From owner-freebsd-platforms Wed Mar 12 07:14:55 1997 Return-Path: Received: (from root@localhost) by freefall.freebsd.org (8.8.5/8.8.5) id HAA18063 for platforms-outgoing; Wed, 12 Mar 1997 07:14:55 -0800 (PST) Received: from rover.village.org (rover.village.org [204.144.255.49]) by freefall.freebsd.org (8.8.5/8.8.5) with SMTP id HAA18057 for ; Wed, 12 Mar 1997 07:14:51 -0800 (PST) Received: from rover.village.org [127.0.0.1] by rover.village.org with esmtp (Exim 0.56 #1) id E0w4piI-0004Pf-00; Wed, 12 Mar 1997 08:13:10 -0700 To: Terry Lambert Subject: Re: Kernel layout reshuffle (Was: Re: FAQ?) Cc: msmith@atrad.adelaide.edu.au (Michael Smith), kim@tinker.com, freebsd-platforms@freebsd.org In-reply-to: Your message of "Tue, 11 Mar 1997 18:45:23 MST." <199703120145.SAA26524@phaeton.artisoft.com> References: <199703120145.SAA26524@phaeton.artisoft.com> Date: Wed, 12 Mar 1997 08:13:10 -0700 From: Warner Losh Message-Id: Sender: owner-platforms@freebsd.org X-Loop: FreeBSD.org Precedence: bulk In message <199703120145.SAA26524@phaeton.artisoft.com> Terry Lambert writes: : FreeBSD needs to let Richard Wackerbarth's ideas be implemented, : even if Richard isn't the person who does it. Among other things. Hmmm, I don't know enough about Richard's stuff to know if it goes this far... : It would be nice if the structure of the source tree did not : actively oppose porting to other platforms. For instance, it would : be nice to have: : : o A build system that will work on all platforms : o Install tools that will work the same on all : platforms. : o Kernel dependent components built along with kernel : o SMP treated as if it were an architecture : o Better seperation of bus code into x86 dependent and : indepedent pieces : o Better seperation of all code : o Seperation of PCI code probe ordering so that on the : DEC Alpha, where the ISA is bridged off the PCI instead : of the other way around I don't know if I agree on all of these. So far I've found that building the kernel seems to work OK on my intel box or my OpenBSD mips box, but building the userland is a nightmare. Work does need to be done to make that sane. I've also found that x86 knows too much about busses and the like. I don't know if NetBSD is the way to go here, but if there were one unified, flawed API, vs many different ones with ours being perfect, you uknow which one I'd pick :-). FreeBSD is a mess to cross compile its userland. I punted quickly after trying once. Too many thing build things that then are used to build the target. OpenBSD is much better about this. Anyway, there are many tasks needed to make FreeBSD truly porting friendly. Warner