13:04:17 #startmeeting pci_passthrough 13:04:18 Meeting started Tue Aug 12 13:04:17 2014 UTC and is due to finish in 60 minutes. The chair is irenab. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:04:19 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 13:04:21 The meeting name has been set to 'pci_passthrough' 13:04:43 hi 13:04:54 baoli sent email that he cannot make it today 13:05:23 `o 13:05:44 2 patches in, great 13:05:52 cool! 13:05:52 I seemed to have missed that email 13:06:27 I am back after a week off, still trying to catch up 13:07:16 Does anyone has some updates? 13:07:47 I am uploading a new patchset today based on current review comments 13:08:10 folks can take a look after this update 13:08:26 sadasu: Did you do some code reuse of NIC switch MD? 13:08:34 great, i will check tomorrow, 13:08:43 sadsu: will review as well 13:08:59 irenab: lets talk about the code reuse 13:09:13 sadsu: sure 13:09:18 we briefly touched upon it the last meeting 13:09:26 not in detail since you weren't there 13:09:54 sadasu: sorry, didn't check the log yet... 13:10:38 so, in the current set of diffs cisco mech driver is inheriting from MechanismDriver 13:10:50 the plan is to go with this until Juno-3 13:11:30 sadasu: so for now there is no reuse, right? 13:11:37 right after Juno 3, we can commit changes to srio-nic-switch mech driver (which is going to be the common one) 13:11:50 to make it more common based on discussion in the ML 13:12:25 my mech driver will inherit from this common srio-nic-switch class as a subsequent bugfix again right after Juno-3 13:12:37 sadasu: to save time for everyone, I'll take a look at the last meeting log and maybe comment on review, ok? 13:13:07 sadasu: generally sounds reasonable 13:14:13 irenab: it is clear that we can re-use code...but just doing that as a seperate change 13:14:56 hoping to make it easier on the reviewers since there is already a lot to come upto speed on here 13:15:30 sadsu: agree, probably most of the features are implemented in the same maner, number of L2 agents is good example... 13:16:15 the common code is about 25% of the changes so it doesn't look like a replica of sriov-nic-switch 13:16:48 sadasu: so for the code review of cisco MD, I am assuming no code reuse 13:17:24 correct....code re-use will be a bugfix within Juno cycle 13:17:37 r u fine with that? 13:18:38 sadasu: I am fine with it. 13:18:50 irenab: cool & thanks 13:19:23 heyongli: Did you have a chance to check the sriov-nic-switch for Intel NICs? 13:21:21 seems that nova patches makeing a lot of progress. 13:21:38 irenab, our CI under internal testing 13:21:53 do you tested? 13:22:22 if not, i need set up a manual testing bed. 13:22:53 heyongli: let me know if there are some issues. Yes, we tested and there is a CI job we run under Mellanox External Testing for this. Currenlty take the nova patches and apply them as part of the devstack deployment 13:23:12 irenab, if you have sriov verify step by step guide, could you share it with me? 13:23:58 heyongli: actually, for now its is just port-create with vnic_type=direct, then nova boot 13:24:04 there are some neutron part setting up, if you can share, that's great help to me 13:24:15 ok, got 13:26:14 any other topic? 13:27:09 heyongli: Ok, will check 13:28:03 heyongli: I can send you the neutron config we use 13:28:11 thanks. 13:28:55 no more topics for me right now, maybe we needto prepare the list of next issues to discuss next meeting, probably need tempest, documentations, ... 13:29:35 I'll start the list on meeting wiki page, will send to all to update 13:29:50 thanks irenab 13:30:09 then end meeting today? 13:30:17 yes..makes sense 13:30:38 great, see you next week and on reviews 13:30:43 #endmeeting