RE: Re: K7 & cpu disconnect

RE: Re: K7 & cpu disconnect

[Grover, Andrew]

> Hi Len,
> 
> one more data point wrt this topic.
> I can enable STPGNT/CPU disconnect on my NForce2 system 
> running WinXP while
> using the tool from http://www.hasw.net/8rdavcore. And guess 
> what, the system
> runs stable for hours, not locking-up at all. Since this is 
> exactly the same
> hardware as well as BIOS configuration, it leads me to 
> beleave the Linux
> ACPI/IO_APIC implementation has an issue.

If the ACPI/IO APIC code has a problem, you should be able to cause the
problem while not enabling the out-of-spec feature. I think you should
feel free to run whatever stress test you can think of, but if the fault
only occurs when you are out of spec, then in my opinion it is asking a
lot to ask people to devote time towards fixing it.

Regards -- Andy


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: Re: K7 & cpu disconnect

[Thomas Herrmann]

> > Hi Len,
> > 
> > one more data point wrt this topic.
> > I can enable STPGNT/CPU disconnect on my NForce2 system 
> > running WinXP while
> > using the tool from http://www.hasw.net/8rdavcore. And guess 
> > what, the system
> > runs stable for hours, not locking-up at all. Since this is 
> > exactly the same
> > hardware as well as BIOS configuration, it leads me to 
> > beleave the Linux
> > ACPI/IO_APIC implementation has an issue.
> 
> If the ACPI/IO APIC code has a problem, you should be able to cause the
> problem while not enabling the out-of-spec feature. I think you should
> feel free to run whatever stress test you can think of, but if the fault
> only occurs when you are out of spec, then in my opinion it is asking a
> lot to ask people to devote time towards fixing it.
> 
> Regards -- Andy
> 
Hi Andy,

is this realy an out-of-spec feature? I mean STPGNT is documented by AMD and
the K7 Barton Core does requests it as a platform/BIOS option.

I just wanted to make aware of an (potential) timing issue in the interrupt
response mechanism of the current ACPI/IO_APIC implementation. The patches
Ross Dickson has posted on the LKML due support this IMHO. Might be it occures
in different scenarios too. Maybe not today, maybe in the future ,who knows.

I'm not asking for a personal favor here. I think you guys have done a great
job getting interrupt routing working fine on the NForce2 & SiS740 platform.

Regards,    Thomas

-- 
+++ GMX - die erste Adresse für Mail, Message, More +++
Bis 31.1.: TopMail + Digicam für nur 29 EUR http://www.gmx.net/topmail



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

Re: K7 & cpu disconnect

[Thomas Herrmann]

Hi Len,

one more data point wrt this topic.

I can enable STPGNT/CPU disconnect on my NForce2 system running WinXP while
using the tool from http://www.hasw.net/8rdavcore. And guess what, the system
runs stable for hours, not locking-up at all. Since this is exactly the same
hardware as well as BIOS configuration, it leads me to beleave the Linux
ACPI/IO_APIC implementation has an issue.

BTW, it seems the acpi-devel list archive on sourceforge.net hangs, no new
entries since 01/24/04.

Regards,    Thomas

-- 
+++ GMX - die erste Adresse für Mail, Message, More +++
Bis 31.1.: TopMail + Digicam für nur 29 EUR http://www.gmx.net/topmail



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

Re: Re: K7 & cpu disconnect

[Arjen Verweij]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=X-UNKNOWN, Size: 2158 bytes --]

Obviously I missed what led up to this message, but are you talking about
your box hanging under heavy io (badblocks, fsck, du etc)? I think most
nforce2 boards have this disconnect function in the bios. Anyway, my box
hangs under heavy io (abit nf7-s barton 2500+) under 2.4.23 with APIC in
the kernel, but not under 2.6.0-test11. In this kernel I have APIC
disabled. If you haven't done so, you kan check lkml archives for messages
by Ross Dickson, he brought this up there.

I think the 2.6.0-test11-mm tree has a patch cycled in with regards to
nforce2 (two actually) I can dig up the URL if you are interested.

Regards,

Arjen

On Thu, 22 Jan 2004, Thomas Herrmann wrote:

> Hi Len,
>
> one more data point wrt this topic.
>
> I can enable STPGNT/CPU disconnect on my NForce2 system running WinXP while
> using the tool from http://www.hasw.net/8rdavcore. And guess what, the system
> runs stable for hours, not locking-up at all. Since this is exactly the same
> hardware as well as BIOS configuration, it leads me to beleave the Linux
> ACPI/IO_APIC implementation has an issue.
>
> BTW, it seems the acpi-devel list archive on sourceforge.net hangs, no new
> entries since 01/24/04.
>
> Regards,    Thomas
>
> --
> +++ GMX - die erste Adresse für Mail, Message, More +++
> Bis 31.1.: TopMail + Digicam für nur 29 EUR http://www.gmx.net/topmail
>
>
>
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
>
>



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn