Re: [Piglit] [PATCH] glx: make sure contexts are destroyed before we exit the test

2019-04-12 Thread Adam Jackson
On Fri, 2019-04-12 at 16:23 +1000, Timothy Arceri wrote: > Without this the last two contexts that are created will not > have been destroyed when exiting the test. This creates a race > condition in Mesa between any threads that might be using > glsl_types and the atexit() callback that destroys

Re: [Piglit] [PATCH] egl: Add tests for EGL_EXT_device_base v2.

2019-04-12 Thread Mathias Fröhlich
Marek, > Reviewed-by: Marek Olšák Thanks for the review! The patch is pushed now. best Mathias > > Marek > > On Wed, Apr 10, 2019 at 5:10 AM wrote: > > > From: Mathias Fröhlich > > > > Hi all, > > > > Below is a new test that should test the about to be arriving egl device > >

Re: [Piglit] [PATCH] glx: make sure contexts are destroyed before we exit the test

2019-04-12 Thread Eric Anholt
Timothy Arceri writes: > Without this the last two contexts that are created will not > have been destroyed when exiting the test. This creates a race > condition in Mesa between any threads that might be using > glsl_types and the atexit() callback that destroys these types. Reviewed-by: Eric

[Piglit] [PATCH] glx: make sure contexts are destroyed before we exit the test

2019-04-12 Thread Timothy Arceri
Without this the last two contexts that are created will not have been destroyed when exiting the test. This creates a race condition in Mesa between any threads that might be using glsl_types and the atexit() callback that destroys these types. --- tests/glx/glx-multithread-shader-compile.c | 1