* [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue
@ 2020-09-11 9:44 Ming Lei
2020-09-11 9:52 ` Christoph Hellwig
2020-09-11 10:01 ` Hannes Reinecke
0 siblings, 2 replies; 3+ messages in thread
From: Ming Lei @ 2020-09-11 9:44 UTC (permalink / raw)
To: Jens Axboe, linux-block, linux-nvme, Christoph Hellwig
Cc: Ming Lei, David Milburn, Ewan D . Milne
NVMe shares tagset between fabric queue and admin queue or between
connect_q and NS queue, so hctx_may_queue() can be called to allocate
request for these queues.
Tags can be reserved in these tagset. Before error recovery, there is
often lots of in-flight requests which can't be completed, and new
reserved request may be needed in error recovery path. However,
hctx_may_queue() can always return false because there is too many
in-flight requests which can't be completed during error handling.
Finally, everything can't move on.
Fix this issue by always allowing reserved tag allocation in
hctx_may_queue(). This ways is reasonable because reserved tag
suppose to be ready any time.
Cc: David Milburn <dmilburn@redhat.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
---
block/blk-mq-tag.c | 3 ++-
block/blk-mq.c | 6 ++++--
2 files changed, 6 insertions(+), 3 deletions(-)
diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
index c31c4a0478a5..aacf10decdbd 100644
--- a/block/blk-mq-tag.c
+++ b/block/blk-mq-tag.c
@@ -76,7 +76,8 @@ void __blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
static int __blk_mq_get_tag(struct blk_mq_alloc_data *data,
struct sbitmap_queue *bt)
{
- if (!data->q->elevator && !hctx_may_queue(data->hctx, bt))
+ if (!data->q->elevator && !(data->flags & BLK_MQ_REQ_RESERVED) &&
+ !hctx_may_queue(data->hctx, bt))
return BLK_MQ_NO_TAG;
if (data->shallow_depth)
diff --git a/block/blk-mq.c b/block/blk-mq.c
index ccb500e38008..91cff275451d 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -1147,15 +1147,17 @@ static bool __blk_mq_get_driver_tag(struct request *rq)
struct sbitmap_queue *bt = rq->mq_hctx->tags->bitmap_tags;
unsigned int tag_offset = rq->mq_hctx->tags->nr_reserved_tags;
int tag;
+ bool reserved = blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags,
+ rq->internal_tag);
blk_mq_tag_busy(rq->mq_hctx);
- if (blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags, rq->internal_tag)) {
+ if (reserved) {
bt = rq->mq_hctx->tags->breserved_tags;
tag_offset = 0;
}
- if (!hctx_may_queue(rq->mq_hctx, bt))
+ if (!reserved && !hctx_may_queue(rq->mq_hctx, bt))
return false;
tag = __sbitmap_queue_get(bt);
if (tag == BLK_MQ_NO_TAG)
--
2.25.2
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue
2020-09-11 9:44 [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue Ming Lei
@ 2020-09-11 9:52 ` Christoph Hellwig
2020-09-11 10:01 ` Hannes Reinecke
1 sibling, 0 replies; 3+ messages in thread
From: Christoph Hellwig @ 2020-09-11 9:52 UTC (permalink / raw)
To: Ming Lei
Cc: Jens Axboe, linux-block, linux-nvme, Christoph Hellwig,
David Milburn, Ewan D . Milne
On Fri, Sep 11, 2020 at 05:44:53PM +0800, Ming Lei wrote:
> unsigned int tag_offset = rq->mq_hctx->tags->nr_reserved_tags;
> int tag;
> + bool reserved = blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags,
> + rq->internal_tag);
>
> blk_mq_tag_busy(rq->mq_hctx);
>
> - if (blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags, rq->internal_tag)) {
> + if (reserved) {
> bt = rq->mq_hctx->tags->breserved_tags;
> tag_offset = 0;
> }
>
> - if (!hctx_may_queue(rq->mq_hctx, bt))
> + if (!reserved && !hctx_may_queue(rq->mq_hctx, bt))
What about:
if (blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags, rq->internal_tag)) {
bt = rq->mq_hctx->tags->breserved_tags;
tag_offset = 0;
} else {
if (!hctx_may_queue(rq->mq_hctx, bt))
return false;
}
which seems a little easier to follow?
Otherwise looks good:
Reviewed-by: Christoph Hellwig <hch@lst.de>
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue
2020-09-11 9:44 [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue Ming Lei
2020-09-11 9:52 ` Christoph Hellwig
@ 2020-09-11 10:01 ` Hannes Reinecke
1 sibling, 0 replies; 3+ messages in thread
From: Hannes Reinecke @ 2020-09-11 10:01 UTC (permalink / raw)
To: Ming Lei, Jens Axboe, linux-block, linux-nvme, Christoph Hellwig
Cc: David Milburn, Ewan D . Milne
On 9/11/20 11:44 AM, Ming Lei wrote:
> NVMe shares tagset between fabric queue and admin queue or between
> connect_q and NS queue, so hctx_may_queue() can be called to allocate
> request for these queues.
>
> Tags can be reserved in these tagset. Before error recovery, there is
> often lots of in-flight requests which can't be completed, and new
> reserved request may be needed in error recovery path. However,
> hctx_may_queue() can always return false because there is too many
> in-flight requests which can't be completed during error handling.
> Finally, everything can't move on.
>
> Fix this issue by always allowing reserved tag allocation in
> hctx_may_queue(). This ways is reasonable because reserved tag
> suppose to be ready any time.
>
> Cc: David Milburn <dmilburn@redhat.com>
> Cc: Ewan D. Milne <emilne@redhat.com>
> Signed-off-by: Ming Lei <ming.lei@redhat.com>
> ---
> block/blk-mq-tag.c | 3 ++-
> block/blk-mq.c | 6 ++++--
> 2 files changed, 6 insertions(+), 3 deletions(-)
>
> diff --git a/block/blk-mq-tag.c b/block/blk-mq-tag.c
> index c31c4a0478a5..aacf10decdbd 100644
> --- a/block/blk-mq-tag.c
> +++ b/block/blk-mq-tag.c
> @@ -76,7 +76,8 @@ void __blk_mq_tag_idle(struct blk_mq_hw_ctx *hctx)
> static int __blk_mq_get_tag(struct blk_mq_alloc_data *data,
> struct sbitmap_queue *bt)
> {
> - if (!data->q->elevator && !hctx_may_queue(data->hctx, bt))
> + if (!data->q->elevator && !(data->flags & BLK_MQ_REQ_RESERVED) &&
> + !hctx_may_queue(data->hctx, bt))
> return BLK_MQ_NO_TAG;
>
> if (data->shallow_depth)
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index ccb500e38008..91cff275451d 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1147,15 +1147,17 @@ static bool __blk_mq_get_driver_tag(struct request *rq)
> struct sbitmap_queue *bt = rq->mq_hctx->tags->bitmap_tags;
> unsigned int tag_offset = rq->mq_hctx->tags->nr_reserved_tags;
> int tag;
> + bool reserved = blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags,
> + rq->internal_tag);
>
> blk_mq_tag_busy(rq->mq_hctx);
>
> - if (blk_mq_tag_is_reserved(rq->mq_hctx->sched_tags, rq->internal_tag)) {
> + if (reserved) {
> bt = rq->mq_hctx->tags->breserved_tags;
> tag_offset = 0;
> }
>
> - if (!hctx_may_queue(rq->mq_hctx, bt))
> + if (!reserved && !hctx_may_queue(rq->mq_hctx, bt))
> return false;
> tag = __sbitmap_queue_get(bt);
> if (tag == BLK_MQ_NO_TAG)
>
Very good point.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Cheers,
Hannes
--
Dr. Hannes Reinecke Kernel Storage Architect
hare@suse.de +49 911 74053 688
SUSE Software Solutions Germany GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), GF: Felix Imendörffer
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2020-09-11 10:02 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-11 9:44 [PATCH] blk-mq: always allow reserved allocation in hctx_may_queue Ming Lei
2020-09-11 9:52 ` Christoph Hellwig
2020-09-11 10:01 ` Hannes Reinecke
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).