netdev - Re: [PATCH 2/2] drivers: net: xgene: Clean up all outstanding tx descriptors

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] [day] [month] [year] [list]

Message-ID: <CAKh23Fn_oqpsKnSH+OUbQMVvN0oyVSsnbpsHb-c_7obGLjXLSw@mail.gmail.com>
Date:   Mon, 28 Aug 2017 10:48:12 -0700
From:   Iyappan Subramanian <isubramanian@....com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     David Miller <davem@...emloft.net>,
        netdev <netdev@...r.kernel.org>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        Dean Nelson <dnelson@...hat.com>,
        Quan Nguyen <qnguyen@....com>, patches <patches@....com>
Subject: Re: [PATCH 2/2] drivers: net: xgene: Clean up all outstanding tx descriptors

Hi Andrew,

On Fri, Aug 25, 2017 at 4:10 PM, Andrew Lunn <andrew@...n.ch> wrote:
> On Fri, Aug 25, 2017 at 03:23:30PM -0700, Iyappan Subramanian wrote:
>> When xgene_enet is rmmod'd and there are still outstanding tx descriptors
>> that have been setup but have not completed, it is possible on the next
>> modprobe of the driver to receive the oldest of such tx descriptors. This
>> results in a kernel NULL pointer dereference.
>>
>> This patch attempts to clean up (by tearing down) all outstanding tx
>> descriptors when the xgene_enet driver is being rmmod'd.
>>
>> Given that, on the next modprobe it should be safe to ignore any such tx
>> descriptors received that map to a NULL skb pointer.
>
> This does not sound correct. Before the module is allowed to be
> removed, everything needs to be finished. You need to wait for all the
> tx descriptors to be returned before unloading. How can you free the
> memory for the descriptor if it is still in use? How can you free the
> skbuf the descriptor points to, if it is still in use...

Thanks for pointing out the issue.  It is an error, we will fix the issue.

Since the two patches are unrelated, I'm going to post them separately.

Thanks,
Iyappan

>
>       Andrew