On 14.06.2009, at 13:57, Avi Kivity <avi@xxxxxxxxxx> wrote:
Ben Sanders wrote:
I'm looking to reproduce the results of Alexander Graf earlier this
year when he said that he was able to get ESX to run a ReactOS guest
all on top of KVM. More information on that can be found here:
http://kerneltrap.org/mailarchive/linux-kvm/2009/1/5/4600354/thread
So far, I can only boot esx into service console mode, not regular
mode or debug mode without a kernel panic (in ESX). I've enabled
nesting, both in the module and in the kvm command on v86, using the
-cpu phenom option as well.
First question is about cpu type. I've been using some dual core AMD
cpu's with svm enabled, but I'm wondering if I actually need a phenom
cpu for all of this to work. I was under the impression that KVM/
QEMU
could emulate a different CPU if the features didn't exist.
No, kvm (mostly) doesn't emulate cpu features. I don't know why -
cpu phenom is required. Copying Alex.
I needed nested paging to have things work, so phenom was mandatory.
Also, because I was running things on a Phenom, I figured the safest
bet is to be truthful in telling the guest what it's running on. Also
ESX checked for specific cpuid revisions and I knew the CPU I was on
is safe.
If you get it running in debug mode, do -serial stdio and send me the
output so I can see if I remember the cure ;)
Alex
Second question is about the host OS. Are the nesting features in
KVM
only supported in an x86 OS? or should x86_64 work as well? I've
tried each (in varying degrees), but if one implementation wouldn't
work (like emulating a phenom, an x86_64 processor on an x86 host
OS),
I'd like to know so I can stay away from that. Also, I've been using
Ubuntu 8.10 and 9.04, but if there's some reason I should switch to a
different distribution, I'd like to know that as well.
x86 generally means i386 and x86_64. Both should work, and x86_64
is recommended.
--
error compiling committee.c: too many arguments to function
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
