Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 04 Jan 2007 06:33:13 -0500
From:      Randall Stewart <rrs@cisco.com>
To:        Ricardo Nabinger Sanchez <rnsanchez@wait4.org>
Cc:        freebsd-current@freebsd.org
Subject:   Re: A stuck system
Message-ID:  <459CE5F9.3020504@cisco.com>
In-Reply-To: <20070103100555.3611b41c.rnsanchez@wait4.org>
References:  <45891FE9.4020700@cisco.com>	<58281AA0-3738-490C-9EA8-7766033713A2@siliconlandmark.com>	<458960F2.9090703@cisco.com>	<200612281756.29949.jhb@freebsd.org>	<4594F282.7080504@cisco.com> <20070103100555.3611b41c.rnsanchez@wait4.org>

next in thread | previous in thread | raw e-mail | index | archive | help
Ricardo Nabinger Sanchez wrote:
> On Fri, 29 Dec 2006 05:48:34 -0500
> Randall Stewart <rrs@cisco.com> wrote:
> 
>> Nope... its just a single port, on-motherboard msk0.
>>
>> It does wake up though if I ping any interface...
>>
>> I suspect it might be a hardware problem.. not sure
>> yet :-0
> 
> How about installing a ping trap in the device driver to generate a dump?
> What I mean is to, whenever the device driver receives a packet, it checks if
> the packet is a special ping packet (with some specific data, like
> "dumpdump..." in the data field), and if so, forces a dump so you can check
> (luckily) where the system came from.
> 
> It's a long shot, but perhaps it gives a hint.
> 
> Does this behavior happens on IA-64 boxes?  If so, the kernel could set up
> the processor to save performance data (specifically the branch history), and
> the special ping (or something else) could be used to print the branch buffer
> history, instead of dumping a core.  Debugging symbols would be a must, I
> believe.
> 
Well... the machine is only a p4d gigabit motherboard...

I am more and more suspecting a hardware problem.

There is a em card in the machine and the motherboard msk card.
The most recent update of the msk card seems to crash the system
at startup.. so I took it out of my load config.. have not
played with it yet..

Previously I could ping the msk net.. and the machine would
wakeup.. now that I don't have the msk card.. guess what.. pinging
the em0 card DOES NOT wake the machine up..

I bet there is some foul-up on the motherboard  causing it to
not deliver interrupts until another one (on the mother board)
comes in... oh well..

R

-- 
Randall Stewart
NSSTG - Cisco Systems Inc.
803-345-0369 <or> 803-317-4952 (cell)



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?459CE5F9.3020504>