[clang] [Clang] Fix linker error for function multiversioning (PR #71706)
via cfe-commits
cfe-commits at lists.llvm.org
Mon Nov 13 10:20:48 PST 2023
================
@@ -4098,8 +4098,26 @@ void CodeGenModule::emitMultiVersionFunctions() {
}
llvm::Constant *ResolverConstant = GetOrCreateMultiVersionResolver(GD);
- if (auto *IFunc = dyn_cast<llvm::GlobalIFunc>(ResolverConstant))
+ if (auto *IFunc = dyn_cast<llvm::GlobalIFunc>(ResolverConstant)) {
ResolverConstant = IFunc->getResolver();
+ // In Aarch64, default versions of multiversioned functions are mangled to
+ // their 'normal' assembly name. This deviates from other targets which
+ // append a '.default' string. As a result we need to continue appending
+ // .ifunc in Aarch64.
+ // FIXME: Should Aarch64 mangling for 'default' multiversion function and
+ // in turn ifunc function match that of other targets?
+ if (FD->isTargetClonesMultiVersion() &&
+ !getTarget().getTriple().isAArch64()) {
+ const CGFunctionInfo &FI = getTypes().arrangeGlobalDeclaration(GD);
+ llvm::FunctionType *DeclTy = getTypes().GetFunctionType(FI);
+ std::string MangledName = getMangledNameImpl(
+ *this, GD, FD, /*OmitMultiVersionMangling=*/true);
+ auto *Alias = llvm::GlobalAlias::create(
+ DeclTy, 0, llvm::Function::WeakODRLinkage, MangledName + ".ifunc",
----------------
elizabethandrews wrote:
I looked into this a bit and I think `WeakODR `is incorrect here. I see we call `getMultiversionLinkage` to set linkage of resolver function. I believe this is to account for internal linkage as well. I think that is the right solution here. I am not too familiar with Linkage Types though and so I could be wrong and any input is appreciated.
https://github.com/llvm/llvm-project/pull/71706
More information about the cfe-commits
mailing list