[Clang][FMV] Stop emitting implicit default version using target_clones. (#141808)

With the current behavior the following example yields a linker error:
"multiple definition of `foo.default'"

// Translation Unit 1
__attribute__((target_clones("dotprod, sve"))) int foo(void) { return 1; }

// Translation Unit 2
int foo(void) { return 0; }
__attribute__((target_version("dotprod"))) int foo(void);
__attribute__((target_version("sve"))) int foo(void);
int bar(void) { return foo(); }

That is because foo.default is generated twice. As a user I don't find
this particularly intuitive. If I wanted the default to be generated in
TU1 I'd rather write target_clones("dotprod, sve", "default")
explicitly.

When changing the code I noticed that the RISC-V target defers the
resolver emission when encountering a target_version definition. This
seems accidental since it only makes sense for AArch64, where we only
emit a resolver once we've processed the entire TU, and only if the
default version is present. I've changed this so that RISC-V immediately
emmits the resolver. I adjusted the codegen tests since the functions
now appear in a different order.

Implements https://github.com/ARM-software/acle/pull/377

1 files changed, 12 insertions, 13 deletions

diff --git a/clang/lib/CodeGen/CodeGenModule.cpp b/clang/lib/CodeGen/CodeGenModule.cpp
index 6d2c705..9383c57 100644
--- a/clang/lib/CodeGen/CodeGenModule.cpp
+++ b/clang/lib/CodeGen/CodeGenModule.cpp
@@ -4237,19 +4237,19 @@ void CodeGenModule::EmitMultiVersionFunctionDefinition(GlobalDecl GD,
       EmitGlobalFunctionDefinition(GD.getWithMultiVersionIndex(I), nullptr);
   } else if (auto *TC = FD->getAttr<TargetClonesAttr>()) {
     for (unsigned I = 0; I < TC->featuresStrs_size(); ++I)
-      // AArch64 favors the default target version over the clone if any.
-      if ((!TC->isDefaultVersion(I) || !getTarget().getTriple().isAArch64()) &&
-          TC->isFirstOfVersion(I))
+      if (TC->isFirstOfVersion(I))
         EmitGlobalFunctionDefinition(GD.getWithMultiVersionIndex(I), nullptr);
-    // Ensure that the resolver function is also emitted.
-    GetOrCreateMultiVersionResolver(GD);
   } else
     EmitGlobalFunctionDefinition(GD, GV);
 
-  // Defer the resolver emission until we can reason whether the TU
-  // contains a default target version implementation.
-  if (FD->isTargetVersionMultiVersion())
-    AddDeferredMultiVersionResolverToEmit(GD);
+  // Ensure that the resolver function is also emitted.
+  if (FD->isTargetVersionMultiVersion() || FD->isTargetClonesMultiVersion()) {
+    // On AArch64 defer the resolver emission until the entire TU is processed.
+    if (getTarget().getTriple().isAArch64())
+      AddDeferredMultiVersionResolverToEmit(GD);
+    else
+      GetOrCreateMultiVersionResolver(GD);
+  }
 }
 
 void CodeGenModule::EmitGlobalDefinition(GlobalDecl GD, llvm::GlobalValue *GV) {
@@ -4351,7 +4351,7 @@ void CodeGenModule::emitMultiVersionFunctions() {
     };
 
     // For AArch64, a resolver is only emitted if a function marked with
-    // target_version("default")) or target_clones() is present and defined
+    // target_version("default")) or target_clones("default") is defined
     // in this TU. For other architectures it is always emitted.
     bool ShouldEmitResolver = !getTarget().getTriple().isAArch64();
     SmallVector<CodeGenFunction::FMVResolverOption, 10> Options;
@@ -4374,12 +4374,11 @@ void CodeGenModule::emitMultiVersionFunctions() {
             TVA->getFeatures(Feats, Delim);
             Options.emplace_back(Func, Feats);
           } else if (const auto *TC = CurFD->getAttr<TargetClonesAttr>()) {
-            if (IsDefined)
-              ShouldEmitResolver = true;
             for (unsigned I = 0; I < TC->featuresStrs_size(); ++I) {
               if (!TC->isFirstOfVersion(I))
                 continue;
-
+              if (TC->isDefaultVersion(I) && IsDefined)
+                ShouldEmitResolver = true;
               llvm::Function *Func = createFunction(CurFD, I);
               Feats.clear();
               if (getTarget().getTriple().isX86()) {