https://gcc.gnu.org/bugzilla/show_bug.cgi?id=119978
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |alias, missed-optimization
--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
h.v = v_3(D);
_12 = &MEM[(struct IntrusiveList *)this_2(D)].node;
MEM[(struct ListNode &)&h].next = _12;
_13 = MEM[(struct IntrusiveList *)this_2(D)].node.prev;
_13->next = &h;
MEM[(struct IntrusiveList *)this_2(D)].node.prev = &h; <--- this one
MEM[(struct ListNode *)_4(D)].prev = _13;
_9 = MEM[(struct ListNode &)&h].next;
MEM[(struct ListNode *)_4(D)].next = _9;
MEM[(struct ListNode &)&h].prev = 0B;
MEM[(struct ListNode &)&h].next = 0B;
It looks like there could be a case where that is still alive at the end of the
function.
There is a missed optimization here where GCC thinks this_2(D) and _4(D) could
point to the same thing but it could not. Note _4(D) is the hidden "return
pointer"
