#####################################################################
May 1, 2025
Attendees
1. Patrick Robb
2. Paul Szczepanek
3. Luca Vizzarro
4. Aaron Conole
#####################################################################
Minutes
=====================================================================
General Announcements
* DPDK Summit Prague is May 8-9
=====================================================================
CI Status
---------------------------------------------------------------------
UNH-IOL Community Lab
* UNH shared out the x86 QAT cryptographic test metrics with Kai Ji, and have enabled reporting to patchwork for this test in CI
* Baremetal updates:
* X86 E810 - it looks like we finally have a working solution for this, which is pairing up two separate E810 NICs (with trex on both) up against 1 dual port e810 NIC acting as the DUT
* It also requires a patch for the single core forwarding test as it pertains to how trex crafts the 64 byte frames. After this we can craft 64 byte frames and send at 250 MPPS, which is more than we need.
* Cody & Patrick can dry run this and then bring it online this week.
* We are trying to add LTS tests for the ARM Grace server, and noticed there is a meson error for 22.11 and 23.11, which the maintainers indicated could be resolved by registering the neoverse v2 SOC
* https://inbox.dpdk.org/dev/bug-1702-3@http.bugs.dpdk.org%2F/
* https://inbox.dpdk.org/dev/bug-1703-3@http.bugs.dpdk.org%2F/
---------------------------------------------------------------------
Intel Lab
* None
---------------------------------------------------------------------
Github Actions
* There was a partial outage 2 weeks ago. The tests still ran, but the emails were not sent out (so, no results on patchwork). If you go to your series URL on github, the results are there. All set now.
* Cirrus CI: work is paused for now. This requires good support for FreeBSD, which is not happening now - builds are flaky and meson tests are also not well supported and will not pass.
* Discussion pertaining to looking at polling for series using the events API. This events endpoint (with series created event) returns info that a series has been created, but returns a limited set of data in the payload, and this necessitates a followup request to patchwork. So, this seems like it would actually increase the amount of requests made to the patchwork server. Some related issues discussed are:
* You cannot query the events endpoint for only events from a particular project (this matters for patchwork instances with many projects under them). For DPDK there are only 4 projects under DPDK patchwork, so it’s not a huge deal, but still a small issue.
* The datetime that the series-created event returns is the datetimes of one of the commits in the series, not the datetime of when the series was submitted. So, this means that if you amend a commit (this does not update commit datetime) and resubmit a patchseries, the datetime on the series-created record will not be “updated”. This can cause us to miss series when polling via the events endpoint.
---------------------------------------------------------------------
AWS
* Their CI test lab is now online/reporting to patchwork, and runs a build + set of meson fast tests for x86 and ARM systems on AWS: http://mails.dpdk.org/archives/test-report/2025-April/874152.html
* Shai (aws) will be at DPDK Prague where we can discuss next steps with him
* Next steps is basically using DTS
* Patrick should encourage Shai to join this meeting periodically
---------------------------------------------------------------------
Loongarch Lab
* None
=====================================================================
DTS Improvements & Test Development
* Rte_flow: Dean sent out his new proposed test plan, based on Thomas Monjalon’s feedback: https://inbox.dpdk.org/dev/CABD7UXPfQcvLt4a5YrxSUaH5NNRJzZHjHUtXs9G-jBnNkiJw8w@mail.gmail.com/
* VF tests: Ian Stokes had said a possible footgun for running these with Intel NICs is that trust mode may have to be set before promiscuous mode works correctly. Patrick has confirmed that although testpmd “show port info” will return promiscuous=enabled without trust mode, it will not work properly until you enable trust like:
* Enable trust: “ip link set dev ethX vf 1 trust on”
* So, the testsuites requiring promisc mode will need this, and it needs to be run before binding to vfio-pci as it is an ip command
* From https://access.redhat.com/solutions/2620231 “Enabling promiscuous mode on an SR-IOV VF requires "trusted" mode.”
* Old testsuites which need to be rebased:
* Port control has been updated and resubmitted: https://patchwork.dpdk.org/project/dpdk/list/?series=35098
* Nick, please test/review and we can merge it
* TREX patchseries:
* https://patchwork.dpdk.org/project/dpdk/list/?series=35097
* Luca is refactoring the pdump packet capture testsuite to use the new dumpcap
=====================================================================
Any other business
* Some lab maintainers have not been joining the CI meetings regularly
* Does it make sense to have a quarterly “all hands” type meeting in which all lab maintainers are expected to join? Major points of lab cooperation could be discussed at these meetings, like patchwork usage policy/strategies etc.
* Next Meeting May 15, 2025