https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113431
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #20 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Though, trying that in a cross to arm, with -march=armv9-a -munaligned-access
it matches (in that case I believe vect_hw_misalign should be true), but it
matches even with -march=armv9-a -mno-unaligned-access (and in that case I
think it should be !vect_hw_misaligned target).
That said, sure, if it is
/* { dg-final { scan-tree-dump-times "optimized: basic block part vectorized" 2
"slp1" { target { vect_int && vect_hw_misaligned } } } } */
then it will simply not test anything on the non-vect_hw_misaligned targets,
rather than say XPASS, so maybe that is ok.
Richi, thoughts on that?
