From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: From: Jianchao Wang To: axboe@kernel.dk Cc: keith.busch@linux.intel.com, ming.lei@redhat.com, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH V2 4/4] blk-mq: fallback to previous nr_hw_queues when updating fails Date: Fri, 12 Oct 2018 18:07:28 +0800 Message-Id: <1539338848-1789-5-git-send-email-jianchao.w.wang@oracle.com> In-Reply-To: <1539338848-1789-1-git-send-email-jianchao.w.wang@oracle.com> References: <1539338848-1789-1-git-send-email-jianchao.w.wang@oracle.com> List-ID: When we try to increate the nr_hw_queues, we may fail due to shortage of memory or other reason, then blk_mq_realloc_hw_ctxs stops and some entries in q->queue_hw_ctx are left with NULL. However, because queue map has been updated with new nr_hw_queues, some cpus have been mapped to hw queue which just encounters allocation failure, thus blk_mq_map_queue could return NULL. This will cause panic in following blk_mq_map_swqueue. To fix it, when increase nr_hw_queues fails, fallback to previous nr_hw_queues and post warning. At the same time, driver's .map_queues usually use completion irq affinity to map hw and cpu, fallback nr_hw_queues will cause lack of some cpu's map to hw, so use default blk_mq_map_queues to do that. Reported-by: syzbot+83e8cbe702263932d9d4@syzkaller.appspotmail.com Signed-off-by: Jianchao Wang --- block/blk-mq.c | 27 ++++++++++++++++++++++++--- 1 file changed, 24 insertions(+), 3 deletions(-) diff --git a/block/blk-mq.c b/block/blk-mq.c index 34f3973..d2ce67a 100644 --- a/block/blk-mq.c +++ b/block/blk-mq.c @@ -2540,7 +2540,7 @@ static struct blk_mq_hw_ctx *blk_mq_alloc_and_init_hctx( static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set, struct request_queue *q) { - int i, j; + int i, j, end; struct blk_mq_hw_ctx **hctxs = q->queue_hw_ctx; /* protect against switching io scheduler */ @@ -2574,8 +2574,20 @@ static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set, break; } } + /* + * Increasing nr_hw_queues fails. Free the newly allocated + * hctxs and keep the previous q->nr_hw_queues. + */ + if (i != set->nr_hw_queues) { + j = q->nr_hw_queues; + end = i; + } else { + j = i; + end = q->nr_hw_queues; + q->nr_hw_queues = set->nr_hw_queues; + } - for (j = i; j < q->nr_hw_queues; j++) { + for (; j < end; j++) { struct blk_mq_hw_ctx *hctx = hctxs[j]; if (hctx) { @@ -2587,7 +2599,6 @@ static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set, } } - q->nr_hw_queues = i; mutex_unlock(&q->sysfs_lock); } @@ -2972,6 +2983,7 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, { struct request_queue *q; LIST_HEAD(head); + int prev_nr_hw_queues; lockdep_assert_held(&set->tag_list_lock); @@ -3000,10 +3012,19 @@ static void __blk_mq_update_nr_hw_queues(struct blk_mq_tag_set *set, blk_mq_sysfs_unregister(q); } + prev_nr_hw_queues = set->nr_hw_queues; set->nr_hw_queues = nr_hw_queues; blk_mq_update_queue_map(set); +fallback: list_for_each_entry(q, &set->tag_list, tag_set_list) { blk_mq_realloc_hw_ctxs(set, q); + if (q->nr_hw_queues != set->nr_hw_queues) { + pr_warn("Increasing nr_hw_queues to %d fails, fallback to %d\n", + nr_hw_queues, prev_nr_hw_queues); + set->nr_hw_queues = prev_nr_hw_queues; + blk_mq_map_queues(set); + goto fallback; + } blk_mq_map_swqueue(q); } -- 2.7.4