Merge branch 'main' into attr-arg-consts

Author: danbrown-amd

cmake/modules/HandleLLVMOptions.cmake

@@ -301,7 +301,6 @@ if( MSVC )
 
   set(msvc_warning_flags
     # Disabled warnings.
-    -wd4146 # Suppress 'unary minus operator applied to unsigned type, result still unsigned'
     -wd4180 # Suppress 'qualifier applied to function type has no meaning; ignored'
     -wd4244 # Suppress ''argument' : conversion from 'type1' to 'type2', possible loss of data'
     -wd4258 # Suppress ''var' : definition from the for loop is ignored; the definition from the enclosing scope is used'

docs/ReleaseNotes.md

@@ -19,7 +19,9 @@ The included licenses apply to the following files:
 
 ### Upcoming Release
 
-Place release notes for the upcoming release below this line and remove this line upon naming this 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`.
+- Header file `dxcpix.h` was added to the release package.
 
 ### Version 1.8.2505
 

docs/SPIR-V.rst

@@ -896,6 +896,13 @@ are translated into SPIR-V ``OpTypeImage``, with parameters:
 The meanings of the headers in the above table is explained in ``OpTypeImage``
 of the SPIR-V spec.
 
+For storage images (e.g. ``RWTexture2D<T>``) and texel buffers (e.g. ``RWBuffer<T>``),
+the image format is typically inferred from the data type ``T``. However, the
+``-fspv-use-unknown-image-format`` command-line option can be used to change
+this behavior. When this option is active, the default format for these
+resources becomes ``Unknown`` if not otherwise specified by the
+``[[vk::image_format]]`` attribute.
+
 Vulkan specific Image Formats
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
@@ -1012,17 +1019,18 @@ right now:
 2. DirectX memory layout rules for uniform buffers and storage buffers:
    they allow packing data on the application side that can be shared with
    DirectX. They can be enabled by ``-fvk-use-dx-layout``.
+   
+   NOTE: This requires ``VK_EXT_scalar_block_layout`` to be enabled on the
+   application side.
 3. Strict OpenGL ``std140`` for uniform buffers and strict OpenGL ``std430``
    for storage buffers: they allow packing data on the application side that
    can be shared with OpenGL. They can be enabled by ``-fvk-use-gl-layout``.
 4. Scalar layout rules introduced via `VK_EXT_scalar_block_layout`, which
    basically aligns all aggregrate types according to their elements'
    natural alignment. They can be enabled by ``-fvk-use-scalar-layout``.
-
-To use scalar layout, the application side need to request
-``VK_EXT_scalar_block_layout``. This is also true for using DirectX memory
-layout since there is no dedicated DirectX layout extension for Vulkan
-(at least for now). So we must request something more permissive.
+   
+   NOTE: This requires ``VK_EXT_scalar_block_layout`` to be enabled on the
+   application side.
 
 In the above, "vector-relaxed OpenGL ``std140``/``std430``" rules mean OpenGL
 ``std140``/``std430`` rules with the following modification for vector type
@@ -1032,7 +1040,7 @@ alignment:
 2. If the above causes an `improper straddle <https://registry.khronos.org/vulkan/specs/latest/html/vkspec.html#interfaces-resources-layout>`_,
    the alignment will be set to 16 bytes.
 
-As an exmaple, for the following HLSL definition:
+As an example, for the following HLSL definition:
 
 .. code:: hlsl
 

external/SPIRV-Headers

@@ -154,6 +154,7 @@ const float kMaxMipLodBias = 15.99f;
 const float kMinMipLodBias = -16.0f;
 
 const unsigned kResRetStatusIndex = 4;
+const unsigned kVecResRetStatusIndex = 1;
 
 /* <py::lines('OLOAD_DIMS-TEXT')>hctdb_instrhelp.get_max_oload_dims()</py>*/
 // OLOAD_DIMS-TEXT:BEGIN

external/SPIRV-Tools

@@ -81,7 +81,7 @@ ModulePass *createResumePassesPass();
 FunctionPass *createMatrixBitcastLowerPass();
 ModulePass *createDxilCleanupAddrSpaceCastPass();
 ModulePass *createDxilRenameResourcesPass();
-ModulePass *createDxilScalarizeVectorLoadStoresPass();
+ModulePass *createDxilScalarizeVectorIntrinsicsPass();
 
 void initializeDxilLowerCreateHandleForLibPass(llvm::PassRegistry &);
 void initializeDxilAllocateResourcesForLibPass(llvm::PassRegistry &);
@@ -116,7 +116,7 @@ void initializeResumePassesPass(llvm::PassRegistry &);
 void initializeMatrixBitcastLowerPassPass(llvm::PassRegistry &);
 void initializeDxilCleanupAddrSpaceCastPass(llvm::PassRegistry &);
 void initializeDxilRenameResourcesPass(llvm::PassRegistry &);
-void initializeDxilScalarizeVectorLoadStoresPass(llvm::PassRegistry &);
+void initializeDxilScalarizeVectorIntrinsicsPass(llvm::PassRegistry &);
 
 ModulePass *createDxilValidateWaveSensitivityPass();
 void initializeDxilValidateWaveSensitivityPass(llvm::PassRegistry &);

include/dxc/DXIL/DxilConstants.h

@@ -462,6 +462,9 @@ const unsigned kHitObjectInvoke_PayloadOpIdx = 2;
 const unsigned kHitObjectFromRayQuery_WithAttrs_AttributeOpIdx = 4;
 const unsigned kHitObjectFromRayQuery_WithAttrs_NumOp = 5;
 
+// HitObject::GetAttributes
+const unsigned kHitObjectGetAttributes_AttributeOpIdx = 2;
+
 // Linear Algebra Operations
 
 // MatVecMul

include/dxc/HLSL/DxilGenerationPass.h

@@ -403,6 +403,12 @@ def fspv_enable_maximal_reconvergence: Flag<["-"], "fspv-enable-maximal-reconver
   HelpText<"Enables the MaximallyReconvergesKHR execution mode for this module.">;
 def fspv_use_vulkan_memory_model: Flag<["-"], "fspv-use-vulkan-memory-model">, Group<spirv_Group>, Flags<[CoreOption, DriverOption]>,
   HelpText<"Generates SPIR-V modules that use the Vulkan memory model instead of GLSL450.">;
+def fspv_use_unknown_image_format
+    : Flag<["-"], "fspv-use-unknown-image-format">,
+      Group<spirv_Group>,
+      Flags<[CoreOption, DriverOption]>,
+      HelpText<"For storage images and texel buffers, sets the default format to 'Unknown' when not specified via the `vk::image_format` attribute. If this option is not used, the format is inferred from the resource's data type.">;
+
 def fvk_auto_shift_bindings: Flag<["-"], "fvk-auto-shift-bindings">, Group<spirv_Group>, Flags<[CoreOption, DriverOption]>,
   HelpText<"Apply fvk-*-shift to resources without an explicit register assignment.">;
 def Wno_vk_ignored_features : Joined<["-"], "Wno-vk-ignored-features">, Group<spirv_Group>, Flags<[CoreOption, DriverOption, HelpHidden]>,

include/dxc/HLSL/HLOperations.h

@@ -71,6 +71,7 @@ struct SpirvCodeGenOptions {
   bool fixFuncCallArguments;
   bool enableMaximalReconvergence;
   bool useVulkanMemoryModel;
+  bool useUnknownImageFormat;
   bool IEEEStrict;
   /// Maximum length in words for the OpString literal containing the shader
   /// source for DebugSource and DebugSourceContinued. If the source code length