diff options
| author | Song Liu <songliubraving@fb.com> | 2020-07-06 14:57:32 -0700 |
|---|---|---|
| committer | Song Liu <songliubraving@fb.com> | 2020-07-14 23:01:31 -0700 |
| commit | c9020e64cf33f2dd5b2a7295f2bfea787279218a (patch) | |
| tree | aec0586fc3450a91410ce822972a21048113ea33 /drivers/md/raid5-cache.c | |
| parent | e1a86dbbbd6a77f73c3d099030495fa31f181e2f (diff) | |
md/raid5-cache: clear MD_SB_CHANGE_PENDING before flushing stripes
In recovery, if we process too much data, raid5-cache may set
MD_SB_CHANGE_PENDING, which causes spinning in handle_stripe().
Fix this issue by clearing the bit before flushing data only
stripes. This issue was initially discussed in [1].
[1] https://www.spinics.net/lists/raid/msg64409.html
Signed-off-by: Song Liu <songliubraving@fb.com>
Diffstat (limited to 'drivers/md/raid5-cache.c')
| -rw-r--r-- | drivers/md/raid5-cache.c | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index 9b6da759dca2..0bea21d81697 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -2430,10 +2430,15 @@ static void r5c_recovery_flush_data_only_stripes(struct r5l_log *log, struct mddev *mddev = log->rdev->mddev; struct r5conf *conf = mddev->private; struct stripe_head *sh, *next; + bool cleared_pending = false; if (ctx->data_only_stripes == 0) return; + if (test_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags)) { + cleared_pending = true; + clear_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags); + } log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_BACK; list_for_each_entry_safe(sh, next, &ctx->cached_list, lru) { @@ -2448,6 +2453,8 @@ static void r5c_recovery_flush_data_only_stripes(struct r5l_log *log, atomic_read(&conf->active_stripes) == 0); log->r5c_journal_mode = R5C_JOURNAL_MODE_WRITE_THROUGH; + if (cleared_pending) + set_bit(MD_SB_CHANGE_PENDING, &mddev->sb_flags); } static int r5l_recovery_log(struct r5l_log *log) |