February 14, 2024
#####################################################################
Attendees
* Patrick Robb
* Juraj Linkeš
* Thomas Monjalon
* Gregory Etelson
* Juraj Linkes
* Luca Vizzarro
* Nicholas Pratte
#####################################################################
Agenda
* Additions to the agenda
* Patch discussions
* DTS Developer documentation
* Bugzilla discussion
#####################################################################
Minutes
=====================================================================
General Announcements
* Added developers for DTS: Nick from UNH is starting on DTS now, and 1-2 more people from UNH will be starting on this project in the near future.
* The first thing Nick will do is build the DTS API docs from Juraj’s patch and provide a review
* He will be joining the meetings going forward
* Recordings of these meetings can be seen at the below url, by clicking on the meeting date, then the 3 dots, then share meeting recording: https://projectadmin.lfx.linuxfoundation.org/project/a0941000005osNmAAI
=====================================================================
Patch discussions
* Docs improvements: Luca’s patch was accepted
* Thomas noted that the schema for DPDK build options was probably not well thought out, and just imported from old dts.
* Noting I remember Juraj mentioned once that cross compile targets probably are not needed
1. This is not a priority - we should wait and see if there is interest later.
2. https://bugs.dpdk.org/show_bug.cgi?id=1360
3. There are some unused values which can be removed
4. Lcores value: change from empty string to “all” (if not specifying a specific list of cores)
5. Port topology: should this be a part of the conf schema?
1. Some testsuites may allow for different port topologies and should be configurable, others may have specific requirements which should be built into the capabilities assessment per testsuite
6. Ports: If we specify port 0 on machine 1 connects to port 0 on machine 2, we don’t then need to specify port 0 on machine 2 connects to port 0 on machine 1 - it’s redundant configuration (and introduces possible setup human error)
* Improved error messages:
* Intent for Luca’s patch is to improve patch reporting, as well as log errors from remote sessions
* Scatter
* Per conversation at the previous meeting, adding a 2nd testcase to the testsuite (one will include the rx offload testpmd flag, one will not)
* Ran into a type error yesterday from the xmlrpc client when running on one of the community lab’s testbeds, but will continue debugging
1. Error is about a list not being callable? Jeremy Spewock will send the error and latest diff to Juraj on Slack
* When running a suite in which testpmd is started for 1 testcase, then stopped, then started again for subsequent testcases, this can lead to a timeout. The timeout comes from paramiko.
* Device capabilities: Only source of truth for collecting capabilities should be any information we can gather via testpmd
* Needs to build a way to mark testsuites with capabilities
* Will build a list of capabilities per device
* Dockerfile patch:
* Patrick Robbneeds to apply this, run from container on one of our baremetal servers, and ask Thomas to merge
* API Docs:
* Needs reviews - Nick will provide one
* Juraj: Small patch for storing the output from remote commands (strips the whitespace)
* Testcase blocking: When a testcase fails, everything under it is blocked. This patch is final from Juraj’s point of view, and Jeremy has provided some comments.
* More reviews are needed. Patrick and Nick can do reviews, also Luca possibly.
=====================================================================
Bugzilla:
* If we can assign configuration file bugzillas to two people, they can provide reviews for one another
* We will discuss bugzilla tickets more tomorrow at the CI meeting
#####################################################################
Any other business
* Juraj is on vacation 4th of March to 29 March.