I'm not sure if it's caused by this binutils upgrade (because openssl in oe-core is also using version script, imported from debian) or some change in cmake/openssl, but today all recipes using cmake are failing with:
| DEBUG: Executing shell function do_configure | cmake: sysroots/x86_64-linux/usr/bin/../lib/libcrypto.so.1.0.0: version `OPENSSL_1.0.0' not found (required by cmake) | cmake: sysroots/x86_64-linux/usr/bin/../lib/libssl.so.1.0.0: version `OPENSSL_1.0.1' not found (required by cmake) | cmake: sysroots/x86_64-linux/usr/bin/../lib/libssl.so.1.0.0: version `OPENSSL_1.0.0' not found (required by cmake) | WARNING: exit code 1 from a shell command. Maybe it's another symptom of missing openssl-native dependency in cmake-native (just like libidn-native issue reported here https://bugzilla.yoctoproject.org/show_bug.cgi?id=9639) - and cmake-native was originally linked against host's openssl which was built with version script and binutils-2.24 (Ubuntu-14.04), but then it's executed with libssl and libcrypto from native sysroot, where the version script didn't work: $ ldd sysroots/x86_64-linux/usr/bin/cmake sysroots/x86_64-linux/usr/bin/cmake: sysroots/x86_64-linux/usr/bin/../lib/libcrypto.so.1.0.0: version `OPENSSL_1.0.0' not found (required by sysroots/x86_64-linux/usr/bin/cmake) sysroots/x86_64-linux/usr/bin/cmake: sysroots/x86_64-linux/usr/bin/../lib/libssl.so.1.0.0: version `OPENSSL_1.0.1' not found (required by sysroots/x86_64-linux/usr/bin/cmake) sysroots/x86_64-linux/usr/bin/cmake: sysroots/x86_64-linux/usr/bin/../lib/libssl.so.1.0.0: version `OPENSSL_1.0.0' not found (required by sysroots/x86_64-linux/usr/bin/cmake) linux-vdso.so.1 => (0x00007ffed2fce000) libdl.so.2 => /lib/x86_64-linux-gnu/libdl.so.2 (0x00007fb631276000) libssl.so.1.0.0 => sysroots/x86_64-linux/usr/bin/../lib/libssl.so.1.0.0 (0x00007fb63100c000) libcrypto.so.1.0.0 => sysroots/x86_64-linux/usr/bin/../lib/libcrypto.so.1.0.0 (0x00007fb630bc1000) libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007fb6308bd000) libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007fb6305b7000) libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007fb6303a1000) libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007fb62ffdc000) /lib64/ld-linux-x86-64.so.2 (0x00007fb63147a000) Which is sort of confirmed by removing libssl.so and libcrypto.so from native sysroot, I can execute cmake binary again. On Tue, Aug 9, 2016 at 5:24 PM, Richard Purdie < richard.pur...@linuxfoundation.org> wrote: > On Tue, 2016-08-09 at 07:42 -0700, Khem Raj wrote: > > I could also see it on ppc. backtrace, shows the segfault is in exit > > path and happens in libc > > at this point, I think the problem is how libc is compiled with > > binutils 2.27, connman itself > > is ok. > > Its the issue here: > > https://sourceware.org/ml/glibc-bugs/2015-01/msg00274.html > https://sourceware.org/bugzilla/show_bug.cgi?id=17908 > > Basically, if you remove the global _IO_stdin_used symbol, it triggers > compatibility code which crashes. > > I've confirmed that if I add that symbol to the version-script in > connman, things work again. > > Any idea how we raise the priority of this issue. There are no comments > on the bug despite it having been posted a while ago :(. > > Cheers, > > Richard > -- > _______________________________________________ > Openembedded-devel mailing list > openembedded-de...@lists.openembedded.org > http://lists.openembedded.org/mailman/listinfo/openembedded-devel >
-- _______________________________________________ Openembedded-core mailing list Openembedded-core@lists.openembedded.org http://lists.openembedded.org/mailman/listinfo/openembedded-core
