bio-integrity: don't restrict the size of integrity metadata

bio_integrity_add_page restricts the size of the integrity metadata to
queue_max_hw_sectors(q). This restriction is not needed because oversized
bios are split automatically. This restriction causes problems with
dm-integrity 'inline' mode - if we send a large bio to dm-integrity and
the bio's metadata are larger than queue_max_hw_sectors(q),
bio_integrity_add_page fails and the bio is ended with BLK_STS_RESOURCE
error.

An example that triggers it:

dd: error writing '/dev/mapper/in2': Cannot allocate memory
1+0 records in
0+0 records out
0 bytes copied, 0.00169291 s, 0.0 kB/s

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Fixes: fb0987682c ("dm-integrity: introduce the Inline mode")
Fixes: 0ece1d649b ("bio-integrity: create multi-page bvecs in bio_integrity_add_page()")
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Anuj Gupta <anuj20.g@samsung.com>
Link: https://lore.kernel.org/r/e41b3b8e-16c2-70cb-97cb-881234bb200d@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Mikulas Patocka 2024-09-03 21:47:59 +02:00 committed by Jens Axboe
parent e58f5142f8
commit b858a36fe9

View File

@ -167,10 +167,6 @@ int bio_integrity_add_page(struct bio *bio, struct page *page,
struct request_queue *q = bdev_get_queue(bio->bi_bdev);
struct bio_integrity_payload *bip = bio_integrity(bio);
if (((bip->bip_iter.bi_size + len) >> SECTOR_SHIFT) >
queue_max_hw_sectors(q))
return 0;
if (bip->bip_vcnt > 0) {
struct bio_vec *bv = &bip->bip_vec[bip->bip_vcnt - 1];
bool same_page = false;