From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnout Vandecappelle Date: Mon, 27 Mar 2017 00:45:17 +0200 Subject: [Buildroot] [autobuild.buildroot.net] Build results for 2017-03-24 In-Reply-To: <20170326122951.GK22954@waldemar-brodkorb.de> References: <20170325072834.09F51207FA@mail.free-electrons.com> <20170326122951.GK22954@waldemar-brodkorb.de> Message-ID: List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On 26-03-17 14:29, Waldemar Brodkorb wrote: > Hi, > Arnout Vandecappelle wrote, > >> >> >> On 25-03-17 08:28, Thomas Petazzoni wrote: >>> or1k | qt-4.8.7 | NOK | http://autobuild.buildroot.net/results/fa37b4cee01301dc01394fb279e7a16f10e3a6b5 | ORPH >> >> This looks to me like a compiler bug. It seems that with -fPIC, static class >> members are generated with a 16-bit GOT entry. That means that they can only be >> used in binaries smaller than 64K... >> >> Looks like Waldemar asked about this at some point on #openrisc but got no >> response. >> >> What to do with this? > > Can we disable this package for or1k? Yes, but it will probably fail for any C++ package that uses a static member function. And for libraries, it will only fail when linking it into something larger than 64K... Perhaps we should check a few other C++ packages. > May be things getting better when gcc7 is out and might include or1k > support. > > https://lists.librecores.org/pipermail/openrisc/2017-February/000501.html I haven't checked but I rather suspect this is a binutils issue... Regards, Arnout -- Arnout Vandecappelle arnout at mind be Senior Embedded Software Architect +32-16-286500 Essensium/Mind http://www.mind.be G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle GPG fingerprint: 7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF