Re: latest pahole breaks libbpf CI and let's talk about staging

From: Andrii Nakryiko <andrii.nakryiko@gmail.com>
To: Arnaldo Carvalho de Melo <arnaldo.melo@gmail.com>
Cc: bpf <bpf@vger.kernel.org>, dwarves@vger.kernel.org, siudin@fb.com
Subject: Re: latest pahole breaks libbpf CI and let's talk about staging
Date: Wed, 16 Jun 2021 15:38:38 -0700	[thread overview]
Message-ID: <CAEf4BzYn31_93G_f924HR8dSW=oGqyFaneRa0fo5Btcg-Y2xJg@mail.gmail.com> (raw)
In-Reply-To: <YMpCDuEO/mItxdR7@kernel.org>

On Wed, Jun 16, 2021 at 11:25 AM Arnaldo Carvalho de Melo
<arnaldo.melo@gmail.com> wrote:
>
> Em Wed, Jun 16, 2021 at 10:40:45AM -0700, Andrii Nakryiko escreveu:
> > On Wed, Jun 16, 2021 at 7:56 AM Arnaldo Carvalho de Melo
> > <arnaldo.melo@gmail.com> wrote:
> > >
> > > Em Tue, Jun 15, 2021 at 04:30:03PM -0700, Andrii Nakryiko escreveu:
> > > > Hey Arnaldo,
> > > >
> > > > Seems like de3a7f912559 ("btf_encoder: Reduce the size of encode_cu()
> > > > by moving function encoding to separate method") break two selftests
> > > > in libbpf CI (see [0]). Please take a look. I suspect some bad BTF,
> > > > because both tests rely on kernel BTF info.
> > > >
> > > > You've previously asked about staging pahole changes. Did you make up
> > > > your mind about branch names and the process overall? Seems like a
> > > > good chance to bring this up ;-P
> > > >
> > > >   [0] https://travis-ci.com/github/libbpf/libbpf/jobs/514329152
> > >
> > > Ok, please add tmp.master as the staging branch, I'll move things to
> > > master only after it passing thru CI.
> > >
> >
> > So I'm thinking about what's the best setup to catch pahole staging
> > problems, but not break main libbpf CI and kernel-patches CI flows.
>
> > How about we keep all the existing CI jobs to use pahole's master.
>
> Agreed.
>
> > Then add a separate job to do full kernel build with pahole built from
> > staging branch. And mark it as non-critical (or whatever the
> > terminology), so it doesn't mark the build red. I'd do that as a cron
> > job that runs every day. That way if you don't have anything urgent,
> > next day you'll get staging tested automatically. If you need to test
> > right now, there is a way to re-trigger previous build and it will
> > re-fetch latest staging (so there is a way for you to proactively
> > test).
> >
> > Basically, I want broken staging pahole to not interrupt anything we
> > are doing. WDYT?
>
> Sounds like a plan, please hand hold me on this, I'm not versed on
> github.

I'll set up everything from my side, and then we'll just setup proper
access rights for you to be able to trigger builds. We are migrating
everything from Travis CI to Github Actions, and I'm not yet too
familiar with Github Actions, so I might need a few iterations.

BTW, while you are investigating pahole regression, can you please
revert the offending commit and push it to master to make out CIs
green again?

>
> - Arnaldo
>
> > > Now looking at that code, must be something subtle...
> > >
> > > - Arnaldo
>
> --
>
> - Arnaldo