Subject: Re: [PATCH v2 5/5] md: protect md_thread with a new disk level spin
 lock
To:     Guoqing Jiang <guoqing.jiang@linux.dev>,
        Yu Kuai <yukuai1@huaweicloud.com>, agk@redhat.com,
        snitzer@kernel.org, song@kernel.org
Cc:     linux-kernel@vger.kernel.org, linux-raid@vger.kernel.org,
        yi.zhang@huawei.com, yangerkun@huawei.com,
        "yukuai (C)" <yukuai3@huawei.com>
References: <20230315061810.653263-1-yukuai1@huaweicloud.com>
 <20230315061810.653263-6-yukuai1@huaweicloud.com>
 <16613534-5482-23c5-fa97-cfaedecab3d7@linux.dev>
From:   Yu Kuai <yukuai1@huaweicloud.com>
Message-ID: <7cc22b63-a2f7-67ed-1a50-9fb415fadb81@huaweicloud.com>
Date:   Wed, 15 Mar 2023 18:02:04 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <16613534-5482-23c5-fa97-cfaedecab3d7@linux.dev>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi,

在 2023/03/15 17:39, Guoqing Jiang 写道:
> 
> 
> On 3/15/23 14:18, Yu Kuai wrote:
>> From: Yu Kuai <yukuai3@huawei.com>
>>
>> Our test reports a uaf for 'mddev->sync_thread':
>>
>> T1                      T2
>> md_start_sync
>>   md_register_thread
>>             raid1d
>>              md_check_recovery
>>               md_reap_sync_thread
>>                md_unregister_thread
>>                 kfree
>>
>>   md_wakeup_thread
>>    wake_up
>>    ->sync_thread was freed
> 
> Better to provide the relevant uaf (user after free perhaps you mean)
> log from the test.
Ok, I'll add uaf report(the report is from v5.10) in the next version.
> 
>> Currently, a global spinlock 'pers_lock' is borrowed to protect
>> 'mddev->thread', this problem can be fixed likewise, however, there might
>> be similar problem for other md_thread, and I really don't like the 
>> idea to
>> borrow a global lock.
>>
>> This patch use a disk level spinlock to protect md_thread in relevant 
>> apis.
> 
> It is array level I think, and you probably want to remove the comment.
> 
> * pers_lockdoes extra service to protect accesses to
> * mddev->thread when the mutex cannot be held.

Yes, I missed this.

Thanks,
Kuai
> 
> Thanks,
> Guoqing
> .
>