[Mesa-dev] [Bug 98172] Concurrent call to glClientWaitSync results in segfault in one of the waiters.

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Wed Oct 19 03:53:22 UTC 2016


--- Comment #28 from Suzuki, Shinji <shinji.suzuki at gmail.com> ---
Yes. I agree with you that we can do without per-sync-object if  we
allow all waiters enter fence_finish() freely.
With that said, per-sync-object mutex has another benefit of
potentially reducing lock contention among waiters on differing sync
objects and with other mesa components that deals with shared
resources. To be fair I also have to mention that ctx->Shared.Mutex is
touched everywhere that trying to optimize in this particular context
only may not make much sense and adding mutex certainly has associated
overhead. Overall, I vote +1 on your strategy  if free execution of
fence_finish() is to be allowed.

On Wed, Oct 19, 2016 at 10:21 AM,  <bugzilla-daemon at freedesktop.org> wrote:
> Comment # 27 on bug 98172 from Michel Dänzer
> Note that if we allow concurrent fence_finish calls, I don't think we need a
> per-sync-object mutex.
> ________________________________
> You are receiving this mail because:
> You reported the bug.
> You are on the CC list for the bug.

You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/mesa-dev/attachments/20161019/3d0e80f5/attachment.html>

More information about the mesa-dev mailing list