https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113107
Andrew Pinski <pinskia at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |needs-bisection
Status|NEW |RESOLVED
Known to work| |15.0
Target Milestone|--- |15.0
Known to fail| |13.3.0, 14.1.0
Resolution|--- |FIXED
--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Fixed on the trunk by inlining back in the part. And the reason why it is
inlinable now is because the "stack" usage is gone down.
It has gone down because the variable f is no longer part of the split.
So this testcase was fixed for the trunk by r15-4334-g60de5585812f59 .
I came up with a testcase where the array f is still used in the split but GCC
does not then move the load of g_56+0x10 up.
So closing as fixed.
