13:00:09 <baoli> #startmeeting PCI Passthrough
13:00:09 <openstack> Meeting started Tue Sep  2 13:00:09 2014 UTC and is due to finish in 60 minutes.  The chair is baoli. Information about MeetBot at http://wiki.debian.org/MeetBot.
13:00:10 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
13:00:12 <openstack> The meeting name has been set to 'pci_passthrough'
13:00:26 <baoli> Hi
13:00:34 <irenab> hi
13:00:54 <baoli> irenab, thanks for coming up with the etherpad
13:01:34 <irenab> baoli: glad to do it
13:01:37 <heyongli> hi
13:01:49 <baoli> heyongli, hi
13:02:08 <irenab> please add your inoputs there, maybe we need priority later on
13:02:54 <baoli> irenab, it's a good list. Did you forget to add the MD refactoring?
13:02:55 <irenab> baoli: any updates on patches? Currently it seems we are going to miss the deadline...
13:03:25 <irenab> baoli: You are right! I'll add it asap
13:03:44 <heyongli> 9.4 coming soon, baoli , what's the status?
13:03:56 <baoli> ok, status is coming up.
13:05:00 <baoli> On Thursday's nova meeting, the cores voted to fixing the two issues we mentioned in last meeting before juno FF (feature freeze), and give feature freeze exception to sr-iov patches
13:05:28 <irenab> baoli: good news!
13:05:33 <heyongli> sound we can make it
13:05:53 <irenab> baoli: Is there any help you need with patches?
13:06:41 <baoli> dansmith, russellb have volunteered to code. So the first issue converting requested_network megatuple to objects is in good shape. Dansmith has done a good job coding it up quickly
13:06:59 <baoli> there are still two patches pending review for this issue
13:08:03 <heyongli> is this means we depending on dansmith's code?
13:08:05 <irenab> baoli: can you put a link for patches for review. We all can help in review
13:08:12 <baoli> The second issue (pci request storage) is being worked on, russellb came up with the initial two patches, and dansmith took the torch afterwards.
13:08:25 <baoli> I need to talk to him on its progress today.
13:09:50 <baoli> https://blueprints.launchpad.net/nova/+spec/pci-passthrough-sriov. the last two are for the pci request storage issue
13:10:09 <heyongli> got it
13:10:19 <irenab> thanks
13:12:39 <baoli> https://review.openstack.org/#/c/115334/, and you can follow the dependency links for all the remaining patches for the requested network megatuple conversion
13:12:48 <irenab> any other updates?
13:13:41 <baoli> I have made the change to use the requested netowrk objects in my local repo, and it works.
13:14:27 <heyongli> cool
13:14:31 <baoli> Once the patches for the two issues are merged, I can update my remaining three patches that incorporate them
13:15:00 <irenab> baoli: so your patches are on hold for now?
13:15:05 <heyongli> core involved in is a good thing, make sure we can make it merge in time
13:15:59 <baoli> https://review.openstack.org/#/c/114286/ has been approved, but due to another patch that had changes in the same area got merged ahead of mine, and caused conflict, I have to rebase manually and test everythine, etc. Now it's waiting for approval again.
13:16:22 <baoli> irenab, three of them are on hold.
13:17:13 <baoli> heyongli, can't complain with them involved and help in coding.
13:17:15 <irenab> baoli: when there is something for review, feel free to ping me to join the review to push it faster
13:17:34 <baoli> irenab, thanks, sure.
13:17:40 <heyongli> baoli, that's really good thing
13:18:38 <baoli> so at this point, all we can do is to work with them and hopefully all the patches can land in juno.
13:19:14 <heyongli> i feel and hope it could
13:19:38 <irenab> then we can plan for next features
13:20:47 <baoli> irenab, yes. Let's spend a little time on it.
13:21:16 <irenab> baoli:sure
13:21:20 * johnthetubaguy is curious how ready to merge https://blueprints.launchpad.net/nova/+spec/pci-passthrough-sriov is, given he is going through the list applying −2s
13:23:35 <irenab> john: we just discussed it few mins ago, do you want a short recap?
13:25:26 <johnthetubaguy> irenab: I see a −1 on a patch, is that some disagreement? wondering if its resolved?
13:25:57 <johnthetubaguy> there was talk of changing the interface with neutron, if I remember correctly, removing a mega tuple, or somesuch
13:26:20 <baoli> johnthetubaguy, per last week's nova meeting, we need to fix two preexisting issues, metatuple is one of them
13:26:34 <baoli> another one is the pci requests storage
13:26:40 <johnthetubaguy> basically, need all code approve by tomorrow, or it will need a FFE (exception)
13:26:49 <johnthetubaguy> OK, how are those patches going?
13:26:54 <johnthetubaguy> are you rebased on top of them yet?
13:27:23 <baoli> john, the megatuple one is almost done. I need to talk to dansmith on the pci requests storage
13:27:25 <dansmith> I'm not done with the storage refactor ones yet
13:27:36 <dansmith> megatuple had almost all merged on friday
13:27:52 <johnthetubaguy> dansmith: OK, so close but we are heading towards FFE I guess
13:27:57 <baoli> dansmith, hi, nice having you here
13:28:08 <dansmith> johnthetubaguy: yeah, the plan is that this gets an FFE if we can get the refactors done by FF
13:28:14 <dansmith> johnthetubaguy: per mr. ptl
13:28:30 <johnthetubaguy> right, I remember reading that in the nova-meeting now
13:28:32 <dansmith> just checked, two more megatuple patches to merge
13:28:44 <johnthetubaguy> dansmith: thanks
13:28:45 <dansmith> just needs +As
13:29:04 <dansmith> https://review.openstack.org/#/q/status:open+project:openstack/nova+branch:master+topic:refactor-megatuple,n,z
13:29:59 <johnthetubaguy> cools
13:30:20 <johnthetubaguy> baoli: sorry for interrupting, that makes sense, I will leave the code up as long as I can.
13:30:45 <baoli> john: thanks.
13:31:02 <johnthetubaguy> my only advice is to rebase early on top of those, and make sure you people are all +1 on all of the changes, if you all agree on them
13:31:09 <johnthetubaguy> that should make it easier at crunch time
13:31:19 <johnthetubaguy> thanks all
13:31:31 * johnthetubaguy bravely runs away, away
13:31:50 <heyongli> johnthetubaguy, sure
13:31:50 <baoli> dansmith, for the second issue, do you have new patches to be uploaded soon?
13:32:44 <dansmith> baoli: it was a holiday weekend here, so I'm just getting back to it.. I put up some patches on friday, but they still need a lot of work
13:32:57 <dansmith> baoli: not quite ready to rebase onto those yet, but I'll keep you updated of course
13:32:59 <johnthetubaguy> (also, please help review dansmith patches to help them get in quicker, forgot that, but I guess its obvious)
13:33:11 <baoli> dansmith, I saw that. Thanks for helping out.
13:33:53 <irenab> guys, please post all patches required to complete the solution, so we all can be on track to review
13:34:38 <baoli> irenab, for the remaining patches, I think that you can just get to them from the BP page
13:35:02 <irenab> baoli: ok, thanks
13:36:01 <baoli> ok, do we want to go back to the kilo feature list?
13:36:22 <irenab> baoli: works for me
13:36:58 <baoli> irenab, the 9th item, can you explain in more detail?
13:38:07 <irenab> baoli: this item is for SR-IOV support for ironic, in case some Hypervisor services (like storage) would run on some VF
13:38:57 <baoli> irenab, I see.
13:39:20 <irenab> baoli: I'll try to add description on etherpad to clarify
13:39:39 <baoli> sounds good
13:40:08 <baoli> at some point, maybe starting next week, we should go over the list one by one.
13:40:13 <irenab> since there are a lot of areas to expand with regards to SR-IOV, we need to see what seems valuable and prioritize
13:40:32 <irenab> baoli: agree
13:40:46 <baoli> irenab, yes. And we'd like to see who wants to work on what.
13:40:58 <irenab> baoli: agree again
13:41:15 <baoli> Ok, anything else?
13:41:26 <heyongli> no from me
13:41:37 <irenab> not foe me
13:41:42 <baoli> thanks everyone, see you next time
13:41:58 <baoli> #endmeeting