From owner-freebsd-current Tue Sep 1 23:34:39 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id XAA24294 for freebsd-current-outgoing; Tue, 1 Sep 1998 23:34:39 -0700 (PDT) (envelope-from owner-freebsd-current@FreeBSD.ORG) Received: from cimlogic.com.au (cimlog.lnk.telstra.net [139.130.51.31]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id XAA24275 for ; Tue, 1 Sep 1998 23:34:34 -0700 (PDT) (envelope-from jb@cimlogic.com.au) Received: (from jb@localhost) by cimlogic.com.au (8.8.8/8.8.7) id QAA27724; Wed, 2 Sep 1998 16:45:36 +1000 (EST) (envelope-from jb) From: John Birrell Message-Id: <199809020645.QAA27724@cimlogic.com.au> Subject: Re: Standardizing a BSD/ELF ABI... In-Reply-To: <19980902154024.E606@freebie.lemis.com> from Greg Lehey at "Sep 2, 98 03:40:24 pm" To: grog@lemis.com (Greg Lehey) Date: Wed, 2 Sep 1998 16:45:36 +1000 (EST) Cc: mike@smith.net.au, tlambert@primenet.com, jb@cimlogic.com.au, freebsd-current@FreeBSD.ORG X-Mailer: ELM [version 2.4ME+ PL40 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Greg Lehey wrote: > > You will have to convince the NetBSD people to stop randomly changing > > their system calls first. John Birrell discussed this at some length > > while I was there last week, and it's about at the point where we are > > going to need an ABI shim for them. > > What about BSDI? Presumably they've been stable for a while. Having > those two compatible would also be an incentive for the NetBSD people. I think we're all compatible because we're using binutils (though not necessarily the same version). This means that an elf32 binary opened by FreeBSD will be recognised as elf32. If it has an interpreter section and we have emulation support for that, then we can probably handle it. If no interpreter section, we have to take a stab at recognising a branded executable. Once we've resolved either the interpreter or the brand, we know the emulation and can therefore (supposedly) deal with the different syscall interface between userland and the kernel. Tracking the syscall design of other operating systems like NetBSD is a pain unless you only work to their releases and forget about -current. I wish the *BSD groups would agree on a standard syscall interface and publish it. Then we wouldn't need to waste effort working on emulating each other. -- John Birrell - jb@cimlogic.com.au; jb@freebsd.org http://www.cimlogic.com.au/ CIMlogic Pty Ltd, GPO Box 117A, Melbourne Vic 3001, Australia +61 418 353 137 To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message