Verify that the rendering done by the high priority thread was also
correct by rendering a reference image at the beginning.

Add a subtest that reuses the same EGLDisplay used in the main thread.
This is just another case we can test and see if there's a difference in
the preemption behavior.

After both the main and the high priority threads have finished, check
their render timestamps to try to figure out if preemption has kicked
in.

Start a new thread right after dispatching the render commands from the
main thread, after glFlush().

This second thread creates a high priority EGL context, and uses it to
render to a frame buffer multiple times. These draw calls are supposed
to finish before the one in the main thread.

v2:
 - skip on high priority failure (Ken)
 - remove useless check (Tapani)

Check if we can create high priority contexts, otherwise just fail the
test immediately. There's no reason to test for preemption if we can't
create high priority contexts.

v2: Require configless context (Tapani)

This is the base of the test, where we use the main piglit context to
render to the framebuffer. It is supposed to be an artificially slow
draw call, that can be preempted by a higher priority context if needed.

Drivers who implement GL_ALPHA textures by swizzling a GL_R texture
needs to be prepared to offer multiple blend-colors, otherwise they
will produce the wrong result, because the blend-color will also
need per-format swizzling.

Similarly, when drivers implement GL_RGB with GL_RGBA and adjust the
blend-factors, they need to have separate blend-factors for all
framebuffer attachments or the results will be wrong.

This reveals a couple of bugs in virgl, where these currently fail.

Signed-off-by: Erik Faye-Lund <erik.faye-lund@collabora.com>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>

Sergii Romantsov authored Oct 24, 2018 and Brian Paul committed Oct 25, 2018

Trivial fix of memory leaking for test 'initialized-vbo'.

v2: Rb added.

Fixes: 6a467e00 (security/initialized-vbo: test contents of uninitialized VBOs)
CC: Brian Paul <brianp@vmware.com>
Signed-off-by: Sergii Romantsov <sergii.romantsov@globallogic.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>

glslangValidator rejects all of the shaders that try to change the
qualifier with errors like:

    ERROR: 0:8: 'redeclaration' : cannot change qualification of gl_ClipDistance

Mesa currently allows all of them to compile.  Closed-source drivers
have not been tested.

glslangValidator rejects the shaders that try to change the type with:

    ERROR: 0:8: 'gl_ClipDistance' : redeclaration of array with a different element type

Mesa currently rejects these with a somewhat less helpful error message.
Closed-source drivers have not been tested.

NOTE: The lone .vert test is here because vertex shader array input were
added in GLSL 1.50.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Cc: Andrii Simiklit <andrii.simiklit@globallogic.com>

glslangValidator rejects all of the shaders that try to change the
qualifier with errors like:

    ERROR: 0:8: 'redeclaration' : cannot change qualification of gl_ClipDistance

Mesa currently allows all of them to compile.  Closed-source drivers
have not been tested.

glslangValidator rejects the shaders that try to change the type with:

    ERROR: 0:8: 'gl_ClipDistance' : redeclaration of array with a different element type

Mesa currently rejects these with a somewhat less helpful error message.
Closed-source drivers have not been tested.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Cc: Andrii Simiklit <andrii.simiklit@globallogic.com>

glslangValidator rejects all of the shaders that try to change the
qualifier with errors like:

    ERROR: 0:8: 'redeclaration' : cannot change qualification of gl_TexCoord

Mesa currently allows all of them to compile.  Closed-source drivers
have not been tested.

glslangValidator rejects the shaders that try to change the type with:

    ERROR: 0:8: 'gl_TexCoord' : redeclaration of array with a different element type

Mesa currently rejects these with a somewhat less helpful error message.
Closed-source drivers have not been tested.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Cc: Andrii Simiklit <andrii.simiklit@globallogic.com>

See https://lists.freedesktop.org/archives/piglit/2013-October/007935.html


for the discussion.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Cc: Andrii Simiklit <andrii.simiklit@globallogic.com>

Don't use C++11 features.

Reported-by: Jan Vesely <jan.vesely@rutgers.edu>

Test extremes of GS amplification under tessellation.

Acked-by: Marek Olšák <marek.olsak@amd.com>

Test the extremes of # of GS output vertices / components per GS input
primitive.

Existing tests only cover cases of high # of components per vertex, but
not cases of low # of components per vertex and high # of output vertices.

Acked-by: Marek Olšák <marek.olsak@amd.com>

Danylo Piliaiev authored Jul 23, 2018 and Nanley Chery committed Oct 09, 2018

Test for the regression which happened when GL_TEXTURE_BUFFER was
allowed to have incompatible format.

v2: Removed unnecessary code duplication - use upload_image instead
     of init_level. (Francisco Jerez)
v3: Removed upload_image call because image is already called
     by init_image. (Francisco Jerez)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106465



Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Francisco Jerez <currojerez@riseup.net>

Eric Engestrom authored Oct 07, 2018 and Eric Engestrom committed Oct 09, 2018

CID: 1439995
Fixes: f787ad0b "Add tests for GL_EXTENSION_STRING vs. old
                              idTech2 / idTech3 games"
Cc: Ian Romanick <ian.d.romanick@intel.com>
Signed-off-by: Eric Engestrom <eric@engestrom.ch>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>

Current mesa fails both of these (in different ways!).

Inspired by WebGL conformance test
conformance2/glsl3/vector-dynamic-indexing-swizzled-lvalue.html

Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108012

As people dig up other games, we can (and should) easily add them.

v2: Fix some comments and reduce the limit from 4k to 2k.  Suggested by
Roland.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Cc: Federico Dossena <info@fdossena.com>

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

I noticed that although the tests ran successfully, the output of
'piglit summary console' did not have proper names.

For example, before:
../../build/debug/piglit/target_api/no_api/tests/tests/fast_color_clear/all-colors: pass

after:
fast_color_clear/all-colors: pass

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

Mesa commit 710b1d2e665ed654fb8d52b146fa22469e1dc3a7 introduced a bug
with downloading void-extent blocks that have channel values between 0
and 4 (exclusive). Test this case.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

This array will be reused in the next patch.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

Reproduces a GPU hang on radeonsi.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107857


Signed-off-by: Józef Kucia <joseph.kucia@gmail.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>

Our sin(0.0) and cos(0.0) were too inaccurate, causing twisting in
SDL2's rendering.

v2: Use vertex shader passthrough.

Reviewed-by: Tapani Pälli <tapani.palli@intel.com>

Fixes: f07e9df4 "sanity: Add various shader_runner tests"
Reported-by: Mark Janes <mark.a.janes@intel.com>
Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>

Piglit fails to build with this commit.

This reverts commit f07e9df4.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Acked-by: Dylan Baker <dylan@pnwbakers.com>

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>

This is the old interface which is now not depended on by any tests.

v2: - remove shader_test interfaces as well

v2: - Serialize no-process-isolation shader runner tests

This converts the two test types that use the fast-skipping mechanism
(glsl parser and shader) to use the new interfaces. This has been
verified by running with PIGLIT_NO_FAST_SKIP=1 and without, and sans 1
test (that is fixed later in the series) the results are the same.

Currently fast skipping is implemented such that it assumes there's a
single version of ES, a single version of desktop, and all extensions
are shared between them. This has basically worked because 1) there are
basically no gles1 tests, and 2) piglit didn't have compat profile. But
worked and correct are two different things.

With the addition of compat profiles it's time to re-evaluate how fast
skipping works. Namely we need to have different attributes for ES1,
ES1+, core, compat, and I've added on for "legacy" (pre-profile), since
waffle supports that.

This maintains legacy interfaces so that existing code continues to
work.

v2: - Fix versions < 3.1 on implementations without core profile

So don't treat it like one. Stops one test from being fast-skipped that
shouldn't be.

Danylo Piliaiev authored Sep 19, 2018 and Timothy Arceri committed Sep 20, 2018

These tests test the case when initialising with incompatible type
changed a type of the variable being initialized.
While main manifestation of the issue is overly verbose and incorrect
error message it did result in a crash in case of second test.

v2: Splitted the test in several ones (Timothy Arceri)

Signed-off-by: Danylo Piliaiev <danylo.piliaiev@globallogic.com>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=107547