Commits · edab510dfad8ae03bb94ba2fee96418e5e9aeec0 · Chen Yisong / swiftshader

08 Apr, 2019 2 commits

authored Apr 08, 2019

Change-Id: I1509e2056f5b06bb956e3968e9d179531cf6eece
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28401Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

edab510d

Replace use of ASSERT(false) with UNIMPLEMENTED · 3bb94905

authored Apr 07, 2019

... along with a message saying what switch case wasn't handled.

Helps identify what needs fixing for the daily Regres reports.

Change-Id: I79448135fc84190d6e803e90e7995ae83058f663
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28400Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

3bb94905

07 Apr, 2019 1 commit

Regres: Update test lists @ · 4ba1911b

authored Apr 07, 2019

Change-Id: Ic0a5efe8c00564198f929f4a6324598b684352e2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28398Reviewed-by: Chris Forbes <chrisforbes@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>

4ba1911b

05 Apr, 2019 13 commits

Implement Android Vulkan HAL module plumbing · b3596a6b

authored Apr 05, 2019

The Android loader looks for the symbol `HMI` which describes the module
and provides pointers to the initial entrypoints. Add a minimal
implementation of that.

Based on docs: https://source.android.com/devices/graphics/implement-vulkan#driver_emun
Bug: b/122837488
Change-Id: I367bac2046368bc8a08334cfac2f339ca2c1eb5f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28410Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

b3596a6b

Adjust Android vulkan build · 0deebed7

authored Apr 05, 2019

- Add initial device-side build configuration for LLVM7
- Made platform headers module available everywhere (host, device+core, device+vendor)
- Added missing source files from src/Vulkan/*
- Adjusted SPIRV-Tools build to be available to vendor code, and static only
  (otherwise we end up with a .so for it in the vendor image)

Bug: b/122837488
Change-Id: I9be0eaf91c90a6967fe9d3f8350c94c530f61122
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28409Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Greg Hartman <ghartman@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

0deebed7

Compilation files for Android · 94e0305a

authored Mar 29, 2019

Bug: b/122837488
Change-Id: I45fce8c0b6ef9f6a42f1d3745d7a0434dd08832e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28231Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

94e0305a

Disable surface and swapchain support for Android · 1d667d6a

authored Apr 05, 2019

On Android we will expose ANDROID_native_buffer instead, and leave the
surface and swapchain support to the android loader.

Bug: b/122837488
Change-Id: I03f0fd617042f49c19227701bbdcd2ccda883c94
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28408Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

1d667d6a

Support pNext chain entries in vkCreateImageView · 60da1580

authored Apr 05, 2019

Added support for these structures in vkCreateImageView:
VK_STRUCTURE_TYPE_IMAGE_VIEW_USAGE_CREATE_INFO_KHR
VK_STRUCTURE_TYPE_SAMPLER_YCBCR_CONVERSION_INFO

These only add validation without modifying imageView behavior.

This should affect a few thousand tests, which now hit the next
UNIMPLEMENTED() call, inside the SPIR-V compiler.

Bug b/119620767

Change-Id: Icd2f9608d6b9ca1500ada8e51621c8d54ff9cd8d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28429Tested-by: Alexis Hétu <sugoi@google.com>
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

60da1580

Sample location properly marked as unsupported · d6fae389

authored Apr 05, 2019

VK_EXT_sample_locations is not supported in SwiftShader, but all
related tests were being marked as unimplemented because
vkGetPhysicalDeviceProperties2 didn't handle
VK_STRUCTURE_TYPE_PHYSICAL_DEVICE_SAMPLE_LOCATIONS_PROPERTIES_EXT.
Explicitly ignored this case (with a proper verification that
VK_EXT_sample_locations is indeed marked as unsupported).

Bug b/117974925

Change-Id: I64a8540c0c30599c8a0a56ce322c563918ba2c31
Tests: dEQP-VK.pipeline.multisample.sample_locations_ext.*
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28428Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

d6fae389

Input aspect support in vkCreateRenderPass · 67806a17

authored Apr 05, 2019

Added support for:
VK_STRUCTURE_TYPE_RENDER_PASS_INPUT_ATTACHMENT_ASPECT_CREATE_INFO
VK_STRUCTURE_TYPE_RENDER_PASS_MULTIVIEW_CREATE_INFO
in vkCreateRenderPass.

These only add extra validations, but don't modify the renderpass' behavior.

Bug b/119620965

Change-Id: I0102a4f7fdd9f57de4ef7b0a813ead951dafe0bb
Tests: dEQP-VK.renderpass.*_use_input_aspect
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28448Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

67806a17

SpirvShader: Implement descriptor set arrays · 6b511347

authored Apr 05, 2019

Introduce a new Object::Kind - DescriptorSet.
This represents a pointer to a vk::DescriptorSet*, which is now
dereferenced to access the buffer data in OpAccessChain or OpLoad / OpStore.
This shuffling is required to handle descriptor array access
as the array index is only known at OpAccessChain time.

Tests: *dynamic*
Bug: b/126330097
Change-Id: Id754d966d8945f4e4fcf7895ed2210ce4f6ba713
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28391Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

6b511347

SpirvShader: Rework pointer types · 484e08e0

authored Apr 05, 2019

Previously Objects of the Variable kind would hold an Object::ID to another pointerBase object, which was combined with a per-lane offset in SpirvRoutine::intermediates.
This is almost exactly the same as PhysicalPointer - except the base address was taken from SpirvRoutine::physicalPointers.

With descriptor indices, we need a dynamic base pointer (only known at emit time) with per lane offsets.

This change transforms the Kind::Variable and Kind::PhysicalPointer kinds into Kind::DivergentPointer and Kind::NonDivergentPointer. This reduces complexity in loads and stores, and better represents the various forms of pointer we care about.

Bug: b/126330097
Change-Id: I514af5962b9cad4109197893066eda6f996be107
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28390Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

484e08e0

Improve the dEQP documentation for Linux. · bd3af855

authored Apr 03, 2019

The original dEQP.md covered Windows pretty well, but was confusing to
follow on Linux. So I've fleshed out the Linux process to be as complete
as the Windows process.

Change-Id: I8373e56db7d8d74bf5bcdfc7c7d23b740a11c25a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28369Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Sean Risser <srisser@google.com>

bd3af855

Fix DescriptorUpdateTemplate::updateDescriptorSet · 72438f16

authored Apr 04, 2019

The logic did not handle desciptor arrays, or the descriptor overflow case.

Bug: b/123244275
Tests: *with_template*
Change-Id: Idf3f70848136865d1d76dc4f2f41b4c35d9230db
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28388Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

72438f16

SpirvShader: Make value names more consistent · 47747616

authored Apr 04, 2019

Rename Object::Kind::Value to Object::Kind::Intermediate.
Rename SpirvRoutine::Value to SpirvRoutine::Variable.

Change-Id: I1dba30687cbf979d80b0deeca7a96e64a88c94db
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28393
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

47747616

Regres: Update test lists @ · d67ee5f5

authored Apr 05, 2019

Change-Id: I45b04d1043a83a3732aab59a9009bdbbb1ee56ba
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28397Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

d67ee5f5

04 Apr, 2019 5 commits

SpirvRoutine: Verify create functions are called once per id · 0e3d328a

authored Apr 04, 2019

Just a safety belt that we're not doing stupid things.

Change-Id: I092b0755d151a559f793e1a9f30b17ef52626d10
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28389
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

0e3d328a

Implement an Android NDK build of SwiftShader using the CMake files. · e6ab01f4

authored Apr 04, 2019

From the build directory, run:

  cmake -DCMAKE_TOOLCHAIN_FILE=../build/android.toolchain.cmake ..

Everything compiles and links. unittests, vk-unittests and ReactorUnitTests are known to work.
A "rundroid.sh" script is provided to upload and run Android binaries.
This CL has contains the first draft of an NDK-based FrameBuffer implementation.
It stubs out the gralloc calls in Image (consequence: EGLImage likely won't work).

NOTE: a small CMake patch is necessary; hopefully we'll find a way around this or land it in
CMake:

diff cmake-3.10/Modules/Platform/Android/Determine-Compiler-NDK.cmake.bak cmake-3.10/Modules/Platform/Android/Determine-Compiler-NDK.cmake
231c231
<   set(_ANDROID_TOOL_C_COMPILER_EXTERNAL_TOOLCHAIN ${CMAKE_ANDROID_NDK}/toolchains/${_ANDROID_TOOL_NAME}/prebuilt/${_ANDROID_HOST_DIR})
---
>   set(_ANDROID_TOOL_C_COMPILER_EXTERNAL_TOOLCHAIN ${CMAKE_ANDROID_NDK}/toolchains/${_ANDROID_TOOL_LLVM_NAME}/prebuilt/${_ANDROID_HOST_DIR})

Bug: b/129942368
Change-Id: I107a2f719256b6477ad105054ca68c676c05ec5c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/27869
Presubmit-Ready: Stephen White <senorblanco@chromium.org>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Stephen White <senorblanco@chromium.org>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

e6ab01f4

Create a recursive version of MutexLock, and apply it to EGL. · d39c96e9

authored Mar 15, 2019

This fixes one deadlock in Android's use of SwiftShader. Now fixed for
ASAN/tsan.

Bug: b/124530765
Test: atest CtsGraphicsTestCases:android.graphics.cts.BitmapTest#testDrawingHardwareBitmapNotLeaking -- --abi x86
Test: b/124530765#comment53
Signed-off-by: Cody Schuffelen <schuffelen@google.com>
Change-Id: Iacb04c0ac515eca1849d365c20efa84d58b75f72
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28308
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

d39c96e9

Renderer.cpp: Refactor setBatchIndices... · 90561c0e

authored Apr 03, 2019

... so we can use it for the non-indexed case.

Bug: b/118386749
Tests: dEQP-VK.draw.basic_draw.*
Change-Id: I21f1616db0b6613f8db6bff61f6aa1dfcf67c253
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28356
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

90561c0e

Regres: Update test lists @ · e3afd70e

authored Apr 04, 2019

Change-Id: Ie0776ebefe7d83eee26b98b9dbc84efa30dd2cad
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28357Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

e3afd70e

03 Apr, 2019 15 commits

Implement dynamic buffer offsets. · 225a1301

authored Apr 02, 2019

Tests: *dynamic*
Bug: b/126330097
Change-Id: I7e4f7e3d921acb72878b7728216415ba66f63ec7
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28249Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

225a1301

Removed SwiftShader's custom DrawType enum · fcbb1450

authored Nov 22, 2018

Replaced DrawType by VkPrimitiveTopology and VkIndexType.

Bug b/118386749

Change-Id: I6cd9a263b823e32c19ab85df7978cd674a757f1f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/22848Tested-by: Alexis Hétu <sugoi@google.com>
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

fcbb1450

Add support for OpKill · 97e9589a

authored Apr 02, 2019

Bug: b/124056625
Change-Id: I2e5650f81516da1e8f18775cb4b10ff5cfbb8a80
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28328
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>

97e9589a

VkPipeline: Add basic optimization passes. · 4fa92a7b

authored Apr 03, 2019

A lot of the Regres timeouts we are seeing are in LLVM code generation.
These appear to happen for shaders with a pathological number of loads and stores - most likely created by glslang.
Each of these loads and stores produce a bunch of branches in the LLVM IR (due to lane masking), which then choke LLVM.

Tests: dEQP-VK.ssbo.layout.*
Change-Id: I0bb654d9001d0d6bdd02244844bed8e73ce8d9d4
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28348
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

4fa92a7b

SpirvShader: Ignore unreachable inputs for Phis. · 5d143aa0

authored Apr 03, 2019

Bug: b/128527271
Change-Id: I4bf01e7915735d7fd3dc827fd67873fff75a756f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28354
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

5d143aa0

Regres: Reduce dEQP logging efforts · 37401472

authored Apr 03, 2019

Nothing reads these, save the SSD R/W cycles!

Change-Id: Ic49a40387672a1e5aaedb169d9d46f61daedc430
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28353Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

37401472

Regres: Remove hack to always start with a daily test · 15f2e8c1

authored Apr 03, 2019

Change-Id: Ia8793c5274985a8474407a42628b054136de8aa9
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28352Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

15f2e8c1

Regres: Include failure reasons in CI message · 24d1166a

authored Apr 03, 2019

Change-Id: I8aed6a72c4b9a2a575e441b846c581a67f82cf9b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28351Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

24d1166a

Regres: Really kill timed-out processes · f55e0a87

authored Apr 03, 2019

Only the parent process was being killed, allowing children to accumulate, eventually bringing down the machine.

Change-Id: I84e43265e53fe17058b3e3568919f6fad704f94f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28350Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

f55e0a87

Regres: Changes to reduce timeouts · e83248c4

authored Apr 03, 2019

* Shuffle the test lists to mix heavy-load tests with lighter ones.
* Bump the test timeout to 10 minutes.

Change-Id: I886b08993a39ce78fd198ac4a1ce06f7378dbb20
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28349Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

e83248c4

Regres: Update test lists @ · ba8b5ab0

authored Apr 03, 2019

Change-Id: I91eb237cd3a501eb06054b3069661de5869d8678
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28255Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

ba8b5ab0

Switch to LLVM 7.0 for Visual Studio solution · 2b9fb4bf

authored Mar 26, 2019

Uses llvm.vcxproj generated by build/cmake.sh

Bug b/115344057

Change-Id: I09041d219f194e5ec97daad6b1b247a565909e98
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28110
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

2b9fb4bf

Default to 64-bit MSVC toolchain · 04756886

authored Mar 28, 2019

Prevents linking issues with LLVM 7.0.

Bug b/115344057

Change-Id: I7270c2bcf6095b1ec9911074b3119f87680a7cad
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28168
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

04756886

Add missing LLVM 7.0 source files · 54542827

authored Mar 26, 2019

Fixes linking for MSVC build.

Bug b/115344057

Change-Id: I9a38d987f0cf8c20c670ccdb1c38dd49681bb941
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/27968
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

54542827

Rename swiftshader_icd.def to libvk_swiftshader.def · c5da6e00

authored Mar 26, 2019

The CMake build derives the .def name from the project name, which is
libvk_swiftshader.

Bug b/115344057
Bug b/116336664

Change-Id: I236839b2da4531e01f5260705c189d9ae8df9769
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/27988
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

c5da6e00

02 Apr, 2019 3 commits

.gitignore: .vscode/settings.json & .vscode/launch.json · 04dd12d8

authored Mar 27, 2019

This is per-user data and can easily contain paths to directories outside of the project.

Also remove .vscode/launch.json.

Change-Id: Iefedc52ec8446d4b6de5e6026761359ac9ac1a0f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28069
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

04dd12d8

Fix materialization of function arguments · 51d98676

authored Apr 02, 2019

Subzero does not appear to preserve function arguments passed in as
registers onto the stack by itself. Any scratch registers can get reused
for local variable register allocation.

This workaround simply materializes each concrete Reactor variable
constructed from an argument on definition. It does not ensure that
arguments turned into variables at a later point are properly preserved.
Also this solution does not centralize the materialization, making it
bug prone when new concrete Reactor types are implemented.

Bug b/129757459

Change-Id: I1007ea0e7204d05e60203b2e896589a03fc515a9
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28309Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>

51d98676

Regres: Update test lists @ · 3f20ba6b

authored Apr 02, 2019

Change-Id: I3d7474e24faccf30a11a9c9bfdae69aa1ae918bf
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28251Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

3f20ba6b

01 Apr, 2019 1 commit

Eliminate Intermediate::replace() · 459453a9

authored Mar 27, 2019

With Reactor variables now merely tracking the last assigned rvalue if
used within a single basic block, we can avoid the unsafe replace()
operation on SpirvShader::Intermediate.

This effectively reverts
https://swiftshader-review.googlesource.com/c/SwiftShader/+/27769

Bug b/129356087
Bug b/128527271

Change-Id: Ibfafc6960ac7e10b898ff8804752b928a7fc1988
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/28109Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>

459453a9