From: Jens Axboe <axboe@kernel.dk>
To: linux-block@vger.kernel.org
Cc: Jens Axboe <axboe@kernel.dk>
Subject: [PATCH 1/4] block: have plug stored requests hold references to the queue
Date: Wed, 3 Nov 2021 12:32:19 -0600 [thread overview]
Message-ID: <20211103183222.180268-2-axboe@kernel.dk> (raw)
In-Reply-To: <20211103183222.180268-1-axboe@kernel.dk>
Requests that were stored in the cache deliberately didn't hold an enter
reference to the queue, instead we grabbed one every time we pulled a
request out of there. That made for awkward logic on freeing the remainder
of the cached list, if needed, where we had to artificially raise the
queue usage count before each free.
Grab references up front for cached plug requests. That's safer, and also
more efficient.
Fixes: 47c122e35d7e ("block: pre-allocate requests if plug is started and is a batch")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
block/blk-core.c | 2 +-
block/blk-mq.c | 7 ++++---
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/block/blk-core.c b/block/blk-core.c
index fd389a16013c..c2d267b6f910 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -1643,7 +1643,7 @@ void blk_flush_plug(struct blk_plug *plug, bool from_schedule)
flush_plug_callbacks(plug, from_schedule);
if (!rq_list_empty(plug->mq_list))
blk_mq_flush_plug_list(plug, from_schedule);
- if (unlikely(!from_schedule && plug->cached_rq))
+ if (unlikely(!rq_list_empty(plug->cached_rq)))
blk_mq_free_plug_rqs(plug);
}
diff --git a/block/blk-mq.c b/block/blk-mq.c
index c68aa0a332e1..5498454c2164 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -410,7 +410,10 @@ __blk_mq_alloc_requests_batch(struct blk_mq_alloc_data *data,
tag_mask &= ~(1UL << i);
rq = blk_mq_rq_ctx_init(data, tags, tag, alloc_time_ns);
rq_list_add(data->cached_rq, rq);
+ nr++;
}
+ /* caller already holds a reference, add for remainder */
+ percpu_ref_get_many(&data->q->q_usage_counter, nr - 1);
data->nr_tags -= nr;
return rq_list_pop(data->cached_rq);
@@ -630,10 +633,8 @@ void blk_mq_free_plug_rqs(struct blk_plug *plug)
{
struct request *rq;
- while ((rq = rq_list_pop(&plug->cached_rq)) != NULL) {
- percpu_ref_get(&rq->q->q_usage_counter);
+ while ((rq = rq_list_pop(&plug->cached_rq)) != NULL)
blk_mq_free_request(rq);
- }
}
static void req_bio_endio(struct request *rq, struct bio *bio,
--
2.33.1
next prev parent reply other threads:[~2021-11-03 18:32 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-03 18:32 [PATCHSET 0/4] Alloc batch fixes Jens Axboe
2021-11-03 18:32 ` Jens Axboe [this message]
2021-11-04 9:01 ` [PATCH 1/4] block: have plug stored requests hold references to the queue Christoph Hellwig
2021-11-04 11:33 ` Jens Axboe
2021-11-03 18:32 ` [PATCH 2/4] block: make blk_try_enter_queue() available for blk-mq Jens Axboe
2021-11-03 18:32 ` [PATCH 3/4] block: move queue enter logic into blk_mq_submit_bio() Jens Axboe
2021-11-04 9:10 ` Christoph Hellwig
2021-11-04 11:41 ` Jens Axboe
2021-11-04 12:14 ` Jens Axboe
2021-11-04 17:32 ` Christoph Hellwig
2021-11-04 17:35 ` Jens Axboe
2021-11-04 17:30 ` Christoph Hellwig
2021-11-04 17:36 ` Jens Axboe
2021-11-03 18:32 ` [PATCH 4/4] block: move plug rq alloc into helper and ensure queue match Jens Axboe
2021-11-04 9:17 ` Christoph Hellwig
2021-11-04 11:35 ` Jens Axboe
2021-11-04 17:33 ` Christoph Hellwig
2021-11-04 17:37 ` Jens Axboe
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20211103183222.180268-2-axboe@kernel.dk \
--to=axboe@kernel.dk \
--cc=linux-block@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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).