Port the recipe from the dunfell poky branch. This version contains the latest
vulnerability fixes, including CVE-2022-0778.
https://onedigi.atlassian.net/browse/DEL-7868
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
(cherry picked from commit be046db4faae911b2a858d748551c6c91fc54043)
The sign mode needed for each platform is invariable, and since the platform
is already a mandatory parameter for the script, we can store this information
implicitly. Reflect this change in every recipe where the script is used, but
keep the variable at the Yocto level since it's still needed in several places.
https://onedigi.atlassian.net/browse/DEL-7862
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Signed image support in U-Boot has been split into two separate configurations:
one that adds artifact authentication support and another that signs the U-Boot
binary at the end of the build. Reflect this change in meta-digi.
https://onedigi.atlassian.net/browse/DEL-7862
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
In meta-openembedded commit 35d249bb9a575a5a48491052896e121266d515f0,
libconfig, one of swupdate's dependencies, was split into two packages: one for
the C library and another for the C++ one. swupdate depends only on the C
library, but the C++ library was implicitly pulling in other dependencies, such
as libgcc.
After the commit mentioned above, libgcc stopped being pulled into our recovery
image, causing the following error to appear when swupdate terminates:
libgcc_s.so.1 must be installed for pthread_cancel to work
Aborting
Even if the action carried out by swupdate is successful, our recovery script
detects this as an error and aborts the update operation. Avoid this by
explicitly adding libgcc as a runtime dependency for swupdate.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This commit adds power safe and reboot safe script for the MCA
and substitute default reboot and poweroff busybox commands.
Signed-off-by: Mike Engel <Mike.Engel@digi.com>
https://onedigi.atlassian.net/browse/DEL-7828
The license checksum and the source branch should be on the properly
versioned bitbake recipes. Otherwise, when we include this common file
in a different version linux recipe (for example, the upcoming 5.10) it
will inherit incorrect values.
While on it, fix the incorrect license checksum values for the linux 5.4
recipes.
Signed-off-by: Javier Viguera <javier.viguera@digi.com>
We originally included this recipe to fix some vulnerabilities in the dnsmasq
version used in zeus (v2.80), but the recipe has been updated since then. The
current version used by default in hardknott is v2.85.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Now that the "--platform" parameter is optional, we don't need to hardcode the
platform anymore, so remove the first patch.
In this version, the FDO platform has been renamed to WL, and since the default
PACKAGECONFIG is meant to work with older cog versions, we need to manually
append the "wl" PACKAGECONFIG to build the wayland platform.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This is a partial revert of commit 3819ee6672.
NXP simplified the profile script when updating to hardknott-5.10.72-2.2.0, but
since the /run/user/0 directory isn't automatically created in sysvinit
environments, the weston desktop fails to start in these cases. The previous
version of the script automatically creates said directory if it's missing, so
use that version instead.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This commit updates the US Board Data File (BDF) to include the changes
required to pass the EN 300 328 V2.2.2 blocking test.
The new BDF files are:
- ccimx6/bdwlan30_US.bin (eb3d3cdafd30b30e837974177a281648)
- ccimx8m/bdwlan30_US.bin (72531e6e0870673e4ba2b02bd551925b)
- ccimx8x/bdwlan30_US.bin (1cf667271da33f6eb207360eeb3d4d63)
https://onedigi.atlassian.net/browse/DEL-7773
Signed-off-by: Arturo Buzarra <arturo.buzarra@digi.com>
This makes it so bitbake obtains the SDMA firmware from the firmware-imx recipe
instead.
This also avoids including all of the linux-firmware packages in i.MX6 SDK
builds, preventing errors when building the ccimx6sbc SDK (linux-firmware
provides the community versions of the ath6k and ar3k firmwares, which clash
with our versions).
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
We already have a copy of NXP's weston.ini, so there's no need to manuallly
append to it in our do_install_append(). Simply add our changes for our custom
desktop background directly in the source file.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
In older Yocto versions, weston was launched via a udev rule, but now it gets
launched as part of the "graphical" system target. Since this target isn't
reached automatically on boot, the weston desktop will only appear if you
manually launch said target (or the weston service itself).
To make the desktop launch automatically, change the target to "multi-user",
which is the default target laucnhed on boot.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This makes it possible to distinguish between two different packages:
* NXP's tensorflow-lite, which contains modifications for i.MX platforms
* Our tensorflow-lite, which is meant to be used with Google Coral
Since tensorflow-lite-coral provides a subset of the files provided by
tensorflow-lite (specifically, the python libraries), make the packages
conflict during runtime.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This package is used by the eIQ recipes, so make sure we have NXP's .bbappend
so it works as expected
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Update a patch so it applies cleanly and remove another one, since it has
already been applied upstream
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This package was originally being added via RDEPENDS, and its removal was
missing when porting the newer file from NXP's meta-imx. Re-incorporate the
removal to avoid including the package in our images, but do so by adding it
to our images' BAD_RECOMMENDATIONS.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Even though our busybox recipe (v1.34.1) is backported from poky's honister
branch, it has many elements in common with the recipe provided from poky's
hardknott branch (v1.33.1). To simplify our backport, re-use poky's busybox
files when possible, replacing them only when strictly necessary.
While at it, remove the busybox-inittab recipe, since there are barely any
differences between the honister and hardknott versions.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This is a small oversight on NXP's part, because this package is pulled in as a
dependency of imx-vpuwrap, but it's only compatible with their EVK board,
making the build fail on any other i.MX8MP board.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This PACKAGECONFIG existed in gatesgarth, but it's included by default in
hardknott and it pulls in libgbm as a dependency. This package isn't available
for i.MX6 platforms, causing the build to fail immediately. Remove this from
the ccimx6 PACKAGECONFIG and leave it how it was on gatesgarth.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
For packages such as weston or opencv, bitbake will sometimes choose the poky
recipes, causing build errors. Explicitly set the NXP version of these packages
as the preferred version by mimicking the layer.conf configuration in the
meta-bsp layer of meta-imx.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
It seems like the version of mkfs.vfat used in Yocto 3.3 has a label name
length limit of 11 characters. All of the labels for our current platforms
surpassed this limit ("Boot ccimxX" alone already occupies 11 characters), so
replace the machine name with "DEY"
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
These variables were originally set in NXP's recipes, but they moved their
definitions to each platform's configuration file. Set the appropriate values
where needed. While at it, replace every mention of the SOC_TARGET variable
with its new name, IMX_BOOT_SOC_TARGET.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
We now use the meta-freescale recipe as a base for our .bbappend, so remove
the filter.
This reverts commit 9e68d61f7c465800d62913c044e43c541f2eacd7.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
The recipe we ported from meta-imx has the same version number (v2.4) and
bitbake chooses the meta-freescale version because said layer comes before
meta-digi-arm in the BBLAYERS. Since the meta-freescale version uses different
platform overrides, the build fails, so filter it out to use our version
instead.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Aside from the file paths changing, there are a few cosmetic differences
between v1.22.14 and v1.30.4 that prevented the patches from applying, so
re-work them.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
The common license file GPL-2.0 is now called GPL-2.0-only in poky, so we need
to reflect this name change to avoid errors
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
This is a generic package that is currently being used by NXP to load firmware
on the target, so include it in our images as well. Since it only works with
systemd, leave it out of builds that use sysvinit.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
poky now provides glibc v2.33 by default, and there aren't any alternative
versions available, so there's no need for this.
This reverts commit e7921dfd93.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
Remove the v3.0.8 version of the recipe to use the v3.0.9 recipe in poky. Also,
adapt our .bbappend since some of its changes are outdated.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>
NXP now uses the meta-freescale version of the imx-boot recipe as a base, and
applies a .bbappend in meta-imx to apply more customizations. Reflect this
in meta-digi-arm and merge said .bbappend with ours. While at it, omit one of
NXP's operations in their do_deploy, since it breaks our build and doesn't look
essential.
Signed-off-by: Gabriel Valcazar <gabriel.valcazar@digi.com>