Date: Wed, 7 May 1997 10:23:39 -0700 (MST) From: Terry Lambert <terry@lambert.org> To: thorpej@nas.nasa.gov Cc: ccsanady@nyx.pr.mcs.net, jkh@time.cdrom.com, brandon@cold.org, FreeBSD-Hackers@FreeBSD.ORG Subject: Re: Alpha questions.. Message-ID: <199705071723.KAA21463@phaeton.artisoft.com> In-Reply-To: <199705070210.TAA23402@lestat.nas.nasa.gov> from "Jason Thorpe" at May 6, 97 07:10:10 pm
next in thread | previous in thread | raw e-mail | index | archive | help
> > Just out of curiosity, couldn't we use a large part of the NetBSD/alpha > > machine dependant code? I would think that this would significantly > > speed up a port.. > > ...if you do that, you'll probably have to use a large part of the > NetBSD machine-independent code, too... since the MD code uses MI > facilities, while the MI facilities also rely on MD support. Yes. This is in line with Steve Passe's comments on SMP Alpha support, and the need to abstract hardware in a formal HAL. Actually, I'd like to see hardware offer "services" to the kernel; this is a bit lower level than a traditional HAL, since it would be things like "this chip offers these three services" and "that chip and this chip in concert offer pcaudio services". This is slightly (not unreconcilably much) different than the NetBSD MI code. One of the hairiest problems back on the orginal FreeBSD PCI Alpha code was getting the time and other services up to support the FreeBSD console code, since they were *extremely* PC architecture dependent. The Alpha bridged it's ISA the other direction off the PCI than a PC (a *much* better design, but what's new there?). 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?199705071723.KAA21463>