On 08/11/2014 07:12 PM, Peter A. Bigot wrote:
On 08/11/2014 02:02 PM, Peter A. Bigot wrote:
The program below built on the target with the MACHINE=beaglebone gcc-4.9.1 compiler from Yocto/OpenEmbedded poky master produces this error:

beaglebone[52]$ g++ -std=c++1y -pthread test.cc && ./a.out
starting
joining
pure virtual method called
terminate called without an active exception
Aborted (core dumped)

When the program is recompiled with the defines for __GCC_HAVE_SYNC_COMPARE_AND_SWAP_X enabled as suggested at https://groups.google.com/d/msg/automatak-dnp3/Jisp_zGhd5I/ck_Cj6nO8joJ it works:

beaglebone[53]$ g++ -std=c++1y -pthread test.cc && ./a.out
starting
joining
doit
done

Preliminary analysis confirms that the built-ins for those defines are not being added by the compiler because it thinks the target doesn't support those operations. Nonetheless, it doesn't use the substitutes that are obviously available.

Can anybody recall anything about the way GCC is built under OE that would explain this?

Not an OE problem. See: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62100

I'm testing a patch locally and awaiting GCC maintainer comment before proposing it for OE.

The GNU folks objected to the workaround and suggested that the behavior is evidence that libstdc++ was not built with the same ABI flags as the compiler uses by default.

I built the release version of gcc 4.9.1 on the beaglebone and confirmed that it works fine.

This is indeed an OE-Core bug.

I've just sent a series of cleanup patches to gcc that may make it easier to detect differences. I've spent some time changing the "obvious" things and have been unable to figure out where the compilation is going wrong. I'm hoping somebody else can help out; as C++11 becomes more prevalent the workaround recommended on stackoverflow and other sites will continue to propagate, and it's not the right solution.

Peter

--
_______________________________________________
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Reply via email to