Commit e0910c8e authored by Mike Snitzer's avatar Mike Snitzer

dm raid: fix discard limits for raid1 and raid10

Block core warned that discard_granularity was 0 for dm-raid with
personality of raid1.  Reason is that raid_io_hints() was incorrectly
special-casing raid1 rather than raid0.

But since commit 29efc390 ("md/md0: optimize raid0 discard
handling") even raid0 properly handles large discards.

Fix raid_io_hints() by removing discard limits settings for raid1.
Also, fix limits for raid10 by properly stacking underlying limits as
done in blk_stack_limits().

Depends-on: 29efc390 ("md/md0: optimize raid0 discard handling")
Fixes: 61697a6a ("dm: eliminate 'split_discard_bios' flag from DM target interface")
Cc: stable@vger.kernel.org
Reported-by: default avatarZdenek Kabelac <zkabelac@redhat.com>
Reported-by: default avatarMikulas Patocka <mpatocka@redhat.com>
Signed-off-by: default avatarMike Snitzer <snitzer@redhat.com>
parent cd746938
...@@ -3730,12 +3730,14 @@ static void raid_io_hints(struct dm_target *ti, struct queue_limits *limits) ...@@ -3730,12 +3730,14 @@ static void raid_io_hints(struct dm_target *ti, struct queue_limits *limits)
blk_limits_io_opt(limits, chunk_size_bytes * mddev_data_stripes(rs)); blk_limits_io_opt(limits, chunk_size_bytes * mddev_data_stripes(rs));
/* /*
* RAID1 and RAID10 personalities require bio splitting, * RAID10 personality requires bio splitting,
* RAID0/4/5/6 don't and process large discard bios properly. * RAID0/1/4/5/6 don't and process large discard bios properly.
*/ */
if (rs_is_raid1(rs) || rs_is_raid10(rs)) { if (rs_is_raid10(rs)) {
limits->discard_granularity = chunk_size_bytes; limits->discard_granularity = max(chunk_size_bytes,
limits->max_discard_sectors = rs->md.chunk_sectors; limits->discard_granularity);
limits->max_discard_sectors = min_not_zero(rs->md.chunk_sectors,
limits->max_discard_sectors);
} }
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment