exposed processor type, unsupported operations, intelppm.sys and windows crashs

exposed processor type, unsupported operations, intelppm.sys and windows crashs

[Gildas]

Hi,

I had a problem when moving my windows XP machine from the ACPI HAL to
the standard PC HAL in order to get better performances (as pointed in
http://kvm.qumranet.com/kvmwiki/Windows_ACPI_Workaround): on reboot
the machine would just BSOD with kvm but work with -no-kvm.

The BSOD message was about intelppm.sys, so I did my homework,
searched on Google and I found a page on Ben Armstrong's Blog:
http://blogs.msdn.com/virtual_pc_guy/archive/2005/10/24/484461.aspx

The very interesting part of the article is:

"In the mean time we have made some subtle changes to the way our
hardware exposes the processor in Virtual Server R2 so that in future
products these drivers should never get loaded inside of virtual
machines."

Should similar tricks be performed in kvm?

Btw, in case you wonder, changing the value of the key Start to 4 in both
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Processor
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\Intelppm
worked for me and the virtual machine is way quicker now.

Cheers,
Gildas

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: exposed processor type, unsupported operations, intelppm.sys and windows crashs

[Izik Eidus]

Gildas wrote:
> Hi,
>
> I had a problem when moving my windows XP machine from the ACPI HAL to
> the standard PC HAL in order to get better performances (as pointed in
> http://kvm.qumranet.com/kvmwiki/Windows_ACPI_Workaround): on reboot
> the machine would just BSOD with kvm but work with -no-kvm.
>
> The BSOD message was about intelppm.sys, so I did my homework,
> searched on Google and I found a page on Ben Armstrong's Blog:
> http://blogs.msdn.com/virtual_pc_guy/archive/2005/10/24/484461.aspx
>
> The very interesting part of the article is:
>
> "In the mean time we have made some subtle changes to the way our
> hardware exposes the processor in Virtual Server R2 so that in future
> products these drivers should never get loaded inside of virtual
> machines."
>
> Should similar tricks be performed in kvm?
>   
kvm have patch that we wrote that let us play with how the cpu expose 
itself by the meaning of the cpuid instruction.
but i believe it is something else ther.e


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/