[Bug tree-optimization/123897] [16 Regression] during GIMPLE pass: widening_mul ICE: verify_gimple failed: non-trivial conversion in unary operation

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123897

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
  vector(32) signed charD.20 _13;
  vector(32) signed charD.20 _22;
  vector(32) signed charD.20 _7;

  _7 = _13 * _22;
  _28 = VIEW_CONVERT_EXPR<vector(32) unsigned charD.21>(_7);
..
  _30 = _28 - _29;


But then we do:
  vector(32) signed charD.20 _25;
  vector(32) unsigned charD.21 _29;
  vector(32) unsigned charD.21 _30;

  _25 = -_29;
  _30 = .FMA (_13, _22, _25);