[Bug tree-optimization/126366] Re-alignment for early break ineffective on x86

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126366

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tnfchris at gcc dot gnu.org
             Blocks|                            |115130

--- Comment #1 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Note that in vect_enhance_data_refs_alignment today we do

  /* Check if we can possibly peel the loop.  */
  if (!vect_can_advance_ivs_p (loop_vinfo)
      || !slpeel_can_duplicate_loop_p (loop, LOOP_VINFO_MAIN_EXIT (loop_vinfo),
                                       loop_preheader_edge (loop))
      || loop->inner
      /* We don't currently maintain the LCSSA for prologue peeled inversed
         loops.  */
      || (LOOP_VINFO_EARLY_BREAKS_VECT_PEELED (loop_vinfo)
          && !LOOP_VINFO_NITERS_UNCOUNTED_P (loop_vinfo)))
    do_peeling = false;

so peeling is disabled for LOOP_VINFO_EARLY_BREAKS_VECT_PEELED but I don't know
why it still is given LOOP_VINFO_NITERS_UNCOUNTED_P are all also
LOOP_VINFO_EARLY_BREAKS_VECT_PEELED so the issue must have been fixed already.

I'll remove this and see.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=115130
[Bug 115130] [meta-bug] early break vectorization