Skip to content

Conversation

@wkz
Copy link
Contributor

@wkz wkz commented Nov 19, 2025

Description

Checklist

Tick relevant boxes, this PR is-a or has-a:

  • Bugfix
    • Regression tests
    • ChangeLog updates (for next release)
  • Feature
    • YANG model change => revision updated?
    • Regression tests added?
    • ChangeLog updates (for next release)
    • Documentation added?
  • Test changes
    • Checked in changed Readme.adoc (make test-spec)
    • Added new test to group Readme.adoc and yaml file
  • Code style update (formatting, renaming)
  • Refactoring (please detail in commit messages)
  • Build related changes
  • Documentation content changes
    • ChangeLog updated (for major changes)
  • Other (please describe):

@wkz wkz force-pushed the post-post-image branch 10 times, most recently from bf65879 to 8d8f587 Compare November 22, 2025 22:04
@wkz wkz force-pushed the post-post-image branch from 8d8f587 to 7b458e2 Compare December 3, 2025 14:12
wkz added 19 commits December 3, 2025 16:45
A bunch of different artifacts should follow the same naming scheme,
depending on a bunch of config settings + whether we're building a
release or not.

Therefore, provide a single definition of this that we can reuse to
name disk images, upgrade packages, etc.
The board/*/*/*.mk is very broad, intended to hit all board specific
definitions, but may also cause duplicate inclusions, e.g., in
board/common.

Let each architecture do the inclusion instead.
Rather than using the creation of a signed image as a proxy for
whether the trusted keys should be installed RAUC/U-Boot's trust
stores, use the dedicated option.
Add scaffolding for breaking out image generation to separate make
targets.
Add a generic image target to build aux.ext4, which can be used both
when creating target-specific SD-card images, and when creating
regular disk images.

While we're here, make sure that we don't need a RAUC bundle in order
to generate aux.ext4 (which mkrauc-status.sh did). This saves us time
on _every_ incremental build.
We have not installed .dtb:s to $O/images/ for quite some time, and
nobody cared. That goes to show that this is not really used.

The image is still useful at times, so if it needed in the future,
then we can resurrect it from the logs and refactor it to an image
package.
Limit support to x86, like we do on the "official" appliances on the
marketplace. Aarch64 has never really been used AFAIK.

Avoid the os-release import, since all that info is not important now
that these appliance files are only for development scenarios. In all
other cases, the official one, based on a proper release, should be
used.
This is an internal artifact, so there is no need to give it a branded
name.
Now that all components are generated from their own fragments, we
have no need for post-image.sh anymore.
Previously, post-image.sh was able to create QEMU images from an
existing release tarball. This was useful when you wanted to test a
new bootloader build without having to wait for a full Infix build.

Restore this capability by adding a separate image target for it, and
then allow image-itb-qcow to source its input images from that instead
of a locally build squash+aux.
This CPU has less emulation overhead than `max`, which is what we
mostly care about.
Use the new image-itb-dl-release to compose a QCOW with Infix and
U-Boot, in the same way that we previously did in post-image.sh.
@wkz wkz force-pushed the post-post-image branch from 7b458e2 to bac11da Compare December 3, 2025 15:46
@wkz wkz marked this pull request as ready for review December 3, 2025 15:46
@wkz wkz requested review from mattiaswal and troglobit December 3, 2025 15:46
Copy link
Contributor

@troglobit troglobit left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Much needed cleanup, nice work!

@wkz
Copy link
Contributor Author

wkz commented Dec 3, 2025

One failing test on styx:

2025-12-03 16:03:42 # Starting (2025-12-03 16:03:42)
2025-12-03 16:03:42 # host:    monza
2025-12-03 16:03:42 #     mgmt:    sw1p1
2025-12-03 16:03:42 # target:  styx1
2025-12-03 16:03:42 #     mgmt:    e28
2025-12-03 16:03:42 #
2025-12-03 16:03:42 # Waiting for DUTs to become reachable...
2025-12-03 16:03:42 # Probing styx1 on port sw1p1 for IPv6LL mgmt address ...
2025-12-03 16:03:42 # Testing using NETCONF
2025-12-03 16:03:42 # Connecting to mgmt IP fe80::96de:eff:fe05:2614%sw1p1:830 ...
2025-12-03 16:03:42 # YANG models downloaded.
2025-12-03 16:03:44 # Waiting for DUTs to become reachable...
2025-12-03 16:03:44 # Probing styx1 on port sw1p1 for IPv6LL mgmt address ...
2025-12-03 16:03:44 ok 1 - Set up topology and attach to target DUT
2025-12-03 16:03:44 # GPG not available on target - skipping encryption tests
2025-12-03 16:03:44 ok 2 - Check for GPG availability on target
2025-12-03 16:03:44 # support collect failed with return code 255
2025-12-03 16:03:44 # stderr:
2025-12-03 16:03:44 not ok 3 - Run support collect with --work-dir and short log tail
2025-12-03 16:03:44 # Exiting (2025-12-03 16:03:44)
2025-12-03 16:03:44 # Traceback (most recent call last):
2025-12-03 16:03:44 #   File "/home/styx/.ghmoon/kernelkit/infix/git/test/case/misc/support_collect/test.py", line 51, in <module>
2025-12-03 16:03:44 #     raise Exception("support collect command failed")
2025-12-03 16:03:44 # Exception: support collect command failed
2025-12-03 16:03:44 #
test error, no plan

Not sure what happened there (or why the report was not uploaded). But seems more likely to be some flaky thing in the test (given that it was added very recently), or what do you think @troglobit?

@troglobit
Copy link
Contributor

One failing test on styx:
[...]
2025-12-03 16:03:44 not ok 3 - Run support collect with --work-dir and short log tail
2025-12-03 16:03:44 # Exiting (2025-12-03 16:03:44)
2025-12-03 16:03:44 # Traceback (most recent call last):
2025-12-03 16:03:44 # File "/home/styx/.ghmoon/kernelkit/infix/git/test/case/misc/

Not sure what happened there (or why the report was not uploaded). But seems more likely to be some flaky thing in the test (given that it was added very recently), or what do you think @troglobit?

Agreed, very likely flaky new test. Created #1303 for this.

@troglobit troglobit merged commit ec0ed82 into main Dec 3, 2025
5 of 6 checks passed
@troglobit troglobit deleted the post-post-image branch December 3, 2025 19:32
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants