merge main into amd-staging

Author: ronlieb

.ci/generate_test_report_lib.py

@@ -92,7 +92,7 @@ def plural(num_tests):
             ]
         )
     elif failures:
-        report.extend(["", "## Failed Tests", "(click to see output)"])
+        report.extend(["", "## Failed Tests", "(click on a test name to see its output)"])
 
         for testsuite_name, failures in failures.items():
             report.extend(["", f"### {testsuite_name}"])

.ci/generate_test_report_lib_test.py

@@ -2,6 +2,9 @@
 # See https://llvm.org/LICENSE.txt for license information.
 # SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 
+# To run these tests:
+# python -m unittest generate_test_report_lib_test.py
+
 import unittest
 from io import StringIO
 from textwrap import dedent
@@ -150,7 +153,7 @@ def test_report_single_file_single_testsuite(self):
           * 2 tests failed
 
           ## Failed Tests
-          (click to see output)
+          (click on a test name to see its output)
 
           ### Bar
           <details>
@@ -182,7 +185,7 @@ def test_report_single_file_single_testsuite(self):
         * 2 tests failed
 
         ## Failed Tests
-        (click to see output)
+        (click on a test name to see its output)
 
         ### ABC
         <details>

clang/include/clang/Basic/AddressSpaces.h

@@ -59,6 +59,7 @@ enum class LangAS : unsigned {
   // HLSL specific address spaces.
   hlsl_groupshared,
   hlsl_constant,
+  hlsl_private,
 
   // Wasm specific address spaces.
   wasm_funcref,

clang/include/clang/Sema/Sema.h

@@ -4328,6 +4328,7 @@ class Sema final : public SemaBase {
   NamedDecl *findLocallyScopedExternCDecl(DeclarationName Name);
 
   void deduceOpenCLAddressSpace(ValueDecl *decl);
+  void deduceHLSLAddressSpace(VarDecl *decl);
 
   /// Adjust the \c DeclContext for a function or variable that might be a
   /// function-local external declaration.

clang/include/clang/Sema/SemaHLSL.h

@@ -154,6 +154,8 @@ class SemaHLSL : public SemaBase {
   bool TransformInitList(const InitializedEntity &Entity,
                          const InitializationKind &Kind, InitListExpr *Init);
 
+  void deduceAddressSpace(VarDecl *Decl);
+
 private:
   // HLSL resource type attributes need to be processed all at once.
   // This is a list to collect them.

clang/lib/AST/Type.cpp

@@ -94,6 +94,10 @@ bool Qualifiers::isTargetAddressSpaceSupersetOf(LangAS A, LangAS B,
          (A == LangAS::Default &&
           (B == LangAS::cuda_constant || B == LangAS::cuda_device ||
            B == LangAS::cuda_shared)) ||
+         // `this` overloading depending on address space is not ready,
+         // so this is a hack to allow generating addrspacecasts.
+         // IR legalization will be required when this address space is used.
+         (A == LangAS::Default && B == LangAS::hlsl_private) ||
          // Conversions from target specific address spaces may be legal
          // depending on the target information.
          Ctx.getTargetInfo().isAddressSpaceSupersetOf(A, B);

clang/lib/AST/TypePrinter.cpp

@@ -2581,6 +2581,8 @@ std::string Qualifiers::getAddrSpaceAsString(LangAS AS) {
     return "groupshared";
   case LangAS::hlsl_constant:
     return "hlsl_constant";
+  case LangAS::hlsl_private:
+    return "hlsl_private";
   case LangAS::wasm_funcref:
     return "__funcref";
   default:

clang/lib/Basic/TargetInfo.cpp

@@ -47,6 +47,8 @@ static const LangASMap FakeAddrSpaceMap = {
     11, // ptr32_uptr
     12, // ptr64
     13, // hlsl_groupshared
+    14, // hlsl_constant
+    15, // hlsl_private
     20, // wasm_funcref
 };
 

clang/lib/Basic/Targets/AArch64.h

@@ -45,6 +45,7 @@ static const unsigned ARM64AddrSpaceMap[] = {
     static_cast<unsigned>(AArch64AddrSpace::ptr64),
     0, // hlsl_groupshared
     0, // hlsl_constant
+    0, // hlsl_private
     // Wasm address space values for this target are dummy values,
     // as it is only enabled for Wasm targets.
     20, // wasm_funcref

clang/lib/Basic/Targets/AMDGPU.cpp

@@ -60,6 +60,9 @@ const LangASMap AMDGPUTargetInfo::AMDGPUDefIsGenMap = {
     llvm::AMDGPUAS::FLAT_ADDRESS,     // ptr64
     llvm::AMDGPUAS::FLAT_ADDRESS,     // hlsl_groupshared
     llvm::AMDGPUAS::CONSTANT_ADDRESS, // hlsl_constant
+    // FIXME(pr/122103): hlsl_private -> PRIVATE is wrong, but at least this
+    // will break loudly.
+    llvm::AMDGPUAS::PRIVATE_ADDRESS, // hlsl_private
 };
 
 const LangASMap AMDGPUTargetInfo::AMDGPUDefIsPrivMap = {
@@ -85,6 +88,7 @@ const LangASMap AMDGPUTargetInfo::AMDGPUDefIsPrivMap = {
     llvm::AMDGPUAS::FLAT_ADDRESS,     // ptr64
     llvm::AMDGPUAS::FLAT_ADDRESS,     // hlsl_groupshared
     llvm::AMDGPUAS::CONSTANT_ADDRESS, // hlsl_constant
+    llvm::AMDGPUAS::PRIVATE_ADDRESS,  // hlsl_private
 };
 } // namespace targets
 } // namespace clang