Re: [Qemu-devel] Plex86 and Qemu

From: "Jim C. Brown" <jma5@umd.edu>
To: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] Plex86 and Qemu
Date: Sun, 13 Feb 2005 17:06:14 -0500	[thread overview]
Message-ID: <20050213220614.GA30662@jbrown.mylinuxbox.org> (raw)
In-Reply-To: <001c01c51203$24d2e510$254d21d1@computername>

On Sun, Feb 13, 2005 at 01:35:10PM -0600, jeebs@yango.us wrote:
> From: "Jim C. Brown" <jma5@umd.edu>
> 
> >> The reason I'm asking is that I use Windows, not Linux. 
> > 
> > I can tell by the length of lines in your email. ;)
> 
> Yeah... Some Linux mail readers do seem to have problems with line lengths.  You'd think they'd fix it and add reasonable line wrapping, but... Heck... even the old dial up BBS newsgroups (Fidonet etc.) in the 80's and 90's had mail readers that could handle long lines and do line wrapping...  [shrug]

Actually, it's the editor I use to reply to emails. I could edit it to support
wrapping long lines, I just haven't gotten around to it yet.

> 
> >> Therefor Fabrice's new module is useless to me.  Nice idea etc., but utterly useless to me.
> > 
> > Mine will be too, unless someone ports plex86 to Windows.
> 
> Plex86 v2 is still limited to running only *modified* linux kernals.
> 
> So even for those who just want to run a copy of Linux aren't really going to be helped.  No Knoppix or any other Live! emergency boot cd.  No experimenting with other distro's, etc.

For the reasons given below, this will not apply to qemu.

> 
> >> Plex86 v2 is also Linux only, and requires patched guest kernels.
> > 
> > My idea is to use the module in cosimulation. This would not require a patch
> > guest kernel and would allow for any guest OS to be used, not just linux.
> 
> I'm not quite sure I follow...
> 
> I'm not quite sure how you are going to do that.
> 

The 2 main reasons that plex86 v2 didn't support anything other than a modified
guest were:

a) It was a huge pain supporting ring 0 virtualization

b) Rather than go thru the trouble of hardware emulation, plex86 used a HAL
(Hardware Abstraction Layer) that the guest OS had to know about.

a) is taken care of by letting user space qemu use dynamic translation on
ring 0 instructions (same as qemu + kqemu works now), and b) is taken care of
by letting qemu emulate the hardware. The plex86 kernel module is needed only
for the virtalization, everything else is taken care of by qemu.

>  
> >> Anyway, the point I'm making (and asking about) is that
> >> the majority of potential users are either going to run Windows
> >> as the guest, or as the host..  (And possibly both.)
> > 
> > Mine will support using Windows as guest, but not as host.
> 
> Then that's still going to be extremely limited.  Effecting probably 95%+ of the potential users.
> 

It will support the same series of users that kqemu does. plex86 doesn't support
2.6 kernel series, but it shouldn't be too difficult to port either.

>  
> >> So... is your idea going to be helpful to the majority of potential users?
> >> 
> > 
> > Yes. It will let linux2.4 users run Windows and Windows apps.
> 
> I wouldn't call that "majority of potential users"...  Maybe "many in this mailing list", but that's not quite the same thing.
> 
> [grimace]
> 
> Oh well.
>  

Feel free to port it to a Windows host then. I noticed that you didn't complain
when kqemu was announced, despite the fact that it only supports linux hosts.
Care to explain why?

> 
> 
> 
> _______________________________________________
> Qemu-devel mailing list
> Qemu-devel@nongnu.org
> http://lists.nongnu.org/mailman/listinfo/qemu-devel
> 

-- 
Infinite complexity begets infinite beauty.
Infinite precision begets infinite perfection.