Re: [PATCH dwarves v4 1/4] dwarf_loader: Receive per-thread data on worker threads.

[Kui-Feng Lee <kuifeng@fb.com>]

On Wed, 2022-03-09 at 16:18 -0800, Andrii Nakryiko wrote:
> On Wed, Mar 9, 2022 at 4:14 PM Andrii Nakryiko
> <andrii.nakryiko@gmail.com> wrote:
> > 
> > On Wed, Mar 9, 2022 at 11:24 AM Arnaldo Carvalho de Melo
> > <acme@kernel.org> wrote:
> > > 
> > > Em Tue, Mar 08, 2022 at 03:45:03PM -0800, Andrii Nakryiko
> > > escreveu:
> > > > On Thu, Jan 27, 2022 at 11:22 AM Arnaldo Carvalho de Melo
> > > > <arnaldo.melo@gmail.com> wrote:
> > > > > 
> > > > > Em Wed, Jan 26, 2022 at 11:55:25AM -0800, Andrii Nakryiko
> > > > > escreveu:
> > > > > > On Wed, Jan 26, 2022 at 11:20 AM Kui-Feng Lee
> > > > > > <kuifeng@fb.com> wrote:
> > > > > > > 
> > > > > > > Add arguments to steal and thread_exit callbacks of
> > > > > > > conf_load to
> > > > > > > receive per-thread data.
> > > > > > > 
> > > > > > > Signed-off-by: Kui-Feng Lee <kuifeng@fb.com>
> > > > > > > ---
> > > > > > 
> > > > > > Please carry over acks you got on previous revisions,
> > > > > > unless you did
> > > > > > some drastic changes to already acked patches.
> > > > > 
> > > > > Yes, please do so.
> > > > > 
> > > > > I'll collect them this time, no need to resend.
> > > > > 
> > > > 
> > > > Hey, Arnaldo!
> > > > 
> > > > Any idea when these patches will make it into master branch? I
> > > > see
> > > > they are in tmp.master right now.
> > > 
> > > I did some minor fixups to the cset comment and to the code in
> > > the
> > > 'pahole --compile' new feature at the head of it and pushed all
> > > up,
> > > please check.
> > > 
> > 
> > I did check locally with latest pahole master, and it seems like
> > something is wrong with generated BTF. I get three selftests
> > failure
> > if I use latest pahole compiled from master.
> > 
> > Kui-Feng, please take a look when you get a chance. Arnaldo, please
> > hold off from releasing a new version for now.

I just figure out the root cuase.
It caused by missing info from percpu_secinfo when collecting data from
threads.  The encoder stores it separatedly from struct btf, and we
have separated different encoders for threads.  They are not merged
together.  I will fixed it ASAP.