[DebugInfo] Make DISubprogram's hashing always produce the same result

A DISubprogram's hashing algorithm takes into account its Scope. A Scope
can be a temporary though which can be replaced later on during
compilation. This means that the hashing algorithm for a DISubprogram
could produce a different hash before/after the Scope has changed. Fix
this by checking the Scope's linkage name instead, which should always
be the same.

rdar://127004707
(cherry picked from commit 0f47349)

Author: augusto2112

llvm/lib/IR/LLVMContextImpl.h

@@ -843,19 +843,26 @@ template <> struct MDNodeKeyImpl<DISubprogram> {
   bool isDefinition() const { return SPFlags & DISubprogram::SPFlagDefinition; }
 
   unsigned getHashValue() const {
+    // Use the Scope's linkage name instead of using the scope directly, as the
+    // scope may be a temporary one which can replaced, which would produce a
+    // different hash for the same DISubprogram.
+    llvm::StringRef ScopeLinkageName;
+    if (auto *CT = dyn_cast_or_null<DICompositeType>(Scope))
+      if (auto *ID = CT->getRawIdentifier())
+        ScopeLinkageName = ID->getString();
+
     // If this is a declaration inside an ODR type, only hash the type and the
     // name.  Otherwise the hash will be stronger than
     // MDNodeSubsetEqualImpl::isDeclarationOfODRMember().
-    if (!isDefinition() && LinkageName)
-      if (auto *CT = dyn_cast_or_null<DICompositeType>(Scope))
-        if (CT->getRawIdentifier())
-          return hash_combine(LinkageName, Scope);
+    if (!isDefinition() && LinkageName &&
+        isa_and_nonnull<DICompositeType>(Scope))
+      return hash_combine(LinkageName, ScopeLinkageName);
 
     // Intentionally computes the hash on a subset of the operands for
     // performance reason. The subset has to be significant enough to avoid
     // collision "most of the time". There is no correctness issue in case of
     // collision because of the full check above.
-    return hash_combine(Name, Scope, File, Type, Line);
+    return hash_combine(Name, ScopeLinkageName, File, Type, Line);
   }
 };
 

llvm/unittests/IR/DebugInfoTest.cpp

@@ -7,6 +7,7 @@
 //===----------------------------------------------------------------------===//
 
 #include "llvm/IR/DebugInfo.h"
+#include "../lib/IR/LLVMContextImpl.h"
 #include "llvm/ADT/APSInt.h"
 #include "llvm/AsmParser/Parser.h"
 #include "llvm/IR/DIBuilder.h"
@@ -19,6 +20,7 @@
 #include "llvm/IR/Verifier.h"
 #include "llvm/Support/SourceMgr.h"
 #include "llvm/Transforms/Utils/Local.h"
+
 #include "gtest/gtest.h"
 
 using namespace llvm;
@@ -231,7 +233,7 @@ TEST(DbgVariableIntrinsic, EmptyMDIsKillLocation) {
   EXPECT_TRUE(DbgDeclare->isKillLocation());
 }
 
-TEST(DIBuiler, CreateFile) {
+TEST(DIBuilder, CreateFile) {
   LLVMContext Ctx;
   std::unique_ptr<Module> M(new Module("MyModule", Ctx));
   DIBuilder DIB(*M);
@@ -730,4 +732,51 @@ TEST(AssignmentTrackingTest, InstrMethods) {
   }
 }
 
+// Test that the hashing function for DISubprograms produce the same result
+// after replacing the temporary scope.
+TEST(DIBuilder, HashingDISubprogram) {
+  LLVMContext Ctx;
+  std::unique_ptr<Module> M = std::make_unique<Module>("MyModule", Ctx);
+  DIBuilder DIB(*M);
+
+  DIFile *F = DIB.createFile("main.c", "/");
+  DICompileUnit *CU =
+      DIB.createCompileUnit(dwarf::DW_LANG_C, F, "Test", false, "", 0);
+
+  llvm::TempDIType ForwardDeclaredType =
+      llvm::TempDIType(DIB.createReplaceableCompositeType(
+          llvm::dwarf::DW_TAG_structure_type, "MyType", CU, F, 0, 0, 8, 8, {},
+          "UniqueIdentifier"));
+
+  // The hashing function is different for declarations and definitions, so
+  // create one of each.
+  DISubprogram *Declaration =
+      DIB.createMethod(ForwardDeclaredType.get(), "MethodName", "LinkageName",
+                       F, 0, DIB.createSubroutineType({}));
+
+  DISubprogram *Definition = DIB.createFunction(
+      ForwardDeclaredType.get(), "MethodName", "LinkageName", F, 0,
+      DIB.createSubroutineType({}), 0, DINode::FlagZero,
+      llvm::DISubprogram::SPFlagDefinition, nullptr, Declaration);
+
+  // Produce the hash with the temporary scope.
+  unsigned HashDeclaration =
+      MDNodeKeyImpl<DISubprogram>(Declaration).getHashValue();
+  unsigned HashDefinition =
+      MDNodeKeyImpl<DISubprogram>(Definition).getHashValue();
+
+  // Instantiate the real scope and replace the temporary one with it.
+  DICompositeType *Type = DIB.createStructType(CU, "MyType", F, 0, 8, 8, {}, {},
+                                               {}, 0, {}, "UniqueIdentifier");
+  DIB.replaceTemporary(std::move(ForwardDeclaredType), Type);
+
+  // Now make sure the hashing is consistent.
+  unsigned HashDeclarationAfter =
+      MDNodeKeyImpl<DISubprogram>(Declaration).getHashValue();
+  unsigned HashDefinitionAfter =
+      MDNodeKeyImpl<DISubprogram>(Definition).getHashValue();
+
+  EXPECT_EQ(HashDeclaration, HashDeclarationAfter);
+  EXPECT_EQ(HashDefinition, HashDefinitionAfter);
+}
 } // end namespace