[9fans] another VM type host

[9fans] another VM type host

[Aharon Robbins]

A friend recently pointed me at:

	 http://www.cl.cam.ac.uk/Research/SRG/netos/xen/

The guest OS runs in x86 ring 1, instead of ring 0, so the OS must actually
be ported to work with it.  Has any 9developer seen this yet?

Arnold

Re: [9fans] another VM type host

[Wes Kussmaul]

Correct me if I'm wrong, but isn't Xen development backed by Microsoft?

Not that that would necessarily be a stopper for anyone, but people should
be aware of it.

Wes Kussmaul


At 04:05 AM 2/2/2004, you wrote:
>A friend recently pointed me at:
>
>         http://www.cl.cam.ac.uk/Research/SRG/netos/xen/
>
>The guest OS runs in x86 ring 1, instead of ring 0, so the OS must actually
>be ported to work with it.  Has any 9developer seen this yet?
>
>Arnold

Re: [9fans] another VM type host

[Aharon Robbins]

They are helping, certainly; the Xen people got Win/XP source to do a port,
but it seems to be very Linux based.

They claim superior performance over vmware.  I personally view it as Yet
Another Research Project that gets 90% there but that I wouldn't want to
bet my business on.

In article <6.0.1.1.2.20040202083618.03a91398@mail.village.com> you write:
>Correct me if I'm wrong, but isn't Xen development backed by Microsoft?
>
>Not that that would necessarily be a stopper for anyone, but people should
>be aware of it.
>
>Wes Kussmaul
>
>
>At 04:05 AM 2/2/2004, you wrote:
>>A friend recently pointed me at:
>>
>>         http://www.cl.cam.ac.uk/Research/SRG/netos/xen/
>>
>>The guest OS runs in x86 ring 1, instead of ring 0, so the OS must actually
>>be ported to work with it.  Has any 9developer seen this yet?
>>
>>Arnold

Re: [9fans] another VM type host

[ron minnich]

I am talking to those guys and trying to find some time to do this port.

If we could get this to go it would be useful -- run a auth, fs, and
terminal on your laptop, for example.

ron

Re: [9fans] another VM type host

[ron minnich]

On Mon, 2 Feb 2004, Wes Kussmaul wrote:

> Correct me if I'm wrong, but isn't Xen development backed by Microsoft?

who the heck cares? It's good work and open source near as I can tell.

ron

Re: [9fans] another VM type host

[Matthias Teege]

> I am talking to those guys and trying to find some time to do this port.

please do so. It would be realy nice to kill
the slow vmware emulation.

Matthias

Re: [9fans] another VM type host

[a]

// It would be realy nice to kill the slow vmware emulation.

perhaps, but i'm skeptical of emulation that involves porting
the guest system. yet another small-user-base port to maintain?
the benefit over vmware had best be pretty darn significant.

not that it matters to me personally. i've still got the *really*
slow emulation of VirtualPC on my Mac (or worse: bochs).
ア

Re: [9fans] another VM type host

[ron minnich]

On Mon, 2 Feb 2004 a@9srv.net wrote:

> the guest system. yet another small-user-base port to maintain?
> the benefit over vmware had best be pretty darn significant.

read the papers, they are quite good and there is a real performance gain
in many cases over vmware, and in one case over native mode (!).

ron

Re: [9fans] another VM type host

[Martin C.Atkins]

On Mon, 2 Feb 2004 10:19:48 -0700 (MST) ron minnich <rminnich@lanl.gov> wrote:
> I am talking to those guys and trying to find some time to do this port.

I suspect you mean "the port to get Plan 9 running under Xen", but
wouldn't the reverse also be rather interesting, viz "the port to get
Xen running under Plan 9"?

This would be a lot easier/possible than porting vmware to Plan 9,
since Xen is open source.

BTW, Xen isn't the only "VM for an incomplete machine"-type of VM
project, the Plex86 project has recently announced that they are
abandoning the full-VM (a la vmware) approach for the cut-down VM
approach. AFAIK, they have also taken a similar approach to device
and network access.

Martin
--
Martin C. Atkins			martin@parvat.com
Parvat Infotech Private Limited		http://www.parvat.com{/,/martin}

Re: [9fans] another VM type host

[Russ Cox]

> A friend recently pointed me at:
>
> 	 http://www.cl.cam.ac.uk/Research/SRG/netos/xen/
>
> The guest OS runs in x86 ring 1, instead of ring 0, so the OS must actually
> be ported to work with it.  Has any 9developer seen this yet?

The guest OS in VMware does too.  The difference is they
punt on x86 compatibility, so that's why the OS must be ported.
They don't virtualize the video either, making it pretty darned
useless to me -- I can already run auth, terminal, and fs on
one laptop by using fossil.

Russ

Re: [9fans] another VM type host

[okamoto]

> I can already run auth, terminal, and fs on
> one laptop by using fossil.

You meant without venti, and which is the configuration
David mentioned a while ago?

Kenji