From owner-svn-src-all@FreeBSD.ORG Mon Aug 15 08:05:14 2011 Return-Path: Delivered-To: svn-src-all@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id DF8F810657D6; Mon, 15 Aug 2011 08:05:14 +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 BB1DB8FC17; Mon, 15 Aug 2011 08:05:14 +0000 (UTC) Received: from fledge.watson.org (fledge.watson.org [65.122.17.41]) by cyrus.watson.org (Postfix) with ESMTPS id 0F91546B32; Mon, 15 Aug 2011 04:05:14 -0400 (EDT) Date: Mon, 15 Aug 2011 09:05:13 +0100 (BST) From: Robert Watson X-X-Sender: robert@fledge.watson.org To: Alexander Best In-Reply-To: <20110814170621.GA4189@freebsd.org> Message-ID: References: <201108131603.p7DG3e80023093@svn.freebsd.org> <4E47EF64.2070405@gmail.com> <20110814170621.GA4189@freebsd.org> User-Agent: Alpine 2.00 (BSF 1167 2008-08-23) MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Cc: svn-src-head@freebsd.org, Niclas Zeising , src-committers@freebsd.org, svn-src-all@freebsd.org Subject: Re: svn commit: r224842 - head/sys/kern X-BeenThere: svn-src-all@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: "SVN commit messages for the entire src tree \(except for " user" and " projects" \)" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 15 Aug 2011 08:05:15 -0000 On Sun, 14 Aug 2011, Alexander Best wrote: >> It might be worth mentioning this in UPDATING or similar, since a kernel >> built with clang (I haven't tried gcc) during this window will make it >> impossible to do a buildworld without first rebuilding the kernel with this >> fix. It also seems to break at least portsnap, and possibly other tools as >> well. > > +1. please also mention the KPI change to fget()! so many people have had > issues, where 3rd party drivers - mostly the nvidia binary driver - got > broken. Now in place. However, I think it's fairly likely that nVidia binary drivers will see further disruption in the future. Other than preventing the driver from loading when it's out of sync with __FreeBSD_version (which I believe we now do?), I'm not sure what further is to be done there. In general, we don't tweak things like fget(9) very often, but we do make changes to fileops with moderate frequency, and the failure modes there are quite a bit more subtle. It would be useful to see a KBI analysis of the nVidia module so we can decide how to moderate that risk. (We've done some work on formalising the relationship between network device drivers and the kernel, which has led is to have specific policies on data structure changes, for example -- but we haven't done symbol dependencies as yet.) Robert