public inbox for stable@vger.kernel.org
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	patches@lists.linux.dev, Yu Kuai <yukuai3@huawei.com>,
	Jens Axboe <axboe@kernel.dk>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.15 39/80] block, bfq: fix procress reference leakage for bfqq in merge chain
Date: Mon, 28 Oct 2024 07:25:19 +0100	[thread overview]
Message-ID: <20241028062253.709311621@linuxfoundation.org> (raw)
In-Reply-To: <20241028062252.611837461@linuxfoundation.org>

5.15-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Yu Kuai <yukuai3@huawei.com>

[ Upstream commit 73aeab373557fa6ee4ae0b742c6211ccd9859280 ]

Original state:

        Process 1       Process 2       Process 3       Process 4
         (BIC1)          (BIC2)          (BIC3)          (BIC4)
          Λ                |               |               |
           \--------------\ \-------------\ \-------------\|
                           V               V               V
          bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
    ref    0               1               2               4

After commit 0e456dba86c7 ("block, bfq: choose the last bfqq from merge
chain in bfq_setup_cooperator()"), if P1 issues a new IO:

Without the patch:

        Process 1       Process 2       Process 3       Process 4
         (BIC1)          (BIC2)          (BIC3)          (BIC4)
          Λ                |               |               |
           \------------------------------\ \-------------\|
                                           V               V
          bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
    ref    0               0               2               4

bfqq3 will be used to handle IO from P1, this is not expected, IO
should be redirected to bfqq4;

With the patch:

          -------------------------------------------
          |                                         |
        Process 1       Process 2       Process 3   |   Process 4
         (BIC1)          (BIC2)          (BIC3)     |    (BIC4)
                           |               |        |      |
                            \-------------\ \-------------\|
                                           V               V
          bfqq1--------->bfqq2---------->bfqq3----------->bfqq4
    ref    0               0               2               4

IO is redirected to bfqq4, however, procress reference of bfqq3 is still
2, while there is only P2 using it.

Fix the problem by calling bfq_merge_bfqqs() for each bfqq in the merge
chain. Also change bfqq_merge_bfqqs() to return new_bfqq to simplify
code.

Fixes: 0e456dba86c7 ("block, bfq: choose the last bfqq from merge chain in bfq_setup_cooperator()")
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Link: https://lore.kernel.org/r/20240909134154.954924-3-yukuai1@huaweicloud.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 block/bfq-iosched.c | 37 +++++++++++++++++--------------------
 1 file changed, 17 insertions(+), 20 deletions(-)

diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index b0bdb5197530a..c985c944fa654 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -2981,10 +2981,12 @@ void bfq_release_process_ref(struct bfq_data *bfqd, struct bfq_queue *bfqq)
 	bfq_put_queue(bfqq);
 }
 
-static void
-bfq_merge_bfqqs(struct bfq_data *bfqd, struct bfq_io_cq *bic,
-		struct bfq_queue *bfqq, struct bfq_queue *new_bfqq)
+static struct bfq_queue *bfq_merge_bfqqs(struct bfq_data *bfqd,
+					 struct bfq_io_cq *bic,
+					 struct bfq_queue *bfqq)
 {
+	struct bfq_queue *new_bfqq = bfqq->new_bfqq;
+
 	bfq_log_bfqq(bfqd, bfqq, "merging with queue %lu",
 		(unsigned long)new_bfqq->pid);
 	/* Save weight raising and idle window of the merged queues */
@@ -3078,6 +3080,8 @@ bfq_merge_bfqqs(struct bfq_data *bfqd, struct bfq_io_cq *bic,
 	bfq_reassign_last_bfqq(bfqq, new_bfqq);
 
 	bfq_release_process_ref(bfqd, bfqq);
+
+	return new_bfqq;
 }
 
 static bool bfq_allow_bio_merge(struct request_queue *q, struct request *rq,
@@ -3113,14 +3117,8 @@ static bool bfq_allow_bio_merge(struct request_queue *q, struct request *rq,
 		 * fulfilled, i.e., bic can be redirected to new_bfqq
 		 * and bfqq can be put.
 		 */
-		bfq_merge_bfqqs(bfqd, bfqd->bio_bic, bfqq,
-				new_bfqq);
-		/*
-		 * If we get here, bio will be queued into new_queue,
-		 * so use new_bfqq to decide whether bio and rq can be
-		 * merged.
-		 */
-		bfqq = new_bfqq;
+		while (bfqq != new_bfqq)
+			bfqq = bfq_merge_bfqqs(bfqd, bfqd->bio_bic, bfqq);
 
 		/*
 		 * Change also bqfd->bio_bfqq, as
@@ -5482,9 +5480,7 @@ bfq_do_early_stable_merge(struct bfq_data *bfqd, struct bfq_queue *bfqq,
 	 * state before killing it.
 	 */
 	bfqq->bic = bic;
-	bfq_merge_bfqqs(bfqd, bic, bfqq, new_bfqq);
-
-	return new_bfqq;
+	return bfq_merge_bfqqs(bfqd, bic, bfqq);
 }
 
 /*
@@ -5916,6 +5912,7 @@ static bool __bfq_insert_request(struct bfq_data *bfqd, struct request *rq)
 	bool waiting, idle_timer_disabled = false;
 
 	if (new_bfqq) {
+		struct bfq_queue *old_bfqq = bfqq;
 		/*
 		 * Release the request's reference to the old bfqq
 		 * and make sure one is taken to the shared queue.
@@ -5931,18 +5928,18 @@ static bool __bfq_insert_request(struct bfq_data *bfqd, struct request *rq)
 		 * then complete the merge and redirect it to
 		 * new_bfqq.
 		 */
-		if (bic_to_bfqq(RQ_BIC(rq), 1) == bfqq)
-			bfq_merge_bfqqs(bfqd, RQ_BIC(rq),
-					bfqq, new_bfqq);
+		if (bic_to_bfqq(RQ_BIC(rq), 1) == bfqq) {
+			while (bfqq != new_bfqq)
+				bfqq = bfq_merge_bfqqs(bfqd, RQ_BIC(rq), bfqq);
+		}
 
-		bfq_clear_bfqq_just_created(bfqq);
+		bfq_clear_bfqq_just_created(old_bfqq);
 		/*
 		 * rq is about to be enqueued into new_bfqq,
 		 * release rq reference on bfqq
 		 */
-		bfq_put_queue(bfqq);
+		bfq_put_queue(old_bfqq);
 		rq->elv.priv[1] = new_bfqq;
-		bfqq = new_bfqq;
 	}
 
 	bfq_update_io_thinktime(bfqd, bfqq);
-- 
2.43.0




  parent reply	other threads:[~2024-10-28  6:32 UTC|newest]

Thread overview: 89+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-10-28  6:24 [PATCH 5.15 00/80] 5.15.170-rc1 review Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 01/80] bpf: Make sure internal and UAPI bpf_redirect flags dont overlap Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 02/80] bpf: devmap: provide rxq after redirect Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 03/80] RDMA/bnxt_re: Fix incorrect AVID type in WQE structure Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 04/80] RDMA/bnxt_re: Add a check for memory allocation Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 05/80] x86/resctrl: Avoid overflow in MB settings in bw_validate() Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 06/80] ARM: dts: bcm2837-rpi-cm3-io3: Fix HDMI hpd-gpio pin Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 07/80] ALSA: hda/cs8409: Fix possible NULL dereference Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 08/80] RDMA/cxgb4: Fix RDMA_CM_EVENT_UNREACHABLE error for iWARP Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 09/80] RDMA/irdma: Fix misspelling of "accept*" Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 10/80] ipv4: give an IPv4 dev to blackhole_netdev Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 11/80] RDMA/bnxt_re: Return more meaningful error Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 12/80] RDMA/bnxt_re: Fix a bug while setting up Level-2 PBL pages Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 13/80] drm/msm/dsi: fix 32-bit signed integer extension in pclk_rate calculation Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 14/80] drm/msm: Avoid NULL dereference in msm_disp_state_print_regs() Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 15/80] drm/msm: Allocate memory for disp snapshot with kvzalloc() Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 16/80] net: usb: usbnet: fix race in probe failure Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 17/80] octeontx2-af: Fix potential integer overflows on integer shifts Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 18/80] macsec: dont increment counters for an unrelated SA Greg Kroah-Hartman
2024-10-28  6:24 ` [PATCH 5.15 19/80] net: ethernet: aeroflex: fix potential memory leak in greth_start_xmit_gbit() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 20/80] net/smc: Fix searching in list of known pnetids in smc_pnet_add_pnetid Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 21/80] net: xilinx: axienet: fix potential memory leak in axienet_start_xmit() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 22/80] net: systemport: fix potential memory leak in bcm_sysport_xmit() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 23/80] tcp/dccp: Dont use timer_pending() in reqsk_queue_unlink() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 24/80] genetlink: hold RCU in genlmsg_mcast() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 25/80] scsi: target: core: Fix null-ptr-deref in target_alloc_device() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 26/80] smb: client: fix OOBs when building SMB2_IOCTL request Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 27/80] usb: typec: altmode should keep reference to parent Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 28/80] s390: Initialize psw mask in perf_arch_fetch_caller_regs() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 29/80] Bluetooth: bnep: fix wild-memory-access in proto_unregister Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 30/80] arm64:uprobe fix the uprobe SWBP_INSN in big-endian Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 31/80] arm64: probes: Fix uprobes for big-endian kernels Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 32/80] KVM: s390: gaccess: Refactor gpa and length calculation Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 33/80] KVM: s390: gaccess: Refactor access address range check Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 34/80] KVM: s390: gaccess: Cleanup access to guest pages Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 35/80] KVM: s390: gaccess: Check if guest address is in memslot Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 36/80] usb: gadget: Add function wakeup support Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 37/80] XHCI: Separate PORT and CAPs macros into dedicated file Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 38/80] usb: dwc3: core: Fix system suspend on TI AM62 platforms Greg Kroah-Hartman
2024-10-28  6:25 ` Greg Kroah-Hartman [this message]
2024-10-28  6:25 ` [PATCH 5.15 40/80] exec: dont WARN for racy path_noexec check Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 41/80] drm/vboxvideo: Replace fake VLA at end of vbva_mouse_pointer_shape with real VLA Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 42/80] ASoC: codecs: lpass-rx-macro: add missing CDC_RX_BCL_VBAT_RF_PROC2 to default regs values Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 43/80] ASoC: fsl_sai: Enable FIFO continue on error FCONT bit Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 44/80] arm64: Force position-independent veneers Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 45/80] udf: fix uninit-value use in udf_get_fileshortad Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 46/80] platform/x86: dell-wmi: Ignore suspend notifications Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 47/80] arm64/uprobes: change the uprobe_opcode_t typedef to fix the sparse warning Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 48/80] ASoC: qcom: sm8250: add qrb4210-rb2-sndcard compatible string Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 49/80] platform/x86: dell-sysman: add support for alienware products Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 50/80] jfs: Fix sanity check in dbMount Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 51/80] tracing: Consider the NULL character when validating the event length Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 52/80] xfrm: extract dst lookup parameters into a struct Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 53/80] xfrm: respect ip protocols rules criteria when performing dst lookups Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 54/80] net/sun3_82586: fix potential memory leak in sun3_82586_send_packet() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 55/80] be2net: fix potential memory leak in be_xmit() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 56/80] net: plip: fix break; causing plip to never transmit Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 57/80] net: dsa: mv88e6xxx: Fix error when setting port policy on mv88e6393x Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 58/80] netfilter: xtables: fix typo causing some targets not to load on IPv6 Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 59/80] net: wwan: fix global oob in wwan_rtnl_policy Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 60/80] net: usb: usbnet: fix name regression Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 61/80] net: sched: fix use-after-free in taprio_change() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 62/80] r8169: avoid unsolicited interrupts Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 63/80] posix-clock: posix-clock: Fix unbalanced locking in pc_clock_settime() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 64/80] bpf,perf: Fix perf_event_detach_bpf_prog error handling Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 65/80] ALSA: firewire-lib: Avoid division by zero in apply_constraint_to_size() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 66/80] ALSA: hda/realtek: Update default depop procedure Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 67/80] btrfs: zoned: fix zone unusable accounting for freed reserved extent Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 68/80] drm/amd: Guard against bad data for ATIF ACPI method Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 69/80] ACPI: resource: Add LG 16T90SP to irq1_level_low_skip_override[] Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 70/80] ACPI: button: Add DMI quirk for Samsung Galaxy Book2 to fix initial lid detection issue Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 71/80] nilfs2: fix kernel bug due to missing clearing of buffer delay flag Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 72/80] openat2: explicitly return -E2BIG for (usize > PAGE_SIZE) Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 73/80] KVM: nSVM: Ignore nCR3[4:0] when loading PDPTEs from memory Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 74/80] ALSA: hda/realtek: Add subwoofer quirk for Acer Predator G9-593 Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 75/80] xfrm: fix one more kernel-infoleak in algo dumping Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 76/80] hv_netvsc: Fix VF namespace also in synthetic NIC NETDEV_REGISTER event Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 77/80] selinux: improve error checking in sel_write_load() Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 78/80] serial: protect uart_port_dtr_rts() in uart_shutdown() too Greg Kroah-Hartman
2024-10-28  6:25 ` [PATCH 5.15 79/80] net: phy: dp83822: Fix reset pin definitions Greg Kroah-Hartman
2024-10-28  6:26 ` [PATCH 5.15 80/80] ASoC: qcom: Fix NULL Dereference in asoc_qcom_lpass_cpu_platform_probe() Greg Kroah-Hartman
2024-10-28 14:21 ` [PATCH 5.15 00/80] 5.15.170-rc1 review Mark Brown
2024-10-28 17:49 ` SeongJae Park
2024-10-28 19:08 ` Florian Fainelli
2024-10-29  2:07 ` [PATCH 5.15] " Hardik Garg
2024-10-29  5:05 ` [PATCH 5.15 00/80] " Harshit Mogalapalli
2024-10-29  7:34 ` Naresh Kamboju
2024-10-29 13:49 ` Muhammad Usama Anjum
2024-10-30  1:46 ` Ron Economos

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=20241028062253.709311621@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=axboe@kernel.dk \
    --cc=patches@lists.linux.dev \
    --cc=sashal@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=yukuai3@huawei.com \
    /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