================
@@ -1989,6 +1989,14 @@ llvm::Value *CodeGenFunction::EmitLoadOfScalar(Address
Addr, bool Volatile,
return EmitAtomicLoad(AtomicLValue, Loc).getScalarVal();
}
+ if (const auto *BIT = Ty->getAs<BitIntType>()) {
+ if (BIT->getNumBits() > 128) {
+ // Long _BitInt has array of bytes as in-memory type.
+ llvm::Type *NewTy = ConvertType(Ty);
----------------
momchil-velikov wrote:
Oh, I see. It looks close to what we are trying to do with
https://github.com/llvm/llvm-project/pull/93495, which is:
* create in-memory representations according to the target ABI
* improve efficiency of loads/stores, e.g. load/store of `i18` in LLVM must
touch just 3 bytes, so a compiler would emit one 16-bit load and one 8-bit
load, but if `i18` comes from `_BitInt(18)` then a single 32-bit load would
work better.
https://github.com/llvm/llvm-project/pull/91364
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
