[Bug tree-optimization/123755] [16 Regression] gcc.dg/vect/pr71823.c ICEs with -march=armv9-a --param aarch64-autovec-preference=sve-only -ftree-vectorize -O2

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

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123755

Jeffrey A. Law <law at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|FIXED                       |---
             Status|RESOLVED                    |REOPENED

--- Comment #10 from Jeffrey A. Law <law at gcc dot gnu.org> ---
More fallout.  In this case the original fix is causing an ICE in
get_expr_operands.  I'd hoped/expected the followup had resolved everything,
but it didn't.  This is reduc-call-2.c from gcc.target/riscv/rvv/autovec/reduc

-O3 -ffast-math -march=rv32gcbv


double foo (double *a, double *b, double *c)
{
  double result = 0.0;
  for (int i = 0; i < 1024; ++i)
    result += i & 1 ? __builtin_fma (a[i], b[i], c[i]) : 0.0;
  return result;
}