[clang] [llvm] [clang][X86] Return Win64 __int128 indirectly (PR #210358)
LIU Hao via cfe-commits
cfe-commits at lists.llvm.org
Wed Nov 4 00:29:00 PST 2026
================
@@ -3586,12 +3586,10 @@ ABIArgInfo WinX86_64ABIInfo::classify(QualType Ty, unsigned &FreeSSERegs,
Align, /*AddrSpace=*/getDataLayout().getAllocaAddrSpace(),
/*ByVal=*/false);
- // Mingw64 GCC returns i128 in XMM0. Coerce to v2i64 to handle that.
- // Clang matches them for compatibility.
if (BT->getKind() == BuiltinType::Int128 ||
----------------
lhmouse wrote:
Does it make sense for this to reuse the code path for `__float128`?
These types were on the same path; the split was done in July: https://github.com/llvm/llvm-project/pull/204887/changes#diff-31d7111f07d665289bbea1c70b4db1e48fbaa19292aa0dbcd4e1cd9d6e1bb2d8R3452
The difference is that your change returns `__int128` in memory even for `__vectorcall`.
https://github.com/llvm/llvm-project/pull/210358
More information about the cfe-commits
mailing list