Re: [head tinderbox] failure on powerpc/powerpc

From: Kostik Belousov <kostikbel_at_gmail.com>
Date: Wed, 29 Apr 2009 15:48:57 +0300
On Wed, Apr 29, 2009 at 01:20:31PM +0100, Bruce Simpson wrote:
> pluknet wrote:
> >...
> >>/src/sys/kern/sched_ule.c: In function 'sched_idletd':
> >>/src/sys/kern/sched_ule.c:2546: error: dereferencing pointer to 
> >>incomplete type
> >>/src/sys/kern/sched_ule.c:2546: error: 'CG_FLAG_THREAD' undeclared (first 
> >>use in this function)
> >>/src/sys/kern/sched_ule.c:2546: error: (Each undeclared identifier is 
> >>reported only once
> >>/src/sys/kern/sched_ule.c:2546: error: for each function it appears in.)
> >>*** Error code 1
> >>
> >>    
> >
> >This is the only  arch without "options SMP" in GENERIC.
> >Global NOTES also define options SCHED_4BSD and thus
> >does not trigger the error (CG_FLAG_THREAD in SMP scope).
> >
> >  
> +1, I just saw this on a 'make universe' run affecting the ARM targets.
> 
> BMS

The following works for me. The tdq_idled() definition may be #ifdef SMP
too, but I do not see a point. THe CG_FLAG_THREAD is used in sched_idletd(),
and adding more #ifdef SMP would make the code more ugly without any gain.

diff --git a/sys/sys/smp.h b/sys/sys/smp.h
index 05c8642..293f07d 100644
--- a/sys/sys/smp.h
+++ b/sys/sys/smp.h
_at__at_ -16,8 +16,6 _at__at_
 
 #ifndef LOCORE
 
-#ifdef SMP
-
 /*
  * Topology of a NUMA or HTT system.
  *
_at__at_ -57,6 +55,7 _at__at_ struct cpu_group {
 #define	CG_FLAG_SMT	0x02		/* New age htt, less crippled. */
 #define	CG_FLAG_THREAD	(CG_FLAG_HTT | CG_FLAG_SMT)	/* Any threading. */
 
+#ifdef SMP
 /*
  * Convenience routines for building topologies.
  */

Received on Wed Apr 29 2009 - 10:49:05 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:46 UTC