netdev - Re: [PATCH linux-next 1/3] selftests: net: udpgso_bench

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: <ef0371ebc094e2c0778badca69ea0043b98589aa.camel@redhat.com>
Date:   Tue, 18 Apr 2023 16:20:12 +0200
From:   Paolo Abeni <pabeni@...hat.com>
To:     Yang Yang <yang.yang29@....com.cn>, davem@...emloft.net,
        edumazet@...gle.com, willemdebruijn.kernel@...il.com
Cc:     kuba@...nel.org, linux-kernel@...r.kernel.org,
        linux-kselftest@...r.kernel.org, netdev@...r.kernel.org,
        shuah@...nel.org, zhang.yunkai@....com.cn, xu.xin16@....com.cn,
        Xuexin Jiang <jiang.xuexin@....com.cn>
Subject: Re: [PATCH linux-next 1/3] selftests: net: udpgso_bench_rx: Fix
 verifty exceptions

On Mon, 2023-04-17 at 20:24 +0800, Yang Yang wrote:
> From: Zhang Yunkai (CGEL ZTE) <zhang.yunkai@....com.cn>
> 
> The verification function of this test case is likely to encounter the
> following error, which may confuse users.
> 
> Executing the following command fails:
> bash# udpgso_bench_tx -l 4 -4 -D "$DST"
> bash# udpgso_bench_tx -l 4 -4 -D "$DST" -S 0
> bash# udpgso_bench_rx -4 -G -S 1472 -v
> udpgso_bench_rx: data[1472]: len 2944, a(97) != q(113)


As noted by Willem, both the commit message and the above command
sequence is quite confusing. Please reorder the commands in the exact
sequence you run them, presumably:

udpgso_bench_rx -4 -G -S 1472 -v &
udpgso_bench_tx -l 4 -4 -D "$DST" -S 0


> This is because the sending buffers are not aligned by 26 bytes, and the
> GRO is not merged sequentially, and the receiver does not judge this
> situation. We do the validation after the data is split at the receiving
> end, just as the application actually uses this feature.

The wording from Willem response is much more clear. If applicable,
please use such text.

BTW I could not reproduce the issue with any permutation of the
suggested commands I could think of, so possibly that section need some
extra clarification.

Thanks,

Paolo