From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759444AbYEFC6o (ORCPT ); Mon, 5 May 2008 22:58:44 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752878AbYEFC6d (ORCPT ); Mon, 5 May 2008 22:58:33 -0400 Received: from mx1.redhat.com ([66.187.233.31]:58576 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752562AbYEFC6c (ORCPT ); Mon, 5 May 2008 22:58:32 -0400 Message-ID: <481FC94C.8090203@redhat.com> Date: Mon, 05 May 2008 19:58:20 -0700 From: Ulrich Drepper Organization: Red Hat, Inc. User-Agent: Thunderbird 2.0.0.12 (X11/20080226) MIME-Version: 1.0 To: Andrew Morton CC: linux-kernel@vger.kernel.org, netdev@vger.kernel.org, davidel@xmailserver.org, mtk.manpages@gmail.com, torvalds@linux-foundation.org Subject: Re: [PATCH 02/18] flag parameters: socket and socketpair References: <200805050342.m453gktl029817@devserv.devel.redhat.com> <20080505185850.789d870a.akpm@linux-foundation.org> <481FC38A.6070805@redhat.com> <20080505194221.4349418b.akpm@linux-foundation.org> In-Reply-To: <20080505194221.4349418b.akpm@linux-foundation.org> X-Enigmail-Version: 0.95.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Andrew Morton wrote: > Sorry to sound slow, but what do we gain from all this? Your covering > description was heavy in references to some earlier discussion to which I > obviously paid little attention and now I'm lost. What we gain is that the only change for a new flag is a new entry in the appropriate table. The reference to non-sequential file descriptors refers to http://lwn.net/Articles/236724/ Not this specific incarnation of the functionality, but something like this. A few more points: - - for the socket functions we can share the table - - in most cases the loop isn't even entered since the flag value is zero. In any case, the loop is ended once all flags have been mapped - - in case this code is really called often, than the code is in the cache and it is shorter than each of possible multiple copies of the unrolled loop Anyway, we spent already far too much time on this. If you want to see it changed tell me. It's really not important. - -- ➧ Ulrich Drepper ➧ Red Hat, Inc. ➧ 444 Castro St ➧ Mountain View, CA ❖ -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.7 (GNU/Linux) iD8DBQFIH8lM2ijCOnn/RHQRAta6AKC39ty8FdiRYjRxxHDoG2T/N1hhywCeKPcZ L/iDtbf9DozMmEfoLY8nlbk= =B+YO -----END PGP SIGNATURE-----