On 8/22/18 2:57 PM, André Draszik wrote: > On Wed, 2018-08-22 at 14:11 +0200, Richard Leitner wrote: >> Hi André, >> >> On 8/22/18 2:04 PM, André Draszik wrote: >>> Hi Richard, >>> >>> I've noticed that I had one more uncommitted but necessary patch lying >>> around to make the aarch32 port work with musl. >> >> Thank you for noticing! >> >> We should definitely add qemu test-images with musl. So similar issues >> pop up during my automated tests ;-) >> >> Do you have a qemu musl setup which could be integrated in our oeqa? > > Well, I have my hacked qemu cortex a7 machine that doesn't work with oeqa. > For musl, I have my own distro, but not sure how the rest of the Yocto > project does the automated musl testing - there might be something more > standard... > > Because of the different hotspot versions, it would really need to test all > three ports, default, aarch64, aarch32 and each with glibc and musl.
ACK. Plus OpenJDK-7... And in the future hopefully 9 or 10 ;-) > > BTW, I just today saw that meta-ivi has a qemu cortex a9 here > http://git.yoctoproject.org/cgit/cgit.cgi/meta-ivi/tree/meta-ivi-bsp > maybe that works better with oeqa? Thanks for that pointer... I hacked together a similar machine based on some staled patches from oe-core [1] and meta-qemuarma9 [2] which works for oeqa. I'll try to get the best out of the three and create a patch for oe-core. regards;Richard.L > > > Cheers, > Andre' > > [1] https://patchwork.openembedded.org/project/oe-core/patches/?q=qemuarma9 [2] https://github.com/WindRiver-Labs/meta-qemuarma9 -- _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.openembedded.org/mailman/listinfo/openembedded-devel
