Date: Tue, 11 Aug 1998 23:23:27 +0200 From: Mats Lofkvist <mal@algonet.se> To: luoqi@watermarkgroup.com Cc: freebsd-smp@FreeBSD.ORG Subject: Re: SMP + Wine => panic Message-ID: <199808112123.XAA09082@kairos> In-Reply-To: <199808102126.RAA23228@lor.watermarkgroup.com> (message from Luoqi Chen on Mon, 10 Aug 1998 17:26:52 -0400 (EDT)) References: <199808102126.RAA23228@lor.watermarkgroup.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Date: Mon, 10 Aug 1998 17:26:52 -0400 (EDT)
From: Luoqi Chen <luoqi@watermarkgroup.com>
> I got a kernel panic when trying to run wine on my SMP current system
> (cvsup'ped July 21, 2xPpro233 on Providence mb, smp+softupdates).
>
> Could it be the USER_LDT stuff not working with SMP?
>
> I have appended the stack trace, if someone wants more info just ask.
>
I doubt it's SMP related. It looked like the page fault occured when
trying to access the user ldt table. Could you look at pcb->pcb_ldt
(long word at 0xf444801c)? Is it 0? You may want to try the patch
I submitted in PR i386/6219 (use the patch in my own follow up, there're
some error in the original submission).
(kgdb) print/x *0xf444801c
$1 = 0x0
If I turn off the second processor with machdep.smp_active=0,
I do _not_ get any panics.
I also tried the pr 6219 fix, doesn't seem to make any difference.
When trying it a few times I noted that the current process is
not wine, but some other process like bash or xclock. Is wine
doing something (e.g. via user ldt) to one of the processors
which is not kept track of when wine wanders to the other cpu?
(I also had problems getting a dump after applying the pr 6219 fix,
the system hangs when trying to sync the disks. Is there a command
that goes more directly to the dumping than 'panic'?)
_
Mats Lofkvist
mal@algonet.se
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199808112123.XAA09082>