4 files changed, 45 insertions, 2 deletions
diff --git a/clang/lib/CodeGen/CGHLSLBuiltins.cpp b/clang/lib/CodeGen/CGHLSLBuiltins.cpp
index 4f2f5a76..384bd59 100644
--- a/clang/lib/CodeGen/CGHLSLBuiltins.cpp
+++ b/clang/lib/CodeGen/CGHLSLBuiltins.cpp
@@ -160,6 +160,16 @@ static Value *handleHlslSplitdouble(const CallExpr *E, CodeGenFunction *CGF) {
   return LastInst;
 }
 
+static Value *emitBufferStride(CodeGenFunction *CGF, const Expr *HandleExpr,
+                               LValue &Stride) {
+  // Figure out the stride of the buffer elements from the handle type.
+  auto *HandleTy =
+      cast<HLSLAttributedResourceType>(HandleExpr->getType().getTypePtr());
+  QualType ElementTy = HandleTy->getContainedType();
+  Value *StrideValue = CGF->getTypeSize(ElementTy);
+  return CGF->Builder.CreateStore(StrideValue, Stride.getAddress());
+}
+
 // Return dot product intrinsic that corresponds to the QT scalar type
 static Intrinsic::ID getDotProductIntrinsic(CGHLSLRuntime &RT, QualType QT) {
   if (QT->isFloatingType())
@@ -372,6 +382,19 @@ Value *CodeGenFunction::EmitHLSLBuiltinExpr(unsigned BuiltinID,
         RetTy, CGM.getHLSLRuntime().getNonUniformResourceIndexIntrinsic(),
         ArrayRef<Value *>{IndexOp});
   }
+  case Builtin::BI__builtin_hlsl_resource_getdimensions_x: {
+    Value *Handle = EmitScalarExpr(E->getArg(0));
+    LValue Dim = EmitLValue(E->getArg(1));
+    llvm::Type *RetTy = llvm::Type::getInt32Ty(getLLVMContext());
+    Value *DimValue = Builder.CreateIntrinsic(
+        RetTy, CGM.getHLSLRuntime().getGetDimensionsXIntrinsic(),
+        ArrayRef<Value *>{Handle});
+    return Builder.CreateStore(DimValue, Dim.getAddress());
+  }
+  case Builtin::BI__builtin_hlsl_resource_getstride: {
+    LValue Stride = EmitLValue(E->getArg(1));
+    return emitBufferStride(this, E->getArg(0), Stride);
+  }
   case Builtin::BI__builtin_hlsl_all: {
     Value *Op0 = EmitScalarExpr(E->getArg(0));
     return Builder.CreateIntrinsic(
diff --git a/clang/lib/CodeGen/CGHLSLRuntime.h b/clang/lib/CodeGen/CGHLSLRuntime.h
index 7c6c285..103b4a9 100644
--- a/clang/lib/CodeGen/CGHLSLRuntime.h
+++ b/clang/lib/CodeGen/CGHLSLRuntime.h
@@ -135,6 +135,7 @@ public:
   GENERATE_HLSL_INTRINSIC_FUNCTION(BufferUpdateCounter, resource_updatecounter)
   GENERATE_HLSL_INTRINSIC_FUNCTION(GroupMemoryBarrierWithGroupSync,
                                    group_memory_barrier_with_group_sync)
+  GENERATE_HLSL_INTRINSIC_FUNCTION(GetDimensionsX, resource_getdimensions_x)
 
   //===----------------------------------------------------------------------===//
   // End of reserved area for HLSL intrinsic getters.
diff --git a/clang/lib/CodeGen/CodeGenTBAA.cpp b/clang/lib/CodeGen/CodeGenTBAA.cpp
index 4e29d8a..cd08f3e 100644
--- a/clang/lib/CodeGen/CodeGenTBAA.cpp
+++ b/clang/lib/CodeGen/CodeGenTBAA.cpp
@@ -609,8 +609,7 @@ llvm::MDNode *CodeGenTBAA::getValidBaseTypeInfo(QualType QTy) {
   // First calculate the metadata, before recomputing the insertion point, as
   // the helper can recursively call us.
   llvm::MDNode *TypeNode = getBaseTypeInfoHelper(Ty);
-  LLVM_ATTRIBUTE_UNUSED auto inserted =
-      BaseTypeMetadataCache.insert({Ty, TypeNode});
+  [[maybe_unused]] auto inserted = BaseTypeMetadataCache.insert({Ty, TypeNode});
   assert(inserted.second && "BaseType metadata was already inserted");
 
   return TypeNode;
diff --git a/clang/lib/CodeGen/Targets/AMDGPU.cpp b/clang/lib/CodeGen/Targets/AMDGPU.cpp
index 0fcbf7e..16d5919 100644
--- a/clang/lib/CodeGen/Targets/AMDGPU.cpp
+++ b/clang/lib/CodeGen/Targets/AMDGPU.cpp
@@ -402,6 +402,26 @@ void AMDGPUTargetCodeGenInfo::setFunctionDeclAttributes(
 
     F->addFnAttr("amdgpu-max-num-workgroups", AttrVal.str());
   }
+
+  if (auto *Attr = FD->getAttr<CUDAClusterDimsAttr>()) {
+    auto GetExprVal = [&](const auto &E) {
+      return E ? E->EvaluateKnownConstInt(M.getContext()).getExtValue() : 1;
+    };
+    unsigned X = GetExprVal(Attr->getX());
+    unsigned Y = GetExprVal(Attr->getY());
+    unsigned Z = GetExprVal(Attr->getZ());
+    llvm::SmallString<32> AttrVal;
+    llvm::raw_svector_ostream OS(AttrVal);
+    OS << X << ',' << Y << ',' << Z;
+    F->addFnAttr("amdgpu-cluster-dims", AttrVal.str());
+  }
+
+  // OpenCL doesn't support cluster feature.
+  const TargetInfo &TTI = M.getContext().getTargetInfo();
+  if ((IsOpenCLKernel &&
+       TTI.hasFeatureEnabled(TTI.getTargetOpts().FeatureMap, "clusters")) ||
+      FD->hasAttr<CUDANoClusterAttr>())
+    F->addFnAttr("amdgpu-cluster-dims", "0,0,0");
 }
 
 void AMDGPUTargetCodeGenInfo::setTargetAttributes(