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;
Subject: Re: [PATCH v2] nvme-multipath: Early exit if no path is available
To:     Hannes Reinecke <hare@suse.de>, Daniel Wagner <dwagner@suse.de>
CC:     <linux-nvme@lists.infradead.org>, Sagi Grimberg <sagi@grimberg.me>,
        <linux-kernel@vger.kernel.org>, Jens Axboe <axboe@fb.com>,
        Keith Busch <kbusch@kernel.org>, Christoph Hellwig <hch@lst.de>
References: <20210127103033.15318-1-dwagner@suse.de>
 <db9baae0-547c-7ff4-8b2c-0b95f14be67c@huawei.com>
 <20210128075837.u5u56t23fq5gu6ou@beryllium.lan>
 <69575290-200e-b4a1-4269-c71e4c2cc37b@huawei.com>
 <8704038d-0b71-5c9c-2d2b-a9ff2166dbe4@suse.de>
From:   Chao Leng <lengchao@huawei.com>
Message-ID: <c91cb4df-6736-bc0f-9d33-ecc3e2f4f510@huawei.com>
Date:   Fri, 29 Jan 2021 09:18:10 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:68.0) Gecko/20100101
 Thunderbird/68.9.0
MIME-Version: 1.0
In-Reply-To: <8704038d-0b71-5c9c-2d2b-a9ff2166dbe4@suse.de>
Content-Type: text/plain; charset="utf-8"; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 8bit
Precedence: bulk


On 2021/1/28 17:23, Hannes Reinecke wrote:
> On 1/28/21 10:18 AM, Chao Leng wrote:
>>
>>
>> On 2021/1/28 15:58, Daniel Wagner wrote:
>>> On Thu, Jan 28, 2021 at 09:31:30AM +0800, Chao Leng wrote:
>>>>> --- a/drivers/nvme/host/multipath.c
>>>>> +++ b/drivers/nvme/host/multipath.c
>>>>> @@ -221,7 +221,7 @@ static struct nvme_ns *nvme_round_robin_path(struct nvme_ns_head *head,
>>>>>        }
>>>>>        for (ns = nvme_next_ns(head, old);
>>>>> -         ns != old;
>>>>> +         ns && ns != old;
>>>> nvme_round_robin_path just be called when !"old".
>>>> nvme_next_ns should not return NULL when !"old".
>>>> It seems unnecessary to add checking "ns".
>>>
>>> The problem is when we enter nvme_round_robin_path() and there is no
>>> path available. In this case the initialization ns = nvme_next_ns(head,
>>> old) could return a NULL pointer."old" should not be NULL, so there is at least one path that is "old".
>> It is impossible to return NULL for nvme_next_ns(head, old).
> 
> No. list_next_or_null_rcu()/list_first_or_null_rcu() will return NULL when then end of the list is reached.
Although list_next_or_null_rcu()/list_first_or_null_rcu() may return
NULL, but nvme_next_ns(head, old) assume that the "old" is in the "head",
so nvme_next_ns(head, old) should not return NULL. If the "old" is not
in the "head", nvme_next_ns(head, old) will run abnormal.
So there is other bug which cause nvme_next_ns(head, old).

I review the code about head->list and head->current_path, I find 2 bugs
may cause nvme_next_ns(head, old) abnormal:
First, I already send the patch. see:
https://lore.kernel.org/linux-nvme/20210128033351.22116-1-lengchao@huawei.com/
Second, in nvme_ns_remove, list_del_rcu is before
nvme_mpath_clear_current_path. This may cause "old" is deleted from the
"head", but still use "old". I'm not sure there's any other
consideration here, I will check it and try to fix it.
> 
> Cheers,
> 
> Hannes