diff options
| author | Nate Dailey <nate.dailey@stratus.com> | 2019-02-07 20:19:01 +0100 |
|---|---|---|
| committer | Song Liu <songliubraving@fb.com> | 2019-02-12 23:06:58 +0100 |
| commit | dfcc34c99f3ebc16b787b118763bf9cb6b1efc7a (patch) | |
| tree | 1019fd0794fb6db7465d218b0a7444d8e68ed223 /drivers/md/dm-log-userspace-base.c | |
| parent | floppy: check_events callback should not return a negative number (diff) | |
| download | linux-dfcc34c99f3ebc16b787b118763bf9cb6b1efc7a.tar.xz linux-dfcc34c99f3ebc16b787b118763bf9cb6b1efc7a.zip | |
md/raid1: don't clear bitmap bits on interrupted recovery.
sync_request_write no longer submits writes to a Faulty device. This has
the unfortunate side effect that bitmap bits can be incorrectly cleared
if a recovery is interrupted (previously, end_sync_write would have
prevented this). This means the next recovery may not copy everything
it should, potentially corrupting data.
Add a function for doing the proper md_bitmap_end_sync, called from
end_sync_write and the Faulty case in sync_request_write.
backport note to 4.14: s/md_bitmap_end_sync/bitmap_end_sync
Cc: stable@vger.kernel.org 4.14+
Fixes: 0c9d5b127f69 ("md/raid1: avoid reusing a resync bio after error handling.")
Reviewed-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Tested-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Nate Dailey <nate.dailey@stratus.com>
Signed-off-by: Song Liu <songliubraving@fb.com>
Diffstat (limited to 'drivers/md/dm-log-userspace-base.c')
0 files changed, 0 insertions, 0 deletions