[Bug tree-optimization/110080] [13/14 Regression] Missed Dead Code Elimination at -Os when using __builtin_unreachable since r13-6945-g429a7a88438

[pinskia at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110080

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
  a.0_1 = aD.2830;
  # RANGE [irange] unsigned int [0, 255] MASK 0xff VALUE 0x0
  _2 = (unsigned intD.9) a.0_1;
  # RANGE [irange] unsigned char [0, 0][4, 4] MASK 0x4 VALUE 0x0
  _6 = a.0_1 & 4;
  # RANGE [irange] int [0, 0][4, 4] MASK 0x4 VALUE 0x0
  _3 = (intD.6) _6;

Those ranges should have been:
a.0_1 [131,131]
_2 [131,131]
_6 [0,0]
_3 [0,0]

Like we got from the (not) unreachable branch:
=========== BB 4 ============
Imports: a.0_1  _3  
Exports: a.0_1  _3  _6  
a.0_1   [irange] unsigned char [131, 131]
_3      [irange] int [0, 0] MASK 0x4 VALUE 0x0


The store to e was not 131 because it was before the unreachable though:
  e = _3;
  if (a.0_1 != 131)
    goto <bb 3>; [0.00%]
  else
    goto <bb 4>; [100.00%]

  <bb 3> [count: 0]:
  __builtin_unreachable ();