diff options
author | Christoph Hellwig <hch@lst.de> | 2022-01-18 08:04:44 +0100 |
---|---|---|
committer | Jens Axboe <axboe@kernel.dk> | 2022-01-18 14:34:05 +0100 |
commit | fd9f4e62a39f09a7c014d7415c2b9d1390aa0504 (patch) | |
tree | aef1b7d1c4a37ba8076ddf6ee349003960f383ab /block | |
parent | block: cleanup q->srcu (diff) | |
download | linux-fd9f4e62a39f09a7c014d7415c2b9d1390aa0504.tar.xz linux-fd9f4e62a39f09a7c014d7415c2b9d1390aa0504.zip |
block: assign bi_bdev for cloned bios in blk_rq_prep_clone
bio_clone_fast() sets the cloned bio to have the same ->bi_bdev as the
source bio. This means that when request-based dm called setup_clone(),
the cloned bio had its ->bi_bdev pointing to the dm device. After Commit
0b6e522cdc4a ("blk-mq: use ->bi_bdev for I/O accounting")
__blk_account_io_start() started using the request's ->bio->bi_bdev for
I/O accounting, if it was set. This caused IO going to the underlying
devices to use the dm device for their I/O accounting.
Set up the proper ->bi_bdev in blk_rq_prep_clone based on the whole
device bdev for the queue the request is cloned onto.
Fixes: 0b6e522cdc4a ("blk-mq: use ->bi_bdev for I/O accounting")
Reported-by: Benjamin Marzinski <bmarzins@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
[hch: the commit message is mostly from a different patch from Benjamin]
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Benjamin Marzinski <bmarzins@redhat.com>
Link: https://lore.kernel.org/r/20220118070444.1241739-1-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Diffstat (limited to 'block')
-rw-r--r-- | block/blk-mq.c | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/block/blk-mq.c b/block/blk-mq.c index a6d4780580fc..b5e35e63adad 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -2976,6 +2976,7 @@ int blk_rq_prep_clone(struct request *rq, struct request *rq_src, bio = bio_clone_fast(bio_src, gfp_mask, bs); if (!bio) goto free_and_out; + bio->bi_bdev = rq->q->disk->part0; if (bio_ctr && bio_ctr(bio, bio_src, data)) goto free_and_out; |