Tracking the latest master of all dependency layers, gcc was recently
updated to gcc8% from gcc7% and linux-stable-4.16.y fails to build.
linux-stable-4.17.y succeeds with gcc8%, so update to it and make it
the default.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Provide a linux-stable recipe for 4.16.2
git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git
Remove 4.12 recipe from
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
It doesn't build, and hasn't been able to build in a long time. The point of
petitboot was to enable booting an image from uSD. Initially the rockchip
bootloaders did not support booting from uSD. However, more recent rockchip
bootloaders do support booting an image from uSD. If your board doesn't
support booting from uSD, try updating your device's in-flash bootloader.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
There is no MACHINE that is using it.
This recipe should have really been called "linux-radxa" since that is from
where it is fetching the sources, i.e. github.com/radxa and not
github.com/rockchip-linux.
The conf/machine/include/rk3{066|188}.inc files were still using the old
kernel recipe name.
Now with RSS, the "prebuilt-gcc-linaro" no longer works. Although I could get
this recipe to work again, trying to build the kernel fails with "as:
unrecognized option -EL".
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Linux v4.11 has been released, bump the version to 4.11.
Signed-off-by: Romain Perier <romain.perier@collabora.com>
Tested-by: Trevor Woerner <twoerner@gmail.com>
Linux kernel 4.11 being released soon, bump recipe to 4.11-rc8.
Signed-off-by: Romain Perier <romain.perier@collabora.com>
Tested-by: Trevor Woerner <twoerner@gmail.com>
Add mainline kernel support. Now the version is 4.10-rc8.
Signed-off-by: Eddie Cai <eddie.cai.linux@gmail.com>
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
Reviewed-by: Romain Perier <romain.perier@collabora.com>
New URL, fixes build failure. It should work with krogoth and morty.
Currently only the radxa rock is supported
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Romain Perier <romain.perier@gmail.com>
These recipes are outdated, no longer build and use the wrong SRC_URI.
That's preferable to drop these recipes and rewrite the stuff from
scratch.
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Romain Perier <romain.perier@gmail.com>
This BSP meta layer should not be used for kernel development. For this
the developer can simply compile his kernel outside of yocto. We prefer
provide recipes for upstream and vendors kernel.
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Romain Perier <romain.perier@gmail.com>
The version needs updating in order to build successfully.
Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Reviewed-by: Romain Perier <romain.perier@gmail.com>
This bbclass is already required by linux-yocto.inc, so remove it as it
is redundant.
Signed-off-by: Romain Perier <romain.perier@free-electrons.com>
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
This hack was required to ask do_kernel_configme to don't pass
option --noallconfig to configme (which had bad effects for
the current configuration). As the task configme is disabled,
this assignment is no longer required.
Signed-off-by: Romain Perier <romain.perier@free-electrons.com>
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
We define the KERNEL_CONFIGCOMMAND variable in order to launch
'make <defconfig' during the configure step of the recipe. As the old
'file://defconfig' is no longer required, we remove it. The task
do_kernel_configme is incompatible with the notion of in-tree defconfig,
so we disable it.
Signed-off-by: Romain Perier <romain.perier@free-electrons.com>
Reviewed-by: Trevor Woerner <twoerner@gmail.com>
It includes firmware and system files for Broadcom bcm4354 required
to work with the linux kernel. These are the files from chromeos.
Signed-off-by: Romain Perier <romain.perier@gmail.com>
By default the git fetcher try to find a tag attached to a branch (master for us).
This is not true with linux-next as only the current tag is attached to master,
other ones are still actives and maintained but detached. Adding argument
"nobranch=1" to SRC_URI fixes this issue. Then, we need to define the commit id
in SRCREV. In this way, we can checkout any active tag from next (as there are
present during 2 months, this is useful).
