Implement explciit binding for metal and wgsl. (#5778)

* Respect explicit bindings in wgsl emit.

* Implement explciit binding generation for metal and wgsl.

* Update toc.

* Fix warnings in tests.

* Fix tests.

---------

Co-authored-by: Ellie Hermaszewska <ellieh@nvidia.com>

Author: csyonghe

docs/user-guide/a2-02-metal-target-specific.md

@@ -267,3 +267,11 @@ Metal requires explicit address space qualifiers. Slang automatically assigns ap
 | RW/Structured Buffers | `device`            |
 | Group Shared          | `threadgroup`       |
 | Parameter Blocks      | `constant`          |
+
+## Explicit Parameter Binding
+
+The HLSL `:register()` semantic is respected when emitting Metal code.
+
+Since metal does not differentiate a constant buffer, a shader resource (read-only) buffer and an unordered access buffer, Slang will map `register(tN)`, `register(uN)` and `register(bN)` to `[[buffer(N)]]` when such `register` semantic is declared on a buffer typed parameter.
+
+`spaceN` specifiers inside `register` semantics are ignored.

docs/user-guide/a2-03-wgsl-target-specific.md

@@ -156,3 +156,9 @@ Matrix type translation
 A m-row-by-n-column matrix in Slang, represented as float`m`x`n` or matrix<T, m, n>, is translated to `mat[n]x[m]` in WGSL, i.e. a matrix with `n` columns and `m` rows.
 The rationale for this inversion of terminology is the same as [the rationale for SPIR-V](a2-01-spirv-target-specific.md#matrix-type-translation).
 Since the WGSL matrix multiplication convention is the normal one, where inner products of rows of the matrix on the left are taken with columns of the matrix on the right, the order of matrix products is also reversed in WGSL. This is relying on the fact that the transpose of a matrix product equals the product of the transposed matrix operands in reverse order.
+
+## Explicit Parameter Binding
+
+The `[vk::binding(index,set)]` attribute is respected when emitting WGSL code, and will translate to `@binding(index) @group(set)` in WGSL.
+
+If the `[vk::binding()]` attribute is not specified by a `:register()` semantic is present, Slang will derive the binding from the `register` semantic the same way as the SPIRV and GLSL backends.

docs/user-guide/toc.html

@@ -233,6 +233,7 @@
 <li data-link="metal-target-specific#value-type-conversion"><span>Value Type Conversion</span></li>
 <li data-link="metal-target-specific#conservative-rasterization"><span>Conservative Rasterization</span></li>
 <li data-link="metal-target-specific#address-space-assignment"><span>Address Space Assignment</span></li>
+<li data-link="metal-target-specific#explicit-parameter-binding"><span>Explicit Parameter Binding</span></li>
 </ul>
 </li>
 <li data-link="wgsl-target-specific"><span>WGSL specific functionalities</span>
@@ -249,6 +250,7 @@
 <li data-link="wgsl-target-specific#pointers"><span>Pointers</span></li>
 <li data-link="wgsl-target-specific#address-space-assignment"><span>Address Space Assignment</span></li>
 <li data-link="wgsl-target-specific#matrix-type-translation"><span>Matrix type translation</span></li>
+<li data-link="wgsl-target-specific#explicit-parameter-binding"><span>Explicit Parameter Binding</span></li>
 </ul>
 </li>
 </ul>

source/slang/slang-emit-wgsl.cpp

@@ -663,7 +663,7 @@ void WGSLSourceEmitter::emitLayoutQualifiersImpl(IRVarLayout* layout)
 
             EmitVarChain chain = {};
             chain.varLayout = layout;
-            auto space = getBindingSpaceForKinds(&chain, kind);
+            auto space = getBindingSpaceForKinds(&chain, LayoutResourceKindFlag::make(kind));
             m_writer->emit("@group(");
             m_writer->emit(space);
             m_writer->emit(") ");

source/slang/slang-parameter-binding.cpp

@@ -984,7 +984,8 @@ static void addExplicitParameterBindings_HLSL(
     //
     // For now we do the filtering on target in a very direct fashion:
     //
-    if (!isD3DTarget(context->getTargetRequest()) && !isMetalTarget(context->getTargetRequest()))
+    bool isMetal = isMetalTarget(context->getTargetRequest());
+    if (!isD3DTarget(context->getTargetRequest()) && !isMetal)
         return;
 
     auto typeLayout = varLayout->typeLayout;
@@ -1018,13 +1019,27 @@ static void addExplicitParameterBindings_HLSL(
         if (kind == LayoutResourceKind::None)
             continue;
 
+
         // TODO: need to special-case when this is a `c` register binding...
 
         // Find the appropriate resource-binding information
         // inside the type, to see if we even use any resources
         // of the given kind.
 
         auto typeRes = typeLayout->FindResourceInfo(kind);
+        if (isMetal && !typeRes)
+        {
+            // Metal doesn't distinguish a unordered access and a readonly/uniform buffer.
+            switch (kind)
+            {
+            case LayoutResourceKind::UnorderedAccess:
+            case LayoutResourceKind::ShaderResource:
+                semanticInfo.kind = LayoutResourceKind::MetalBuffer;
+                typeRes = typeLayout->FindResourceInfo(LayoutResourceKind::MetalBuffer);
+                break;
+            }
+        }
+
         LayoutSize count = 0;
         if (typeRes)
         {
@@ -1073,7 +1088,7 @@ static void addExplicitParameterBindings_GLSL(
     // so that we are able to distinguish between
     // Vulkan and OpenGL as targets.
     //
-    if (!isKhronosTarget(context->getTargetRequest()))
+    if (!isKhronosTarget(context->getTargetRequest()) && !isWGPUTarget(context->getTargetRequest()))
         return;
 
     auto typeLayout = varLayout->typeLayout;

tests/bugs/gh-471.slang

@@ -22,7 +22,7 @@ int test(int inVal)
 	return x * 16;
 }
 
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/bugs/gh-775.slang

@@ -18,7 +18,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[9 9 9 9], stride=4):out,name outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/bugs/static-method.slang

@@ -11,7 +11,7 @@ struct S
 }
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 int test(int t)
 {

tests/bugs/static-var.slang

@@ -8,7 +8,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[9 9 9 9], stride=4):out,name outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/bugs/texture2d-gather.hlsl

@@ -3,8 +3,9 @@
 //TEST_INPUT: Texture2D(size=16, content=chessboard, format=R32_FLOAT):name g_texture
 //TEST_INPUT: Sampler :name g_sampler
 
-Texture2D<float> g_texture : register(t0);
-SamplerState g_sampler : register(s0);
+Texture2D<float> g_texture;
+
+SamplerState g_sampler;
 
 cbuffer Uniforms
 {

tests/bugs/type-legalize-bug-1.slang

@@ -5,7 +5,7 @@
 //TEST_INPUT:type_conformance A:IFoo=0
 //TEST_INPUT:type_conformance B:IFoo=1
 
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 interface IFoo
 {
     associatedtype T : IFoo;

tests/compute/break-stmt.slang

@@ -16,7 +16,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/continue-stmt.slang

@@ -21,7 +21,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/default-initializer.slang

@@ -23,7 +23,7 @@ int test(int value)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/explicit-this-expr.slang

@@ -16,7 +16,7 @@ struct A
 };
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
-RWStructuredBuffer<float> outputBuffer : register(u0);
+RWStructuredBuffer<float> outputBuffer;
 
 
 float test(float inVal)

tests/compute/generics-constrained.slang

@@ -28,6 +28,7 @@ float testHelp(T helper)
 }
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
+[vk::binding(0, 0)]
 RWStructuredBuffer<float> outputBuffer : register(u0);
 
 

tests/compute/global-init.slang

@@ -12,7 +12,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/implicit-generic-app.slang

@@ -30,7 +30,7 @@ int test(int val)
 }
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/implicit-this-expr.slang

@@ -15,7 +15,7 @@ struct A
 };
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
-RWStructuredBuffer<float> outputBuffer : register(u0);
+RWStructuredBuffer<float> outputBuffer;
 
 float test(float inVal)
 {

tests/compute/init-list-defaults.slang

@@ -24,7 +24,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[9 9 9 9], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/inout.slang

@@ -36,7 +36,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/multiple-continue-sites.slang

@@ -28,7 +28,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/struct-default-init.slang

@@ -26,7 +26,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[9 9 9 9], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/switch-stmt.slang

@@ -30,7 +30,7 @@ int test(int inVal)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3 4 5 6 7], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(8, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/this-type.slang

@@ -36,7 +36,7 @@ int test(int value)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/compute/user-defined-initializer.slang

@@ -28,7 +28,7 @@ int test(int value)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/language-feature/properties/property-in-interface.slang

@@ -46,7 +46,7 @@ int test(int value)
 }
 
 //TEST_INPUT:ubuffer(data=[0 1 2 3], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 [numthreads(4, 1, 1)]
 void computeMain(uint3 dispatchThreadID : SV_DispatchThreadID)

tests/preprocessor/line-macro.slang

@@ -1,7 +1,7 @@
 //TEST:COMPARE_COMPUTE(filecheck-buffer=CHECK): -output-using-type
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
-RWStructuredBuffer<int> outputBuffer : register(u0);
+RWStructuredBuffer<int> outputBuffer;
 
 #define T(x) x
 #define LL T(__LINE__)

tests/serialization/std-lib-serialize.slang

@@ -12,7 +12,7 @@ struct A
 };
 
 //TEST_INPUT:ubuffer(data=[0 0 0 0], stride=4):out,name=outputBuffer
-RWStructuredBuffer<float> outputBuffer : register(u0);
+RWStructuredBuffer<float> outputBuffer;
 
 
 float test(float inVal)

tests/wgsl/explicit-binding.slang

@@ -0,0 +1,23 @@
+//TEST:SIMPLE(filecheck=METAL): -target metal
+//TEST:SIMPLE(filecheck=CHECK): -target wgsl -entry computeMain -stage compute
+
+// CHECK-DAG: @binding(9) @group(7)
+// CHECK-DAG: @binding(3) @group(4)
+// CHECK-DAG: @binding(1) @group(2)
+
+// METAL-DAG: buffer(9)
+// METAL-DAG: texture(7)
+
+[vk::binding(1, 2)]
+Texture2D texA : register(t7);
+
+[vk::binding(3, 4)]
+ConstantBuffer<float> cb;
+
+RWStructuredBuffer<float> ob : register(u9, space7);
+
+[numthreads(1,1,1)]
+void computeMain()
+{
+    ob[0] = cb + texA.Load(int3(0)).x;
+}