Re: Performance regressions in networking & storage benchmarks in Linux kernel 5.8

[Abdul Anshad Azeez]

Hello Thomas,

Thank you very much for your comments.

Since the performance regressions were fixed when we tested version 5.9-rc4, we 
were not reporting it as an issue and our intention was just to share this as an
 information only.

Thanks,
Abdul Anshad A


From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
Sent: Tuesday, September 22, 2020 04:55 PM
To: Abdul Anshad Azeez <aazees@xxxxxxxxxx>; linux-kernel@xxxxxxxxxxxxxxx <linux-kernel@xxxxxxxxxxxxxxx>; x86@xxxxxxxxxx <x86@xxxxxxxxxx>; netdev@xxxxxxxxxxxxxxx <netdev@xxxxxxxxxxxxxxx>; linux-fsdevel@xxxxxxxxxxxxxxx <linux-fsdevel@xxxxxxxxxxxxxxx>
Cc: rostedt@xxxxxxxxxxx <rostedt@xxxxxxxxxxx>
Subject: Re: Performance regressions in networking & storage benchmarks in Linux kernel 5.8 
 
Abdul,

On Tue, Sep 22 2020 at 08:51, Abdul Anshad Azeez wrote:
> Part of VMware's performance regression testing for Linux Kernel upstream rele
> ases we compared Linux kernel 5.8 against 5.7. Our evaluation revealed perform
> ance regressions mostly in networking latency/response-time benchmarks up to 6
> 0%. Storage throughput & latency benchmarks were also up by 8%.
> In order to find the fix commit, we bisected again between 5.8 and 5.9-rc4 and
>  identified that regressions were fixed from a commit made by the same author 
> Thomas Gleixner, which unbreaks the interrupt affinity settings - "e027fffff79
> 9cdd70400c5485b1a54f482255985(x86/irq: Unbreak interrupt affinity setting)".
>
> We believe these findings would be useful to the Linux community and wanted to
>  document the same.

thanks for letting us know, but the issue is known already and the fix
has been backported to the stable kernel version 5.8.6 as of Sept. 3rd.

Please always check the latest stable version.

Thanks,

        tglx