[Bug tree-optimization/126919] [17 Regression] ICE in gimple_phi_arg_def_from_edge in gcc/gimpl.h

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126919

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |17.0
          Component|c                           |tree-optimization
                 CC|                            |aldyh at gcc dot gnu.org,
                   |                            |amacleod at redhat dot com
   Last reconfirmed|                            |2026-08-18
             Status|UNCONFIRMED                 |WAITING
           Keywords|                            |ice-on-valid-code
     Ever confirmed|0                           |1
            Summary|ICE in                      |[17 Regression] ICE in
                   |gimple_phi_arg_def_from_edg |gimple_phi_arg_def_from_edg
                   |e in gcc/gimpl.h            |e in gcc/gimpl.h

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
I suspect that there isn't an edge from prev to bb, but w/o preprocessed source
it's hard to say.  Probably prev_bb () is odd?

  basic_block bb = gimple_bb (phi);
  basic_block prev = prev_bb ();
  edge e_in = find_edge (prev, bb);
  // The incoming edge the path supplies is never abnormal, so the
  // argument on it is a valid value for the PHI result even when the
  // result occurs in an abnormal PHI.
  gcc_checking_assert (!(e_in->flags & EDGE_ABNORMAL));
  tree arg = PHI_ARG_DEF_FROM_EDGE (phi, e_in);