================
@@ -670,6 +670,20 @@ bool X86TargetLowering::CanLowerReturn(
CallingConv::ID CallConv, MachineFunction &MF, bool isVarArg,
const SmallVectorImpl<ISD::OutputArg> &Outs, LLVMContext &Context,
const Type *RetTy) const {
+ // Mingw64 GCC returns f128 via sret, which matches the documentation of the
+ // Windows x64 calling convention:
+ //
+ //
https://learn.microsoft.com/en-us/cpp/build/x64-calling-convention?view=msvc-170#return-values
+ //
+ // > Otherwise, the caller must allocate memory for the return value and pass
+ // a pointer to it as the first argument.
+ //
+ // Return false, which will perform sret demotion.
+ if (Subtarget.isCallingConvWin64(CallConv) &&
----------------
folkertdev wrote:
I think it's just "is this calling convention ever valid on win64"
https://github.com/llvm/llvm-project/blob/9f578bc595ca7c1a5ce35b17c483fd6c557d11ef/llvm/lib/Target/X86/X86Subtarget.h#L353-L378
But I agree, clearly it's not tested at all either.
https://github.com/llvm/llvm-project/pull/204887
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
