From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755511Ab0E1GG1 (ORCPT ); Fri, 28 May 2010 02:06:27 -0400 Received: from mail-gy0-f174.google.com ([209.85.160.174]:56176 "EHLO mail-gy0-f174.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1758866Ab0E1GFv convert rfc822-to-8bit (ORCPT ); Fri, 28 May 2010 02:05:51 -0400 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type:content-transfer-encoding; b=KuHMCnFuxaS59LI0z81/2QSnhMB2pLyOsBLBIMAnC20YHO3uhizXP79dUqPjYQc1U0 YA1f+oZDK5cBZF0ZUw7SA5J++BpXVW/jNzTRYLOs3kQi4gVfw1VXsw8gY8cEll2/XS8Z 1Ju/yDjrTfRgjNFBuwn5ZANr6ppoZ+lyjyDMk= MIME-Version: 1.0 In-Reply-To: <1274863506-14168-1-git-send-email-vapier@gentoo.org> References: <1274815467-14584-1-git-send-email-vapier@gentoo.org> <1274863506-14168-1-git-send-email-vapier@gentoo.org> From: Mike Frysinger Date: Fri, 28 May 2010 02:05:27 -0400 Message-ID: Subject: Re: [PATCH 1/2 v2] FLAT: split the stack & data alignments To: linux-kernel@vger.kernel.org Cc: uclinux-dev@uclinux.org, David Howells , David McCullough , Greg Ungerer , Paul Mundt , uclinux-dist-devel@blackfin.uclinux.org, microblaze-uclinux@itee.uq.edu.au, Michal Simek , linux-m32r@ml.linux-m32r.org, Hirokazu Takata , Yoshinori Sato 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 On Wed, May 26, 2010 at 04:45, Mike Frysinger wrote: > The stack and data have different alignment requirements, so don't force > them to wear the same shoe.  Increase the data alignment to match that > which the elf2flt linker script has always been using: 0x20 bytes.  Not > only does this bring the kernel loader in line with the toolchain, but > it also fixes a swath of gcc tests which try to force larger alignment > values but randomly fail when the FLAT loader fails to deliver. btw, a follow up patch might be to move the shared lib identifiers from the start of the data section to the end of it so that the re-aligning isnt necessary (we'd get a 4k page alignment from mmap and such). but i cant seem to figure out how these identifiers are being read/written. otherwise, the fact that we're force aligning to 0x20 bytes means that there is always room for 8 identifiers ... no point in flipping between 1 or 4, at least from this point of view ... -mike