https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125838
--- Comment #9 from Iain Sandoe <iains at gcc dot gnu.org> ---
(In reply to Iain Sandoe from comment #6)
> (In reply to Alfie Richards from comment #5)
> Do we have a test case that exercises that problematical scenario?
If we do:
int foo [[gnu::preserve_none]] (int, ...);
....
__attribute__((noipa, noinline))
int foo [[VEND::preserve_none]] (int a, ...)
{
va_list args;
va_start(args, a);
...
}
It seems we then generate wrong code .. so we do have something of a similar
issue to the original clang one.
