| *** svenkat has joined #openstack-powervm | 00:07 | |
| *** thorst has joined #openstack-powervm | 00:09 | |
| *** thorst has quit IRC | 00:13 | |
| *** thorst has joined #openstack-powervm | 00:44 | |
| *** thorst has quit IRC | 00:45 | |
| *** edmondsw has joined #openstack-powervm | 01:13 | |
| *** edmondsw has quit IRC | 01:18 | |
| *** thorst has joined #openstack-powervm | 01:46 | |
| *** thorst has quit IRC | 01:51 | |
| *** svenkat has quit IRC | 02:01 | |
| *** thorst has joined #openstack-powervm | 02:47 | |
| *** thorst has quit IRC | 02:51 | |
| *** edmondsw has joined #openstack-powervm | 03:02 | |
| *** edmondsw has quit IRC | 03:06 | |
| *** chhavi has joined #openstack-powervm | 03:48 | |
| *** thorst has joined #openstack-powervm | 03:48 | |
| *** thorst has quit IRC | 03:52 | |
| *** tonyb has quit IRC | 03:53 | |
| *** tonyb has joined #openstack-powervm | 04:00 | |
| *** edmondsw has joined #openstack-powervm | 04:50 | |
| *** edmondsw has quit IRC | 04:54 | |
| *** thorst has joined #openstack-powervm | 05:49 | |
| *** thorst has quit IRC | 05:54 | |
| *** k0da has joined #openstack-powervm | 06:03 | |
| *** thorst has joined #openstack-powervm | 06:51 | |
| *** mawi has joined #openstack-powervm | 06:53 | |
| *** thorst has quit IRC | 06:55 | |
| *** k0da has quit IRC | 07:06 | |
| *** thorst has joined #openstack-powervm | 07:51 | |
| *** thorst has quit IRC | 07:56 | |
| *** edmondsw has joined #openstack-powervm | 08:26 | |
| *** k0da has joined #openstack-powervm | 08:30 | |
| *** edmondsw has quit IRC | 08:30 | |
| *** thorst has joined #openstack-powervm | 08:52 | |
| *** thorst has quit IRC | 08:56 | |
| *** edmondsw has joined #openstack-powervm | 10:14 | |
| *** edmondsw has quit IRC | 10:19 | |
| *** thorst has joined #openstack-powervm | 10:54 | |
| *** thorst has quit IRC | 10:58 | |
| efried | miltonm Yup, we've been tracking that. esberglu and I are +1 on it - just waiting for devstack cores to decide it's good to go. | 11:12 |
|---|---|---|
| efried | Think they may be a little gun-shy considering the snafu it caused last time :) | 11:12 |
| efried | and/or busy with the release. | 11:13 |
| efried | and/or everyone's on vacation last week and this. | 11:13 |
| *** openstackgerrit has quit IRC | 11:17 | |
| *** chhavi has quit IRC | 11:45 | |
| *** edmondsw has joined #openstack-powervm | 11:55 | |
| *** thorst has joined #openstack-powervm | 11:57 | |
| *** svenkat has joined #openstack-powervm | 12:18 | |
| *** svenkat has quit IRC | 12:22 | |
| *** chhavi has joined #openstack-powervm | 12:23 | |
| *** svenkat has joined #openstack-powervm | 12:23 | |
| *** AndyWojo has quit IRC | 12:26 | |
| *** AndyWojo has joined #openstack-powervm | 12:27 | |
| *** edmondsw has quit IRC | 12:29 | |
| *** esberglu has joined #openstack-powervm | 13:15 | |
| *** apearson has joined #openstack-powervm | 13:28 | |
| *** edmondsw has joined #openstack-powervm | 13:29 | |
| *** apearson has quit IRC | 13:47 | |
| *** apearson has joined #openstack-powervm | 13:48 | |
| *** edmondsw has quit IRC | 16:05 | |
| *** edmondsw has joined #openstack-powervm | 16:07 | |
| *** edmondsw has quit IRC | 16:11 | |
| *** k0da has quit IRC | 16:20 | |
| *** miltonm has quit IRC | 16:21 | |
| *** miltonm has joined #openstack-powervm | 16:34 | |
| *** apearson_ has joined #openstack-powervm | 16:39 | |
| *** apearson has quit IRC | 16:40 | |
| *** edmondsw has joined #openstack-powervm | 17:39 | |
| *** edmondsw has quit IRC | 17:43 | |
| *** edmondsw has joined #openstack-powervm | 17:52 | |
| *** chhavi has quit IRC | 17:58 | |
| *** mawi has quit IRC | 18:13 | |
| edmondsw | efried what does a drc_name look like? | 19:09 |
| efried | edmondsw Like a location code. | 19:10 |
| efried | edmondsw U5791.001.CEC1202-P02-C027 kind of thing | 19:10 |
| edmondsw | it's a unique id for a slot? | 19:10 |
| efried | yes | 19:10 |
| efried | But so is the DRC index. | 19:10 |
| thorst | efried: is DRC the PCI identifier for us? | 19:10 |
| efried | thorst We're generating a pseudo PCI address from the DRC index. | 19:11 |
| thorst | hmm | 19:11 |
| efried | Was just about to post up that change set. | 19:11 |
| efried | Yes, let's discuss. | 19:11 |
| efried | Cause there's not a good way to do this. | 19:11 |
| efried | And this is the least awful way I've figured. | 19:11 |
| thorst | I'm just not a huge fan of DRC's | 19:14 |
| thorst | we had that in the NL installer and it never worked out there | 19:14 |
| thorst | but I think that was because pvmctl wasn't boot strapped yet | 19:14 |
| efried | thorst edmondsw 5755 | 19:14 |
| efried | The only thing we need it for is to uniquely identify the slot in an environment (like nova) that's expecting a device to have a PCI address in the format of a PCI address. | 19:15 |
| efried | Serendipitous that y'all were asking about this right as I was putting it together - where did the question come from? | 19:17 |
| edmondsw | efried I'm looking at 5749 | 19:20 |
| edmondsw | the exception there says "slot at <drc_name>" and that just sounds odd to someone used to thinking of "slot <index>" | 19:21 |
| efried | edmondsw I think people are a lot more used to looking at DRC names (i.e. location codes) than anything else, when it comes to Power slots/cards. | 19:21 |
| edmondsw | efried probably... POWER isn't my background | 19:22 |
| edmondsw | hence the question | 19:23 |
| edmondsw | left you some more comments on 5749... not quite sold yet | 19:23 |
| efried | They're definitely not interested in looking at PCI addresses, cause there ain't no such thing. And my hope is to keep these pseudo PCI addresses hidden from the user. | 19:24 |
| efried | I debated whether to put the pseudo PCI address generator in nova instead. | 19:24 |
| efried | We should still discuss that. | 19:24 |
| efried | Thing that tips me over into having it in pypowervm is, I'm not sure we won't eventually need to use it in our neutron plugins as well. | 19:25 |
| edmondsw | efried what would be the reason to put it in pypowervm? Do we think something besides nova would find that useful? | 19:25 |
| edmondsw | ic | 19:26 |
| efried | That said, if we do, it may or may not be in the context of this SlotInfo thingy. | 19:26 |
| efried | So I just dunno. | 19:26 |
| edmondsw | efried make it a utility method not part of SlotInfo? | 19:27 |
| efried | right | 19:27 |
| edmondsw | efried or do it on IOSlot... | 19:36 |
| efried | edmondsw Then I'd have to implement it twice, or make a base class. | 19:37 |
| efried | Which is part of the ew of this whole thing. | 19:37 |
| *** k0da has joined #openstack-powervm | 19:41 | |
| edmondsw | esberglu ceilometer stable/pike is out | 19:50 |
| esberglu | Of course the day the workaround gets in... | 19:51 |
| esberglu | I thought that they were waiting to release with the official date? | 19:52 |
| edmondsw | they weren't supposed to be... I think they were confused | 19:52 |
| edmondsw | esberglu seems like our emails on the ML helped move that along | 19:53 |
| efried | I also pestered 'em in -release again earlier today. | 19:54 |
| edmondsw | esberglu you might want to look at the last email in that ML thread for future reference when we're using releases | 19:54 |
| efried | I had been tempted to propose the release change set myself over the weekend. | 19:54 |
| efried | edmondsw For 5755, I deliberately left the explanation out of the code comments. I don't really want to expose that much of our dirty undercarriage to the world. But I'm happy to explain it to you here. | 19:56 |
| efried | 1) Why do we ignore the first byte? seroyer led me to believe it's always the same (I've got test data that seems to belie that, but who knows where the test data came from?) | 19:57 |
| efried | 2) The other pieces are being used as they are because they seem to map to right-sized pieces on the system (again according to seroyer). | 19:58 |
| efried | A "domain" is two bytes and a "bus" is one; and the DRC index is apparently constructed as 0x21aabbbb. So I used the two-byte part for the two-byte part, and the one-byte part for the one-byte part. | 19:58 |
| efried | I didn't use the one-byte part for the slot because the slot of a PCI address is max 0x1F and that 'aa' part of the DRC index can apparently be more than that. | 19:59 |
| edmondsw | efried ok. And why don't we have PCI addresses in POWER? Or maybe we do, but just buried at a layer we don't work at? | 20:00 |
| edmondsw | I assume the latter | 20:00 |
| efried | Finally, the 'func' (which has a max of 0x7) at least begins to relate to something I've dealt with before, which would have to be an SR-IOV physical function - for which we may be able to support passthrough at some point, whereupon we may use that field. | 20:01 |
| efried | edmondsw Right, because PHYP ain't Linux. | 20:01 |
| efried | We have a similar concept, which maps to physlocs (and presumably there's *some* correlation between physlocs and DRC indices). | 20:02 |
| efried | Power has, like, CEC/Drawer/Slot/pport/vport or something. | 20:03 |
| edmondsw | efried should we run this by someone on the PHYP team and see what they think of it? | 20:03 |
| efried | edmondsw Like seroyer? | 20:03 |
| efried | Cause that's where this came from. | 20:03 |
| edmondsw | ok, cool | 20:03 |
| edmondsw | efried seems like if this is a place where nova is requiring a Linux-ism, we could go fix that in nova | 20:05 |
| efried | edmondsw Yes, I would like to do that eventually. | 20:06 |
| edmondsw | efried cool... keeping a list somewhere? | 20:06 |
| efried | Getting rid of it entirely is likely to be a fight. | 20:07 |
| efried | edmondsw At the moment, local chicken-scratch. Was considering chucking it into an etherpad. You interested in reading it? | 20:07 |
| edmondsw | yeah | 20:07 |
| efried | edmondsw https://etherpad.openstack.org/p/powervm-pci-passthrough-notes | 20:09 |
| edmondsw | tx | 20:09 |
| *** apearson_ has quit IRC | 20:27 | |
| *** apearson has joined #openstack-powervm | 20:30 | |
| *** edmondsw_ has joined #openstack-powervm | 20:47 | |
| *** apearson has quit IRC | 20:47 | |
| *** edmondsw has quit IRC | 20:49 | |
| *** apearson has joined #openstack-powervm | 21:00 | |
| *** thorst has quit IRC | 21:04 | |
| *** edmondsw_ has quit IRC | 21:04 | |
| *** k0da has quit IRC | 21:38 | |
| *** k0da has joined #openstack-powervm | 21:41 | |
| *** svenkat has quit IRC | 21:47 | |
| miltonm | efried, part of the mismatch comes from linux only giving 16 bits to pci domain but phyp using 32 and 64 bit identifiers for their api | 22:02 |
| miltonm | that, and phyp only presents an abstract machine to any partition | 22:03 |
| *** apearson has quit IRC | 22:26 | |
| *** thorst has joined #openstack-powervm | 22:35 | |
| *** thorst has quit IRC | 22:53 | |
| efried | miltonm Good to know. | 23:15 |
| *** thorst has joined #openstack-powervm | 23:33 | |
| *** thorst has quit IRC | 23:33 | |
| *** k0da has quit IRC | 23:34 | |
| *** thorst has joined #openstack-powervm | 23:59 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!