From owner-freebsd-new-bus Sat Feb 26 3: 1:37 2000 Delivered-To: freebsd-new-bus@freebsd.org Received: from sasami.jurai.net (sasami.jurai.net [63.67.141.99]) by hub.freebsd.org (Postfix) with ESMTP id CA4D137BC1B for ; Sat, 26 Feb 2000 03:01:30 -0800 (PST) (envelope-from winter@jurai.net) Received: from localhost (winter@localhost) by sasami.jurai.net (8.9.3/8.8.7) with ESMTP id GAA77157; Sat, 26 Feb 2000 06:01:26 -0500 (EST) Date: Sat, 26 Feb 2000 06:01:26 -0500 (EST) From: "Matthew N. Dodd" To: Jeroen Ruigrok/Asmodai Cc: new-bus@FreeBSD.ORG Subject: Re: New version of the newbus draft In-Reply-To: <20000226113136.S79013@daemon.ninth-circle.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-new-bus@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG "Newbus is the new bus abstraction layer architecture which saw its introduction in FreeBSD 4.0." Actually it came in with the Alpha port IIRC. Check 3.4. "Newbus is the new bus..." sounds clunky, as does "abstraction layer architecture". If a sentence costs you more than $1.50 then you're using too many 50c words. :) - Suggest using a bullet list for enumeration of multiple items. (p1s3) - The Alpha doesn't have 'nexus' (p3s1) - A device is really the sum of its methods. A device is a device; a bus is a device with bus methods. I'm not sure what you'd call a 'bridge' in newbus terms. (p4s1) - 'map its resources' and the implications of the text following aren't quite clear. Resources are reserved and allocated (which isn't very good either since we really want to reserve them, then activate them, but that distinction is yet implemented in a coherent manner.) In addition it is implied that the behavior of resource methods is -always- to call the parent when that isn't always the case. Explaining default methods and the goal of resource allocation/reservation/foo would be a better focus for this paragraph since you could demonstrate the action that each layer takes on resource allocation/reservation/foo. The existing example is a pretty good start though. - Newbus doesn't really have anything to do with bus_space at this point. We wish they were more intimate but they are really separate. Same thing with bus_dma. - Newbus doesn't 'allows for definitions of interface methods...'; thats the entire ball of wax. bus_if.m and device_if.m provide the basic structure to implement a hierarchy of devices. This functionality is implicit. A more top down approach that brings the method definitions and rules of hierarchy and inheritance into focus early in the document would be good. Anyhow, you've got a good start. -- | Matthew N. Dodd | '78 Datsun 280Z | '75 Volvo 164E | FreeBSD/NetBSD | | winter@jurai.net | 2 x '84 Volvo 245DL | ix86,sparc,pmax | | http://www.jurai.net/~winter | This Space For Rent | ISO8802.5 4ever | To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-new-bus" in the body of the message