Merge branch 'main' of github.com:microsoft/DirectXShaderCompiler into devshFixes

Author: AnastaZIuk

azure-pipelines.yml

@@ -36,7 +36,7 @@ stages:
       submodules: true
     - script: |
         call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
-        call utils\hct\hctbuild.cmd -$(platform) -$(configuration) -show-cmake-log -spirvtest
+        call utils\hct\hctbuild.cmd -$(platform) -$(configuration) -show-cmake-log -spirvtest -warp-nuget-version 1.0.16.1
       displayName: 'Building'
     - script: |
         call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
@@ -45,7 +45,23 @@ stages:
     - script: |
         call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
         call utils\hct\hcttest.cmd -$(configuration) exec
-      displayName: 'DXIL Execution Tests'
+      displayName: 'DXIL Execution Tests'        
+    - script: |
+        call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
+        call utils\hct\hcttest.cmd -$(configuration) compat-suite 1.6
+      displayName: 'DXIL Compat Suite Tests (1.6 release)'
+      condition: succeededOrFailed()
+    - script: |
+        call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
+        call utils\hct\hcttest.cmd -$(configuration) compat-suite 1.7
+      displayName: 'DXIL Compat Suite Tests (1.7 release)'
+      condition: succeededOrFailed()
+    - script: |
+        call utils\hct\hctstart.cmd %HLSL_SRC_DIR% %HLSL_BLD_DIR%
+        call utils\hct\hcttest.cmd -$(configuration) compat-suite 1.8
+      displayName: 'DXIL Compat Suite Tests (1.8 release)'
+      condition: succeededOrFailed()
+    
 
   - job: Nix
     timeoutInMinutes: 165
@@ -80,6 +96,7 @@ stages:
           configuration: Debug
           CC: gcc-9
           CXX: g++-9
+          CMAKE_OPTS: -DLLVM_USE_SPLIT_DWARF=On
         MacOS_Clang_Release:
           image: ${{ variables.macOS }}
           configuration: Release
@@ -117,6 +134,11 @@ stages:
       displayName: 'Installing dependencies'
       condition: eq(variables['image'], variables['linux'])
 
+    - bash: |
+        sudo apt-get install g++-9 -y
+      displayName: 'Installing g++-9'
+      condition: and(eq(variables['CXX'], 'g++-9'), eq(variables['image'], variables['linux']))
+
     - bash: |
         brew update
         brew install ninja
@@ -143,4 +165,3 @@ stages:
         testResultsFormat: 'JUnit'
         testResultsFiles: '**/testresults.xunit.xml'
       condition: succeededOrFailed()
-

cmake/modules/AddLLVM.cmake

@@ -222,11 +222,35 @@ function(set_output_directory target bindir libdir)
     set(moddir ${libdir})
   endif()
   if(NOT "${CMAKE_CFG_INTDIR}" STREQUAL ".")
+    # HLSL change begin: work around broken MSVC_BUILD_AS_X
+    get_target_property(target_type ${target} TYPE)
+    if(target_type STREQUAL "SHARED_LIBRARY" AND MSVC_BUILD_AS_X EQUAL 1)
+      set(arm64x_workaround 1)
+      message(NOTICE "Working around ARM64X / MSVC_BUILD_AS_X bug for ${target}")
+    else()
+      set(arm64x_workaround 0)
+    endif()
+    # HLSL change end
+
     foreach(build_mode ${CMAKE_CONFIGURATION_TYPES})
       string(TOUPPER "${build_mode}" CONFIG_SUFFIX)
       string(REPLACE ${CMAKE_CFG_INTDIR} ${build_mode} bi ${bindir})
       string(REPLACE ${CMAKE_CFG_INTDIR} ${build_mode} li ${libdir})
       string(REPLACE ${CMAKE_CFG_INTDIR} ${build_mode} mi ${moddir})
+
+      # HLSL change begin: work around broken MSVC_BUILD_AS_X
+      #
+      # MSVC_BUILD_AS_X generates a vcxproj that builds for both ARM64 and
+      # ARM64EC, but these two platforms are only visible in msbuild. It also
+      # currently configures the import libraries to all be output to the same
+      # location, which means it's a race for if we get the ARM64 or ARM64X
+      # version there.  The appended $(PLATFORM) is evaluated by msbuild at
+      # build time to separate the ARM64EC from ARM64 builds.
+      if(arm64x_workaround EQUAL 1)
+        string(APPEND li "/$(PLATFORM)")
+      endif()
+      # HLSL change end
+
       set_target_properties(${target} PROPERTIES "RUNTIME_OUTPUT_DIRECTORY_${CONFIG_SUFFIX}" ${bi})
       set_target_properties(${target} PROPERTIES "ARCHIVE_OUTPUT_DIRECTORY_${CONFIG_SUFFIX}" ${li})
       set_target_properties(${target} PROPERTIES "LIBRARY_OUTPUT_DIRECTORY_${CONFIG_SUFFIX}" ${mi})

cmake/modules/FindTAEF.cmake

@@ -44,7 +44,7 @@ set(TAEF_SDK_BIN ${TAEF_INCLUDE_DIR}/../../Runtimes/TAEF)
 if ((CMAKE_GENERATOR_PLATFORM STREQUAL "x64") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" STREQUAL "x64"))
   set(TAEF_BIN_ARCH "amd64")
   set(TAEF_ARCH "x64")
-elseif ((CMAKE_GENERATOR_PLATFORM STREQUAL "x86") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" STREQUAL "x86"))
+elseif ((CMAKE_GENERATOR_PLATFORM STREQUAL "Win32") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" STREQUAL "X86"))
   set(TAEF_BIN_ARCH "x86")
   set(TAEF_ARCH "x86")
 elseif ((CMAKE_GENERATOR_PLATFORM MATCHES "ARM64.*") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" MATCHES "ARM64.*"))
@@ -53,7 +53,12 @@ elseif ((CMAKE_GENERATOR_PLATFORM MATCHES "ARM64.*") OR ("${CMAKE_C_COMPILER_ARC
 elseif ((CMAKE_GENERATOR_PLATFORM MATCHES "ARM.*") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" MATCHES "ARM.*"))
   set(TAEF_BIN_ARCH "arm")
   set(TAEF_ARCH "arm")
-endif((CMAKE_GENERATOR_PLATFORM STREQUAL "x64") OR ("${CMAKE_C_COMPILER_ARCHITECTURE_ID}" STREQUAL "x64"))
+else ()
+  message(FATAL_ERROR
+    "Unable to determine TAEF architecture - CMAKE_GENERATOR_PLATFORM is "
+    "${CMAKE_GENERATOR_PLATFORM}, CMAKE_C_COMPILER_ARCHITECTURE_ID is "
+    "${CMAKE_C_COMPILER_ARCHITECTURE_ID}.")
+endif()
 
 set (TAEF_ARCH ${TAEF_ARCH} CACHE INTERNAL "arch for taef test")
 find_program(TAEF_EXECUTABLE te.exe PATHS

cmake/modules/HandleLLVMOptions.cmake

@@ -567,9 +567,19 @@ if(LLVM_USE_SANITIZER)
   endif()
 endif()
 
-# Turn on -gsplit-dwarf if requested
-if(LLVM_USE_SPLIT_DWARF)
-  add_definitions("-gsplit-dwarf")
+# Turn on -gsplit-dwarf if requested in debug builds.
+if (LLVM_USE_SPLIT_DWARF AND
+    ((uppercase_CMAKE_BUILD_TYPE STREQUAL "DEBUG") OR
+     (uppercase_CMAKE_BUILD_TYPE STREQUAL "RELWITHDEBINFO")))
+  # Limit to clang and gcc so far. Add compilers supporting this option.
+  if (CMAKE_CXX_COMPILER_ID MATCHES "Clang" OR
+      CMAKE_CXX_COMPILER_ID STREQUAL "GNU")
+    add_compile_options($<$<COMPILE_LANGUAGE:C,CXX>:-gsplit-dwarf>)
+    include(CheckLinkerFlag)
+    check_linker_flag(CXX "-Wl,--gdb-index" LINKER_SUPPORTS_GDB_INDEX)
+    append_if(LINKER_SUPPORTS_GDB_INDEX "-Wl,--gdb-index"
+      CMAKE_EXE_LINKER_FLAGS CMAKE_MODULE_LINKER_FLAGS CMAKE_SHARED_LINKER_FLAGS)
+  endif()
 endif()
 
 add_llvm_definitions( -D__STDC_CONSTANT_MACROS )

docs/BuildingAndTestingDXC.rst

@@ -36,7 +36,7 @@ Building on windows additionally requires:
 * [Visual Studio 2019 or later](https://www.visualstudio.com/downloads) - select the following workloads: 
     * Universal Windows Platform Development
     * Desktop Development with C++
-* [Windows SDK](https://developer.microsoft.com/en-US/windows/downloads/windows-10-sdk) - version 10.0.18362.0 or newer
+* [Windows SDK](https://developer.microsoft.com/en-US/windows/downloads/windows-sdk) - version 10.0.26100.0 or newer
 * [Windows Driver Kit](https://docs.microsoft.com/en-us/windows-hardware/drivers/download-the-wdk) - same version as the SDK
 
 Note: DXC uses submodules for some external dependencies. You must initialize

docs/DXIL.rst

@@ -616,7 +616,7 @@ Each HLSL signature parameter is defined via C-like type, interpolation mode, an
 
 SV semantics add specific meanings and constraints to associated parameters. A parameter may be supplied by the hardware, and is then known as a System Generated Value (SGV). Alternatively, a parameter may be interpreted by the hardware and is then known as System Interpreted Value (SIV).  SGVs and SIVs are pipeline-stage dependent; moreover, some participate in signature packing and some do not. Non-SV semantics always participate in signature packing.
 
-Most System Generated Values (SGV) are loaded using special Dxil intrinsic functions, rather than loading the input from a signature.  These usually will not be present in the signature at all.  Their presence may be detected by the declaration and use of the special instrinsic function itself.  The exceptions to this are notible.  In one case they are present and loaded from the signature instead of a special intrinsic because they must be part of the packed signature potentially passed from the prior stage, allowing the prior stage to override these values, such as for SV_PrimitiveID and SV_IsFrontFace that may be written in the the Geometry Shader.  In another case, they identify signature elements that still contribute to DXBC signature for informational purposes, but will only use the special intrinsic function to read the value, such as for SV_PrimitiveID for GS input and SampleIndex for PS input.
+Most System Generated Values (SGV) are loaded using special Dxil intrinsic functions, rather than loading the input from a signature.  These usually will not be present in the signature at all.  Their presence may be detected by the declaration and use of the special instrinsic function itself.  The exceptions to this are notable.  In one case they are present and loaded from the signature instead of a special intrinsic because they must be part of the packed signature potentially passed from the prior stage, allowing the prior stage to override these values, such as for SV_PrimitiveID and SV_IsFrontFace that may be written in the the Geometry Shader.  In another case, they identify signature elements that still contribute to DXBC signature for informational purposes, but will only use the special intrinsic function to read the value, such as for SV_PrimitiveID for GS input and SampleIndex for PS input.
 
 The classification of behavior for various system values in various signature locations is described in a table organized by SemanticKind and SigPointKind.  The SigPointKind is a new classification that uniquely identifies each set of parameters that may be input or output for each entry point.  For each combination of SemanticKind and SigPointKind, there is a SemanticInterpretationKind that defines the class of treatment for that location.
 

docs/ReleaseNotes.md

@@ -19,9 +19,10 @@ The included licenses apply to the following files:
 
 ### Upcoming Release
 
-- Fix regression: [#7510](https://github.com/microsoft/DirectXShaderCompiler/issues/7510) crash when calling `sizeof` on templated type.
-- Fix regression: [#7508](https://github.com/microsoft/DirectXShaderCompiler/issues/7508) crash when calling `Load` with `status`.
+- Fixed regression: [#7510](https://github.com/microsoft/DirectXShaderCompiler/issues/7510) crash when calling `sizeof` on templated type.
+- Fixed regression: [#7508](https://github.com/microsoft/DirectXShaderCompiler/issues/7508) crash when calling `Load` with `status`.
 - Header file `dxcpix.h` was added to the release package.
+- Moved Linear Algebra (Cooperative Vector) DXIL Opcodes to experimental Shader Model 6.10
 
 ### Version 1.8.2505
 

docs/SPIR-V.rst

@@ -285,6 +285,21 @@ Right now the following ``<builtin>`` are supported:
 Please see Vulkan spec. `15.9. Built-In Variables <https://registry.khronos.org/vulkan/specs/latest/html/vkspec.html#interfaces-builtin-variables>`_
 for detailed explanation of these builtins.
 
+Helper Lane Support
+~~~~~~~~~~~~~~~~~~~
+
+Shader Model 6.7 introduces the `[WaveOpsIncludeHelperLanes]` attribute. When this
+attribute is applied to a shader entry point, the SPIR-V backend will:
+
+1. Add the ``SPV_KHR_maximal_reconvergence`` and ``SPV_KHR_quad_control``
+   extensions to the module.
+2. Add the ``QuadControlKHR`` capability.
+3. Add the ``MaximallyReconvergesKHR`` and ``RequireFullQuadsKHR`` execution modes
+   to the entry point.
+
+This ensures that helper lanes are included in wave operations, which is the
+behavior required by the HLSL specification.
+
 Supported extensions
 ~~~~~~~~~~~~~~~~~~~~
 

external/SPIRV-Tools

@@ -29,7 +29,7 @@ namespace DXIL {
 const unsigned kDxilMajor = 1;
 /* <py::lines('VALRULE-TEXT')>hctdb_instrhelp.get_dxil_version_minor()</py>*/
 // VALRULE-TEXT:BEGIN
-const unsigned kDxilMinor = 9;
+const unsigned kDxilMinor = 10;
 // VALRULE-TEXT:END
 
 inline unsigned MakeDxilVersion(unsigned DxilMajor, unsigned DxilMinor) {
@@ -1087,6 +1087,7 @@ enum class OpCode : unsigned {
   NumOpCodes_Dxil_1_6 = 222,
   NumOpCodes_Dxil_1_7 = 226,
   NumOpCodes_Dxil_1_8 = 258,
+  NumOpCodes_Dxil_1_9 = 312,
 
   NumOpCodes = 312 // exclusive last value of enumeration
 };
@@ -1422,6 +1423,7 @@ enum class OpCodeClass : unsigned {
   NumOpClasses_Dxil_1_6 = 149,
   NumOpClasses_Dxil_1_7 = 153,
   NumOpClasses_Dxil_1_8 = 174,
+  NumOpClasses_Dxil_1_9 = 196,
 
   NumOpClasses = 196 // exclusive last value of enumeration
 };