From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1766002AbYDNSis (ORCPT ); Mon, 14 Apr 2008 14:38:48 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1760174AbYDNSik (ORCPT ); Mon, 14 Apr 2008 14:38:40 -0400 Received: from wx-out-0506.google.com ([66.249.82.229]:17779 "EHLO wx-out-0506.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1760013AbYDNSij (ORCPT ); Mon, 14 Apr 2008 14:38:39 -0400 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=kIbLGMx0G1LkiK4D3qnlw6/s1S6Eu52cXl+4OvBAymBKnGJ/BLTd/s4mAGXNd3zqUVVpYziNrOPlWv4tVcac5cppEg+idozYKwr8l6FdUCJEUhQuwHfQLYzyDQbhShJ3wxHm0YsR1Z2Jnx1yZ39Qp6kJ2W1uel1oqW/dg8FTEaI= Message-ID: Date: Mon, 14 Apr 2008 20:38:32 +0200 From: "Bart Van Assche" To: "Peter Zijlstra" Subject: Re: API documentation (was [PATCH] Replace completions with semaphores) Cc: "Jonathan Corbet" , "Matthew Wilcox" , "Roland Dreier" , "Ingo Oeser" , "Daniel Walker" , linux-kernel@vger.kernel.org, "Linus Torvalds" , "Ingo Molnar" In-Reply-To: <1208194395.7129.13.camel@twins> MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Content-Disposition: inline References: <16641.1208193124@vena.lwn.net> <1208194395.7129.13.camel@twins> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Apr 14, 2008 at 7:33 PM, Peter Zijlstra wrote: > Books should only be used to obtain the general picture, any details > will be instantly-obsolete, esp at the pace Linux changes. > > Most of the concepts from LDD3 are still valid, many of the details are > dead wrong. > > Can't we make LDD4 a high level book, explcitly mentioning how people > should go about obtaining details? Like go ask on #kernelnewbies and the > sorts. > > The thing I always tell #kernelnewbies people is to look at a related > driver (of course that kite doesn't always fly). Another good way to > learn stuff is to just read the implementation. > > A 'trick' that is often useful is to look in git to see how something > was changed, provided you knew how to do it some time in the past. What I remember from the first time I wrote a network driver is that reading the chapter in LDD3 about network drivers was the fastest way to get started. Learning how to write a network driver by reverse engineering existing drivers would have required much more time. My opinion is that mentioning all relevant details in a book makes sense. Bart.