Closed shenki closed 5 years ago
TODO:
[x] The guests cannot cause qemu to exit. I've submitted some qemu patches [1, 2] to add support for this, I would need to backport them and do a new qemu build.
[x] Work out how to build additional targets in order to have the device trees built. The solution in driver: build default target
doesn't work, as building modules for the multi_v7_defconfig goes over the Travis time limit
[x] Drop travis: Disable non-arm32 builds
. This is here for testing
[x] Look at refactoring the addtional config options (as is used to disable ftrace). We could instead allow targets to create their own fragments (or check fragments into the tree). For example:
echo "CONFIG_PPC_PSERIES=y" >> fragment.config
echo "CONFIG_PPC_POWERNV=y" >> fragment.config
echo "CONFIG_MTD=y" >> fragment.config
echo "CONFIG_MTD_BLOCK=y" >> fragment.config
echo "CONFIG_MTD_POWERNV_FLASH=y" >> fragment.config
ARCH=powerpc scripts/kconfig/merge_config.sh \
arch/powerpc/configs/ppc64_defconfig \
arch/powerpc/configs/le.config \
fragment.config
- [ ] Work out how to build additional targets in order to have the device trees built. The solution in
driver: build default target
doesn't work, as building modules for the multi_v7_defconfig goes over the Travis time limit
Would the dtbs
target not work?
- [ ] Look at refactoring the addtional config options (as is used to disable ftrace). We could instead allow targets to create their own fragments (or check fragments into the tree). For example:
echo "CONFIG_PPC_PSERIES=y" >> fragment.config echo "CONFIG_PPC_POWERNV=y" >> fragment.config echo "CONFIG_MTD=y" >> fragment.config echo "CONFIG_MTD_BLOCK=y" >> fragment.config echo "CONFIG_MTD_POWERNV_FLASH=y" >> fragment.config ARCH=powerpc scripts/kconfig/merge_config.sh \ arch/powerpc/configs/ppc64_defconfig \ arch/powerpc/configs/le.config \ fragment.config
Alternatively, have CONFIGS_TO_APPEND
bash array that we iterate through that architectures can define.
* [ ] The guests cannot cause qemu to exit. I've submitted some qemu patches [[1](http://patchwork.ozlabs.org/patch/1010823/), [2](http://patchwork.ozlabs.org/patch/1010825/)] to add support for this, I would need to backport them and do a new qemu build.
I use "reboot" in the guest and the "-no-reboot" qemu command line option.
* [ ] The guests cannot cause qemu to exit. I've submitted some qemu patches [[1](http://patchwork.ozlabs.org/patch/1010823/), [2](http://patchwork.ozlabs.org/patch/1010825/)] to add support for this, I would need to backport them and do a new qemu build.
I use "reboot" in the guest and the "-no-reboot" qemu command line option.
Thanks for the suggestion. I've successfully switched to this for the ppc32 config, I'll do the same here.
This renames the existing 32-bit ARM target to arm32_v7, and adds two more to test ARMv5 and ARMv6. A rootfs that can be shared by the other ISAs was added previously, so we don't need a new rootfs.