https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99739
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|--- |11.0
--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Whether we handle this or not at -O1 is quite dependent on short-circuiting
anyway - we only have jump threading at our hands at -O1 (we could consider
enabling non-iterating EVRP).
So I'm not sure this is an important regression. At -O2 we end up with
<bb 2> [local count: 1073741824]:
_7 = i_2(D) != 0;
_8 = j_3(D) != 0;
_9 = _7 & _8;
if (_9 != 0)
goto <bb 5>; [50.00%]
else
goto <bb 3>; [50.00%]
<bb 3> [local count: 536870911]:
_23 = k_4(D) != 0;
_24 = _7 & _23;
_11 = _8 & _24;
if (_11 != 0)
goto <bb 4>; [94.27%]
else
goto <bb 5>; [5.73%]
<bb 4> [count: 0]:
__builtin_abort ();
<bb 5> [local count: 1073741824]:
return;
where reassociation leaves redundant _9 on the plate.