From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261612AbVFHUui (ORCPT ); Wed, 8 Jun 2005 16:50:38 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261613AbVFHUuh (ORCPT ); Wed, 8 Jun 2005 16:50:37 -0400 Received: from ns1.suse.de ([195.135.220.2]:23182 "EHLO mx1.suse.de") by vger.kernel.org with ESMTP id S261612AbVFHUuV (ORCPT ); Wed, 8 Jun 2005 16:50:21 -0400 From: Marcus Meissner To: "David S. Miller" , linux-kernel@vger.kernel.org Subject: Re: [PATCH] ppc64: Fix PER_LINUX32 behaviour In-Reply-To: <20050608.121950.104038734.davem@davemloft.net.suse.lists.linux.kernel> X-Newsgroups: suse.lists.linux.kernel User-Agent: tin/1.6.2-20030910 ("Pabbay") (UNIX) (Linux/2.6.11.10-3-ppc64 (ppc64)) Message-Id: <20050608205017.861757D1D4@grape.suse.de> Date: Wed, 8 Jun 2005 22:50:17 +0200 (CEST) Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org In article <20050608.121950.104038734.davem@davemloft.net.suse.lists.linux.kernel> you wrote: > From: Arnd Bergmann > Date: Wed, 8 Jun 2005 20:49:48 +0200 > >> With the LINUX32 personality, you can build 32 bit binaries through >> autoconf, rpmbuild, or the kernel without pretending to be >> cross-compiling. It may not be the best solution, but people seem to >> rely on it and the patch brings ppc64 in line with how it works on >> the other architectures. > > I totally agree, this has a large precedence on many platforms > and there are even gcc frontends that check the uname output > to decide what code model to output by default. Actually what you then do (which is the standard way) is powerpc32 bash --login which gives you a shell with 32bit personality and work from this. Ciao, Marcus