Skip site navigation (1)Skip section navigation (2)
Date:      Sat, 24 Aug 2002 12:49:00 +0930
From:      Greg 'groggy' Lehey <grog@FreeBSD.org>
To:        Patrick O'Reilly <bsd@perimeter.co.za>
Cc:        freebsd-hackers@freebsd.org
Subject:   Re: ! Fatal trap 12: System Crash twice in last 2 days
Message-ID:  <20020824031900.GS87563@wantadilla.lemis.com>
In-Reply-To: <07c801c24ac2$3fb0d6e0$b50d030a@PATRICK>
References:  <025201c24833$3144f550$b50d030a@PATRICK> <20020821052147.GF78608@wantadilla.lemis.com> <07c801c24ac2$3fb0d6e0$b50d030a@PATRICK>

next in thread | previous in thread | raw e-mail | index | archive | help
On Friday, 23 August 2002 at 18:29:21 +0200, Patrick O'Reilly wrote:
> From: "Greg 'groggy' Lehey" <grog@FreeBSD.org>
>
>
>>> ---
>>> Fatal trap 12: page fault while in kernel mode
>>> ---
>>> Unfortunately this means nothing to me :(
>>
>> It doesn't mean much to me by itself.
>>
>> None of this is much help.  You need to build a debug kernel and get a
>> dump and a stack trace, then we can see what's going on.
>>
>> Greg
>
> Greg,
>
> After learning how to do this stuff, I'm back :)
>
> I am now at 10 crashes since Sunday, and counting.
>
> Anyway, I don't know much about gdb, but here is what I was able to get
> from my most recently dumped kernel:
>
> http://docs.perimeter.co.za/crash/20020823_1623_d

Hmm.  This shows a known crash in Vinum.  I've never been able to
reproduce it, so it's possible that it's I/O failure related.  It's
very likely that it's related to this bug report:

  http://www.freebsd.org/cgi/query-pr.cgi?pr=41740

This report contains a patch.  Could you try that, please, and see if
it solves the problem.  I'm particularly interested in crashes with
the kind of backtrace in
http://docs.perimeter.co.za/crash/20020823_1623_d:


#5  0xc026dd47 in trap (frame={tf_fs = 0x10, tf_es = 0x10, tf_ds = 0x10, tf_edi = 0xc139b400, tf_esi = 0x1, 
      tf_ebp = 0xcc970d4c, tf_isp = 0xcc970d24, tf_ebx = 0xc1315a20, tf_edx = 0xc1178800, tf_ecx = 0x0, tf_eax = 0x0, 
      tf_trapno = 0xc, tf_err = 0x0, tf_eip = 0xc10c075d, tf_cs = 0x8, tf_eflags = 0x10286, tf_esp = 0xc1178800, 
      tf_ss = 0xc10b0000}) at /usr/src/sys/i386/i386/trap.c:458
#6  0xc10c075d in launch_requests (rq=0xc122b940, reviveok=0x0)
    at /usr/src/sys/modules/vinum/../../dev/vinum/vinumrequest.c:396
#7  0xc10c0582 in vinumstart (bp=0xc593a348, reviveok=0x0)
    at /usr/src/sys/modules/vinum/../../dev/vinum/vinumrequest.c:291
#8  0xc10c03ea in vinumstrategy (bp=0xc593a348) at
      /usr/src/sys/modules/vinum/../../dev/vinum/vinumrequest.c:164

You've had others, but getting rid of this one would be worthwhile.

> There was a suggestion that the fault might lie in the fact that one
> of my IDE controllers is "falling back to PIO" because I don't have
> the ATA100 compliant cable on device ata1.  I'm hoping to get to the
> server on Monday to replace that.

I'm pretty sure it's not the cable.  We've seen random IDE problems
with 4.x and AMD chipsets, and that's what I currently suspect.  My
guess is that the Vinum problem is triggered by another bug.

Greg
--
See complete headers for address and phone numbers

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20020824031900.GS87563>