blk-mq: update hardware and software queues for sleeping alloc
authorJens Axboe <axboe@fb.com>
Thu, 27 Oct 2016 15:49:19 +0000 (09:49 -0600)
committerJens Axboe <axboe@fb.com>
Fri, 28 Oct 2016 01:13:56 +0000 (19:13 -0600)
commit7dd2fb6877b601a0b9c7284534ca4916cd7ac4bd
tree47535b50e12507af176a503ebb364938c2fa3651
parentff26956875c2f05e12ecec9938411a2c7dfc767d
blk-mq: update hardware and software queues for sleeping alloc

If we end up sleeping due to running out of requests, we should
update the hardware and software queues in the map ctx structure.
Otherwise we could end up having rq->mq_ctx point to the pre-sleep
context, and risk corrupting ctx->rq_list since we'll be
grabbing the wrong lock when inserting the request.

Reported-by: Dave Jones <davej@codemonkey.org.uk>
Reported-by: Chris Mason <clm@fb.com>
Tested-by: Chris Mason <clm@fb.com>
Fixes: 63581af3f31e ("blk-mq: remove non-blocking pass in blk_mq_map_request")
Signed-off-by: Jens Axboe <axboe@fb.com>
block/blk-mq.c