[Bug tree-optimization/121772] [13/14/15/16 Regression] aarch64: ICE (invalid gimple) with vcombine since r13-1836-g8a1e05b7618fed

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121772

--- Comment #5 from Alex Coplan <acoplan at gcc dot gnu.org> ---
Here's a fully-preprocessed reduced testcase; not for the testsuite, but just
to enable debugging with a cc1 cross:

typedef __Int16x4_t int16x4_t;
typedef __Int16x8_t int16x8_t;
short f(int16x4_t va) {
  int16x4_t vz = (int16x4_t){};
  int16x8_t vcomb = __builtin_aarch64_combinev4hi(va, vz);
  return __builtin_aarch64_reduc_plus_scal_v8hi(vcomb);
}