On Thu, Nov 26, 2020 at 10:28:37AM +0800, Jeffle Xu wrote:
> iopoll is initially for small size, latency sensitive IO. It doesn't
> work well for big IO, especially when it needs to be split to multiple
> bios. In this case, the returned cookie of __submit_bio_noacct_mq() is
> indeed the cookie of the last split bio. The completion of *this* last
> split bio done by iopoll doesn't mean the whole original bio has
> completed. Callers of iopoll still need to wait for completion of other
> split bios.
>
> Besides bio splitting may cause more trouble for iopoll which isn't
> supposed to be used in case of big IO.
>
> iopoll for split bio may cause potential race if CPU migration happens
> during bio submission. Since the returned cookie is that of the last
> split bio, polling on the corresponding hardware queue doesn't help
> complete other split bios, if these split bios are enqueued into
> different hardware queues. Since interrupts are disabled for polling
> queues, the completion of these other split bios depends on timeout
> mechanism, thus causing a potential hang.
>
> iopoll for split bio may also cause hang for sync polling. Currently
> both the blkdev and iomap-based fs (ext4/xfs, etc) support sync polling
> in direct IO routine. These routines will submit bio without REQ_NOWAIT
> flag set, and then start sync polling in current process context. The
> process may hang in blk_mq_get_tag() if the submitted bio has to be
> split into multiple bios and can rapidly exhaust the queue depth. The
> process are waiting for the completion of the previously allocated
> requests, which should be reaped by the following polling, and thus
> causing a deadlock.
>
> To avoid these subtle trouble described above, just disable iopoll for
> split bio and return BLK_QC_T_NONE in this case. The side effect is that
> non-HIPRI IO also returns BLK_QC_T_NONE now. It should be acceptable
> since the returned cookie is never used for non-HIPRI IO.
>
> Suggested-by: Ming Lei <ming.lei@xxxxxxxxxx>
> Signed-off-by: Jeffle Xu <jefflexu@xxxxxxxxxxxxxxxxx>
> ---
> block/blk-merge.c | 8 ++++++++
> block/blk-mq.c | 5 +++++
> 2 files changed, 13 insertions(+)
>
> diff --git a/block/blk-merge.c b/block/blk-merge.c
> index bcf5e4580603..8a2f1fb7bb16 100644
> --- a/block/blk-merge.c
> +++ b/block/blk-merge.c
> @@ -279,6 +279,14 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
> return NULL;
> split:
> *segs = nsegs;
> +
> + /*
> + * Bio splitting may cause subtle trouble such as hang when doing sync
> + * iopoll in direct IO routine. Given performance gain of iopoll for
> + * big IO can be trival, disable iopoll when split needed.
> + */
> + bio->bi_opf &= ~REQ_HIPRI;
> +
> return bio_split(bio, sectors, GFP_NOIO, bs);
> }
>
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index 55bcee5dc032..580fd570be23 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -2157,6 +2157,7 @@ blk_qc_t blk_mq_submit_bio(struct bio *bio)
> unsigned int nr_segs;
> blk_qc_t cookie;
> blk_status_t ret;
> + bool hipri;
>
> blk_queue_bounce(q, &bio);
> __blk_queue_split(&bio, &nr_segs);
> @@ -2173,6 +2174,8 @@ blk_qc_t blk_mq_submit_bio(struct bio *bio)
>
> rq_qos_throttle(q, bio);
>
> + hipri = bio->bi_opf & REQ_HIPRI;
> +
> data.cmd_flags = bio->bi_opf;
> rq = __blk_mq_alloc_request(&data);
> if (unlikely(!rq)) {
> @@ -2265,7 +2268,9 @@ blk_qc_t blk_mq_submit_bio(struct bio *bio)
> blk_mq_sched_insert_request(rq, false, true, true);
> }
>
> + cookie = hipri ? cookie : BLK_QC_T_NONE;
> return cookie;
> +
> queue_exit:
> blk_queue_exit(q);
> return BLK_QC_T_NONE;
> --
> 2.27.0
>
Reviewed-by: Ming Lei <ming.lei@xxxxxxxxxx>
Thanks,
Ming
