Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 27 Jul 2021 18:06:46 +0800
From:   Ming Lei <ming.lei@redhat.com>
To:     John Garry <john.garry@huawei.com>
Cc:     axboe@kernel.dk, osandov@fb.com, linux-block@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] blk-mq-sched: Fix blk_mq_sched_alloc_tags() error
 handling
Message-ID: <YP/atlyuacbHF/sp@T590>
References: <1627378373-148090-1-git-send-email-john.garry@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <1627378373-148090-1-git-send-email-john.garry@huawei.com>
Precedence: bulk

On Tue, Jul 27, 2021 at 05:32:53PM +0800, John Garry wrote:
> If the blk_mq_sched_alloc_tags() -> blk_mq_alloc_rqs() call fails, then we
> call blk_mq_sched_free_tags() -> blk_mq_free_rqs().
> 
> It is incorrect to do so, as any rqs would have already been freed in the
> blk_mq_alloc_rqs() call.
> 
> Fix by calling blk_mq_free_rq_map() only directly.
> 
> Fixes: 6917ff0b5bd41 ("blk-mq-sched: refactor scheduler initialization")
> Signed-off-by: John Garry <john.garry@huawei.com>

Not sure it is one fix, because blk_mq_free_rqs() does nothing when
->static_rqs[] isn't filled, so 'Fixes' tag isn't needed, IMO.

> 
> diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c
> index c838d81ac058..0f006cabfd91 100644
> --- a/block/blk-mq-sched.c
> +++ b/block/blk-mq-sched.c
> @@ -515,17 +515,6 @@ void blk_mq_sched_insert_requests(struct blk_mq_hw_ctx *hctx,
>  	percpu_ref_put(&q->q_usage_counter);
>  }
>  
> -static void blk_mq_sched_free_tags(struct blk_mq_tag_set *set,
> -				   struct blk_mq_hw_ctx *hctx,
> -				   unsigned int hctx_idx)
> -{
> -	if (hctx->sched_tags) {
> -		blk_mq_free_rqs(set, hctx->sched_tags, hctx_idx);
> -		blk_mq_free_rq_map(hctx->sched_tags, set->flags);
> -		hctx->sched_tags = NULL;
> -	}
> -}
> -
>  static int blk_mq_sched_alloc_tags(struct request_queue *q,
>  				   struct blk_mq_hw_ctx *hctx,
>  				   unsigned int hctx_idx)
> @@ -539,8 +528,10 @@ static int blk_mq_sched_alloc_tags(struct request_queue *q,
>  		return -ENOMEM;
>  
>  	ret = blk_mq_alloc_rqs(set, hctx->sched_tags, hctx_idx, q->nr_requests);
> -	if (ret)
> -		blk_mq_sched_free_tags(set, hctx, hctx_idx);
> +	if (ret) {
> +		blk_mq_free_rq_map(hctx->sched_tags, set->flags);
> +		hctx->sched_tags = NULL;
> +	}

The patch itself looks fine:

Reviewed-by: Ming Lei <ming.lei@redhat.com>


Thanks,
Ming