| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
| |
Not sure why this only started cropping up now.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
| |
repo cannot do without right now
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
| |
Was missed during review
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
| |
Signed-off-by: Kostiantyn Bushko <kbushko@intellias.com>
|
|
|
|
| |
Signed-off-by: Kostiantyn Bushko <kbushko@intellias.com>
|
|
|
|
| |
Signed-off-by: Kostiantyn Bushko <kbushko@intellias.com>
|
|
|
|
| |
Signed-off-by: Kostiantyn Bushko <kbushko@intellias.com>
|
|
|
|
| |
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
|
| |
Also should work for other bootloaders. This is intended to help with
keeping around older versions of images with a similarly old bootloader.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
| |
Signed-off-by: Mykhaylo Sul <ext-mykhaylo.sul@here.com>
|
|
|
|
| |
Signed-off-by: Mykhaylo Sul <ext-mykhaylo.sul@here.com>
|
|
|
|
|
|
|
|
|
|
| |
Originally I was just trying to remove the reference to DEPLOY_DIR_IMAGE
because it depends on MACHINE, but then I realized that this recipe is
just plain not worth keeping. It isn't viable for anything except the
most limited testing, since it bakes device-specific credentials into an
image, but we weren't even using it for testing.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
| |
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
| |
No use printing all that if there's a problem anyway.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
|
|
|
|
| |
Anything else will cause problems because of the hackery with keeping
backup copies of the image and the u-boot rom.
Also warn about supplying a different image name after the overlay has
been created. That doesn't work and it will be ignored.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Since bitbake can remove old images that an overlay was non-obviously
dependent on, the safest thing to do is make a copy of the image and
keep it alongside the overlay. When using the overlay later,
automatically use that image. Also do the same thing with the U-Boot
rom. This should also make moving the overlay file to another machine
much easier.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
|
|
| |
By default qemu monitor gets attached to stdio,
and that interferes with the test output.
We also don't need vga options in nographic mode.
Signed-off-by: Eugene Smirnov <evgenii.smirnov@here.com>
|
|
|
|
| |
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
| |
Hacky but works
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
|
| |
So that `before_script` can be used for additional steps when
sub-classing
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
| |
Was killed for good
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
|
|
| |
(use docker 18)
See https://gitlab.com/gitlab-org/gitlab-runner/issues/4499
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
| |
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
| |
Was causing problem with cache when updater-repo was updated
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
| |
It didn't work if the manifest already had a revision specified
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
| |
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Squashed:
* Oe-selftest GitLab pipeline stage
* Use credentials for CI's oe-selftest
* Setup kvm trick for docker on CI
The gid of the kvm group needs to match the one from the host
* Run ci scripts from Docker images
* Template out jobs and split .gitlab-ci.yml
To be easily reused in other branches and projects
Rely on this gitlab feature: https://docs.gitlab.com/ee/ci/yaml/#extends
* More flexible checkout script for CI
- can work without $CURRENT_PROJECT (checks out everything)
- can take a list of pinned versions
* Add optional CI jobs
ptest and other oe-selftests
* Publish bitbaked images as artifacts
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Including these points (initially as separate commits):
* Try caching updater-repo in gitlab
* Only patch aktualizr recipe in CI if needed
* Run under user 4321 in bitbake docker image
* Add ability to supply METADIR to envsetup.sh
So that we can bypass the "relative path from the script" trick that
doesn't work with symlinks.
* Checkout the PR base branch
* do not run on pushes
* run on runners with "gitlab" tag
* Print all revisions after repo checkout on CI
* Pull CI image from the base branch instead of master
* More robust checkout in CI
* Only store updater-repo artifacts for one day
* Use "stretch" instead of "stable" for Debian release in Dockerfile
(in fact, stable just changed to buster)
* Move UPDATER_REPO variable out of .gitlab-ci.yml
To be defined in the project's settings
* Make docker CI stage buildable on shared runners
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
| |
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|\
| |
| | |
Log persistence
|
| |
| |
| |
| | |
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|/
|
|
|
|
|
|
|
|
|
|
| |
"Autoprovisioning" or "automatic provisioning" is now known as "shared
credential provisioning". "Implicit provisioning" is now known as
"device credential provisioning". "HSM provisioning" was always a
misnomer, so it is now refered to as "device credential provisioning
with an HSM".
This includes a bump of the aktualizr version as well.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
| |
Signed-off-by: Mike Sul <ext-mykhaylo.sul@here.com>
|
|
|
|
|
|
| |
Expose a fake hardware rng on the guest, linked to the host's /dev/urandom.
Signed-off-by: Laurent Bonnans <laurent.bonnans@here.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
do_image_wic task is a standalone task that depending on
do_image_ota_ext4, so it's possible that do_image_wic runs
(taskhash contaminated) but do_image_ota_ext4 does not (taskhash not
contaminated), in which case, the otaimage would be in DEPLOY_DIR_IMAGE
instead of in IMGDEPLOYDIR, so we need check both of them.
Also, the logger.error is not supposed to raise a error, it just
prints out a error message, so we should use WicError.
And another typo is it checks IMGDEPLOYDIR but reports DEPLOY_DIR_IMAGE
not exist, also fix that.
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
|
|
|
|
| |
Signed-off-by: Manikandan Kandasamy <manikandan.kandasamy@here.com>
|
|
|
|
|
|
| |
Also remove all references to aktualizr-example-interface. It has been removed.
Signed-off-by: Patrick Vacek <patrickvacek@gmail.com>
|
|
|
|
|
|
|
|
|
|
|
| |
Do not write directly to the IMAGE_DEPLOY_DIR but to the new
deploy directory IMGDEPLOYDIR.
Also change the image ending to match the image type name (from
otaimg to ota-ext4). This allows to use the automatic symlinking
and output artefact management too.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com>
|
|
|
|
|
| |
In a separate Jenkinsfile right now, with a lot of duplication. Let's
hope to find a cleaner way
|
| |
|
| |
|
|
|
|
| |
For qemu tests
|
| |
|
| |
|
|\
| |
| | |
Feat/better dependencies
|
| | |
|
| | |
|
| |
| |
| |
| |
| | |
More generic, hopefully a bit clearer. The input must be a recipe, but
recipes depend on packages.
|
| | |
|
| | |
|