[Bug tree-optimization/124902] sum reduction of 2 masked_loads not vectorized

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124902

Drea Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
            Summary|[LoopVectorize][AArch64]    |sum reduction of 2
                   |GCC fails to vectorize      |masked_loads not vectorized
                   |conditional reduction with  |
                   |`if` statement that Clang   |
                   |vectorizes with SVE         |
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2026-04-16
          Component|target                      |tree-optimization

--- Comment #1 from Drea Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

Simplified testcase:
```
unsigned int foo(
    const unsigned * __restrict__ a,
    const unsigned * __restrict__ b,
    const unsigned * __restrict__ c,
    int n
) {
    unsigned sum = 0;
    for (int i = 0; i<n; i++)
    {
        if (a[i] != 0) {
            sum += a[i];
            sum += b[i];
        }
    }
    return sum;
}
```

GCC does not detect this as a sum reduction.

This is because ifcvt does not produce COND_ADD here even though it should.
We get from ifcvt:
  _27 = _4 != 0;
  sum_13 = _4 + sum_19;
...
  sum_15 = _6 + sum_13;
  sum_7 = _27 ? sum_15 : sum_19;

But that is two COND_ADD and ifcvt only tries to detect one COND_ADD here.