From 9f16a66733c90b5f33f624b0b0e36a345b0aaf93 Mon Sep 17 00:00:00 2001 From: Ming Lei Date: Fri, 13 Nov 2020 21:44:48 +0800 Subject: [PATCH] block: mark flush request as IDLE when it is really finished For avoiding use-after-free on flush request, we call its .end_io() from both timeout code path and __blk_mq_end_request(). When flush request's ref doesn't drop to zero, it is still used, we can't mark it as IDLE, so fix it by marking IDLE when its refcount drops to zero really. Fixes: 65ff5cd04551 ("blk-mq: mark flush request as IDLE in flush_end_io()") Signed-off-by: Ming Lei Cc: Yi Zhang Signed-off-by: Jens Axboe --- block/blk-flush.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-) diff --git a/block/blk-flush.c b/block/blk-flush.c index e32958f0b68750..fd5cee9f1a3bec 100644 --- a/block/blk-flush.c +++ b/block/blk-flush.c @@ -225,13 +225,18 @@ static void flush_end_io(struct request *flush_rq, blk_status_t error) /* release the tag's ownership to the req cloned from */ spin_lock_irqsave(&fq->mq_flush_lock, flags); - WRITE_ONCE(flush_rq->state, MQ_RQ_IDLE); if (!refcount_dec_and_test(&flush_rq->ref)) { fq->rq_status = error; spin_unlock_irqrestore(&fq->mq_flush_lock, flags); return; } + /* + * Flush request has to be marked as IDLE when it is really ended + * because its .end_io() is called from timeout code path too for + * avoiding use-after-free. + */ + WRITE_ONCE(flush_rq->state, MQ_RQ_IDLE); if (fq->rq_status != BLK_STS_OK) error = fq->rq_status;