linux-kernel - Re: [PATCH] block: Fix potential deadlock warning in blk_mq_mark_tag

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <6e729890-7374-4335-ab7d-ead00775057e@kernel.dk>
Date: Wed, 14 Aug 2024 14:22:20 -0600
From: Jens Axboe <axboe@...nel.dk>
To: Bart Van Assche <bvanassche@....org>,
 Li Lingfeng <lilingfeng@...weicloud.com>, hch@....de, jack@...e.cz,
 linux-block@...r.kernel.org, linux-kernel@...r.kernel.org
Cc: yukuai1@...weicloud.com, yukuai3@...wei.com, houtao1@...wei.com,
 yi.zhang@...wei.com, yangerkun@...wei.com, lilingfeng3@...wei.com
Subject: Re: [PATCH] block: Fix potential deadlock warning in
 blk_mq_mark_tag_wait

On 8/14/24 1:52 PM, Bart Van Assche wrote:
> On 8/14/24 4:35 AM, Li Lingfeng wrote:
>> When interrupt is turned on while a lock holding by spin_lock_irq it
>> throws a warning because of potential deadlock.
> 
> Which tool reported the warning? Please mention this in the patch
> description.
>>
>> blk_mq_mark_tag_wait
>>   spin_lock_irq(&wq->lock)
>>        --> turn off interrupt and get lockA
>>   blk_mq_get_driver_tag
>>    __blk_mq_tag_busy
>>     spin_lock_irq(&tags->lock)
>>     spin_unlock_irq(&tags->lock)
>>        --> release lockB and turn on interrupt accidentally
> 
> The above call chain does not match the code in Linus' master tree.
> Please fix this.
> 
>> Fix it by using spin_lock_irqsave to get lockB instead of spin_lock_irq.
>> Fixes: 4f1731df60f9 ("blk-mq: fix potential io hang by wrong 'wake_batch'")
>> Signed-off-by: Li Lingfeng <lilingfeng3@...wei.com>
> 
> Please leave a blank line between the patch description and the section
> with tags.

Please just include the actual lockdep trace rather than a doctored up
one, it's a lot more descriptive. And use the real lock names rather
than turn it into hypotheticals.

-- 
Jens Axboe