From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from moutng.kundenserver.de ([212.227.17.9]) by metis.ext.pengutronix.de with esmtp (Exim 4.72) (envelope-from ) id 1RxaqY-0004BM-Sr for ptxdist@pengutronix.de; Wed, 15 Feb 2012 10:08:20 +0100 Message-ID: <4F3B75FD.20508@corscience.de> Date: Wed, 15 Feb 2012 10:08:13 +0100 From: =?ISO-8859-1?Q?Andreas_Bie=DFmann?= MIME-Version: 1.0 References: <2236550.JcqvJ5vtdt@ws-stein> <10381918.aAGBVVShah@ws-stein> <4F3B54E8.8070202@biessmann.de> <1846417.Lzzx8S0Ecs@ws-stein> In-Reply-To: <1846417.Lzzx8S0Ecs@ws-stein> Subject: Re: [ptxdist] Detect if x86 target is 32 or 64bit? Reply-To: ptxdist@pengutronix.de List-Id: PTXdist Development Mailing List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Sender: ptxdist-bounces@pengutronix.de Errors-To: ptxdist-bounces@pengutronix.de To: ptxdist@pengutronix.de Cc: =?ISO-8859-1?Q?Andreas_Bie=DFmann?= , Klaus.Hachmeister@l-3com.com Hi Alexander, On 15.02.2012 08:30, Alexander Stein wrote: > Hello Andreas, > = > Am Mittwoch, 15. Februar 2012, 07:47:04 schrieb Andreas Bie=DFmann: >> is there any conclusion to this discussion? >> I startet to build a x86_64 rootfs (on top of i586-qemu-bsp) and failed >> to build sylinux due to some wrong decisions in syslinux.make (another >> discussion ..). >> I started defining ARCH_X86_64 as optimization of ARCH_X86. Alexander, >> are you building x86_64 stuff too ore did you only ask cause of the >> ability to build gnu-efi for ia32, x86_32 and ia64? > = > I didn't work on this topic again. But yes, I asked only for gnu-efi stuf= f. Until now we are only building 32bit. Ok. >>> Another problem prio to this is a 64bit toolchain in ptxdist. Up to now= there is no OSELAStoolchain for x86_64. >> >> Can you please test my OSELAS.Toolchain x86_64 patches? > = > IIRC from other mails, do you need a x86_64 host to use a x86_64 toolchai= n? Well, I'm running a x86_64 host, so this wouldn't be a problem. No, that is a misunderstanding here. You can build the x86_64-unknown-linux-gnu OSELAS.Toolchain on any host you want (even Mac OS X as I did). If you have a host that has the same GNU triplet as the target (and that may be the fact for some users with x86 build host and x86 target toolchain but not only he x86_64 toolchain) the autotools mechanism will end up in deciding for native build which means e.g. gcc without prefixed arch triplet. Therefore one of my patches adds explicit '--build=3D' switch to cross-tools configure to fool the autotools guessing mechanism by providing a non-guessable arch triplet (e.g. x86_64-host-linux-gnu -> see GNU_HOST in /rules/pre/Rules.make). This let the autoconf logic think it has a x86_64-host-linux-gnu build system building a x86_64-unknown-linux-gnu target toolchain running on a x86_64-unknown-linux-gnu host system which would be 'cross compiling'. The resulting tools have the arch triplet prefixed and everything should be fine. I hope this is clear now. Feel free to test the OSELAS.Toolchain patches, would be great to get some feedback. best regards Andreas Bie=DFmann -- = ptxdist mailing list ptxdist@pengutronix.de