[Bug tree-optimization/116133] New: Missing mult_overflow detection for aarch64

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116133

            Bug ID: 116133
           Summary: Missing mult_overflow detection for aarch64
           Product: gcc
           Version: 15.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: enhancement
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pinskia at gcc dot gnu.org
  Target Milestone: ---
            Target: aarch64-linux-gnu

Using the testcase from PR 95852:
```
bool f(unsigned x, unsigned y, unsigned* res)
{
    *res = x * y;
    return x && ((*res / x) != y);
}
```

This should be optimized to just:
```
  _14 = (int) x_7(D);
  _15 = (int) y_8(D);
  _16 = .MUL_OVERFLOW (_14, _15);
  _17 = REALPART_EXPR <_16>;
  _1 = (unsigned int) _17;
  _18 = IMAGPART_EXPR <_16>;
  *res_10(D) = _1;
  _12 = _18 != 0;
  _13 = (int) _12;
```

Like it is on x86_64 But it looks like widening multiply is not considered if
MUL_OVERFLOW will be better than the original IR.