Re: NForce2 pseudoscience stability testing (2.6.0-test11)

From: ross . alexander
Date: Tue Dec 02 2003 - 05:15:13 EST


I upgraded the BIOS about a week ago to 1007. I personally found it to be
stable than 1006. I don't believe it is a problem with my hardware
since it has been stable for long periods of time. I was running the SMP
simply because I (wrongly) presumed a) you needed it to get the IO-APIC
and b) it didn't do any harm.

It is clear that the UP kernel is considerable more stable than the SMP
kernel. This
is a very useful fact since it suggests that it is not a problem with the
IDE device
driver per se. The whole purpose of my testing is to try to determine
which options
increased the stability and hence highlight where the problem could be.

One of the reasons I don't like ACPI is the huge amount of additional
it adds and the amount of stuff it could screw up. Now I have not heard
that any
of the VIA KTxxx based motherboards have any problems. If this is true
then the
problem does not lie with the LAPIC, since that is in the processor, not
the MB.
The fact that it seems to only occur with the NForce2 chipset means it
well be some interrupt coming into the LAPIC from Interrupt Bus. However
I certainly don't claim to be an expert on this so I could well be talking

Conclusion: More testing required.



Ross Alexander "We demand clearly defined
MIS - NEC Europe Limited boundaries of uncertainty and
Work ph: +44 20 8752 3394 doubt."

Alistair John Strachan <s0348365@xxxxxxxxxxxx>
28/11/2003 04:46 p.m.

To: ross.alexander@xxxxxxxxxxxxx, "Brendan Howes"
cc: linux-kernel@xxxxxxxxxxxxxxx
Subject: Re: NForce2 pseudoscience stability testing

On Friday 28 November 2003 15:13, ross.alexander@xxxxxxxxxxxxx wrote:
> The conclusion to this is the problem is in Local APIC with SMP. I'm
> saying this is actually true
> only that is what the data suggests. If anybody wants me to try some
> other stuff feel free to suggest
> ideas.
> Cheers,
> Ross

It's evidently a configuration problem, albeit BIOS, mainboard revision,
memory quality, etc. because I and many others like me are able to run
2.4/2.6 with all the options you tested and still achieve absolute
on the nForce 2 platform.

My system is an EPOX 8RDA+, with an Athlon 2500+ (Barton) overclocked to
2.2Ghz, and 2x256MB TwinMOS PC3200 dimms. FSB is at 400Mhz, and the ram
timings are 4,2,2,2. One might expect such a configuration to be unstable,

but it is not.

I'm currently running 2.6.0-test10-mm1 with full ACPI (+ routing), APIC
local APIC, no preempt, UP, and everything has been rock-solid, despite
machine being under constant 100% CPU load and fairly active IO load.

Also, many others have found that just disabling local apic (and the MPS
setting in the BIOS) as well as ACPI solves their problem, so I'm
that SMP really causes *nForce 2 specific* instability.


personal: alistair()devzero!co!uk
university: s0348365()sms!ed!ac!uk
student: CS/AI Undergraduate
contact: 7/10 Darroch Court,
University of Edinburgh.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at