Re: [PATCH v2 7/7] tools/testing/selftests/bpf: replace open-coded 16 with TASK_COMM_LEN

From: Yafang Shao
Date: Sat Feb 11 2023 - 22:20:33 EST


On Sun, Feb 12, 2023 at 12:51 AM Qais Yousef <qyousef@xxxxxxxxxxx> wrote:
>
> On 02/09/23 23:37, Yafang Shao wrote:
> > On Thu, Feb 9, 2023 at 10:28 PM Kajetan Puchalski
> > <kajetan.puchalski@xxxxxxx> wrote:
> > >
> > > On Thu, Feb 09, 2023 at 02:20:36PM +0800, Yafang Shao wrote:
> > >
> > > [...]
> > >
> > > Hi Yafang,
> > >
> > > > Many thanks for the detailed analysis. Seems it can work.
> > > >
> > > > Hi John,
> > > >
> > > > Could you pls. try the attached fix ? I have verified it in my test env.
> > >
> > > I tested the patch on my environment where I found the issue with newer
> > > kernels + older Perfetto. The patch does improve things so that's nice.
> >
> > Thanks for the test. I don't have Perfetto in hand, so I haven't
> > verify Perfetto.
>
> FWIW, perfetto is not android specific and can run on normal linux distro setup
> (which I do but haven't noticed this breakage).
>
> It's easy to download the latest release (including for android though I never
> tried that) from github
>
> https://github.com/google/perfetto/releases
>

Thanks for the information. I will try to run it on my test env.
I suspect the "systrace_parse_failure" error is caused by the field we
introduced into struct ftrace_event_field in the proposed patch, but I
haven't taken a deep look at the perfetto src code yet.

> Kajetan might try to see if he can pick the latest version which IIUC contains
> a workaround.
>
> If this simple patch can be tweaked to make it work again against older
> versions that'd be nice though.
>
> HTH.
>
>
> Cheers
>
> --
> Qais Yousef



--
Regards
Yafang