From owner-freebsd-hackers Thu Jan 7 15:24:57 1999 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id PAA23917 for freebsd-hackers-outgoing; Thu, 7 Jan 1999 15:24:57 -0800 (PST) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from iquest3.iquest.net (iquest3.iquest.net [209.43.20.203]) by hub.freebsd.org (8.8.8/8.8.8) with SMTP id PAA23912 for ; Thu, 7 Jan 1999 15:24:56 -0800 (PST) (envelope-from toor@y.dyson.net) Received: (qmail 1565 invoked from network); 7 Jan 1999 23:24:24 -0000 Received: from dyson.iquest.net (HELO y.dyson.net) (198.70.144.127) by iquest3.iquest.net with SMTP; 7 Jan 1999 23:24:24 -0000 Received: (from root@localhost) by y.dyson.net (8.9.1/8.9.1) id SAA01200; Thu, 7 Jan 1999 18:24:23 -0500 (EST) Message-Id: <199901072324.SAA01200@y.dyson.net> Subject: Re: questions/problems with vm_fault() in Stable In-Reply-To: <199901072306.PAA35328@apollo.backplane.com> from Matthew Dillon at "Jan 7, 99 03:06:21 pm" To: dillon@apollo.backplane.com (Matthew Dillon) Date: Thu, 7 Jan 1999 18:24:23 -0500 (EST) Cc: tlambert@primenet.com, dyson@iquest.net, pfgiffun@bachue.usc.unal.edu.co, freebsd-hackers@FreeBSD.ORG From: "John S. Dyson" Reply-To: dyson@iquest.net X-Mailer: ELM [version 2.4ME+ PL38 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Matthew Dillon said: > > Huh? Where's the cache? What? No cache? Are you nuts? Or are you > assuming the client-side caches the object locally. But then, where's > the cache coherency? Right out the window. > .... > > There ISN'T only one object. I will repeat that a thousand times. > Your entire model assumes that there is only one object or that all the > VFS layers can be collapsed. Neither assumption works. > Those ARE the key points. The cool thing (and I am agreeing with you Matt) is that the VM method is an abstraction that fully supports the file I/O model also. The file I/O method doesn't support an mmap model at all (at least in a coherent way.) Rather than spending lots of time trying to make a (conventional) VFS scheme work, and adding lots of hacks to simulate coherency, we can start with a hybrid Heidemann/VM framework with proper invalidation protocols and have everything work from the beginning!!! I don't think that any of us are suggesting that the Heidemann framework be ignored. We are suggesting that it should be applied in a different way that gives us alot more flexibility and consistancy. This would allow the infrastructure to do more of the work (almost automatically supporting coherency) rather than adding hackery to a filesystem scheme to force VM mmap coherency. This scheme that Matt and I propose would simplify and truely define the interfaces. -- John | Never try to teach a pig to sing, dyson@iquest.net | it makes one look stupid jdyson@nc.com | and it irritates the pig. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message