lkml.org 
[lkml]   [2003]   [Sep]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
SubjectFwd: Re: Kernel NMI error
Date
NMI can occur for a number fo reasons:

I/O check from a card
Parity error from memory
DMA timeout
Watchdog timer expiry
INT 2 instruction

To investigate this further you need to look back though your Syslog for any
messages that might indicate intmittant h/w related errors or watchdog
reports.

One way to cause an intermittent NMI is by having power management enabled
with devices that don't support it. They see the world caving in when power
goes away and sometimes generate an NMI. If you're in a power saving process
then you have probably switched to System Management Mode, in which case the
NMI will be held pending the processor returning from SMM.

Do you have a crash dump associtated with this problem?

--
Richard J Moore
IBM Linux Technology Centre

On Tue 16 September 2003 11:38 am, msrinath wrote:
> Hello Everybody,
>
> Can anyone help me on this?
>
> Recently one of our servers running RedHat linux 7.2 with 2.4.7-10 SMP
> kernel generated the following log and the system rebooted. This system has
> 2 CPUs.
>
> Sep 16 01:34:24 cbesc ftpd[30753]: FTP LOGIN FROM 16.128.157.7
> [16.128.157.7], scuser
> Sep 16 01:36:48 cbesc ftpd[30753]: FTP session closed
> Sep 16 01:54:30 cbesc kernel: Uhhuh. NMI received for unknown reason 35.
> Sep 16 01:54:30 cbesc kernel: Dazed and confused, but trying to continue
> Sep 16 01:54:30 cbesc kernel: Do you have a strange power saving mode
> enabled?
> Sep 16 01:54:30 cbesc kernel: eth0: card reports no resources.
> Sep 16 01:58:09 cbesc syslogd 1.4.1: restart.
> Sep 16 01:58:09 cbesc syslog: syslogd startup succeeded
>
> This is the first time we have faced this problem. The ethernet card used
> is intel eepro 100. The details are shown below.
>
> Sep 16 07:33:53 cbesc kernel: eepro100.c:v1.09j-t 9/29/99 Donald Becker
> http://cesdis.gsfc.nasa.gov/linux/drivers/eepro100.html
> Sep 16 07:33:53 cbesc kernel: eepro100.c: $Revision: 1.36 $ 2000/11/17
> Modified by Andrey V. Savochkin <saw@saw.sw.com.sg> and others
> Sep 16 07:33:53 cbesc kernel: eth0: Intel Corporation 82557 [Ethernet Pro
> 100], 00:A0:C9:A0:B7:71, IRQ 16.
> Sep 16 07:33:53 cbesc kernel: Receiver lock-up bug exists -- enabling
> work-around.
> Sep 16 07:33:53 cbesc kernel: Board assembly 668081-004, Physical
> connectors present: RJ45
> Sep 16 07:33:53 cbesc kernel: Primary interface chip i82555 PHY #1.
> Sep 16 07:33:54 cbesc kernel: General self-test: passed.
> Sep 16 07:33:54 cbesc kernel: Serial sub-system self-test: passed.
> Sep 16 07:33:54 cbesc kernel: Internal registers self-test: passed.
> Sep 16 07:33:54 cbesc kernel: ROM checksum self-test: passed
> (0x3c15c8f1). Sep 16 07:33:54 cbesc kernel: Receiver lock-up workaround
> activated.
>
>
> Please let me know why this happened and whether it indicates any hardware
> problem in the system.
>
> Please send a CC to my email address, since I have not subscribed to the
> list.
>
> Thanks & Regards,
>
> - Srinath.

-------------------------------------------------------

--
Richard J Moore
IBM Linux Technology Centre

-------------------------------------------------------

--
Richard J Moore
IBM Linux Technology Centre
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:48    [W:0.037 / U:0.068 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site