From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750761AbWJPNqY (ORCPT ); Mon, 16 Oct 2006 09:46:24 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750763AbWJPNqY (ORCPT ); Mon, 16 Oct 2006 09:46:24 -0400 Received: from ns2.suse.de ([195.135.220.15]:54751 "EHLO mx2.suse.de") by vger.kernel.org with ESMTP id S1750761AbWJPNqX (ORCPT ); Mon, 16 Oct 2006 09:46:23 -0400 To: "Robert P. J. Day" Cc: linux-kernel@vger.kernel.org Subject: Re: thoughts on potential cleanup of semaphores? References: From: Andi Kleen Date: 16 Oct 2006 15:46:13 +0200 In-Reply-To: Message-ID: User-Agent: Gnus/5.09 (Gnus v5.9.0) Emacs/21.3 MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org "Robert P. J. Day" writes: > after submitting one patch related to semaphores and before i submit > any others, any thoughts whether any of the following clean-ups are > valid and/or worthwhile? (some are admittedly simply aesthetic but > better aesthetics is never a bad thing.) Semaphores and rwsems need a lot of cleanup. Some time ago we put spinlocks which are a lot more time critical than semaphores out of line. So the same could be done for semaphores too. This had the advantages that most of the assembler voodoo wouldn't be needed anymore and they could be simple straight C implementations shared by all architectures. -Andi