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, 1 Sep 2020 09:39:27 -0700
From:   Keith Busch <kbusch@kernel.org>
To:     Xianting Tian <tian.xianting@h3c.com>
Cc:     axboe@fb.com, hch@lst.de, sagi@grimberg.me,
        linux-nvme@lists.infradead.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH] [v2] nvme-pci: check req to prevent crash in
 nvme_handle_cqe()
Message-ID: <20200901163927.GC2831246@dhcp-10-100-145-180.wdl.wdc.com>
References: <20200831105553.1621-1-tian.xianting@h3c.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20200831105553.1621-1-tian.xianting@h3c.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Aug 31, 2020 at 06:55:53PM +0800, Xianting Tian wrote:
> As blk_mq_tag_to_rq() may return null, so it should be check whether it is
> null before using it to prevent a crash.

It may return NULL if the command id exceeds the number of tags. We
already have a check for a valid command id value, so something is not
adding up here if we're still getting NULL.

>  	req = blk_mq_tag_to_rq(nvme_queue_tagset(nvmeq), cqe->command_id);
> +	if (unlikely(!req)) {
> +		dev_warn(nvmeq->dev->ctrl.device,
> +			"req is null(tag:%d) on queue %d\n",
> +			cqe->command_id, le16_to_cpu(cqe->sq_id));
> +		return;
> +	}