From owner-freebsd-arch@FreeBSD.ORG Fri Aug 26 08:59:12 2011 Return-Path: Delivered-To: freebsd-arch@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 72E51106564A for ; Fri, 26 Aug 2011 08:59:12 +0000 (UTC) (envelope-from rwatson@FreeBSD.org) Received: from cyrus.watson.org (cyrus.watson.org [65.122.17.42]) by mx1.freebsd.org (Postfix) with ESMTP id 2C57C8FC15 for ; Fri, 26 Aug 2011 08:59:12 +0000 (UTC) Received: from fledge.watson.org (fledge.watson.org [65.122.17.41]) by cyrus.watson.org (Postfix) with ESMTPS id 52FDA46B0D; Fri, 26 Aug 2011 04:59:11 -0400 (EDT) Date: Fri, 26 Aug 2011 09:59:11 +0100 (BST) From: Robert Watson X-X-Sender: robert@fledge.watson.org To: Xin LI In-Reply-To: Message-ID: References: <1313785806.56747.YahooMailClassic@web113503.mail.gq1.yahoo.com> <175270279.98941.1313793515391.JavaMail.root@erie.cs.uoguelph.ca> User-Agent: Alpine 2.00 (BSF 1167 2008-08-23) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: vadim_nuclight@mail.ru, freebsd-arch@freebsd.org Subject: Debian/kFreeBSD (was: Re: FreeBSD problems and preliminary ways to solve) X-BeenThere: freebsd-arch@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Discussion related to FreeBSD architecture List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 26 Aug 2011 08:59:12 -0000 On Thu, 25 Aug 2011, Xin LI wrote: >>> It seems to me that FreeBSD should do what it can to support this effort. >>> Why? Well, I suspect a lot of why organizations >> >> No. FreeBSD has limited resources. We need to: >> >> a) fix our own problems >> b) develop our own unique features (see my message to Robert Watson) >> >> And supporting Debian will mean wasting resources which could be spent to >> these. This will effectively kill FreeBSD as a separate entity if our >> problems will stay unfixed and get worse due to this. >> >> Leave this work to Debian: they already have a wide community and many >> resources. Any Linux distro has more resources than BSD because they just >> only pack someone's software, and we have to also actually *develop* those >> software (most of all, kernel and libc). > > I consider Debian GNU/kFreeBSD as a platform where we can see what the > actual {performance,functionality} difference between our code and GNU ones, > which is valuable. Supporting their effort might be time consuming and > maybe not our priority, but it's good not to make their work harder because > more importantly, their existence gives us more exposure and more eyes on > our codebase as well. Debian GNU/kFreeBSD offers a number of interesting prospects: (1) Side-by-side comparison between FreeBSD and Linux kernels + features using a mature and complete userspace (2) Allow Debian users easy access to features hard to get with a stock Linux kernel: pf, 802.11 stack, Netgraph, DTrace, ZFS, Capsicum, etc. (3) Allow easier hosting of mature Linux environments on top of hybrid FreeBSD/Linux hosting: provide Linux jails on FreeBSD ISPs but with the many benefits of using the native ABI. (4) Potential stepping stone on the path from Linux to FreeBSD (or, to be fair, vice versa!). (5) Increase exposure of FreeBSD kernel features and approaches in the broder OS community. (6) Provide another motivation for third-party application developers to consider FreeBSD-specific kernel features viable for use. For example, if Debian/kFreeBSD makes developing Capsicum-aware applications on a Linux-like platform easy, then FreeBSD ports wins as well. (7) Provide an additional set of interested hands when it comes to improving the clarity of definition and robustness of our system call ABI. This is something we've been increasing our focus on over time, and especially if the Debian folk track our development HEAD, they can provide us much earlier feedback either when we mess up, or when there's something we can do to make things easier for them (and therefore, likely, us in the future). (8) Use of the Debian "brand" to promote FreeBSD as a vibrant and interesting thing for the broader community of Linux users. I think we should be supporting the Debian/kFreeBSD folks in their effort -- hence inviting them to join our developer summits, accepting patches, etc. As both the FreeBSD and the Debian communities pride themselves in being a bit persnickety about the spelling and details, there will inevitably be some disagreements about approach. However, we may find that we can make their lives a lot easier by making relatively small changes to our kernel, and that they have extremely useful feedback to bring us on our kernel implementation. Robert