Subject: Re: CVS commit: syssrc (turning off idle page zeroing)
To: Erik E. Fair <firstname.lastname@example.org>
From: Matthew Jacob <email@example.com>
Date: 09/14/2000 19:44:02
> At 6:44 -0700 9/12/00, Paul Kranenburg wrote:
> >Module Name: syssrc
> >Committed By: pk
> >Date: Tue Sep 12 13:43:56 UTC 2000
> >Modified Files:
> > syssrc/sys/arch/sparc/sparc [netbsd-1-5]: cpu.c
> >Log Message:
> >Pullup rev. 1.106->1.107 (approved by thorpej):
> >> revision 1.107
> >> date: 2000/09/10 11:32:04; author: pk; state: Exp; lines: +17 -2
> >> Turn off page zeroing in the idle loop; an unidentified bug causes (very
> >> sporadic) user process corruption.
> >(actually applies to Microsparc 1 machines only).
> There are only two microSPARC I CPU systems that Sun ever made: the
> SPARC Classic, and the SPARC LX (well, and the "X terminal"
> configurations of those two). Everything else since is microSPARC II,
> including the Voyager, SS4, and SS5.
> Can't this be shut off at boot after the CPU type is detected (or
> turned on for all CPUs except the microSPARC I)?
You most certainly should be doign page zeroing in the idle loop, too.
And I agree with you on this one. Has anyone picked up the Linux/PPC hack for
doing pagezero's bypassing the cache?