Stabilization on the ARM: an unrolling bug

[Michael Hayes]

Jeffrey A Law writes:
 > I went ahead and installed this.  I need to get some additional
 > loop unrolling tests done :-)

I'm not sure I'm looking forward to this ;-(

Anyway, I was unable to install the patch last night since the CVS
server wouldn't talk to me.  I suggest we also add the following
comment.

Index: unroll.c
===================================================================
RCS file: /egcs/carton/cvsfiles/egcs/gcc/unroll.c,v
retrieving revision 1.43
diff -c -3 -p -r1.43 unroll.c
*** unroll.c	1999/01/15 10:05:56	1.43
--- unroll.c	1999/01/16 03:06:33
*************** loop_iterations (loop_start, loop_end, l
*** 3497,3504 ****
        return 0;
      }
  
!   /* If there is a more than a single jump to the top of the loop
!      we cannot (easily) determine the iteration count.  */
    if (LABEL_NUSES (JUMP_LABEL (last_loop_insn)) > 1)
      {
        if (loop_dump_stream)
--- 3497,3507 ----
        return 0;
      }
  
!   /* If there is a more than a single jump to the top of the loop we
!      cannot (easily) determine the iteration count.  Note that the
!      jump optimizer will not coalesce labels separated by a loop note.
!      Thus the label at the top of the loop is only reached from back
!      edges.  */
    if (LABEL_NUSES (JUMP_LABEL (last_loop_insn)) > 1)
      {
        if (loop_dump_stream)