| Message ID | 20230624130105.1443879-1-ming.lei@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | blk-mq: fix two misuses on RQF_USE_SCHED | expand |
Hi, Ming Could not hit the error after apply your patch. good job! Ming Lei <ming.lei@redhat.com> 于2023年6月24日周六 21:01写道： > > Request allocated from sched tags can't be issued via ->queue_rqs() > directly, since driver tag isn't allocated yet. This is the 1st misuse > of RQF_USE_SCHED for figuring out plug->has_elevator. > > Request allocated from sched tags can't be ended by > blk_mq_end_request_batch() too, fix the 2nd RQF_USE_SCHED misuse > in blk_mq_add_to_batch(). > > Without this patch, NVMe uring cmd passthrough IO workload can run into > hang easily with real io scheduler. > > Fixes: dd6216bb16e8 ("blk-mq: make sure elevator callbacks aren't called for passthrough request") > Reported-by: Guangwu Zhang <guazhang@redhat.com> > Closes: https://lore.kernel.org/linux-block/CAGS2=YrBjpLPOKa-gzcKuuOG60AGth5794PNCDwatdnnscB9ug@mail.gmail.com/ > Cc: Christoph Hellwig <hch@lst.de> > Signed-off-by: Ming Lei <ming.lei@redhat.com> > --- > block/blk-mq.c | 6 +++++- > include/linux/blk-mq.h | 6 +++++- > 2 files changed, 10 insertions(+), 2 deletions(-) > > diff --git a/block/blk-mq.c b/block/blk-mq.c > index 1628873d7587..0aba08a9ded6 100644 > --- a/block/blk-mq.c > +++ b/block/blk-mq.c > @@ -1302,7 +1302,11 @@ static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq) > > if (!plug->multiple_queues && last && last->q != rq->q) > plug->multiple_queues = true; > - if (!plug->has_elevator && (rq->rq_flags & RQF_USE_SCHED)) > + /* > + * Any request allocated from sched tags can't be issued to > + * ->queue_rqs() directly > + */ > + if (!plug->has_elevator && (rq->rq_flags & RQF_SCHED_TAGS)) > plug->has_elevator = true; > rq->rq_next = NULL; > rq_list_add(&plug->mq_list, rq); > diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h > index fa265e85d753..12103d5654f5 100644 > --- a/include/linux/blk-mq.h > +++ b/include/linux/blk-mq.h > @@ -852,7 +852,11 @@ static inline bool blk_mq_add_to_batch(struct request *req, > struct io_comp_batch *iob, int ioerror, > void (*complete)(struct io_comp_batch *)) > { > - if (!iob || (req->rq_flags & RQF_USE_SCHED) || ioerror || > + /* > + * blk_mq_end_request_batch() can't end request allocated from > + * sched tags > + */ > + if (!iob || (req->rq_flags & RQF_SCHED_TAGS) || ioerror || > (req->end_io && !blk_rq_is_passthrough(req))) > return false; > > -- > 2.40.1 > -- Guangwu Zhang Thanks
On Sat, 24 Jun 2023 21:01:05 +0800, Ming Lei wrote: > Request allocated from sched tags can't be issued via ->queue_rqs() > directly, since driver tag isn't allocated yet. This is the 1st misuse > of RQF_USE_SCHED for figuring out plug->has_elevator. > > Request allocated from sched tags can't be ended by > blk_mq_end_request_batch() too, fix the 2nd RQF_USE_SCHED misuse > in blk_mq_add_to_batch(). > > [...] Applied, thanks! [1/1] blk-mq: fix two misuses on RQF_USE_SCHED commit: c6b7a3a26e809c9d2a51ae303764c1d2994f31cf Best regards,
diff --git a/block/blk-mq.c b/block/blk-mq.c index 1628873d7587..0aba08a9ded6 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1302,7 +1302,11 @@ static void blk_add_rq_to_plug(struct blk_plug *plug, struct request *rq) if (!plug->multiple_queues && last && last->q != rq->q) plug->multiple_queues = true; - if (!plug->has_elevator && (rq->rq_flags & RQF_USE_SCHED)) + /* + * Any request allocated from sched tags can't be issued to + * ->queue_rqs() directly + */ + if (!plug->has_elevator && (rq->rq_flags & RQF_SCHED_TAGS)) plug->has_elevator = true; rq->rq_next = NULL; rq_list_add(&plug->mq_list, rq); diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index fa265e85d753..12103d5654f5 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -852,7 +852,11 @@ static inline bool blk_mq_add_to_batch(struct request *req, struct io_comp_batch *iob, int ioerror, void (*complete)(struct io_comp_batch *)) { - if (!iob || (req->rq_flags & RQF_USE_SCHED) || ioerror || + /* + * blk_mq_end_request_batch() can't end request allocated from + * sched tags + */ + if (!iob || (req->rq_flags & RQF_SCHED_TAGS) || ioerror || (req->end_io && !blk_rq_is_passthrough(req))) return false;
Request allocated from sched tags can't be issued via ->queue_rqs() directly, since driver tag isn't allocated yet. This is the 1st misuse of RQF_USE_SCHED for figuring out plug->has_elevator. Request allocated from sched tags can't be ended by blk_mq_end_request_batch() too, fix the 2nd RQF_USE_SCHED misuse in blk_mq_add_to_batch(). Without this patch, NVMe uring cmd passthrough IO workload can run into hang easily with real io scheduler. Fixes: dd6216bb16e8 ("blk-mq: make sure elevator callbacks aren't called for passthrough request") Reported-by: Guangwu Zhang <guazhang@redhat.com> Closes: https://lore.kernel.org/linux-block/CAGS2=YrBjpLPOKa-gzcKuuOG60AGth5794PNCDwatdnnscB9ug@mail.gmail.com/ Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Ming Lei <ming.lei@redhat.com> --- block/blk-mq.c | 6 +++++- include/linux/blk-mq.h | 6 +++++- 2 files changed, 10 insertions(+), 2 deletions(-)