From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755678AbbA2MSA (ORCPT ); Thu, 29 Jan 2015 07:18:00 -0500 Received: from mail-oi0-f50.google.com ([209.85.218.50]:64389 "EHLO mail-oi0-f50.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753152AbbA2MR5 (ORCPT ); Thu, 29 Jan 2015 07:17:57 -0500 From: Ming Lei To: Jens Axboe , linux-kernel@vger.kernel.org Cc: Christoph Hellwig , Bart Van Assche , Sasha Levin , Ming Lei Subject: [PATCH 2/2] blk-mq: release mq's kobjects in blk_release_queue() Date: Thu, 29 Jan 2015 20:17:27 +0800 Message-Id: <1422533847-31726-3-git-send-email-ming.lei@canonical.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1422533847-31726-1-git-send-email-ming.lei@canonical.com> References: <1422533847-31726-1-git-send-email-ming.lei@canonical.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The kobject memory inside blk-mq hctx/ctx shouldn't have been freed before the kobject is released because driver core can access it freely before its release. We can't do that in all ctx/hctx/mq_kobj's release handler because it can be run before blk_cleanup_queue(). Given mq_kobj shouldn't have been introduced, this patch simply moves mq's release into blk_release_queue(). Reported-by: Sasha Levin Signed-off-by: Ming Lei --- block/blk-mq.c | 29 ++++++++++++++++++++++------- block/blk-mq.h | 2 ++ block/blk-sysfs.c | 2 ++ 3 files changed, 26 insertions(+), 7 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index a880b6c..9b250da 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -1644,10 +1644,8 @@ static void blk_mq_free_hw_queues(struct request_queue *q, struct blk_mq_hw_ctx *hctx; unsigned int i; - queue_for_each_hw_ctx(q, hctx, i) { + queue_for_each_hw_ctx(q, hctx, i) free_cpumask_var(hctx->cpumask); - kfree(hctx); - } } static int blk_mq_init_hctx(struct request_queue *q, @@ -1872,6 +1870,27 @@ static void blk_mq_add_queue_tag_set(struct blk_mq_tag_set *set, mutex_unlock(&set->tag_list_lock); } +/* + * It is the actual release handler for mq, but we do it from + * request queue's release handler for avoiding use-after-free + * and headache because q->mq_kobj shouldn't have been introduced, + * but we can't group ctx/kctx kobj without it. + */ +void blk_mq_release(struct request_queue *q) +{ + struct blk_mq_hw_ctx *hctx; + unsigned int i; + + /* hctx kobj stays in hctx */ + queue_for_each_hw_ctx(q, hctx, i) + kfree(hctx); + + kfree(q->queue_hw_ctx); + + /* ctx kobj stays in queue_ctx */ + free_percpu(q->queue_ctx); +} + struct request_queue *blk_mq_init_queue(struct blk_mq_tag_set *set) { struct blk_mq_hw_ctx **hctxs; @@ -2005,12 +2024,8 @@ void blk_mq_free_queue(struct request_queue *q) percpu_ref_exit(&q->mq_usage_counter); - free_percpu(q->queue_ctx); - kfree(q->queue_hw_ctx); kfree(q->mq_map); - q->queue_ctx = NULL; - q->queue_hw_ctx = NULL; q->mq_map = NULL; mutex_lock(&all_q_mutex); diff --git a/block/blk-mq.h b/block/blk-mq.h index 4f4f943..6a48c4c 100644 --- a/block/blk-mq.h +++ b/block/blk-mq.h @@ -62,6 +62,8 @@ extern void blk_mq_sysfs_unregister(struct request_queue *q); extern void blk_mq_rq_timed_out(struct request *req, bool reserved); +void blk_mq_release(struct request_queue *q); + /* * Basic implementation of sparser bitmap, allowing the user to spread * the bits over more cachelines. diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 935ea2a..faaf36a 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -517,6 +517,8 @@ static void blk_release_queue(struct kobject *kobj) if (!q->mq_ops) blk_free_flush_queue(q->fq); + else + blk_mq_release(q); blk_trace_shutdown(q); -- 1.7.9.5