Age | Commit message (Collapse) | Author | Files | Lines |
|
Change the dimension of the texture to test how i965 handles having a
row pitch too large for the BLT engine.
v2: Query the maximum supported texture width (Ilia Mirkin)
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
We'd like to increase the pitch of the texture in the next patch.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
This will enable dynamically allocating the texture data with variable
texture dimensions.
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
Reviewed-by: Lionel Landwerlin <lionel.g.landwerlin@intel.com>
|
|
Instead of doing 900 piglit_proble_pixel_rgba calls (one for each pixel)
which each call glReadPixels, use the new helper which only invokes
glReadPixels once.
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
|
|
Test sRGB texturing on a cleared texture. This case is of interest
to Intel GPUs, which will fast-clear the texture.
|
|
Continuing to run the tests after one fails provides helpful information
when debugging.
Suggested-by: Ian Romanick <idr@freedesktop.org>
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
|
|
Enable this test to run on more system configurations.
Suggested-by: Ian Romanick <idr@freedesktop.org>
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
|
|
-128 and -127 represent exactly the same value (-1.0) for snorm8 values,
as do -32768/-32767 for snorm16. Therefore it seems quite reasonable that an
implementation would return the other representation (in particular r600 will
do this with a blit, and the snorm->float->snorm conversion implied by this
will never return the -128/-32768 values).
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
This is a problem of all texturing tests, they cannot exercise the tesselation
stages. (But I'm too lazy to fix the others, and too lazy to hack something up
for tcs stage, I only need it to verify a bug/fix with r600 buffer textures.)
v2: add to all.py so it is actually run (all in all there's about 40
addtional tests run)
Reviewed-by: Fabian Bieler <fabianbieler@fastmail.fm>
|
|
Replace GLenum stringification macros with calls to piglit_get_gl_enum_name.
V2: Make sure GL 1.0 legacy internal formats '3' and '4' are logged
correctly.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum stringification macros with calls to piglit_get_gl_enum_name.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum stringification macros with calls to piglit_get_gl_enum_name.
Also replace string comparison between command line argument and GLenum name
with call to piglit_get_gl_enum_from_name and integer comparision.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum stringification macros with calls to piglit_get_gl_enum_name.
Also replace string comparison between command line argument and GLenum name
with call to piglit_get_gl_enum_from_name and integer comparision.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum stringification macros with calls to piglit_get_gl_enum_name.
Also replace string comparison between command line argument and GLenum name
with call to piglit_get_gl_enum_from_name and integer comparision.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum string literals with calls to piglit_get_gl_enum_name.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum string literals with calls to piglit_get_gl_enum_name.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Replace GLenum string literals with calls to piglit_get_gl_enum_name.
This commit modifies all tests were the replacement is simple and
straightforward.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
The existing repeat and clamp modes are easy to implement. The
mirror_repeat (GL_MIRRORED_REPEAT) and mirror_clamp
(GL_MIRROR_CLAMP_TO_EDGE) modes however have some very interesting
differences to "gather is just the same as bilinear filtering without
the actual filtering". In particular, a bilinear filter implementation
can cut a lot of corners wrt texel selection as it doesn't need to
maintain texel order (as long as the weights follow the same order) and
can even pick different samples (as long as they have weight 0).
So, test these modes too. I used this for fixing llvmpipe, passes with
nvidia blob too (the test actually always hits exactly the cases where
one texel would have weight 0).
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
GL_MAX_PROGRAM_TEXTURE_GATHER_COMPONENTS_ARB = 1
Expected results were with respect to
GL_MAX_PROGRAM_TEXTURE_GATHER_COMPONENTS_ARB = 4. This patch fixes
it for GL_MAX_PROGRAM_TEXTURE_GATHER_COMPONENTS_ARB = 1 as well.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
supported
Fixes failure on radeon and probably others.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
v2: Remove ARB_texture_non_power_of_two requirement.
Cc: Kenneth Graunke <kenneth@whitecape.org>
Cc: Mark Janes <mark.a.janes@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=100925
Signed-off-by: Nanley Chery <nanley.g.chery@intel.com>
|
|
|
|
|
|
There is still a huge number of tests to classify but this gives us
a good number of tests to start testing with.
|
|
There is still a huge number of tests to classify but this gives us
a good number of tests to start testing with.
|
|
The test generates a mipmap by "copying" a client-array of pixels to
miplevel 0 with glDrawPixels, then calling glGenerateMipmap. The test
expects the rendered level 0 to match the client-array source. Dithering
during glDrawPixels inteferes with that expectation.
Fixes on Intel Skylake:
spec.ext_framebuffer_object.getteximage-formats init-by-clear-and-render
spec.ext_framebuffer_object.getteximage-formats init-by-render
Regressed by Mesa:
commit c4b87f129eb036c9615df3adcc1cebd9df10fc84
Author: Chad Versace <chadversary@chromium.org>
Subject: meta: Disable dithering during glGenerateMipmap
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99266
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Tested-by: Mark Janes <mark.a.janes@intel.com>
|
|
You don't get to specify depth cubes until 3.0 (GL_EXT_gpu_shader4
merge).
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
|
|
Use of gl_FragColor is deprecated since glsl 1.3
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
|
|
According to GLSLLangSpec: Use of gl_FragData and gl_FragColor
is deprecated in version 1.3, using user-defined out instead.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
|
|
1. It seems with NVIDIA's driver that using a proxy texture isn't a
sure fire way to know that a given texture format/size can actually be
created. Update the find_max_tex3d_size() function to actually try
creating a texture with glTexImage3D/glTexStorage3D and see if it
works.
2. Improve the speed of texture initialization by copying the first
3D slice to the other slices with glCopyImageSubData().
3. Use glTexStorage3D when GL_ARB_texture_storage is supported.
4. In addition to GL_RGBA8, test GL_INTENSITY8 and GL_RGBA32F formats.
5. Before testing the largest possible texture, try a couple smaller
sizes as a sanity check.
6. Loosen the piglit probe tolerance by one bit to account for inaccuracy
caused by GL_NEAREST filtering.
Tested with NVIDIA driver, VMware driver and llvmpipe.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
Fixes spurious failure on vc4, which only does unorm depth.
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
|
|
Not really needed as all the inspected pixels are overwritten
after the clear but before getting read back.
Allows one to reproduce 91926 on SKL with 100% probability:
texsubimage pbo manual GL_TEXTURE_2D GL_RGB8 6 10 0 94 53 0 -auto -fbo
6,10,0: test = 40,12,0,255 ref = 12,0,40,255 (comparing 8 bits)
texsubimage failed
target: GL_TEXTURE_2D
internal format: GL_RGB8
region: 6, 10 94 x 53
PIGLIT: {"result": "fail" }
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=91926
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Signed-off-by: Topi Pohjolainen <topi.pohjolainen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
And require an argument of either 'size' or 'format'.
Reviewed-by: Vinson Lee <vlee@freedesktop.org>
|