linux-stable.git/drivers/md, branch linux-5.19.y

md/raid5: Wait for MD_SB_CHANGE_PENDING in raid5d

2022-10-24T07:58:26+00:00

[ Upstream commit 5e2cf333b7bd5d3e62595a44d598a254c697cd74 ]

A complicated deadlock exists when using the journal and an elevated
group_thrtead_cnt. It was found with loop devices, but its not clear
whether it can be seen with real disks. The deadlock can occur simply
by writing data with an fio script.

When the deadlock occurs, multiple threads will hang in different ways:

 1) The group threads will hang in the blk-wbt code with bios waiting to
    be submitted to the block layer:

        io_schedule+0x70/0xb0
        rq_qos_wait+0x153/0x210
        wbt_wait+0x115/0x1b0
        io_schedule+0x70/0xb0
        rq_qos_wait+0x153/0x210
        wbt_wait+0x115/0x1b0
        __rq_qos_throttle+0x38/0x60
        blk_mq_submit_bio+0x589/0xcd0
        wbt_wait+0x115/0x1b0
        __rq_qos_throttle+0x38/0x60
        blk_mq_submit_bio+0x589/0xcd0
        __submit_bio+0xe6/0x100
        submit_bio_noacct_nocheck+0x42e/0x470
        submit_bio_noacct+0x4c2/0xbb0
        ops_run_io+0x46b/0x1a30
        handle_stripe+0xcd3/0x36b0
        handle_active_stripes.constprop.0+0x6f6/0xa60
        raid5_do_work+0x177/0x330

    Or:
        io_schedule+0x70/0xb0
        rq_qos_wait+0x153/0x210
        wbt_wait+0x115/0x1b0
        __rq_qos_throttle+0x38/0x60
        blk_mq_submit_bio+0x589/0xcd0
        __submit_bio+0xe6/0x100
        submit_bio_noacct_nocheck+0x42e/0x470
        submit_bio_noacct+0x4c2/0xbb0
        flush_deferred_bios+0x136/0x170
        raid5_do_work+0x262/0x330

 2) The r5l_reclaim thread will hang in the same way, submitting a
    bio to the block layer:

        io_schedule+0x70/0xb0
        rq_qos_wait+0x153/0x210
        wbt_wait+0x115/0x1b0
        __rq_qos_throttle+0x38/0x60
        blk_mq_submit_bio+0x589/0xcd0
        __submit_bio+0xe6/0x100
        submit_bio_noacct_nocheck+0x42e/0x470
        submit_bio_noacct+0x4c2/0xbb0
        submit_bio+0x3f/0xf0
        md_super_write+0x12f/0x1b0
        md_update_sb.part.0+0x7c6/0xff0
        md_update_sb+0x30/0x60
        r5l_do_reclaim+0x4f9/0x5e0
        r5l_reclaim_thread+0x69/0x30b

    However, before hanging, the MD_SB_CHANGE_PENDING flag will be
    set for sb_flags in r5l_write_super_and_discard_space(). This
    flag will never be cleared because the submit_bio() call never
    returns.

 3) Due to the MD_SB_CHANGE_PENDING flag being set, handle_stripe()
    will do no processing on any pending stripes and re-set
    STRIPE_HANDLE. This will cause the raid5d thread to enter an
    infinite loop, constantly trying to handle the same stripes
    stuck in the queue.

    The raid5d thread has a blk_plug that holds a number of bios
    that are also stuck waiting seeing the thread is in a loop
    that never schedules. These bios have been accounted for by
    blk-wbt thus preventing the other threads above from
    continuing when they try to submit bios. --Deadlock.

To fix this, add the same wait_event() that is used in raid5_do_work()
to raid5d() such that if MD_SB_CHANGE_PENDING is set, the thread will
schedule and wait until the flag is cleared. The schedule action will
flush the plug which will allow the r5l_reclaim thread to continue,
thus preventing the deadlock.

However, md_check_recovery() calls can also clear MD_SB_CHANGE_PENDING
from the same thread and can thus deadlock if the thread is put to
sleep. So avoid waiting if md_check_recovery() is being called in the
loop.

It's not clear when the deadlock was introduced, but the similar
wait_event() call in raid5_do_work() was added in 2017 by this
commit:

    16d997b78b15 ("md/raid5: simplfy delaying of writes while metadata
                   is updated.")

Link: https://lore.kernel.org/r/7f3b87b6-b52a-f737-51d7-a4eec5c44112@deltatee.com
Signed-off-by: Logan Gunthorpe 
Signed-off-by: Song Liu 
Signed-off-by: Sasha Levin

bcache: fix set_at_max_writeback_rate() for multiple attached devices

2022-10-24T07:58:25+00:00

[ Upstream commit d2d05b88035d2d51a5bb6c5afec88a0880c73df4 ]

Inside set_at_max_writeback_rate() the calculation in following if()
check is wrong,
	if (atomic_inc_return(&c->idle_counter) <
	    atomic_read(&c->attached_dev_nr) * 6)

Because each attached backing device has its own writeback thread
running and increasing c->idle_counter, the counter increates much
faster than expected. The correct calculation should be,
	(counter / dev_nr) < dev_nr * 6
which equals to,
	counter < dev_nr * dev_nr * 6

This patch fixes the above mistake with correct calculation, and helper
routine idle_counter_exceeded() is added to make code be more clear.

Reported-by: Mingzhe Zou 
Signed-off-by: Coly Li 
Acked-by: Mingzhe Zou 
Link: https://lore.kernel.org/r/20220919161647.81238-6-colyli@suse.de
Signed-off-by: Jens Axboe 
Signed-off-by: Sasha Levin

md/raid5: Remove unnecessary bio_put() in raid5_read_one_chunk()

2022-10-24T07:57:55+00:00

[ Upstream commit c66a6f41e09ad386fd2cce22b9cded837bbbc704 ]

When running chunk-sized reads on disks with badblocks duplicate bio
free/puts are observed:

   =============================================================================
   BUG bio-200 (Not tainted): Object already free
   -----------------------------------------------------------------------------
   Allocated in mempool_alloc_slab+0x17/0x20 age=3 cpu=2 pid=7504
    __slab_alloc.constprop.0+0x5a/0xb0
    kmem_cache_alloc+0x31e/0x330
    mempool_alloc_slab+0x17/0x20
    mempool_alloc+0x100/0x2b0
    bio_alloc_bioset+0x181/0x460
    do_mpage_readpage+0x776/0xd00
    mpage_readahead+0x166/0x320
    blkdev_readahead+0x15/0x20
    read_pages+0x13f/0x5f0
    page_cache_ra_unbounded+0x18d/0x220
    force_page_cache_ra+0x181/0x1c0
    page_cache_sync_ra+0x65/0xb0
    filemap_get_pages+0x1df/0xaf0
    filemap_read+0x1e1/0x700
    blkdev_read_iter+0x1e5/0x330
    vfs_read+0x42a/0x570
   Freed in mempool_free_slab+0x17/0x20 age=3 cpu=2 pid=7504
    kmem_cache_free+0x46d/0x490
    mempool_free_slab+0x17/0x20
    mempool_free+0x66/0x190
    bio_free+0x78/0x90
    bio_put+0x100/0x1a0
    raid5_make_request+0x2259/0x2450
    md_handle_request+0x402/0x600
    md_submit_bio+0xd9/0x120
    __submit_bio+0x11f/0x1b0
    submit_bio_noacct_nocheck+0x204/0x480
    submit_bio_noacct+0x32e/0xc70
    submit_bio+0x98/0x1a0
    mpage_readahead+0x250/0x320
    blkdev_readahead+0x15/0x20
    read_pages+0x13f/0x5f0
    page_cache_ra_unbounded+0x18d/0x220
   Slab 0xffffea000481b600 objects=21 used=0 fp=0xffff8881206d8940 flags=0x17ffffc0010201(locked|slab|head|node=0|zone=2|lastcpupid=0x1fffff)
   CPU: 0 PID: 34525 Comm: kworker/u24:2 Not tainted 6.0.0-rc2-localyes-265166-gf11c5343fa3f #143
   Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.13.0-1ubuntu1.1 04/01/2014
   Workqueue: raid5wq raid5_do_work
   Call Trace:
    
    dump_stack_lvl+0x5a/0x78
    dump_stack+0x10/0x16
    print_trailer+0x158/0x165
    object_err+0x35/0x50
    free_debug_processing.cold+0xb7/0xbe
    __slab_free+0x1ae/0x330
    kmem_cache_free+0x46d/0x490
    mempool_free_slab+0x17/0x20
    mempool_free+0x66/0x190
    bio_free+0x78/0x90
    bio_put+0x100/0x1a0
    mpage_end_io+0x36/0x150
    bio_endio+0x2fd/0x360
    md_end_io_acct+0x7e/0x90
    bio_endio+0x2fd/0x360
    handle_failed_stripe+0x960/0xb80
    handle_stripe+0x1348/0x3760
    handle_active_stripes.constprop.0+0x72a/0xaf0
    raid5_do_work+0x177/0x330
    process_one_work+0x616/0xb20
    worker_thread+0x2bd/0x6f0
    kthread+0x179/0x1b0
    ret_from_fork+0x22/0x30
    

The double free is caused by an unnecessary bio_put() in the
if(is_badblock(...)) error path in raid5_read_one_chunk().

The error path was moved ahead of bio_alloc_clone() in c82aa1b76787c
("md/raid5: move checking badblock before clone bio in
raid5_read_one_chunk"). The previous code checked and freed align_bio
which required a bio_put. After the move that is no longer needed as
raid_bio is returned to the control of the common io path which
performs its own endio resulting in a double free on bad device blocks.

Fixes: c82aa1b76787c ("md/raid5: move checking badblock before clone bio in raid5_read_one_chunk")
Signed-off-by: David Sloan 
Signed-off-by: Logan Gunthorpe 
Reviewed-by: Christoph Hellwig 
Acked-by: Guoqing Jiang 
Signed-off-by: Song Liu 
Signed-off-by: Sasha Levin

md/raid5: Ensure stripe_fill happens on non-read IO with journal

2022-10-24T07:57:55+00:00

[ Upstream commit e2eed85bc75138a9eeb63863d20f8904ac42a577 ]

When doing degrade/recover tests using the journal a kernel BUG
is hit at drivers/md/raid5.c:4381 in handle_parity_checks5():

  BUG_ON(!test_bit(R5_UPTODATE, &dev->flags));

This was found to occur because handle_stripe_fill() was skipped
for stripes in the journal due to a condition in that function.
Thus blocks were not fetched and R5_UPTODATE was not set when
the code reached handle_parity_checks5().

To fix this, don't skip handle_stripe_fill() unless the stripe is
for read.

Fixes: 07e83364845e ("md/r5cache: shift complex rmw from read path to write path")
Link: https://lore.kernel.org/linux-raid/e05c4239-41a9-d2f7-3cfa-4aa9d2cea8c1@deltatee.com/
Suggested-by: Song Liu 
Signed-off-by: Logan Gunthorpe 
Signed-off-by: Song Liu 
Signed-off-by: Sasha Levin

md: Replace snprintf with scnprintf

2022-10-24T07:57:55+00:00

[ Upstream commit 1727fd5015d8f93474148f94e34cda5aa6ad4a43 ]

Current code produces a warning as shown below when total characters
in the constituent block device names plus the slashes exceeds 200.
snprintf() returns the number of characters generated from the given
input, which could cause the expression “200 – len” to wrap around
to a large positive number. Fix this by using scnprintf() instead,
which returns the actual number of characters written into the buffer.

[ 1513.267938] ------------[ cut here ]------------
[ 1513.267943] WARNING: CPU: 15 PID: 37247 at /lib/vsprintf.c:2509 vsnprintf+0x2c8/0x510
[ 1513.267944] Modules linked in:  
[ 1513.267969] CPU: 15 PID: 37247 Comm: mdadm Not tainted 5.4.0-1085-azure #90~18.04.1-Ubuntu
[ 1513.267969] Hardware name: Microsoft Corporation Virtual Machine/Virtual Machine, BIOS Hyper-V UEFI Release v4.1 05/09/2022
[ 1513.267971] RIP: 0010:vsnprintf+0x2c8/0x510
<-snip->
[ 1513.267982] Call Trace:
[ 1513.267986]  snprintf+0x45/0x70
[ 1513.267990]  ? disk_name+0x71/0xa0
[ 1513.267993]  dump_zones+0x114/0x240 [raid0]
[ 1513.267996]  ? _cond_resched+0x19/0x40
[ 1513.267998]  raid0_run+0x19e/0x270 [raid0]
[ 1513.268000]  md_run+0x5e0/0xc50
[ 1513.268003]  ? security_capable+0x3f/0x60
[ 1513.268005]  do_md_run+0x19/0x110
[ 1513.268006]  md_ioctl+0x195e/0x1f90
[ 1513.268007]  blkdev_ioctl+0x91f/0x9f0
[ 1513.268010]  block_ioctl+0x3d/0x50
[ 1513.268012]  do_vfs_ioctl+0xa9/0x640
[ 1513.268014]  ? __fput+0x162/0x260
[ 1513.268016]  ksys_ioctl+0x75/0x80
[ 1513.268017]  __x64_sys_ioctl+0x1a/0x20
[ 1513.268019]  do_syscall_64+0x5e/0x200
[ 1513.268021]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

Fixes: 766038846e875 ("md/raid0: replace printk() with pr_*()")
Reviewed-by: Michael Kelley 
Acked-by: Guoqing Jiang 
Signed-off-by: Saurabh Sengar 
Signed-off-by: Song Liu 
Signed-off-by: Sasha Levin

md: Flush workqueue md_rdev_misc_wq in md_alloc()

2022-09-15T08:47:09+00:00

[ Upstream commit 5e8daf906f890560df430d30617c692a794acb73 ]

A race condition still exists when removing and re-creating md devices
in test cases. However, it is only seen on some setups.

The race condition was tracked down to a reference still being held
to the kobject by the rdev in the md_rdev_misc_wq which will be released
in rdev_delayed_delete().

md_alloc() waits for previous deletions by waiting on the md_misc_wq,
but the md_rdev_misc_wq may still be holding a reference to a recently
removed device.

To fix this, also flush the md_rdev_misc_wq in md_alloc().

Signed-off-by: David Sloan 
[logang@deltatee.com: rewrote commit message]
Signed-off-by: Logan Gunthorpe 
Signed-off-by: Song Liu 
Signed-off-by: Sasha Levin

md: call __md_stop_writes in md_stop

2022-08-31T15:18:18+00:00

commit 0dd84b319352bb8ba64752d4e45396d8b13e6018 upstream.

From the link [1], we can see raid1d was running even after the path
raid_dtr -> md_stop -> __md_stop.

Let's stop write first in destructor to align with normal md-raid to
fix the KASAN issue.

[1]. https://lore.kernel.org/linux-raid/CAPhsuW5gc4AakdGNdF8ubpezAuDLFOYUO_sfMZcec6hQFm8nhg@mail.gmail.com/T/#m7f12bf90481c02c6d2da68c64aeed4779b7df74a

Fixes: 48df498daf62 ("md: move bitmap_destroy to the beginning of __md_stop")
Reported-by: Mikulas Patocka 
Signed-off-by: Guoqing Jiang 
Signed-off-by: Song Liu 
Signed-off-by: Greg Kroah-Hartman

Revert "md-raid: destroy the bitmap after destroying the thread"

2022-08-31T15:18:18+00:00

commit 1d258758cf06a0734482989911d184dd5837ed4e upstream.

This reverts commit e151db8ecfb019b7da31d076130a794574c89f6f. Because it
obviously breaks clustered raid as noticed by Neil though it fixed KASAN
issue for dm-raid, let's revert it and fix KASAN issue in next commit.

[1]. https://lore.kernel.org/linux-raid/a6657e08-b6a7-358b-2d2a-0ac37d49d23a@linux.dev/T/#m95ac225cab7409f66c295772483d091084a6d470

Fixes: e151db8ecfb0 ("md-raid: destroy the bitmap after destroying the thread")
Signed-off-by: Guoqing Jiang 
Signed-off-by: Song Liu 
Signed-off-by: Greg Kroah-Hartman

drivers:md:fix a potential use-after-free bug

2022-08-25T09:45:47+00:00

[ Upstream commit 104212471b1c1817b311771d817fb692af983173 ]

In line 2884, "raid5_release_stripe(sh);" drops the reference to sh and
may cause sh to be released. However, sh is subsequently used in lines
2886 "if (sh->batch_head && sh != sh->batch_head)". This may result in an
use-after-free bug.

It can be fixed by moving "raid5_release_stripe(sh);" to the bottom of
the function.

Signed-off-by: Wentao_Liang 
Signed-off-by: Song Liu 
Signed-off-by: Jens Axboe 
Signed-off-by: Sasha Levin

md/raid5: Make logic blocking check consistent with logic that blocks

2022-08-25T09:45:47+00:00

[ Upstream commit 6e3f50d30af847bebce072182bd735e90a294c6a ]

The check in raid5_make_request differs very slightly from the logic
that causes it to block lower down. This likely does not cause a bug
as the check is fuzzy anyway (as reshape may move on between the first
check and the subsequent check). However, make it consistent so it can
be cleaned up in a subsequent patch.

The condition which causes the schedule is:

 !(mddev->reshape_backwards ? logical_sector < conf->reshape_progress :
   logical_sector >= conf->reshape_progress) &&
  (mddev->reshape_backwards ? logical_sector < conf->reshape_safe :
   logical_sector >= conf->reshape_safe)

The condition that causes the early bailout is made to match this.

Signed-off-by: Logan Gunthorpe 
Signed-off-by: Song Liu 
Signed-off-by: Jens Axboe 
Signed-off-by: Sasha Levin