On Dec 20, 2015 2:06 PM, "William Hermans" <[email protected]> wrote: >> >> The elf errors are just an annoyance.. >> >> The real problem, the headers package doesn't build correctly when cross built.. > > Really ? Never installed them myself, as *if* I'm already cross compiling the kernel, wtf do I need the header files on the target for ? ;)
That's my thought too.. You already have the full source, build you module against that.. For some reason, when cross built, a 'few' things get built for the 'host'. It's been getting better.. 3.8 was terrible, 3.14 usable, 4.1 (haven't tested as the driver I was testing went mainline...) > > > > On Sun, Dec 20, 2015 at 11:21 AM, Robert Nelson <[email protected]> wrote: >> >> The elf errors are just an annoyance.. >> >> The real problem, the headers package doesn't build correctly when cross built.. >> >> On Dec 20, 2015 11:54 AM, "Fabian Dalbert" <[email protected]> wrote: >> > >> > These wrong ELF class errors are recognised as a fakeroot bug, it seems. Not that this would help me... >> > >> > https://bugs.launchpad.net/ubuntu/+source/fakeroot/+bug/1001129 >> > >> > >> > On Sunday, December 20, 2015 at 6:51:44 PM UTC+1, Fabian Dalbert wrote: >> >> >> >> I installed these i386 libraries as described, so that can't be the problem. The build_kernel.sh script downloads it's own toolchain, which is gcc-linaro-4.9-2015.05-x86_64_arm-linux-gnueabihf, everything compiles without error, but as stated the architecture problem appears after installing the .deb files. Now, when I manually download the gcc-linaro-arm-linux-gnueabihf-4.9-2014.09_linux toolchain as described in Robert's guide and set the system.sh to use this one, I get a constant stream of errors: >> >> >> >> ERROR: ld.so: object 'libfakeroot-sysv.so' from LD_PRELOAD cannot be preloaded (wrong ELF class: ELFCLASS64): ignored. >> >> >> >> installing fakeroot:i386 does not help at all, the error just changes to: >> >> >> >> ERROR: ld.so: object 'libfakeroot-sysv.so' from LD_PRELOAD cannot be preloaded (wrong ELF class: ELFCLASS32): ignored. >> >> >> >> I don't know if using the x86_64 toolchain causes the "./scripts/recordmcount: 1: ./scripts/recordmcount: Syntax error: end of file unexpected" architecture mismatch, but I cannot get using the toolchain recommended in Robert's guide to work. >> > >> > -- >> > For more options, visit http://beagleboard.org/discuss >> > --- >> > You received this message because you are subscribed to the Google Groups "BeagleBoard" group. >> > To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. >> > For more options, visit https://groups.google.com/d/optout. >> >> -- >> For more options, visit http://beagleboard.org/discuss >> --- >> You received this message because you are subscribed to the Google Groups "BeagleBoard" group. >> To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. >> For more options, visit https://groups.google.com/d/optout. > > > -- > For more options, visit http://beagleboard.org/discuss > --- > You received this message because you are subscribed to the Google Groups "BeagleBoard" group. > To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. > For more options, visit https://groups.google.com/d/optout. -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
