KVM paravirtualization on non-VMX/SVM CPUs?

KVM paravirtualization on non-VMX/SVM CPUs?

[Tomasz Chmielewski]

In January, Ingo Molnar announced "KVM paravirtualization for Linux"[1].

The announcement mentions "support for the hardware cr3-cache feature of 
Intel-VMX CPUs" and "i only tested this on 32-bit VMX".

Because of this, I'm not sure how to understand it:

1) KVM paravirtualization means that it will be possible to use KVM on 
non-VMX/SVM CPUs (much like Xen)?

2) KVM paravirtualization means paravirtualized drivers only (net, disk 
etc.), but it will still run only on VMX/SVM-capable CPUs?


Which one is true? First, or second?



[1] http://people.redhat.com/mingo/kvm-paravirt-patches/ANNOUNCE


-- 
Tomasz Chmielewski
http://wpkg.org

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

Re: KVM paravirtualization on non-VMX/SVM CPUs?

[Hugo Mills]

[-- Attachment #1.1: Type: text/plain, Size: 1363 bytes --]

On Mon, Mar 19, 2007 at 05:37:21PM +0100, Tomasz Chmielewski wrote:
> In January, Ingo Molnar announced "KVM paravirtualization for Linux"[1].
> 
> The announcement mentions "support for the hardware cr3-cache feature of 
> Intel-VMX CPUs" and "i only tested this on 32-bit VMX".
> 
> Because of this, I'm not sure how to understand it:
> 
> 1) KVM paravirtualization means that it will be possible to use KVM on 
> non-VMX/SVM CPUs (much like Xen)?
> 
> 2) KVM paravirtualization means paravirtualized drivers only (net, disk 
> etc.), but it will still run only on VMX/SVM-capable CPUs?
> 
> 
> Which one is true? First, or second?

   The latter. I asked this question several weeks ago. The response
was that it "would be nice" for KVM to support non-VMX machines, but
that it wasn't a priority in any way for any of the current
developers.

   My understanding is that the paravirtualisation support that Ingo
wrote is purely for performance purposes, bypassing the hardware
emulation layer in qemu and going directly to the host VM's device
layer.

   Hugo.

-- 
=== Hugo Mills: hugo@... carfax.org.uk | darksatanic.net | lug.org.uk ===
  PGP key: 1C335860 from wwwkeys.eu.pgp.net or http://www.carfax.org.uk
      --- If it's December 1941 in Casablanca,  what time is it ---      
                              in New York?                               

[-- Attachment #1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

[-- Attachment #2: Type: text/plain, Size: 345 bytes --]

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV

[-- Attachment #3: Type: text/plain, Size: 186 bytes --]

_______________________________________________
kvm-devel mailing list
kvm-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Re: KVM paravirtualization on non-VMX/SVM CPUs?

[Avi Kivity]

Tomasz Chmielewski wrote:
> In January, Ingo Molnar announced "KVM paravirtualization for Linux"[1].
>
> The announcement mentions "support for the hardware cr3-cache feature of 
> Intel-VMX CPUs" and "i only tested this on 32-bit VMX".
>
> Because of this, I'm not sure how to understand it:
>
> 1) KVM paravirtualization means that it will be possible to use KVM on 
> non-VMX/SVM CPUs (much like Xen)?
>   

No.  That is possible (and kqemu demonstrates it can be done), but it is 
not implemented and to my knowledge no one is working on it.


> 2) KVM paravirtualization means paravirtualized drivers only (net, disk 
> etc.), but it will still run only on VMX/SVM-capable CPUs?
>
>
>   

Yes.  In addition, it enables platform-level optimization (like the cr3 
cache).


-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys-and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV