[Bug tree-optimization/126997] New: [17 Regression] ICE exact_div, at poly-int.h:2179 with SLP vectorisation

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126997

            Bug ID: 126997
           Summary: [17 Regression] ICE exact_div, at poly-int.h:2179 with
                    SLP vectorisation
           Product: gcc
           Version: 17.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ktkachov at gcc dot gnu.org
                CC: rguenth at gcc dot gnu.org
  Target Milestone: ---
            Target: aarch64

unsigned short
f (unsigned int hl, unsigned int scheme, unsigned int n)
{
  unsigned short tot = 0;
  for (unsigned short i = 0; i < n; i++)
    tot += hl == 1 ? (scheme != 0) : (i == 0);
  return tot;
}

ICEs on aarch64 with -O3:
during GIMPLE pass: vect
<source>: In function 'f':
<source>:2:1: internal compiler error: in exact_div, at poly-int.h:2179
    2 | f (unsigned int hl, unsigned int scheme, unsigned int n)
      | ^

Bisection points to g:bb1344be44e364da8377603697933a668b651c4e