From owner-freebsd-current Thu Oct 23 20:03:39 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id UAA16059 for current-outgoing; Thu, 23 Oct 1997 20:03:39 -0700 (PDT) (envelope-from owner-freebsd-current) Received: from word.smith.net.au (word.smith.net.au [202.0.75.3]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id UAA16053 for ; Thu, 23 Oct 1997 20:03:32 -0700 (PDT) (envelope-from mike@word.smith.net.au) Received: from word.smith.net.au (localhost.smith.net.au [127.0.0.1]) by word.smith.net.au (8.8.7/8.8.5) with ESMTP id MAA00673; Fri, 24 Oct 1997 12:28:55 +0930 (CST) Message-Id: <199710240258.MAA00673@word.smith.net.au> X-Mailer: exmh version 2.0zeta 7/24/97 To: John-Mark Gurney cc: Mike Smith , FreeBSD Current Subject: Re: Doug Rabson's kernel linker code.. In-reply-to: Your message of "Thu, 23 Oct 1997 17:01:30 MST." <19971023170130.55831@hydrogen.nike.efn.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Fri, 24 Oct 1997 12:28:55 +0930 From: Mike Smith Sender: owner-freebsd-current@freebsd.org X-Loop: FreeBSD.org Precedence: bulk > > It should be possible at that stage to reference the various symbols > > set when the kernel is loaded and started; looking at create_pagetables > > in i386/i386/locore.S I see that KERNend is set to the end of the > > kernel plus any symbol table. > > hmm.. any machine independant way to get this information? Not really, no. It's reasonably dependant on the load format and technique used by a particular architecture. > yeh.. well.. as I stated about.. the file that needs this is > kern/link_aout.c... and I don't really want to reference machine specific > symbols... unless we require all machines to contain these symbols.. This is no better or worse than using a compile-time manifest constant. > the other option is to do something were we move the running of SYSINIT > into kern_linker.c... and then at boot time we simply "link" in the > kernel as we do with kld modules... this would require extensions to > kern_linker.c to support linking of a memory address.. but this wouldn't > be hard to do... If I read you correctly here, this is basically the first step in making the kernel boot-time linkable. I think that everyone that's ever been interested in this issue is standing on their chairs yelling "GO!" at you about now... mike