Re: [musl] Support SIGEV_THREAD_ID

[Florian Weimer <fweimer@redhat.com>]

* Michael Forney:

> On 2019-08-01, Florian Weimer <fweimer@redhat.com> wrote:
>> * James Y. Knight:
>>
>>> There seems to be some debate in glibc over whether this API should be
>>> supported, due to the long-standing debate about "pthread_t" vs
>>> "kernel tid" APIs. (And this API uses kernel tids, of course.)
>>
>> The debate is over and has been decided in favor of supporting TIDs.  We
>> just have a backlog of interfaces for which we need to add support.
>
> Hi Florian,
>
> Am I interpreting you correctly that glibc intends to add a
> sigev_notify_thread_id define to access this field in struct sigevent?
> I plan to send a patch to qemu to use sigev_notify_thread_id over
> _sigev_un.tid if it is available, and I just want to confirm that
> glibc intends to follow suit before I mention that in the commit
> message.

I do not expect that we are going to add a sigev_notify_thread_id member
once glibc implements the struct member.  On common configurations, it's
likely that we are going to use a different approach, using anonymous
structs and unions.

I think you will have to use a different mechanism for detecting the
presence of the struct member.

Thanks,
Florian
-- 
Red Hat GmbH, https://de.redhat.com/ , Registered seat: Grasbrunn,
Commercial register: Amtsgericht Muenchen, HRB 153243,
Managing Directors: Charles Cachera, Brian Klemm, Laurie Krebs, Michael O'Neill