[Bug nptl/30587] FAIL: nptl/tst-cleanupx4 on hppa

["dave.anglin at bell dot net" <sourceware-bugzilla@sourceware.org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=30587

--- Comment #8 from dave.anglin at bell dot net ---
On 2023-06-27 12:02 p.m., fberat at redhat dot com wrote:
> https://sourceware.org/bugzilla/show_bug.cgi?id=30587
>
> --- Comment #4 from Frédéric Bérat <fberat at redhat dot com> ---
> Since I don't have the access yet, I'm looking at the logs:
>
> test 0
> clh (2)
> global = 2, expected 15
>
> This executes fn2(), which should basically do:
>    pthread_cleanup_push (clh, (void *) 3l);
>    _pthread_cleanup_push (&b, clh, (void *) 2l);
>    pthread_cleanup_push (clh, (void *) 1l);
>
> Whereas, from the output, only `_pthread_cleanup_pop (&b, 1);` seems to be
> executed.
>
> The same goes for the following iterations. It looks like only
> _pthread_cleanup_pop actions get executed on cancellation.
>
> I don't see how this could be induced by the aforementioned patch.
It seems to me that the sequencing of the cleanups must be timing dependent. 
Simply changing
the read in tst-cleanup4.c to xread changes the test result from PASS to FAIL.

The test doesn't fail if xread is moved into tst-cleanup4.c.

-- 
You are receiving this mail because:
You are on the CC list for the bug.