Acheron writes: “Spotted this on Deadly.org today. Four University of Waterloo students are working on adding SMP support to the OpenBSD kernel as their 4th year Computer Engineering Design Project. Their website can be found at Spinlocks.org. Best of luck to them.”
OpenBSD SMP Project Started
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
What does Theo think of all this? It’s basically never going to be part of the official distro without his sanction.
Well, this should be good, Apple hardware!
http://spinlocks.org/images/hacking_gibson.jpg
About fsckin time.
Not trolling, I’ve just literallying been waiting YEARS for this. I prefer NetBSD and OpenBSD (an opinion thing to FreeBSD) but I have an SMP box so I keep running FBSD. Though, chances are I’ll be too happy with FreeBSD 5.0
Cool to see a project like that started by the University of Waterloo, should help them get on the map of good computer schools better, seeing how its basically the best Canada has to offer, now if only they hadnt rejected my application….(oh im bitter).
That’s the question. Will these changes ever get into the main OpenBSD CVS ? IIRC Many OpenBSD developers have said thiy don’t care at all about SMP, that it just complicates to much, and makes the well known code audition they do much harder.
That’s interesting, given that there’s already an OpenBSD SMP cvs tree:
http://www.openbsd.org/cgi-bin/cvsweb/src/sys/?only_with_tag=SMP
-fp
Seems like OpenBSD will eventually have SMP support, eh Bascule?
Ha, you do know that Uwaterloo is the _only_ school that Microsoft recruits from, right? Trust me, they are already on the map, after years of Software developed for the mac specifically.
That’s interesting, given that there’s already an OpenBSD SMP cvs tree:
Take a closer look at the age on most of those files… the official SMP project is stagnant, IMO. Taking nothing away from the core developers… its just that SMP has not been an immediate priority for them. Security and stability have always been valued first.
That said, there is a large user base pining for just this sort of thing. I think SMP is inevitable for OpenBSD, whether it comes in the form of the Spinlocks project, or elsewhere.
In fact, given Theo’s hardon for the UltraSparc III (to take advantage of security-related features), I’d say that its actually pretty unavoidable.
it’s about time. and i think that theo will love it and implent it, if it’s good enough.
finaly i can convert those linsux users at my computerclub that obsd is THE o/s to be used.
SMP for OpenBSD sounds great but without a reworked threading model isn’t it just going down the same road that FreeBSD has been on?
I think it’s generally accepted that Linux’s threading model is better and that it scales much better (also to be improved further in 2.6 kernel).
Perhaps I’m talking through my hat? Someone correct me if I’m wrong.
Cheers