blk-mq may schedule to call queue's complete function on remote CPU via IPI, but doesn't provide any way to synchronize the request's complete fn. The current queue freeze interface can't provide the synchonization because aborted requests stay at blk-mq queues during EH. In some driver's EH(such as NVMe), hardware queue's resource may be freed & re-allocated. If the completed request's complete fn is run finally after the hardware queue's resource is released, kernel crash will be triggered. Prepare for fixing this kind of issue by introducing blk_mq_tagset_wait_completed_request().
Showing the changes to include/linux (1 of 2 files). Show all
Unified · Split
| @@ -322,6 +322,7 @@ bool blk_mq_run_hw_queue(struct blk_mq_hw_ctx *hctx, bool async); · 321 unmodified lines | ||
| 322 | 322 | void blk_mq_run_hw_queues(struct request_queue *q, bool async); |
| 323 | 323 | void blk_mq_tagset_busy_iter(struct blk_mq_tag_set *tagset, |
| 324 | 324 | busy_tag_iter_fn *fn, void *priv); |
| 325 | +void blk_mq_tagset_wait_completed_request(struct blk_mq_tag_set *tagset); | |
| 325 | 326 | void blk_mq_freeze_queue(struct request_queue *q); |
| 326 | 327 | void blk_mq_unfreeze_queue(struct request_queue *q); |
| 327 | 328 | void blk_freeze_queue_start(struct request_queue *q); |