From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932854AbWFMEEE (ORCPT ); Tue, 13 Jun 2006 00:04:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932857AbWFMEEE (ORCPT ); Tue, 13 Jun 2006 00:04:04 -0400 Received: from cantor2.suse.de ([195.135.220.15]:10387 "EHLO mx2.suse.de") by vger.kernel.org with ESMTP id S932854AbWFMEED (ORCPT ); Tue, 13 Jun 2006 00:04:03 -0400 From: Andi Kleen To: Alan Cox Subject: Re: broken local_t on i386 Date: Tue, 13 Jun 2006 06:02:59 +0200 User-Agent: KMail/1.9.3 Cc: Christoph Lameter , Ingo Molnar , Michal Piotrowski , Andrew Morton , linux-kernel@vger.kernel.org References: <20060609214024.2f7dd72c.akpm@osdl.org> <200606122011.52841.ak@suse.de> <1150143161.25462.26.camel@localhost.localdomain> In-Reply-To: <1150143161.25462.26.camel@localhost.localdomain> MIME-Version: 1.0 Content-Disposition: inline Message-Id: <200606130602.59504.ak@suse.de> Content-Type: text/plain; charset="iso-8859-15" Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Monday 12 June 2006 22:12, Alan Cox wrote: > Ar Llu, 2006-06-12 am 20:11 +0200, ysgrifennodd Andi Kleen: > > The segment register needs an offset. So you need the linker to generate > > the offset from the base of the per cpu segment somehow. At compile time the > > address is not known so it cannot be done then. > > What happens if you put a section at zero and a section at non-fixed > address (aligned). In the asm macros you stick the variable in both, > using the zero based one for linker symbols and the non zero based one > for data, then discard the zero based one. Interesting idea. Maybe it'll work. > That used to work for old binutils which didn't care/spot if you were > discarding material you actually linked against. Not sure what todays > binutils does with it. AFAIK it warns at least. Might get noisy. -Andi