Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 27 Dec 2006 15:50:20 -0500
From:      Jan Knepper <jan@digitaldaemon.com>
To:        Mike Meyer <mwm@mired.org>
Cc:        freebsd-isp@freebsd.org, Tim McCullagh <tim@halenet.com.au>, Mario Theodoridis <mario-dated-1167617290.71c9c2@schmut.com>, FreeBSD Hackers <FreeBSD-Hackers@freebsd.org>
Subject:   Re: 6.1-RELEASE / 6.2 Kernel Crash...
Message-ID:  <4592DC8C.4040704@digitaldaemon.com>
In-Reply-To: <17810.54292.139339.466850@bhuda.mired.org>
References:  <45918F6E.90006@digitaldaemon.com>	<004c01c7293b$d5e03b40$6500a8c0@laptopt>	<4591CB3C.1060902@digitaldaemon.com>	<200612261808.08984.mario@schmut.com>	<4592690B.7080800@digitaldaemon.com> <17810.54292.139339.466850@bhuda.mired.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Understood... and exactly as I wrote...
Not impossible, but not that likely...

Thanks!
Jan



Mike Meyer wrote:
> In <4592690B.7080800@digitaldaemon.com>, Jan Knepper <jan@digitaldaemon.com> typed:
>   
>> FreeBSD 5.x branch run on that machine for almost 2 years without a 
>> problem and magically the same time period in *hours* that I upgrade the 
>> machine I get hardware problems too? Not an impossible coincidence, but 
>> not very likely...
>>     
>
> Or it could be that you have a hardware problem in hardware that
> wasn't used by 5.x but is by 6.x.
>
> <Storytime>
> I had an 11/750 that ran BSD 4.2 for years with no problems. When I
> tried to upgrade it to BSD 4.3, it would reliably panic in namei
> during the boot process. We had about a dozen 750s, and this was our
> test machine - so none of them were going to be upgraded until this
> got fixed, deadline or no.
>
> Stepping through namei in the debugger showed that one of the
> instructions in the function prelude was changing a high bit in a
> register it wasn't supposed to touch at all. In 4.2, the register was
> unused, because namei got passed a handful of arguments. In 4.3, it
> got passed a pointer to a struct with some of that information in it,
> and the pointer wound up in said register. Dereferencing the pointer
> caused the panic.
>
> A motherboard replacement solved the problem.
>
> </Storytime>
>
> 	<mike
>   



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4592DC8C.4040704>