Google Benchmark by default only measures time spent in the main thread.
Since our Vulkan queue implementation executes command buffers on a
separate thread, this can lead to very little "CPU" time being
reported. Use MeasureProcessCPUTime() instead to use the total CPU usage
of the process by all the threads: https://github.com/google/benchmark#cpu-timers

Bug: b/177624844
Change-Id: I071cd31095a83a5afec51120fb78e42decda2083
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52628Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>

... to allow rolling SwiftShader for Android.

external/swiftshader/third_party/llvm-10.0/llvm/lib/Support/
ManagedStatic.cpp:21:30: error: unused variable 'ManagedStaticMutex'

external/swiftshader/third_party/llvm-10.0/llvm/lib/Support/
ManagedStatic.cpp:22:24: error: unused variable 'mutex_init_flag'

Bug: b/171498948
Test: build inside Android repo
Change-Id: Idc3ab2ef4de759c14499454f5d841b750124686b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52648
Presubmit-Ready: Jason Macnak <natsu@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>

This reverts commit 3e9b79ff.

Reason for revert: Android does not allow CoW mappings to be EXEC because it assumes they are text relocations (when selinux enforcing

Change-Id: I1227a2caa18bdad1a46873e4ec420a83cb21821b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52630
Presubmit-Ready: Jason Macnak <natsu@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>
Tested-by: Jason Macnak <natsu@google.com>

On Linux-based machines, a shared library's path may be resolved in
multiple ways, such as from LD_LIBRARY_PATH. For these OSes, use dlopen
to see if it exists. Note that trying to create a vk::DynamicLoader with
an invalid path throws an exception, and enabling exceptions is not
always possible, so we want to detect the error case beforehand.

Bug: b/177624844
Change-Id: I5f27cfe485afbac0e4f6bda14eed2809669c2510
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52629Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Presubmit-Ready: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>

This change moves the VulkanBenchmark and DrawBenchmark classes to
VulkanWrapper so that they can be used from other unit tests -- namely,
vk-unittests. In doing so, it became clear that using these as base
classes wasn't great for writing googletests, as text fixtures are
classes themselves, and this resulted in messy multiple inheritance. So
I modified the two classes to use callback registration instead of
virtual functions.

Apart from reworking existing tests (e.g. see TriangleBenchmark.cpp), I
also added a new DrawTests.cpp to vk-unittests with a unit test to make
sure we don't crash when leaving out "gl_Position", a bug that sugoi@
fixed in swiftshader-cl/51808. This is a good example of how easy it can
be to write such unit tests now.

List of changes:

* Moved VulkanBenchmark and DrawBenchmark to VulkanWrapper, and renamed
  VulkanTester and DrawTester respectively.
* ClearImageBenchmark refactored to aggregate a VulkanTester. This is an
  example where using a class is fine as we can still use the testers
  via aggregation.
* TriangleBenchmark tests refactored to use DrawTester and register
  callbacks.
* Moved compute tests to a ComputeTests.cpp.
* Moved the other tests to BasicTests.cpp.
* Added DrawTests.cpp with new DrawTests.VertexShaderNoPositionOutput
  test.
* CMake: add VulkanWrapper target for unittests as well as benchmarks.
* CMake: change FOLDER to better organize the tests and benchmarks for
  VS.

Bug: b/176981107
Change-Id: Ib1a0b85b3df787d2e39da08930414f9a14954a73
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52348
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>

Use the constexpr constructor for _MSC_VER >= 1925, otherwise
ManagedStatic will have a dynamic initializer, which depending on init
order, results in the Ptr field being overwritten with 0. This
eventually leads to multiple instances of the same ManagedStatic
instance in the StaticList, and asserts when double destroying during
llvm_shutdown.

I reported this bug [here](https://bugs.llvm.org/show_bug.cgi?id=49027),
and learned that this bug had already been fixed in upstream LLVM.

Note that llvm_subzero already has a similar change, though for _MSC_VER
>= 1920. According to the LLVM comment, the VC++ compiler up until 1925
may still emit a dynamic initializer for a constexpr constructor, but I
suppose we never ran into that for Subzero, so I'll leave this as is.

Bug: b/175782868
Change-Id: Ice3944f67e496aa94f1a7ed7502b49e763d702b4
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52508
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>

The LLVM JIT fails loudly when attempting to parse this command line
option when targeting non-x86 CPUs.

Bug: b/157555596
Change-Id: Ic5ddccbdbc86c2f03ded5f4004369ece0100c031
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52488
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Fuzzing tests generate shaders with large arrays or very high numbers of
local variables, which can cause stack overflow. We need to limit the
allowable stack memory usage of generated routines.

Note this change does not yet gracefully deal with routines which exceed
this limit. They will cause a null pointer dereference instead of a
stack overflow.

The default stack size limit of 1 MiB at the Subzero level is to ensure
we catch cases of excessive stack sizes even in the case no explicit
limit was set. At the Reactor level we reduce it to 512 KiB to prevent
actual stack overflow for a 1 MiB stack, assuming some earlier calls
might want to use the stack. Also, our legacy 'ASM' compiler for GLSL
allocates 4096 'registers' of 4 components for 128-bit SIMD, which
already requires 256 KiB.

Bug: b/157555596
Change-Id: I474285eecc786496edffbaef29719ca0cdf03f7d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52329
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>

Fuzzing tests generate shaders with large arrays or very high numbers of
local variables, which can cause stack overflow. We need to limit the
allowable stack memory usage of generated routines.

Note this change does not yet gracefully deal with routines which exceed
this limit. They will cause a null pointer dereference instead of a
stack overflow.

The 512 KiB stack size limit is chosen to prevent actual stack overflow
for a 1 MiB stack, assuming some earlier calls might want to use the
stack. Also, our legacy 'ASM' compiler for GLSL allocates 4096
'registers' of 4 components for 128-bit SIMD, which already requires
256 KiB.

Bug: b/157555596
Change-Id: I25c57420f6d2af323ce98faf515feca0aa834a4a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51548
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>

This fixes the following issues when cross compiling SwiftShader
for arm64 macOS using CMake:

- Sets architecture based on CMAKE_OSX_ARCHITECTURES settings.
- Add missing dependency to LLVM.
- Add a workaround for assembly file compilation on macOS in marl.

Bug: fix build
Test: cmake .. -DCMAKE_OSX_ARCHITECTURES=arm64

Change-Id: I8abba9e39907d8f4d59f51c57f09cd55d4edc148
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51890
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Yilong Li <liyl@google.com>
Commit-Queue: Yilong Li <liyl@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

These SPIR-V execution modes can be used to optimize early depth test,
but are currently ignored, which is fine from a correctness point of
view.

We need to handle these execution modes because they can appear in any
shader, without the implementation having to support using them for
optimizations. Also, we can keep the SpirvShader::Modes fields that
correspond to them for debugging purposes (e.g. an application might
incorrectly rely on early depth test optimization to happen).

Bug: b/177915067
Change-Id: I851fdfbcdde5be6287bc4ee689fdbd6549e2900f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51908
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

VectorType::getNumElements() will be removed in LLVM 12. Only
FixedVectorType will support this method. Cast to the latter where
appropriate.

Since LLVM 10 does not support FixedVectorType, treat it as an alias
for the legacy VectorType implementation.

Bug: b/354893276
Change-Id: Ie8160478819ff1c1bcd7fbe4425d2264869770e9
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52308Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Clang produces:
"error: code will never be executed [-Werror,-Wunreachable-code]"
For code following an unconditional `break`. But it hints that:
"note: silence by adding parentheses to mark code as explicitly dead"

Hence the dead code can follow an "if((true)) break;" without producing
the warning/error.

Note that in this instance we could have just removed the dead code, but
I wanted to make it clear that this is an example of an unfinished
Reactor function.

Bug: b/178929463
Change-Id: I1ea4dac91877fcede189d09e3953568b2813129e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52288Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Currently, VkImages with VK_IMAGE_TILING_OPTIMAL tiling actually
also have a linear texel layout in our current implementation,
thus it can be sampled or used as blit source. And we would
like to add support to sampling features for linear images when
calling vkGetPhysicalDeviceImageFormatProperties() to query image
format properties.

Exceptions include images with compressed formats and images
created to be used as cube maps.

This will unblock SwiftShader users like AEMU/FEMU from sampling
host-visible memory shared between guest processes.

Bug: b/171299814
Bug: fuchsia:54153
Bug: fuchsia:68365
Test: dEQP-VK.*

Change-Id: Id9019fc9d9239fc85d0d2b086d4efd468844d254
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/49108
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Yilong Li <liyl@google.com>
Tested-by: Yilong Li <liyl@google.com>

Previously each constant would cause new memory to be allocated. This
change makes us search through the existing constants to check if a
suitable one already exists.

Bug: b/178661423
Change-Id: I1315c80f019ff51e96245238100aa672d5e74930
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52229
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>

This enables generating multiple functions within the same scope, as
excercised by the Trampoline unit test.

Also adds a unit test for functions which are never finalized (useful
for when a code path is not supported and we want to fall back to a C++
implementation).

Also adds a unit test for the common pattern of deriving from Function<>
and using Reactor variables as members.

Bug: b/177999324
Change-Id: I269946e9c8a791eab84ff756c9c588ca8ce26337
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51988Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

LLVM has a known leak, which is addressed by this patch:
https://reviews.llvm.org/. With this patch, and making sure that
llvm::shutdown() is called, AppVerifier on Windows, and LSAN on Linux,
report no more load/unload leaks.

Bug: b/175782868
Bug: chromium:972686
Bug: angleproject:5377
Change-Id: If34a00c3a4b845604445f22bdb5889f51bb5a1d6
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52168
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

Naming 'anonymous' mmap memory, which we use for Reactor's executable
memory, enables profilers to attribute JIT-compiled routine execution
to SwiftShader.

Also suppress a GCC warning about not using the result of ftruncate().

Bug: b/174801963
Bug: b/178614275
Change-Id: Iac4938f332f0e75a9b1d995c83f0cbaaf3532651
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52228Tested-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Naming 'anonymous' mmap memory, which we use for Reactor's executable
memory, enables profilers to attribute JIT-compiled routine execution
to SwiftShader.

Also rename `swiftshader_reactor_private_config` to
`swiftshader_llvm_reactor_private_config` to clarify that it is only used
for building Reactor with the LLVM backend.

Bug: b/174801963
Change-Id: I4a6cb471b00f8403978870fd0be4608ee2798e54
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52208
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

This change restricts the effect of REACTOR_ANONYMOUS_MMAP_NAME to
platforms which define the __linux__ macro, since this feature makes
use of mmap(), syscall(), and posix_memalign().

Note Android is considered Linux.

Bug: b/174801963
Change-Id: Id194aca837b2c9c23f5c7419c926bc2872b77d1b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52209
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

This cl adds a new utility function to count
the number of components used by an input.

Bug: b/171415086
Change-Id: I348959ccd8945d2f336384cf7d905d3140231e4e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51736
Commit-Queue: Alexis Hétu <sugoi@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

The sampleRateShading feature adds new builtins variables which have
to be supported: BuiltInSampleId and BuiltInSamplePosition. This cl
sets the values for these builtins properly and turns on per sample
rendering where either of these is persent in the shader.

Bug: b/171415086
Change-Id: I959e80722493d12213be349019a0cc3a108a7596
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51735Tested-by: Alexis Hétu <sugoi@google.com>
Commit-Queue: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

When the shader contains interpolation instruction, interpolation
related data needs to be stored in the routine in order to be able
to perform the interpolation operations in the middle of running
the spirv shader.

Bug: b/171415086
Change-Id: I08fc6befdca1df92fa50326d9047926ac3417aa5
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51734Tested-by: Alexis Hétu <sugoi@google.com>
Commit-Queue: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

This cl introduces per sample shading in the fragment shader.
Rather than call the fragment shader multiple times per sample,
this cl adds a potential loop in the fragment shader where each
sample is processes in one of the loop's iteration.

- Each multisample related loop now processes either all samples,
  like before, or the current sample, if per sample shading is
  enabled
- A new per sample PixelProgram::maskAny() function was added
- emitEpilog() now has an option not to clear phis in order to be
  able to only clear them on the last sample
- The routine's fragCoord values are set per sample, with the
  proper sample offsets
- Similarly, the xxxx and yyyy values used for interpolation are
  now offset with the proper sample offsets when per sample
  shading is enabled

Bug: b/171415086
Change-Id: Ibd0c1bad23e2d81f7fa97240ebb50f88f1fee36e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51733Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Alexis Hétu <sugoi@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Naming 'anonymous' mmap memory, which we use for Reactor's executable
memory, enables profilers to attribute JIT-compiled routine execution
to SwiftShader.

Bug: b/174801963
Change-Id: I3b1558fb89d1cac42f65be0609cfa5ac2a8319ef
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52129Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Jason Macnak <natsu@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

* Move out VulkanBenchmark and DrawBenchmark classes to their own
  hpp/cpps.
* Move out ClearImageBenchmark and TriangleBenchmark classes and tests
  to their own cpps.
* Add non-template DrawBenchmark::addVertexBuffer that takes the size of
  the vertex type. This allows for the implementation to be moved to the
  cpp file, keeping the header light and easy to understand.

Bug: b/176981107
Change-Id: I657bd112c5523aa56dc5617082d6744b9edac03a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52130
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>

VkPipelineMultisampleStateCreateInfo contains parameters specific
to the sampleRateShading feature. This cl handles these new
parameters (without enabling the sampleRateShading feature).

Bug: b/171415086
Change-Id: I8e895483ab5d27643a2fc59347bb41a51570602d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51732Tested-by: Alexis Hétu <sugoi@google.com>
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Some of the interpolation logic is required to properly implement
the GLSLstd450Interpolate* operations. In order to prevent code
duplication and to make sure the logic is always identical between
the regular fragment input interpolation and the one performed by
the GLSLstd450Interpolate* operations, the interpolateCentroid()
function was moved to SpirvShader and renamed with a more generic
name, interpolateXY(), as it will be used to interpolate all modes
(centroid, sample and offset).

Bug: b/171415086
Change-Id: Ifbffa2b395a0b8c22fa120b36643db05f867a306
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51731Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

The sample qualifier forces the use of per sample rendering,
so this cl adds a mode for it and detects whether the sample
decoration is used in the shader.

Bug: b/171415086
Change-Id: Ic59eb45c41f019859eff01de4bf467284de53c0a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51730Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Bug: b/176981107
Test: build latest merge inside of Android repo
Change-Id: I3b3d87ac118157d103b33afd1a885116f659662e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52128
Presubmit-Ready: Jason Macnak <natsu@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Jason Macnak <natsu@google.com>
Commit-Queue: Jason Macnak <natsu@google.com>

* Renamed TriangleBenchmark to DrawBenchmark
* Added a set of hooks (virtual functions) that are invoked at certain
points during execution of DrawBenchmark::init and renderFrame. These
functions are prefixed with "do" to distinguish them.
* Added some resource management functions in DrawBenchmark, such as
addImage and addSampler for child classes to create these resources. We
want these owned by the base class so that resources can be properly
disposed of in the right order.
* Removed enum class FragShadeType, and replaced with three classes
derived from DrawBenchmark that implement the necessary hooks.
* DrawBenchmark tracks the number of vertices provided via
addVertexBuffer so that it can pass it to vk::CommandBuffer::draw().
Derived types can therefore provide as many (triangle) vertices as they
like from doCreateVertexBuffers.

Bug: b/176981107
Change-Id: I687f3d5ca09f7f93a3d6d7f68871a95bd083bf89
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52108Tested-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Generalized TriangleBenchmark to support rendering a triangle of a solid
color (hard-coded in shader), interpolated input colors, and
interpolated colors with texture sampling. Added benchmark captures for
each type, along with multi-sampled versions of each.

Bug: b/176981107
Change-Id: Iadef1f0f9a339038198f8ef0e492c4c2909a96b2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/51889Tested-by: Antonio Maiorano <amaiorano@google.com>
Commit-Queue: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

SwANGLE tests detected that SwiftShader's sample positions were 180
degree flipped from Vulkan's exact sample locations. Most multisampling
tests we've run with SwiftShader so far relied on resolved outputs,
so having the samples in the wrong order didn't matter, as long as we
had the correct 4 samples (which we do, since the samples are
essentially rotated by 90 degrees around the center of the pixel from
each other).

Bug: b/170962186
Bug: b/141380498
Change-Id: I65552863e1fe38b321b84995e038984488512e1c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52088Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Commit-Queue: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

The first devices with M1 'Apple Silicon' AArch64 processors were
launched with macOS 11.0, so it's safe to increase the target OS beyond
the 10.11 support level needed by Chrome.

This helps enable C++17 by addressing the lack of aligned heap
allocation support in previous macOS versions:

third_party/marl/include/marl/pool.h:236:27: error: aligned deallocation function of type 'void (void *, std::align_val_t) noexcept' is only available on macOS 10.14 or newer
src/Vulkan/VkPipeline.hpp:83:10: error: aligned deallocation function of type 'void (void *, std::align_val_t) noexcept' is only available on macOS 10.14 or newer

C++17 invokes aligned new/delete because of our use of alignas(16) for
types like sw::float4, which is used for member fields in
sw::DrawCall::BatchData and vk::GraphicsState. While this could be
addressed by using custom new/delete operators for these classes (and
classes which in turn use them as members), misalignment is just a minor
performance issue, so we can leave things as-is and work around it with
a build flag (note 16-byte alignment is needed for some vectors accessed
in Reactor code, but that's handled for those individual cases).

Bug: b/174843857
Change-Id: Iaa40a3ac6aa7d12c088bf93a55803d8146bd289c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/52068Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Commit-Queue: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

* changes:
  Update src/Pipeline/SpirvShaderInstructions.inl
  Update SPIR-V Tools to f37547c73
  Squashed 'third_party/SPIRV-Tools/' changes from b0e22d28f5e..f37547c73a9
  Update SPIR-V Headers to 8bb2420b4
  Squashed 'third_party/SPIRV-Headers/' changes from 104ecc356c1..8bb2420b44a
  Mend git merge history for SPIRV-Headers and SPIRV-Tools
  Mend git merge history for SPIRV-Headers and SPIRV-Tools
  Revert 9dff6a3b to mend git merge history for SPIRV-Headers and SPIRV-Tools
  Squashed 'third_party/SPIRV-Tools/' changes from a61d07a7276..b0e22d28f5e
  Squashed 'third_party/SPIRV-Headers/' changes from 7845730cab6..104ecc356c1

for SPIRV-Tools update

Bug: b/123642959
Change-Id: I6c21b03ddb5f04e20fb72c0e6b280ca9caaf1f7b

Changes:
    f37547c73a9 spirv-val: Add Vulkan EXT builtins (#4115)
    1bd539b9bfc Support pending Intel extensions (#4116)
    56f8ed48efd Validate Sampled=1 for Vulkan ImageQuerySizeLod, ImageQueryLevels, ImageQueryLod (#4103)
    e25db023c47 spirv-val: Add Vulkan Memory Scope VUs (#4106)
    8383bd5d6f2 Migrate all Kokoro build scripts over to use the docker VM image (#4114)
    ee39b5db5f1 spirv-val: Add Vulkan Addressing Model check (#4107)
    9150cd441f4 Remove WebGPU support (#4108)
    b2cfc5d1cec spirv-val: Vulkan atomic storage class (#4079)
    cec658c1160 Avoid integrity check failures caused by propagating line instructions (#4096)
    b1507d0d2ba Linker usability improvements (#4084)
    7bbe1a31641 Revert "Generate differentiated error codes for buffer oob checking (#4097)" (#4100)
    c32277c0ba0 Generate differentiated error codes for buffer oob checking (#4097)
    f3ccb633dfd use std::string::empty() to test for emptiness (#4098)
    ad77ed7a8d5 spirv-val: Label standalone Vulkan VUID (#4091)
    aa005e8bd43 spirv-val: Add Vulkan decroation VUID (#4090)
    4ed1f4fce9e Fix binding number calculation in desc sroa (#4095)
    af3a3d481f6 Build deps: dump ini from 1.3.5 to 1.3.7 in tools/sva (#4092)
    94d1a80159e spirv-val: Add Vulkan FP Mode VUID (#4088)
    6d05ed8410b spirv-val: Fix Vulkan image sampled check (#4085)
    37c03859de9 spirv-val: Add Vulkan ForwardPointer VUID (#4089)
    d630e5f8c10 spirv-val: Add Vulkan ImageTexelPointer format check (#4087)
    1bb80d2778a spirv-val: Add Vulkan Group Operation VUID (#4086)
    17ffa89097b spirv-val: Add first StandAlone VUID 04633 (#4077)
    8f4b35c332e spirv-val: Add Subgroup VUIDs (#4074)
    4e31fdd4aa1 spirv-fuzz: Fix OpPhi handling in DuplicateRegionWithSelection (#4065)
    ad898cb9498 validation: validate return type of OpImageRead (#4072)
    305caff2ebb validation: tighter validation of multisampled images (#4059)
    a0370efd589 validate OpTypeImage Sampled values for environemnts (#4064)
    09d1fea5890 Force using Python 3 git-sync-deps (#4067)
    7170218b8de validate StorageImageMultisampled capability (#4062)
    3b85234542c spirv-val: Add last TessLevelOuter and TessLevelInner VUID (#4055)
    171290703a6 spirv-val: Add last ClipDistance and CullDistance VUID (#4054)
    1a6b4053fa0 spirv-val: Add last ViewportIndex and Layer VUID (#4053)
    9df5225e672 spirv-val: Add last Position VUID (#4052)
    862d44a86ed Allow forward pointer to be used in types generally (#4044)
    bda102d7a76 opt: Run DCE when SPV_KHR_shader_clock is used (#4049)
    b27b1afd12d Update CHANGES to include latest ray tacing fixes.
    cd050786621 Take new (raytracing) termination instructions into account. (#4050)
    21f7c5675da Start SPIRV-Tools v2020.7
    4c2f34a5048 Finalize SPIRV-Tools v2020.6
    c9c1f54330d Update CHANGES
    10e0ae7946c Do run DCE if SPV_KHR_ray_query is used. (#4047)
    84546647a0d Update CHANGES
    c1d5a045f6f Change ref_analysis to RefAnalysis to follow coding standards. (#4045)
    c502a15f25c Handle 8-bit index in elim dead member (#4043)
    32573bb2168 Add validation support for the ray tracing built-in variables (#4041)
    7046c05d2fb Add texel buffer out-of-bounds checking instrumentation (#4038)
    cf2d1e7afcd Update spirv-header deps (#4040)
    1299436c8f1 Reject SPIR-V that applies void to OpUndef, OpCopyObject, OpPhi (#4036)
    2c458414c08 BuildModule: optionally avoid adding new OpLine instructions (#4033)
    a79aa038ec1 Remove prototype for unimplemented method (#4031)

Commands:
    ./third_party/update-spirvtools.sh

Bug: b/123642959
Change-Id: Ic6136d02f11f153a36ab9b9bb8019b4b18b3b5f3

f37547c73a9 spirv-val: Add Vulkan EXT builtins (#4115)
1bd539b9bfc Support pending Intel extensions (#4116)
56f8ed48efd Validate Sampled=1 for Vulkan ImageQuerySizeLod, ImageQueryLevels, ImageQueryLod (#4103)
e25db023c47 spirv-val: Add Vulkan Memory Scope VUs (#4106)
8383bd5d6f2 Migrate all Kokoro build scripts over to use the docker VM image (#4114)
ee39b5db5f1 spirv-val: Add Vulkan Addressing Model check (#4107)
9150cd441f4 Remove WebGPU support (#4108)
b2cfc5d1cec spirv-val: Vulkan atomic storage class (#4079)
cec658c1160 Avoid integrity check failures caused by propagating line instructions (#4096)
b1507d0d2ba Linker usability improvements (#4084)
7bbe1a31641 Revert "Generate differentiated error codes for buffer oob checking (#4097)" (#4100)
c32277c0ba0 Generate differentiated error codes for buffer oob checking (#4097)
f3ccb633dfd use std::string::empty() to test for emptiness (#4098)
ad77ed7a8d5 spirv-val: Label standalone Vulkan VUID (#4091)
aa005e8bd43 spirv-val: Add Vulkan decroation VUID (#4090)
4ed1f4fce9e Fix binding number calculation in desc sroa (#4095)
af3a3d481f6 Build deps: dump ini from 1.3.5 to 1.3.7 in tools/sva (#4092)
94d1a80159e spirv-val: Add Vulkan FP Mode VUID (#4088)
6d05ed8410b spirv-val: Fix Vulkan image sampled check (#4085)
37c03859de9 spirv-val: Add Vulkan ForwardPointer VUID (#4089)
d630e5f8c10 spirv-val: Add Vulkan ImageTexelPointer format check (#4087)
1bb80d2778a spirv-val: Add Vulkan Group Operation VUID (#4086)
17ffa89097b spirv-val: Add first StandAlone VUID 04633 (#4077)
8f4b35c332e spirv-val: Add Subgroup VUIDs (#4074)
4e31fdd4aa1 spirv-fuzz: Fix OpPhi handling in DuplicateRegionWithSelection (#4065)
ad898cb9498 validation: validate return type of OpImageRead (#4072)
305caff2ebb validation: tighter validation of multisampled images (#4059)
a0370efd589 validate OpTypeImage Sampled values for environemnts (#4064)
09d1fea5890 Force using Python 3 git-sync-deps (#4067)
7170218b8de validate StorageImageMultisampled capability (#4062)
3b85234542c spirv-val: Add last TessLevelOuter and TessLevelInner VUID (#4055)
171290703a6 spirv-val: Add last ClipDistance and CullDistance VUID (#4054)
1a6b4053fa0 spirv-val: Add last ViewportIndex and Layer VUID (#4053)
9df5225e672 spirv-val: Add last Position VUID (#4052)
862d44a86ed Allow forward pointer to be used in types generally (#4044)
bda102d7a76 opt: Run DCE when SPV_KHR_shader_clock is used (#4049)
b27b1afd12d Update CHANGES to include latest ray tacing fixes.
cd050786621 Take new (raytracing) termination instructions into account. (#4050)
21f7c5675da Start SPIRV-Tools v2020.7
4c2f34a5048 Finalize SPIRV-Tools v2020.6
c9c1f54330d Update CHANGES
10e0ae7946c Do run DCE if SPV_KHR_ray_query is used. (#4047)
84546647a0d Update CHANGES
c1d5a045f6f Change ref_analysis to RefAnalysis to follow coding standards. (#4045)
c502a15f25c Handle 8-bit index in elim dead member (#4043)
32573bb2168 Add validation support for the ray tracing built-in variables (#4041)
7046c05d2fb Add texel buffer out-of-bounds checking instrumentation (#4038)
cf2d1e7afcd Update spirv-header deps (#4040)
1299436c8f1 Reject SPIR-V that applies void to OpUndef, OpCopyObject, OpPhi (#4036)
2c458414c08 BuildModule: optionally avoid adding new OpLine instructions (#4033)
a79aa038ec1 Remove prototype for unimplemented method (#4031)

git-subtree-dir: third_party/SPIRV-Tools
git-subtree-split: f37547c73a981bcf3f8b1a471e3000e606dd41a8

Changes:
    8bb2420b44a Merge pull request #176 from MrSidims/private/MrSidims/OtherExtensions
    0dd29fb93ac Apply suggestions to Intel extensions PR
    b690b3e2174 Update generated files
    f9797ef9c1b Add SPV_INTEL_long_constant_composite extension
    8987518da9e Add SPV_INTEL_loop_fuse extension
    3fad4b35436 Add SPV_INTEL_fpga_cluster_attributes and SPV_INTEL_fp_fast_math_mode
    b6be746fc77 Update SPV_INTEL_fpga_loop_controls extension
    ad00ca3b55b Update SPV_INTEL_kernel_attributes extension
    0c290562414 Update SPV_INTEL_function_pointers extension
    ea791133ee4 Upstream SPV_INTEL_float_controls2 extension
    24e9c63e6dd Upstream SPV_INTEL_vector_compute extension
    2b76471587f Upstream SPV_INTEL_fpga_memory_accesses extension
    0e756817218 Upstream SPV_INTEL_io_pipes extension
    ef87dd18967 Upstream SPV_INTEL_variable_length_array extension
    e202eb89141 Upstream SPV_INTEL_usm_storage_classes extension
    a001134e417 Upstream SPV_INTEL_arbitrary_precision_integers extensions
    7669e05273b Upstream SPV_INTEL_inline_assembly extension
    d5035de3d31 Upstream SPV_INTEL_fpga_buffer_location extension
    f027d53ded7 Merge pull request #184 from dgkoch/dkoch_remove_hittkhr
    bdd2aa34c45 remove HitTKHR

Commands:
    ./third_party/update-spirvheaders.sh

Bug: b/123642959
Change-Id: I21d11818df5a17704b930408b509f1ca3cfdefbc

8bb2420b44a Merge pull request #176 from MrSidims/private/MrSidims/OtherExtensions
0dd29fb93ac Apply suggestions to Intel extensions PR
b690b3e2174 Update generated files
f9797ef9c1b Add SPV_INTEL_long_constant_composite extension
8987518da9e Add SPV_INTEL_loop_fuse extension
3fad4b35436 Add SPV_INTEL_fpga_cluster_attributes and SPV_INTEL_fp_fast_math_mode
b6be746fc77 Update SPV_INTEL_fpga_loop_controls extension
ad00ca3b55b Update SPV_INTEL_kernel_attributes extension
0c290562414 Update SPV_INTEL_function_pointers extension
ea791133ee4 Upstream SPV_INTEL_float_controls2 extension
24e9c63e6dd Upstream SPV_INTEL_vector_compute extension
2b76471587f Upstream SPV_INTEL_fpga_memory_accesses extension
0e756817218 Upstream SPV_INTEL_io_pipes extension
ef87dd18967 Upstream SPV_INTEL_variable_length_array extension
e202eb89141 Upstream SPV_INTEL_usm_storage_classes extension
a001134e417 Upstream SPV_INTEL_arbitrary_precision_integers extensions
7669e05273b Upstream SPV_INTEL_inline_assembly extension
d5035de3d31 Upstream SPV_INTEL_fpga_buffer_location extension
f027d53ded7 Merge pull request #184 from dgkoch/dkoch_remove_hittkhr
bdd2aa34c45 remove HitTKHR

git-subtree-dir: third_party/SPIRV-Headers
git-subtree-split: 8bb2420b44af3c514480e63ac24439044706b583