Use the SPIRV-Headers files from the subtree copy in third_party,
instead of the ones in include/spirv.

Bug b/123642959

Change-Id: I79f77d2f659d5623583b2de40536dec0c79e2f7f
Reviewed-on: https://swiftshader-review.googlesource.com/c/24348Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

Clang doesn't support the unused-lambda-capture warning until version
5.0: http://releases.llvm.org/5.0.0/tools/clang/docs/DiagnosticsReference.html#wunused-lambda-capture

Bug b/123933266

Change-Id: I6c5681a64e08ab479d726679ec18447dec926630
Reviewed-on: https://swiftshader-review.googlesource.com/c/24448Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

clang: warning: argument unused during compilation: '-s' [-Wunused-command-line-argument]

Bug: b/123933266
Change-Id: I809458fa08b9b6b3af7d31c19a90206462d0b0c6
Reviewed-on: https://swiftshader-review.googlesource.com/c/24429
Kokoro-Presubmit: Ben Clayton <bclayton@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

Bug: b/123894424
Change-Id: I9d3d24241d929cb80f9f28062537b50093009530
Reviewed-on: https://swiftshader-review.googlesource.com/c/24408
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Ben Clayton <bclayton@google.com>

There is no sense in passing -Wno-error=header-hygiene
if -Wno-header-hygiene is not passed (gcc case). Thus,
use -Wno-error=header-hygiene only with clang.

Bug: chromium:819294
Change-Id: I01bd9771352ab2e8827f1fdd493eb89bb719dd0f
Reviewed-on: https://swiftshader-review.googlesource.com/c/24428
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>

Seen building on Ubuntu 18.04.1:

```
libllvm.a(DynamicLibrary.cpp.o): In function `llvm::sys::DynamicLibrary::HandleSet::~HandleSet()':
DynamicLibrary.cpp:(.text._ZN4llvm3sys14DynamicLibrary9HandleSetD2Ev+0x1c): undefined reference to `dlclose'
DynamicLibrary.cpp:(.text._ZN4llvm3sys14DynamicLibrary9HandleSetD2Ev+0x2c): undefined reference to `dlclose'
libllvm.a(DynamicLibrary.cpp.o): In function `llvm::sys::DynamicLibrary::HandleSet::DLSym(void*, char const*)':
DynamicLibrary.cpp:(.text._ZN4llvm3sys14DynamicLibrary9HandleSet5DLSymEPvPKc+0x1): undefined reference to `dlsym'
```

Also removes --no-as-needed hack.

Bug: b/123360006
Change-Id: Icef1078e68cbf68af991e52f7fbb6e8d449e0809
Reviewed-on: https://swiftshader-review.googlesource.com/c/24310Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

We have two kinds of "variable" -- those that are directly backed by an
lvalue, and those that are subsets of an indexed location, etc space.

Split them so we don't have to rediscover it later at emit time.

Bug: b/120799499
Change-Id: I7403f76d615d7afa1ee50c014172f07b90262c40
Reviewed-on: https://swiftshader-review.googlesource.com/c/24374Tested-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Buffer usage was ignored in memory requirements. Usage support
was added for the purpose of computing the proper required alignment.

Fixes:
dEQP-VK.memory.requirements.core.buffer.regular

Bug b/118383648

Change-Id: Ib66e8d85f1b468765e24edba598fa06f4b5888c9
Reviewed-on: https://swiftshader-review.googlesource.com/c/23928
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

OpVariable handling needs to be split out; variables are clearly not
present in the type tree. Take care to apply the OpVariable's
decorations in the right place.

Bug: b/123779863

Change-Id: I2761a604683cf8b025f6b9ba1efa5383f53d7ed1
Reviewed-on: https://swiftshader-review.googlesource.com/c/24372Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>

This implementation of vkCmdSetEvent and vkCmdResetEvent assumes:
- There's only one queue (which currently is the case)
- Everything in the queue is executed linearly (which is currently the case)
- VkPipelineStageFlags can be ignored for now

Taking VkPipelineStageFlags into account could be used to avoid unnecessary
stalls or cache flushing. This will come at the optimization phase after
full conformance is achieved.

Bug b/117835459

Change-Id: Icb08a8b8e3fe63e827c9ba994a5e4353603e4ac9
Reviewed-on: https://swiftshader-review.googlesource.com/c/24371Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

Implemented vkResetCommandPool and vkTrimCommandPool.
The reset command out all command buffers in the command
pool back to the initial state. It also frees up the entire pool.

The trim command is a potential future memory usage optimization.
It is currently a noop.

Passes a few tests in dEQP-VK.api.command_buffers.

Bug b/119827933

Change-Id: Ic974f63a8dfca014462ac904218b9dcc3035fc8a
Reviewed-on: https://swiftshader-review.googlesource.com/c/24370Tested-by: Alexis Hétu <sugoi@google.com>
Kokoro-Presubmit: kokoro <noreply+kokoro@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

BUG b/123360006

Change-Id: Ib986b6409d7ac1d2ced6ea3b9995b026f960fa3e
Reviewed-on: https://swiftshader-review.googlesource.com/c/24289Tested-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

Bug: b/73656151
Change-Id: I93a4ad7c28f2177224ba3b3ae39c9eaba6c72f39
Reviewed-on: https://swiftshader-review.googlesource.com/c/23690Tested-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

This cl properly parses VkDeviceGroupDeviceCreateInfo,
but only allows the already supported default case.

Fixes all tests in:
dEQP-VK.api.object_management.*.device_group

Bug b/117974925

Change-Id: I42851f10bb9fc94848414bbd09b65ab8b807ae0e
Reviewed-on: https://swiftshader-review.googlesource.com/c/23968Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Older version of gcc mistakenly missed out on linking libdl without it.

Change-Id: I5c12972c140348fa504e4ff4acb81655b445b4a5
Reviewed-on: https://swiftshader-review.googlesource.com/c/24268Reviewed-by: Chris Forbes <chrisforbes@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>

Currently VMOs on Fuchsia are allocated with execute rights, but soon
that will not be the case, so the calls later to re-map memory as
executable will fail. Add zx_vmo_replace_as_executable call to ensure
they don't.

Change-Id: I45c5f5527bc36ed898b29095788006239bb2b0b3
Reviewed-on: https://swiftshader-review.googlesource.com/c/24188Reviewed-by: Wez <wez@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Matthew Dempsky <mdempsky@google.com>

Adds commit 'd14db341b834cfb3c574a258c331b3a6b1c2cbc5' from
https://github.com/KhronosGroup/SPIRV-Tools as 'third_party/SPIRV-Tools'

Commands:
git subtree add --prefix third_party/SPIRV-Tools
https://github.com/KhronosGroup/SPIRV-Tools
d14db341b834cfb3c574a258c331b3a6b1c2cbc5 --squash

Bug: b/123642959
Change-Id: I9de28ed72f80a4ed6280ac1b990a57eed92b9667

git-subtree-dir: third_party/SPIRV-Tools
git-subtree-split: d14db341b834cfb3c574a258c331b3a6b1c2cbc5

Adds commit '79b6681aadcb53c27d1052e5f8a0e82a981dbf2f' from
https://github.com/KhronosGroup/SPIRV-Headers as
'third_party/SPIRV-Headers'

Commands:
git subtree add --prefix third_party/SPIRV-Headers
https://github.com/KhronosGroup/SPIRV-Headers
79b6681aadcb53c27d1052e5f8a0e82a981dbf2f --squash

Bug: b/123642959
Change-Id: I39d6f1aff817ec9ed527c01eb631c99d81a17a95

git-subtree-dir: third_party/SPIRV-Headers
git-subtree-split: 79b6681aadcb53c27d1052e5f8a0e82a981dbf2f

Added the implementation for vkDeviceWaitIdle, which calls
into a noop vkQueueWaitIdle for now, until we have a non
noop implementation of Fence objects.

This allows many tests to run properly.

Bug b/117835459

Change-Id: I39be52f9462d87591bbb0bccee1e9e41b5a744b9
Reviewed-on: https://swiftshader-review.googlesource.com/c/24048Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

LLVM was updated to the latest 7.0.1 revision, which includes the proper
fix.

Reverts "Fixed build on Arm 64"
This reverts commit fed67899.

Bug b/123026156
Bug chromium:922089

Change-Id: I1f6ee4d3869e642e6ce2bd8cbf2c8f7553e4a607
Reviewed-on: https://swiftshader-review.googlesource.com/c/24150Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

* changes:
  Update to latest LLVM 7.0 release branch revision.
  Squashed 'third_party/llvm-7.0/llvm/' changes from 65ce2e56889..cd98f42d074

The simple Draw command now supports all of the input parameters.

Bug b/118619338

Change-Id: I51c3680dcc6497296e00000ad3e80f302629af3c
Reviewed-on: https://swiftshader-review.googlesource.com/c/24049Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

Since PipelineCache is a potential optimization, unmarking it as
UNIMPLEMENTED() and leaving a TODO instead. This can be addressed
much later, but for now, this shouldn't make tests fail.

Bug b/118386749 b/123588002

Change-Id: I5bdc15aedd9a01e35a457eef1a70c450083fd577
Reviewed-on: https://swiftshader-review.googlesource.com/c/23969Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Nicolas Capens <nicolascapens@google.com>

Merge commit '6583d4f1'

Commands:
git remote add llvm https://git.llvm.org/git/llvm.git
git subtree pull --prefix third_party/llvm-7.0/llvm llvm release_70 --squash

Bug b/123026156
Bug chromium:922089

Change-Id: If2de3cbde48d6acf02db68844bd16317869bf548

cd98f42d074 Merging r348444:
4a684f7170a Merging r346203:
06a6c330784 Merging r340125:
cc39aee5288 Merging r348462:
fa099fd85c1 Merging r348461:
a3ee448a5f6 Merging r348181:
3315ebed51e Merging r343369:
08f7943cf05 Merging r344589:
e8af9b4c407 Merging r339260:
f8ee49e3d8b Merging r347431:
92101a335c5 Merging r342865:
5993754bccb Merging r345353:
05998067a2e Merging r344591:
7515784defa Merging r344516:
7d87789bb6f Merging r342946:
2673179dd31 Merging r342884:
6ef29028d44 Merging r341919:
b7cd010ff5e Merging r341221:
aaf6ddfa1c7 Merging r340932:
b023def3de6 Merging r340931:
fe809519b1b Merging r340927:
d0abf8be7d1 Bump version to 7.0.1
2e8411d207a Merging r342354:
5ab8235732b Merging r344454, r344455, r344645:
4c946b7981d Merging r344325:
eb0722e28e7 Merging r342461:
888225e7bee Merging r343373:
b2504eaa1be Merging r343428:
9a0352f0c5e Merging r343443:
6c2e5a132cd Merging r343347:

git-subtree-dir: third_party/llvm-7.0/llvm
git-subtree-split: cd98f42d0747826062fc3d2d2fad383aedf58dd6

Bug b/123360006

Change-Id: I90ebde7968135f136f010d1f526aca26bc2b9418
Reviewed-on: https://swiftshader-review.googlesource.com/c/23988Reviewed-by: Chris Forbes <chrisforbes@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>

Respect loadOp when clearing attachments in beginRenderPass().

Bug b/119621736

Change-Id: Id9ad0f1ad9dc6116264782be3370c37c078d6846
Reviewed-on: https://swiftshader-review.googlesource.com/c/24050Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

The ZX_VM_FLAG_PERM_* names have been deprecated in favour of the
ZX_VM_PERM_* form, and will shortly be removed from the Fuchsia SDK.

Bug: chromium:925597
Change-Id: Ibb2d5c364fa62c9f7a0a9bb0cd715c41fd757522
Reviewed-on: https://swiftshader-review.googlesource.com/c/24068Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Wez <wez@google.com>

LLVM 3.0's raw_fd_ostream doesn't take an std::string as file name.

Bug swiftshader:123
Bug b/123193054

Change-Id: I994f6fe5ce98c573d4be48737934b7c6c0127434
Reviewed-on: https://swiftshader-review.googlesource.com/c/24088Tested-by: Nicolas Capens <nicolascapens@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Implemented subregion clears.
The quad layout had to be disabled to depth/stencil textures,
as it was not expected by dEQP.

Passes all tests in api.image_clearing.dedicated_allocation.
- partial_clear_color_attachment.*
- partial_clear_depth_stencil_attachment.*

Bug b/119621736

Change-Id: I9e854eceeef4a72b2ae507d62dbdbfa0b56c96f3
Reviewed-on: https://swiftshader-review.googlesource.com/c/23829Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Corentin Wallez <cwallez@google.com>

This is the implementation for vkCmdClearAttachments.

It includes the new command along with the plumbing to
get the information to the Image clear function.

The only non trivial behavior here is that the provided
baseArrayLayer for the clear command is to be offset by
the ImageView's subresourceRange's baseArrayLayer so
that it clears the correct layers.

Bug b/119621736

Passes all tests in:
api.image_clearing.dedicated_allocation.clear_color_attachment.*

Change-Id: I19f86b63239ca2fb4d53f2c6c9cb75185794f061
Reviewed-on: https://swiftshader-review.googlesource.com/c/23748Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

RenderPass information is required for RenderPass related commands,
like clearing attachments. This cl adds storage for attachments,
subpasses and dependencies related information in the RenderPass
objects and stores all this information in new class members.

Bug b/119620965

Change-Id: I19d8b8465c807755881de6471700a119082575f8
Reviewed-on: https://swiftshader-review.googlesource.com/c/23788Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

Added QueryPool object.

Fixes all tests in:
dEQP-VK.api.object_management.*.query_pool

Change-Id: Ie564848b616714c39961e7a80bc421ec7bb0ef58
Reviewed-on: https://swiftshader-review.googlesource.com/c/23948Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>
Reviewed-by: Corentin Wallez <cwallez@google.com>

The Image class now contains all the required functions to fill
out the VkSubresourceLayout structure, so this simply hooks the
proper calls to implement vkGetImageSubresourceLayout.

Bug b/119620767

Change-Id: I104811be4a5be31ef5db7ad55965ba263dd110ed
Reviewed-on: https://swiftshader-review.googlesource.com/c/24051Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Chris Forbes <chrisforbes@google.com>

When using TransformFeedback with DrawArraysInstanced, a crash
would occur if no varyings were set to use transform feedback
due to a null transform feedback buffer. Added a null pointer
check and an assert to fix this.

Bug b/117080493

Change-Id: Ieb226bfb70ae837bdb206bdbefe84f62dc1f2204
Reviewed-on: https://swiftshader-review.googlesource.com/c/23569Tested-by: Alexis Hétu <sugoi@google.com>
Reviewed-by: Corentin Wallez <cwallez@google.com>

Bug b/120382288

Change-Id: I506920a1ff5f68c2ece7c3e8327525c0bd5609ce
Reviewed-on: https://swiftshader-review.googlesource.com/c/23889Tested-by: Lingfeng Yang <lfy@google.com>
Reviewed-by: Alexis Hétu <sugoi@google.com>

Previously this would just repeatedly overwrite the one file, which
makes it difficult to see what's happening when you compile a pipeline
that needs 3 routines.

Change-Id: Ibab115f5ccbb17333d504e4e7948be54ca304f9d
Bug: b/123193054
Reviewed-on: https://swiftshader-review.googlesource.com/c/23469Reviewed-by: Alexis Hétu <sugoi@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>

Routine names were passed as wide strings for the integration with
CodeAnalyst, which no longer exists.

- Rip out the remnants of the CodeAnalyst support
- Change Reactor interface to take routine names as const char *

Bug: b/123193048
Change-Id: I919ce3a55c59c3a08057f85cac994fbffad37614
Reviewed-on: https://swiftshader-review.googlesource.com/c/23908Reviewed-by: Nicolas Capens <nicolascapens@google.com>
Tested-by: Chris Forbes <chrisforbes@google.com>