This commit changes wgpu::Device::CreateSwapChain to take an additional
wgpu::Surface argument. Passing nullptr is enough to stay on the
previous swapchain implementation, until the new one is ready.
In order to support both the "old" implementation-based swapchains and
the "new" surface-based swapchains. SwapChainBase is now split into
three abstract classes:
- SwapChainBase that has a virtual method for each of the
wgpu::SwapChain methods.
- OldSwapChainBase that corresponds to the implementation-based
swapchains.
- NewSwapChainBase that will contain the surface-based swapchain
implementation and will eventually just be renamed to SwapChainBase.
The interaction of the surface-based swapchains with the Surface objects
aren't implemented yet, neither are the swapchain methods. Only creation
works.
Validation tests for surface-based swapchain creation are added in the
end2end test target because they need to create OS windows.
Bug: dawn:269
Change-Id: I7e07d6c666479867b9a16d7b1b8c181d5dbd69a0
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15281
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
The _BitScanReverse64 intrinsic only exists when compiling for 64bit.
Replace it by two calls to _BitScanReverse (the 32bit one) when on
32bit.
Bug:
Change-Id: Ie294327ec914b0ca4a73732e4b78c1f2a08f100f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15321
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
- Fix a warning on 32bit for converting uint64_t to size_t for the
buffer mapped range.
- Fix a macro redifinition caused by including glfw3.h before windows.h
Bug:
Change-Id: I3897af55679d943b1dfc34b93bba7bd25d8fb7d6
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15320
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Roll third_party/SPIRV-Tools/ f8d7df760..323a81fc5 (2 commits)
f8d7df760c..323a81fc5e
$ git log f8d7df760..323a81fc5 --date=short --no-merges --format='%ad %ae %s'
2019-09-04 alanbaker Validate Vulkan 1.2 capabilities
2019-08-28 dneto Add basic support for Vulkan 1.2: SPV_ENV_VULKAN_1_2
Roll third_party/glslang/ d203754bc..3ed344dd7 (4 commits)
d203754bc1..3ed344dd78
$ git log d203754bc..3ed344dd7 --date=short --no-merges --format='%ad %ae %s'
2020-01-16 cepheus Fix#2059, and also attempt to skip test bots [skip ci]
2019-12-24 laddoc Add Error check flag in io mapper
2020-01-15 cepheus Update to latest SPIRV-Tools, supporting Vulkan 1.2.
2020-01-15 cepheus SPV/Vulkan: Add support for Vulkan 1.2, which defaults to SPIR-V 1.5.
Roll third_party/shaderc/ df0e683b5..f53792645 (2 commits)
df0e683b58..f53792645f
$ git log df0e683b5..f53792645 --date=short --no-merges --format='%ad %ae %s'
2020-01-15 rharrison Use the correct comparison when checking for not MSL (#960)
2020-01-15 dneto Support Vulkan 1.2 (#958)
Roll third_party/spirv-cross/ 172e39f03..f9818f080 (7 commits)
172e39f039..f9818f0804
$ git log 172e39f03..f9818f080 --date=short --no-merges --format='%ad %ae %s'
2020-01-16 post Update license headers to 2020.
2020-01-16 post Run format_all.sh.
2020-01-16 post MSL: Add support for force-activating IAB resources.
2020-01-16 post Deal with illegal names in types as well.
2020-01-15 post Reflection: Add array stride/matrix stride reflection.
2020-01-15 post Reflection: Emit reflection information for array size literalness.
2020-01-15 post GLSL: Implement geometry shader passthrough extension.
Created with:
roll-dep third_party/SPIRV-Tools third_party/glslang third_party/shaderc third_party/spirv-cross third_party/spirv-headers
Change-Id: Icf361e327598702ecfc0a4cdd3014d0060f2c083
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15200
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ryan Harrison <rharrison@chromium.org>
Handle DeviceLostCallback once DeviceLost error occurs.
Disallow any other commands or actions on device to happen after device
has been lost.
Bug: dawn:68
Change-Id: Icbbbadf278cae5e6213050d00439118789c863dc
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/12801
Commit-Queue: Natasha Lee <natlee@microsoft.com>
Reviewed-by: Austin Eng <enga@chromium.org>
This is a reland of 2b3975f808
The previous CL failed to retain autoreleased ObjC objects which
should live longer than the autoreleasepool block. This reland fixes
the issue and adds tests for it.
Original change's description:
> Metal: Add CommandRecordingContext
>
> Introduces the idea of a CommandRecordingContext to the Metal backend,
> similar to other backends. This is a class to track which Metal encoder
> is open on the device-global pending MTLCommandBuffer.
> It will be needed to open/close encoders for lazy clearing.
>
> Bug: dawn:145
> Change-Id: Ief6b71a079d73943677d2b61382d1c36b88a4f87
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14780
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Kai Ninomiya <kainino@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
Bug: dawn:145
Change-Id: I67494b35225ce8f6443a3fa9787d054522e5d422
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15042
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
This is another step to implement webgpu.h swapchains, Surface is
essentially a union type of all the types of windows that can be used to
create swapchains.
Changes to allow implementing wgpu::Surface and test its creation are:
- Add GLFWUtils.cpp/.h/_metal.mm that contains helpers used to use
WebGPU with GLFW. This deprecates BackendBinding.h that will be removed
when the NXT swapchain is removed.
- Add a `dawn_use_x11` GN variable to factor all the places in BUILD.gn
where we checked whether we should use X11.
- Add a `supports_glfw_for_windowing` GN variable in the main BUILD.gn
file to control which configuration tests and samples using GLFW can be
built.
- Add a ObjCUtils.h to contain some ObjC functionality that we'd need
in files that otherwise would be C++ (so that they can be compiled on
all platforms).
Bug: dawn:269
Change-Id: I25548142a1d1d1f05b0f4d71aa3bdc4698d19622
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15081
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
struct WGPUChainedStruct {
WGPUChainedStruct const * nextInChain;
WGPUSType sType;
};
And changes all the nextInChain to point to such structures. This adds
more type safety to extension structs and requires less casting to check
sTypes and friends.
Bug: dawn:269
Change-Id: I443f363cdb55dbec7c7f6e897245d4a7ea0ebe70
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15080
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
This reverts commit 2b3975f808.
Reason for revert: causes the failure in crbug.com/1041358
Original change's description:
> Metal: Add CommandRecordingContext
>
> Introduces the idea of a CommandRecordingContext to the Metal backend,
> similar to other backends. This is a class to track which Metal encoder
> is open on the device-global pending MTLCommandBuffer.
> It will be needed to open/close encoders for lazy clearing.
>
> Bug: dawn:145
> Change-Id: Ief6b71a079d73943677d2b61382d1c36b88a4f87
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14780
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Kai Ninomiya <kainino@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
TBR=cwallez@chromium.org,kainino@chromium.org,enga@chromium.org
# Not skipping CQ checks because original CL landed > 1 day ago.
Bug: dawn:145
Bug: chromium:1041358
Change-Id: I05c76cd96f723230d05cff65127dc8513d5e03c5
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15060
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Now ComputeCopyStorageBufferTests.SizedArrayOfStruct can pass on all
backends with the fix in the latest SPIRV-Cross.
BUG=chromium:1037829
TEST=dawn_end2end_tests
Change-Id: I20103c3b0848da394e82912f59548a71a3273ac1
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/15020
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
The way in which the Result class is used in Dawn can be fragile
with respect to memory management because the caller of AcquireError
must know they need to delete the returned pointer or a memory leak
will occur. We've had a couple of instances where developers have
accidentally left out the delete call and managed to get past code
review.
This CL changes the Result class so that it assumes the error is
allocated on the heap and forces the caller to use unique_ptr when
calling AcquireError.
Bug:dawn:320
Change-Id: I13ec953b0c37eaafbd6ce93c2f719b4743676acb
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14960
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Rafael Cintron <rafael.cintron@microsoft.com>
The dawn_native::Adapter::GetPCIInfo/GetBackendType/GetDeviceType
methods are now deprecated in favor of a method returning a webgpu.h
AdapterProperties structure. Deprecated function are still available to
avoid breaking Chromium or Skia compilation.
This reduces the difference between dawn.json and webgpu.h
BUG=dawn:160
Change-Id: Ib68fe1c4d1d87676c01c212c91f80fdd26056c56
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14541
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Roll third_party/SPIRV-Tools/ 96354f504..fb2e81981 (17 commits)
96354f5047..fb2e819819
$ git log 96354f504..fb2e81981 --date=short --no-merges --format='%ad %ae %s'
2020-01-08 dneto Fix GN build for OpenCL.DebugInfo.100 update (#3134)
2020-01-08 bclayton Fix bad parameter names in error message (#3129)
2020-01-07 alanbaker Revert PR #3093 (#3131)
2020-01-07 alanbaker Disallow forward references in arrays (#3093)
2020-01-07 afdx spirv-fuzz: Add fuzzer pass to perform module donation (#3117)
2020-01-06 dneto GetOperandConstants operand can be const (#3126)
2019-12-27 dneto Avoid pessimizing std::move (#3124)
2019-12-27 kburjack Fix typo in validation message (#3122)
2019-12-27 greg Change default version for CreatInstBindlessCheckPass to 2 (#3119)
2019-12-20 greg Fix convert-relax-to-half invalid code (#3099) (#3106)
2019-12-19 dneto Support OpenCL.DebugInfo.100 extended instruction set (#3080)
2019-12-19 afdx spirv-fuzz: Always add new globals to entry point interfaces (#3113)
2019-12-19 afdx spirv-fuzz: Transformation to add a new function to a module (#3114)
2019-12-19 afdx spirv-fuzz: Avoid passing access chains as parameters (#3112)
2019-12-18 dneto Add support for SPV_KHR_non_semantic_info (#3110)
2019-12-16 afdx spirv-fuzz: Transformations to add types, constants and variables (#3101)
2019-12-16 greg Make Instrumentation format version 2 the default (Step 1) (#3096)
Roll third_party/glslang/ ebf634bca..518136706 (13 commits)
ebf634bcaa..5181367062
$ git log ebf634bca..518136706 --date=short --no-merges --format='%ad %ae %s'
2020-01-07 cepheus Fix#1829: Add "--" command-line options for macro def/undef.
2020-01-08 laddoc Move symbol builtin check to grammar stage
2020-01-06 lryer Add missing extension defination
2020-01-06 cepheus Bump revision
2020-01-06 laddoc Add builtin constants
2019-11-26 laddoc Add support for ARB_gpu_shader_fp64
2019-12-26 cepheus HLSL: Fix#2037: Integer dot used incorrect input for adds.
2019-12-25 laddoc atomic counter offset should align to 4
2019-12-25 laddoc Modify atomic_uint binding check
2019-11-26 laddoc Add support for ARB_uniform_buffer_object
2019-11-26 laddoc Add support for ARB_texture_multisample
2019-11-26 laddoc Add support for ARB_sample_shading
2019-12-20 cepheus Command-line: Give better error messages. From #1829.
Roll third_party/shaderc/ 6d3e0ee67..0b9a2992c (3 commits)
6d3e0ee67f..0b9a2992c7
$ git log 6d3e0ee67..0b9a2992c --date=short --no-merges --format='%ad %ae %s'
2020-01-07 rharrison Rolling 5 dependencies and updating expectations (#946)
2019-12-18 rharrison Adding in missing APIs for MSL support in Dawn (#943)
2019-12-16 rharrison Rolling 4 dependencies (#942)
Roll third_party/spirv-cross/ f912c3289..172e39f03 (20 commits)
f912c32898..172e39f039
$ git log f912c3289..172e39f03 --date=short --no-merges --format='%ad %ae %s'
2020-01-09 post HLSL: Add a resource remapping API similar to MSL.
2020-01-09 post MSL: Deal with sign on wave min/max.
2020-01-09 post HLSL: Deal with casting for WaveActiveMin/Max.
2020-01-09 post GLSL: Deal with sign in subgroup Min/Max operations.
2020-01-08 post Run format_all.sh.
2020-01-08 post HLSL: Fix bug when reading and writing structs from SSBO.
2020-01-08 post HLSL: Implement stores for complex composites in ByteAddressBuffers.
2020-01-08 post HLSL: Support loading complex composites from ByteAddressBuffer.
2020-01-08 post Run format_all.sh.
2020-01-07 post MSL: Deal with padded fragment output + Component decoration.
2020-01-07 post MSL: Explicitly don't support component packing for tessellation.
2020-01-07 post MSL: Don't set OrigID when emitting component packed vectors.
2020-01-07 post MSL: Deal with packing vectors for vertex input/fragment output.
2020-01-07 post MSL: Add trivial tests for Component decoration.
2020-01-06 post Fix Clang warnings.
2020-01-06 post Roll custom versions of isalpha/isalnum.
2020-01-06 post Add test shader for OpCopyLogical with packing/unpacking.
2020-01-06 post Go through access chain path for OpCopyLogical.
2020-01-06 post Basic implementation of OpCopyLogical.
2019-12-21 dm86.jang Add debug prefix on Windows
Created with:
roll-dep third_party/SPIRV-Tools third_party/glslang third_party/shaderc third_party/spirv-cross third_party/spirv-headers
BUG=1037829
Change-Id: I0dc87160c607e615be6e6f2f6dd4a892b010de51
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14943
Reviewed-by: Ryan Harrison <rharrison@chromium.org>
Reviewed-by: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Ryan Harrison <rharrison@chromium.org>
This is a reland of d08a14b709
It gates dawn_use_swiftshader on dawn_enable_vulkan because the
Vulkan backend isn't supported on Mac. We may support Swiftshader
Vulkan on Mac in the future, but that will first require changes
to how we initialize the Vulkan backend.
Original change's description:
> Use Swiftshader by default when fuzzing
>
> Note: This still requires Chromium to have set dawn_swiftshader_dir
> to take effect.
>
> Bug: dawn:295
> Change-Id: I0c953196630dd2694f68d290e6f3f1e427c20a2f
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14821
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Reviewed-by: Kai Ninomiya <kainino@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
Bug: dawn:295
Change-Id: Ia125005b35951021702104a99a78712bcd47dd1e
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14941
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Introduces the idea of a CommandRecordingContext to the Metal backend,
similar to other backends. This is a class to track which Metal encoder
is open on the device-global pending MTLCommandBuffer.
It will be needed to open/close encoders for lazy clearing.
Bug: dawn:145
Change-Id: Ief6b71a079d73943677d2b61382d1c36b88a4f87
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14780
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Due to significant performance regressions on Intel Gen11 Graphics,
temporarily disable use of the D3D12 Render Pass API until a workaround
infrastructure can be implemented.
Bug: dawn:310
Change-Id: I994a2c2a0f6a3b61c48b083c73d6e0d3f8910dfa
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14663
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Brandon Jones <brandon1.jones@intel.com>
This patch moves Instance initialization and adapter discovery so that
it is done once globally, and not for every fuzz input. This is to work
around a bug where destructing the instance at the end of a run breaks
when fuzzing with Swiftshader.
Bug: dawn:295, chromium:1038952
Change-Id: Iabfe178f40b9df85d47a6353f16cd2ef26f39966
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14822
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
This is a reland of f58f69f66b
The whitebox dawn_end2end_tests are updated to link statically against
libdawn_native. This is required because the test link against and use
libdawn_native as sources. It is an error with MSVC to both import and
export symbols from libdawn_native.
Original change's description:
> fuzzing: Add supportsErrorInjection option to DawnWireServerFuzzer
>
> This option will be used by backends that support error injection so
> that errors can be injected into a "clean" corpus to generate a seed
> corpus with good examples of injected error conditions.
>
> Bug: dawn:295
> Change-Id: I837acdde6dd4274adb56edf8e4307427f8d6333b
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14681
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
Bug: dawn:295
Change-Id: Ifa092d28aa7ac57cfb197aa4daeb8408f8036d4a
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14820
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
This reverts commit f58f69f66b.
Reason for revert: This breaks the MSVC build because the whitebox end2end tests try to both import and export the error injector symbols from libdawn_native
Original change's description:
> fuzzing: Add supportsErrorInjection option to DawnWireServerFuzzer
>
> This option will be used by backends that support error injection so
> that errors can be injected into a "clean" corpus to generate a seed
> corpus with good examples of injected error conditions.
>
> Bug: dawn:295
> Change-Id: I837acdde6dd4274adb56edf8e4307427f8d6333b
> Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14681
> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
> Commit-Queue: Austin Eng <enga@chromium.org>
TBR=cwallez@chromium.org,kainino@chromium.org,enga@chromium.org
Change-Id: I14a15fcd094d431cbb8a29d5642a4a7fe6a11f4c
No-Presubmit: true
No-Tree-Checks: true
No-Try: true
Bug: dawn:295
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14741
Reviewed-by: Austin Eng <enga@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
jsoncpp causes some flaky failures because of inconsistencies in nested
DEPS. We barely use it for JSON serialization so the simplest solution
is to just manually write the serialization for perf test tracing data.
Bug: dawn:208
Change-Id: Ie5e4b5436e2c9e32e6817d64c6e95c774cb8751f
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14720
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Commit-Queue: Austin Eng <enga@chromium.org>
Results should be printed METRIC: STORY, not the other way around.
Also, story names shouldn't have slashes as it will parse as separate
chart segments and won't be allowed when we switch to using histograms.
Bug: dawn:208, dawn:311
Change-Id: Ifc893e5aa94eddcb3a08c0d4aff66b7a0f41620b
Reviewed-on: https://dawn-review.googlesource.com/c/dawn/+/14740
Commit-Queue: Austin Eng <enga@chromium.org>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
