netdev - Re: [PATCH] net: tcp: ratelimit warnings in tcp

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]

Date:   Mon, 9 Nov 2020 16:39:24 +0100
From:   Eric Dumazet <eric.dumazet@...il.com>
To:     Menglong Dong <menglong8.dong@...il.com>,
        Eric Dumazet <edumazet@...gle.com>
Cc:     David Miller <davem@...emloft.net>,
        Alexey Kuznetsov <kuznet@....inr.ac.ru>,
        Hideaki YOSHIFUJI <yoshfuji@...ux-ipv6.org>,
        Jakub Kicinski <kuba@...nel.org>,
        netdev <netdev@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Menglong Dong <dong.menglong@....com.cn>
Subject: Re: [PATCH] net: tcp: ratelimit warnings in tcp_recvmsg



On 11/9/20 3:48 PM, Menglong Dong wrote:
> On Mon, Nov 9, 2020 at 9:36 PM Eric Dumazet <edumazet@...gle.com> wrote:
>>
>> I do not think this patch is useful. That is simply code churn.
>>
>> Can you trigger the WARN() in the latest upstream version ?
>> If yes this is a serious bug that needs urgent attention.
>>
>> Make sure you have backported all needed fixes into your kernel, if
>> you get this warning on a non pristine kernel.
> 
> Theoretically, this WARN() shouldn't be triggered in any branches.
> Somehow, it just happened in kernel v3.10. This really confused me. I
> wasn't able to keep tracing it, as it is a product environment.
> 
> I notice that the codes for tcp skb receiving didn't change much
> between v3.10 and the latest upstream version, and guess the latest
> version can be triggered too.
> 
> If something is fixed and this WARN() won't be triggered, just ignore me.
> 

Yes, I confirm this WARN() should not trigger.

The bug is not in tcp recvmsg(), that is why you do not see obvious
fix for this issue in 3.10