Eyal Lavee (Mellanox)
Joseph Gasparakis (Intel)
Paul Carver (AT&T)
Henry Fowler (AT&T)
Edward Ting (Lenovo)
Kiran KN (Juniper)
Simon Horman (Netronome)
- Release schedule for 5.1
- Patches from Netronome
- If we have an interface in TF where a vendor can implement some functionality in form of a binary (which source code will not be distributed in any shape or form) that dynamically links in kernel space, would that be ok from legal perspective?
- And if so, is there any restriction in the license that this binary (that will be loaded in Linux kernel space) will have?
Ultimately, it seems like whichever party is going to have an issue will have to do the legal research is this part Juniper and/or whoever distributes a commercial version of TF?
- Any other concerns? Most concerns were discussed in last week's SmartNIC meeting on 2018-10-18
- TSC/TC Elections: voting now for two members -vs- voting now for all TSC
- Tools for Release Notes - Ed Ting
- 5.1 release review
- Dev complete is now November 15th
- Release targeted for December 15th
- Discussion of smartNIC offload, licensing and code changes
- There might be some legal issues with whoever distributes the binary but it seems it doesn't affect TF community. Can we get feedback from LF's legal expert on Friday's meeting?
- About CI we need a 3rd party CI that will require each vendor's hardware (which is fine and expected) but there is some discussion on how this will integrate with the rest of the CI.
- There are questions about documentation. Frikke from Netronome said that they can provide it.
- Joseph brought up the question of tainted kernel. (Joseph can you clarify your concern here?)
- Eyal raised the point that there is a tension between having the community work on a common interface, and enabling vendor-specific interfaces which may reduce the motivation for a common industry standard. He added that it is understood that it will take time to develop the common interface, and that creates pressure for interim (vendor-specific) solutions.
- Jan raised the point that this approach would be the first proposal for the full offload, and would enable OpenStack work to begin sooner.
- Joseph proposed that the entire TSC be voted on - not just the replacements. A vote is needed on Joseph's proposal.
- Paul pointed out that we do not have a list of who voters are. We have rules in the by-laws, but have not applied the rules to create a list of people.
- The official by-laws are in a PDF, per Casey (no longer editable, of course). It was agreed that this should be posted in a conspicuous location.