From owner-cvs-all Wed Sep 30 01:06:10 1998 Return-Path: Received: (from daemon@localhost) by hub.freebsd.org (8.8.8/8.8.8) id BAA19818 for cvs-all-outgoing; Wed, 30 Sep 1998 01:06:10 -0700 (PDT) (envelope-from owner-cvs-all) Received: from spinner.netplex.com.au (spinner.netplex.com.au [202.12.86.3]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id BAA19664; Wed, 30 Sep 1998 01:05:46 -0700 (PDT) (envelope-from peter@netplex.com.au) Received: from spinner.netplex.com.au (localhost [127.0.0.1]) by spinner.netplex.com.au (8.9.1/8.9.1/Spinner) with ESMTP id QAA10857; Wed, 30 Sep 1998 16:02:04 +0800 (WST) (envelope-from peter@spinner.netplex.com.au) Message-Id: <199809300802.QAA10857@spinner.netplex.com.au> X-Mailer: exmh version 2.0.2 2/24/98 To: John Birrell cc: dima@best.net, jb@FreeBSD.ORG, cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG Subject: Re: cvs commit: www/en/alpha Makefile alpha.sgml bootstrapping.sgml current.sgml In-reply-to: Your message of "Wed, 30 Sep 1998 17:54:19 +1000." <199809300754.RAA03825@cimlogic.com.au> Date: Wed, 30 Sep 1998 16:02:03 +0800 From: Peter Wemm Sender: owner-cvs-all@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk John Birrell wrote: > Peter Wemm wrote: > > Going by what I've seen, the boot blocks look OK, but they don't (yet) > > preload kld modules - it just loads kernels at the moment. kld itself > > should work though. > > At what point will it be possible to make a decision about whether 3.0 (i386) > goes out with an elf kernel? I was going to do some MI elf load and i386 and alpha elf startup code last night but ran out of steam while figuring out whether it was going to be possible to skip the need for forcing everything to be linked as a dynamic module (not PIC, just dynamic!) since we can for the most part get all the information from elsewhere (eg: section headers). Interestingly, the alpha-elf loader pulls in the section headers already and ddb/db_elf.c uses them extensively. The drawback of using the symbols and data pointed to by the section headers is that it's removed by strip(1) while the information we get from the tables intended for dynamic linkers survives a strip(). We could [in theory] run with a stripped kernel and still have internal symbol support, modules and have kvm_nlist() use an ioctl interface or something. Cheers, -Peter