https://gcc.gnu.org/bugzilla/show_bug.cgi?id=123657
--- Comment #7 from Victor Do Nascimento <victorldn at gcc dot gnu.org> ---
in `vect_analyze_loop_form', we have:
if (!expr_invariant_in_loop_p (loop, inner.number_of_iterations))
return opt_result::failure_at (vect_location,
"not vectorized: inner-loop count not"
" invariant.\n");
where `expr_invariant_in_loop_p (loop, inner.number_of_iterations))' evaluates
to `true' when `inner.number_of_iterations == chrec_dont_know'.
This leads to the incorrect assumption that outer loop vectorization is okay in
this situation.
I've got a patch for the fix and I'm running the tests for regressions.
