From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Tue, 13 May 2014 23:51:26 +0200 Subject: [Buildroot] [autobuild.buildroot.net] Build results for 2014-05-10 In-Reply-To: <20140511094535.GF4096@tarshish> References: <20140511063009.DE62C100F87@stock.ovh.net> <20140511094535.GF4096@tarshish> Message-ID: <20140513235126.212289d9@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Baruch Siach, On Sun, 11 May 2014 12:45:35 +0300, Baruch Siach wrote: > On Sun, May 11, 2014 at 08:30:09AM +0200, Thomas Petazzoni wrote: > > arm | ltrace-0.7.3 | NOK | http://autobuild.buildroot.net/results/eefa8efbe6e6a6b203c02c2977d29be9c77846cb/ > > Build log says: > > checking for backtrace in -lunwind... no > configure: error: Couldn't find or use libunwind. > > But I can't reproduce it here. Thomas, can you look into config.log? To reproduce the problem, you needed to have: libatomic_ops, libunwind, and ltrace selected, and an architecture like ARMv4 or ARMv5. The problem is that on ARMv4 and ARMv5, libatomic_ops doesn't provide real atomic operations. But libunwind did not use the libatomic_ops stuff properly, and therefore libatomic_ops was not providing emulated atomic operations for those architectures. Later on, when ltrace tried to link against libunwind, there were unresolved symbols due to this problem. I've sent a patch to the Buildroot mailing list to fix this. You're Cc'ed on the patch. Thanks, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com