1. 20 Dec, 2016 4 commits
    • D3D11: Fix stale SRVs with tiny depth textures. · 408293f8
      Jamie Madill authored
      When the tiny depth texture would get re-created on change, it was
      not freeing any cached SRVs which were pointing to the old tex.
      Fix this by erasing the cache entry if the texure is recreated.
      
      Also include a test which hooks into the workarounds to force
      testing on every platform.
      
      Also narrow the workaround to only apply to depth/stencil textures
      which use these mips in GL, rather than D3D, where we always create
      a full mip chain.
      
      BUG=angleproject:1664
      
      Change-Id: If0ac396b8847e1bf9b50165e5332da573e9bb3e4
      Reviewed-on: https://chromium-review.googlesource.com/421567
      Commit-Queue: Jamie Madill <jmadill@chromium.org>
      Reviewed-by: 's avatarCorentin Wallez <cwallez@chromium.org>
    • Rename MalformedShaderTest to ShaderValidationTest · 842f23f8
      Olli Etuaho authored
      Lots of these tests aren't actually testing malformed shaders, so the
      name was misleading. ShaderValidationTest reflects the content of the
      tests better.
      
      Some test classes that needlessly used different API spec versions are
      also merged.
      
      BUG=angleproject:1673
      TEST=angle_unittests
      
      Change-Id: Ib842af153a09d3887f9bb857709fd0cd29bcc9c4
      Reviewed-on: https://chromium-review.googlesource.com/422368Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
      Reviewed-by: 's avatarCorentin Wallez <cwallez@chromium.org>
      Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
    • Vulkan: Hook up loader code. · 327ba857
      Jamie Madill authored
      This integrates the build files for the loader SDK, and tests the
      compilation by calling InitInstance. There's no current way to
      test the runtime behaviour since there's no way for the tests to
      initialize the Vulkan back-end, that will come in the next CL.
      
      BUG=angleproject:1319
      
      Change-Id: Ia8bf96ca068eaf40744c9753b59ffaaa5ada8a73
      Reviewed-on: https://chromium-review.googlesource.com/367519
      Commit-Queue: Jamie Madill <jmadill@chromium.org>
      Reviewed-by: 's avatarGeoff Lang <geofflang@chromium.org>
    • D3D: Ensure compiler is loaded in applyShaders. · f0470974
      Jamie Madill authored
      We're finding that D3Dcompile is being called with a null
      compiler handle. This means at some point we successfully linked
      a program, using a valid compiler handle, and likely we went
      through some traumatic event like a GPU device lost. This unloaded
      the compiler DLL and hence leads to a null deref.
      
      It's not clear how this is happening since a device lost should
      trigger a restart of the browser process in all cases with Chrome,
      but this should fix the NULL deref and lead to a user-visible
      program link error.
      
      BUG=chromium:671162
      
      Change-Id: Ibf85f19d4f099d2f00921c230ee0a52b46bfd8cf
      Reviewed-on: https://chromium-review.googlesource.com/422553Reviewed-by: 's avatarJamie Madill <jmadill@chromium.org>
      Commit-Queue: Jamie Madill <jmadill@chromium.org>
  2. 19 Dec, 2016 9 commits
  3. 16 Dec, 2016 5 commits
  4. 15 Dec, 2016 4 commits
  5. 14 Dec, 2016 3 commits
  6. 13 Dec, 2016 9 commits
  7. 12 Dec, 2016 6 commits