================
@@ -201,22 +201,22 @@ int call_wide_char(Pair2 p, WideChar w) { return vf(p,
w); }
// declared parameter.
int ell_bitint17(Pair2 p, _BitInt(17) b) { return vf(p, b); }
-// CIR-LABEL: cir.func {{.*}}@ell_bitint17(%arg0: !u64i loc({{.+}}), %arg1:
!cir.int<s, 17, bitint> {llvm.signext} loc({{.+}})) -> !s32i
+// CIR-LABEL: cir.func {{.*}}@ell_bitint17(%arg0: !u64i loc({{.+}}), %arg1:
!cir.int<s, 17, bitint> {llvm.noundef, llvm.signext} loc({{.+}})) -> !s32i
----------------
erichkeane wrote:
These changes are interesting... Apparently by getting these 'correct' we start
emitting the 'noundef' because there is a 'full size' to the bitints? It isn't
perfectly clear to me, but this seems to put us in alignment(hah!) with LLVM-IR.
https://github.com/llvm/llvm-project/pull/219487
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
