From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261274AbULSG5n (ORCPT ); Sun, 19 Dec 2004 01:57:43 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261275AbULSG5m (ORCPT ); Sun, 19 Dec 2004 01:57:42 -0500 Received: from parcelfarce.linux.theplanet.co.uk ([195.92.249.252]:51655 "EHLO www.linux.org.uk") by vger.kernel.org with ESMTP id S261273AbULSG5k (ORCPT ); Sun, 19 Dec 2004 01:57:40 -0500 Date: Sun, 19 Dec 2004 06:57:39 +0000 From: Matthew Wilcox To: linux-kernel@vger.kernel.org Cc: linux-fsdevel@vger.kernel.org Subject: Removing i_sock Message-ID: <20041219065739.GE7113@parcelfarce.linux.theplanet.co.uk> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.4.1i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org The i_sock member of struct inode is redundant with S_ISSOCK(i_mode). Here's a series of 6 patches to remove it: http://ftp.linux.org.uk/pub/linux/willy/patches/i_sock/ I'm sending them to their respective maintainers now, and I'll submit a series after 2.6.10 is out. Removing i_sock from struct inode actually doesn't reduce the size of struct inode -- it fits into a gap between two other members. But it may help someone else put struct inode on a diet, and it reduces confusion. -- "Next the statesmen will invent cheap lies, putting the blame upon the nation that is attacked, and every man will be glad of those conscience-soothing falsities, and will diligently study them, and refuse to examine any refutations of them; and thus he will by and by convince himself that the war is just, and will thank God for the better sleep he enjoys after this process of grotesque self-deception." -- Mark Twain