On 01.08.2010, at 16:02, Avi Kivity wrote:

On 07/29/2010 03:47 PM, Alexander Graf wrote:

quoted

On PPC we run PR=3D0 (kernel mode) code in PR=3D1 (user mode) and =

don't use the

quoted

hypervisor extensions.
=20
While that is all great to show that virtualization is possible, =

there are

quoted

quite some cases where the emulation overhead of privileged =

instructions is

quoted

killing performance.
=20
This patchset tackles exactly that issue. It introduces a paravirtual =

framework

quoted

using which KVM and Linux share a page to exchange register state =

with. That

quoted

way we don't have to switch to the hypervisor just to change a value =

of a

quoted

privileged register.
=20
To prove my point, I ran the same test I did for the MMU =

optimizations against

quoted

the PV framework. Here are the results:
=20
[without]
=20
debian-powerpc:~# time for i in {1..1000}; do /bin/echo hello>  =

/dev/null; done

quoted

=20
real    0m14.659s
user    0m8.967s
sys     0m5.688s
=20
[with]
=20
debian-powerpc:~# time for i in {1..1000}; do /bin/echo hello>  =

/dev/null; done

quoted

=20
real    0m7.557s
user    0m4.121s
sys     0m3.426s
=20
=20
So this is a significant performance improvement! I'm quite happy how =

fast this

quoted

whole thing becomes :)
=20
I tried to take all comments I've heard from people so far about such =

a PV

quoted

framework into account. In case you told me something before that is =

a no-go

quoted

and I still did it, please just tell me again.
=20
To make use of this whole thing you also need patches to qemu and =

openbios. I

quoted

have them in my queue, but want to see this set upstream first before =

I start

quoted

sending patches to the other projects.
=20
Now go and have fun with fast VMs on PPC! Get yourself a G5 on ebay =

and start

quoted

experiencing the power yourself. - heh
=20
v1 ->  v2:
=20
  - change hypervisor calls to use r0 and r3
  - make crit detection only trigger in supervisor mode
  - RMO ->  PAM
  - introduce kvm_patch_ins
  - only flush icache when patching
  - introduce kvm_patch_ins_b
  - update documentation
=20
v2 ->  v3:
=20
  - use pPAPR conventions for hypercall interface
  - only use r0 as magic sc number
  - remove PVR detection
  - remove BookE shared page mapping support
  - combine book3s-64 and -32 magic page ra override
  - add self-test check if the mapping works to guest code
  - add safety check for relocatable kernels
=20

=20
Looks reasonable.  Since it's fair to say I understand nothing about =

powerpc, I'd like someone who does to review it and ack, please, with an =
emphasis on the interfaces.

Sounds good. Preferably someone with access to the ePAPR spec :).


Alex