On Wed, 2018-01-17 at 23:36 -0500, Mike Snitzer wrote:
> On Wed, Jan 17 2018 at 11:06pm -0500,
> Ming Lei <ming.lei@xxxxxxxxxx> wrote:
>
> > If we run into blk_mq_request_direct_issue(), when queue is busy,
> > we
> > don't want to dispatch this request into hctx->dispatch_list, and
> > what we need to do is to return the queue busy info to caller, so
> > that caller can deal with it well.
> >
> > Fixes: 396eaf21ee ("blk-mq: improve DM's blk-mq IO merging via
> > blk_insert_cloned_request feedback")
> > Reported-by: Laurence Oberman <loberman@xxxxxxxxxx>
> > Reviewed-by: Mike Snitzer <snitzer@xxxxxxxxxx>
> > Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx>
> > ---
> > block/blk-mq.c | 22 ++++++++++------------
> > 1 file changed, 10 insertions(+), 12 deletions(-)
> >
> > diff --git a/block/blk-mq.c b/block/blk-mq.c
> > index 4d4af8d712da..1af7fa70993b 100644
> > --- a/block/blk-mq.c
> > +++ b/block/blk-mq.c
> > @@ -1856,15 +1856,6 @@ static blk_status_t
> > __blk_mq_issue_directly(struct blk_mq_hw_ctx *hctx,
> > return ret;
> > }
> >
> > -static void __blk_mq_fallback_to_insert(struct request *rq,
> > - bool run_queue, bool
> > bypass_insert)
> > -{
> > - if (!bypass_insert)
> > - blk_mq_sched_insert_request(rq, false, run_queue,
> > false);
> > - else
> > - blk_mq_request_bypass_insert(rq, run_queue);
> > -}
> > -
> > static blk_status_t __blk_mq_try_issue_directly(struct
> > blk_mq_hw_ctx *hctx,
> > struct request
> > *rq,
> > blk_qc_t *cookie,
> > @@ -1873,9 +1864,16 @@ static blk_status_t
> > __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
> > struct request_queue *q = rq->q;
> > bool run_queue = true;
> >
> > - /* RCU or SRCU read lock is needed before checking
> > quiesced flag */
> > + /*
> > + * RCU or SRCU read lock is needed before checking
> > quiesced flag.
> > + *
> > + * When queue is stopped or quiesced, ignore
> > 'bypass_insert' from
> > + * blk_mq_request_direct_issue(), and return BLK_STS_OK to
> > caller,
> > + * and avoid driver to try to dispatch again.
> > + */
> > if (blk_mq_hctx_stopped(hctx) || blk_queue_quiesced(q)) {
> > run_queue = false;
> > + bypass_insert = false;
> > goto insert;
> > }
> >
> > @@ -1892,10 +1890,10 @@ static blk_status_t
> > __blk_mq_try_issue_directly(struct blk_mq_hw_ctx *hctx,
> >
> > return __blk_mq_issue_directly(hctx, rq, cookie);
> > insert:
> > - __blk_mq_fallback_to_insert(rq, run_queue, bypass_insert);
> > if (bypass_insert)
> > return BLK_STS_RESOURCE;
> >
> > + blk_mq_sched_insert_request(rq, false, run_queue, false);
> > return BLK_STS_OK;
> > }
>
> OK so you're just leveraging blk_mq_sched_insert_request()'s
> ability to resort to__blk_mq_insert_request() if !q->elevator.
I tested this against Mike's latest combined tree and its stable.
This fixes the list corruption issue.
Many Thanks Ming and Mike.
I will apply it to Bart's latest SRP/SRPT tree tomorrow as its very
late here but it will clearly fix the issue in Bart's tree too.
Tested-by: Laurence Oberman <loberman@xxxxxxxxxx>
