Commit 160e8737 ("lava: amd64-upsquared6000: Use v2022 1st stage image
with firmware pkgs") set the UP Squared 6000 1st stage image to a daily
build as it included the required firmware pkgs. As the daily image will
get deleted soon, let's use a weekly one until the next v2022 release it
out.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Edmund Smith authored 2 years ago and Sjoerd Simons committed 2 years ago

These changes are made to demonstrate to client repository owners how
to use the new tool for best effect. The tests from this repository
are not now run as standard, and the base configuration cannot be run
because the Apertis build ids are unset. Leaving the tests in an
obsolete state seems actively harmful; the alternative is simply to
remove the testing infrastructure entirely, but that makes it more
difficult to find a fully worked example of how all the tools this
repo provides fit together.

Up until now, the approach has been to run two commands back-to-back:
first generate the tests (as YAML), then submit them to Lava with
lava-submit.py. Using the new generate-test-pipeline.py tool, we
generate a pipeline which will run all the generated test files. We
can then define a new trigger job for every existing submit job to
execute the generated pipeline.

Edmund Smith authored 2 years ago and Sjoerd Simons committed 2 years ago

generate-test-pipeline.py can replace lava-submit.py. The latter will
submit a batch of test files to Lava, and return the ids for further
monitoring as required. This model, however, prevents the Lava jobs
themselves from making use of short-lived tokens like CI_JOB_TOKEN,
because the job tokens available at the point of submission will
expire with the completion of the submit job.

This new tool produces a Gitlab YAML pipeline file, suitable for use
in a subsequent trigger job. Each test file will have its own Gitlab
job in the pipeline, and that Gitlab job will run using the Lava
runner, which means live logs will be available, and all tokens will
remain valid for the duration of the job.

Note that we are still generating this pipeline in a job that has
already terminated before the Lava jobs begin, so we cannot use our
own CI_JOB_TOKEN in the test files. However, we can use the new
templating feature of the lava runner to request that the individual
test job's own CI_JOB_TOKEN be substituted into the Lava job before
it is submitted by the Lava runner.

If you are generating jobs in the standard way, using generate-jobs.py,
you need to escape the templating you want the Lava runner to perform,
so URLs might look like:

   https://gitlab-ci-token:{{ '{{job.CI_JOB_TOKEN}}' }}@my.gitlab.example/my/repo.git

After template substitution by generate-jobs.py, the Lava runner will
now see a URL of the form

   https://gitlab-ci-token:{{job.CI_JOB_TOKEN}}@my.gitlab.example/my/repo.git

and will substitute its own CI_JOB_TOKEN before submitting the
job. This will work for any of a job's associated CI variables, as
needed.

The latest Apertis NFS images now include firmware pkgs required to
enable audio and the Intel AX210 wifi module on the UP Squared 6000
board.

Without these firmware pkgs, drivers will retry several times. Thus,
lots of kernel msgs are printed to the serial console, which makes LAVA
test fail to parse the login prompt.

Since we are currently in the release period and this is adding a lot
noise to the results, let's use the latest daily image. But note that
this should be updated quickly to a weekly release, before it becomes
unavailable.

Fixes: infrastructure/apertis-issues#171



Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Many comments in the test cases mention the Minnowboard as the target
reference board. This is now retired and replaced by the UP Squared 6000
board, so replace any occurrence.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The lava test templates have some comments on the timeouts referring to
the Minnowboard, now retired. Let's remove this old comments, but keep
the used timeouts, as that's the current setup working on LAVA.

A follow up task should revisit these timeouts and set them more tightly
if needed.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The Apertis v2022 image used to flash the testing images is reporting an
issue related to the wifi driver:

See https://lava.collabora.dev/scheduler/job/7001988#L1209
```
[   27.837322] iwlwifi 0000:02:00.0: minimum version required: iwlwifi-ty-a0-gf-a0-39
[   27.837324] iwlwifi 0000:02:00.0: maximum version supported: iwlwifi-ty-a0-gf-a0-66
[   27.837325] iwlwifi 0000:02:00.0: check git://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git


[   27.847366] stack segment: 0000 [#1] SMP NOPTI
[   27.952703] CPU: 3 PID: 85 Comm: kworker/3:1 Tainted: G            E     5.15.0-trunk-amd64 #1  Debian 5.15.22-0~apertis1
[   27.964951] Hardware name: AAEON UPN-EHL01/UPN-EHL01, BIOS UNEHAM19 04/07/2022
[   27.973021] Workqueue: events request_firmware_work_func
[   27.978961] RIP: 0010:kfree+0x64/0x250
[   27.983150] Code: 80 48 01 dd 0f 82 f5 01 00 00 48 c7 c0 00 00 00 80 48 2b 05 fe 68 f0 00 48 01 c5 48 c1 ed 0c 48 c1 e5 06 48 03 2d dc 68 f0 00 <48> 8b 45 08 48 8d 50 ff a8 01 48 0f 45 ea 48 8b 55 08 48 8d 42 ff
[   28.004139] RSP: 0018:ffffb8e48023fcc0 EFLAGS: 00010207
[   28.009974] RAX: 00006f5940000000 RBX: 36415f5f0004000f RCX: 00000000000006bc
[   28.017946] RDX: 0000000000000000 RSI: ffffffffc0e16e86 RDI: 36415f5f0004000f
[   28.025916] RBP: 00d8e24923001000 R08: ffffffff9fa41940 R09: ffffb8e48023fbd8
[   28.033888] R10: ffff90a742874000 R11: 0000000000000000 R12: ffff90a7425572b8
[   28.041861] R13: ffff90a8a41b5900 R14: 0000000000000000 R15: ffff90a742557000
[   28.049834] FS:  0000000000000000(0000) GS:ffff90a8a4180000(0000) knlGS:0000000000000000
[   28.058874] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   28.065294] CR2: 0000565365b792b0 CR3: 00000001075a6000 CR4: 0000000000350ee0
[   28.073266] Call Trace:
[   28.075997]  <TASK>
[   28.078341]  iwl_dealloc_ucode+0x36/0x110 [iwlwifi]
[   28.083809]  iwl_req_fw_callback+0x2cf/0x23d0 [iwlwifi]
[   28.089650]  ? free_fw_priv+0x9f/0x110
[   28.093835]  ? _request_firmware+0x42f/0x640
[   28.098604]  ? kfree+0x1f8/0x250
[   28.102206]  ? _request_firmware+0x42f/0x640
[   28.106974]  request_firmware_work_func+0x4a/0x90
[   28.112229]  process_one_work+0x1ee/0x390
[   28.116707]  worker_thread+0x53/0x3e0
[   28.120793]  ? process_one_work+0x390/0x390
[   28.125454]  kthread+0x124/0x150
[   28.129059]  ? set_kthread_struct+0x40/0x40
[   28.133731]  ret_from_fork+0x1f/0x30
[   28.137727]  </TASK>
[   28.140163] Modules linked in: sha512_generic(E) snd_intel_sdw_acpi(E) snd_hda_codec(E) ctr(E) snd_hda_core(E) drbg(E+) aesni_intel(E+) snd_hwdep(E) asus_nb_wmi(E) iwlwifi(E) asus_wmi(E) crypto_simd(E) snd_pcm(E) ansi_cprng(E) cryptd(E) iTCO_wdt(E) platform_profile(E) mmc_block(E) intel_cstate(E) mei_hdcp(E) evdev(E) snd_timer(E) intel_pmc_bxt(E) battery(E) cfg80211(E) snd(E) ecdh_generic(E+) sparse_keymap(E) iTCO_vendor_support(E) intel_wmi_thunderbolt(E) wmi_bmof(E) soundcore(E) efi_pstore(E) pcspkr(E) rfkill(E) watchdog(E) ecc(E) ti_adc128s052(E) crc16(E) industrialio(E) i915(E) ttm(E) drm_kms_helper(E) sdhci_pci(E) cec(E) cqhci(E) ahci(E) rc_core(E) libahci(E) sdhci(E) drm(E) libata(E) mei_me(E) mmc_core(E) scsi_mod(E) mei(E) scsi_common(E) intel_pmc_core(E) acpi_tad(E) acpi_pad(E) button(E) fuse(E) configfs(E) efivarfs(E) ip_tables(E) x_tables(E) autofs4(E) nfsv3(E) nfs_acl(E) nfs(E) lockd(E) grace(E) sunrpc(E) fscache(E) netfs(E) spi_pxa2xx_platform(E) dw_dmac(E) dw_dmac_core(E)
[   28.140232]  crc32_pclmul(E) crc32c_intel(E) igb(E) i2c_i801(E) i2c_smbus(E) i2c_algo_bit(E) dca(E) igc(E) xhci_pci(E) ptp(E) xhci_hcd(E) pps_core(E) usbcore(E) intel_lpss_pci(E) intel_lpss(E) usb_common(E) idma64(E) fan(E) wmi(E) video(E) pinctrl_elkhartlake(E)
[   28.263736] ---[ end trace 2553bae38d1333f9 ]---
```

To overcome this, use the following image as it includes a newer kernel
v5.15.34, where the iwlwifi issue is fixed: weekly/v2022/20220803.0154

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Use a more recent image, as there is no support for the UP Squared 6000
board in v2019.1.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The recently added UP Squared 6000 device uses the eMMC for deployment,
which is different from the default setting (sd-card). Let's add a
parameter so each LAVA device can set the deploy device in each case.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

The Apertis amd64 reference platform -MinnowBoard Turbot board- went out
of production a while ago and are no longer available. The UP Squared
6000 board has been chosen as the new reference platform, so let's
replace it.

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Ryan Gonzalez authored 2 years ago and Walter Lozano committed 2 years ago

This should help prevent issues in the future where the release is
changed without adding it to the ref-bindings.

https://phabricator.apertis.org/T8979



Signed-off-by: Ryan Gonzalez <ryan.gonzalez@collabora.com>

Ariel D'Alessandro authored 2 years ago and Walter Lozano committed 2 years ago

RPi AUM tests were disabled on Lava due to an issue on RPi4 U-boot.

The origin of this issue was that the same U-boot's defconfig
`rpi_arm64_defconfig` is used for both RPi3 and RPi4. This defconfig
sets the bootcount mmc dev/part to `0:1`. For RPi3 this works fine, as
mmc0 is the SD card, but on RPi4 mmc0 is used to control the SDIO for
wireless.

This has been fixed already by setting the aliases on RPi4 dtb for its
mmc devices. Thus, AUM tests can be re-enabled on RPi.

This reverts commit  4d06a3d1:

  Author: Frédéric Danis <frederic.danis@collabora.com>
  Date:   Wed Feb 16 17:07:05 2022 +0100

      lava: Prevent AUM tests to be generated for RPi board

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Now that QA Report App renders the test cases in its interface the deploy
using apertis-test-cases-web is not longer needed.

Drop the deploy of this repo and also the deploy in forks to avoid confusions.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>

Dylan Aïssi authored 2 years ago and Ryan Gonzalez committed 2 years ago

folks-inspect has been removed from Apertis and cannot be used anymore to
test bustle. Instead, we can use the dbus test tool to generate traffic.

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Signed-off-by: Apertis CI <devel@lists.apertis.org>

The goal of this test is to ensure we can execute the getting-started
instructions in the apertis-image-recipes README from inside the Apertis
SDK running on Virtualbox.

We don't actually have Virtualbox in LAVA, so it must be simulated.
As Virtualbox doesn't support nested KVM virtualization, let's remove
the KVM module. This should trigger the chroot-based Debos backend
instead of the fakemachine one"

Signed-off-by: Ariel D'Alessandro <ariel.dalessandro@collabora.com>

Dylan Aïssi authored 2 years ago and Walter Lozano committed 2 years ago

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Dylan Aïssi authored 2 years ago and Walter Lozano committed 2 years ago

folks-inspect has been removed from Apertis and cannot be used anymore to
test d-feet. Instead, we can use the dbus test tool to generate traffic.

Signed-off-by: Dylan Aïssi <dylan.aissi@collabora.com>

Walter Lozano authored 3 years ago and Detlev Casanova committed 2 years ago

In the past due to the reduce availability of amd64 hardware AUM tests were
disabled for that arch. However, it makes sense to at leas run a very basic
set of AUM test to confirm that the functionality works in amd64 arch.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>

Walter Lozano authored 3 years ago and Detlev Casanova committed 2 years ago

In the past it was decided to run AUM test only on Fixed Function images to
make a better use of LAVA, following this approach avoid running
aum-offline-upgrade-collection_id on HMI.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>

Now that a stable release has been published, use it as the starting
point for the upgrade compatibility tests.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Running LAVA tests shows random failures on some boards, causing several
tests to fail. After investigating it, this is most likely to happen
when using v2021 as first stage image to flash the device.

Since the v2022 is the first release which properly supports RPi4 switch
to it as first stage release.

Signed-off-by: Walter Lozano <walter.lozano@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Andre Moreira Magalhaes committed 3 years ago

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Detlev Casanova authored 3 years ago and Emanuele Aina committed 3 years ago

The grilo test installs 2 packages as dependencies. On OSTree, apt is
not available, so the test cannot be run on these images.

Signed-off-by: Detlev Casanova <detlev.casanova@collabora.com>

Add usage of boot-delay service to be able to know and have time to reset
the board before boot completes.

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>

Emanuele Aina authored 3 years ago and Frederic Danis committed 3 years ago

Merge the amd64 steps in the main testcase to avoid duplication.

Signed-off-by: Emanuele Aina <emanuele.aina@collabora.com>

Signed-off-by: Apertis CI <devel@lists.apertis.org>

Akshay M authored 3 years ago and Frederic Danis committed 3 years ago

Signed-off-by: Akshay M <M.Akshay@in.bosch.com>

LAVA version 2021.11 introduces a change in the way paths in compressed
overlays are handled breaking Apertis tests.

Fix the issue by specifying the path to be use to extract the compressed
overlay.

Additional information:

https://gitlab.collabora.com/lava/lava/-/commit/07048f042c65202eb75ac874cc93eabd2d96cd54



Signed-off-by: Walter Lozano <walter.lozano@collabora.com>

aum-offline-upgrade-branch is not generated with following error:
template variable not defined in aum-offline-upgrade-branch: 'osname' is undefined

Signed-off-by: Frédéric Danis <frederic.danis@collabora.com>