I'm looking at dEQP-GLES2.functional.clipping.line.wide*

I sent the updated client side solution to masked clear problem to Mesa dev.  This brings the number of failures down to 136 on CrOS on top of nvidia host drivers.

['dEQP-GLES2.functional.shaders.linkage.uniform_struct_partial_ivec2_ivec2']

['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec2_yyyy_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec2_tttt_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec2_gggg_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_z_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_xz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_zz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_zzz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_zzy_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_xzx_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_zzzz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_p_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_sp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_pp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_ppp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_ppt_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_sps_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_pppp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_b_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_rb_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_bb_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_bbb_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_bbg_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_rbr_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec3_bbbb_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_w_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_w_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wx_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_www_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_www_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_yyw_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wzy_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_yyyy_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wwww_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wwww_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wzzw_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_wwwy_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_zzwz_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_q_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_q_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qs_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qqq_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qqq_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_ttq_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qpt_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_tttt_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qqqq_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qqqq_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qppq_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_qqqt_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_ppqp_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_a_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_a_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_ar_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_ab_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_aaa_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_aaa_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_gga_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_abg_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_gggg_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_aaaa_vertex']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_aaaa_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_abba_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_aaag_fragment']
['dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_ivec4_bbab_fragment']

['dEQP-GLES2.functional.shaders.return.output_write_dynamic_vertex']

['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_nicest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_nicest']

['dEQP-GLES2.functional.fragment_ops.random.3']
['dEQP-GLES2.functional.fragment_ops.random.9']
['dEQP-GLES2.functional.fragment_ops.random.13']
['dEQP-GLES2.functional.fragment_ops.random.18']
['dEQP-GLES2.functional.fragment_ops.random.31']
['dEQP-GLES2.functional.fragment_ops.random.37']
['dEQP-GLES2.functional.fragment_ops.random.61']

['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.5']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.7']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.8']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.9']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.10']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.17']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.19']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.21']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.23']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.34']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.35']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.39']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.41']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.43']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.57']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.60']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.62']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.64']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.72']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.73']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.74']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.75']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.77']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.78']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.81']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.83']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.87']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.91']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.94']

['dEQP-GLES2.functional.fbo.render.recreate_colorbuffer.rebind_rbo_rgba4']
['dEQP-GLES2.functional.fbo.render.recreate_colorbuffer.no_rebind_rbo_rgba4']
['dEQP-GLES2.functional.fbo.render.recreate_colorbuffer.no_rebind_rbo_rgba4_stencil_index8']
['dEQP-GLES2.functional.fbo.render.recreate_depthbuffer.rebind_rbo_rgba4_depth_component16']
['dEQP-GLES2.functional.fbo.render.recreate_depthbuffer.no_rebind_rbo_rgba4_depth_component16']
['dEQP-GLES2.functional.fbo.render.recreate_stencilbuffer.rebind_rbo_rgba4_stencil_index8']
['dEQP-GLES2.functional.fbo.render.recreate_stencilbuffer.no_rebind_rbo_rgba4_stencil_index8']

['dEQP-GLES2.functional.depth_range.write.reverse']
['dEQP-GLES2.functional.depth_range.write.zero_to_half']
['dEQP-GLES2.functional.depth_range.write.half_to_one']
['dEQP-GLES2.functional.depth_range.write.half_to_zero']
['dEQP-GLES2.functional.depth_range.write.one_to_half']
['dEQP-GLES2.functional.depth_range.write.third_to_0_8']
['dEQP-GLES2.functional.depth_range.write.0_8_to_third']
['dEQP-GLES2.functional.depth_range.write.zero_to_zero']
['dEQP-GLES2.functional.depth_range.compare.reverse']
['dEQP-GLES2.functional.depth_range.compare.zero_to_half']
['dEQP-GLES2.functional.depth_range.compare.half_to_one']
['dEQP-GLES2.functional.depth_range.compare.one_to_half']
['dEQP-GLES2.functional.depth_range.compare.third_to_0_8']
['dEQP-GLES2.functional.depth_range.compare.0_8_to_third']

['dEQP-GLES2.functional.state_query.integers.viewport_getinteger']
['dEQP-GLES2.functional.state_query.integers.viewport_getfloat']

['dEQP-GLES2.functional.clipping.line.wide_line_z_clip']
['dEQP-GLES2.functional.clipping.line.wide_line_z_clip_viewport_center']
['dEQP-GLES2.functional.clipping.line.wide_line_z_clip_viewport_corner']
['dEQP-GLES2.functional.clipping.line.wide_line_clip']
['dEQP-GLES2.functional.clipping.line.wide_line_attrib_clip']

djmk@, the patch I sent to mesa-dev fixes dEQP-GLES2.functional.depth_stencil_clear.depth_stencil_scissored too.  Whatever solution upstream accepts will likely fix both failures. 

As such, here are the updated assignments of common failures:

djmk@
dEQP-GLES2.functional.depth_range.write.reverse

tutankhamen@
dEQP-GLES2.functional.shaders.return.output_write_dynamic_vertex

@marcheu
dEQP-GLES2.functional.clipping.line.wide_line_z_clip
dEQP-GLES2.functional.clipping.line.wide_line_z_clip_viewport_center

elie.tournier@
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_nicest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_nicest']


jakob.bornecrantz@
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.5']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.7']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.8']
....

collabora@
['dEQP-GLES2.functional.fragment_ops.random.3']
['dEQP-GLES2.functional.fragment_ops.random.9']

+Zel -- this is the bug for tracking deqp on virgl

FYI, I'm looking into dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.mediump_*

patches sent upstream for dEQP-GLES2.functional.clipping.line.wide_line -- I'm going to look at dEQP-GLES2.functional.fbo.render.recreate_* now.

dEQP-GLES2.functional.fbo.render.recreate_* seems to be failing on virgl+nvidia because nvidia doesn't downsample the RGBA8 texture to RGBA4 the way deqp expects.

Note that however, the test passes on nvidia+glx desktop if I build deqp on my workstation:
cmake . -DDEQP_TARGET=x11_glx
make -j
modules/gles2/deqp-gles2 --deqp-case=dEQP-GLES2.functional.fbo.render.recreate_colorbuffer.rebind_rbo_rgba4 --deqp-log-images=enable --deqp-watchdog=disable --deqp-surface-width=256 --deqp-surface-height=256 --deqp-log-filename=/tmp/gles2.log
...but that's because it is running a 8-bit accuracy there, because nvidia converts RGBA4444 render targets to RGBA8888 so everything is RGBA8888.

I am not sure what the answer is here, and whether the 8-to-4 conversion is legit or not; it might be a test problem. Need to look more on Monday.

Hello,

We noticed that cts-runner and glcts produce different results.
For example, most of the 'dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.*' fail with cts-runner but if we run it through glcts, only the test 61 fail.
(Our config: Ubuntu guest with Oibaf mesa)

We can maybe reduce the number of fail test thanks to that.
Can you copy/paste the output of
'./glcts --deqp-case=dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.*'?

I will provide a list of our failing test later or tomorrow.

We don't use cts-runner or glcts.  We've installed the dEQP executable on the guest image and use a python script (https://chromium.googlesource.com/chromiumos/third_party/autotest/+/master/client/site_tests/graphics_dEQP/graphics_dEQP.py) to run a test list.  After parsing the Python results, we then run tests failing tests individually when debugging, i.e:

/usr/local/deqp/modules/gles2/deqp-gles2 --deqp-case=dEQP-GLES2.functional.depth_stencil_clear.depth --deqp-surface-type=pbuffer --deqp-gl-config-name=rgba8888d24s8ms0 --deqp-log-images=enable --deqp-watchdog=disable --deqp-surface-width=256 --deqp-surface-height=256 --deqp-log-filename=/tmp/gles2.qpa

I recommend you try to run test cases individually.  Maybe there's a difference in case-lists or gl-configs used by glcts and cts-runner?  For us, 30 dEQP-GLES2.functional.fragment_ops.interaction.basic_shader* tests fail, which are listed in comment #8.

Dave fixed dEQP-GLES2.functional.shaders.linkage.uniform_struct_partial_ivec2_ivec2 and dEQP-GLES2.functional.shaders.swizzles.vector_swizzles.* with https://patchwork.freedesktop.org/patch/204436/.  There are 64 remaining GLES2 failures with CrOS + Nvidia drivers.

I'm looking into dEQP-GLES2.functional.state_query.integers.viewport now, btw ...

I checked, and dEQP-GLES2.functional.fbo.render.recreate_rebind_rbo_rgba4 is failing because the deqp limits are assuming rounding to nearest, which isn't specified. Most platforms upcast RGBA4 to RGBA8 or have different rounding modes.

I'll see if I can fix the test.

I'm looking at dEQP-GLES3.functional.fragment_out.random.86 now

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/36133208f44d6b53d814a32526c4b6674811723f

commit 36133208f44d6b53d814a32526c4b6674811723f
Author: Gurchetan Singh <gurchetansingh@chromium.org>
Date: Thu Feb 15 08:10:47 2018

media-libs/mesa: sync virgl dEQP effort with upstream

We should keep our Mesa in sync with upstream as we try to fix
dEQP on virgl.

BUG= chromium:807873 
TEST=dEQP-GLES2.functional.state_query.integers.viewport_getinteger
     dEQP-GLES2.functional.state_query.integers.viewport_getfloat

Change-Id: Ideaa0834f813b7c928f97cf8abd1a264469d3f57
Reviewed-on: https://chromium-review.googlesource.com/920793
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[modify] https://crrev.com/36133208f44d6b53d814a32526c4b6674811723f/media-libs/mesa/mesa-17.2.3.ebuild
[modify] https://crrev.com/36133208f44d6b53d814a32526c4b6674811723f/media-libs/mesa/mesa-9999.ebuild
[rename] https://crrev.com/36133208f44d6b53d814a32526c4b6674811723f/media-libs/mesa/mesa-17.2.3-r20.ebuild
[add] https://crrev.com/36133208f44d6b53d814a32526c4b6674811723f/media-libs/mesa/files/17.3-mesa-don-t-clamp-just-based-on-ARB_viewport_array-ex.patch

Patch sent for dEQP-GLES3.functional.fragment_out.random.86. I am looking at dEQP-GLES3.functional.texture.units.all_units.mixed.8.

FYI Elie,

['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_nicest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_nicest']

has been fixed.  I believe it was "vrend: Emit dithering state" that did the trick, though I haven't tested to see which of the newer patches fixed it.

Indeed, issues have been resolved using the GL Desktop backend.
But if we use GLES on the host, we can still see the failure.

Attached are the latest GLES3 results.  We have 1421 GLES3 failures, though they can be grouped into 19 categories.  We can't run the entire GLES3 suite everytime, due to segfaults.

Notes:

1) The number of failures has reduced from 2582 (2/14/2018) to 1421 (2/15/2018).

2) However, they were regressions in these tests:

dEQP-GLES3.functional.buffer.map.write.usage_hints.uniform_stream_copy
dEQP-GLES3.functional.buffer.copy.basic.uniform_copy_read

dEQP-GLES3.functional.occlusion_query.scissor_stencil_clear
dEQP-GLES3.functional.occlusion_query.conservative_depth_clear
dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_clear
dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_depth_clear_stencil_write

I'm currently investigating these regressions at the moment.

3) The GLES3 suite doesn't complete all the time due to segfaults.  Specifically, running dEQP-GLES3.functional.draw.draw_elements.triangle_fan.default_attribute should cause virglrenderer and your VM to crash.  Nobody is investigating this at the moment.

4) marcheu@ is currently investigating the dEQP-GLES3.functional.shaders.texture_functions.*.sampler2darray failures.

Deleted: deqp-gles3-2.14.2018 210 KB

deqp-gles3-2.14.2018 210 KB Download

Deleted: deqp-gles3-2.15.2018.txt 108 KB

deqp-gles3-2.15.2018.txt 108 KB View Download

Okay, I can't reproduce the regressions consistently.  Maybe they are flaky.  For now, let's focus on consistently failing tests.

marcheu@ is looking at the crash, and I'm looking at dEQP-GLES3.functional.ubo.*

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/0fab168546b477420307653f2b31fdb107927522

commit 0fab168546b477420307653f2b31fdb107927522
Author: Gurchetan Singh <gurchetansingh@chromium.org>
Date: Sat Feb 17 05:41:59 2018

media-libs/mesa: backport remap query types to hw support fix

This fixes over 900 dEQP-GLES3.functional.transform_feedback.* tests
on virgl.

BUG= chromium:807873 
TEST=emerge-bettyvirgl mesa

Change-Id: Ic324a72046330ec8672acbc633f40f09b39f67de
Reviewed-on: https://chromium-review.googlesource.com/923399
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[modify] https://crrev.com/0fab168546b477420307653f2b31fdb107927522/media-libs/mesa/mesa-17.2.3.ebuild
[modify] https://crrev.com/0fab168546b477420307653f2b31fdb107927522/media-libs/mesa/mesa-9999.ebuild
[add] https://crrev.com/0fab168546b477420307653f2b31fdb107927522/media-libs/mesa/files/17.3-virgl-remap-query-types-to-hw-support.patch
[rename] https://crrev.com/0fab168546b477420307653f2b31fdb107927522/media-libs/mesa/mesa-17.2.3-r21.ebuild

The crashes are fixed by https://lists.freedesktop.org/archives/virglrenderer-devel/2018-February/000251.html

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02

commit 2e9c5b185881fa57ba9e66b63f9aaa1ede10be02
Author: Gurchetan Singh <gurchetansingh@chromium.org>
Date: Fri Feb 23 11:37:14 2018

media-libs/mesa: backport v2 struct fixes

All of these patches haven't landed in upstream Mesa yet, but
probably will soon.

This fixes dEQP-GLES3.functional.ubo.*

BUG= chromium:807873 
TEST=dEQP-GLES3.functional.ubo.single_struct.single_buffer.shared_instance_array_vertex
     on Virgl

Change-Id: If9e5244c125ed6f57f4035ab78e12879c87141d0
Reviewed-on: https://chromium-review.googlesource.com/933592
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[add] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/files/17.3-virgl-reduce-some-default-capset-limits.patch
[modify] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/mesa-17.2.3.ebuild
[add] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/files/17.3-virgl-handle-getting-new-capsets.patch
[modify] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/mesa-9999.ebuild
[rename] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/mesa-17.2.3-r22.ebuild
[add] https://crrev.com/2e9c5b185881fa57ba9e66b63f9aaa1ede10be02/media-libs/mesa/files/17.3-virgl-add-offset-alignment-values-to-to-v2-caps-stru.patch

Here are the latest results:

dEQP GLES3: 1290 failures
dEQP GLES2: 51 failures.  

This is based on ToT virglrenderer and Mesa with the following extra patches:

https://patchwork.freedesktop.org/patch/206134/
https://patchwork.freedesktop.org/patch/206345/
https://patchwork.freedesktop.org/patch/206361/
https://patchwork.freedesktop.org/patch/206360/

Notes:

---  The following GLES2 tests regressed or were exposed by our recent patches. 

['dEQP-GLES2.functional.texture.mipmap.2d.generate.a8_fastest']
['dEQP-GLES2.functional.texture.mipmap.2d.generate.a8_nicest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.a8_fastest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.a8_nicest']


['dEQP-GLES2.functional.texture.specification.basic_copyteximage2d.cube_alpha']
['dEQP-GLES2.functional.texture.specification.basic_copytexsubimage2d.cube_alpha']

I'll try to figure out why this happened.

--- dEQP-GLES3.functional.flush_finish.finish_wait failed this time, but that test has always been a little wonky (see 59f4fc in your dEQP tree).

Also, here is an updated list of who's working on what.  Let me know if this is incorrect:

marcheu@
'dEQP-GLES2.functional.fbo.render.recreate* (suspected dEQP issue, need to fix test)
dEQP-GLES3.functional.shaders.texture_functions.*.sampler*

gurchetansingh@
Figure out GLES2 regressions mentioned above
dEQP-GLES3.functional.texture.specification*

djmk@
dEQP-GLES2.functional.depth_range* (with latest patch, all test pass except dEQP-GLES2.functional.depth_range.write.zero_to_zero)

tutankhamen@
dEQP-GLES2.functional.shaders.return.output_write_dynamic_vertex

elie.tournier@ (for GLES on GLES case)
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.2d.generate.la88_nicest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_fastest']
['dEQP-GLES2.functional.texture.mipmap.cube.generate.la88_nicest']


jakob.bornecrantz@
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.5']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.7']
['dEQP-GLES2.functional.fragment_ops.interaction.basic_shader.8']
....

collabora@
['dEQP-GLES2.functional.fragment_ops.random.3']
['dEQP-GLES2.functional.fragment_ops.random.9']
..

Deleted: deqp-gles2-2.23.2018..txt 3.4 KB

deqp-gles2-2.23.2018..txt 3.4 KB View Download

GLES3 here:

Deleted: deqp-gles3-2.23.2018.txt 103 KB

deqp-gles3-2.23.2018.txt 103 KB View Download

For dEQP-GLES3.functional.flush_finish.finish_wait there is https://buganizer.corp.google.com/issues/71858368 so I think it's a non-problem

The depth_range fix is out for review. I am still looking at the depth_range.write.zero_to_zero failure.  I am starting dEQP-GLES3.functional.shaders.builtin_functions.precision.refract.* now.

working on dEQP-GLES3.functional.fragment_ops.random.18

Sent a fix for the dEQP-GLES3.functional.shaders.texture_functions.texture.isampler2darray.* failures: 
https://lists.freedesktop.org/archives/virglrenderer-devel/2018-February/000262.html

I'm looking at dEQP-GLES3.functional.texture.shadow.*

With the iovec depth changes, the number of GLES3 failures is 574 with no regressions (besides flaky tests which we've encountered before).  GLES2 results remain largely unchanged from the last run.

Deleted: deqp-gles3-2.28.2018.txt 47.3 KB

deqp-gles3-2.28.2018.txt 47.3 KB View Download

all dEQP-GLES3.functional.fragment_ops.random.* tests are fixed

all dEQP-GLES3.functional.fragment_ops.interaction.basic_shader.* tests are fixed (these tests had the same problems as fragment_ops.random.*)

dEQP-GLES3.functional.shaders.builtin_functions.precision.refract.* issues appear to be hardware related, I've ordered a different graphics card to verify.  In the meantime, I'll have a look at dEQP-GLES3.functional.shaders.invariance.* and remaining gles2 failures.

Working on dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba8 now

I sent patches to fix dEQP-GLES3.functional.texture.filtering.cube.* (about 100 tests), note that this does *not* fix dEQP-GLES3.functional.texture.filtering.3d.*.

I know what the problem is for dEQP-GLES3.functional.texture.filtering.3d.*., I'll send patches shortly.

Another GLES3 run.  Note the seamless cube changes aren't reflected in this change.  Also there's a simple fix for dEQP-GLES3.functional.texture.specification.texstorage3d.format.*2d_array* which hasn't been sent out.

Deleted: gles3-3.19.txt 25.4 KB

gles3-3.19.txt 25.4 KB View Download

dEQP-GLES3.functional.shaders.texture_functions.texturelod.samplercube* already works

dEQP-GLES3.functional.texture.mipmap.cube* also works.

I'm going to look into dEQP-GLES3.functional.texture.swizzle.multi_channel.alpha_all_red and friends

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/ce92a4ac21962988e7dec19702f746b6b2642bd4

commit ce92a4ac21962988e7dec19702f746b6b2642bd4
Author: Gurchetan Singh <gurchetansingh@chromium.org>
Date: Tue Mar 20 07:00:25 2018

media-libs/mesa: backport seamless cubemap changes

This fixes hundreds of dEQP GLES3 on Betty Virgl.

Example test:
dEQP-GLES3.functional.texture.filtering.cube.formats.rgba16f_nearest
dEQP-GLES3.functional.texture.filtering.cube.formats.rgb9_e5_nearest_mipmap_nearest

BUG= chromium:807873 
TEST=emerge-bettyvirgl mesa

Change-Id: Ib75ddb03e8faaffeb4a4b97c38a26b6ba6992ef8
Reviewed-on: https://chromium-review.googlesource.com/969947
Commit-Ready: Gurchetan Singh <gurchetansingh@chromium.org>
Tested-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>

[rename] https://crrev.com/ce92a4ac21962988e7dec19702f746b6b2642bd4/media-libs/mesa/mesa-17.2.3-r23.ebuild
[modify] https://crrev.com/ce92a4ac21962988e7dec19702f746b6b2642bd4/media-libs/mesa/mesa-17.2.3.ebuild
[add] https://crrev.com/ce92a4ac21962988e7dec19702f746b6b2642bd4/media-libs/mesa/files/17.3-virgl-Implement-seamless-cube-maps.patch
[modify] https://crrev.com/ce92a4ac21962988e7dec19702f746b6b2642bd4/media-libs/mesa/mesa-9999.ebuild

Here are the latest results:

dEQP GLES3: 85 failures
dEQP GLES2: 8 failures.  

No regressions.

Also, here is an updated list of who's working on what.  Let me know if this is incorrect:

marcheu@
'dEQP-GLES2.functional.fbo.render.recreate* (suspected dEQP issue, need to fix test)
dEQP-GLES3.functional.texture.swizzle.multi_channel.*

gurchetansingh@
dEQP-GLES3.functional.texture.specification.texsubimage3d_unpack_params.*complex1*

djmk@
dEQP-GLES2.functional.depth_range.write.zero_to_zero
dEQP-GLES3.functional.shaders.invariance.*
dEQP-GLES3.functional.shaders.builtin_functions.precision.refract.* (suspected hardware bug, ordered new GPU)



tutankhamen@
dEQP-GLES3.functional.fbo.blit.*


collabora@
GLES on GLES
dEQP-GLES3.functional.fbo.completeness.renderable.texture.color0.*_unsigned_int_2_10_10_10_rev* (patch sent upstream, but not merged)

Deleted: deqp-gles3-3.20.txt 6.6 KB

deqp-gles3-3.20.txt 6.6 KB View Download

Sent a patch for dEQP-GLES3.functional.texture.swizzle.multi_channel.*

Patch sent for
dEQP-GLES3.functional.texture.specification.texsubimage3d_unpack_params.*complex1*, working on dEQP-GLES3.functional.fbo.depth.depth_write_clamp.depth_component32f and related tests.  FYI marcheu@ is working on dEQP-GLES3.functional.fbo.invalidate.whole.unbind_blit_msaa_color

I just sent a patch for the GLES3.functional.fbo.invalidate.* failures

Filed https://b.corp.google.com/issues/76443054 for the dEQP-GLES2.functional.fbo.render.recreate* failures.

Patch sent for these tests:
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba8
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgb10_a2
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba4
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgb5_a1
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgb8
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgb565
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rg8
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r8

Waiting for an approval from Dave Airlie to send a patch for these tests:
'dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgba8']
['dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8']
['dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgb10_a2']
['dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgba4']

Working on these tests:
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba32f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba16f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r11f_g11f_b10f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rg32f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rg16f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r32f']
['dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r16f']

Patch sent for these tests:
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgba8
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgb10_a2
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_rgba4

These tests are fixed too:
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba32f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rgba16f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rg32f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_rg16f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r32f
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_r16f

working on:
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8

Patch sent for:
dEQP-GLES3.functional.fbo.blit.conversion.srgb8_alpha8_to_srgb8_alpha8

chadversary@ got dEQP-GLES3.functional.shaders.builtin_functions.precision.refract.* and dEQP-GLES3.functional.fbo.depth.depth_write_clamp.depth_component32f passing on his Skylake laptop with Bettyvirgl, confirming our suspicions these are Nvidia driver bugs.

Looking into:
dEQP-GLES3.functional.rasterization.fbo.rbo_multisample_max.primitives.points

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/chromite/+/0f0cab3eae122da42bc872926d720fc68193ecb7

commit 0f0cab3eae122da42bc872926d720fc68193ecb7
Author: Ilja H. Friedel <ihf@chromium.org>
Date: Fri Apr 06 22:02:06 2018

chromeos_config: add bettyvirgl builder.

Also prepare a paladin to keep the build (but not the tests) in
working order.

BUG= chromium:807873 
TEST=chromeos_config_unittest

Change-Id: I081f58f9e918ff539d309000c05146e86bb6ab37
Reviewed-on: https://chromium-review.googlesource.com/996921
Tested-by: Ilja H. Friedel <ihf@chromium.org>
Trybot-Ready: Ilja H. Friedel <ihf@chromium.org>
Reviewed-by: Stéphane Marchesin <marcheu@chromium.org>
Reviewed-by: Bernie Thompson <bhthompson@chromium.org>

[modify] https://crrev.com/0f0cab3eae122da42bc872926d720fc68193ecb7/cbuildbot/config_dump.json
[modify] https://crrev.com/0f0cab3eae122da42bc872926d720fc68193ecb7/cbuildbot/waterfall_layout_dump.txt
[modify] https://crrev.com/0f0cab3eae122da42bc872926d720fc68193ecb7/cbuildbot/chromeos_config.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b690a8a742db6cb14594e1f1fe2f3039037c7531

commit b690a8a742db6cb14594e1f1fe2f3039037c7531
Author: chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Date: Sat Apr 07 01:33:59 2018

Roll src/third_party/chromite/ c41a0f5ed..0f0cab3ea (3 commits)

https://chromium.googlesource.com/chromiumos/chromite.git/+log/c41a0f5eda5f..0f0cab3eae12

$ git log c41a0f5ed..0f0cab3ea --date=short --no-merges --format='%ad %ae %s'
2018-04-04 ihf chromeos_config: add bettyvirgl builder.
2018-04-04 dgarrett chromeos_config: Add two informationals to swarming.
2018-04-05 khmel arc: Mirror zipalign to ARC bucket.

Created with:
  roll-dep src/third_party/chromite
BUG= chromium:807873 , chromium:829169 


The AutoRoll server is located here: https://chromite-chromium-roll.skia.org

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=chrome-os-gardeners@chromium.org

Change-Id: Icae5a4e5edf06d06d465a123eb814b593ef0c0a3
Reviewed-on: https://chromium-review.googlesource.com/1000673
Commit-Queue: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Reviewed-by: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#549009}
[modify] https://crrev.com/b690a8a742db6cb14594e1f1fe2f3039037c7531/DEPS

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/chromite/+/c23684514c93ed552d280b25b1280738eb42acfc

commit c23684514c93ed552d280b25b1280738eb42acfc
Author: Ilja H. Friedel <ihf@chromium.org>
Date: Fri Apr 13 18:24:09 2018

Revert "chromeos_config: add bettyvirgl builder."

This reverts commit 0f0cab3eae122da42bc872926d720fc68193ecb7.

Apparently it isn't needed anymore. And also it never worked.

BUG= chromium:807873 

Change-Id: I66edfc2248c0f11a6fd5002cc376c0e218ff8c74
Reviewed-on: https://chromium-review.googlesource.com/1011824
Tested-by: Ilja H. Friedel <ihf@chromium.org>
Trybot-Ready: Ilja H. Friedel <ihf@chromium.org>
Reviewed-by: Gurchetan Singh <gurchetansingh@chromium.org>
Reviewed-by: Don Garrett <dgarrett@chromium.org>

[modify] https://crrev.com/c23684514c93ed552d280b25b1280738eb42acfc/cbuildbot/config_dump.json
[modify] https://crrev.com/c23684514c93ed552d280b25b1280738eb42acfc/cbuildbot/waterfall_layout_dump.txt
[modify] https://crrev.com/c23684514c93ed552d280b25b1280738eb42acfc/cbuildbot/chromeos_config.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/800802bb8800771951b0f4ec910e3ed1c5a29b2e

commit 800802bb8800771951b0f4ec910e3ed1c5a29b2e
Author: chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Date: Fri Apr 13 20:19:23 2018

Roll src/third_party/chromite/ b5d176293..c23684514 (1 commit)

https://chromium.googlesource.com/chromiumos/chromite.git/+log/b5d17629393d..c23684514c93

$ git log b5d176293..c23684514 --date=short --no-merges --format='%ad %ae %s'
2018-04-13 ihf Revert "chromeos_config: add bettyvirgl builder."

Created with:
  roll-dep src/third_party/chromite
BUG= chromium:807873 


The AutoRoll server is located here: https://chromite-chromium-roll.skia.org

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=chrome-os-gardeners@chromium.org

Change-Id: Ic52798de6b856ce362f5d1ed42b0eef29e79395a
Reviewed-on: https://chromium-review.googlesource.com/1012598
Commit-Queue: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Reviewed-by: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#550743}
[modify] https://crrev.com/800802bb8800771951b0f4ec910e3ed1c5a29b2e/DEPS

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/800802bb8800771951b0f4ec910e3ed1c5a29b2e

commit 800802bb8800771951b0f4ec910e3ed1c5a29b2e
Author: chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Date: Fri Apr 13 20:19:23 2018

Roll src/third_party/chromite/ b5d176293..c23684514 (1 commit)

https://chromium.googlesource.com/chromiumos/chromite.git/+log/b5d17629393d..c23684514c93

$ git log b5d176293..c23684514 --date=short --no-merges --format='%ad %ae %s'
2018-04-13 ihf Revert "chromeos_config: add bettyvirgl builder."

Created with:
  roll-dep src/third_party/chromite
BUG= chromium:807873 


The AutoRoll server is located here: https://chromite-chromium-roll.skia.org

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.


TBR=chrome-os-gardeners@chromium.org

Change-Id: Ic52798de6b856ce362f5d1ed42b0eef29e79395a
Reviewed-on: https://chromium-review.googlesource.com/1012598
Commit-Queue: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Reviewed-by: Chromite Chromium Autoroll <chromite-chromium-autoroll@skia-buildbots.google.com.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#550743}
[modify] https://crrev.com/800802bb8800771951b0f4ec910e3ed1c5a29b2e/DEPS

Patches sent for:
dEQP-GLES2.functional.shaders.builtin_variable.pointcoord
CI support

I also submit a series for the QEMU integration. Patches were accepted and will be in the next graphics PR.

Hello, 

I've tested android-gles3 in qemo on an Intel kabylake host and on an r600 barts (nosb) host, attached is the list of common failures using the collabora hacks branch of virglrenderer and mesa git-193d615917. 

I intent to look into 

dEQP-GLES3.functional.fragment_out.basic.float.*
dEQP-GLES3.functional.fragment_out.array.float.*

Deleted: fail-gles3-on-hosts-r600-barts-nosb-and-intel.txt 22.2 KB

fail-gles3-on-hosts-r600-barts-nosb-and-intel.txt 22.2 KB View Download

There are results I was able to get on Intel Skylake:


dEQP-GLES2 fails:

dEQP-GLES2.functional.shaders.invariance.highp.loop_4
dEQP-GLES2.functional.shaders.invariance.mediump.loop_4
dEQP-GLES2.functional.shaders.invariance.lowp.loop_4
dEQP-GLES2.functional.depth_range.write.zero_to_zero

dEQP-GLES3 fails:

dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.scalar
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec2
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec3
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec4
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.scalar
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec2
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec3
dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec4
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.scalar
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec2
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec3
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec4
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.scalar
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec2
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec3
dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec4
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_3
dEQP-GLES3.functional.shaders.invariance.highp.loop_4
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_3
dEQP-GLES3.functional.shaders.invariance.mediump.loop_4
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_1
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_2
dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_3
dEQP-GLES3.functional.shaders.invariance.lowp.loop_4
dEQP-GLES3.functional.texture.specification.basic_teximage3d.rgb8i_3d
dEQP-GLES3.functional.texture.specification.basic_teximage3d.rgb8ui_3d
dEQP-GLES3.functional.fbo.render.resize.rbo_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.render.recreate_color.rbo_r11f_g11f_b10f_depth_stencil_rbo_depth24_stencil8
dEQP-GLES3.functional.fbo.color.clear.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.color.repeated_clear.sample.tex2d.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.color.repeated_clear.blit.tex2d.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.color.repeated_clear.blit.rbo.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.conversion.rgba32f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.rgba16f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.r11f_g11f_b10f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.rg32f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.rg16f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.r32f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.conversion.r16f_to_r11f_g11f_b10f
dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8_linear_out_of_bounds_blit_to_default
dEQP-GLES3.functional.fbo.blit.default_framebuffer.rgb8_linear_out_of_bounds_blit_to_default
dEQP-GLES3.functional.fbo.msaa.2_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.2_samples.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.msaa.4_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.4_samples.r11f_g11f_b10f
dEQP-GLES3.functional.fbo.msaa.8_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.8_samples.r11f_g11f_b10f
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_float
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec2
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec3
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec4
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_float
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec2
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec3
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec4
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_float
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec2
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec3
dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec4
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_float
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec2
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec3
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec4
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_float
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec2
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec3
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec4
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_float
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec2
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec3
dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec4
dEQP-GLES3.functional.fragment_out.random.26
dEQP-GLES3.functional.fragment_out.random.32
dEQP-GLES3.functional.fragment_out.random.46
dEQP-GLES3.functional.fragment_out.random.63
dEQP-GLES3.functional.fragment_out.random.80
dEQP-GLES3.functional.fragment_out.random.81

P.S.> I was unable to get any results for dEQP-GLES3.ubo.* because theses tests lead to crash at random points, so, I just skip them.

Here's another update for GLES on an Nividia GL backend:

dEQP GLES3: 41 failures
dEQP GLES2: 8 failures.  

Some regressions [dEQP-GLES3.functional.fbo.msaa.*_samples.rgb8]

Also, here is an updated list of who's working on what.  Let me know if this is incorrect:

marcheu@
dEQP-GLES2.functional.fbo.render.recreate* (possible dEQP issue, but https://b.corp.google.com/issues/76443054 was marked as "Won't Fix" so we may need additional triage)


djmk@
dEQP-GLES2.functional.depth_range.write.zero_to_zero
dEQP-GLES3.functional.shaders.invariance.*  -- One patch landed in virglrenderer, but another still needs to land in upstream Mesa

https://patchwork.freedesktop.org/patch/216080/


tutankhamen@
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8
dEQP-GLES3.functional.fbo.msaa.2_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.4_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.8_samples.rgb8
dEQP-GLES3.functional.rasterization.fbo.rbo_multisample_max.primitives.points


The following are considered to be Nvidia GL driver bugs since they pass on Intel, and debugging indicated no problems on the virgl side:

dEQP-GLES3.functional.shaders.builtin_functions.precision.refract.*
dEQP-GLES3.functional.fbo.depth.*.{depth_component32f, depth32f_stencil8}

Flaky tests that pass when run individually:

dEQP-GLES3.functional.occlusion_query.conservative_depth_write_depth_clear_stencil_write

Deleted: deqp-gles-5.1.2018.txt 4.0 KB

deqp-gles-5.1.2018.txt 4.0 KB View Download

Hi, 

I've submitted a patch for mesa that fixes the tests involving *.r11f_g11f_b10f*

https://patchwork.freedesktop.org/series/42414/

If you have the time please review and test. 

I'm now looking into failures with 

  dEQP-GLES3.functional.rasterizer_discard.fbo.* 

that pass when run individually but fail semi-randomly when run in batches. 

> P.S I was unable to get any results for dEQP-GLES3.ubo.* because theses tests lead to crash at random points, so, I just skip them.

In my case the tests don't crash, but I am seeing failures when I run the dEQP-GLES3.ubo.* in hasty/batch mode only. I have been working on a possible solution and if all goes well I will push the fix to the virglrenderer ML list soon. It would be interesting to see if the fix for the failures also fixes the crashes for you, or the crashes are a separate issue.

alexandros@
dEQP-GLES3.ubo.* failures in hasty/batch mode

> alexandros@
> dEQP-GLES3.ubo.* failures in hasty/batch mode

Patch has been sent for review, see:

https://patchwork.freedesktop.org/patch/220212/

The reason why the *.rasterizer_discard.* fail is because virgl doesn't update the host render state when the clear command is issued. 

A fix to comment on is submitted to mesa: 

  https://patchwork.freedesktop.org/patch/220214/

A better fix would probably be directly in virgl, but as of now I don't really know how to do this, because in virgl_clear I don't seem to have access to the render state.  

This is what I've got on Fedora 27 (skylake):

GLES3:

X  1125 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_vertex [Fail]
X  1126 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_fragment [Fail]
X  1127 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_vertex [Fail]
X  1128 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_fragment [Fail]
X 17168 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.scalar [Fail]
X 17169 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec2 [Fail]
X 17170 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec3 [Fail]
X 17171 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_vertex.vec4 [Fail]
X 17172 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.scalar [Fail]
X 17173 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec2 [Fail]
X 17174 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec3 [Fail]
X 17175 : dEQP-GLES3.functional.shaders.builtin_functions.precision.sin.highp_fragment.vec4 [Fail]
X 17184 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.scalar [Fail]
X 17185 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec2 [Fail]
X 17186 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec3 [Fail]
X 17187 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_vertex.vec4 [Fail]
X 17188 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.scalar [Fail]
X 17189 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec2 [Fail]
X 17190 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec3 [Fail]
X 17191 : dEQP-GLES3.functional.shaders.builtin_functions.precision.cos.highp_fragment.vec4 [Fail]
X 18145 : dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_1 [Fail]
X 18146 : dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_2 [Fail]
X 18147 : dEQP-GLES3.functional.shaders.invariance.highp.common_subexpression_3 [Fail]
X 18155 : dEQP-GLES3.functional.shaders.invariance.highp.loop_4 [Fail]
X 18157 : dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_1 [Fail]
X 18158 : dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_2 [Fail]
X 18159 : dEQP-GLES3.functional.shaders.invariance.mediump.common_subexpression_3 [Fail]
X 18167 : dEQP-GLES3.functional.shaders.invariance.mediump.loop_4 [Fail]
X 18169 : dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_1 [Fail]
X 18170 : dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_2 [Fail]
X 18171 : dEQP-GLES3.functional.shaders.invariance.lowp.common_subexpression_3 [Fail]
X 18179 : dEQP-GLES3.functional.shaders.invariance.lowp.loop_4 [Fail]
X 24245 : dEQP-GLES3.functional.texture.specification.basic_teximage3d.rgb8i_3d [Fail]
X 24247 : dEQP-GLES3.functional.texture.specification.basic_teximage3d.rgb8ui_3d [Fail]
X 29321 : dEQP-GLES3.functional.fbo.render.resize.rbo_r11f_g11f_b10f [Fail]
X 29393 : dEQP-GLES3.functional.fbo.render.recreate_color.rbo_r11f_g11f_b10f_depth_stencil_rbo_depth24_stencil8 [Fail]
X 29471 : dEQP-GLES3.functional.fbo.color.clear.r11f_g11f_b10f [Fail]
X 29665 : dEQP-GLES3.functional.fbo.color.repeated_clear.sample.tex2d.r11f_g11f_b10f [Fail]
X 29701 : dEQP-GLES3.functional.fbo.color.repeated_clear.blit.tex2d.r11f_g11f_b10f [Fail]
X 29737 : dEQP-GLES3.functional.fbo.color.repeated_clear.blit.rbo.r11f_g11f_b10f [Fail]
X 29954 : dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8 [Fail]
X 30142 : dEQP-GLES3.functional.fbo.blit.conversion.rgba32f_to_r11f_g11f_b10f [Fail]
X 30158 : dEQP-GLES3.functional.fbo.blit.conversion.rgba16f_to_r11f_g11f_b10f [Fail]
X 30174 : dEQP-GLES3.functional.fbo.blit.conversion.r11f_g11f_b10f_to_r11f_g11f_b10f [Fail]
X 30190 : dEQP-GLES3.functional.fbo.blit.conversion.rg32f_to_r11f_g11f_b10f [Fail]
X 30206 : dEQP-GLES3.functional.fbo.blit.conversion.rg16f_to_r11f_g11f_b10f [Fail]
X 30222 : dEQP-GLES3.functional.fbo.blit.conversion.r32f_to_r11f_g11f_b10f [Fail]
X 30238 : dEQP-GLES3.functional.fbo.blit.conversion.r16f_to_r11f_g11f_b10f [Fail]
X 30266 : dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8 [Fail]
X 30272 : dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8_linear_out_of_bounds_blit_to_default [Fail]
X 30300 : dEQP-GLES3.functional.fbo.blit.default_framebuffer.rgb8_linear_out_of_bounds_blit_to_default [Fail]
X 30370 : dEQP-GLES3.functional.fbo.msaa.2_samples.rgb8 [Fail]
X 30376 : dEQP-GLES3.functional.fbo.msaa.2_samples.r11f_g11f_b10f [Fail]
X 30392 : dEQP-GLES3.functional.fbo.msaa.4_samples.rgb8 [Fail]
X 30398 : dEQP-GLES3.functional.fbo.msaa.4_samples.r11f_g11f_b10f [Fail]
X 30414 : dEQP-GLES3.functional.fbo.msaa.8_samples.rgb8 [Fail]
X 30420 : dEQP-GLES3.functional.fbo.msaa.8_samples.r11f_g11f_b10f [Fail]
X 36209 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_float [Fail]
X 36210 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec2 [Fail]
X 36211 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec3 [Fail]
X 36212 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_lowp_vec4 [Fail]
X 36213 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_float [Fail]
X 36214 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec2 [Fail]
X 36215 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec3 [Fail]
X 36216 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_mediump_vec4 [Fail]
X 36217 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_float [Fail]
X 36218 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec2 [Fail]
X 36219 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec3 [Fail]
X 36220 : dEQP-GLES3.functional.fragment_out.basic.float.r11f_g11f_b10f_highp_vec4 [Fail]
X 36629 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_float [Fail]
X 36630 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec2 [Fail]
X 36631 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec3 [Fail]
X 36632 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_lowp_vec4 [Fail]
X 36633 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_float [Fail]
X 36634 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec2 [Fail]
X 36635 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec3 [Fail]
X 36636 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_mediump_vec4 [Fail]
X 36637 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_float [Fail]
X 36638 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec2 [Fail]
X 36639 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec3 [Fail]
X 36640 : dEQP-GLES3.functional.fragment_out.array.float.r11f_g11f_b10f_highp_vec4 [Fail]
X 37051 : dEQP-GLES3.functional.fragment_out.random.26 [Fail]
X 37057 : dEQP-GLES3.functional.fragment_out.random.32 [Fail]
X 37071 : dEQP-GLES3.functional.fragment_out.random.46 [Fail]
X 37088 : dEQP-GLES3.functional.fragment_out.random.63 [Fail]
X 37105 : dEQP-GLES3.functional.fragment_out.random.80 [Fail]
X 37106 : dEQP-GLES3.functional.fragment_out.random.81 [Fail]
X 37347 : dEQP-GLES3.functional.occlusion_query.conservative_scissor [Fail]
X 37348 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write [Fail]
X 37349 : dEQP-GLES3.functional.occlusion_query.conservative_depth_clear [Fail]
X 37350 : dEQP-GLES3.functional.occlusion_query.conservative_stencil_write [Fail]
X 37351 : dEQP-GLES3.functional.occlusion_query.conservative_stencil_clear [Fail]
X 37352 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write [Fail]
X 37353 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_clear [Fail]
X 37354 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_stencil_write [Fail]
X 37355 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_stencil_clear [Fail]
X 37356 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_depth_clear [Fail]
X 37357 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_stencil_write [Fail]
X 37358 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_stencil_clear [Fail]
X 37359 : dEQP-GLES3.functional.occlusion_query.conservative_depth_clear_stencil_write [Fail]
X 37360 : dEQP-GLES3.functional.occlusion_query.conservative_depth_clear_stencil_clear [Fail]
X 37361 : dEQP-GLES3.functional.occlusion_query.conservative_stencil_write_stencil_clear [Fail]
X 37362 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_depth_clear [Fail]
X 37363 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_stencil_write [Fail]
X 37364 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_stencil_clear [Fail]
X 37365 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_clear_stencil_write [Fail]
X 37366 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_clear_stencil_clear [Fail]
X 37367 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_stencil_write_stencil_clear [Fail]
X 37368 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_depth_clear_stencil_write [Fail]
X 37369 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_depth_clear_stencil_clear [Fail]
X 37370 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_stencil_write_stencil_clear [Fail]
X 37371 : dEQP-GLES3.functional.occlusion_query.conservative_depth_clear_stencil_write_stencil_clear [Fail]
X 37372 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_depth_clear_stencil_write [Fail]
X 37373 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_depth_clear_stencil_clear [Fail]
X 37374 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_write_stencil_write_stencil_clear [Fail]
X 37375 : dEQP-GLES3.functional.occlusion_query.conservative_scissor_depth_clear_stencil_write_stencil_clear [Fail]
X 37376 : dEQP-GLES3.functional.occlusion_query.conservative_depth_write_depth_clear_stencil_write_stencil_clear [Fail]
X 37377 : dEQP-GLES3.functional.occlusion_query.conservative_all_occluders [Fail]

Total 118 failed test. It's more than on Ubuntu, but at least it doesn't hang whole system and work more reliable. 

GLES2 tests on Fedora 27 on skylake:

X  6650 : dEQP-GLES2.functional.shaders.invariance.highp.loop_4 [Fail]
X  6656 : dEQP-GLES2.functional.shaders.invariance.mediump.loop_4 [Fail]
X  6663 : dEQP-GLES2.functional.shaders.invariance.lowp.loop_4 [Fail]
X 12804 : dEQP-GLES2.functional.depth_range.write.zero_to_zero [Fail]

Only 4 tests are failing

New patch for mesa for fixing dEQP-GLSES3.functiona.rasterizer_discard.fbo.* failures when run batched: 
 
   https://patchwork.freedesktop.org/series/42796/

I will now look into the dEQP-GLES3.functional.occlusion_query.conservative* tests. 

Partial fixes for dEQP-GLES3.functional.occlusion_query.conservative*
submitted: 
  mesa: https://patchwork.freedesktop.org/patch/221719/
  virglrenderer: https://patchwork.freedesktop.org/series/42951/

Yesterday sent a patch to Mesa that should fix a bunch of tests in dEQP-GLES3.functional.fbo.color.clear:

https://patchwork.freedesktop.org/patch/222684/

Will start looking now at dEQP-GLES3.functional.fbo.msaa.4_samples.stencil_index8.2_samples and dEQP-GLES3.functional.fbo.msaa.4_samples.stencil_index8.4_samples

In regards to:

X  1125 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_vertex [Fail]
X  1126 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_expression_fragment [Fail]
X  1127 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_vertex [Fail]
X  1128 : dEQP-GLES3.functional.shaders.preprocessor.builtin.line_and_file_expression_fragment [Fail]

We've had a workaround for these tests in our Mesa tree for a while:

https://chromium.googlesource.com/chromiumos/overlays/chromiumos-overlay/+/master/media-libs/mesa/files/17.0-glcpp-Hack-to-handle-expressions-in-line-di.patch

Upstream is aware of the problem but has marked it won't fix:

https://bugs.freedesktop.org/show_bug.cgi?id=98242

> Will start looking now at dEQP-GLES3.functional.fbo.msaa.4_samples.stencil_index8.2_samples and dEQP-GLES3.functional.fbo.msaa.4_samples.stencil_index8.4_samples

That was meant to be GLES3.functional.fbo.msaa.2_samples.stencil_index8 and GLES3.functional.fbo.msaa.4_samples.stencil_index8, when running with GLES on the host.

Have sent this patch for virglrenderer that should fix these and other test cases that depend on MSAA support:

https://patchwork.freedesktop.org/patch/222883/

I'm going to work on gles3 on gle failure: 

 dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_1
 dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_2

I have an experimental patch that fixes 

 dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_1
 dEQP-GLES3.functional.shaders.metamorphic.bubblesort_flag.variant_2
 dEQP-GLES3.functional.shaders.builtin_variable.fragcoord_xyz 

https://patchwork.freedesktop.org/series/43319/

It seems to reult in some regressions though. 

With this submitted I'll work on 
  dEQP-GLES3.functional.multisample.fbo_max_samples.*

@djmk Are you still working on the gles2 failures?
I would be happy to help if needed.

Elie, djmk@ says it's all yours :-)

@collabora,

Can anyone share a Ubuntu guest image (with desktop GL) with me?  You can upload it to Google drive (15 GB is drive limit) and share with gurchetansingh@chromium.org

I tried the directions here:

https://gitlab.collabora.com/virgl-es/documentation

And made the appropriate changes (i.e, debootstrap bionic ubuntu-vm/ http://archive.ubuntu.com/ubuntu/).  However, I get stuck in the Ubuntu boot process.  I also tried Ubuntu Cloud images (https://cloud-images.ubuntu.com/bionic/), but they have issues of their own.

gurchetansingh@, the Ubuntu image is coming.

Other subject, can you share an update of who is working on what?

I've worked on 

  dEQP-GLES3.functional.clipping.point.wide_point*

and submitted a patch to virglrenderer

  https://patchwork.freedesktop.org/series/43525/

Thanks for the image.  Here's the update on who's been assigned which GLES bugs:

tournier.elie@
dEQP-GLES2.functional.depth_range.write.zero_to_zero

djmk@
dEQP-GLES3.functional.shaders.invariance.*  -- One patch landed in virglrenderer, but another still needs to land in upstream Mesa

tutankhamen@
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8
dEQP-GLES3.functional.fbo.msaa.2_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.4_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.8_samples.rgb8
dEQP-GLES3.functional.rasterization.fbo.rbo_multisample_max.primitives.points

I can't say for sure if people are actively working on them :-) (for example, I know tutankhamen@ has been looking at a pretty bad memory leak recently).

Okay, I will look into 

    dEQP-GLES3.functional.texture.shadow.*

I'm starting to look at dEQP-GLES3.functional.texture.units.2_units.mixed.5

tutankhamen@

A recent investigation to fix some piglit regressions also fixed the following:

dEQP-GLES3.functional.fbo.msaa.2_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.4_samples.rgb8
dEQP-GLES3.functional.fbo.msaa.8_samples.rgb8

See second patch in: https://patchwork.freedesktop.org/series/43777/

With the patch tomeu send 
  https://patchwork.freedesktop.org/series/43751/
the   dEQP-GLES3.functional.texture.shadow.* tests are now passing. 

I'll look into dEQP-GLES3.functional.fbo.invalidate.* failures. 
As I understand it, the patch e4b943eeaf1 by marcheu@ was for glö hosts.  

I've submitted a patch to fix the dEQP-GLES3.functional.fbo.invalidate.*msaa failures on GLES: 
  https://patchwork.freedesktop.org/series/43525/

As requested offline, here are the latest failures (ran on 6/12)


djmk@
dEQP-GLES3.functional.shaders.invariance.*  -- One patch landed in virglrenderer, but another still needs to land in upstream Mesa

https://patchwork.freedesktop.org/patch/216080/


tutankhamen@
dEQP-GLES3.functional.fbo.blit.conversion.rgb8_to_srgb8_alpha8
dEQP-GLES3.functional.fbo.blit.default_framebuffer.srgb8_alpha8
dEQP-GLES3.functional.rasterization.fbo.rbo_multisample_max.primitives.points

I haven't ran GLES2, but Elie's patch looked fine and we should be fully compliant on Intel (Nvidia has unfixable bugs).

djmk@
Sorry, for some reason I missed that patch.
I added my r-b.

Regarding GLES2 on Intel, I still see the following failure
dEQP-GLES2.functional.rasterization.limits.points

This test isn't passing because of nVidia driver bug:

* dEQP-GLES3.functional.rasterization.fbo.rbo_multisample_max.primitives.points

it renders reference picture incorrectly. This test passes with intel gpu.

tournier.elie@
I don't see dEQP-GLES2.functional.rasterization.limits.points test in gles2 master list.

tutankhamen@
dEQP-GLES2.functional.rasterization.limits.points was add to VK-GL-CTS
with commit b2931a8957.

@tournier.elie, dEQP-GLES2.functional.rasterization.limits.points is passing for me (using gl-4.3 virglrenderer and mesa branches).  Perhaps a Mesa update will fix this?

GLES3 + GLES2 pass conformance with virgl3D now.