<p>Thanks for the clarification.</p>
<p>On 17 Aug 2010 17:57, "Pasi Kärkkäinen" <<a href="mailto:pasik@iki.fi">pasik@iki.fi</a>> wrote:<br type="attribution">> On Tue, Aug 17, 2010 at 09:35:24AM +0100, James Hogarth wrote:<br>>> ><br>
>> > Uh.. are you mixing Xen and KVM now?<br>>> ><br>>> > I think KVM *always* requires the kernel module, aka CPU support for hardware virtualization.<br>>> ><br>>> > -- Pasi<br>
>> ><br>>> <br>>> You might be right... having trouble googling something... but I<br>>> thought that kvm without -enable-kvm (or with -no-kvm) and with<br>>> -kernel, -append and -initrd specified could be used for a<br>
>> paravirtualized guest .... However that might just be fully emulated<br>>> on qemu which might or might not help rather than<br>>> paravirtualization...<br>>> <br>> <br>> Without KVM Qemu runs everything on pure software, <br>
> so it's still full virtualization, but damn slow.<br>> <br>>> My apologies if I err ^^ been a while since I had a system without<br>>> available an vmx/svm interface on the CPU....<br>>> <br>
>> I know for a fact that libvirt doesn't support PV guests under KVM...<br>>> virt-install (which uses libvirt) will refuse to try it etc.<br>>> <br>> <br>> Yeah, because PV guests are *Xen* guests. <br>
> To run PV guests you need to be running Xen hypervisor.<br>> <br>>> I'll test it later this week to satisfy my curiosity ;)<br>>> <br>> <br>> "Xenner" is a separate tool that is able to run some Xen PV guests on KVM,<br>
> but it's experimental and it's not developed anymore.<br>> <br>> Xenner requires KVM, and thus CPU virtualization extensions.<br>> <br>> Only way to run Xen PV guests on a hardware without CPU virtualization support<br>
> is to actually use Xen hypervisor :)<br>> <br>> -- Pasi<br>> <br>> _______________________________________________<br>> CentOS-virt mailing list<br>> <a href="mailto:CentOS-virt@centos.org">CentOS-virt@centos.org</a><br>
> <a href="http://lists.centos.org/mailman/listinfo/centos-virt">http://lists.centos.org/mailman/listinfo/centos-virt</a><br></p>