Re: Fatal trap 12: page fault while in kernel mode

From: Bjoern A. Zeeb <bzeeb-lists_at_lists.zabbadoz.net>
Date: Tue, 22 Jul 2003 18:44:42 +0000 (UTC)
On Tue, 22 Jul 2003, M. Warner Losh wrote:

Hi,

> In message: <Pine.BSF.4.53.0307211602420.45284_at_e0-0.zab2.int.zabbadoz.net>
>             "Bjoern A. Zeeb" <bzeeb-lists_at_lists.zabbadoz.net> writes:
> : On Mon, 21 Jul 2003, Noriyoshi Kawano wrote:
> :
> : > I have similar problem.
> : > disable re-route interrupts.
> : > It's works fine.
> : >
> : > --- /sys/dev/pci/pci.c.orig	Tue Jul  1 23:08:32 2003
> : > +++ /sys/dev/pci/pci.c	Mon Jul 21 11:04:55 2003
> : > _at__at_ -800,7 +800,7 _at__at_
> : >  	}
> : >
> : >  	if (cfg->intpin > 0 && PCI_INTERRUPT_VALID(cfg->intline)) {
> : > -#if defined(__ia64__) || (defined(__i386__) && !defined(SMP))
> : > +#if defined(__ia64__)
> : >  		/*
> : >  		 * Try to re-route interrupts. Sometimes the BIOS or
> : >  		 * firmware may leave bogus values in these registers.
> :
> :
> : Thanks. This works fine. Is there any "global" solution to the problem
> : so that I won't need to patch again the time 5.2R comes out ?
>
> I'm just catching the tail end of this problem report.  What problem
> is solved by not rotuing interrupts?

${SUBJECT}
the machine boots again and doesn't panic at boot time.

for more information please se my initial post.
_at_see http://lists.freebsd.org/pipermail/freebsd-current/2003-July/007155.html


> I'm guessing it is the "we don't setup things correctly to call some
> BIOSes" or "some older pci bioses lie to us so we shoot our selves
> when we believe the lies" or something like that.

I never debugged why this started to happen between 5.1R and HEAD.
The machine is a Compaq DeskPro or s.th. like this; ~ 2 years - so
not that old but perhaps with a "strange" BIOS. It had run fine with
4.x and 5.0R and booted a 5.1R for building HEAD.

Noriyoshi Kawano came up with the patch that made it boot again so I
am happy but as this doesn't seem to happen for many people and the
patch is not mainline I asked if there is any better solution to make
it work on b0rken machines and mainline.

-- 
Bjoern A. Zeeb				bzeeb at Zabbadoz dot NeT
56 69 73 69 74				http://www.zabbadoz.net/
Received on Tue Jul 22 2003 - 09:44:57 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:16 UTC