[Bug rtl-optimization/64164] [4.9/5 Regression] one more stack slot used due to one less inlining level

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64164

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
That is, it is good

  # _30 = PHI <0(2), _10(8), value_33(7), value_33(3), value_33(6)>
  return _30;

vs bad

  # _30 = PHI <0(2), _10(8), _33(7), _33(3), _33(6)>
  return _30;

where it thinks that coalescing _30 and _33 is more important than
coalescing _30 and _10 (which looks reasonable).

The question is why RTL opts make such large difference out of this.