Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 06 Sep 2006 17:33:17 +0100
From:      Gavin Atkinson <gavin.atkinson@ury.york.ac.uk>
To:        Klaus Robert Suetterlin <robert@mpe.mpg.de>
Cc:        freebsd-stable@freebsd.org
Subject:   Re: my kernel spews out a lot of garbage to syslog (long /var/log/messages)
Message-ID:  <1157560397.64218.11.camel@buffy.york.ac.uk>
In-Reply-To: <0D903548-5089-44F1-82BA-6B2D53B91BF4@mpe.mpg.de>
References:  <0D903548-5089-44F1-82BA-6B2D53B91BF4@mpe.mpg.de>

next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, 2006-09-06 at 15:47 +0200, Klaus Robert Suetterlin wrote:
> Hi,
> 
> I have FreeBSD6.1-p5 (PAE+SMP+COMPAT_LINUX+LINPROCFS) system running  
> that is producing a lot of guberish in /var/log/messages through  
> syslog (about 100-1000 bytes per second), I give example below.  Any  
> ideas on this are welcome, especially any hints on diagnostics...
> 
> The giberish started 11:30 local time and still (15:46) continues...
> 
> Here is some example giberish:
> 
> Sep  6 10:48:51 circe kernel: Trying to mount root from ufs:/dev/da0s1a
> Sep  6 11:02:30 circe su: krs to root on /dev/ttyp1
> Sep  6 11:30:19 circe kernel: <NNM2M>NIIMNIM  II IS ISAISA SA  A  
> 302202,00,,  ,
> EE EIIEISSISAASA  Aff fff
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: ff
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: <<22>>f
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: NMNINM MNIIMI  IISSIA AS I AS A32 00,2 , 
> 0 ,EE2 IIE
> S0SIA,AS   fAfEf fIf
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: Sf
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: <<22>>A
> Sep  6 11:30:19 circe kernel:
> Sep  6 11:30:19 circe kernel: ff
> Sep  6 11:30:20 circe kernel: <N<2M2>I> NNIMSMIAIk  I eI3SSrA0An ,   
> eE33l00,I  ,
> SE tIASE ArfI fafSf

It's not exactly garbage, it's four copies of (almost) the same message
being printed at the same time.  What's happening is that you are
receiving an NMI, and all four processors are acting on it.  The default
handler simply prints out a message and ignores the interrupt.  I
believe the actual messages are multiple copies each of:

NMI ISA 30, EISA ff\n
NMI ISA 20, EISA ff\n

Also looks like there's the start of a "kernel trap" message in the
11:30:20 message.

As to what's causing it, it's hard to say, but given they have only just
started, the hardware is trying to tell you something.  My money is on
either overheating or failing RAM (Failed parity checks usually manifest
itself as an NMI, but looking at the code in /usr/src/sys/i386/isa/nmi.c
an attempt is made to try and detect them).  It may be worth trying to
find the rest of that "kernel trap" message.

Gavin



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1157560397.64218.11.camel>