From nobody Mon Nov 25 23:37:37 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id EE55719F13F for ; Wed, 23 Oct 2024 09:55:05 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677307; cv=none; b=DSeFxr1Vv2tYYDSOPtLrWFTTuYdXXJKrlS2NLdK8DQ4bgOKhjUNOY3ebOMgSjc9pUAJf+7HJ/ynnrpIafU1zakmwc2W7RBjHyY5Xw3NS//x7ONszbtuG8VCPKrVNvWdzvKpT/Q+pgsW0FvZ8JDRU5eSzp040tm9rgupfAL/ei9c= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677307; c=relaxed/simple; bh=dEtJywqzX9NlCIsXbk6wbkz14aF8BZqWYkvg6LKGRZk=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=pll2ASJmSa31a8CKgfCuGTwH3O9zRJQeRA34JxIgH/JAeFkDSYQAVAQse9AU2vV4BSMFYobR7ROzj6NAczJuZ6t28knMpkjCNlngck3jWjoSLs138wvz4LwGTh390+mRwlrk1sZNomltkLCBja35CnpTaW5Fj9oEXhPTl8n3Ad8= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=ULlnjceg; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="ULlnjceg" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1729677305; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=RDWZQpe65meGR6CH+Ygs8rTGaxpZEfZwq23nZiUncY8=; b=ULlnjceg72UiAq4tdPx7tVxxbd4XSt0WFxk/J0Hef952w2foQMcRC+jqgMKNMQvd/0nK/P x5qR3yalPAjmyyY5zPIpbl3g7hUOgEF2ikAvMo1YN8iLTUEB8HyJVhhFhw8r+76pc9PV2h FqsGAh740aQYwHxTKcy37aq853XIyCY= Received: from mx-prod-mc-01.mail-002.prod.us-west-2.aws.redhat.com (ec2-54-186-198-63.us-west-2.compute.amazonaws.com [54.186.198.63]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-13-dw2ZXFFvPeS1l5XK2Hi2tg-1; Wed, 23 Oct 2024 05:55:01 -0400 X-MC-Unique: dw2ZXFFvPeS1l5XK2Hi2tg-1 Received: from mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.4]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-01.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id CAD9419560AF; Wed, 23 Oct 2024 09:54:59 +0000 (UTC) Received: from localhost (unknown [10.72.116.171]) by mx-prod-int-01.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id 0EA06300018D; Wed, 23 Oct 2024 09:54:56 +0000 (UTC) From: Ming Lei To: Jens Axboe , linux-block@vger.kernel.org Cc: Christoph Hellwig , Peter Zijlstra , Waiman Long , Boqun Feng , Ingo Molnar , Will Deacon , linux-kernel@vger.kernel.org, Bart Van Assche , Ming Lei Subject: [PATCH 1/3] blk-mq: add non_owner variant of start_freeze/unfreeze queue APIs Date: Wed, 23 Oct 2024 17:54:33 +0800 Message-ID: <20241023095438.3451156-2-ming.lei@redhat.com> In-Reply-To: <20241023095438.3451156-1-ming.lei@redhat.com> References: <20241023095438.3451156-1-ming.lei@redhat.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.4.1 on 10.30.177.4 Content-Type: text/plain; charset="utf-8" Add non_owner variant of start_freeze/unfreeze queue APIs, so that the caller knows that what they are doing, and we can skip lockdep support for non_owner variant in per-call level. Prepare for supporting lockdep for freezing/unfreezing queue. Suggested-by: Christoph Hellwig Signed-off-by: Ming Lei Reviewed-by: Christoph Hellwig --- block/blk-mq.c | 19 +++++++++++++++++++ include/linux/blk-mq.h | 2 ++ 2 files changed, 21 insertions(+) diff --git a/block/blk-mq.c b/block/blk-mq.c index 4b2c8e940f59..2c84c2d2510d 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -196,6 +196,25 @@ void blk_mq_unfreeze_queue(struct request_queue *q) } EXPORT_SYMBOL_GPL(blk_mq_unfreeze_queue); =20 +/* + * non_owner variant of blk_freeze_queue_start + * + * The queue needn't to be unfreeze in current task, and non-owner use is + * fragile and easy to cause race, please try your best to avoid it + */ +void blk_freeze_queue_start_non_owner(struct request_queue *q) +{ + blk_freeze_queue_start(q); +} +EXPORT_SYMBOL_GPL(blk_freeze_queue_start_non_owner); + +/* non_owner variant of blk_mq_unfreeze_queue */ +void blk_mq_unfreeze_queue_non_owner(struct request_queue *q) +{ + __blk_mq_unfreeze_queue(q, false); +} +EXPORT_SYMBOL_GPL(blk_mq_unfreeze_queue_non_owner); + /* * FIXME: replace the scsi_internal_device_*block_nowait() calls in the * mpt3sas driver such that this function can be removed. diff --git a/include/linux/blk-mq.h b/include/linux/blk-mq.h index 4fecf46ef681..c5063e0a38a0 100644 --- a/include/linux/blk-mq.h +++ b/include/linux/blk-mq.h @@ -925,6 +925,8 @@ void blk_freeze_queue_start(struct request_queue *q); void blk_mq_freeze_queue_wait(struct request_queue *q); int blk_mq_freeze_queue_wait_timeout(struct request_queue *q, unsigned long timeout); +void blk_mq_unfreeze_queue_non_owner(struct request_queue *q); +void blk_freeze_queue_start_non_owner(struct request_queue *q); =20 void blk_mq_map_queues(struct blk_mq_queue_map *qmap); void blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, int nr_hw_queu= es); --=20 2.46.0 From nobody Mon Nov 25 23:37:37 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 10C481A7271 for ; Wed, 23 Oct 2024 09:55:09 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677311; cv=none; b=RYfempuRqz0nVEwoqBw5jOghRq7nSmwUUh0UriUb8s9zgUI/jjYsKAWNUx0aRhSuc0oy0kMD0HeGuo0C4Nniaj95OX3t8HwckvgNF/8XLfB8VTFJY5DK/VDgd9Y+gBdkAv4Pk6IhL9kexW+JY943xilZfWty03CEP5WcwFK/ovI= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677311; c=relaxed/simple; bh=zlIJnV19+kvfCZ475p6XXehMNRmYT+ypXvfSmPyDANE=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=WWAPUwMi1FffjuiEoBqXejl9IuVnSnYJ9vUFOOL2nax2Pba4yNgfK6nuoysPGk/jFXCbB/grNFmOthS/RfGw19KZqAOBmkZ8msENQvcd4nqDRhaMHAK2sAtGXbskCVitQvr9BmNmXRaFiKDHBAAEJU1RckeTV0ZqtMLc+IVPTAs= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=JWt00fWp; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="JWt00fWp" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1729677309; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=3frOuUOwN8tRn7DUXsZgTp8QrBlHIkulSdEobuUctI4=; b=JWt00fWpQn88yJOeuajV31sl9B7zupf2Ogd0ulxEeGuFE9OVGRzm215CVpiC+q/GaQM1Nc iH2PC0lwHD2DB9At7z1ybAakPHh1NQ12d35sieGfVoCRkjF9tZ4+gHL9i/pufl/YyLQ7VM 1nCt9flVp0Ypb2I0LNIswFnLhAkypR8= Received: from mx-prod-mc-01.mail-002.prod.us-west-2.aws.redhat.com (ec2-54-186-198-63.us-west-2.compute.amazonaws.com [54.186.198.63]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-517-BheH5eESPUSHqLTG0XYwCg-1; Wed, 23 Oct 2024 05:55:05 -0400 X-MC-Unique: BheH5eESPUSHqLTG0XYwCg-1 Received: from mx-prod-int-04.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-04.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.40]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-01.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id 738E71955F3C; Wed, 23 Oct 2024 09:55:04 +0000 (UTC) Received: from localhost (unknown [10.72.116.171]) by mx-prod-int-04.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id D8C4019560AE; Wed, 23 Oct 2024 09:55:02 +0000 (UTC) From: Ming Lei To: Jens Axboe , linux-block@vger.kernel.org Cc: Christoph Hellwig , Peter Zijlstra , Waiman Long , Boqun Feng , Ingo Molnar , Will Deacon , linux-kernel@vger.kernel.org, Bart Van Assche , Ming Lei Subject: [PATCH 2/3] nvme: core: switch to non_owner variant of start_freeze/unfreeze queue Date: Wed, 23 Oct 2024 17:54:34 +0800 Message-ID: <20241023095438.3451156-3-ming.lei@redhat.com> In-Reply-To: <20241023095438.3451156-1-ming.lei@redhat.com> References: <20241023095438.3451156-1-ming.lei@redhat.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.0 on 10.30.177.40 Content-Type: text/plain; charset="utf-8" nvme_start_freeze() and nvme_unfreeze() may be called from same context, so switch them to call non_owner variant of start_freeze/unfreeze queue. Signed-off-by: Ming Lei Reviewed-by: Christoph Hellwig --- drivers/nvme/host/core.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c index ba6508455e18..06c1e4e8456f 100644 --- a/drivers/nvme/host/core.c +++ b/drivers/nvme/host/core.c @@ -4871,7 +4871,7 @@ void nvme_unfreeze(struct nvme_ctrl *ctrl) =20 srcu_idx =3D srcu_read_lock(&ctrl->srcu); list_for_each_entry_rcu(ns, &ctrl->namespaces, list) - blk_mq_unfreeze_queue(ns->queue); + blk_mq_unfreeze_queue_non_owner(ns->queue); srcu_read_unlock(&ctrl->srcu, srcu_idx); clear_bit(NVME_CTRL_FROZEN, &ctrl->flags); } @@ -4913,7 +4913,7 @@ void nvme_start_freeze(struct nvme_ctrl *ctrl) set_bit(NVME_CTRL_FROZEN, &ctrl->flags); srcu_idx =3D srcu_read_lock(&ctrl->srcu); list_for_each_entry_rcu(ns, &ctrl->namespaces, list) - blk_freeze_queue_start(ns->queue); + blk_freeze_queue_start_non_owner(ns->queue); srcu_read_unlock(&ctrl->srcu, srcu_idx); } EXPORT_SYMBOL_GPL(nvme_start_freeze); --=20 2.46.0 From nobody Mon Nov 25 23:37:37 2024 Received: from us-smtp-delivery-124.mimecast.com (us-smtp-delivery-124.mimecast.com [170.10.129.124]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CFB9A1A7AF7 for ; Wed, 23 Oct 2024 09:55:14 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=170.10.129.124 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677317; cv=none; b=h64J0s6tgmuqPTw6W7HxZkPPeq+9pDHsX+3ohOlY71SiBUFk1Ri+Uj0UtlManT2xsVdYL0iqRCKn3DNHAJvgKXglnzr4n9FuBG2cqh1szUsGbot0Q8q4zEtKdW3a2aax6XwkclIbLnXXL4ytYQUawVoQU9g4jgs1k4fQ3VkQKdU= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1729677317; c=relaxed/simple; bh=+FVdnVre7ofwoYH68Sy4BzXcOlIiJ2nUaR3uwNdr1gw=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=s37bvGHTS38ukKcKUaCvG/6BXxa10G0/LTdaClaT3AHfSIg7kiE9yNGeYQ/ldbs+vTrKqL64lVUcTBPwf8zzC6HkWGrcw8OYcldf8ojQg1Z7olphxN5hEWDX8YbCshwiugVsn/palp4MrMDsDQHjdW4S3RMZvFy8YLAxUMORANo= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com; spf=pass smtp.mailfrom=redhat.com; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b=hqnjyEJ7; arc=none smtp.client-ip=170.10.129.124 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=redhat.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=redhat.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=redhat.com header.i=@redhat.com header.b="hqnjyEJ7" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=redhat.com; s=mimecast20190719; t=1729677313; h=from:from:reply-to:subject:subject:date:date:message-id:message-id: to:to:cc:cc:mime-version:mime-version: content-transfer-encoding:content-transfer-encoding: in-reply-to:in-reply-to:references:references; bh=6fChrjJ4PBXmjzA74oq7U7p9U5IYVMMbrVBQiegw/Jo=; b=hqnjyEJ7Ncf+mC+m6K6VCj9eDQHlYWmzA96uI9dnA91mMMzX0+JSwyP2JvNgca+roQ9smM cOeKfjxTXRjzgxft3IoINhl/G7RhpEf8AJygZ8oipICRr+rtSDr38Iyb13j0j2WVG/Kmao Wg5PPoM2bYIDGiR8kl+LiTc74RqMueU= Received: from mx-prod-mc-05.mail-002.prod.us-west-2.aws.redhat.com (ec2-54-186-198-63.us-west-2.compute.amazonaws.com [54.186.198.63]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-655-JsdOgAHqOP2Uo6xuvAd7IA-1; Wed, 23 Oct 2024 05:55:10 -0400 X-MC-Unique: JsdOgAHqOP2Uo6xuvAd7IA-1 Received: from mx-prod-int-05.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-05.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.17]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-05.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id BA19E1956083; Wed, 23 Oct 2024 09:55:08 +0000 (UTC) Received: from localhost (unknown [10.72.116.171]) by mx-prod-int-05.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id 9D8D81956056; Wed, 23 Oct 2024 09:55:07 +0000 (UTC) From: Ming Lei To: Jens Axboe , linux-block@vger.kernel.org Cc: Christoph Hellwig , Peter Zijlstra , Waiman Long , Boqun Feng , Ingo Molnar , Will Deacon , linux-kernel@vger.kernel.org, Bart Van Assche , Ming Lei Subject: [PATCH 3/3] block: model freeze & enter queue as lock for supporting lockdep Date: Wed, 23 Oct 2024 17:54:35 +0800 Message-ID: <20241023095438.3451156-4-ming.lei@redhat.com> In-Reply-To: <20241023095438.3451156-1-ming.lei@redhat.com> References: <20241023095438.3451156-1-ming.lei@redhat.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-Scanned-By: MIMEDefang 3.0 on 10.30.177.17 Content-Type: text/plain; charset="utf-8" Recently we got several deadlock report[1][2][3] caused by blk_mq_freeze_qu= eue and blk_enter_queue(). Turns out the two are just like acquiring read/write lock, so model them as read/write lock for supporting lockdep: 1) model q->q_usage_counter as two locks(io and queue lock) - queue lock covers sync with blk_enter_queue() - io lock covers sync with bio_enter_queue() 2) make the lockdep class/key as per-queue: - different subsystem has very different lock use pattern, shared lock class causes false positive easily - freeze_queue degrades to no lock in case that disk state becomes DEAD bec= ause bio_enter_queue() won't be blocked any more - freeze_queue degrades to no lock in case that request queue becomes dying because blk_enter_queue() won't be blocked any more 3) model blk_mq_freeze_queue() as acquire_exclusive & try_lock - it is exclusive lock, so dependency with blk_enter_queue() is covered - it is trylock because blk_mq_freeze_queue() are allowed to run concurrent= ly 4) model blk_enter_queue() & bio_enter_queue() as acquire_read() - nested blk_enter_queue() are allowed - dependency with blk_mq_freeze_queue() is covered - blk_queue_exit() is often called from other contexts(such as irq), and it can't be annotated as lock_release(), so simply do it in blk_enter_queue(), this way still covered cases as many as possible With lockdep support, such kind of reports may be reported asap and needn't wait until the real deadlock is triggered. For example, the following lockdep report can be triggered in the report[3]. [ 31.671822] =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D [ 31.673169] WARNING: possible circular locking dependency detected [ 31.674456] 6.11.0_nbd+ #411 Not tainted [ 31.675220] ------------------------------------------------------ [ 31.676379] bash/1425 is trying to acquire lock: [ 31.676861] ffff990b8ea27530 (&q->limits_lock){+.+.}-{3:3}, at: queue_wc= _store+0x8e/0x180 [ 31.677268] but task is already holding lock: [ 31.677548] ffff990b8ea27410 (&q->sysfs_lock){+.+.}-{3:3}, at: queue_att= r_store+0x75/0xc0 [ 31.677931] which lock already depends on the new lock. [ 31.678315] the existing dependency chain (in reverse order) is: [ 31.678664] -> #2 (&q->sysfs_lock){+.+.}-{3:3}: [ 31.678951] __mutex_lock+0xad/0xb20 [ 31.679157] queue_attr_store+0x75/0xc0 [ 31.679366] kernfs_fop_write_iter+0x15c/0x210 [ 31.679608] vfs_write+0x2a7/0x540 [ 31.679801] ksys_write+0x75/0x100 [ 31.679999] do_syscall_64+0x95/0x180 [ 31.680209] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 31.680488] -> #1 (&q->q_usage_counter(queue)#2){++++}-{0:0}: [ 31.680839] blk_queue_enter+0x195/0x1d0 [ 31.681060] blk_mq_alloc_request+0x136/0x2d0 [ 31.681301] scsi_execute_cmd+0x9c/0x4c0 [ 31.681528] read_capacity_16+0x116/0x410 [ 31.681765] sd_revalidate_disk.isra.0+0x54d/0x2f00 [ 31.682044] sd_probe+0x2ec/0x520 [ 31.682238] really_probe+0xd3/0x390 [ 31.682445] __driver_probe_device+0x78/0x150 [ 31.682682] driver_probe_device+0x1f/0x90 [ 31.682908] __device_attach_driver+0x89/0x110 [ 31.683161] bus_for_each_drv+0x95/0xf0 [ 31.683377] __device_attach_async_helper+0xa7/0xf0 [ 31.683639] async_run_entry_fn+0x31/0x130 [ 31.683875] process_one_work+0x212/0x700 [ 31.684100] worker_thread+0x1ce/0x380 [ 31.684308] kthread+0xd2/0x110 [ 31.684490] ret_from_fork+0x31/0x50 [ 31.684700] ret_from_fork_asm+0x1a/0x30 [ 31.684922] -> #0 (&q->limits_lock){+.+.}-{3:3}: [ 31.685499] __lock_acquire+0x15c0/0x23e0 [ 31.685872] lock_acquire+0xd8/0x300 [ 31.686207] __mutex_lock+0xad/0xb20 [ 31.686535] queue_wc_store+0x8e/0x180 [ 31.686877] queue_attr_store+0x84/0xc0 [ 31.687231] kernfs_fop_write_iter+0x15c/0x210 [ 31.687594] vfs_write+0x2a7/0x540 [ 31.687907] ksys_write+0x75/0x100 [ 31.688219] do_syscall_64+0x95/0x180 [ 31.688534] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 31.688910] other info that might help us debug this: [ 31.689621] Chain exists of: &q->limits_lock --> &q->q_usage_counter(queue)#2 --> &q->s= ysfs_lock [ 31.690549] Possible unsafe locking scenario: [ 31.691060] CPU0 CPU1 [ 31.691389] ---- ---- [ 31.691716] lock(&q->sysfs_lock); [ 31.691999] lock(&q->q_usage_counter(queu= e)#2); [ 31.692460] lock(&q->sysfs_lock); [ 31.692863] lock(&q->limits_lock); [ 31.693155] *** DEADLOCK *** [ 31.693746] 6 locks held by bash/1425: [ 31.694043] #0: ffff990b8007e420 (sb_writers#4){.+.+}-{0:0}, at: ksys_w= rite+0x75/0x100 [ 31.694543] #1: ffff990bcf1a3288 (&of->mutex#2){+.+.}-{3:3}, at: kernfs= _fop_write_iter+0x115/0x210 [ 31.695119] #2: ffff990b91888378 (kn->active#166){.+.+}-{0:0}, at: kern= fs_fop_write_iter+0x11e/0x210 [ 31.695685] #3: ffff990b8ea26ee8 (&q->q_usage_counter(io)#2){++++}-{0:0= }, at: queue_attr_store+0x60/0xc0 [ 31.696269] #4: ffff990b8ea26f20 (&q->q_usage_counter(queue)#2){++++}-{= 0:0}, at: queue_attr_store+0x60/0xc0 [ 31.696846] #5: ffff990b8ea27410 (&q->sysfs_lock){+.+.}-{3:3}, at: queu= e_attr_store+0x75/0xc0 [ 31.697381] stack backtrace: [ 31.697826] CPU: 9 UID: 0 PID: 1425 Comm: bash Not tainted 6.11.0_nbd+ #= 411 [ 31.698285] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.1= 6.3-1.fc39 04/01/2014 [ 31.698807] Call Trace: [ 31.699058] [ 31.699289] dump_stack_lvl+0x93/0xf0 [ 31.699598] print_circular_bug+0x26e/0x340 [ 31.699924] check_noncircular+0x16c/0x190 [ 31.700251] ? lock_acquire+0x2a1/0x300 [ 31.700561] __lock_acquire+0x15c0/0x23e0 [ 31.700877] lock_acquire+0xd8/0x300 [ 31.701181] ? queue_wc_store+0x8e/0x180 [ 31.701502] __mutex_lock+0xad/0xb20 [ 31.701806] ? queue_wc_store+0x8e/0x180 [ 31.702128] ? queue_wc_store+0x8e/0x180 [ 31.702446] ? queue_wc_store+0x8e/0x180 [ 31.702761] queue_wc_store+0x8e/0x180 [ 31.703084] ? __mutex_lock+0xad/0xb20 [ 31.703385] ? __mutex_lock+0x6e4/0xb20 [ 31.703691] ? mark_held_locks+0x40/0x70 [ 31.704004] ? queue_attr_store+0x75/0xc0 [ 31.704317] queue_attr_store+0x84/0xc0 [ 31.704643] kernfs_fop_write_iter+0x15c/0x210 [ 31.704987] vfs_write+0x2a7/0x540 [ 31.705274] ksys_write+0x75/0x100 [ 31.705559] do_syscall_64+0x95/0x180 [ 31.705864] ? do_user_addr_fault+0x361/0x790 [ 31.706239] ? trace_hardirqs_off+0x4b/0xc0 [ 31.706564] ? clear_bhb_loop+0x25/0x80 [ 31.706966] ? clear_bhb_loop+0x25/0x80 [ 31.707272] ? clear_bhb_loop+0x25/0x80 [ 31.707568] entry_SYSCALL_64_after_hwframe+0x76/0x7e [ 31.707927] RIP: 0033:0x7fb69d85e174 [ 31.708227] Code: 89 02 48 c7 c0 ff ff ff ff eb bd 66 2e 0f 1f 84 00 00 = 00 00 00 90 f3 0f 1e fa 80 3d 6d b4 0d 00 00 74 13 b8 01 00 00 00 0f 05 <48= > 3d 00 f0 ff ff 77 54 c3 0f 1f 00 55 48 89 e5 48 83 ec 20 48 89 [ 31.709343] RSP: 002b:00007ffed933fb48 EFLAGS: 00000202 ORIG_RAX: 000000= 0000000001 [ 31.709834] RAX: ffffffffffffffda RBX: 0000000000000005 RCX: 00007fb69d8= 5e174 [ 31.710311] RDX: 0000000000000005 RSI: 000055713d6fb7d0 RDI: 00000000000= 00001 [ 31.710779] RBP: 00007ffed933fb70 R08: 0000000000000073 R09: 00000000000= 00001 [ 31.711263] R10: 0000000000000000 R11: 0000000000000202 R12: 00000000000= 00005 [ 31.711755] R13: 000055713d6fb7d0 R14: 00007fb69d932780 R15: 00000000000= 00005 [ 31.712242] [1] occasional block layer hang when setting 'echo noop > /sys/block/sda/qu= eue/scheduler' https://bugzilla.kernel.org/show_bug.cgi?id=3D219166 [2] del_gendisk() vs blk_queue_enter() race condition https://lore.kernel.org/linux-block/20241003085610.GK11458@google.com/ [3] queue_freeze & queue_enter deadlock in scsi https://lore.kernel.org/linux-block/ZxG38G9BuFdBpBHZ@fedora/T/#u Cc: Christoph Hellwig Signed-off-by: Ming Lei Reviewed-by: Christoph Hellwig --- block/blk-core.c | 18 ++++++++++++++++-- block/blk-mq.c | 26 ++++++++++++++++++++++---- block/blk.h | 29 ++++++++++++++++++++++++++--- block/genhd.c | 15 +++++++++++---- include/linux/blkdev.h | 6 ++++++ 5 files changed, 81 insertions(+), 13 deletions(-) diff --git a/block/blk-core.c b/block/blk-core.c index bc5e8c5eaac9..09d10bb95fda 100644 --- a/block/blk-core.c +++ b/block/blk-core.c @@ -261,6 +261,8 @@ static void blk_free_queue(struct request_queue *q) blk_mq_release(q); =20 ida_free(&blk_queue_ida, q->id); + lockdep_unregister_key(&q->io_lock_cls_key); + lockdep_unregister_key(&q->q_lock_cls_key); call_rcu(&q->rcu_head, blk_free_queue_rcu); } =20 @@ -278,18 +280,20 @@ void blk_put_queue(struct request_queue *q) } EXPORT_SYMBOL(blk_put_queue); =20 -void blk_queue_start_drain(struct request_queue *q) +bool blk_queue_start_drain(struct request_queue *q) { /* * When queue DYING flag is set, we need to block new req * entering queue, so we call blk_freeze_queue_start() to * prevent I/O from crossing blk_queue_enter(). */ - blk_freeze_queue_start(q); + bool freeze =3D __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); + + return freeze; } =20 /** @@ -321,6 +325,8 @@ int blk_queue_enter(struct request_queue *q, blk_mq_req= _flags_t flags) return -ENODEV; } =20 + rwsem_acquire_read(&q->q_lockdep_map, 0, 0, _RET_IP_); + rwsem_release(&q->q_lockdep_map, _RET_IP_); return 0; } =20 @@ -352,6 +358,8 @@ int __bio_queue_enter(struct request_queue *q, struct b= io *bio) goto dead; } =20 + rwsem_acquire_read(&q->io_lockdep_map, 0, 0, _RET_IP_); + rwsem_release(&q->io_lockdep_map, _RET_IP_); return 0; dead: bio_io_error(bio); @@ -441,6 +449,12 @@ struct request_queue *blk_alloc_queue(struct queue_lim= its *lim, int node_id) PERCPU_REF_INIT_ATOMIC, GFP_KERNEL); if (error) goto fail_stats; + lockdep_register_key(&q->io_lock_cls_key); + lockdep_register_key(&q->q_lock_cls_key); + lockdep_init_map(&q->io_lockdep_map, "&q->q_usage_counter(io)", + &q->io_lock_cls_key, 0); + lockdep_init_map(&q->q_lockdep_map, "&q->q_usage_counter(queue)", + &q->q_lock_cls_key, 0); =20 q->nr_requests =3D BLKDEV_DEFAULT_RQ; =20 diff --git a/block/blk-mq.c b/block/blk-mq.c index 2c84c2d2510d..60c58142819e 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -120,17 +120,29 @@ void blk_mq_in_flight_rw(struct request_queue *q, str= uct block_device *part, inflight[1] =3D mi.inflight[1]; } =20 -void blk_freeze_queue_start(struct request_queue *q) +bool __blk_freeze_queue_start(struct request_queue *q) { + int freeze; + mutex_lock(&q->mq_freeze_lock); if (++q->mq_freeze_depth =3D=3D 1) { percpu_ref_kill(&q->q_usage_counter); mutex_unlock(&q->mq_freeze_lock); if (queue_is_mq(q)) blk_mq_run_hw_queues(q, false); + freeze =3D true; } else { mutex_unlock(&q->mq_freeze_lock); + freeze =3D false; } + + return freeze; +} + +void blk_freeze_queue_start(struct request_queue *q) +{ + if (__blk_freeze_queue_start(q)) + blk_freeze_acquire_lock(q, false, false); } EXPORT_SYMBOL_GPL(blk_freeze_queue_start); =20 @@ -176,8 +188,10 @@ void blk_mq_freeze_queue(struct request_queue *q) } EXPORT_SYMBOL_GPL(blk_mq_freeze_queue); =20 -void __blk_mq_unfreeze_queue(struct request_queue *q, bool force_atomic) +bool __blk_mq_unfreeze_queue(struct request_queue *q, bool force_atomic) { + int unfreeze =3D false; + mutex_lock(&q->mq_freeze_lock); if (force_atomic) q->q_usage_counter.data->force_atomic =3D true; @@ -186,13 +200,17 @@ void __blk_mq_unfreeze_queue(struct request_queue *q,= bool force_atomic) if (!q->mq_freeze_depth) { percpu_ref_resurrect(&q->q_usage_counter); wake_up_all(&q->mq_freeze_wq); + unfreeze =3D true; } mutex_unlock(&q->mq_freeze_lock); + + return unfreeze; } =20 void blk_mq_unfreeze_queue(struct request_queue *q) { - __blk_mq_unfreeze_queue(q, false); + if (__blk_mq_unfreeze_queue(q, false)) + blk_unfreeze_release_lock(q, false, false); } EXPORT_SYMBOL_GPL(blk_mq_unfreeze_queue); =20 @@ -204,7 +222,7 @@ EXPORT_SYMBOL_GPL(blk_mq_unfreeze_queue); */ void blk_freeze_queue_start_non_owner(struct request_queue *q) { - blk_freeze_queue_start(q); + __blk_freeze_queue_start(q); } EXPORT_SYMBOL_GPL(blk_freeze_queue_start_non_owner); =20 diff --git a/block/blk.h b/block/blk.h index c718e4291db0..832e54c5a271 100644 --- a/block/blk.h +++ b/block/blk.h @@ -4,6 +4,7 @@ =20 #include #include +#include #include /* for max_pfn/max_low_pfn */ #include #include @@ -35,8 +36,9 @@ struct blk_flush_queue *blk_alloc_flush_queue(int node, i= nt cmd_size, void blk_free_flush_queue(struct blk_flush_queue *q); =20 void blk_freeze_queue(struct request_queue *q); -void __blk_mq_unfreeze_queue(struct request_queue *q, bool force_atomic); -void blk_queue_start_drain(struct request_queue *q); +bool __blk_mq_unfreeze_queue(struct request_queue *q, bool force_atomic); +bool blk_queue_start_drain(struct request_queue *q); +bool __blk_freeze_queue_start(struct request_queue *q); int __bio_queue_enter(struct request_queue *q, struct bio *bio); void submit_bio_noacct_nocheck(struct bio *bio); void bio_await_chain(struct bio *bio); @@ -69,8 +71,11 @@ static inline int bio_queue_enter(struct bio *bio) { struct request_queue *q =3D bdev_get_queue(bio->bi_bdev); =20 - if (blk_try_enter_queue(q, false)) + if (blk_try_enter_queue(q, false)) { + rwsem_acquire_read(&q->io_lockdep_map, 0, 0, _RET_IP_); + rwsem_release(&q->io_lockdep_map, _RET_IP_); return 0; + } return __bio_queue_enter(q, bio); } =20 @@ -734,4 +739,22 @@ void blk_integrity_verify(struct bio *bio); void blk_integrity_prepare(struct request *rq); void blk_integrity_complete(struct request *rq, unsigned int nr_bytes); =20 +static inline void blk_freeze_acquire_lock(struct request_queue *q, bool + disk_dead, bool queue_dying) +{ + if (!disk_dead) + rwsem_acquire(&q->io_lockdep_map, 0, 1, _RET_IP_); + if (!queue_dying) + rwsem_acquire(&q->q_lockdep_map, 0, 1, _RET_IP_); +} + +static inline void blk_unfreeze_release_lock(struct request_queue *q, bool + disk_dead, bool queue_dying) +{ + if (!queue_dying) + rwsem_release(&q->q_lockdep_map, _RET_IP_); + if (!disk_dead) + rwsem_release(&q->io_lockdep_map, _RET_IP_); +} + #endif /* BLK_INTERNAL_H */ diff --git a/block/genhd.c b/block/genhd.c index 1c05dd4c6980..6ad3fcde0110 100644 --- a/block/genhd.c +++ b/block/genhd.c @@ -581,13 +581,13 @@ static void blk_report_disk_dead(struct gendisk *disk= , bool surprise) rcu_read_unlock(); } =20 -static void __blk_mark_disk_dead(struct gendisk *disk) +static bool __blk_mark_disk_dead(struct gendisk *disk) { /* * Fail any new I/O. */ if (test_and_set_bit(GD_DEAD, &disk->state)) - return; + return false; =20 if (test_bit(GD_OWNS_QUEUE, &disk->state)) blk_queue_flag_set(QUEUE_FLAG_DYING, disk->queue); @@ -600,7 +600,7 @@ static void __blk_mark_disk_dead(struct gendisk *disk) /* * Prevent new I/O from crossing bio_queue_enter(). */ - blk_queue_start_drain(disk->queue); + return blk_queue_start_drain(disk->queue); } =20 /** @@ -641,6 +641,7 @@ void del_gendisk(struct gendisk *disk) struct request_queue *q =3D disk->queue; struct block_device *part; unsigned long idx; + bool start_drain, queue_dying; =20 might_sleep(); =20 @@ -668,7 +669,10 @@ void del_gendisk(struct gendisk *disk) * Drop all partitions now that the disk is marked dead. */ mutex_lock(&disk->open_mutex); - __blk_mark_disk_dead(disk); + start_drain =3D __blk_mark_disk_dead(disk); + queue_dying =3D blk_queue_dying(q); + if (start_drain) + blk_freeze_acquire_lock(q, true, queue_dying); xa_for_each_start(&disk->part_tbl, idx, part, 1) drop_partition(part); mutex_unlock(&disk->open_mutex); @@ -725,6 +729,9 @@ void del_gendisk(struct gendisk *disk) if (queue_is_mq(q)) blk_mq_exit_queue(q); } + + if (start_drain) + blk_unfreeze_release_lock(q, true, queue_dying); } EXPORT_SYMBOL(del_gendisk); =20 diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 50c3b959da28..57f1ee386b57 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -25,6 +25,7 @@ #include #include #include +#include =20 struct module; struct request_queue; @@ -471,6 +472,11 @@ struct request_queue { struct xarray hctx_table; =20 struct percpu_ref q_usage_counter; + struct lock_class_key io_lock_cls_key; + struct lockdep_map io_lockdep_map; + + struct lock_class_key q_lock_cls_key; + struct lockdep_map q_lockdep_map; =20 struct request *last_merge; =20 --=20 2.46.0