=========================================== #fdio-meeting: "FD.io CSIT Project Meeting" =========================================== Meeting started by mackonstan at 15:04:07 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/fdio-meeting/2024/_fd_io_csit_project_meeting_/fdio-meeting-_fd_io_csit_project_meeting_.2024-02-07-15.04.log.html . Meeting summary --------------- * Vratko Polak. (vrpolak, 15:05:13) * Agenda bashing (mackonstan, 15:06:20) * Mohit - have questions about documentation - about the tests and framework. (mackonstan, 15:09:14) * VPP update (mackonstan, 15:09:29) * TSC update (mackonstan, 15:09:39) * HW accelerator - SPR QAT2.0 (mackonstan, 15:10:32) * Kai - waiting for Peter to provide update on configuration change for QAT for queue and worker configurations (mackonstan, 15:11:32) * Peter - has some code, will be testing in the next few days. (mackonstan, 15:11:53) * Peter - likely the test cases will not have to be updated (mackonstan, 15:12:13) * Kai - had lots of feedback from VPP community, users were not aware of changes required to support QAT2 (mackonstan, 15:13:14) * HW accelerator - SNR QAT2.0 (mackonstan, 15:13:23) * Kai - similar story to SPR QAT2.0 (mackonstan, 15:13:43) * Kai - once we get guidelines working for SPR QAT2.0, they can be applied to SNR QAT 2.0 (mackonstan, 15:14:12) * Peter - single core/cpu tests are working, it's multi-core/cpu that are not working (mackonstan, 15:14:33) * HW accelerator - SPR DSA (mackonstan, 15:14:40) * Kai, Peter - the grub configuration is working for enabling DSA driver (mackonstan, 15:15:07) * Kai - tests are passing, specifically DSA memif tests, need to add them to job_specs (mackonstan, 15:15:54) * Kai - two known issues with DSA related to multiple DSA accelerator instance (mackonstan, 15:16:53) * Kai - for now DSA memif tests will be running master trending (mackonstan, 15:17:36) * Kai - currently not planning for DSA hoststack tests (mackonstan, 15:19:01) * CSIT-2402 release (mackonstan, 15:19:17) * Tibor - rc1 tests done, and almost done with vpp-23.10 retest (mackonstan, 15:20:00) * Vratko - looked at rc1 and retest results, mostly good, but one substantial regression (10-40%) on AVF on all Xeons and some other combinations (mackonstan, 15:24:39) * Vratko - related to vpp patch 40219: buffers: add compile-time option to set buffer alignment | https://gerrit.fd.io/r/c/vpp/+/40219 (mackonstan, 15:26:27) * LINK: 40219: buffers: add compile-time option to set buffer alignment | https://gerrit.fd.io/r/c/vpp/+/40219 (mackonstan, 15:26:30) * Physical Infrastructure (mackonstan, 15:27:02) * Virtual Infrastructure, CI/CD (mackonstan, 15:27:17) * Environment, Framework (mackonstan, 15:27:29) * Vratko - asking Juraj when it's ready for 1n-alt vpp_device to start voting (mackonstan, 15:28:30) * Juraj - currently waiting for mlx cx6 cabling to be completed, followed by testing, and then if stable will enable voting - worst case 2 weeks, best case much sooner (mackonstan, 15:29:38) * Peter - cables arrived, some questions from Ali, Juraj to follow up (mackonstan, 15:30:13) * Other topics (mackonstan, 15:30:53) * Kai - question about the RSS keys patch in vpp (mackonstan, 15:31:05) * Maciek - re Mohit's documentation questions - let's follow up over email, pls send email to csit-dev@lists.fd.io (mackonstan, 15:32:19) * Mohit - Questions about CSIT documentation. Vratko gave some answers. (vrpolak, 15:43:12) Meeting ended at 15:43:17 UTC. People present (lines said) --------------------------- * mackonstan (35) * collab-meetbot (10) * vrpolak (3) * dwallacelf (0) * jlinkes (0) * Kai17 (0) * tifrank (0) Generated by `MeetBot`_ 0.1.4