Dear Logan, dear David,
Thank you for the patch.
Am 02.09.22 um 19:16 schrieb Logan Gunthorpe:
From: David Sloan <david.sloan@xxxxxxxxxxxxx>
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
[…]
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 he move that is no longer needed as
*t*he move
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")
This commit was added to Linux in v5.14-rc1, so it’s not a recent commit
– or I misunderstood the cover letter.
Kind regards,
Paul
Signed-off-by: David Sloan <david.sloan@xxxxxxxxxxxxx>
[logang@xxxxxxxxxxxx: minor rewording of commit message]
Signed-off-by: Logan Gunthorpe <logang@xxxxxxxxxxxx>
---
drivers/md/raid5.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 4e6d865a6456..734f92e75f85 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -5538,7 +5538,6 @@ static int raid5_read_one_chunk(struct mddev *mddev, struct bio *raid_bio)
if (is_badblock(rdev, sector, bio_sectors(raid_bio), &first_bad,
&bad_sectors)) {
- bio_put(raid_bio);
rdev_dec_pending(rdev, mddev);
return 0;
}
