Date: Mon, 2 Dec 1996 13:47:06 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: smp@csn.net (Steve Passe) Cc: ccsanady@friley216.res.iastate.edu, freebsd-smp@freefall.freebsd.org Subject: Re: cvs commit: sys/i386/include pmap.h Message-ID: <199612022047.NAA11277@phaeton.artisoft.com> In-Reply-To: <199612022044.NAA22589@clem.systemsix.com> from "Steve Passe" at Dec 2, 96 01:44:18 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> Long term, I think I agree, I go back and forth on just what the model > should look like. For the short term I have many reasons for my > "minimal impact" model: [ ... expedient reasons ... ] > making an SMP kernel that boots & runs on a uniprocessor board is currently > impossible for a lot of 'architectural' reasons, ie alot would have to change > in the mainline code to make this work. Look at how we pervert things > like curproc for an example. The main line code should be changed, regardless. It would definitely benefit from the changes, and the interrupt architectures for non-Intel machines dictate a similar change be made for porting in any case. I would like to see the change driven by the SMP requirements, since it would be much harder to go back and retrofit SMP if the changes only considered the porting requirements, and not the SMP. The SMP code integration is going to be A Big Deal. If you can fit some additional (good!) changes in under that umbrella, so much the better. Regards, Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199612022047.NAA11277>