Skip to content

A
angle
Switch branch/tag
  1. 24 Jun, 2016 1 commit
  3. 23 Jun, 2016 3 commits
  5. 22 Jun, 2016 6 commits
  7. 21 Jun, 2016 2 commits
  9. 20 Jun, 2016 10 commits
  11. 17 Jun, 2016 8 commits
  13. 16 Jun, 2016 6 commits
    • D3D11: Fix readback of BGRA-backed formats. · 230d9561
      Jamie Madill authored 
      For some BGRA-backed formats (RGBA4, R5G6B5, RGB5A1), our ReadPixels
implementation wasn't aware the BGRA format didn't exactly match the
RGBA format. For these it would do the 'fast path' memcpy method, when
it should stop and do the slow pixel-by-pixel packing method.

BUG=angleproject:1407
BUG=chromium:616176

Change-Id: Ie24758513af6f9ef87f0aa503135456c96493701
Reviewed-on: https://chromium-review.googlesource.com/352252Reviewed-by: 's avatarGeoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
    • perf_tests: Work around stalls in gl_null test teardown. · cd1b1226
      Jamie Madill authored 
      The Clear command in the draw call benchmark was not being properly
no-op'ed. This would cause the driver to get a huge queue of GL
command, which it would only flush when the test was torn down. This
manifested as a 5+ second stall on test teardown.

Looking at the clear command, it does some work with binding
and then calling GL functions, so properly filtering out the driver
commands would be a bit more complicated. For now, we can hack in a
workaround for the draw call test (the only one that uses the GL
NULL renderer at the moment) to not call clear when it would give
us problems.

BUG=620687

Change-Id: I64a2ed1883e8d3bf0d4c423c8bce0902ae9bf8df
Reviewed-on: https://chromium-review.googlesource.com/352828Reviewed-by: 's avatarGeoff Lang <geofflang@chromium.org>
    • Fix unpacking overlapping unpack buffer rows on NVIDIA GL · 218cf9ee
      Olli Etuaho authored 
      When unpack parameters are set so that rows being read overlap in
the unpack buffer stored in GPU memory, NVIDIA GL driver may not
upload the last pixels of the last one or more rows of a texture. The
driver may also crash when the amount of overlap is high.

This issue affects both TexImage* and TexSubImage* calls.

Work around the issue by uploading textures row by row when the rows
being read overlap in the unpack buffer. The workaround could possibly
be optimized by uploading several of the first rows with a single call
in some cases where the amount of overlap is low, but this is expected
to be a rarely used corner case, so the added complexity that the
optimization would create seems like a bad tradeoff.

The issue does not seem to be triggered when the layers (images) of a
3D texture overlap, as long as the rows inside the images don't.

The workaround has been ported from Chromium.

This patch adds setting dirty bits when unpack state is set in
StateManagerGL.

The included test case also reveals some issue in the D3D backend, but
this is left to be addressed later.

BUG=angleproject:1376
TEST=angle_end2end_tests

Change-Id: I7dbe73ebb70bbbc284fa92381546f4f2f832d333
Reviewed-on: https://chromium-review.googlesource.com/346430Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
    • D3D11: Prevent DXGI hooking the Alt+Enter keystroke · cd9aa12a
      James Ross-Gowan authored 
      When ANGLE is using D3D11 and not using DirectComposition, DXGI will hook
the window's message loop and override Alt+Enter to trigger a transition
to exclusive fullscreen mode. This disables that behaviour, since it's
inconsistent between backends and, at least during my testing, the
application did not even work correctly after the transition to fullscreen
mode.

This commit should be similar to the workarounds implemented in Qt and mpv
to prevent this behaviour and will hopefully make those workarounds
unnecessary. Only the DXGI_MWA_NO_ALT_ENTER flag is used, since the other
flags didn't seem to have an effect and probably only matter in fullscreen
mode.

BUG=angleproject:1414

Change-Id: Icb70c2c1405225619dcb27c5ffc9472fddd1157b
Reviewed-on: https://chromium-review.googlesource.com/352950Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
    • Validate that unpack skip is in bounds · 989cac34
      Olli Etuaho authored 
      Unpack skip needs to be taken into account when determining which part
of the unpack buffer is read. This is now done in the out-of-bounds
check when validating texture upload calls.

Unpack skip code is removed from D3D9 backend, since skip is not
supported in GLES2.

BUG=angleproject:1411
TEST=angle_end2end_tests

Change-Id: I0db4db0877a352613c57e2820e5b650edb5a73ab
Reviewed-on: https://chromium-review.googlesource.com/352450Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
    • Don't allow returning a struct containing an array in ESSL 1.00 · e29324ff
      Olli Etuaho authored 
      ESSL 1.00.17 section 6.1 forbids this. Returning arrays was already
covered by the checks that disallow declaring array types in ESSL
1.00.

BUG=angleproject:1015
TEST=dEQP-GLES2.functional.shaders.functions.* (2 new tests pass)
     dEQP-GLES3.functional.shaders.functions.* (no regression)

Change-Id: Iaffa1631c0c940afb57819221e7e8603b2305021
Reviewed-on: https://chromium-review.googlesource.com/352920Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
Reviewed-by: 's avatarCorentin Wallez <cwallez@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
  15. 15 Jun, 2016 4 commits