Date: Wed, 26 Feb 1997 13:41:21 -0700 (MST) From: Nate Williams <nate@mt.sri.com> To: <SimsS@IBM.Net> Cc: "Nate Williams" <nate@mt.sri.com>, "Hackers" <Hackers@freebsd.org> Subject: Re: Building PAO kernel on non-PAO system Message-ID: <199702262041.NAA00507@rocky.mt.sri.com> In-Reply-To: <199702262032.UAA71641@out1.ibm.net> References: <199702262032.UAA71641@out1.ibm.net>
next in thread | previous in thread | raw e-mail | index | archive | help
> I think you missed my point Yep, I did. :) >- It's not possible (or I'm not smart enough to > know how) to build, say, 2.1.7, 2.2 *AND* -current kernels on a single "super > server" and then blow them out amongst the unsuspecting users depending on > their needs (and possibly their threshold for pain). Can such a thing be > done? Sure, but it really goes against the grain. Why can't these unsuspecting users all use the GENERIC kernel and be happy with it? It supports everything they need, and there's no need for the super-server. That's supposedly the purpose of the GENERIC kernel. > This is the point I'm trying to make in my own inept fashion - I can't build > a PAO-enabled kernel to distribute to another machine on a (non-PAO > configured) desktop that's running a different rev of FreeBSD source. Or can > I? One could argue that you shouldn't be placing PAO-kernels on unsuspecting users. You should only run *release* code on machines except where people are aware of the problems, and then they should build their own systems so they can 'fix' bugs that arise. (This isn't a solution, but an explanation of the thinking.) In any case, you can build kernels for any release like you can with *any* software project. You just need to setup the environment that each release requires. Man 'chroot' for some hints. Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199702262041.NAA00507>