Re: [patch] Fix for bz14333 -- race between atexit() and exit()

[Paul Pluzhnikov <ppluzhnikov at google dot com>]

Anoop removed from CC list since his e-mail no longer works.

On Tue, Jul 11, 2017 at 5:22 AM, Joseph Myers <joseph@codesourcery.com> wrote:
>
> Is there such a race for quick_exit / at_quick_exit (which postdate the
> previous patch), or not?

Yes, the test from bz14333#c5 with s/atexit/at_quick_exit/ and
s/exit/quick_exit/ also crashes the same way, and the patch here fixes
that crash as well.

>  If there is, it should also be fixed (all
> atexit-like functions should be fixed).

They all share the implementation, so this is already done here :-)

> Is it possible to add a test / tests for this issue to the glibc
> testsuite?

Sure. I was going to add it after the "basic" atexit ordering test,
but since that patch requires revisions, I will send a patch to add
the multithreaded test for atexit/at_quick_exit/__cxa_at_quick_exit
separately (or should it be part of this patch?).

Thanks,
-- 
Paul Pluzhnikov