public inbox for gcc@gcc.gnu.org
 help / color / mirror / Atom feed
* GCC pending_chain contains only (nil)
@ 2004-09-05  9:30 Hsin-Hsien Wu
  2004-09-06 21:38 ` James E Wilson
  0 siblings, 1 reply; 3+ messages in thread
From: Hsin-Hsien Wu @ 2004-09-05  9:30 UTC (permalink / raw)
  To: gcc

Hello,
        I am currently porting gcc to a new target. After I write my own
machine description,
I got a ICE (segmentation fault) . So, I use GDB to debug, and what I got is
shown bellow,

(gdb) l emit_queue
502
503     /* Perform all the pending incrementations.  */
504
505     void
506     emit_queue ()
507     {
508       rtx p;
509       while ((p = pending_chain))
510         {
511           rtx body = QUEUED_BODY (p);
(gdb) l
512
513           switch (GET_CODE (body))                     <------ 
*SEGEMENTATION FLAULT
514             {
515             case INSN:
516             case JUMP_INSN:
517             case CALL_INSN:
518             case CODE_LABEL:
519             case BARRIER:
520             case NOTE:
521               QUEUED_INSN (p) = body;
(gdb) call debug_rtx(p)
(queued:SI (mem/s:SI (reg:SI 53) [20 <variable>.count+0 S4 A32])
    (nil)
    (nil)
    (nil)
    (nil))
(gdb)

The pending_chain contains only (nil).

What would cause the pending_chain containg only (nil) ?



Thanks

Mike Wu

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: GCC pending_chain contains only (nil)
  2004-09-05  9:30 GCC pending_chain contains only (nil) Hsin-Hsien Wu
@ 2004-09-06 21:38 ` James E Wilson
  2004-09-07  8:17   ` Paolo Bonzini
  0 siblings, 1 reply; 3+ messages in thread
From: James E Wilson @ 2004-09-06 21:38 UTC (permalink / raw)
  To: Hsin-Hsien Wu; +Cc: gcc

Hsin-Hsien Wu wrote:
> The pending_chain contains only (nil).
> What would cause the pending_chain containg only (nil) ?

If pending_chain contained only nil, then you couldn't have reached
here.   The problem here is that pending_chain contains a queue with
QUEUED_BODY nil.

pending_chain is only set in a few places, and the few places that set
it are only called from a few places.  It should be easy to figure out
which one is causing the problem.  Use grep to find these places, and
put breakpoints there.  I'd guess that one of your md file patterns is
failing.
-- 
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: GCC pending_chain contains only (nil)
  2004-09-06 21:38 ` James E Wilson
@ 2004-09-07  8:17   ` Paolo Bonzini
  0 siblings, 0 replies; 3+ messages in thread
From: Paolo Bonzini @ 2004-09-07  8:17 UTC (permalink / raw)
  To: James E Wilson; +Cc: gcc

> pending_chain is only set in a few places, and the few places that set
> it are only called from a few places.  It should be easy to figure out
> which one is causing the problem.  Use grep to find these places, and
> put breakpoints there.  I'd guess that one of your md file patterns is
> failing.

Also, it is gone in mainline GCC.  If it is an option, consider updating 
your port to 3.5 as some complicated bits of RTL have gone away (for 
example RTX_UNCHANGING_P).

Paolo

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2004-09-07  8:17 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2004-09-05  9:30 GCC pending_chain contains only (nil) Hsin-Hsien Wu
2004-09-06 21:38 ` James E Wilson
2004-09-07  8:17   ` Paolo Bonzini

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).