Hi.
I am getting a kernel panic when loading the kvm-amd module. I have talked a
little about this with aliguori on IRC and according to him the panic occurs
when KVM enables EFER.SVME.
Computer information:
AMD Athlon64 X2 4400+, 65nm, Brisbane.
Abit KN9 motherboard (nforce4)
Ubuntu 7.04 x86_6
Pär Andersson wrote:
> Hi.
>
> I am getting a kernel panic when loading the kvm-amd module. I have talked a
> little about this with aliguori on IRC and according to him the panic occurs
> when KVM enables EFER.SVME.
>
I find this very odd. Setting EFER.SVME shouldn't do any harm even if
it
Anthony Liguori wrote:
> Pär Andersson wrote:
>
>> Hi.
>>
>> I am getting a kernel panic when loading the kvm-amd module. I have talked a
>> little about this with aliguori on IRC and according to him the panic occurs
>> when KVM enables EFER.SVME.
>>
>>
>
> I find this very odd. Sett
SVM looks available on your machine. If possible, could you provide dmidecode
information (type "dmidecode") for further verification? I know that some
vendors disabled SVM feature in BIOS last year. But this was pretty rare. Maybe
you can check whether BIOS has an option for SVM.
-Wei
Avi Ki
Huang2, Wei wrote:
> SVM looks available on your machine. If possible, could you provide dmidecode
> information (type "dmidecode") for further verification? I know that some
> vendors disabled SVM feature in BIOS last year. But this was pretty rare.
> Maybe you can check whether BIOS has an opt
As far as I know, there is no special way to detect BIOS-disabled SVM. BIOS
should always trap #GP on writing SVME bit. So the way you suggested should
work.
Pär, are you sure that EFER.SVME bit has been set when the error happened? Our
architecture design team confirmed that enabling EFER.SVME
Huang2, Wei wrote:
> As far as I know, there is no special way to detect BIOS-disabled SVM. BIOS
> should always trap #GP on writing SVME bit. So the way you suggested should
> work.
>
>
Okay. How does the bios disable svm?
--
Do not meddle in the internals of kernels, for they are subtle
On Tuesday 29 May 2007 20:55:03 you wrote:
> Pär, are you sure that EFER.SVME bit has been set when the error happened?
> Our architecture design team confirmed that enabling EFER.SVME twice won't
> cause any problem (There will be some performance loss though).
No, I am not sure, this is a bit ab
On Friday 01 June 2007 07:55:32 you wrote:
> Pär,
>
> Your CPU is AMD Athlon Rev G, which should support SVM. I happened to
> have a Rev G CPU at hand. I will install Ubuntu 64 myself tomorrow and
> see whether I can re-produce the problem.
>
> By any chance, could you can try the following tests?
