2003-03-09 22:35:27

by Shawn Starr

[permalink] [raw]
Subject: [BUG][2.5.64bk4] Weird problem with 2 PCs


I don't have an oops/panic to give because the box won't dump one on this
issue.

I have two PCs, an A7M266-D Athlon MP 2000+ and an IBM 300PL 6892-N2U PIII
(Katmai) 450Mhz. If I leave the IBM machine on for a few hours or so, it's
idle and just handling my mail/dns. The problem happens when I turn on the
A7M266-D the other box locks up and reboots (due to panic=80). ksyslog/klogd
aren't dumping any oopes to my logfile so I'm unable to capture the bug.

1. The PCs are connected via a router to my network.

2. The PCs are connected via serial port together (possibly causing the oops
with interrupts?)

Note, when I turn off and on the A7M266-D the IBM will not panic. It's only
when the IBM is powered on for a few hours (if thats any help in narrowing
down things).

When It did panic, It reported __run_timers and a held spinlock by
linux/timer.c but the EIP was 00000000 (garbage).

Anyone else noticing strange things like this? This is really annoying :-(


2003-03-09 22:56:34

by Lionel Bouton

[permalink] [raw]
Subject: Re: [BUG][2.5.64bk4] Weird problem with 2 PCs

Shawn Starr wrote:

>I don't have an oops/panic to give because the box won't dump one on this
>issue.
>
>I have two PCs, an A7M266-D Athlon MP 2000+ and an IBM 300PL 6892-N2U PIII
>(Katmai) 450Mhz. If I leave the IBM machine on for a few hours or so, it's
>idle and just handling my mail/dns. The problem happens when I turn on the
>A7M266-D the other box locks up and reboots (due to panic=80). ksyslog/klogd
>aren't dumping any oopes to my logfile so I'm unable to capture the bug.
>
>1. The PCs are connected via a router to my network.
>
>2. The PCs are connected via serial port together (possibly causing the oops
>with interrupts?)
>
>

Doesn't seem like a software pb to me...
Are the 2 PCs grounded ?
If you own a voltmeter, remove the serial link and try measuring the
tension between the ground pins of the 2 serial interfaces you use for
your serial link. If my guess is right your IBM (or its serial port) is
not correctly grounded and might act like an accumulator that discharges
itself when the other PC is switched on !

>Note, when I turn off and on the A7M266-D the IBM will not panic. It's only
>when the IBM is powered on for a few hours (if thats any help in narrowing
>down things).
>
>

99,9% chances of an hardware problem.

LB.

2003-03-10 02:43:54

by Shawn Starr

[permalink] [raw]
Subject: Re: [BUG][2.5.64bk4] Weird problem with 2 PCs

Can't be, everything is grounded. If i disconnect the serial and reboot the
A7M266-D it doesn't hang the IBM. I can only suspect the serial ground pin is
bad. I will tomorrow try 2.4.2x and see if this occurs. If it does NOT. Then
we have a problem here and finding it is going to be tough :-(

Shawn.

On Sunday 09 March 2003 5:49 pm, Shawn Starr wrote:
> I don't have an oops/panic to give because the box won't dump one on this
> issue.
>
> I have two PCs, an A7M266-D Athlon MP 2000+ and an IBM 300PL 6892-N2U PIII
> (Katmai) 450Mhz. If I leave the IBM machine on for a few hours or so, it's
> idle and just handling my mail/dns. The problem happens when I turn on the
> A7M266-D the other box locks up and reboots (due to panic=80).
> ksyslog/klogd aren't dumping any oopes to my logfile so I'm unable to
> capture the bug.
>
> 1. The PCs are connected via a router to my network.
>
> 2. The PCs are connected via serial port together (possibly causing the
> oops with interrupts?)
>
> Note, when I turn off and on the A7M266-D the IBM will not panic. It's only
> when the IBM is powered on for a few hours (if thats any help in narrowing
> down things).
>
> When It did panic, It reported __run_timers and a held spinlock by
> linux/timer.c but the EIP was 00000000 (garbage).
>
> Anyone else noticing strange things like this? This is really annoying :-(