https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103010
Bug ID: 103010
Summary: Extra move to x0 for non-POD returns
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: enhancement
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: pinskia at gcc dot gnu.org
Target Milestone: ---
Take the following C++ code:
void ll() noexcept(true);
struct k {
~k() noexcept(true);
};
k ice(k *a)
{
k v;
ll();
return v;
}
---- CUT ----
Currently GCC produces:
stp x29, x30, [sp, -32]!
mov x29, sp
str x19, [sp, 16]
mov x19, x8
bl _Z2llv
mov x0, x19
ldr x19, [sp, 16]
ldp x29, x30, [sp], 32
But the ABI does say x0 need to be set here at all.
So really the call to _Z2llv could even be a tail call.