Subject: Re: [PATCH] xen: avoid deadlock in xenbus
To: Boris Ostrovsky <boris.ostrovsky@oracle.com>,
        linux-kernel@vger.kernel.org, xen-devel@lists.xenproject.org
References: <20170728145355.17258-1-jgross@suse.com>
 <71819268-dd53-a60f-5759-b3ae1ce486d4@oracle.com>
From: Juergen Gross <jgross@suse.com>
Message-ID: <09616301-1587-d5d5-a1dd-10ea74992339@suse.com>
Date: Fri, 28 Jul 2017 17:16:34 +0200
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101
 Thunderbird/52.2.1
MIME-Version: 1.0
In-Reply-To: <71819268-dd53-a60f-5759-b3ae1ce486d4@oracle.com>
Content-Type: text/plain; charset=windows-1252
Content-Language: de-DE
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1188
Lines: 44

On 28/07/17 17:14, Boris Ostrovsky wrote:
> On 07/28/2017 10:53 AM, Juergen Gross wrote:
>> When starting the xenwatch thread a theoretical deadlock situation is
>> possible:
>>
>> xs_init() contains:
>>
>>     task = kthread_run(xenwatch_thread, NULL, "xenwatch");
>>     if (IS_ERR(task))
>>         return PTR_ERR(task);
>>     xenwatch_pid = task->pid;
>>
>> And xenwatch_thread() does:
>>
>>     mutex_lock(&xenwatch_mutex);
>>     ...
>>     event->handle->callback();
>>     ...
>>     mutex_unlock(&xenwatch_mutex);
>>
>> The callback could call unregister_xenbus_watch() which does:
>>
>>     ...
>>     if (current->pid != xenwatch_pid)
>>         mutex_lock(&xenwatch_mutex);
>>     ...
>>
>> In case a watch is firing before xenwatch_pid could be set and the
>> callback of that watch unregisters a watch, then a self-deadlock would
>> occur.
>>
>> Avoid this by setting xenwatch_pid in xenwatch_thread().
>>
>> Signed-off-by: Juergen Gross <jgross@suse.com>
> 
> Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
> 
> +stable?

As this problem is purely theoretical, I don't think the patch is
appropriate for stable (at least the stable rules tell me so).


Juergen