      -1
sys_sched_getaffinity                        197     195      -2
sys_sched_setaffinity                        183     176      -7
compat_sys_sched_getaffinity                 179     172      -7

Link: http://lkml.kernel.org/r/20171204165531.GA8221@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/fork.c: add comment about usage of CLONE_FS flags and namespaces
Marcos Paulo de Souza [Tue, 6 Feb 2018 23:39:34 +0000 (15:39 -0800)]
kernel/fork.c: add comment about usage of CLONE_FS flags and namespaces

All other places that deals with namespaces have an explanation of why
the restriction is there.

The description added in this commit was based on commit e66eded8309e
("userns: Don't allow CLONE_NEWUSER | CLONE_FS").

Link: http://lkml.kernel.org/r/20171112151637.13258-1-marcos.souza.org@gmail.com
Signed-off-by: Marcos Paulo de Souza <marcos.souza.org@gmail.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokernel/fork.c: check error and return early
Marcos Paulo de Souza [Tue, 6 Feb 2018 23:39:30 +0000 (15:39 -0800)]
kernel/fork.c: check error and return early

Thus reducing one indentation level while maintaining the same rationale.

Link: http://lkml.kernel.org/r/20171117002929.5155-1-marcos.souza.org@gmail.com
Signed-off-by: Marcos Paulo de Souza <marcos.souza.org@gmail.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years ago<asm-generic/siginfo.h>: fix language in comments
Randy Dunlap [Tue, 6 Feb 2018 23:39:27 +0000 (15:39 -0800)]
<asm-generic/siginfo.h>: fix language in comments

Fix grammar and add an omitted word.

Link: http://lkml.kernel.org/r/1a5a021c-0207-f793-7f07-addca26772d5@infradead.org
Fixes: f9886bc50a8e ("signal: Document the strange si_codes used by ptrace event stops")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agohfsplus: honor setgid flag on directories
Ernesto A. Fernandez [Tue, 6 Feb 2018 23:49:02 +0000 (15:49 -0800)]
hfsplus: honor setgid flag on directories

When creating a file inside a directory that has the setgid flag set, give
the new file the group ID of the parent, and also the setgid flag if it is
a directory itself.

Link: http://lkml.kernel.org/r/20171204192705.GA6101@debian.home
Signed-off-by: Ernesto A. Fernandez <ernesto.mnd.fernandez@gmail.com>
Reviewed-by: Vyacheslav Dubeyko <slava@dubeyko.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agonilfs2: use time64_t internally
Arnd Bergmann [Tue, 6 Feb 2018 23:39:21 +0000 (15:39 -0800)]
nilfs2: use time64_t internally

The superblock and segment timestamps are used only internally in nilfs2
and can be read out using sysfs.

Since we are using the old 'get_seconds()' interface and store the data
as timestamps, the behavior differs slightly between 64-bit and 32-bit
kernels, the latter will show incorrect timestamps after 2038 in sysfs,
and presumably fail completely in 2106 as comparisons go wrong.

This changes nilfs2 to use time64_t with ktime_get_real_seconds() to
handle timestamps, making the behavior consistent and correct on both
32-bit and 64-bit machines.

The on-disk format already uses 64-bit timestamps, so nothing changes
there.

Link: http://lkml.kernel.org/r/20180122211050.1286441-1-arnd@arndb.de
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Ryusuke Konishi <konishi.ryusuke@lab.ntt.co.jp>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agokallsyms: let print_ip_sym() print raw addresses
Huacai Chen [Tue, 6 Feb 2018 23:39:18 +0000 (15:39 -0800)]
kallsyms: let print_ip_sym() print raw addresses

print_ip_sym() is mostly used for debugging, so I think it should print
the raw addresses.

Link: http://lkml.kernel.org/r/1514519382-405-1-git-send-email-chenhc@lemote.com
Signed-off-by: Huacai Chen <chenhc@lemote.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Fuxin Zhang <zhangfx@lemote.com>
Cc: "Tobin C. Harding" <me@tobin.cc>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>