intermittent ia64 build hang

[Richard Henderson]

On Thu, May 23, 2002 at 04:46:30PM -0700, Janis Johnson wrote:
> If I revert this patch, the build completes (or at least
> it did with yesterday's CVS):
> 
> 2002-05-16  Richard Henderson  <rth@redhat.com>
> 
>         * config/ia64/ia64.c (ia64_reorg): Rebuild bb_for_insn before
>         splitting.  Use split_all_insns; update_life_info_in_dirty_blocks.

Hmm.  Well, I've reverted this section of the patch, since I
know I ran into problems without the splitting changes.


r~


        * config/ia64/ia64.c (ia64_reorg): Use update_life_info instead
        of update_life_info_in_dirty_blocks.

Index: ia64.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/ia64/ia64.c,v
retrieving revision 1.170
diff -c -p -d -u -r1.170 ia64.c
--- ia64.c	23 May 2002 19:23:51 -0000	1.170
+++ ia64.c	24 May 2002 22:51:35 -0000
@@ -6739,8 +6982,9 @@ ia64_reorg (insns)
   if (optimize == 0)
     split_all_insns (0);
 
-  update_life_info_in_dirty_blocks (UPDATE_LIFE_GLOBAL_RM_NOTES,
-				    PROP_DEATH_NOTES);
+  /* ??? update_life_info_in_dirty_blocks fails to terminate during
+     non-optimizing bootstrap.  */
+  update_life_info (NULL, UPDATE_LIFE_GLOBAL_RM_NOTES, PROP_DEATH_NOTES);
 
   if (ia64_flag_schedule_insns2)
     {