From: Thomas Petazzoni Date: Fri, 5 Jul 2013 15:01:37 +0200 Subject: [Buildroot] [PATCH] gcc: apply PowerPC patch when needed In-Reply-To: <51D6C2AA.8070703@zacarias.com.ar> References: <1373013652-4162-1-git-send-email-thomas.petazzoni@free-electrons.com> <51D6C2AA.8070703@zacarias.com.ar> Message-ID: <20130705150137.773e3ad7@skate> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Gustavo Zacarias, On Fri, 05 Jul 2013 09:57:14 -0300, Gustavo Zacarias wrote: > On 07/05/2013 05:40 AM, Thomas Petazzoni wrote: > > > When converting gcc to the package infrastructure, a specific thing > > for PowerPC was forgotten: applying the conditional patch > > powerpc-link-with-math-lib.patch.conditional. This breaks the build of > > some PowerPC toolchains, with failures such as: > > > > : undefined reference to `copysignl' > > > > Signed-off-by: Thomas Petazzoni > > Acked-by: Gustavo Zacarias > Tested-by: Gustavo Zacarias Thanks :) > > Anyone knows why we have this conditional patch? Can't it be made a > > normal patch, with the appropriate modifications? Why isn't that > > upstream since all those years? > > http://gcc.gnu.org/ml/gcc-patches/2008-10/msg00292.html for the > history/thread. I remember reading this thread a long time ago. But my question is more: why hasn't the problem been fixed since 5 years? :-) > A sample failing config without the patch is > qemu_ppc_virtex_ml507_defconfig (a powerpc 440 core, soft float). Indeed, that's the one that was failing in our Jenkins builds, as can be seen from the URL above. Best regards, Thomas -- Thomas Petazzoni, Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com