- Jul 29, 2022
-
-
Jonas Höppner authored
The build for the imx8mpguf hit the limit.
-
- Jul 28, 2022
-
-
The post-install steps for setting a custom devicetree have been changed for the i.MX8 devices. For the custom devicetree to work again, the name of the devicetree needs to be written to the devicetree file in the boot partition. BCS 746-000877 (cherry picked from commit 714adb35)
-
To add qgif to the Qt environment add the gif PACKAGECONFIG. (cherry picked from commit cd3c70f1)
-
The default prepare-kernel script didn't support the i.MX8. Tweaked the script, to support the custom requirements of the i.MX8. The i.MX6 specific parts for the generation of the overlays (from the shared XML) and the gf_platform detection has been moved to new scripts. BCS 746-000877 (cherry picked from commit 40d7ba85) fng-install:prepare-image: Added new scripts and clean-up The new scripts (prepare-overlays and prepare-xml2dto) for the post-install steps have beend added. Also wrong indentations have been replaced and the lists have been sorted. BCS 746-000414 BCS 746-000877 (cherry picked from commit 3edf9fc8) fng-install:prepare-kernel: Removed i.MX8 specific script The i.MX8 used a custom post-install script for the prepare-kernel step. Since the default script has been adopted for the i.MX8, the custom script is removed. BCS 746-000877 (cherry picked from commit 3a397d44) fng-install:rtc: Added RTC detection for the TANARO The TANARO will be produced with one of two RTC variants. We detect the RTC by querying the first register during the post-install and activate one of two devicetree overlays. BCS 746-000877 (cherry picked from commit 32b8f0c2) fng-install:post-install: Added comments to scripts Extended the documentation and comments in the post-install scripts. (cherry picked from commit c14d978b) fng-install:prepare-dt: Replaced loop with find Fixes SC2044 (cherry picked from commit 2cc3f539) fng-install:prepare-image: Convert recipe's syntax to kirkstone
-
Overlays that have been installed by Yocto are now moved into /boot/overlays. This is due to the inability of Yocto to easily handle devicetree overlays during the build. Overlays that are provided via the parameters of fng-install (--DTBO) are also placed inside /boot/overlays. Additionally, they are appended to the overlays variable inside the overlays.txt file. The overlays.txt file is parsed during the prepare-kernel post-install step. The loaded overlays are then applied to the default devicetree. BCS 746-000414 BCS 746-000877 (cherry picked from commit 5f97c8e6)
-
The lsb-release command can be used to query detailed information regarding the Yocto image/build (from the userspace). (cherry picked from commit 883e6834)
-
The lsb-release command is used by Lava to query system information. (cherry picked from commit f584eef6)
-
Jonas Höppner authored
We did remove the kernel modules for FNGSystem to reduce the image size. Compared to FNGSystem 27 the kernelmodules are adding about 2 MB, which seems to be OK for me. (cherry picked from commit 38c8f73c)
-
Jonas Höppner authored
(cherry picked from commit f9407345)
-
Jonas Höppner authored
The scripts failed if run in an installed fngsystem. This is needed in FT2, when new xml files are installed after the FNGSystem installation. (cherry picked from commit 920b5c91)
-
Replaced absolute card number by the ID of the sgtl5000 audio card. Added additional alsa configuration files for imx6ull as it does not have the sgtl5000 codec. BCS 746-000548 (cherry picked from commit b3c8ca1b)
-
Jonas Höppner authored
Also convert recipe's syntax to kirkstone (cherry picked from commit 98409477) (cherry picked from commit 325cd3c2)
-
Jonas Höppner authored
When switching from scheme 1 to 2, there was a bug leading to an unformated partition. Additionally I removed the partition message for FT2. This is only used during the yocto installation and was in the wrong position anyway. (cherry picked from commit c0871d1e)
-
Jonas Höppner authored
The implmentation was not really working when used in a yocto installation, as it would have needed a relocation of FNGSystem. Now the explicit change of the partition scheme is only be possible together with the installation of FNGSystem. (cherry picked from commit a4ecd599)
-
Jonas Höppner authored
When FNGSystem 15.1 was just installed and only one partition was setup on the device the installation failed. This is fixed here. Also: in A-B installation mode is the used target (A or B) reported after the installation. (cherry picked from commit e811145d)
-
The nfs-statd systemd service is disabled by default. BCS 746-000876 (cherry picked from commit 96288601)
-
For imx8m it is necessary to specify drm device 0 to run the qt-compositor with eglfs backend. BCS 746-000787 (cherry picked from commit 5058c033)
-
This allows us to draw the progress bar directly on top of the kernel splash logo. By default, psplash only displays the progress bar. BCS 746-000786 (cherry picked from commit ae1922ba)
-
Jonas Höppner authored
Normally we would use network-manager, but for the gplv3 free variants it is not available. In these cases we keep systemd-networkd as network tools available. Also adapt the call to gf-network-conf in the postinsts step during the first boot to automatically determine the used network tooling. This way the configuration for systemd-networkd is generated from the xml setting as needed. Until now the MACAdress was not configured if systemd-networkd was used. Now the MACAddress is added to the newwork config file like it is done when network-manager is used. (cherry picked from commit 91b1c14e)
-
The weston-start script for xwayland overrides the module loading of the notify library. Without notify, Weston doesn't report a successful start back so systemd. As a result, systemd stops Weston after hitting the timeout period. Without the module loading directive in the start script, xwayland needs to be activated in the weston.ini. (cherry picked from commit 510cc193)
-
If x11 and wayland are both activated in the DISTRO_FEATURES, we add xwayland and its xserver implementation to the image. (cherry picked from commit 0c01033f)
-
The tools recipes still use SSH as protocol for the checkout. Since we have opend our repository to the public, SSH isn't available anymore. Because of this, we switch to HTTPS as protocol. (cherry picked from commit e3c2325c)
-
imx6ull uses the linuxfb backend, imx6 and imx8m* using eglfs. (cherry picked from commit e382ac22)
-
(cherry picked from commit bb708bf0)
-
The qt-multi-screen-compositor is based on examples from QT to implement a compositor based on QtWaylandCOmpositor. The implementation is done mostly in qml. (cherry picked from commit 98725129)
-
Jonas Höppner authored
We have now platforms with custom rootfs names in the config. It is easier to install if the correct name is already in the install script. With this change, the image name for the install script can be adapted in a bbappend file. (cherry picked from commit 8fc62521)
-
Jonas Höppner authored
To find the touch the capabilities entries in the sysfs are checked. On Nallino the format contained a space, which made the script exit with error. This change now (hopefully) decodes it correctly. (cherry picked from commit 8ca830a4)
-
Jonas Höppner authored
(cherry picked from commit 548e591f)
-
Previously, iptables was only included via dependencies. (cherry picked from commit 0d782145)
-
Rename the distribution to SECO Northern Europe distribution and reference the manifest as the main source of information. BCS 746-000659 (cherry picked from commit 81bd01a9)
-
We borrow some recipes form public layers. However, most of the public layers are licensed under the MIT license. To make the licenses clear copy the corresponding license files into the relevant directories next to the recipe in question. Recipes in question: - rtl8812au - kbd - beep - mxt-app BCS 746-000658 (cherry picked from commit c402d855)
-
The "initial-draft" branch is deprecated. Use the branch "master" instead. Currently this should not change the content. (cherry picked from commit 32b7490b)
-
According to our company licensing policy for software the Yocto layers should be licensed under the Apache 2.0 license. Adapt the LICENSE-2.0.txt file from https://www.apache.org/licenses/LICENSE-2.0.txt as described in the APPENDIX. Add the license as LICENSE.txt file to the layer repository. BCS 746-000657 (cherry picked from commit d8fabd95)
-
Jonas Höppner authored
The script was relying in the /dev/input/touchscreen link which was not available in FNGSystem previous to version 16. This fix now loops though all available devices and takes the first device with absolute axis.
-
- Jul 12, 2022
-
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@075aeccb Refactoring: share build and test job basics for ci-test and Yocto This shares YAML code for the following procedures: - Build environment - Source code checkout via repo - SSH key and known hosts setup - LAVA test submission - Docker check if build folder is empty
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@6508e800 CI: Add hardcoded artifact's path generation also for fng system
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@c503a6d4 CI: pipeline yocto: Change scope of CI_PARAM_IMAGE, ... to fix fngsystem build The variables CI_PARAM_IMAGE, CI_PARAM_DISTRO, CI_MACHINE have been defined outside the job in the '.gitlab-ci.yml' file. This has a lower precedence then variables defined inside the job, but it seems that these variables are used in the scope of 'Trigger variables' when available in the trigger job. These override the value set in the template. As main results this made FNGSystem jobs build the normal yocto.
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@ea3aa816 CI: Readd allow_failure:true to fix not finished pipelines According to the docs a manual job always has allow_failure set to true, but we see a lot of jobs staying in an unfinished state since the removal of the allow_failure: true keyword for the sdk and test jobs. This should fix the broken behaviour.
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@dc0503b6 CI: Hardcode the artifacts path in the generated job description Some gitlab update(?) seem to have changed the behaviour in variable passing. In the scripts the variables are available but in the artifacts path does not resolve them any more. May be it has something to do with the introduction of the trigger:forward keyword.
-
-- Commit: seco-ne/yocto/infrastructure/gitlab-ci@f39e65ab Manifest: really include yamllint in the full pipeline Previous commit did not have the effect due to wrong extend order.
-
