Community CI Meeting Minutes - January 8, 2026

Community CI Meeting Minutes - January 8, 2026

[Patrick Robb]

[-- Attachment #1: Type: text/plain, Size: 4137 bytes --]

January 8, 2026

Attendees
1. Patrick Robb
2. Aaron Conole
3. Paul Szczepanek
4. Andrew Bailey

Minutes

General Announcements
* Patchwork has been slow yesterday/today and timing out on some API
requests, which may be affecting the CI Labs. Patrick emailed Ali asking
about this.
   * In December Ali made various suggestions to CI labs regarding how they
could reduce their requests being made to patchwork, and to the best of our
knowledge all the labs completed these improvements.

CI Status

AWS Lab
* Running normally.

UNH-IOL Community Lab
* UNH missed testing on a significant amount of patchseries over the
Holiday break, which it is now backfilling. There are still about 10
patchseries remaining, but we should be caught up on results tomorrow.
* DTS:
   * Nic_single_core_fwd: Moving CI runs of this testsuite from legacy DTS
to new DTS. This should be all done by the end of next week.
   * Other than that, the only remaining DTS testing coming from the Legacy
framework will be cryptodev, but Andrew Bailey is working on this currently
for 26.03, so we should be completely off of the Legacy framework after
that.
* New Distros:
   * Fedora 43
   * Rhel 10
* DPDK LTS build failures on Fedora43
   * Stephen Hemminger mentioned on Bugzilla there will not be patches
backported to fedora 43 to resolve this
* We have moved all the testing over from our Windows 2022 machine to
Windows 2025. Andre Muezerie is trying to get us a Windows 2025 License key.
   * Our unit testing pipeline for Windows is failing for 23.11 and 24.11
LTS due to the fact that it is using the new devx environment variables.
Resolution is probably simply switching the environment variable name based
on whether we are testing <25.11 or >=25.11

Intel Lab
* Akhil mentioned on Slack it is problematic that they dont publish logs
alongside failures
* David mentioned this has been mentioned in the past and that they have
not had the bandwidth to address it in the past.

Github Actions Robot
* Sonnet Claude Code Review:
   * Did a pass against a DPDK patch Stephen Hemminger sent:
      *
https://patches.dpdk.org/project/dpdk/list/?series=36859&state=%2A&archive=both
   * Coderabbit (the agent previously tested) was just creating noise, but
Claude seems to be performing better
* The SQL syntax issue for the new SQLite version and Ozlabs https
patchwork issues from December has been resolved, so the testing is back
online

Loongson Lab
* Patrick to check in with Bibo about whether they have reduced their
requests being made to patchwork for the recheck framework scripting.
DTS Improvements & Test Development
* Cryptodev:
   * Should the cryptodev stats printing methods be staticmethods or module
level functions?
   * Patrick has sent reviews on ⅔ patches, will send the final review today
   * Need to automate creating the virtual functions on the crypto devices.
      * Also, strategy is to create the max number of possible VFs for each
cryptodev physical function based on the value within sriov_max_vfs
* Flow Offload:
   * Dean sent the RFC for the new extensive flow offload testsuite
      * Uses the generic generator for creating flow rules and the
associated packets for DROP, QUEUE, and MODIFY actions
      * Uses explicitly written rules and assertions for the JUMP rule
testing
* Virtio:
   * RFC has been sent
* Misc:
   * Patches to be merged:
      *
https://patchwork.dpdk.org/project/dpdk/patch/20251106181355.209706-1-dmarx@iol.unh.edu/
      *
https://patchwork.dpdk.org/project/dpdk/patch/20251126141650.367633-1-abailey@iol.unh.edu/
      *
https://patchwork.dpdk.org/project/dpdk/patch/20251215214049.694604-1-abailey@iol.unh.edu/
      *
https://patchwork.dpdk.org/project/dpdk/patch/20251217165716.705128-1-abailey@iol.unh.edu/

Any other business
   * Do we have an ARM machine with E810 in the labs (driver patch)? There
is an IAVF issue with ARM and E810.
   * Yes, we do. Patrick will connect with Jay from ARM on how the UNH lab
can verify their fix on E810.
   * Next meeting is January 21, 2026

[-- Attachment #2: Type: text/html, Size: 4986 bytes --]