md: don't check MD_SB_CHANGE_CLEAN in md_allow_write

Only MD_SB_CHANGE_PENDING should be used to wait for transition from
clean to dirty. Checking also MD_SB_CHANGE_CLEAN is unnecessary and can
race with e.g. md_do_sync(). This sporadically causes a hang when
changing consistency policy during resync:

INFO: task mdadm:6183 blocked for more than 30 seconds.
      Not tainted 4.14.0-rc3+ #391
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
mdadm           D12752  6183   6022 0x00000000
Call Trace:
 __schedule+0x93f/0x990
 schedule+0x6b/0x90
 md_allow_write+0x100/0x130 [md_mod]
 ? do_wait_intr_irq+0x90/0x90
 resize_stripes+0x3a/0x5b0 [raid456]
 ? kernfs_fop_write+0xbe/0x180
 raid5_change_consistency_policy+0xa6/0x200 [raid456]
 consistency_policy_store+0x2e/0x70 [md_mod]
 md_attr_store+0x90/0xc0 [md_mod]
 sysfs_kf_write+0x42/0x50
 kernfs_fop_write+0x119/0x180
 __vfs_write+0x28/0x110
 ? rcu_sync_lockdep_assert+0x12/0x60
 ? __sb_start_write+0x15a/0x1c0
 ? vfs_write+0xa3/0x1a0
 vfs_write+0xb4/0x1a0
 SyS_write+0x49/0xa0
 entry_SYSCALL_64_fastpath+0x18/0xad

Fixes: 2214c260c7 ("md: don't return -EAGAIN in md_allow_write for external metadata arrays")
Cc: <stable@vger.kernel.org>
Signed-off-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Signed-off-by: Shaohua Li <shli@fb.com>
This commit is contained in:
Artur Paszkiewicz 2017-10-26 15:56:54 +02:00 committed by Shaohua Li
parent f0e230ad87
commit b90f6ff080
1 changed files with 0 additions and 1 deletions

View File

@ -8125,7 +8125,6 @@ void md_allow_write(struct mddev *mddev)
sysfs_notify_dirent_safe(mddev->sysfs_state); sysfs_notify_dirent_safe(mddev->sysfs_state);
/* wait for the dirty state to be recorded in the metadata */ /* wait for the dirty state to be recorded in the metadata */
wait_event(mddev->sb_wait, wait_event(mddev->sb_wait,
!test_bit(MD_SB_CHANGE_CLEAN, &mddev->sb_flags) &&
!test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags)); !test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags));
} else } else
spin_unlock(&mddev->lock); spin_unlock(&mddev->lock);