Attendees: Philippe Gerum, Jan Kiszka (Siemens), Florent Pirou (Intel), Hongzhan Chen (Intel), Fino Meng (Intel), Florian Bezdeka (Siemens), Jeroen Van den Keybus (National Instruments), Steven, Sonnypoint, Yang Yao (Intel), Dao Zhang (Intel), Jianbo Cao (Intel), Rick Wang (Intel)
Agenda: *Community call logistics - Rick *Community update - Jan *Dovetail, CXP and EVL update - Philippe *Clarify Xenomai Roadmap - Rick *Open *Community call logistics ----------------------------------- Biweekly every odd week (week #1 - 2021.1.1) Next call is on Jan 27th. Europe and Asia friendly time: 7:00-8:00 UTC | 8:00-9:00 Munich | 15:00-16:00 Shanghai We can't find a good time for all geos and hard to split the call. For people can't attend the regular call, it's possible to setup special sessions per request when needed. Dial in info: https://xenomai.org/pipermail/xenomai/2020-December/043906.html *Community update ---------------------------- Current stable 3.1 has no new patches in the past two weeks. Master branch is working on 3.2. Lava CI is doing well with test matrix growing. *Dovetail, CXP and EVL update ---------------------------------------- Dovetail 5.11 is WIP. EVL network stack is WIP. It will not be RTNet. Lack of feedback in mailing list to CXP. CXP spec definition plan to complete in Apr 2021. Porting for Dovetail in Xenomai is expected to complete by end of Jan. Thanks Hongzhan Chen from Intel for the hard work! >From Feb, move on to CXP RTDM, POSIX library, etc. Clarify Xenomai Roadmap ---------------------------------- 3.1 - Current stable based on I-pipe 3.2 - Add dovetail to Xenomai. Cobalt core will run on top of an abstraction layer which interface to both I-Pipe and Dovetail. 3.3 - Common Xenomai Platform. Libcobalt and Cobalt core will be updated to prepare for transition to EVL. 4.0 - EVL. Both libevl and libcobalt (CXP) will be supported. Cobalt core will be replaced by EVL core and a POSIX driver layer on top of it. Open ---------------- Discussion on EVL network stack. Philippe is working on a new design to implement the RT network driver alternative to XDP or DPDK approaches. Thanks, Rick