If we are restoring a lost device we should have deleted all Contexts.
This would have freed any bound textures.

Remove the release call and replace it with an ASSERT. One usage of the
Proxy Context remains: Display::destroyImage.

Bug: angleproject:2714
Change-Id: Id3e4b7b0341092a8fae37705a6bb171167196fe2
Reviewed-on: https://chromium-review.googlesource.com/1137880
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This frees up one more use of ProxyContext. It will also keep the
bound surface alive until it is unbound from a Texture.

Two usages of the proxy context remains.

* DisplayD3D::restoreLostDevice.
* Display::destroyImage.

Bug: angleproject:2714
Change-Id: Ied72c6ebe060d7fc1743b3313e162d540fcbfe02
Reviewed-on: https://chromium-review.googlesource.com/1137878
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

After the first test using the Mock ICD runs, an environment
variable VK_ICD_FILENAMES is set when previously it did not
exists. The cleanup of that test, rather than deleting the
environment variable, set it to an empty string, which the
Vulkan loader would use for subsequent Vulkan tests, causing
issues with Vulkan not finding the correct extensions.

Had to disable two more tests since they started failing once
those backends were enabled.

Bug: angleproject:2604
Change-Id: Ie103756eed60d255d1dd92d7ef3b726183feeb1e
Reviewed-on: https://chromium-review.googlesource.com/1137342Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Omar El Sheikh <theoking@google.com>

Only the read framebuffer should affect the flipping for ReadPixels.

BUG=angleproject:2673

Change-Id: I3733bb823f5f3ebd55df63a36cc99b510dc0b1aa
Reviewed-on: https://chromium-review.googlesource.com/1138817Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

This encapsulates most of the logic for a Program into a helper class.
Now we can store multiple instances of the Program's back-end to
implement different shader behaviour at draw time.

This will be useful for shader patching for OpenGL line segment raster.

Bug: angleproject:2598
Change-Id: I800a737088574e28f3a4ec23b91c0cb2647e4e12
Reviewed-on: https://chromium-review.googlesource.com/1127302Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This would cause internal errors to trigger a nullptr dereference.
Bug introduced in "Add gl::ErrorSet and angle::Result." 6b873dd7

Bug: angleproject:2491
Change-Id: Ia98d59d27614912631d41ff0823ee5aa50769733
Reviewed-on: https://chromium-review.googlesource.com/1138408Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Rendering from disabled attributes is implemented using small dynamic
buffers. The buffers use a stride of zero so the same vertex data is
pulled for every index.

This fixes all the disable attribute tests in dEQP.

Bug: angleproject:2444
Change-Id: I04fe139076da3e3ff723bed8eb17e333b4cb0ddf
Reviewed-on: https://chromium-review.googlesource.com/1136664
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Add two members to vk::Format:
- vertex data copy function
- flag indicating if the function converts or not
Use the function when streaming vertex data so it gets converted if needed.
Add fallbacks for integer formats.  These formats will now work everywhere,
as long as they are in client memory, not a buffer object.
Adjust test expectations accordingly.

BUG=angleproject:2405

Change-Id: I677221219d933c35740633a0ab7694293e218177
Reviewed-on: https://chromium-review.googlesource.com/1084328Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The program textures update needs to call into the incomplete textures
cache. Moving this step into an update mechanism allows us to avoid
some complexity of calling the incomplete textures cache inside the
ProgramVk class.

This enables further refactors for draw call shader patching and line
segment raster.

Bug: angleproject:2598
Change-Id: I0fd81ca58d2641572a2dd444992ccddd99ebc823
Reviewed-on: https://chromium-review.googlesource.com/1127301
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This will be useful for line segment rasterization.

Bug: angleproject:2598
Change-Id: I84912c976665ecb32903181fa820b8db88786a28
Reviewed-on: https://chromium-review.googlesource.com/1127299
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2732

Change-Id: I1fa3c46654c7e6d446ec3bb43eb3a6c01a15e8f9
Reviewed-on: https://chromium-review.googlesource.com/1138234Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>

Unclear if these tests are still failing.

Bug: angleproject:1098
Change-Id: If64a0a046d2b21b67f67aba1efb92d33a30b96c3
Reviewed-on: https://chromium-review.googlesource.com/1137390Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Pointer was freed before using it's contents to set the
value of the ImageExtent to be returned

Bug: 863603
Change-Id: Ibdb26c862f3216e7314516f7aee64942467b4add
Reviewed-on: https://chromium-review.googlesource.com/1137349Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This refactors a few methods to enable draw call shader patching. The
shader serials in the Pipeline description are inserted right before
we query the pipeline cache. This is done during a draw call. Also
renames the 'QueueSerial' member of the ObjectAndSerial class to just
'Serial' to more accurately reflect it usage in ShaderAndSerial.

Also changes the GlslangWrapper class to have all static methods. If we
need to store state we can revert these changes at some point.

Also splits the GlslangWrapper link call into two static calls. One
call is called to get the linked source code. The second call compiles
the linked sources into shader code. Only the second call will be
necessary for draw call shader patching to implement OpenGL line
rasterization in Vulkan.

Bug: angleproject:2598
Change-Id: I7bad3c3eeab1fb062c15a840836db4a28f841a26
Reviewed-on: https://chromium-review.googlesource.com/1127158
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Generate code for buffer fallbacks as well as texture fallbacks.
No functional change.

BUG=angleproject:2405

Change-Id: I9f30a2cbb3cd9ba1d18474f99cba434b030b0232
Reviewed-on: https://chromium-review.googlesource.com/1113026
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Using the non-GTest dEQP targets led to crashes when deferencing a
nullptr error logging method. Fix this by skipping platform init if
there is a nullptr logError function.

Bug: angleproject:2552
Change-Id: I161e35b49f7e9a7592cec968e3540a2e8b625f94
Reviewed-on: https://chromium-review.googlesource.com/1136663
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>

Bug: angleproject:2592
Change-Id: Iec3d111df4d8e5cef205c8afb177f0514ffac5c8
Reviewed-on: https://chromium-review.googlesource.com/1135448
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Bug: angleproject:2728
Bug: angleproject:2726
Bug: angleproject:2673
Bug: angleproject:2729
Change-Id: Iad657a116bf9fcd06c3d4cf1918aaacba4f1fd38
Reviewed-on: https://chromium-review.googlesource.com/1131515Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

Introduces a vk::Context class to contain an error handler and Renderer
pointer. This abtracts the common code path for ContextVk + DisplayVk.

Removes vk::Error in favor of the POD angle::Result class. There are a
few remaining usages of gl::Error that will have to be cleaned up when
we can change the front-end APIs.

Bug: angleproject:2713

Change-Id: I5e68f223d595c6c561b59d6a85759e5738ed43c6
Reviewed-on: https://chromium-review.googlesource.com/1128924
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

gl::ErrorSet can handle errors and is accessible from ContextImpl.
It allows the implementation to mutate the error set without using
the gl::Context directly.

angle::Result is the faster POD return value class. It should generate
optimal code. It can also be used seamlessly with the ANGLE_TRY macro.

Also introduces an internal enum as a workaround for generating and
consuming errors in the back-end. When the internal enum is used as
an error return value the error is not consumed in the front-end. This
is a temporary workaround only.

Bug: angleproject:2491
Bug: angleproject:2713
Change-Id: I6cbdaadd075ccbdf241844cbcbc4ed5c3be40a8b
Reviewed-on: https://chromium-review.googlesource.com/1133200
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Bug: angleproject:1578
Change-Id: I1ce739f596b74397b625fee3806f2e1d219b29b6
Reviewed-on: https://chromium-review.googlesource.com/1135573
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This started failing with the driver uniform rework:

Vulkan: Add driver uniforms for point coords flipping

https://chromium-review.googlesource.com/1133584

Suppress until we can investigate.

Bug: angleproject:2727
Change-Id: Id1d1fe713ac058a0bb45230ba9e292f012755bcc
Reviewed-on: https://chromium-review.googlesource.com/1135729Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

- This makes all end2end tests pass successfully on Win/Nvidia
with viewport flipping enabled.

Bug: angleproject:2673
Change-Id: I031c0a0f5fb0aede58b59479e57c2c127bcb964c
Reviewed-on: https://chromium-review.googlesource.com/1133703Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

This new node type can be used to inject preprocessor directives into
the intermediate tree. Outputting the preprocessor directive is easy
for the GLSL back-end. This can be used to implement workarounds such
as the OpenGL line rasterization rules on Vulkan.

Also fixes the build that was broken on Win/x64 with a prior change and
makes more methods in IntermNode.h const.

Bug: angleproject:2598
Change-Id: Ifd6d0ac7912ccf2137997bb9a0187fd063023d5e
Reviewed-on: https://chromium-review.googlesource.com/1133420Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This new driver uniform allows us to dynamically decide if we
flip the y coordinate of the gl_PointCoord variable.

We only flip the coordinate if we render downside-up, and otherwise
we cancel the flipping since the y coordinate doesn't need
modification anymore.

This completes all end2end and deqp gles2 tests success when viewport
flipping is enabled on Windows Nvidia.

Bug: angleproject: 2673
Change-Id: I570e01ef4965933428375f6ebd8eac83ef16fbff
Reviewed-on: https://chromium-review.googlesource.com/1133584
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This marks the internal format pixel bytes compute helpers and the
Context scratch and zero buffer helpers as "no discard". This prevents
sketchy code where we don't catch error values.

Bug: angleproject:2713
Change-Id: Ic6d285897dc36ad529cd5361d2763338578df61f
Reviewed-on: https://chromium-review.googlesource.com/1134207Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

- Preliminary change for gl_PointCoords fix for Vulkan.

Bug: angleproject:2673
Change-Id: Ie43e8aeccf75a8daa4a137b98292d848522ea2f0
Reviewed-on: https://chromium-review.googlesource.com/1134877Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

The traversal logic for many node types is essentially the same. Use
a single traverse() function for all simple node types instead of
having different ones for each node type. Special traversal code is
only needed for those node types where the traversal logic is
overridden in specific traversers or which do special bookkeeping.

This makes traverser behavior a bit more consistent: InVisit calls are
now done for all node types, including if/else, ternary and loop
nodes. Also false returned from visit function will always skip
traversing the next children of that node.

This reduces shader_translator binary size on Windows by 8 kilobytes.
The added helper functions will also make it easier to implement
alternative more efficient traversers.

Unfortunately this also regresses compiler perf tests by around 2-3%.

BUG=angleproject:2662
TEST=angle_unittests, angle_end2end_tests

Change-Id: I3cb1256297b66e1db4b133b8fb84a24c349a9e29
Reviewed-on: https://chromium-review.googlesource.com/1133009Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

This removes usages of ErrorOrResult from SurfaceVk and FramebufferVk.
This will enable the Vulkan error handling refactor.

Bug: angleproject:2713
Change-Id: I69430dba51390a5cdc18b7b47ca423b9635e65de
Reviewed-on: https://chromium-review.googlesource.com/1128923
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

* Adds a nodiscard to the Error class.
* Fixes the places where Error was discard.
* Uses more vk::Error instead of gl::Error.

This enables the Vulkan back-end Error refactoring.

Bug: angleproject:2713
Change-Id: I09606d965e8b7cfd341ad7fb95cf029b9d694ef4
Reviewed-on: https://chromium-review.googlesource.com/1128922
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

We were previously creating our device *before* initializing the
FeaturesVk fields. This means we weren't requesting the MAINTENANCE1
extension correctly. Moving feature init before the first createDevice
call fixes the ordering issue.

This unblocks the viewport flip behaviour fixes.

Bug: angleproject:2673
Change-Id: Iae6973b57bcb4da78134a17b0644cd248bfb4981
Reviewed-on: https://chromium-review.googlesource.com/1133920Reviewed-by: Luc Ferron <lucferron@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This refactor will allow us to generate different error types in
different backends. This makes Vulkan happy because it won't have to
generate gl::Errors and can stay with vk::Error.

Bug: angleproject:2713
Change-Id: I981402450f3b519d4f79851982547695d583355a
Reviewed-on: https://chromium-review.googlesource.com/1128921
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This removes the use of the ErrorOrResult class from these methods.
This will enable more performant Error handling. Also cleans up the
ANGLE_TRY_CHECKED_MATH macro to be more general.

Bug: angleproject:2713
Change-Id: I349947d320907839ca88ec1f9251e6ddc3858a08
Reviewed-on: https://chromium-review.googlesource.com/1128920
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

It was possible for ANGLE to call eglMakeCurrent with EGL_NO_SURFACE and
a valid context when the client called eglMakeCurrent with
EGL_NO_SURFACE and EGL_NO_CONTEXT.  Fix this by always binding a surface
when the driver has no native surfaceless support.

Don't expose the surfaceless extension when it's not possible to support
it (unvirtualized contexts and no native surfaceless support).

BUG=860800
BUG=angleproject:2464

Change-Id: Id8af9638d4356dbd710c453c9f196b9f25a2bbf9
Reviewed-on: https://chromium-review.googlesource.com/1131555Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

The ErrorOrResult pattern doesn't help much. Removing it enables the
Error passing refactor in the Vulkan back-end.

Bug: angleproject:2713
Change-Id: I4e8277ad856c785bf22b4d37b7ae880b534ef005
Reviewed-on: https://chromium-review.googlesource.com/1128919
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Bug: angleproject:2494
Change-Id: Ia8e374846427b7140ab2565ae5b9b18409a76d96
Reviewed-on: https://chromium-review.googlesource.com/1117323
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>

- Improving the tests revealed a bug in D3D 11 Fast Path rendering.
- These changes here are preliminary to implementing the blit for
depth/stencil in Vulkan when using the viewport flipping.

Bug: angleproject:2673
Bug: angleproject:2719
Change-Id: I6d55084e559d3110c8eeb0e7acb4e6fb09b6c1b5
Reviewed-on: https://chromium-review.googlesource.com/1132125Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

This patch is the first one of the implementation of atomic
functions in D3D11.

There are mainly two differences in the usage of GLSL and HLSL
atomic functions:
1. All GLSL atomic functions have return values, which all
   represent the original value of the shared or ssbo variable;
   while all HLSL atomic functions don't, and the original value
   can be stored in the last parameter of the function call.
2. For HLSL atomic functions, the last parameter that stores the
   original value is optional except for InterlockedExchange and
   InterlockedCompareExchange. Missing original_value in the call
   of InterlockedExchange and InterlockedCompareExchange results
   in a compile error from HLSL compiler.

To handle these differences, we plan to implement the translation
in two steps:
1. Support direct translations from GLSL atomic functions to HLSL
   ones.
Direct translation can only handle the following two situations:
(1) The sentence is a GLSL atomic function call without requesting
    a return value and it is not atomicExchange or atomicCompSwap:
    e.g.
    GLSL: atomicAdd(mem, value);
 -> HLSL: InterlockedAdd(mem, value);
(2) The sentence is a simple assignment expression: its right is
    a GLSL atomic function call and its left is a declared variable.
    e.g.
    GLSL: oldValue = atomicAdd(mem, value);
 -> HLSL: InterlockedAdd(mem, value, oldValue);

2. Support atomic functions in the situations that don't support
   direct translations.
We will modify the intermediate tree to make direct translation work
on all these situations.
e.g.
   atomicExchange(mem, value);
-> int oldValue;
   oldValue = atomicExchange(mem, value);

   int oldValue = atomicAdd(mem, value);
-> int oldValue;
   oldValue = atomicAdd(mem, value);

   return atomicAdd(mem, value);
-> int temp;
   temp = atomicAdd(mem, value);
   return temp;

   for (i = 0; i < atomicAdd(mem, value); ++i)
-> int temp;
   temp = atomicAdd(mem, value);
   for (i = 0; i < temp; ++i)
   {
       ...
       temp = atomicAdd(mem, value);
   }

   int result = isTrue ? atomicAdd(mem, value) : 0;
-> int result;
   if (isTrue)
   {
       result = atomicAdd(mem, value);
   }
   else
   {
       result = 0;
   }

This patch completes Step 1 which mainly focus on the translation
from GLSL atomic functions to HLSL ones.

BUG=angleproject:2682
TEST=angle_end2end_tests

Change-Id: I3b655b6e286dad4fd97f255f7fe87521c94db30c
Reviewed-on: https://chromium-review.googlesource.com/1121835
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>

Also downgrades several Vulkan UNIMPLEMENTED() to WARN().
Also downgrades a couple D3D-specific errors to warnings.
Also downgrades an undefined behaviour integer clear error to warning.
Also includes suppressions for failing D3D11 ES 3.1 SSBO tests.
Also includes suppressions for failing Android GLES format tests.
Also includes suppressions for failing Android Vulkan buffer tests.

Bug: angleproject:2552
Bug: angleproject:2567
Bug: angleproject:1951
Bug: angleproject:2405
Change-Id: Ie619085021d42012cd578b669f7ff4252ca41a58
Reviewed-on: https://chromium-review.googlesource.com/1062791
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Open files with mode 'r' instead of 'rb' so the hash calculation
produces the same results on Linux and Windows.
Recalculate the hashes and proc_table_autogen.cpp which was out of date.

BUG=angleproject:2711

Change-Id: I31562c96ce36e6df009c44c565fe9a3f1b5ba6c4
Reviewed-on: https://chromium-review.googlesource.com/1128549Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>