[vlc-devel] Defect #22230

[Ori Pessach]

I've seen this happen with five fewer than 18 instances, actually - as few
as 2 or 3. I've also reproduced this on different graphics card (and hence,
different drivers.)

The code was written to use a large number of instances because it crashes
faster.

And yes, without hardware acceleration the code is stable. Or should I say
- stabler. I'm chasing down other apparent race conditions in other areas
of the code, but those are much harder to reproduce. I thought I'd start
with an easy one. :)

Ori

On Fri, May 3, 2019 at 3:10 PM Sean McGovern <gseanmcg at gmail.com> wrote:

> Hi Ori,
>
> I don't claim to be an expert on any of this but from just looking at your
> sample code, are there any practical limitations of D3D11 that creating 18
> player instances would cause it to occasionally choke?
>
> Is it crash free without the hardware acceleration? Maybe you are hitting
> a driver bug.
>
> Could it be a weird TOCTOU thing?
>
> -- Sean McGovern
> *From:* mail at oripessach.com
> *Sent:* May 3, 2019 11:10 AM
> *To:* vlc-devel at videolan.org
> *Reply-to:* vlc-devel at videolan.org
> *Subject:* [vlc-devel] Defect #22230
>
> Hi,
>
> I was wondering if there's anything I can do to help triage and resolve
> defect #22230 - the problem seems to be very repeatable in my tests, and my
> test code is attached to the defect.
>
> Would it be helpful at all if I included the entire Visual Studio solution
> with the test code?
>
> Thanks,
>
> Ori Pessach
> _______________________________________________
> vlc-devel mailing list
> To unsubscribe or modify your subscription options:
> https://mailman.videolan.org/listinfo/vlc-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.videolan.org/pipermail/vlc-devel/attachments/20190504/38cb2f8d/attachment.html>