From owner-freebsd-hackers Fri Oct 25 12:05:18 1996 Return-Path: owner-hackers Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id MAA08893 for hackers-outgoing; Fri, 25 Oct 1996 12:05:18 -0700 (PDT) Received: from rocky.mt.sri.com (rocky.mt.sri.com [206.127.76.100]) by freefall.freebsd.org (8.7.5/8.7.3) with ESMTP id MAA08884 for ; Fri, 25 Oct 1996 12:05:16 -0700 (PDT) Received: (from nate@localhost) by rocky.mt.sri.com (8.7.5/8.7.3) id NAA11438; Fri, 25 Oct 1996 13:01:30 -0600 (MDT) Date: Fri, 25 Oct 1996 13:01:30 -0600 (MDT) Message-Id: <199610251901.NAA11438@rocky.mt.sri.com> From: Nate Williams To: Terry Lambert Cc: msmith@atrad.adelaide.edu.au (Michael Smith), hackers@freebsd.org Subject: Re: DOS emulation (was Re: Networking in PCEMU (1/2)) In-Reply-To: <199610251842.LAA13997@phaeton.artisoft.com> References: <199610250230.MAA16152@genesis.atrad.adelaide.edu.au> <199610251842.LAA13997@phaeton.artisoft.com> Sender: owner-hackers@freebsd.org X-Loop: FreeBSD.org Precedence: bulk Terry Lambert writes: > > > http://world.std.com/~bochs/ > > > > Yeah, well. It builds trivially under 2.1 (haven't tried on -current, > > expect it will there too) with an include fix, but it's a bit limited > > functionality-wise. It's also dog-slow, even compared to PCemu. > > > > What we _need_ is for the kernel vm86 stuff to happen so that DOScmd > > can go ahead; this is basically predicated on Sean having the time to > > do it, as none of the other x86 gurus seemed interested in helping us > > deal with the problem we hit last time. The emulation code in DOScmd > > is actually pretty good, and having brutally cleaned it, it's _much_ > > easier to read and work on than the Linux dosemu code. > > > > Once _that_ works, we can do some _serious_ DOS emulation. > > Which only works on Intel platforms. 8-(. Yep. To expect more than that is foolishness. Insignia software has spents *millions* of dollars on their x86 emulation software, and it still isn't very impressive. Expecting that a single-individual or group of individuals could produce something that is even on par with their product is silly. You'd have better luck building 'translator' code that changes the x86 opcodes into native instructions on each platform, which DEC and IBM have done. Nate