https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85673
--- Comment #2 from Arseny Solokha <asolokha at gmx dot com> ---
The testcase from comment 0 doesn't fail for me anymore w/ the current trunk
snapshots (as of 266255), while the following one does w/ -mavx -O1 (-O2, -O3,
-Ofast, -Og) -fexpensive-optimizations -fschedule-insns -fselective-scheduling
-fno-dce -fno-tree-dce -fno-tree-ter --param selsched-max-lookahead=5:
int
b3 (ks)
{
int ot;
ot = (ks == 1) == (b3 (ks + 1) + 1);
(void) ot;
return ks;
}
But isn't this PR basically a duplicate of PR88070?
