From: Kemeng Shi <shikemeng@huawei.com>
To: <paolo.valente@linaro.org>, <axboe@kernel.dk>
Cc: <linux-block@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
<shikemeng@huawei.com>
Subject: [PATCH 13/20] block,bfq: remove redundant nonrot_with_queueing check in bfq_setup_cooperator
Date: Tue, 1 Nov 2022 17:34:10 +0800 [thread overview]
Message-ID: <20221101093417.10540-14-shikemeng@huawei.com> (raw)
In-Reply-To: <20221101093417.10540-1-shikemeng@huawei.com>
Commit 430a67f9d6169 ("block, bfq: merge bursts of newly-created queues")
add stable merge logic in bfq_setup_cooperator and will only be executed
for !nonrot_with_queueing device. Actually, bfq_setup_cooperator is
designed for only !nonrot_with_queueing and has already returned NULL
before doing real work if device is nonrot_with_queueing. We can add
stable merge after existing nonrot_with_queueing check and no need to
re-check nonrot_with_queueing.
Signed-off-by: Kemeng Shi <shikemeng@huawei.com>
---
block/bfq-iosched.c | 97 ++++++++++++++++++++++-----------------------
1 file changed, 47 insertions(+), 50 deletions(-)
diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index a46e49de895a..b8af0bb98d66 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -2886,56 +2886,6 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
if (bfqq->new_bfqq)
return bfqq->new_bfqq;
- /*
- * Check delayed stable merge for rotational or non-queueing
- * devs. For this branch to be executed, bfqq must not be
- * currently merged with some other queue (i.e., bfqq->bic
- * must be non null). If we considered also merged queues,
- * then we should also check whether bfqq has already been
- * merged with bic->stable_merge_bfqq. But this would be
- * costly and complicated.
- */
- if (unlikely(!bfqd->nonrot_with_queueing)) {
- /*
- * Make sure also that bfqq is sync, because
- * bic->stable_merge_bfqq may point to some queue (for
- * stable merging) also if bic is associated with a
- * sync queue, but this bfqq is async
- */
- if (bfq_bfqq_sync(bfqq) && bic->stable_merge_bfqq &&
- !bfq_bfqq_just_created(bfqq) &&
- time_is_before_jiffies(bfqq->split_time +
- msecs_to_jiffies(bfq_late_stable_merging)) &&
- time_is_before_jiffies(bfqq->creation_time +
- msecs_to_jiffies(bfq_late_stable_merging))) {
- struct bfq_queue *stable_merge_bfqq =
- bic->stable_merge_bfqq;
- int proc_ref = min(bfqq_process_refs(bfqq),
- bfqq_process_refs(stable_merge_bfqq));
-
- /* deschedule stable merge, because done or aborted here */
- bfq_put_stable_ref(stable_merge_bfqq);
-
- bic->stable_merge_bfqq = NULL;
-
- if (!idling_boosts_thr_without_issues(bfqd, bfqq) &&
- proc_ref > 0) {
- /* next function will take at least one ref */
- struct bfq_queue *new_bfqq =
- bfq_setup_merge(bfqq, stable_merge_bfqq);
-
- if (new_bfqq) {
- bic->stably_merged = true;
- if (new_bfqq->bic)
- new_bfqq->bic->stably_merged =
- true;
- }
- return new_bfqq;
- } else
- return NULL;
- }
- }
-
/*
* Do not perform queue merging if the device is non
* rotational and performs internal queueing. In fact, such a
@@ -2976,6 +2926,53 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
if (likely(bfqd->nonrot_with_queueing))
return NULL;
+ /*
+ * Check delayed stable merge for rotational or non-queueing
+ * devs. For this branch to be executed, bfqq must not be
+ * currently merged with some other queue (i.e., bfqq->bic
+ * must be non null). If we considered also merged queues,
+ * then we should also check whether bfqq has already been
+ * merged with bic->stable_merge_bfqq. But this would be
+ * costly and complicated.
+ * Make sure also that bfqq is sync, because
+ * bic->stable_merge_bfqq may point to some queue (for
+ * stable merging) also if bic is associated with a
+ * sync queue, but this bfqq is async
+ */
+ if (bfq_bfqq_sync(bfqq) && bic->stable_merge_bfqq &&
+ !bfq_bfqq_just_created(bfqq) &&
+ time_is_before_jiffies(bfqq->split_time +
+ msecs_to_jiffies(bfq_late_stable_merging)) &&
+ time_is_before_jiffies(bfqq->creation_time +
+ msecs_to_jiffies(bfq_late_stable_merging))) {
+ struct bfq_queue *stable_merge_bfqq =
+ bic->stable_merge_bfqq;
+ int proc_ref = min(bfqq_process_refs(bfqq),
+ bfqq_process_refs(stable_merge_bfqq));
+
+ /* deschedule stable merge, because done or aborted here */
+ bfq_put_stable_ref(stable_merge_bfqq);
+
+ bic->stable_merge_bfqq = NULL;
+
+ if (!idling_boosts_thr_without_issues(bfqd, bfqq) &&
+ proc_ref > 0) {
+ /* next function will take at least one ref */
+ struct bfq_queue *new_bfqq =
+ bfq_setup_merge(bfqq, stable_merge_bfqq);
+
+ if (new_bfqq) {
+ bic->stably_merged = true;
+ if (new_bfqq->bic)
+ new_bfqq->bic->stably_merged =
+ true;
+ }
+ return new_bfqq;
+ } else
+ return NULL;
+ }
+
+
/*
* Prevent bfqq from being merged if it has been created too
* long ago. The idea is that true cooperating processes, and
--
2.30.0
next prev parent reply other threads:[~2022-11-01 9:34 UTC|newest]
Thread overview: 21+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-01 9:33 [PATCH 00/20] A few bugfix and cleanup patches for bfq-iosched Kemeng Shi
2022-11-01 9:33 ` [PATCH 01/20] block, bfq: fix typo in comment Kemeng Shi
2022-11-01 9:33 ` [PATCH 02/20] block, bfq: Update bfqd->max_budget with bfqd->lock held Kemeng Shi
2022-11-01 9:34 ` [PATCH 03/20] block, bfq: correct bfq_max_budget and bfq_min_budget Kemeng Shi
2022-11-01 9:34 ` [PATCH 04/20] block, bfq: simpfy computation of bfqd->budgets_assigned Kemeng Shi
2022-11-01 9:34 ` [PATCH 05/20] block, bfq: recover the "service hole" if enough budget is left Kemeng Shi
2022-11-01 9:34 ` [PATCH 06/20] block, bfq: correct interactive weight-raise check in bfq_set_budget_timeout Kemeng Shi
2022-11-01 9:34 ` [PATCH 07/20] block, bfq: simpfy check for interactive bfqq in bfq_update_wr_data Kemeng Shi
2022-11-01 9:34 ` [PATCH 08/20] block, bfq: do srt filtering for interactive queues in bfq_completed_request Kemeng Shi
2022-11-01 9:34 ` [PATCH 09/20] block, bfq: remove redundant check if (bfqq->dispatched > 0) Kemeng Shi
2022-11-01 9:34 ` [PATCH 10/20] block, bfq: define and use soft_rt, in_burst and wr_or_deserves_wr only low_latency is enable Kemeng Shi
2022-11-01 9:34 ` [PATCH 11/20] block, bfq: remove unnecessary "wr" part of wr_or_deserves_wr Kemeng Shi
2022-11-01 9:34 ` [PATCH 12/20] block, bfq: start/restart service_from_wr accumulating correctly Kemeng Shi
2022-11-01 9:34 ` Kemeng Shi [this message]
2022-11-01 9:34 ` [PATCH 14/20] block, bfq: remove redundant oom_bfqq check for bfqq from bfq_find_close_cooperator Kemeng Shi
2022-11-01 9:34 ` [PATCH 15/20] block, bfq: some cleanups for function bfq_pos_tree_add_move Kemeng Shi
2022-11-01 9:34 ` [PATCH 16/20] block, bfq: remove unnecessary goto tag in __bfq_weights_tree_remove Kemeng Shi
2022-11-01 9:34 ` [PATCH 17/20] block, bfq: remove unnecessary traverse in bfq_add_to_burst Kemeng Shi
2022-11-01 9:34 ` [PATCH 18/20] block, bfq: remove unnecessary bfqq->next_rq = NULL in bfq_remove_request Kemeng Shi
2022-11-01 9:34 ` [PATCH 19/20] block, bfq: remove unnecessary local variable __bfqq in bfq_setup_merge Kemeng Shi
2022-11-01 9:34 ` [PATCH 20/20] block, bfq: remove unncessary process_ref check for merged queue " Kemeng Shi
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=20221101093417.10540-14-shikemeng@huawei.com \
--to=shikemeng@huawei.com \
--cc=axboe@kernel.dk \
--cc=linux-block@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=paolo.valente@linaro.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).