Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 2 Oct 1998 14:06:58 +0930 (CST)
From:      Mark Newton <newton@atdot.dotat.org>
To:        mike@smith.net.au (Mike Smith)
Cc:        viren@rstcorp.com, freebsd-chat@FreeBSD.ORG, newton@dotat.org
Subject:   Re: SVR4 emulation?
Message-ID:  <199810020436.OAA24871@atdot.dotat.org>
In-Reply-To: <199810011952.MAA00964@dingo.cdrom.com> from "Mike Smith" at Oct 1, 98 12:52:21 pm

next in thread | previous in thread | raw e-mail | index | archive | help
Mike Smith wrote:

 > > I just followed a link on slashdot.org to someone's work on SVR4
 > > emulation for FreeBSD. It is at
 > >    ftp://slash.dotat.org/pub/freebsd-svr4/
 > > Has anyone tried this out? Are there any plans for it to be
 > > incorporated in FreeBSD?

A couple of people have tried it as far as I can tell, but I've had
precious little feedback.  Reading anguished requests for beta testers
in -hackers leads me to believe that that state of affairs is in no
way unusual for kernel-related work :-(

 > It was just a bit too late for 3.0, but it will be incorporated shortly 
 > therafter.
 > 
 > > Tought I'd ask before trying it out myself.
 > 
 > We'd love to hear your results; please keep us informed.

Indeed.  Please give it a try and let me know how you go with it.
It isn't complete, by any means, but I've had mixed amounts of
success with relatively complicated applications.  Most of the
software on the Solaris 2.5.1 CD-ROM seems to work ok, anyway.

Things which aren't done include:

   - Light-Weight Processes, otherwise known as kernel threads (duh!)
   - Interfacing BSD's emulation of SysV shared memory to SysVR4
     syscalls
   - I think there's still a small bug in sigreturn() which might
     explain why Netscape doesn't work and Solaris vi dumps core
     (it never used to dump core until I reimplemented sendsig/sigreturn,
     so the problem has to be in there somewhere).
   - various other things on a to-do list on the webpage.

Since this is going to -chat I guess it'd make sense to explain why
we'd want SysVR4:  The (abortive?) UltraSPARC port and the somewhat
more successful Alpha port have had the beneficial side-effect of
getting people to work on making FreeBSD 64-bit clean, placing us in
a good position for a Merced port if Merced ever gets released.

SysV vendors such as Sun, Silicon Graphics, SCO and HP have announced
plans to port Solaris, IRIX, OpenServer and HP-UX to Merced.

Now -- If we run on Merced and can run Merced SysVR4 executables, 
that means we'll be able to run programs written for Solaris, IRIX,
OpenServer and HP-UX.

... which is rather more comprehensive than, "Yeah, we can run most
Linux code, but Oracle still doesn't work."

At the moment, of course, I'm saying, "Yeah, we can run most Solaris
and some SCO code, but Oracle still doesn't work," so maybe we haven't
made much headway yet :-)

Personally I think Merced is a Microsoft plot:  Distract all the
UNIX vendors by making them rush around porting their OSs to a
processor that'll never be released while NT powers ahead on Pentium
descendants :-)  But just in case I'm wrong I've done the SysVR4 stuff. 

There's a web page for the project at 
http://slash.dotat.org/~newton/freebsd-svr4 (essentially a HTMLized
version of the README file, 'cept a bit more up-to-date).  Download
it, join the mailing list, try it out with a Solaris or SCO CD, and
tell us what you think.

Thanks,

    - mark

--------------------------------------------------------------------
I tried an internal modem,                    newton@atdot.dotat.org
     but it hurt when I walked.                          Mark Newton
----- Voice: +61-4-1958-3414 ------------- Fax: +61-8-83034403 -----

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199810020436.OAA24871>