[Bug tree-optimization/122419] [16 Regression] ice in vect_analyze_slp_reduc_chain again

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122419

--- Comment #6 from Richard Biener <rguenth at gcc dot gnu.org> ---
t.c:4:10: note:   Analyze phi: a_lsm.6_6 = PHI <_3(7), a_lsm.6_10(3)>
t.c:4:10: note:   reduction path: _3 _2 a_lsm.6_6 
t.c:4:10: note:   reduction: detected reduction

huh.

  # a_lsm.6_6 = PHI <_3(7), a_lsm.6_10(3)>
  _2 = (char) a_lsm.6_6;
  _3 = (int) _2;  

and STMT_VINFO_REDUC_CODE is NOP_EXPR.  OK, somewhat reasonable.  I have
a patch.