================
@@ -794,4 +794,78 @@ TEST_F(AArch64TargetInfoTest,
ClassifyArgumentOveralignedHFAAlign) {
}
}
+// Empty records and zero-size types are ignored as returns under all AArch64
----------------
smithp35 wrote:
I know this is existing clang behaviour but it looks wrong for C++ by the
definition in:
https://github.com/ARM-software/abi-aa/blob/main/cppabi64/cppabi64.rst#41summary-of-differences-from-and-additions-to-the-generic-c-abi
> For the purposes of parameter passing in
> [[AAPCS64](https://github.com/ARM-software/abi-aa/releases)], a parameter
> whose type is an empty class shall be treated as if its type were an
> aggregate with a single member of type unsigned byte.
and
https://github.com/ARM-software/abi-aa/blob/main/aapcs64/aapcs64.rst#69result-return
> If the type, T, of the result of a function is such that
>
> void func(T arg)
> would require that arg be passed as a value in a register (or set of
> registers) according to the rules in [Parameter
> passing](https://github.com/ARM-software/abi-aa/blob/main/aapcs64/aapcs64.rst#parameter-passing),
> then the result is returned in the same registers as would be used for such
> an argument.
My interpretation there is that if an Empty type is treated as if it was an
aggregate of a single member of type unsigned byte, then it would be returned
in w0.
>From what I can tell GCC does return a value for Empty types. For example for
```
struct Empty {};
Empty ret_empty() { return {}; }
```
I get from GCC
```
.global _Z9ret_emptyv
.type _Z9ret_emptyv, %function
_Z9ret_emptyv:
.LFB0:
.cfi_startproc
mov w0, 0
ret
.cfi_endproc
.LFE0:
.size _Z9ret_emptyv, .-_Z9ret_emptyv
```
but clang has just got the `ret`.
I think you've got no choice but follow existing clang behaviour here though.
I'll see if I can chase up internally if this has come up before.
https://github.com/llvm/llvm-project/pull/224476
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
