| Message ID | 20250310091610.2010623-1-ming.lei@redhat.com (mailing list archive) |
|---|---|
| State | New |
| Headers | show |
| Series | block: make sure ->nr_integrity_segments is cloned in blk_rq_prep_clone | expand |
On Mon, Mar 10, 2025 at 05:16:10PM +0800, Ming Lei wrote: > Make sure ->nr_integrity_segments is cloned in blk_rq_prep_clone(), > otherwise zero ->nr_integrity_segments will be observed in > sg_alloc_table_chained(), in which BUG() is hit. Maybe reword this to: ... otherwise requests cloned by device-mapper multipath will not have the prope nr_integrity_segments values set? The fact that sg_alloc_table_chained BUGā½)s is just a symptom of that. > Cc: stable@vger.kernel.org ... and replace this with a proper Fixes tag?
diff --git a/block/blk-mq.c b/block/blk-mq.c index 40490ac88045..005c520d3498 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -3314,6 +3314,7 @@ int blk_rq_prep_clone(struct request *rq, struct request *rq_src, rq->special_vec = rq_src->special_vec; } rq->nr_phys_segments = rq_src->nr_phys_segments; + rq->nr_integrity_segments = rq_src->nr_integrity_segments; if (rq->bio && blk_crypto_rq_bio_prep(rq, rq->bio, gfp_mask) < 0) goto free_and_out;
Make sure ->nr_integrity_segments is cloned in blk_rq_prep_clone(), otherwise zero ->nr_integrity_segments will be observed in sg_alloc_table_chained(), in which BUG() is hit. Cc: stable@vger.kernel.org Signed-off-by: Ming Lei <ming.lei@redhat.com> --- block/blk-mq.c | 1 + 1 file changed, 1 insertion(+)