Received-SPF: pass (google.com: domain of linux-kernel+bounces-183398-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
From: Yang Yang <yang.yang@vivo.com>
To: Jens Axboe <axboe@kernel.dk>,
	linux-block@vger.kernel.org,
	linux-kernel@vger.kernel.org
Cc: Yang Yang <yang.yang@vivo.com>
Subject: [RFC PATCH] blk-mq: fix potential I/O hang caused by batch wakeup
Date: Mon, 20 May 2024 11:38:46 +0800
Message-Id: <20240520033847.13533-1-yang.yang@vivo.com>
Content-Transfer-Encoding: 8bit
Content-Type: text/plain
Precedence: bulk
MIME-Version: 1.0
X-MS-Exchange-AntiSpam-MessageData-ChunkCount: 1
X-MS-Exchange-AntiSpam-MessageData-0:
	=?us-ascii?Q?ltEmVa/oT3ONGIkccdYCz/y7aLmyuaOt+Fcee7M3QjKbnOSDQA56nxvambAp?=
 =?us-ascii?Q?HV+vHZNbLei1IaRMwRBnHq3ejeAV+b8RoNu8tvVTDMOxp2aHyXBq5gZbw1Nm?=
 =?us-ascii?Q?XofIjAoiFkhvetbqeUjblw5aZ290yVnlnZiKEFaOYxXgJ3isYZ4BtT7QtZlu?=
 =?us-ascii?Q?0dTMwkDRR9KxmYbbrFszGZeu35FPmJkIymL3G7z323TtoQCUwFPjV3M6vsin?=
 =?us-ascii?Q?1NZfQrydpHCKvjF5TglB2lmDmZoL/zjuqas+aA3nZ88Fjc8rz5yqG1zZD6xC?=
 =?us-ascii?Q?9F6jvgJR3lCOVKsuthDUcsLNUwCI0fGQklRU0JZoN60mq9sSJ7D6UJ7p+3hh?=
 =?us-ascii?Q?mfzVBAmexp5vOjJcFtjQLQyftDSpmBgiK2lIaqVx6DsjymYozO7c7JI45dLQ?=
 =?us-ascii?Q?Aku0+yEPwbKrsbcUuaRRrr54Ar91gUp4JGZatkAQ1tDbsbrxH+A7bDgzWSGD?=
 =?us-ascii?Q?ewsY2Ky+nl/x5CX/hVm4NFo1FLGbgzknErIhhF8u9kRiE/a5OuRfE+rdaItF?=
 =?us-ascii?Q?jAOUKOa0N5lIPmzFC+rtBrCUl0nnLRPKj11JJ9gHNJhImWl6Mr8RwEQEruEI?=
 =?us-ascii?Q?wXEO0ybPSVgqv8Ra8xMpVKiO/4+Mh0wzLQ/0VUFsyCM6jVi64tvU/nr28TlV?=
 =?us-ascii?Q?MK/CC0VjG3DWbp+i1SBjlFg15AsMpD+LvSVyt8KN8AfN1dHRT0Zueuv8opyy?=
 =?us-ascii?Q?6uZ0s7fcZjh/n2PiW5tNv6BzdRcqYu5l6558+WHjXwWeoYZVENpHWmnWEe0z?=
 =?us-ascii?Q?gz1IDWivzlxykRYqccUsr8AfhSML3f8FKm0mfTpk8Mpha/hBPjOrFlPzl2Z6?=
 =?us-ascii?Q?uwt++nDJ0kwslZwkb44ais0xGFaicXq58KlkZttmqVj8Vls7yrL37LOc27yQ?=
 =?us-ascii?Q?M6QbYDdHSCz2VEHGtLnCjNftDPpzV5fUZjSFL4426AipzehPe/8O0IF4b1nN?=
 =?us-ascii?Q?mf3vgWOXYgwJlGlYmNDtSaj+hLdUD62DLeVhE7ZyLPNFaeI6HBfrn2mfhORl?=
 =?us-ascii?Q?eFxIjVzyOZFNpZoTHZwqcLNFqd10mIKtuYx/c31YDpUgrXgQO1+evnmS3mvx?=
 =?us-ascii?Q?l4QZKvdx+BtAYrHRasUGEJ0HAXlw0x8Ue16b/ImWIAXEbjM/HLFuWy4JuJIS?=
 =?us-ascii?Q?vd4JYMhRrW24GipFiFHU4MXd6oRT9FYihxKIcmEnb48thD3kBOcana6Vp3Tl?=
 =?us-ascii?Q?hidQNE5YUlafj8PxzvHW7xL7v5889aw9F/hN7qw3m5mHKql0sGgHQyApYMbw?=
 =?us-ascii?Q?RgmBq/er0b9PdAbzAhwnYDeWmNz8OwnQlwVLEdSWeniWc1Gz5nZ9PFmF3RsY?=
 =?us-ascii?Q?TNo7IEhHNf08ADiTJygGixwKhZITmGKimCQCdAg8wXNIrBv3utSuqks5K0Iy?=
 =?us-ascii?Q?mM7gNd4i6HNUpXCofj3weRltU3iOoUHJTzDRQhwGzxwUz8qSXQ7JANXcUmA2?=
 =?us-ascii?Q?ByEWqE/JycyfJSvmIxEjpJ+x+mumMVyLYMNXC78uvHniGrxSUSunxijTPU/o?=
 =?us-ascii?Q?iDcfWFkik8ksF1YcKb7NrOfbhmNSl0E1kPt+V2S+cX0E9960BD45kqwfgxkg?=
 =?us-ascii?Q?HD0v67pisj620TnlDYkGXsTFlu8M6rYH87DWD3+e?=
X-OriginatorOrg: vivo.com
X-MS-Exchange-CrossTenant-Network-Message-Id: 5ee7c794-761d-4d40-efc2-08dc787e6d04
X-MS-Exchange-CrossTenant-AuthSource: KL1PR06MB7401.apcprd06.prod.outlook.com
X-MS-Exchange-CrossTenant-AuthAs: Internal
X-MS-Exchange-CrossTenant-OriginalArrivalTime: 20 May 2024 03:39:17.1493
 (UTC)
X-MS-Exchange-CrossTenant-FromEntityHeader: Hosted
X-MS-Exchange-CrossTenant-Id: 923e42dc-48d5-4cbe-b582-1a797a6412ed
X-MS-Exchange-CrossTenant-MailboxType: HOSTED
X-MS-Exchange-CrossTenant-UserPrincipalName: VdHBC140FjscUC+x08sOgYh8RFHaeUQe9MhFu6bRHcWBpeKwXwmMyvtHf8Zz2+iSXEq/aJdesiMGIor+ifoSkw==
X-MS-Exchange-Transport-CrossTenantHeadersStamped: SEZPR06MB5023

The depth is 62, and the wake_batch is 8. In the following situation,
the task would hang forever.

  t1:                 t2:                          t3:
  blk_mq_get_tag      .                            .
  io_schedule         .                            .
                      elevator_switch              .
                      blk_mq_freeze_queue          .
                      blk_freeze_queue_start       .
                      blk_mq_freeze_queue_wait     .
                                                   blk_mq_submit_bio
                                                   __bio_queue_enter

Fix this issue by waking up all the waiters sleeping on tags after
freezing the queue.

Signed-off-by: Yang Yang <yang.yang@vivo.com>
---
 block/blk-core.c | 2 --
 block/blk-mq.c   | 4 +++-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index a16b5abdbbf5..e1eacfad6e5b 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -298,8 +298,6 @@ void blk_queue_start_drain(struct request_queue *q)
 	 * prevent I/O from crossing blk_queue_enter().
 	 */
 	blk_freeze_queue_start(q);
-	if (queue_is_mq(q))
-		blk_mq_wake_waiters(q);
 	/* Make blk_queue_enter() reexamine the DYING flag. */
 	wake_up_all(&q->mq_freeze_wq);
 }
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 4ecb9db62337..9eb3139e713a 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -125,8 +125,10 @@ void blk_freeze_queue_start(struct request_queue *q)
 	if (++q->mq_freeze_depth == 1) {
 		percpu_ref_kill(&q->q_usage_counter);
 		mutex_unlock(&q->mq_freeze_lock);
-		if (queue_is_mq(q))
+		if (queue_is_mq(q)) {
+			blk_mq_wake_waiters(q);
 			blk_mq_run_hw_queues(q, false);
+		}
 	} else {
 		mutex_unlock(&q->mq_freeze_lock);
 	}
-- 
2.34.1