Skip to content

L
linux
Switch branch/tag
  1. 07 Oct, 2019 20 commits
  3. 05 Oct, 2019 20 commits
    • Greg Kroah-Hartman's avatar
      Linux 4.19.77 · 6cad9d0c
      Greg Kroah-Hartman authored
      6cad9d0c
    • Kai-Heng Feng's avatar
      drm/amd/display: Restore backlight brightness after system resume · 2c60da90
      Kai-Heng Feng authored 
      commit bb264220 upstream.

Laptops with AMD APU doesn't restore display backlight brightness after
system resume.

This issue started when DC was introduced.

Let's use BL_CORE_SUSPENDRESUME so the backlight core calls
update_status callback after system resume to restore the backlight
level.

Tested on Dell Inspiron 3180 (Stoney Ridge) and Dell Latitude 5495
(Raven Ridge).

Cc: <stable@vger.kernel.org>
Signed-off-by: default avatarKai-Heng Feng <kai.heng.feng@canonical.com>
Signed-off-by: default avatarAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      2c60da90
    • Yafang Shao's avatar
      mm/compaction.c: clear total_{migrate,free}_scanned before scanning a new zone · 4d8bdf7f
      Yafang Shao authored 
      [ Upstream commit a94b5252 ]

total_{migrate,free}_scanned will be added to COMPACTMIGRATE_SCANNED and
COMPACTFREE_SCANNED in compact_zone().  We should clear them before
scanning a new zone.  In the proc triggered compaction, we forgot clearing
them.

[laoar.shao@gmail.com: introduce a helper compact_zone_counters_init()]
  Link: http://lkml.kernel.org/r/1563869295-25748-1-git-send-email-laoar.shao@gmail.com
[akpm@linux-foundation.org: expand compact_zone_counters_init() into its single callsite, per mhocko]
[vbabka@suse.cz: squash compact_zone() list_head init as well]
  Link: http://lkml.kernel.org/r/1fb6f7da-f776-9e42-22f8-bbb79b030b98@suse.cz
[akpm@linux-foundation.org: kcompactd_do_work(): avoid unnecessary initialization of cc.zone]
Link: http://lkml.kernel.org/r/1563789275-9639-1-git-send-email-laoar.shao@gmail.com
Fixes: 7f354a54 ("mm, compaction: add vmstats for kcompactd work")
Signed-off-by: default avatarYafang Shao <laoar.shao@gmail.com>
Signed-off-by: default avatarVlastimil Babka <vbabka@suse.cz>
Reviewed-by: default avatarVlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Cc: Yafang Shao <shaoyafang@didiglobal.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: default avatarSasha Levin <sashal@kernel.org>
      4d8bdf7f
    • Eric Biggers's avatar
      fuse: fix deadlock with aio poll and fuse_iqueue::waitq.lock · 5bead06b
      Eric Biggers authored 
      [ Upstream commit 76e43c8c ]

When IOCB_CMD_POLL is used on the FUSE device, aio_poll() disables IRQs
and takes kioctx::ctx_lock, then fuse_iqueue::waitq.lock.

This may have to wait for fuse_iqueue::waitq.lock to be released by one
of many places that take it with IRQs enabled.  Since the IRQ handler
may take kioctx::ctx_lock, lockdep reports that a deadlock is possible.

Fix it by protecting the state of struct fuse_iqueue with a separate
spinlock, and only accessing fuse_iqueue::waitq using the versions of
the waitqueue functions which do IRQ-safe locking internally.

Reproducer:

	#include <fcntl.h>
	#include <stdio.h>
	#include <sys/mount.h>
	#include <sys/stat.h>
	#include <sys/syscall.h>
	#include <unistd.h>
	#include <linux/aio_abi.h>

	int main()
	{
		char opts[128];
		int fd = open("/dev/fuse", O_RDWR);
		aio_context_t ctx = 0;
		struct iocb cb = { .aio_lio_opcode = IOCB_CMD_POLL, .aio_fildes = fd };
		struct iocb *cbp = &cb;

		sprintf(opts, "fd=%d,rootmode=040000,user_id=0,group_id=0", fd);
		mkdir("mnt", 0700);
		mount("foo",  "mnt", "fuse", 0, opts);
		syscall(__NR_io_setup, 1, &ctx);
		syscall(__NR_io_submit, ctx, 1, &cbp);
	}

Beginning of lockdep output:

	=====================================================
	WARNING: SOFTIRQ-safe -> SOFTIRQ-unsafe lock order detected
	5.3.0-rc5 #9 Not tainted
	-----------------------------------------------------
	syz_fuse/135 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire:
	000000003590ceda (&fiq->waitq){+.+.}, at: spin_lock include/linux/spinlock.h:338 [inline]
	000000003590ceda (&fiq->waitq){+.+.}, at: aio_poll fs/aio.c:1751 [inline]
	000000003590ceda (&fiq->waitq){+.+.}, at: __io_submit_one.constprop.0+0x203/0x5b0 fs/aio.c:1825

	and this task is already holding:
	0000000075037284 (&(&ctx->ctx_lock)->rlock){..-.}, at: spin_lock_irq include/linux/spinlock.h:363 [inline]
	0000000075037284 (&(&ctx->ctx_lock)->rlock){..-.}, at: aio_poll fs/aio.c:1749 [inline]
	0000000075037284 (&(&ctx->ctx_lock)->rlock){..-.}, at: __io_submit_one.constprop.0+0x1f4/0x5b0 fs/aio.c:1825
	which would create a new lock dependency:
	 (&(&ctx->ctx_lock)->rlock){..-.} -> (&fiq->waitq){+.+.}

	but this new dependency connects a SOFTIRQ-irq-safe lock:
	 (&(&ctx->ctx_lock)->rlock){..-.}

	[...]

Reported-by: syzbot+af05535bb79520f95431@syzkaller.appspotmail.com
Reported-by: syzbot+d86c4426a01f60feddc7@syzkaller.appspotmail.com
Fixes: bfe4037e ("aio: implement IOCB_CMD_POLL")
Cc: <stable@vger.kernel.org> # v4.19+
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: default avatarEric Biggers <ebiggers@google.com>
Signed-off-by: default avatarMiklos Szeredi <mszeredi@redhat.com>
Signed-off-by: default avatarSasha Levin <sashal@kernel.org>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      5bead06b
    • NeilBrown's avatar
      md/raid0: avoid RAID0 data corruption due to layout confusion. · bbe3e205
      NeilBrown authored 
      [ Upstream commit c84a1372 ]

If the drives in a RAID0 are not all the same size, the array is
divided into zones.
The first zone covers all drives, to the size of the smallest.
The second zone covers all drives larger than the smallest, up to
the size of the second smallest - etc.

A change in Linux 3.14 unintentionally changed the layout for the
second and subsequent zones.  All the correct data is still stored, but
each chunk may be assigned to a different device than in pre-3.14 kernels.
This can lead to data corruption.

It is not possible to determine what layout to use - it depends which
kernel the data was written by.
So we add a module parameter to allow the old (0) or new (1) layout to be
specified, and refused to assemble an affected array if that parameter is
not set.

Fixes: 20d0189b ("block: Introduce new bio_split()")
cc: stable@vger.kernel.org (3.14+)
Acked-by: default avatarGuoqing Jiang <guoqing.jiang@cloud.ionos.com>
Signed-off-by: default avatarNeilBrown <neilb@suse.de>
Signed-off-by: default avatarSong Liu <songliubraving@fb.com>
Signed-off-by: default avatarSasha Levin <sashal@kernel.org>
      bbe3e205
    • Pavel Shilovsky's avatar
      CIFS: Fix oplock handling for SMB 2.1+ protocols · 4290a9e5
      Pavel Shilovsky authored 
      commit a016e279 upstream.

There may be situations when a server negotiates SMB 2.1
protocol version or higher but responds to a CREATE request
with an oplock rather than a lease.

Currently the client doesn't handle such a case correctly:
when another CREATE comes in the server sends an oplock
break to the initial CREATE and the client doesn't send
an ack back due to a wrong caching level being set (READ
instead of RWH). Missing an oplock break ack makes the
server wait until the break times out which dramatically
increases the latency of the second CREATE.

Fix this by properly detecting oplocks when using SMB 2.1
protocol version and higher.

Cc: <stable@vger.kernel.org>
Signed-off-by: default avatarPavel Shilovsky <pshilov@microsoft.com>
Signed-off-by: default avatarSteve French <stfrench@microsoft.com>
Reviewed-by: default avatarRonnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      4290a9e5
    • Murphy Zhou's avatar
      CIFS: fix max ea value size · a3a15089
      Murphy Zhou authored 
      commit 63d37fb4 upstream.

It should not be larger then the slab max buf size. If user
specifies a larger size, it passes this check and goes
straightly to SMB2_set_info_init performing an insecure memcpy.
Signed-off-by: default avatarMurphy Zhou <jencce.kernel@gmail.com>
Reviewed-by: default avatarAurelien Aptel <aaptel@suse.com>
CC: Stable <stable@vger.kernel.org>
Signed-off-by: default avatarSteve French <stfrench@microsoft.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      a3a15089
    • Chris Brandt's avatar
      i2c: riic: Clear NACK in tend isr · a0f7fd38
      Chris Brandt authored 
      commit a71e2ac1 upstream.

The NACKF flag should be cleared in INTRIICNAKI interrupt processing as
description in HW manual.

This issue shows up quickly when PREEMPT_RT is applied and a device is
probed that is not plugged in (like a touchscreen controller). The result
is endless interrupts that halt system boot.

Fixes: 310c18a4 ("i2c: riic: add driver")
Cc: stable@vger.kernel.org
Reported-by: default avatarChien Nguyen <chien.nguyen.eb@rvc.renesas.com>
Signed-off-by: default avatarChris Brandt <chris.brandt@renesas.com>
Signed-off-by: default avatarWolfram Sang <wsa@the-dreams.de>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      a0f7fd38
    • Laurent Vivier's avatar
      hwrng: core - don't wait on add_early_randomness() · fec38267
      Laurent Vivier authored 
      commit 78887832 upstream.

add_early_randomness() is called by hwrng_register() when the
hardware is added. If this hardware and its module are present
at boot, and if there is no data available the boot hangs until
data are available and can't be interrupted.

For instance, in the case of virtio-rng, in some cases the host can be
not able to provide enough entropy for all the guests.

We can have two easy ways to reproduce the problem but they rely on
misconfiguration of the hypervisor or the egd daemon:

- if virtio-rng device is configured to connect to the egd daemon of the
host but when the virtio-rng driver asks for data the daemon is not
connected,

- if virtio-rng device is configured to connect to the egd daemon of the
host but the egd daemon doesn't provide data.

The guest kernel will hang at boot until the virtio-rng driver provides
enough data.

To avoid that, call rng_get_data() in non-blocking mode (wait=0)
from add_early_randomness().
Signed-off-by: default avatarLaurent Vivier <lvivier@redhat.com>
Fixes: d9e79726 ("hwrng: add randomness to system from rng...")
Cc: <stable@vger.kernel.org>
Reviewed-by: default avatarTheodore Ts'o <tytso@mit.edu>
Signed-off-by: default avatarHerbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      fec38267
    • Chao Yu's avatar
      quota: fix wrong condition in is_quota_modification() · 06098609
      Chao Yu authored 
      commit 6565c182 upstream.

Quoted from
commit 3da40c7b ("ext4: only call ext4_truncate when size <= isize")

" At LSF we decided that if we truncate up from isize we shouldn't trim
  fallocated blocks that were fallocated with KEEP_SIZE and are past the
 new i_size.  This patch fixes ext4 to do this. "

And generic/092 of fstest have covered this case for long time, however
is_quota_modification() didn't adjust based on that rule, so that in
below condition, we will lose to quota block change:
- fallocate blocks beyond EOF
- remount
- truncate(file_path, file_size)

Fix it.

Link: https://lore.kernel.org/r/20190911093650.35329-1-yuchao0@huawei.com
Fixes: 3da40c7b ("ext4: only call ext4_truncate when size <= isize")
CC: stable@vger.kernel.org
Signed-off-by: default avatarChao Yu <yuchao0@huawei.com>
Signed-off-by: default avatarJan Kara <jack@suse.cz>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      06098609
    • Theodore Ts'o's avatar
      ext4: fix punch hole for inline_data file systems · 091c754d
      Theodore Ts'o authored 
      commit c1e8220b upstream.

If a program attempts to punch a hole on an inline data file, we need
to convert it to a normal file first.

This was detected using ext4/032 using the adv configuration.  Simple
reproducer:

mke2fs -Fq -t ext4 -O inline_data /dev/vdc
mount /vdc
echo "" > /vdc/testfile
xfs_io -c 'truncate 33554432' /vdc/testfile
xfs_io -c 'fpunch 0 1048576' /vdc/testfile
umount /vdc
e2fsck -fy /dev/vdc

Cc: stable@vger.kernel.org
Signed-off-by: default avatarTheodore Ts'o <tytso@mit.edu>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      091c754d
    • Rakesh Pandit's avatar
      ext4: fix warning inside ext4_convert_unwritten_extents_endio · 775e3e73
      Rakesh Pandit authored 
      commit e3d550c2 upstream.

Really enable warning when CONFIG_EXT4_DEBUG is set and fix missing
first argument.  This was introduced in commit ff95ec22 ("ext4:
add warning to ext4_convert_unwritten_extents_endio") and splitting
extents inside endio would trigger it.

Fixes: ff95ec22 ("ext4: add warning to ext4_convert_unwritten_extents_endio")
Signed-off-by: default avatarRakesh Pandit <rakesh@tuxera.com>
Signed-off-by: default avatarTheodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      775e3e73
    • Tetsuo Handa's avatar
      /dev/mem: Bail out upon SIGKILL. · caa6926d
      Tetsuo Handa authored 
      commit 8619e5bd upstream.

syzbot found that a thread can stall for minutes inside read_mem() or
write_mem() after that thread was killed by SIGKILL [1]. Reading from
iomem areas of /dev/mem can be slow, depending on the hardware.
While reading 2GB at one read() is legal, delaying termination of killed
thread for minutes is bad. Thus, allow reading/writing /dev/mem and
/dev/kmem to be preemptible and killable.

  [ 1335.912419][T20577] read_mem: sz=4096 count=2134565632
  [ 1335.943194][T20577] read_mem: sz=4096 count=2134561536
  [ 1335.978280][T20577] read_mem: sz=4096 count=2134557440
  [ 1336.011147][T20577] read_mem: sz=4096 count=2134553344
  [ 1336.041897][T20577] read_mem: sz=4096 count=2134549248

Theoretically, reading/writing /dev/mem and /dev/kmem can become
"interruptible". But this patch chose "killable". Future patch will make
them "interruptible" so that we can revert to "killable" if some program
regressed.

[1] https://syzkaller.appspot.com/bug?id=a0e3436829698d5824231251fad9d8e998f94f5eSigned-off-by: default avatarTetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: stable <stable@vger.kernel.org>
Reported-by: default avatarsyzbot <syzbot+8ab2d0f39fb79fe6ca40@syzkaller.appspotmail.com>
Link: https://lore.kernel.org/r/1566825205-10703-1-git-send-email-penguin-kernel@I-love.SAKURA.ne.jpSigned-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      caa6926d
    • Denis Kenzior's avatar
      cfg80211: Purge frame registrations on iftype change · bd3a11af
      Denis Kenzior authored 
      commit c1d3ad84 upstream.

Currently frame registrations are not purged, even when changing the
interface type.  This can lead to potentially weird situations where
frames possibly not allowed on a given interface type remain registered
due to the type switching happening after registration.

The kernel currently relies on userspace apps to actually purge the
registrations themselves, this is not something that the kernel should
rely on.

Add a call to cfg80211_mlme_purge_registrations() to forcefully remove
any registrations left over prior to switching the iftype.

Cc: stable@vger.kernel.org
Signed-off-by: default avatarDenis Kenzior <denkenz@gmail.com>
Link: https://lore.kernel.org/r/20190828211110.15005-1-denkenz@gmail.comSigned-off-by: default avatarJohannes Berg <johannes.berg@intel.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      bd3a11af
    • NeilBrown's avatar
      md: only call set_in_sync() when it is expected to succeed. · 5dc86e95
      NeilBrown authored 
      commit 480523fe upstream.

Since commit 4ad23a97 ("MD: use per-cpu counter for
writes_pending"), set_in_sync() is substantially more expensive: it
can wait for a full RCU grace period which can be 10s of milliseconds.

So we should only call it when the cost is justified.

md_check_recovery() currently calls set_in_sync() every time it finds
anything to do (on non-external active arrays).  For an array
performing resync or recovery, this will be quite often.
Each call will introduce a delay to the md thread, which can noticeable
affect IO submission latency.

In md_check_recovery() we only need to call set_in_sync() if
'safemode' was non-zero at entry, meaning that there has been not
recent IO.  So we save this "safemode was nonzero" state, and only
call set_in_sync() if it was non-zero.

This measurably reduces mean and maximum IO submission latency during
resync/recovery.
Reported-and-tested-by: default avatarJack Wang <jinpu.wang@cloud.ionos.com>
Fixes: 4ad23a97 ("MD: use per-cpu counter for writes_pending")
Cc: stable@vger.kernel.org (v4.12+)
Signed-off-by: default avatarNeilBrown <neilb@suse.com>
Signed-off-by: default avatarSong Liu <songliubraving@fb.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      5dc86e95
    • NeilBrown's avatar
      md: don't report active array_state until after revalidate_disk() completes. · 598a2cda
      NeilBrown authored 
      commit 9d4b45d6 upstream.

Until revalidate_disk() has completed, the size of a new md array will
appear to be zero.
So we shouldn't report, through array_state, that the array is active
until that time.
udev rules check array_state to see if the array is ready.  As soon as
it appear to be zero, fsck can be run.  If it find the size to be
zero, it will fail.

So add a new flag to provide an interlock between do_md_run() and
array_state_show().  This flag is set while do_md_run() is active and
it prevents array_state_show() from reporting that the array is
active.

Before do_md_run() is called, ->pers will be NULL so array is
definitely not active.
After do_md_run() is called, revalidate_disk() will have run and the
array will be completely ready.

We also move various sysfs_notify*() calls out of md_run() into
do_md_run() after MD_NOT_READY is cleared.  This ensure the
information is ready before the notification is sent.

Prior to v4.12, array_state_show() was called with the
mddev->reconfig_mutex held, which provided exclusion with do_md_run().

Note that MD_NOT_READY cleared twice.  This is deliberate to cover
both success and error paths with minimal noise.

Fixes: b7b17c9b ("md: remove mddev_lock() from md_attr_show()")
Cc: stable@vger.kernel.org (v4.12++)
Signed-off-by: default avatarNeilBrown <neilb@suse.com>
Signed-off-by: default avatarSong Liu <songliubraving@fb.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      598a2cda
    • Xiao Ni's avatar
      md/raid6: Set R5_ReadError when there is read failure on parity disk · e8323e0d
      Xiao Ni authored 
      commit 143f6e73 upstream.

7471fb77 ("md/raid6: Fix anomily when recovering a single device in
RAID6.") avoids rereading P when it can be computed from other members.
However, this misses the chance to re-write the right data to P. This
patch sets R5_ReadError if the re-read fails.

Also, when re-read is skipped, we also missed the chance to reset
rdev->read_errors to 0. It can fail the disk when there are many read
errors on P member disk (other disks don't have read error)

V2: upper layer read request don't read parity/Q data. So there is no
need to consider such situation.

This is Reported-by: kbuild test robot <lkp@intel.com>

Fixes: 7471fb77 ("md/raid6: Fix anomily when recovering a single device in RAID6.")
Cc: <stable@vger.kernel.org> #4.4+
Signed-off-by: default avatarXiao Ni <xni@redhat.com>
Signed-off-by: default avatarSong Liu <songliubraving@fb.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      e8323e0d
    • Filipe Manana's avatar
      Btrfs: fix race setting up and completing qgroup rescan workers · bacff03b
      Filipe Manana authored 
      commit 13fc1d27 upstream.

There is a race between setting up a qgroup rescan worker and completing
a qgroup rescan worker that can lead to callers of the qgroup rescan wait
ioctl to either not wait for the rescan worker to complete or to hang
forever due to missing wake ups. The following diagram shows a sequence
of steps that illustrates the race.

        CPU 1                                                         CPU 2                                  CPU 3

 btrfs_ioctl_quota_rescan()
  btrfs_qgroup_rescan()
   qgroup_rescan_init()
    mutex_lock(&fs_info->qgroup_rescan_lock)
    spin_lock(&fs_info->qgroup_lock)

    fs_info->qgroup_flags |=
      BTRFS_QGROUP_STATUS_FLAG_RESCAN

    init_completion(
      &fs_info->qgroup_rescan_completion)

    fs_info->qgroup_rescan_running = true

    mutex_unlock(&fs_info->qgroup_rescan_lock)
    spin_unlock(&fs_info->qgroup_lock)

    btrfs_init_work()
     --> starts the worker

                                                        btrfs_qgroup_rescan_worker()
                                                         mutex_lock(&fs_info->qgroup_rescan_lock)

                                                         fs_info->qgroup_flags &=
                                                           ~BTRFS_QGROUP_STATUS_FLAG_RESCAN

                                                         mutex_unlock(&fs_info->qgroup_rescan_lock)

                                                         starts transaction, updates qgroup status
                                                         item, etc

                                                                                                           btrfs_ioctl_quota_rescan()
                                                                                                            btrfs_qgroup_rescan()
                                                                                                             qgroup_rescan_init()
                                                                                                              mutex_lock(&fs_info->qgroup_rescan_lock)
                                                                                                              spin_lock(&fs_info->qgroup_lock)

                                                                                                              fs_info->qgroup_flags |=
                                                                                                                BTRFS_QGROUP_STATUS_FLAG_RESCAN

                                                                                                              init_completion(
                                                                                                                &fs_info->qgroup_rescan_completion)

                                                                                                              fs_info->qgroup_rescan_running = true

                                                                                                              mutex_unlock(&fs_info->qgroup_rescan_lock)
                                                                                                              spin_unlock(&fs_info->qgroup_lock)

                                                                                                              btrfs_init_work()
                                                                                                               --> starts another worker

                                                         mutex_lock(&fs_info->qgroup_rescan_lock)

                                                         fs_info->qgroup_rescan_running = false

                                                         mutex_unlock(&fs_info->qgroup_rescan_lock)

							 complete_all(&fs_info->qgroup_rescan_completion)

Before the rescan worker started by the task at CPU 3 completes, if
another task calls btrfs_ioctl_quota_rescan(), it will get -EINPROGRESS
because the flag BTRFS_QGROUP_STATUS_FLAG_RESCAN is set at
fs_info->qgroup_flags, which is expected and correct behaviour.

However if other task calls btrfs_ioctl_quota_rescan_wait() before the
rescan worker started by the task at CPU 3 completes, it will return
immediately without waiting for the new rescan worker to complete,
because fs_info->qgroup_rescan_running is set to false by CPU 2.

This race is making test case btrfs/171 (from fstests) to fail often:

  btrfs/171 9s ... - output mismatch (see /home/fdmanana/git/hub/xfstests/results//btrfs/171.out.bad)
#      --- tests/btrfs/171.out     2018-09-16 21:30:48.505104287 +0100
#      +++ /home/fdmanana/git/hub/xfstests/results//btrfs/171.out.bad      2019-09-19 02:01:36.938486039 +0100
#      @@ -1,2 +1,3 @@
#       QA output created by 171
#      +ERROR: quota rescan failed: Operation now in progress
#       Silence is golden
#      ...
#      (Run 'diff -u /home/fdmanana/git/hub/xfstests/tests/btrfs/171.out /home/fdmanana/git/hub/xfstests/results//btrfs/171.out.bad'  to see the entire diff)

That is because the test calls the btrfs-progs commands "qgroup quota
rescan -w", "qgroup assign" and "qgroup remove" in a sequence that makes
calls to the rescan start ioctl fail with -EINPROGRESS (note the "btrfs"
commands 'qgroup assign' and 'qgroup remove' often call the rescan start
ioctl after calling the qgroup assign ioctl,
btrfs_ioctl_qgroup_assign()), since previous waits didn't actually wait
for a rescan worker to complete.

Another problem the race can cause is missing wake ups for waiters,
since the call to complete_all() happens outside a critical section and
after clearing the flag BTRFS_QGROUP_STATUS_FLAG_RESCAN. In the sequence
diagram above, if we have a waiter for the first rescan task (executed
by CPU 2), then fs_info->qgroup_rescan_completion.wait is not empty, and
if after the rescan worker clears BTRFS_QGROUP_STATUS_FLAG_RESCAN and
before it calls complete_all() against
fs_info->qgroup_rescan_completion, the task at CPU 3 calls
init_completion() against fs_info->qgroup_rescan_completion which
re-initilizes its wait queue to an empty queue, therefore causing the
rescan worker at CPU 2 to call complete_all() against an empty queue,
never waking up the task waiting for that rescan worker.

Fix this by clearing BTRFS_QGROUP_STATUS_FLAG_RESCAN and setting
fs_info->qgroup_rescan_running to false in the same critical section,
delimited by the mutex fs_info->qgroup_rescan_lock, as well as doing the
call to complete_all() in that same critical section. This gives the
protection needed to avoid rescan wait ioctl callers not waiting for a
running rescan worker and the lost wake ups problem, since setting that
rescan flag and boolean as well as initializing the wait queue is done
already in a critical section delimited by that mutex (at
qgroup_rescan_init()).

Fixes: 57254b6e ("Btrfs: add ioctl to wait for qgroup rescan completion")
Fixes: d2c609b8 ("btrfs: properly track when rescan worker is running")
CC: stable@vger.kernel.org # 4.4+
Reviewed-by: default avatarJosef Bacik <josef@toxicpanda.com>
Signed-off-by: default avatarFilipe Manana <fdmanana@suse.com>
Signed-off-by: default avatarDavid Sterba <dsterba@suse.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      bacff03b
    • Qu Wenruo's avatar
      btrfs: qgroup: Fix reserved data space leak if we have multiple reserve calls · b5c42ef0
      Qu Wenruo authored 
      commit d4e20494 upstream.

[BUG]
The following script can cause btrfs qgroup data space leak:

  mkfs.btrfs -f $dev
  mount $dev -o nospace_cache $mnt

  btrfs subv create $mnt/subv
  btrfs quota en $mnt
  btrfs quota rescan -w $mnt
  btrfs qgroup limit 128m $mnt/subv

  for (( i = 0; i < 3; i++)); do
          # Create 3 64M holes for latter fallocate to fail
          truncate -s 192m $mnt/subv/file
          xfs_io -c "pwrite 64m 4k" $mnt/subv/file > /dev/null
          xfs_io -c "pwrite 128m 4k" $mnt/subv/file > /dev/null
          sync

          # it's supposed to fail, and each failure will leak at least 64M
          # data space
          xfs_io -f -c "falloc 0 192m" $mnt/subv/file &> /dev/null
          rm $mnt/subv/file
          sync
  done

  # Shouldn't fail after we removed the file
  xfs_io -f -c "falloc 0 64m" $mnt/subv/file

[CAUSE]
Btrfs qgroup data reserve code allow multiple reservations to happen on
a single extent_changeset:
E.g:
	btrfs_qgroup_reserve_data(inode, &data_reserved, 0, SZ_1M);
	btrfs_qgroup_reserve_data(inode, &data_reserved, SZ_1M, SZ_2M);
	btrfs_qgroup_reserve_data(inode, &data_reserved, 0, SZ_4M);

Btrfs qgroup code has its internal tracking to make sure we don't
double-reserve in above example.

The only pattern utilizing this feature is in the main while loop of
btrfs_fallocate() function.

However btrfs_qgroup_reserve_data()'s error handling has a bug in that
on error it clears all ranges in the io_tree with EXTENT_QGROUP_RESERVED
flag but doesn't free previously reserved bytes.

This bug has a two fold effect:
- Clearing EXTENT_QGROUP_RESERVED ranges
  This is the correct behavior, but it prevents
  btrfs_qgroup_check_reserved_leak() to catch the leakage as the
  detector is purely EXTENT_QGROUP_RESERVED flag based.

- Leak the previously reserved data bytes.

The bug manifests when N calls to btrfs_qgroup_reserve_data are made and
the last one fails, leaking space reserved in the previous ones.

[FIX]
Also free previously reserved data bytes when btrfs_qgroup_reserve_data
fails.

Fixes: 52472553 ("btrfs: qgroup: Introduce btrfs_qgroup_reserve_data function")
CC: stable@vger.kernel.org # 4.4+
Signed-off-by: default avatarQu Wenruo <wqu@suse.com>
Signed-off-by: default avatarDavid Sterba <dsterba@suse.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      b5c42ef0
    • Qu Wenruo's avatar
      btrfs: qgroup: Fix the wrong target io_tree when freeing reserved data space · c521bfa8
      Qu Wenruo authored 
      commit bab32fc0 upstream.

[BUG]
Under the following case with qgroup enabled, if some error happened
after we have reserved delalloc space, then in error handling path, we
could cause qgroup data space leakage:

From btrfs_truncate_block() in inode.c:

	ret = btrfs_delalloc_reserve_space(inode, &data_reserved,
					   block_start, blocksize);
	if (ret)
		goto out;

 again:
	page = find_or_create_page(mapping, index, mask);
	if (!page) {
		btrfs_delalloc_release_space(inode, data_reserved,
					     block_start, blocksize, true);
		btrfs_delalloc_release_extents(BTRFS_I(inode), blocksize, true);
		ret = -ENOMEM;
		goto out;
	}

[CAUSE]
In the above case, btrfs_delalloc_reserve_space() will call
btrfs_qgroup_reserve_data() and mark the io_tree range with
EXTENT_QGROUP_RESERVED flag.

In the error handling path, we have the following call stack:
btrfs_delalloc_release_space()
|- btrfs_free_reserved_data_space()
   |- btrsf_qgroup_free_data()
      |- __btrfs_qgroup_release_data(reserved=@reserved, free=1)
         |- qgroup_free_reserved_data(reserved=@reserved)
            |- clear_record_extent_bits();
            |- freed += changeset.bytes_changed;

However due to a completion bug, qgroup_free_reserved_data() will clear
EXTENT_QGROUP_RESERVED flag in BTRFS_I(inode)->io_failure_tree, other
than the correct BTRFS_I(inode)->io_tree.
Since io_failure_tree is never marked with that flag,
btrfs_qgroup_free_data() will not free any data reserved space at all,
causing a leakage.

This type of error handling can only be triggered by errors outside of
qgroup code. So EDQUOT error from qgroup can't trigger it.

[FIX]
Fix the wrong target io_tree.
Reported-by: default avatarJosef Bacik <josef@toxicpanda.com>
Fixes: bc42bda2 ("btrfs: qgroup: Fix qgroup reserved space underflow by only freeing reserved ranges")
CC: stable@vger.kernel.org # 4.14+
Reviewed-by: default avatarNikolay Borisov <nborisov@suse.com>
Signed-off-by: default avatarQu Wenruo <wqu@suse.com>
Signed-off-by: default avatarDavid Sterba <dsterba@suse.com>
Signed-off-by: default avatarGreg Kroah-Hartman <gregkh@linuxfoundation.org>
      c521bfa8
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7