My system has an Intel Core processor based on the Haswell architecture, and
I'm using the corei7 tuning in my image. As it's currently implemented, the
runqemu-internal script utilizes the core2dup cpu option when running the
qemux86-64 QEMUARCH. However, in order to run my image with QEMU, I need to use
the Haswell cpu option. While it's possible to run my image by adding
'qemuparams="-cpu Haswell"' to the runqemu command, I would like to change the
cpu option in the runqemu-internal script as part of the nativesdk build so
that users of the SDK don't have to mess with anything. I think this can be
accomplished by appending the nativesdk-qemu-helper recipe to include a patch
to make the simple change. However, it gets a little strange because the recipe
includes the runqemu-internal script from the ${COREBASE} as such:
SRC_URI = "file://${COREBASE}/scripts/runqemu-internal"
Because of the dynamic nature of ${COREBASE}, I'm not exactly sure how to
prepare my patch since the file will be located in some arbitrary folder within
the ${WORKDIR}. I'm assuming I can do something creative in do_patch(), but I'm
not sure where to start. Any suggestions?
Thanks,
Thomas
--
_______________________________________________
yocto mailing list
[email protected]
https://lists.yoctoproject.org/listinfo/yocto
