From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from omta03sl.mx.bigpond.com (omta03sl.mx.bigpond.com [144.140.92.155]) by ozlabs.org (Postfix) with ESMTP id F3262DDEEC for ; Sat, 30 Jun 2007 20:57:42 +1000 (EST) Date: Sat, 30 Jun 2007 20:27:41 +0930 From: Alan Modra To: Paul Mackerras Subject: Re: Please pull from 'for_paulus' branch Message-ID: <20070630105741.GD497@bubble.grove.modra.org> References: <18052.42546.78964.805791@cargo.ozlabs.ibm.com> <18053.65260.136499.154352@cargo.ozlabs.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <18053.65260.136499.154352@cargo.ozlabs.ibm.com> Cc: Arnd Bergmann , linuxppc-dev@ozlabs.org List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Sat, Jun 30, 2007 at 04:57:48PM +1000, Paul Mackerras wrote: > Andreas Schwab writes: > > > The assembler should be called with -a64, which has the effect of > > defaulting to -mppc64. > > Yes, the assembler is called with -a64, but that doesn't appear to > have the effect of defaulting to -mppc64. > > [Alan: this thread is about the fact that compiling C code with "gcc > -m64 -mcpu=powerpc64" will give assembler errors, complaining about > instructions like sldi, whereas compiling with "gcc -m64" doesn't.] Eh? For a long time, powerpc gcc has passed "-many" to the assembler, which should avoid this sort of problem. Broken toolchain, I'd say. -- Alan Modra IBM OzLabs - Linux Technology Centre