[Plex86-devel] Re: Plex86-devel Digest, Vol 3, Issue 3

[Drew Northup N1XIM]

> 
> ----------------------------------------------------------------------
> 
> Date: Thu, 13 Feb 2003 12:22:51 -0500
> From: Steve Coleman <address@hidden>
> 
> In the light of the recent announcement of the new "light weight" Plex86 
> VM I have a few questions...
> 
> Can anyone tell me what "user privilege" might mean in terms of running 
> selinux (http://www.nsa.gov/selinux/) or MOSIX (http://www.mosix.org/) 
> Linux derivatives under the new VM? I had been thinking of looking into 
> VMware to virtualize them for some experiments I was wanting to do but 
> if Plex86 is getting ready for prime time then perhaps I should just 
> help out where I can (not sure where, but I do love an educational 
> challenge). If the same PVI compilation macro trick will work for 
> selinux and/or mosix then I'd much rather spend my time learning and 
> growing with Plex86 as it matures.
> 
> Also, is the older heavyweight MV code being put out to pasture? Or is 
> this basically a planned fork in the road between two different camps? 
> Just curious if the old camp still has a critical mass growing out 
> there, or what the older VM had to offer that might be lost to the light 
> weight version other than support for other non-x86 platforms.
> 
> Thanks,
> 
>  Steve
> 
I'm keeping the old code up on the web for now.

> ------------------------------
> 
> Date: Thu, 13 Feb 2003 13:25:14 -0600
> From: "Bob Diepenbrock" <address@hidden>
> To: <address@hidden>
> Subject: [Plex86-devel] Light Weight Plex86 Questions
> Message-ID: <address@hidden>
> Content-Type: multipart/alternative; boundary="=_DD82A531.C5A48398"
> MIME-Version: 1.0
> Precedence: list
> Message: 3
> 
> This is a MIME message. If you are reading this text, you may want to 
> consider changing to a mail reader or gateway that understands how to 
> properly handle MIME multipart messages.
> 
> --=_DD82A531.C5A48398
> Content-Type: text/plain; charset="us-ascii"
> Content-Transfer-Encoding: quoted-printable
> 
> >From my reading the Lite Plex 86 will be restricted to Linux as Guest OS so=
>  you won't be able to run Windows, Sco or Solaris with this.
> 
> As to the "performance" of Plex86.  I believe there is a long way to go bef=
> ore this project will be "on par" with the VMWARE product.
> 
> Plex86 suffers from a few serious problems at this time..
> 
> 1.  Apparently there is some fragmentation of the development efforts and t=
> his project lacks a clear vision for attainable development.
>      A.  This new "light" version just starting.
>      B.  Win-32 port being worked
>      C.  The use of BOCS stuff has sucked some development energy towards t=
> hat  project.
>      D.  The founding developer is now moved on and there is not much getti=
> ng done. (See 5 and 6 below)
Yes, all of the above really hurts.

> 2.  User Level Documentation is poor and lacks enough detail to get the nov=
> ice up and running his guest OS.=20
I've actually held back on re-writing all of it myself.  I guess that if
I want this to stay afloat I'll just have to do it all myself, and
modify the CONTRIBUTORS file to suit.

> 3.  Nobody is maintaining install packages for various supported platforms,=
>  you basically have to build this stuff yourself and there are a lot of thi=
> ngs you will have to install before that is going to work.
We don't do packages yet for a reason.  This is ALPHA CODE.  I don't
believe the philosophy that some at RedHat do--that any cool looking
code is ready for the masses of newbies out there--and I act
accordingly.

> 4.  Plex 86 does not yet support some important features, including floppy =
> drive access, CD Rom Access in a manner that makes it easy to install a new=
>  guest OS.
I can install just about anything that I want, thank you very much.
CDROM = /dev/cdrom (usually) Sure looks like a file to me.
Floppy = /dev/fd0  Also, sure looks like a file.
Just give that user permission of access those devices raw.  I know it
looks like a security hole (and it can be), but if you know what you're
doing then it is not a problem.

> 5.  Development level documentation is non-existent, so there is a very ste=
> ep learning curve for new developers and no guiding principles for adding n=
> ew features.
I've been working on that, slowly.

> 6.  Internal documentation in the source is also non-existent for the most =
> part, from my quick review of the code.
Definitely a problem.  See my /. posts.

> 
> This is not intended to offend the current developers of this project.  Thi=
> s is what I see as the major issues here.
No offence taken, at least not by me.

> 
> I'm greatly concerned about this new "Plex86 - Light" project.  I don't thi=
> nk it serves the Plex86 project and will further detract from it by dividin=
> g the attention of the available developers.  I would also encourage the de=
> velopers for the Light side to make a clear statement as to their purpose a=
> nd how it will differ from the purpose of Plex86.  I would also like them t=
> o make that clear by altering their project's name, perhaps removing the Pl=
> ex86 or including "Light" in the name.
That's what I suggest.  We'll see if Kevin likes the idea.

> 
> -=3D Bob =3D-
> 
> ------------------------------
> 
> Date: Thu, 13 Feb 2003 14:53:49 -0500
> From: Eric Laberge <address@hidden>
> 
> On Wed February 12 2003 19:31, Perry E. Metzger wrote:
> > Could someone clue me in on the deal with the sourceforge plex86
> > vs. the savannah plex86, and the (apparently two) mailing lists?
> 
> Aloha!
> 
> I'm still working on the Savannah Plex86 project for my degree, but my 
> current 
> code is unusable for now. I have IO working, and running real-mode code 
> through vm86 calls. You can find some sample of my work in the "files" 
> section (EL-Exp-VM). I got the BIOS booting all the way to accessing the 
> floppy drive, with hardware interrupt handling stopping me right now.
> 
> Just to say "don't kill the Savannah project" right now. There still is a 
> need 
> for a "heavy weight" Plex86.
> 
> - -- 
> Eric Laberge
> ICQ: 1261355
> MSN: address@hidden
> Jabber: address@hidden
> 
> GCS/M/S d-(++) s-: a-- C++(++++) UL+++>$ !P L+++ E--- W++ N+ w++ !O M
> !V PS PE Y PGP-(+++) t++(-) X- R* tv+ b+ DI+++  D+ G e++>+++++ h! 
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.2.1 (GNU/Linux)
> 
> iD8DBQE+S/fQ1Q3tcr/gnPMRAhXNAJwLG4RAVV40LuXTAe/BLNAQ1fIsXACghIKO
> FJj8rPtTcaG3gJ2zuCkC8bQ=
> =S2UV
> -----END PGP SIGNATURE-----
> 
Well, I don't plan on killing the savannah site.  So that shouldn't be a
problem for you.

> 
> 
> ------------------------------
> 
> Date: Thu, 13 Feb 2003 12:38:35 -0800 (PST)
> From: Kevin Lawton <address@hidden>
> Subject: [Plex86-devel] Re: [Plex86-developers] light weight Plex86 questions
> 
> --- Steve Coleman <address@hidden> wrote:
> > Can anyone tell me what "user privilege" might mean in terms of running 
> > selinux (http://www.nsa.gov/selinux/) or MOSIX (http://www.mosix.org/) 
> > Linux derivatives under the new VM? I had been thinking of looking into 
> > VMware to virtualize them for some experiments I was wanting to do but 
> > if Plex86 is getting ready for prime time then perhaps I should just 
> > help out where I can (not sure where, but I do love an educational 
> > challenge). If the same PVI compilation macro trick will work for 
> > selinux and/or mosix then I'd much rather spend my time learning and 
> > growing with Plex86 as it matures.
> 
> I'm supporting Linux 2.4 and up for guests.  I've tried 2.4.20
> and 2.5.59 (getting close to 2.6).  Either works fine with the
> simple macro mods.  Anything based on those should be doable all
> the same as a general rule.  You have to configure out useless
> hardware since I bail on probing of unsupported hardware at the
> moment.  Probably could use the same guest drivers I'm making.
> [and soon hopefully which other people who are truly talented
> at coding drivers will be making :^) ]
> 
> Logical mods to the kernel aren't really seen by plex86.  Only
> access to hardware.  So, without knowing much more, I'd say
> the answer is likely yet to your question for selinux.
> 
I'm still not convinced that the full VM is dead for running binary only
OS's.  But, there are things that can be done.

> For mosix, being clustering software, I'm not sure what it's
> additional hardware requirements are, latencies and all that.  If you want
> to point me at the most salient URL which describes the stuff
> I need to know, I'll look at it and take a guess.
> 
> I'm quite interested in plex86 being supported by the security world,
> so if the selinux following needs a little help getting it working, I'll
> try to help where I have the resources to.
> 
This is something where a concept talked about a good while ago would be
worth pursuing.  Multiple clients inside of one framework.

> 
> > Also, is the older heavyweight MV code being put out to pasture? Or is 
> > this basically a planned fork in the road between two different camps? 
> > Just curious if the old camp still has a critical mass growing out 
> > there, or what the older VM had to offer that might be lost to the light 
> > weight version other than support for other non-x86 platforms.
> 
> I don't manage the old fork, so I'll let the others make comments there.
> My personal feeling is that it would be best to retire it, and
> redirect to the new ones.  This new effort will is much more clean,
> simple and easy to manage, and I would certainly enjoy if all the
> interested parties from the old fork would join forces on this new
> effort.  I added a lot of performance to bochs as well, so bochs
> in conjunction with this new plex86 will be useful for binary
> only OSes.
I think that we can manage to still work together.  The two forks seem
to have different aims at the moment.  This does not preclude code and
idea sharing, however.

> 
> And hey, I'm even adding comments to the code.  :^)  (a fun jab
> in response to some Slashdot commentary)
> 
> 
> -Kevin

Thank GOD!!!

> 
> __________________________________________________
> 
> Date: Thu, 13 Feb 2003 15:42:39 -0600
> From: "Bob Diepenbrock" <address@hidden>
> 
> I agree, don't kill it now because the "Lite" project does not have the sam=
> e goals.
> 
> -=3D bob =3D-
> 
> _______________________________________________
> 
> Date: 13 Feb 2003 19:17:43 -0500
> From: "Perry E. Metzger" <address@hidden>
> 
> James Pellow <address@hidden> writes:
> > I am excited about the new code, but it seems there still is a
> > usefullness to the "Heavy weight" old system.  I would love to be
> > able to run win2k for example and be rid of vmware.  Will the
> > bochs-plex86 solution provide the performance of vmware?  I doesn't
> > seem like it will, but I could be wrong.  I for one hope to old
> > project continues.
> 
> I would like it too, but I think Kevin has different plans.
> 
> Perry
> 
As said, I run Plex86 now, at least this fork.  Kevin runs the other
one.  For at least the time being I'm going to keep this code easily
available for people to play with.

> 
> ------------------------------
> 
> Date: Fri, 14 Feb 2003 07:44:12 +0100
> From: Lars Segerlund <address@hidden>
> 
>   HI,
> 
>    I think there is motivation for both, so if the active developers 
> backport Kevins stuff that is usefull for the heavy wheight and submit 
> some patches that might be useful for the lightwheight everyone will be 
> better off.
> 
>    I think the heavywheight is trying to do a VMWARE, and the 
> lightwheight is trying to be a developement tool.
> 
>    The important thing is not to stop coding, and to keep the CVS up to 
> date !
> 
>   / regards, and good luck, Lars Segerlund.
> 
Exactly.  I'm not worried about coding right now.  I'm worried about
documenting the great code already there.  Once that is at least
reasonably done I think that other things can be worried about (like a
clean X fullscreen mode, for instance).

--Drew Northup, N1XIM