From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 5 Mar 2003 20:58:26 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 5 Mar 2003 20:58:26 -0500 Received: from to-wiznet.redhat.com ([216.129.200.2]:61692 "EHLO touchme.toronto.redhat.com") by vger.kernel.org with ESMTP id ; Wed, 5 Mar 2003 20:58:25 -0500 Date: Wed, 5 Mar 2003 21:08:56 -0500 From: Benjamin LaHaise To: Ulrich Drepper Cc: Andi Kleen , Linux Kernel Subject: Re: Better CLONE_SETTLS support for Hammer Message-ID: <20030305210856.B16093@redhat.com> References: <3E664836.7040405@redhat.com> <20030305190622.GA5400@wotan.suse.de> <3E6650D4.8060809@redhat.com> <20030305212107.GB7961@wotan.suse.de> <3E668267.5040203@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5.1i In-Reply-To: <3E668267.5040203@redhat.com>; from drepper@redhat.com on Wed, Mar 05, 2003 at 03:04:07PM -0800 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, Mar 05, 2003 at 03:04:07PM -0800, Ulrich Drepper wrote: > I wouldn't want to either in your position. You caused this whole mess > and now you're not willing to fix it. I disagree. We don't need to use a segment register to get the thread library equivalent of "current". Changing the segment registers on a per process basis is a waste of time in the context switch. Instead, making x86-64 TLS support based off of the stack pointer, or even using a fixed per-cpu segment register such that gs:0 holds the pointer to the thread "current" would be better. Make the users of threads suffer, not every single application and syscall in the system. -ben -- Junk email? aart@kvack.org