Message-ID: <1513156889.26281.4.camel@suse.com>
Subject: Re: [BUG] kaweth: a possible sleep-in-atomic bug in
 kaweth_start_xmit
From: Oliver Neukum <oneukum@suse.com>
To: Jia-Ju Bai <baijiaju1990@gmail.com>,
        David Miller <davem@davemloft.net>, tklauser@distanz.ch,
        arvind.yadav.cs@gmail.com, edumazet@google.com,
        johannes.berg@intel.com
Cc: Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        linux-usb@vger.kernel.org, netdev@vger.kernel.org
Date: Wed, 13 Dec 2017 10:21:29 +0100
In-Reply-To: <384f1b94-1fd1-cd38-f8ab-fcba872c77e9@gmail.com>
References: <384f1b94-1fd1-cd38-f8ab-fcba872c77e9@gmail.com>
Content-Type: text/plain; charset="UTF-8"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 797
Lines: 26

Am Mittwoch, den 13.12.2017, 16:57 +0800 schrieb Jia-Ju Bai:
> According to drivers/net/usb/kaweth.c, the driver may sleep under a 
> spinlock.
> The function call path is:
> kaweth_start_xmit (acquire the spinlock)
>    kaweth_async_set_rx_mode
>      kaweth_control
>        kaweth_internal_control_msg
>          usb_start_wait_urb
>            wait_event_timeout --> may sleep
>            usb_kill_urb --> may sleep
> 
> I do not find a good way to fix it, so I only report.
> This possible bug is found by my static analysis tool (DSAC) and checked 
> by my code review.
> 

Hi,

thanks for reporting. I need to get out my old test device.
It will take a few days. The obvious fix would be to set this
filter only on initialization. Unfortunately this needs to
be tested.

	Regards
		Oliver