https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93940
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |NEW
Last reconfirmed| |2020-02-26
CC| |marxin at gcc dot gnu.org
Component|middle-end |ipa
Target Milestone|--- |8.4
Summary|crazy codegen bug on Og |[8/9/10 Regression] crazy
|with builtin_constant_p, |codegen bug on Og with
|coverage, and pthread |builtin_constant_p,
| |coverage, and pthread
Ever confirmed|0 |1
--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Take:
using uint16_t = unsigned short;
struct a {
uint16_t b = 0;
};
struct c {
short d;
};
class e {
public:
void f();
void init_session(c);
};
auto htons = [](uint16_t s) {
if (__builtin_constant_p(s)) {
return uint16_t(uint16_t(s >> 8) | uint16_t(s << 8));
}
return uint16_t(uint16_t(s >> 8) | uint16_t(s << 8));
};
struct g {
e h;
void i(a k) {
h.f();
auto j = c();
j.d = htons(k.b);
h.init_session(j);
}
};
void test() {
g().i({});
}
---- CUT ----
This is to remove __builtin_unreachable being emitted from the front-end.
__builtin_unreachable is only added by the inliner.
I have not looked into why though.
