[Bug bootstrap/21168] bootstrap failed on Linux/ia64

nathan at codesourcery dot com gcc-bugzilla@gcc.gnu.org
Sat Apr 23 08:47:00 GMT 2005


------- Additional Comments From nathan at codesourcery dot com  2005-04-23 08:47 -------
Subject: Re:  New: bootstrap failed on Linux/ia64

hjl at lucon dot org wrote:
> This patch
> 
> http://gcc.gnu.org/ml/gcc-patches/2005-04/msg02301.html
> 
> caused
> 
> /net/gnu-9/export/gnu/src/gcc-next/gcc/gcc/config/ia64/crtfastmath.c: In
> function \uffff\uffff\uffff__ia64_set_fast_math
> /net/gnu-9/export/gnu/src/gcc-next/gcc/gcc/config/ia64/crtfastmath.c:37:
> internal compiler error: in schedule_block, at haifa-sched.c:2111

oops.  de morgan snafu.  I must have been confused by the original form
which used inconsistent != notation :)  It used to be
!   if (current_sched_info->queue_must_finish_empty && q_size != 0)
!       abort ();
and this patch correctly implements that as
!   gcc_assert (!current_sched_info->queue_must_finish_empty || !q_size);

Fixed with the attached patch, committed as obvious

nathan

2005-04-23  Nathan Sidwell  <nathan@codesourcery.com>

	* haifa-sched.c (schedule_block): Fix thinko in previous
	assertification patch.

Index: haifa-sched.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/haifa-sched.c,v
retrieving revision 1.254
diff -c -3 -p -r1.254 haifa-sched.c
*** haifa-sched.c	22 Apr 2005 16:14:55 -0000	1.254
--- haifa-sched.c	23 Apr 2005 08:43:31 -0000
*************** schedule_block (int b, int rgn_n_insns)
*** 2108,2114 ****
  
    /* Sanity check -- queue must be empty now.  Meaningless if region has
       multiple bbs.  */
!   gcc_assert (!current_sched_info->queue_must_finish_empty || q_size);
  
    /* Update head/tail boundaries.  */
    head = NEXT_INSN (prev_head);
--- 2108,2114 ----
  
    /* Sanity check -- queue must be empty now.  Meaningless if region has
       multiple bbs.  */
!   gcc_assert (!current_sched_info->queue_must_finish_empty || !q_size);
  
    /* Update head/tail boundaries.  */
    head = NEXT_INSN (prev_head);


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21168



More information about the Gcc-bugs mailing list