public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* Fix early inliner ICE
@ 2009-10-22 14:06 Jan Hubicka
  2010-10-27 13:35 ` H.J. Lu
  0 siblings, 1 reply; 2+ messages in thread
From: Jan Hubicka @ 2009-10-22 14:06 UTC (permalink / raw)
  To: gcc-patches

Hi,
in previous patch I mistakely comitted testcase for 40556 that is fixed by this patch.
Problem is that loop doing incremental inlining for fixed number of iterations has
wrong exit condition so we decide on inlining but do not apply in last iteration.

Bootstrapped/regtested x86_64-linux.  I will also update PR info of the previous commit
while comitting this.

	PR tree-optimize/40556                                                                                                                                         
	* ipa-inline.c (cgraph_early_inlining): Fix iterations condition.
Index: ipa-inline.c
===================================================================
*** ipa-inline.c	(revision 153449)
--- ipa-inline.c	(working copy)
*************** cgraph_early_inlining (void)
*** 1589,1598 ****
  
    if (sorrycount || errorcount)
      return 0;
!   while (cgraph_decide_inlining_incrementally (node,
!   					       iterations
! 					       ? INLINE_SIZE_NORECURSIVE : INLINE_SIZE, 0)
! 	 && iterations < PARAM_VALUE (PARAM_EARLY_INLINER_MAX_ITERATIONS))
      {
        timevar_push (TV_INTEGRATION);
        todo |= optimize_inline_calls (current_function_decl);
--- 1589,1598 ----
  
    if (sorrycount || errorcount)
      return 0;
!   while (iterations < PARAM_VALUE (PARAM_EARLY_INLINER_MAX_ITERATIONS)
!          && cgraph_decide_inlining_incrementally (node,
!   					          iterations
! 					          ? INLINE_SIZE_NORECURSIVE : INLINE_SIZE, 0))
      {
        timevar_push (TV_INTEGRATION);
        todo |= optimize_inline_calls (current_function_decl);

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Fix early inliner ICE
  2009-10-22 14:06 Fix early inliner ICE Jan Hubicka
@ 2010-10-27 13:35 ` H.J. Lu
  0 siblings, 0 replies; 2+ messages in thread
From: H.J. Lu @ 2010-10-27 13:35 UTC (permalink / raw)
  To: Jan Hubicka; +Cc: gcc-patches

On Thu, Oct 22, 2009 at 6:30 AM, Jan Hubicka <hubicka@ucw.cz> wrote:
> Hi,
> in previous patch I mistakely comitted testcase for 40556 that is fixed by this patch.
> Problem is that loop doing incremental inlining for fixed number of iterations has
> wrong exit condition so we decide on inlining but do not apply in last iteration.
>
> Bootstrapped/regtested x86_64-linux.  I will also update PR info of the previous commit
> while comitting this.
>
>        PR tree-optimize/40556
>        * ipa-inline.c (cgraph_early_inlining): Fix iterations condition.

This and/or

http://gcc.gnu.org/ml/gcc-patches/2009-10/msg01418.html

caused:

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46194

I can't tell which one since 2 changes were checked in as one
commit.

-- 
H.J.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2010-10-27 10:26 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-10-22 14:06 Fix early inliner ICE Jan Hubicka
2010-10-27 13:35 ` H.J. Lu

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).