I need this in order to support depth formats that arent D32F.

Bug: b/128363727
Change-Id: I7016351207e433e668053ebaf4dbeb9f2f7936f2
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26888Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>

- Many cases where sampleCounts is required to be VK_SAMPLE_COUNT_1_BIT by spec
- Missing feature flags for ETC2 formats
- Max array layers must be 1 for 3D

See 38.1.1 Supported Sample Counts for details of sample count
requirements

Bug: b/128344974
Test: dEQP-VK.api.info.image_format_properties*
Change-Id: I3ffe747a8cc8d06e0d33aa0937bd2b7ecd045ff5
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26871Tested-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>

In order to eventually remove sw::Surface, some utility
functions used by vk::Image were copied to vk::Format.

Bug b/126883332

Change-Id: Ie8404b70adc3336d536dcd5c0ec26b63e46c5174
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26872Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

We weren't looking at the most recent change.
Kinda important when that's typically the new test list you're trying to export.

Change-Id: I0e312dbde0ae54feb0641a658eebc73bbdb47eaa
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26928Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: b/126871859
Change-Id: I0b3db6c033419a2ad54453d470960330d4f337cc
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/25909
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>

This is by no means finished:
* ComputeProgram holds the codegen and the invoke logic. ComputeProgram::run() should probably be moved somewhere else (maybe Renderer?).
* ComputeProgram::run() is currently fully single threaded.
* The compute routines are currently not cached.

With that said, this is enough to start passing a whole bunch of dEQP compute tests.

Test: *.compute.*
Bug: b/126871859
Change-Id: Ic6a76826b2fec0d34d54e0bed564b360ea0610c0
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26551
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>

Change-Id: I6e695dd5221da6d9f7f198bc2af8aee00eda8881
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26808
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>

Change-Id: I15fb0452789a7a3e3f84453a1b2d9a1b969d510f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26908Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

The sw::Surface object was the intermediate representation of an
image used between the vk::Image and the Blitter. This cl removes
the need for an intermediate representation by having the Blitter
use the vk::Image directly.

There should be no regression for the clear and blit tests.

Bug b/126883332

Change-Id: Icbc15470e3ad112ed78f4f62d6d82c66e3e37a20
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/25928Tested-by: Alexis Hétu <sugoi@google.com>
Presubmit-Ready: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

The idea here is that we can build a number of fine granularity SPIR-V tests, which we can also verify against any optional system Vulkan hardware.

This CL adds support for building and running the Vulkan unit tests on mac and linux.

Bug: b/123749916
Change-Id: I01a4e2a914ffd5473d271013957fa3dddbc43645
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/24328
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Bug b/123360006

Change-Id: I4132ebf474f8055554126529839c33fd87d75499
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26768Tested-by: Nicolas Capens <nicolascapens@google.com>
Presubmit-Ready: Ben Clayton <bclayton@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

- Add clarifying comments.
- Add self-explanatory helper functions.
- Fix typeSize to handle all Reactor types.
- Use uintptr_t base type for emulated type enum.
  Prevents 0x########00000000 from being interpreted as emulated type.

Bug b/126028338

Change-Id: Ib38cf9b59f58c0f6046f9c84a6152849791bb4ed
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26648Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>

Test: *smod*
Bug: b/126873455
Change-Id: If946e72d23aa9386b85cffe96b02ab63195f0a24
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26549Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Exports the latest test data to a Google Sheets document.

Change-Id: Ia1b38464daf7117da571d536e7ff029023b9de58
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26748Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This causes Visual Studio to default to the x64 configuration, which is
64-bit instead of 32-bit.

Change-Id: Ifb1246216c726e6ad0dcc7891f2f3a8f89aeff7c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26649Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Nicolas Capens <nicolascapens@google.com>

Fixes tests that turned from PASS -> CRASH by adding compute shader support.

Bug: b/120799499
Change-Id: I72190eef98223851cba3d61a98d54ab2f6e2a0a7
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26553Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

Change-Id: If451dc09203631737739b30d385fbf8a215b2b46
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26728Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

The test list from the latest patchset is now run on the parent.
Missing tests are a bug.

Change-Id: I149e235968acad4988d0fcff78d8a8796aaf118c
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26559Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Change-Id: Ib0beec96a6efafa0ee6fb8a43e15142206d3cd4d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26558Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

This is better than running potentially two different test lists for both.

Requires caching the parent test results with the testlist's hash as a cache key.

Change-Id: Ic153e65395fc1c3f02a66bbb1df60731ce28901b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26557Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Bug: b/123933266
Change-Id: I4585c881ea7328b03b7c98bd5b4c255a399a0a70
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26532Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Change-Id: I6cd54eefee8e4fe0a20799bf6bce85ddd53df9d5
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26708Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

This gets us the final few cases in the glsl multiplies group.

Test: dEQP-VK.glsl.operator.binary_operator.mul.*
Bug: b/126873455
Change-Id: Id9ec4ad1754a7e63c8d7e0aa0a0b0b156a03b024
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26651
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

- FMin
- FMax
- SMin
- SMax
- UMin
- UMax
- Step
- SmoothStep
- FMix
- FClamp
- SClamp
- UClamp
- FSign
- SSign

Test: dEQP-VK.glsl.operator.common_functions.*
Bug: b/127804400
Change-Id: Ie136d8f0386f6a1db9acbf4e261851ce696e2d95
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26650Tested-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>

Mostly trivial; exception is RoundEven, for which I borrowed the
implementation from the GLES side.

Test: dEQP-VK.glsl.operator.common_functions.*
Bug: b/127804400
Change-Id: I889c3750ebdbc078723890488cf44d89ea77103b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26628Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Test: dEQP-VK.glsl.operator.bool_compare.*
Test: dEQP-VK.glsl.operator.binary_operator.equal.*
Test: dEQP-VK.glsl.operator.binary_operator.not_equal.*
Bug: b/126870789
Change-Id: I9d33444d67792274d8027b7d1632235533cfc079
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26574Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Bug: b/127804400
Test: dEQP-VK.glsl.operator.common_functions.floor.*
Change-Id: Iaefeef8f006673e55d4f85184d7782ac0aafb314
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26572
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>

Bug swiftshader:125

Change-Id: If24564a16d880a11e782e5f6b4f035e612255076
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26573Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

The blitter has a cache which can allow us to reuse already used
routines, so it should be kept in a central location. For that
reason, the blitter ownership was moved from the Image object to
the Device object.

Bug b/117974925

Change-Id: I825853c381dcbc04701b1d9e7dfa376108294221
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/25728Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Fixes dEQP-VK.glsl.operator.geometric.cross.*

Bug: b/127804400
Change-Id: I62337ba547048426c76adb1e90a1ebc4079e2a3d
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26571Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Basic implementation of descriptor update template.
Only supports VK_DESCRIPTOR_UPDATE_TEMPLATE_TYPE_DESCRIPTOR_SET template type.

Bug b/123244275

Change-Id: Iaf7c1e52bee6d2683a2f34bd2f780396aa953442
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26568Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

There is a subtle trap here -- almost all SPIRV instructions require vector
widths to match, but for OpVectorShuffle, the widths of the result
vector and the two sources are completely independent.

Fixes dEQP-VK.glsl.operator.unary_operator.plus.highp_vec2_vertex and many others.

Bug: b/127959969
Change-Id: Iaa2cc09fb510fa2ab07a6f53599af8dade553a9a
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26570
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Change-Id: I0fba3c2969cd61c8e7b6e30628473944f2e5a986
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26555Tested-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Fixes many tests in dEQP-VK.glsl.*

Bug: b/126873455
Change-Id: If949066bbe1d333755d1d83ab175bc39dcc91d91
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26608Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

A step toward being able to run many of dEQP-VK.glsl.*

Bug: b/126870789
Change-Id: Id53d5a511d7f0fa10994c46254529027ad773542
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26569Tested-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>

These decorations are used for all explicit layout storage classes --
Uniform, StorageBuffer, PushConstant.

Bug: b/120799499
Change-Id: Ida7ee6e2a5145d09792acf39b6e33d039431578b
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26489Reviewed-by: Ben Clayton <bclayton@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Change-Id: Ie4e2c11c495f17bf628b3c828ce81596e3d73e5f
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26588Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Level count was off by one for the size.
Also change vk::MAX_IMAGE_ARRAY_LAYERS to be the actual value rather
than log2 -- this was a bug waiting to happen somewhere.

Bug: b/118429780
Change-Id: I783a11571f90c34302a4ead8476009d309880eb6
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26508Tested-by: Chris Forbes <chrisforbes@google.com>
Presubmit-Ready: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

The lexer doesn't actually keep the string for literals.

Bug chromium:939239

Change-Id: Ib8b28e75e36d1c6beff8afa580fc4c29c23b6eb0
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26428Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Change-Id: I32f5d800fdbc0c2e672c3b92b64745fe34d23573
Reviewed-on: https://swiftshader-review.googlesource.com/c/SwiftShader/+/26528Reviewed-by: Ben Clayton <bclayton@google.com>
Tested-by: Ben Clayton <bclayton@google.com>