--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@5e3dfaea

CI:deploy: fix fngsystem deployment path

The match to detect fngsystem still contained 'guf'. Changed to 'seco'.

BCS 746-001043

The lq057q3dc12 runs with a jeida-24 data mapping.

BCS 746-000951

Added an optional node with panel-dt settings for the
lq057q3dc12.

BCS 746-000951

We use a custom Simple Panel driver that allows a more detailed description of
panels in the device tree. This is discouraged by the mainline kernel but has
shown useful for our specific applications.

For a more detailed explanation see the discussion https://patchwork.kernel.org/
project/dri-devel/patch/20190307101030.3822-1-maxime.ripard@bootlin.com/ on
the DRI mailing list and the original commit https://git.seco.com/seco-ne/
kernel/linux-guf/-/commit/bc113aef63149d38a61f618b40f91e22d53fe5cd on the i.MX6
kernel.

Note: Replaced all occurences of the old G&F name to SECO and also fixed
some smaller typos.

BCS 746-000951

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@f8d6de86

Make check error message more verbose

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@21e7edeb

Improve ci-test build

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@76bd81d8

Fix check job rules

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@a6721243

Move check job to parent pipeline

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@593784e1

Refactor check_if_integration_branch_is_up_to_date

- Use "source" and "target" terminology in the script, just like in the
  GitLab CI variables.
- Spefify data types for function arguments.
- Rename function arguments to better reflect their data types.
- Move determination of integration branch into main function, so that
  the successfull status message can be more verbose.
- Remove parent_merge_request feature and checking for gitlab-ci
  integrations, because the check job does not run on gitlab-ci
  integration branches at all.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@45b49ee3

Remove build stage on master

We don't need to explicitly trigger a build after merging changes in
gitlab-ci. A master pipeline on the target projects runs automatically.

Set the usb-c port to device mode as it should be used as peripheral
at first. Added an entry for ptn5150 cc logic IC and put it in UFP
mode.

BCS 746-000948

The include file enables the various HDMI related nodes of the
i.MX8MP.


(cherry picked from commit 43274696)

The AC97 part of the MC3 DTs is the current state of the
development/testing.

BCS 746-000955

Moved the lq057q3dc12 to a separate devicetree include and
added a new devicetree for the MC3 with lq057q3dc12.

BCS 746-000951

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@a0f7e84f

Rename pipeline stages to reflect YAML filenames

Activate the USB nodes and regulators for the VBUS voltages.The OC
detection function is not yet enabled. The USB hub is now enabled
by default. At the moment the connector of the backplane is untested.

BCS 746-000947

Due to an unknown graphics pipeline issue we cannot use the correct
display settings for the SCX1001255GGU06 display. It seems that the
clocking and synchronization of the whole graphics pipeline consisting
of the LCDIF, the MIPI DSI, the MIPI PHY and the MIPI to LVDS
transceiver cannot be applied by simply changing the clock-frequency of
the attached panel.

Though trail and error we were not able to modify the pipeline clocking
in such a way that it works with the correct frequency.

Thus use the frequency from the fg0700w0dsswagl1 instead.

The display has a reduced frame rate, but otherwise works as expected.

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

Temporarily disabled the SAI/AC97 to test the GPIO order.

BCS 746-000955

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

The LVDS receiver requires an input clock of at least 20 MHz.
Since the display supports VGA and QVGA, the higher clock (for VGA)
is used for now.

Note: Further work on the display timings is required.

BCS 746-000951
BCS 746-000944

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

Configured the panel to the correct display settings.
The lvds interface is internally connected to lcdif2.
The GPIO for the backlight enable was wrong.

BCS 746-000951

Also use the GPIO_PAD macros to cleanup the devicetree for can related
gpios.

The GPIO_PAD macros make the source code more readible but
have been available for trizeps8 boards only.
Moving the macros one step up, make them available for the other
boards too.

BCS 746-000945

Felix Gerking authored 2 years ago and Jonas Höppner committed 2 years ago

BCS 746-000943

Felix Gerking authored 2 years ago and Jonas Höppner committed 2 years ago

Now the correct phy reg is set explicitly. Use the driver specific bindings
for the phy reset and set the correct reset timings.

BCS 746-000932

BCS 746-000936

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

First draft of the integration of the LCDIF and the PWM backlight.
The changes haven't been tested yet.

BCS 746-000944
BCS 746-000951

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

BCS 746-000953

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

BCS 746-000952

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

The MC3 uses an AC97 codec for audio/touch input.
Added a first version of the AC97 codec to the devicetree.

Note: The communication couldn't be tested yet and might not work.

Tobias Kahlki authored 2 years ago and Jonas Höppner committed 2 years ago

The eMMC is connected to usdhc3, but we need it on mmc0.
Re-ordered the mmcs to reflect this change.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@d25cdebf

Really fix potential error message in retrigger job

TypeError: 'NoneType' object is not subscriptable

Example:
https://git.seco.com/seco-ne/yocto/manifest/-/jobs/129201

Add the device tree for the touch display set SCX1001255GGU06 including
the LCD panel FG1001L0DSSWMG01 and an SIS touch.
The panel is connected via LVDS.

The current state does not work as expected. It seems the MIPI signals
are not clocked out correctly.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@3eef782e

deploy_gitlab_ci: integrate only correct projects into the manifest

Previously we integrated all projects into the manifest, even those that
belonged to a different manifest branch. That led to all manifests
having the same project revisions, even if they should be different due
to different intgrated project branches.

Save a list of integration sources for each manifest branch separately
now, so that we can check which project belongs to which manifest branch
later.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@2b851ac6

Fix potential error message in retrigger job

TypeError: 'NoneType' object is not subscriptable
Example:
https://git.seco.com/seco-ne/yocto/infrastructure/ci-test/minimal-manifest/-/jobs/127610

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@c3fd0ecc

Check if variable is defined before checking it value

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@c1be0f4b

Fix jinja2 "or" operator

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@8da534ec

Run integration pipeline for target branch only

When we're integrating a single branch into a single manifest branch,
there is no need to run the intgeration pipeline for all configured
integrations, but only for those that match the target branch of the
current merge request.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@8c94b725

ci-test: add job to check for correct foo branch

This should catch an error where during a gitlab-ci merge process the
wrong foo branch is integrated into the manifest, which happened in
Yocto.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@163c7e1e

Add yml extension to jinja2 templates

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@59523aa0

check job: disable verbose output

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@28bf0065

retrigger: add job states to log message

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@fd742017

deploy_gitlab_ci: fix target branch in integration merge request

We need to pass the correct target branch to the merge request creation
instead of using the project's default branch.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@38379173

Run yamllint job on gitlab-ci integrations

Otherwise no pipeline is created at all in these merge requests, which
prevents the MR from being merged.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@dce4f8a3

changelog_generator: Fix machine project ID

We have deleted the original project due to git corruption. Replace with
the new project's ID.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@a7aa7c26

deploy_gitlab_ci: fix project argument passing

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@8a38f194

Put all manifest intgeration jobs into a single stage

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@a09564c1

deploy_gitlab_ci: explicitly pass gitlab-ci project and branch

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@172e7625

Improve error messages

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@f0f5a370

Improve naming of functions and variables

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@c8b484ca

Retrigger MRs from all integrating projects

Instead of getting the projects from the manifest file, search them via
their INTEGRATION variable.

Improve log messages on this occasion: print URL instead of just IDs.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@5e75bfc1

Deploy gitlab-ci to multiple branches

Add different integration jobs for different manifest branches. The
merge stage only has one job for each manifest, though, because
otherwise we cannot guarantee a consistent state for all branches.

Extend the deploy script for this purpose, so that it can deploy to
multiple manifest branches at once.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@771dcff1

Fix temporary dir path

We were passing the complete TemporaryDirectory object to the repo clone
function instead of just the path string, resulting in the repo being
cloned into a local dir "<TemporaryDirectory '/tmp/tmphwakypf8'>".

Fix this to actually use the generated temp dir.

This change makes it necessary to keep the TemporaryDirectory object
reference until we don't need the directory anymore, otherwise it will
be removed immediately.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@dce219c8

Print less important output in grey color

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@5e36261d

deploy_gitlab_ci: remove hard-coded integration targets

Get targets dynamically using the INTEGRATION variable instead.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@94cd6781

Recurse to child pipelines on retrigger

This makes it possible to retrigger jobs that are not part of the
given pipeline but in any nested child pipeline of it.

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@fb1219a7

Ignore __pycache__ in scripts directory

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@0c29c995

Refactoring: rename arguments for consistency

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@d461090b

Remove hard coded project and branch variables

MANIFEST_PROJECT, MASTER_BRANCH_MANIFEST, MASTER_BRANCH_PROJECT

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@a4aae414

Add rules for integration pipeline

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@bfa38b24

Improve log messages

--

Commit: seco-ne/yocto/infrastructure/gitlab-ci@1c9670d9

commit_and_push: do not pass branch

Branch can be implicitly determined via the repo.

BCS 746-000931

The pull-up of the CMD line of the eMMC isn't placed on the hardware.
Therefore, the internal pull-up of the pad is required, for the CMD line
to function properly.
Also increased the clock rate to 400 MHz (see NXP EVK) and assigned the
correct regulators (the eMMC uses fixed power rails for 1.8 V and 3.3 V).

Note: The eMMC isn't working yet.

Added nodes for the generic voltages and three nodes for the eMMC and
SD-Card (the SD-Card requires a 1V8 and 3V3 rail for the high-speed mode).

What works:
* fngsystem (kirkstone) ramfs boot via uuu
* for UART2 an alias to serial0 is set
* ethernet, at the moment no phy-reg is set explictily
* RTC, we need to check if specific backup source settings are required:
root@FLASH-N-GO:~# dmesg | grep rtc
[    1.519835] snvs_rtc 30370000.snvs:snvs-rtc-lp: registered as rtc1
[    1.960681] rtc-rv8803 0-0032: Voltage low, temperature compensation stopped.
[    1.967833] rtc-rv8803 0-0032: Voltage low, data loss detected.
[    1.977571] rtc-rv8803 0-0032: Voltage low, data is invalid.

BCS 746-000818