Date: Thu, 11 Sep 1997 10:41:39 +0930 From: Greg Lehey <grog@lemis.com> To: Sean Eric Fagan <sef@Kithrup.COM> Cc: gallatin@CS.Duke.EDU, emulation@FreeBSD.ORG, jkh@time.cdrom.com, mike@smith.net.au, sos@sos.freebsd.dk Subject: Re: Net posting: SCO gets Linux emulation Message-ID: <19970911104139.00351@lemis.com> In-Reply-To: <199709110105.SAA09151@kithrup.com>; from Sean Eric Fagan on Wed, Sep 10, 1997 at 06:05:10PM -0700 References: <199709110105.SAA09151@kithrup.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, Sep 10, 1997 at 06:05:10PM -0700, Sean Eric Fagan wrote:
>>> Well, its somewhat interesting because it runs entirely in userland
>>> and traps system calls via a SEGV handler.
>> Ugh. Is this what we have come to expect of SCO?
>
> That is both unfair and rude. (Bit of a warning here... I've known the
> author ever since I interviewed at SCO, and I happen to like him.)
Well, in fact, I left the question open. I was wondering. Sure, it
works (well, the concept works, and I assume the emulator does), but
it's not exactly the way to get high performance.
> I looked at the program; it's interesting, but not terribly exciting.
> However, I might have done the iBCS2 emulation the same way, if I could
> have -- however, trapping the system call vector in a user-mode program is
> hard. If 386BSD had used a different entry vector...
It's a question of performance. If I understand it correctly, you do
a SIGSEGV for every system call.
> Mike also probably would have done the program using LDT manipulation,
> except for the fact that the IDT is global.
Yes, I'm sure that there are constraints which made him choose this
method.
>>> And because of this, I imagine that its a good bit slower. Also,
>>> their '$LINUX_ROOT' path remapping is interesting if only for its
>>> flexibility, but their choice of what paths to remap is very
>>> haphazard compared with the {Free,Net}BSD approach.
>> Doesn't sound like a serious implementation effort to me.
>
> To a large degree, it isn't. He did it as a quick&dirty way to be able to
> run the Linux port of Acrobat Reader (there is/was no SCO version), and
> apparantly convinced SCO to allow him to release it under a Berkeley-style
> license.
>
> All told, it probably took Mike about a day to write this.
OK, basically you're agreeing with me. My thoughts were "this is a
cheap way to get functionality. It won't bring performance". I've
written plenty of that category in my time, and some of them have been
great successes.
Greg
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?19970911104139.00351>