Bug: b/168821209
Change-Id: I1892581045ec37abc4060cf4de2c45cb2fb6208a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48568
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Nicolas Caramelli <caramelli.devel@gmail.com>

Bug: b/162574690
Change-Id: I825af0494737ec9090da9580dfd876db88e3c289
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/47368
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Caramelli <caramelli.devel@gmail.com>

Under normal operations, a framebuffer is created with each VkImageView
it will need to use. This extension allows users to specify a flag that
tells the driver to wait until render pass begin time. The user must
also request this feature during device creation, this behavior is not
enabled by default.

Bug: b/167223759
Tests: dEQP-VK.imageless_framebuffer.*
Change-Id: I2be5cec4d8174e20372eb9e148a05034dd338341
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48488Tested-by: Sean Risser <srisser@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

This is a major reworking of the way the debugger is implemented.

Instead of generating shader code that is continually driving `vk::dbg` state, the debugger now:
* Maintains a full shadow copy of all Intermediate values (SSA)
* Only calls out to C++ whenever a trap is set
* Only constructs and updates the `vk::dbg` state when a trap is hit

The main goal of this reworking is to properly support OpenCL.Debug.100's DebugValue instructions that may use deref expressions
(See https://www.khronos.org/registry/spir-v/specs/unified1/OpenCL.DebugInfo.100.html#_debug_operations_a_id_operation_a)
These are remarkably hard to implement without having everything backed with real memory.

The resulting reimplementation is:
* Much, much faster (>1000x faster for sample apps I've tried). Only ~2x slower than non-debug enabled SwiftShader now.
* Uses much less runtime memory.
* Much cleaner, removing a load of template magic for setting variable state that was extremely hard to follow
* More thoroughly commented, explaining each component and method.

Bug: b/145351270
Change-Id: I1b48e9defae095109ebfaf90dfb7c7e22d335816
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48729Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Implement `operator<()`, `std::hash<>`

Bug: b/145351270
Change-Id: I9ebcfccd3a1da36dc9286f6ed6d50c73c8c3bac3
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48698
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Non-power of two vector types were previously not supported as they would be misaligned.
Use a `constexpr` function to ensure the `alignas` is rounded to the next POT.

While the memory layout for vec3 will be identical to a vec4, this allows us to specialize templates to distinguish between `vec3` and `vec4`.
This is used by the debugger to display `vec3` locals without showing an undefined `w` component.

Bug: b/145351270
Change-Id: I61f61e42e69d367b1a625c262be970994b287433
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48708Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

All 3 issues listed have now been fixed.
It is safe to reenable this.

Fixes: b/147726513
Change-Id: Ie76b77cc67f0ab74f45c73da1a998bce5dffd4fa
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48728Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Useful for lazily building a new `File` if it hasn't been registered already.

Also assign the file to the location of a new frame. This is a sensible default to have.

Bug: b/145351270
Change-Id: I7c0abff22a0010923428ff5ed0760d6ae63b0c6b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48697
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Change `windowSpacePosition` from c-array to `std::array`. Other invocation state fields use `std::array`, and this simplifies some template logic in the new debugger implementation.

Add / fix comments in `SpirvShaderControlFlow.cpp`.
Also add a missing call to `dbgEndEmitInstruction()`. This is typically invoked by `SpirvShader::EmitInstruction()`, but the special handling of `OpPhi` in `EmitLoop()` previously skipped this.

Bug: b/145351270
Change-Id: Id6463d5e157b28ef61ce8fc48644bcbc97a2aa55
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48696Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Remove the `Context::Lock` parameter from `enter()`. We don't want to force the lock to be held for the entire duration of the call.

Add `UpdateFrame` type alias.

Add the `UpdateFrame` parameter to `enter()` - you usually want to update the frame as soon as you've pushed the stack.

Switch `pauseAtFrame` from a `shared_ptr` to a `weak_ptr`.
If the frame is released before the thread finds it to pause, then the thread will instead pause at the next frame (following the logic of a null `pauseAtFrame`).

Fixed the assignment of `pauseAtFrame` in `stepOut()` - it was looking at the current frame, not the next one up.

Bug: b/145351270
Change-Id: I1cb85af58c666a7793145480abc95a4ec82e1859
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48695Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

A `SetFunctionBreakpointsRequest` contains the full list of breakpoints. Clear the list in the context before adding the new list.

Also add a `getFunctionBreakpoints()` accessor so you can enumerate the list of function breakpoints.

Bug: b/145351270
Change-Id: Ia7fad403eace7a95a15671e2ce90b8e083da3ce7
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48694Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

`Value` had a number of methods that were never used (like `set()`), so just remove them.
Remove `Type`, this was also unused, and was unncessarily complex to maintain.

Add `Variables` interface that allows other composite value implementations that are not `VariableContainer`.

Break the inheritance of `VariableContainer` from `Value`, as this forces pointer casting and doesn't work with the `Variables` interface.

Add `Struct` which is an implementation of `Value` that implements the new `children()` method to return the provided `Variables`.

Sets the groundwork for the overhauled debugger implementation.

Note: The changes to `SpirvShaderDebugger.cpp` are a least-effort set of changes to make things compile.
A significant amount of this file (along with these changes) will be reimplemented in a followup change.

Bug: b/145351270
Change-Id: Ic3a641a246737b1f82c786fa8c1ec75700b2a71a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48693
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

TSAN will catch this, but when its disabled, everything simply freezes and you have to go hunting across all the threads in the debugger.
Just explode immediately with a sensible message.

Bug: b/145351270
Change-Id: Iecef77a47bd54daabb489c5ca70e44b68c0c1e1f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48692
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Simple accessor for the breakpoint set.

Bug: b/145351270
Change-Id: Ifb534482009b5842d8bd2a796c082625b770e298
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48691
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Into `ServerEventListener` and `ClientEventListener`.

The new debugger implementation wants to listen to breakpoints being set so that traps can be dynamically enabled.

Bug: b/145351270
Change-Id: Iafe64426874da3aa109d4b59b34cb1e36bd06828
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48690
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

`limit()` is the same for all lanes, so use `rr::Int`, not `SIMD::Int`.
This was using an unqualified scope, so likely just a screwup when I first wrote this.

Bug: None. A simple fix.
Change-Id: I0df67fb88f9973dd522a5555888173ba1f5ae594
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48689Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This was uninitialized, showing garbage in the debugger

Bug: b/145351270
Change-Id: I952faabed4b1cf8b61dc4b9a5deb8c7a4e59db90
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48688Reviewed-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

set_sources_assignment_filter() is considered as a mis-feature
of gn (as it makes it difficult to reason about the BUILD.gn
files individually).

Change BUILD.gn in SwiftShader to check the platform before
adding plaform-specific source files to "sources" variable.
This will allow removing the global sources assignment filter
from Chromium codebase.

Bug: chromium:1018739
Change-Id: Icf357de347b09dccce8c031d84ae7d8ce0ec396b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48668Tested-by: Corentin Wallez <cwallez@google.com>
Reviewed-by: Corentin Wallez <cwallez@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Trying to build trout_arm64 on git_master results in linker errors
due to certain library symbols not supporting branch and link relocation

Apply the fix used elsewhere in the codebase (e.g. ag/12429028) to
overcome this issue

Bug: b/155835175
Test: build trout_arm64-userdebug
Change-Id: Ib83fae561f50dbe3080bd3018690844efebe67df
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48648
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Added a case in vkCreateImage to explicitly allow for this extension.
Nothing else needs to be done, because Swiftshader doesn't need to do
anything with the format list.

Also expose the extension string in the device extensions.

Bug: b/167557828
Change-Id: I8af6a63068cf1bea28df38738e258a918f0e5582
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48528Tested-by: Sean Risser <srisser@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Currently is_linux GN variable is set to true on Chrome OS build,
but is is planned to set false. This is the preparation to
keep the compatibility.

Bug: chromium:1110266
Test: Built locally.
Change-Id: I87e8f1312c85f85a3f882698f1dce3d65e0cfab6
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48348
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Hidehiko Abe <hidehiko@google.com>

Bug: b/146515640
Bug: b/161909468
Test: run Cuttlefish w/ SwiftShader w/ Gralloc4
Change-Id: I60c806f460ef4fd0b4de5e3993731c1054030159
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48448
Presubmit-Ready: Jason Macnak <natsu@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Jason Macnak <natsu@google.com>

Reactor backend: Subzero

Change-Id: Idbdc8cad8a31d60e02db59e8dd1d2bf325ee5f21
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/47748
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>

Rather than set the FOLDER property from within certain third-party
CMakeLists.txt, do so from the root, and make them all go to a
"third_party" folder.

Bug: b/145758253
Change-Id: I7b9f69a8f16be8ef6e0d153fc28ac4f6c3d64e5e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48368
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Antonio Maiorano <amaiorano@google.com>

Prints each debug instruction as it is defined. Yet another tool in debugging the debugger.

Also rename `PRINT_EACH_PROCESSED_INSTRUCTION` to `PRINT_EACH_EMITTED_INSTRUCTION` as that's really what it meant.

Bug: b/148401179
Change-Id: I9ddf90eab1919deefb5f4557756364478f06375a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48430Tested-by: Ben Clayton <bclayton@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

This debug instruction partially updates a local variable with SSA values.

This is typically used to update a `DebugLocalVariable` of a composite type, which holds structure member offsets from a base address.
To handle these, we allocate shadow memory to hold a copy of the entire variable in contiguous memory and have the `DebugLocalVariable` point to this memory. Whenever we encounter a `DebugValue`, we copy the necessary fields to the
shadow memory.

This approach is simplier and more performant than attempting to fiddle with `vk::dbg::Value`s for each `DebugValue`.

Bug: b/148401179
Change-Id: I6aafce1f7553f4a3a97bbf84e6e9dac97ce71a68
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48429Reviewed-by: Jaebaek Seo <jaebaek@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Previously the way to create `VariableContainer`s was to call
`createVariableContainer()` on the `Context::Lock`. This was required so that debug-client requests to inspect the container's contents can be done with a simple map lookup.
Creating variable containers is a high frequency operation, and obtaining a `Context::Lock` (unsurprisingly) requires acquiring a mutex lock, which across multiple threads results in significant mutex lock contention.

To dramatically help with debugger performance, we can now construct `VariableContainer`s outside of the lock. This is achieved by registering the variable container in the context map only when it is handed to the client. This only occurs when there's a DAP transaction, which is orders of magnitude less frequent.

Bug: b/148401179
Change-Id: I58eca33aac8fd4711dd9756072a6a06efdacc671
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48428
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Jaebaek Seo <jaebaek@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

In GLSL ES 3.00, float literals are allowed to have an 'f' or 'F'
suffix, but they don't change the parsing process since they're always
interpreted as single-precision IEEE-754 floating-point values.

The std::istringstream used for parsing the literal's string can have
implementation dependent behavior depending on the presence of the
suffix, so this change removes it. This matches what ANGLE does.

Fixes: b/168250854
Change-Id: I203833a2e817ca698894db5669941fb435b97868
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48388
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Based on OpenCL.DebugInfo.100 spec, if the function has no return
value, 'Return Type' operand of DebugTypeFunction is OpTypeVoid.

Bug: b/148401179
Change-Id: Ib29c7a569ff2ec24ba4667523766b0bdb0d6f260
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48128
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Jaebaek Seo <jaebaek@google.com>

A VulkanGlass synchronization fix now allows us to run it in debug
without asserting on a validation layer error, so updating to ToT.

Bug: b/141361933
Change-Id: I8b2ebff6b5d822370d7b62afdf6845d1736c1a1b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48289
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Constant arrays are allocated in the same 'register file' as the one
for per-program uniforms, to support dynamic indexing. While we had
a special case for that in OutputASM::registerType(), in registerIndex()
we were actually allocating them in the temporaries register file. This
led to assigning the same register index range twice, leading to
overwriting the per-program uniforms with constant array data in the
shader.

Bug: b/168060171
Change-Id: If42944d186ed1c1328a8d477edca4662aa1e8e54
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48248
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Unlike image views, buffer views don't have a component mapping field,
so we were setting the mapping to identity swizzles for uniform texel
buffers. However, we also use the component mapping for assigning 0 or 1
to components not provided by the texel format:
https://swiftshader-review.googlesource.com/c/SwiftShader/+/31388

This change ensures we still get the 0 and 1 in the desired channels for
uniform texel buffers.

Bug: b/168052622
Test: dEQP-VK.texture.texel_buffer.uniform.packed.b10g11r11-ufloat-pack32
Change-Id: I40885f998a53e371e1e2358ed197fc3446916d8c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48268Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Added missing case for a single sample with alphaToCoverage enabled

Change-Id: Iad9234fdd30d64c60ae189c123564a0423ab12df
Tests: dEQP-VK.rasterization.frag_side_effects.color_at_*
Bug: b/168046415
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48209
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Commit
"Merge vk-gl-cts/vulkan-cts-1.2.4 into vk-gl-cts/master"
Revision
https://chromium.googlesource.com/external/github.com/KhronosGroup/VK-GL-CTS/+/5e85f689b91102b012ab4c4278f46efeb6214381
Branch
https://chromium.googlesource.com/external/github.com/KhronosGroup/VK-GL-CTS/+/refs/heads/master

Includes fix for flaky failure of dEQP-VK.wsi.*.swapchain.acquire.too_man*:
* Fix creation of signaled fences in swapchain tests https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2505

New notable test coverage:
* Add coverage for negative float to clear color values https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2360
* Execute the graphics/compute pipeline https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2325
* Add tests for shader reuse with differing layout https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2324
* Enable random writes for dynamic storage buffers https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/1973
* Check per-sample copy results in multisample copy tests https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2448
* 1D and 3D image copy and blitting tests https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2391
* Support for VkPhysicalDeviceVulkan1{1,2}Properties in framework https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2425
* Test 3D image blit with differing depth https://gitlab.khronos.org/Tracker/vk-gl-cts/-/issues/2263

Bug: b/167692239
Fixes: b/162249856
Change-Id: I9d8933999e29a2702ce0990c7f426f6350c9ed3d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48149Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

https://gitlab.khronos.org/vulkan/vulkan/-/merge_requests/3399 clarified
that when identity component swizzle is expected, one can also use the
corresponding component enum.

Fixes: b/167706626
Change-Id: I2c94f6ac108be5269d56f397ecf3da58cd310e5e
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48168Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Sean Risser <srisser@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>

Bug: b/162351826
Change-Id: Idd8e6c869ffa99945e7a6bc7701a7ab854f18432
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/47188
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

DescriptorPool::computeTotalFreeSize() was adding the size
of the last element to its available memory size, rather
than subtracting it, which was causing the DescriptorPool
to return the wrong error (VK_ERROR_FRAGMENTED_POOL
instead of VK_ERROR_OUT_OF_POOL_MEMORY).

Bug: b/167617619
Change-Id: Ie60c2fe521b0d233b32504b446af1715361f5c21
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/47549Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

This cl fixes GCC warnings caused by using clang specific directives.
Also ran git cl format on modified files.

Bug: chromium:1122889
Change-Id: Ibdaa84ab8162228ae40934fc9be14379ba89ca59
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48148
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

The SPIR-V spec states that the Sample operand of OpImageTexelPointer
"must be a valid <id> for the value 0 if the OpTypeImage has MS of 0."

This change optimizes for this common case by no performing texel
address sample offset and bounds check calculations.

Bug: b/163142358
Change-Id: I75932c264d45df1012fa8451fc4c57671191b59b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/47808
Presubmit-Ready: Nicolas Capens <nicolascapens@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Reviewed-by: Antonio Maiorano <amaiorano@google.com>

Due to a defect in older versions of libstdc++, the std library could
not support hashing of enums. So each enum used in a hash is now
explicitly used as a uint32_t.

Bug: b/166302706
Change-Id: Id7de96f97bef0e6f104dc783115a416e3aa47b7c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/48108
Kokoro-Result: kokoro <noreply+kokoro@google.com>
Tested-by: Sean Risser <srisser@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>