LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Sasha Levin <sashal@kernel.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Paolo Valente <paolo.valente@linaro.org>,
	Davide Zini <davidezini2@gmail.com>, Jens Axboe <axboe@kernel.dk>,
	Sasha Levin <sashal@kernel.org>,
	linux-block@vger.kernel.org
Subject: [PATCH AUTOSEL 5.13 24/29] block, bfq: honor already-setup queue merges
Date: Sat, 11 Sep 2021 09:12:28 -0400	[thread overview]
Message-ID: <20210911131233.284800-24-sashal@kernel.org> (raw)
In-Reply-To: <20210911131233.284800-1-sashal@kernel.org>

From: Paolo Valente <paolo.valente@linaro.org>

[ Upstream commit 2d52c58b9c9bdae0ca3df6a1eab5745ab3f7d80b ]

The function bfq_setup_merge prepares the merging between two
bfq_queues, say bfqq and new_bfqq. To this goal, it assigns
bfqq->new_bfqq = new_bfqq. Then, each time some I/O for bfqq arrives,
the process that generated that I/O is disassociated from bfqq and
associated with new_bfqq (merging is actually a redirection). In this
respect, bfq_setup_merge increases new_bfqq->ref in advance, adding
the number of processes that are expected to be associated with
new_bfqq.

Unfortunately, the stable-merging mechanism interferes with this
setup. After bfqq->new_bfqq has been set by bfq_setup_merge, and
before all the expected processes have been associated with
bfqq->new_bfqq, bfqq may happen to be stably merged with a different
queue than the current bfqq->new_bfqq. In this case, bfqq->new_bfqq
gets changed. So, some of the processes that have been already
accounted for in the ref counter of the previous new_bfqq will not be
associated with that queue.  This creates an unbalance, because those
references will never be decremented.

This commit fixes this issue by reestablishing the previous, natural
behaviour: once bfqq->new_bfqq has been set, it will not be changed
until all expected redirections have occurred.

Signed-off-by: Davide Zini <davidezini2@gmail.com>
Signed-off-by: Paolo Valente <paolo.valente@linaro.org>
Link: https://lore.kernel.org/r/20210802141352.74353-2-paolo.valente@linaro.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 block/bfq-iosched.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index eccbe2aed7c3..f9638c77dac2 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -2623,6 +2623,15 @@ bfq_setup_merge(struct bfq_queue *bfqq, struct bfq_queue *new_bfqq)
 	 * are likely to increase the throughput.
 	 */
 	bfqq->new_bfqq = new_bfqq;
+	/*
+	 * The above assignment schedules the following redirections:
+	 * each time some I/O for bfqq arrives, the process that
+	 * generated that I/O is disassociated from bfqq and
+	 * associated with new_bfqq. Here we increases new_bfqq->ref
+	 * in advance, adding the number of processes that are
+	 * expected to be associated with new_bfqq as they happen to
+	 * issue I/O.
+	 */
 	new_bfqq->ref += process_refs;
 	return new_bfqq;
 }
@@ -2685,6 +2694,10 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 {
 	struct bfq_queue *in_service_bfqq, *new_bfqq;
 
+	/* if a merge has already been setup, then proceed with that first */
+	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
@@ -2784,9 +2797,6 @@ bfq_setup_cooperator(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 	if (bfq_too_late_for_merging(bfqq))
 		return NULL;
 
-	if (bfqq->new_bfqq)
-		return bfqq->new_bfqq;
-
 	if (!io_struct || unlikely(bfqq == &bfqd->oom_bfqq))
 		return NULL;
 
-- 
2.30.2


  parent reply	other threads:[~2021-09-11 13:16 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-09-11 13:12 [PATCH AUTOSEL 5.13 01/29] dt-bindings: mtd: gpmc: Fix the ECC bytes vs. OOB bytes equation Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 02/29] remoteproc: qcom: wcnss: Fix race with iris probe Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 03/29] mfd: db8500-prcmu: Adjust map to reality Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 04/29] PCI: Add ACS quirks for NXP LX2xx0 and LX2xx2 platforms Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 05/29] fuse: fix use after free in fuse_read_interrupt() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 06/29] PCI: tegra194: Fix handling BME_CHGED event Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 07/29] PCI: tegra194: Fix MSI-X programming Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 08/29] PCI: tegra: Fix OF node reference leak Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 09/29] mfd: Don't use irq_create_mapping() to resolve a mapping Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 10/29] PCI: rcar: Fix runtime PM imbalance in rcar_pcie_ep_probe() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 11/29] PCI: rcar: Add L1 link state fix into data abort hook Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 12/29] KVM: arm64: Make hyp_panic() more robust when protected mode is enabled Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 13/29] tracing/probes: Reject events which have the same name of existing one Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 14/29] PCI: cadence: Use bitfield for *quirk_retrain_flag* instead of bool Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 15/29] PCI: cadence: Add quirk flag to set minimum delay in LTSSM Detect.Quiet state Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 16/29] PCI: j721e: Add PCIe support for J7200 Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 17/29] PCI: j721e: Add PCIe support for AM64 Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 18/29] PCI: Add ACS quirks for Cavium multi-function devices Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 19/29] watchdog: Start watchdog in watchdog_set_last_hw_keepalive only if appropriate Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 20/29] octeontx2-af: Add additional register check to rvu_poll_reg() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 21/29] Set fc_nlinfo in nh_create_ipv4, nh_create_ipv6 Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 22/29] flow: fix object-size-mismatch warning in flowi{4,6}_to_flowi_common() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 23/29] net: usb: cdc_mbim: avoid altsetting toggling for Telit LN920 Sasha Levin
2021-09-11 13:12 ` Sasha Levin [this message]
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 25/29] PCI: ibmphp: Fix double unmap of io_mem Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 26/29] ethtool: Fix an error code in cxgb2.c Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 27/29] NTB: Fix an error code in ntb_msit_probe() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 28/29] NTB: perf: Fix an error code in perf_setup_inbuf() Sasha Levin
2021-09-11 13:12 ` [PATCH AUTOSEL 5.13 29/29] net: phylink: add suspend/resume support Sasha Levin

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=20210911131233.284800-24-sashal@kernel.org \
    --to=sashal@kernel.org \
    --cc=axboe@kernel.dk \
    --cc=davidezini2@gmail.com \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=paolo.valente@linaro.org \
    --cc=stable@vger.kernel.org \
    --subject='Re: [PATCH AUTOSEL 5.13 24/29] block, bfq: honor already-setup queue merges' \
    /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

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).