Age | Commit message (Collapse) | Author |
|
We have a bunch of yaml files that we are carrying around
whose display-name was set to 'DELETE ME' to remove them
from Jenkins, but the yaml file was never removed.
Since the jobs are no longer in Jenkins, let's go ahead and
remove them now.
Signed-off-by: Kelley Spoon <kelley.spoon@linaro.org>
Change-Id: Ic0bcbc1c8d14064309e38b0224c3eb207cd25e2b
|
|
Now that lkft jobs have been migrated to gitlab, we need
to clean up the old Jenkins jobs left behind. This change
deletes any job that is using a node with a label that
build-lkft-01 serves (ie, docker-lkft, or docker-buster-lkft)
LKFT AOSP jobs remain untouched.
Signed-off-by: Kelley Spoon <kelley.spoon@linaro.org>
Change-Id: Ic3a3d7a31a7b884467a114b6631ae7d1aa6113b5
Signed-off-by: Kelley Spoon <kelley.spoon@linaro.org>
|
|
There appears to be a problem with the MultiConfig plugin
where it doesn't recognize "built-in" as label. This is
causing a large number of jobs to block while waiting for
an executor to become available, so let's revert the node
requirement back to 'master' while we look for a long
term solution.
Signed-off-by: Kelley Spoon <kelley.spoon@linaro.org>
Change-Id: Icf8cb3490cc04fdb3f33231d3c2248a3e290e72b
|
|
This change updates all configs that are intended to
run on the former 'master' node to use the 'built-in'
label exclusively as per the new Jenkins naming conventions.
Signed-off-by: Kelley Spoon <kelley.spoon@linaro.org>
Change-Id: Iece59a578329bd4cd038455ec4966a1e874a75f3
|
|
Starting in 2.307.* of Jenkins, the naming scheme for nodes
is migrating to 'built-in' to replace master. This change
makes a mass commit to add the 'built-in' as a potential
label for jobs that are currently tied to the 'master' node.
Change-Id: I51bdae6282de863d1bc1ddaaab7964ef288ab480
|
|
The previous label, docker-lkft, pointed to Debian 9.13
(Stretch), which included Python 3.5, insufficiently new for
`repo`, which requires Python 3.6.
Change-Id: I1af17037f20d72cbe59c07924e01fe5009434e40
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Change-Id: I0dd8a36a0503aad3d006d2d60d4ea42dcb63a4a8
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
stable-rc LAVA job priority updated according to the review cycles.
The most often builds are higher kernel version and which need to
run fast and fast turn around time for reporting.
Signed-off-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Change-Id: Icad8018f19496cc5141f71f46f4fed7a22ac31c0
|
|
Kselftest update to version 5.3 which need new way to parse kselftest results
by using test definitions master branch parser.
Setting up TDEFINITIONS_REVISION=master
Signed-off-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Change-Id: Ieef2a0d01c54a007b7ff825ed4ea30b0d750f24c
|
|
This reverts commit d9ffb8a596d11f4310f32812133becf4f8f3af9b.
Change-Id: I3101b98f831eae7bdaaf019be8e87210b5022181
|
|
Change-Id: I564a10e813899dc52a7263a249b4506100bc7970
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
This is already running successfully on mainline and next. Deploy to
remaining active linux lkft branches.
Change-Id: Iafe66a822cc3a0152b6d70bb81e097aad765c704
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
The original implementation of build reporting in a3cdfdc98ed4 ("LKFT:
Send build status as a test result to squad") failed to realize that
DEVICE_TYPE gets poisoned by qemu-testjobs.sh, which resulted in all
build results being reported incorrectly as QEMU build results.
To fix this, change how qemu jobs are generated. Instead of calling
qemu-testjobs.sh to change DEVICE_TYPE and then recall
submit_for_testing.sh, handle the qemu jobs directly in
submit_for_testing.sh.
This simplifies the jenkins job files by removing three steps, at the
cost of making submit_for_testing.sh slightly more complicated.
test_submit_for_testing.sh has been modified to use an explicit list of
devices to better mimic what jenkins is actually doing. It is a bit
unfortunate to have to maintain the list there, but there is no place
the list exists except in each jenkins job definition file.
The template changes have been tested with test_submit_for_testing.sh
and verified not to regress functionality. However, the author lacks an
environment to test all of these changes together comprehensively.
Change-Id: I5c9d76250b6ffd0c35de0d70a759bdbfeae2decf
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
Change-Id: I141836019c143c1b189de6c5bcd3a858330891ea
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
This puts all parameters being used into all of the LKFT main
jobs:
* Stable RC's
* Maineline
* Next
Change-Id: I975840ad047b11a7fa6717f33f30437d5a6ec7a6
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
The TEST_SUITES parameter controls what test suites to run
through qa-reports. It can generally be one of these:
* kselftests
* libhugetlbfs
* ltp
or, alternatively:
* all
* sanity
* none
or even blank, which will be translated as "all", or just the
name of the specific test suite, like:
* ltp-syscalls
The "sanity" test suite is implied with "all". Specifying just
"sanity" will run those test plans and nothing else.
Change-Id: I4e0a34dce32622fa995ef57d74b40bd376cd6cd3
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Current priority: 75 for all stable branches, 80 for stable/sanity
branches. Instead of waiting for all results let one build finish first.
Following changes helps in getting results fast for each branch at once.
- 4.4: 78
- 4.9: 77
- 4.14: 76
- 4.19: 75
- 4.20: 70
Change-Id: Id50f1ce729a26c267728dd7835ffe3551a1d7dc8
Signed-off-by: Naresh Kamboju <naresh.kamboju@linaro.org>
|
|
This reverts commit 538fbd94cacb8c8d7c22bd54e382abe7300a3b2b.
While the partitioning of the builders.sh script is fine, a
few things need to exist in the environment for each phase to
function properly. This is a good opportunity to document
those things, so that will happen in an upcoming commit.
Change-Id: I3590b0fdf9ed1340cb0dbe2cdd105a5e2192642e
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Rearrange job definition in order to simplify things:
* add script to install dependencies,
* separate build and publishing preparation
Change-Id: Ic0bfb8ca9afb52000cb77b7b93eca8d33ab20d80
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
This reverts commit 55b4dad531d920d83910790a979a860f881154a8.
Change-Id: Ib0330fa4087c1c4f329a92d06570e4a1225a9411
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
The publishing destination used to contain the (correct)
manifest branch, which is "rocko", but that was recently
changed to "lkft/rocko", which buried the artifacts one level
deeper in the snapshots directory structure.
This change removes the "lkft/" part from the $MANIFEST_BRANCH
variable of the publishing destination, so that the artifacts
are published where they used to.
Change-Id: I6668a30df7a98f857cb65e4f0b2e81b9a3401eea
Reported-by: Anders Roxell <anders.roxell@linaro.org>
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
The LKFT distro setup has worked nicely on next and mainline.
We can now switch to it entirely.
Change-Id: I3f9050682788af80a8c3c394cb9a5f3517261f9f
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Change-Id: I0c090733162b105d619afa1604b14c99c493dfff
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Setting the manifest URL can help in debugging using different
versions or testing new packages.
Change-Id: I9273301924caa26dbcd87c7d65048046cc343bc2
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Change-Id: Ib58bd54efe50960a43c36f5021d0c43ecb94c7ee
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
Change-Id: I254a05d7ce57ac255b0ed19ae4d9dacef4729101
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
Add the machine corei7 for lkft's x86 builds.
Change-Id: I63188f57fcf4d797241e513c981485a419ce65a1
Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
|
|
Per KV-120, this implements a second test project per branch which performs a
minimal test run against a subset of boards.
Change-Id: I95778715653a24056beda49e530f719e6105bcdc
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
Rocko has now been tested on linux-next for some time, and the
outcome has been positive as far as we can tell. We have
garnered enough confidence in updating the baseline to Rocko,
so this does that in all jobs that are currently under the
radar for KVT:
* openembedded-lkft-linux-mainline
* openembedded-lkft-linaro-hikey-stable-rc-4.4
* openembedded-lkft-linux-stable-rc-4.4
* openembedded-lkft-linux-stable-rc-4.9
* openembedded-lkft-linux-stable-rc-4.14
* openembedded-lkft-linux-stable-rc-4.15
* openembedded-lkft-linux-stable-rc-4.16
* openembedded-lkft-linux-stable-rc-4.17
Change-Id: I395618d547e208d9417dd83fb106d67b3956253b
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
This change should've gone with fa0c45f4, as this image,
rpb-console-image-lkft, is the one that will be used for
testing.
Change-Id: I0885437b6e345ae9e6f40b070b94d7216774fe74
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
Change-Id: I4f793c83967834261cb498fd4a8e0c930ad93d95
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
The code is generic and can be used by any builds as long as PUB_DEST job
parameter is used. Move it under rpb-openembedded folder since it isn't LT
QCOM specific either.
Change-Id: Id8f02487beca7d28d0699f2ce0fadf9788db1dfa
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
As of lava 2018.04, lava supports job priority 0-100, where 0 is 'low',
50 is 'medium', and 100 is 'high'. Change LKFT jobs that are 'low' to
25, and jobs that are 'high' to 75. This allows more flexibility in
scheduling.
The staging jobs were exporting LAVA_JOB_PRIORITY twice.
Change-Id: I720718b23673ac52621ad13282062c870b290d2c
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
the step is modified with the inclusion of set -ex, in order to help debug
future failures (as observed in https://projects.linaro.org/browse/CTT-1107).
Change-Id: I2cc785cb79e8a7a0eb1fe3d4c79278a2de739c3c
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Enable x86-64 qemu jobs, which boot reliably. Keep arm64 (hikey) qemu
disabled until https://projects.linaro.org/browse/CTT-997 is resolved.
Change-Id: Ia33cd550c4d63baaebd69c074ac596800854b118
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
Change-Id: I074f4d89c78dff266706491d337b92e0b424ab97
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
Expose the build matrix parameters as a build parameter. Very handy!
Tested @ https://ci.linaro.org/job/lkft-staging-no-skiplist/
Change-Id: I5971384e01090fccc1fdedd6a4145d8b89e0b059
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
QEMU is not yet working/stable. Leave it enabled for branch 4.14, but
disable it for the others. This commit may be reverted once 4.14 is
stable and baselined in production.
Change-Id: I9b9723c6fee929928e249226630e2929fb7f6904
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
Change-Id: I7c4c0eca3a393544da8b6f20fe19a0c0d8db19e7
Signed-off-by: Milosz Wasilewski <milosz.wasilewski@linaro.org>
|
|
There's an SLA for getting stable kernel results.
Move them on a dedicated slave to ensure we deliver the builds in a
reasonable amount of time.
Change-Id: I2b658d89e026092f1dc2bebd4ac027083e2e7f8d
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
This continues what 7543e1e started.
Applied to:
* linux-4.4
* linux-4.4-rc
* linux-4.9
* linux-4.9-rc
* linux-4.13
* linux-4.13-rc
Change-Id: I0118a2f6e99bcf344ee099eb571b2e58961ea3d5
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org>
|
|
The publishing destination is now exposed via a job string parameter.
It's the best practice to use the environment variable PUB_DEST to store
its value.
Rename SNAPSHOTS_URL to BASE_URL, which also makes it more consistent with
OE RPB jobs.
Adjust the relevant jobs, scripts and templates to match the new behavior.
Get rid of the switch/case in the common builders script, since it is a
per build configuration and not a shared common code.
Change-Id: If8aa200aab44c71328927f1d131d40089c18197b
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
Email notifications are too noisy with successful builds included.
Change-Id: I9d5fef5eb64297de73a9a83b74bb30c617aef764
Signed-off-by: Dan Rue <dan.rue@linaro.org>
|
|
It isn't a token. Use email-ext groovy support to achieve the same
intended string manipulation.
Change-Id: I58399132256f0d539a78d66ce94c338730c5360c
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
All emails are now delivered via lkft-maintainers and lkft-triage
mailing lists.
Change-Id: Id29e5aa0097ca1c526521cb53db7fb32735069e2
Signed-off-by: Milosz Wasilewski <milosz.wasilewski@linaro.org>
|
|
Send notification about failed builds only.
Use a shorter email subject.
Change-Id: Iccfc0c838d6bcb32da741905f333c5c87344af44
Signed-off-by: Vishal Bhoj <vishal.bhoj@linaro.org>
|
|
kselftest skipfiles created as per ${DEVICE_TYPE} and ${KERNEL_VERSION}
Remove build parameter "KSELFTEST_SKIPLIST"
Ref:
LKFT: move kselftest skiplists to git
https://projects.linaro.org/browse/QA-1759
Change-Id: Ie13cee2140ae3d0a815e66260582af85673cbc83
Signed-off-by: Naresh Kamboju <naresh.kamboju@linaro.org>
|
|
Add missing comma.
Change-Id: I2a28fc8e2783b98688c0811f93d0d31944204346
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|
|
On each completed build the email is sent to lkft-triage mailing list.
Change-Id: If5376a0c8320a5af33d410db7fe6ddca07474f65
Signed-off-by: Milosz Wasilewski <milosz.wasilewski@linaro.org>
|
|
LKFT started on HiKey but the kernel recipes are now generic.
They aren't HiKey specific and the name is misleading.
Rename linux-hikey-* to linux-generic-*
Change-Id: I35d0089f463f95df3c88a0cb012b0bbd2a976c25
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
|