[Piglit] [PATCH] glsl-1.50: add test that emits max number of output components

2016-11-02 Thread Ilia Mirkin
This triggers an error presently on at least nouveau. It's an adaptation of the existing end-primitive test, but passes the end_prim_offset via a separate varying (and always picks the maximum number of vertices possible for that number of output components). Signed-off-by: Ilia Mirkin --- The i

Re: [Piglit] [PATCH 2/2] ext_framebuffer_multisample: Add a test for alpha to coverage with no draw buffer zero write

2016-11-02 Thread Anuj Phogat
On Wed, Nov 2, 2016 at 3:05 PM, Brian Paul wrote: > On 11/02/2016 03:36 PM, Anuj Phogat wrote: >> >> In a situation when there are multiple render targets with alpha to >> coverage enabled, if fragment shader doesn't write to draw buffer >> zero, alpha value used for coverage will be undefined. Su

Re: [Piglit] [PATCH 2/2] ext_framebuffer_multisample: Add a test for alpha to coverage with no draw buffer zero write

2016-11-02 Thread Brian Paul
On 11/02/2016 03:36 PM, Anuj Phogat wrote: In a situation when there are multiple render targets with alpha to coverage enabled, if fragment shader doesn't write to draw buffer zero, alpha value used for coverage will be undefined. Such case should not cause a GPU hang. Signed-off-by: Anuj Phoga

[Piglit] [PATCH 2/2] ext_framebuffer_multisample: Add a test for alpha to coverage with no draw buffer zero write

2016-11-02 Thread Anuj Phogat
In a situation when there are multiple render targets with alpha to coverage enabled, if fragment shader doesn't write to draw buffer zero, alpha value used for coverage will be undefined. Such case should not cause a GPU hang. Signed-off-by: Anuj Phogat --- Not everything in allocate_data_arrays

[Piglit] [PATCH 1/2] ext_framebuffer_multisample: Add an option to disable draw buffer zero write

2016-11-02 Thread Anuj Phogat
Signed-off-by: Anuj Phogat --- .../alpha-to-coverage-dual-src-blend.cpp| 3 ++- .../alpha-to-coverage-no-draw-buffer-zero.cpp | 3 ++- .../alpha-to-one-dual-src-blend.cpp | 3 ++- .../alpha-to-one-msaa-disabled.cpp

Re: [Piglit] [PATCH v2 2/2] util: implement eventloop for wayland platform

2016-11-02 Thread Tapani Pälli
On 11/02/2016 01:16 PM, Tapani Pälli wrote: v2: lots of fixes based on review from Pekka Paalanen - fix leak of fd in keymap handler - fix leak of wayland registry - just call wl_display_dispatch in process_events - act when key pressed (matches x11 backend behavior) - just

[Piglit] [PATCH v2 2/2] util: implement eventloop for wayland platform

2016-11-02 Thread Tapani Pälli
v2: lots of fixes based on review from Pekka Paalanen - fix leak of fd in keymap handler - fix leak of wayland registry - just call wl_display_dispatch in process_events - act when key pressed (matches x11 backend behavior) - just pass any keys, now tests that do special stuff

[Piglit] [PATCH v2 1/2] cmake: require libxkbcommon when building with Wayland

2016-11-02 Thread Tapani Pälli
This enables us to support keys properly on Wayland backend. Signed-off-by: Tapani Pälli --- CMakeLists.txt| 7 +++ tests/util/CMakeLists.txt | 5 + 2 files changed, 12 insertions(+) diff --git a/CMakeLists.txt b/CMakeLists.txt index 4002fe9..3964dd8 100644 --- a/CMakeLists.

[Piglit] [PATCH v2 0/2] eventloop for Wayland platform

2016-11-02 Thread Tapani Pälli
Round 2 *gonggg* Here's another try to implement eventloop for Wayland. I've done fixes based on Pekka's review. I did not attempt to support multiseat or multiple keyboards, such configurations are considered exotic for running Piglit and support can be added later. Main thing is to try to be con