| jgilaber | sean-k-mooney, melwitt can you take a look at the nova patches related to mdevs https://review.opendev.org/q/topic:%22cyborg-vgpu-support%22 when you have some time? thanks | 12:44 |
|---|---|---|
| jgilaber | I'd like to ensure they are in a good place before leaving on PTO | 12:45 |
| jgilaber | I tested them end to end in https://review.opendev.org/c/openstack/cyborg/+/997364 | 12:46 |
| sean-k-mooney | jgilaber: sure ill try and take a look at them this week | 13:09 |
| melwitt | jgilaber: sure, will do | 13:58 |
| jgilaber | thanks! | 14:04 |
| opendevreview | Gihong Lee proposed openstack/cyborg master: Migrate accelerator drivers to shared lspci helpers https://review.opendev.org/c/openstack/cyborg/+/998817 | 14:50 |
| melwitt | sean-k-mooney: fyi this is the KNOWN_NICS patch https://review.opendev.org/c/openstack/cyborg/+/997782 that came out of the smartnic spike https://gist.github.com/melwitt/bb7979d9b1e06cba4a0702e90b4211ab | 16:49 |
| sean-k-mooney | yes i saw it brifely last week | 16:49 |
| melwitt | ah ok cool | 16:50 |
| sean-k-mooney | long term i think we shoudl look to deprecat that driver and the nic driver type in general | 16:50 |
| sean-k-mooney | but for now that is cool | 16:50 |
| sean-k-mooney | were you able to test this end to end? | 16:50 |
| melwitt | yeah I was thinking that too and briefly discussed during the last cyborg meeting | 16:51 |
| melwitt | yes the end to end test worked | 16:51 |
| melwitt | (it's in that github gist) | 16:51 |
| sean-k-mooney | wiht neutron or just with reqing it via a device profile | 16:51 |
| sean-k-mooney | so there are 2 wasy to consume this right now, you can create a neutron port of type acclerator-direct | 16:52 |
| melwitt | with neutron | 16:52 |
| sean-k-mooney | and the sriov nic agent shoudl be able to manage it and you shoudl be able to use the vf via a neutron network | 16:52 |
| sean-k-mooney | ok cool | 16:52 |
| sean-k-mooney | it shoudl also work via the flavor | 16:52 |
| melwitt | there is a minimalistic patch I could do to duplicate the logic from the pci driver (I think?) to use a whitelist conf instead of the KNOWN_NICS list in the shorter term since removing the nic driver would be a larger change | 16:53 |
| sean-k-mooney | melwitt: so looking at https://review.opendev.org/c/openstack/cyborg/+/997782 i think its ok | 16:53 |
| sean-k-mooney | it may be nice to ether add docs to this patch or to have a seperate followup | 16:53 |
| sean-k-mooney | melwitt: well so if you jsut used teh whitelist tha woudl work with the pci driver today without code changes | 16:54 |
| sean-k-mooney | the biter question is will that still work with neutron | 16:54 |
| sean-k-mooney | the intel driver in theory allwos use to encode info about the pshenet as traits on the RP | 16:55 |
| sean-k-mooney | i think you can do that with the generic driver manually | 16:55 |
| sean-k-mooney | the larger question i had was does neutron or nova care | 16:55 |
| sean-k-mooney | i.e. is anythign today enfocing the VF and phsnet match end to end | 16:55 |
| melwitt | sean-k-mooney: you mean like "how to" kind of docs? | 16:56 |
| melwitt | I can add that, just wasn't sure if I should or if the support matrix is all that was wanted | 16:57 |
| sean-k-mooney | yes a how to use the intel driver to provide cyborg manged neutron ports | 16:57 |
| sean-k-mooney | and if it worked with the generic drive we coudl add that later | 16:57 |
| sean-k-mooney | i assume the generic pci driver will fail in nova | 16:58 |
| sean-k-mooney | becuase in teh nic driver we encode the physnet info in the attachmetn handel https://github.com/openstack/cyborg/blob/master/cyborg/accelerator/drivers/nic/intel/sysinfo.py#L236-L243 | 16:58 |
| melwitt | hm ok, I can give that a try | 16:58 |
| sean-k-mooney | so assume that is checked somewhere | 16:59 |
| sean-k-mooney | but maybe it is not | 16:59 |
| melwitt | gotcha | 16:59 |
| sean-k-mooney | melwitt: basiclly what i want to confirm is if we have 2 nic ports on phsynet A and B | 17:00 |
| sean-k-mooney | do we actully passh the phsyent info end to end and enforce it | 17:00 |
| melwitt | yeah, last meeting we were chatting about the idea of "moving" the nic driver stuff into the pci driver and just have a little bit of "if device type" to handle them both | 17:00 |
| sean-k-mooney | or was that missed when this was orgianlly devleoped | 17:00 |
| sean-k-mooney | melwitt: ya that could work | 17:00 |
| sean-k-mooney | or like in nova we coudl add teh physical_network tag to the whitelist config option | 17:01 |
| melwitt | ok I can look into that with the physnet | 17:01 |
| sean-k-mooney | cool for now i think we can proceed with your small patch | 17:01 |
| sean-k-mooney | adn then we can followup with docs and more testing | 17:01 |
| sean-k-mooney | melwitt: does that work for you? | 17:02 |
| melwitt | cool sounds good | 17:02 |
| sean-k-mooney | melwitt: jgilaber incase this passed ye by this is an interisting patch https://review.opendev.org/c/openstack/cyborg/+/998726 | 17:09 |
| sean-k-mooney | i was going to look at this as part fo large rework next cycle | 17:09 |
| sean-k-mooney | but even without that the cleanup form conslitdating woudl be nice to have | 17:09 |
| sean-k-mooney | my incliation is to allow that to prpoceed with a fairly narrow scope this cycle and build on it next cycle to intoduce a PCIBusManager with real objects instead of just strings | 17:10 |
| melwitt | oh yeah, that's cool | 17:10 |
| melwitt | makes sense | 17:11 |
| sean-k-mooney | amoralej: so while the spec was proposed in yoga the keyston bootstrap command only created the manager role form bobcat | 17:27 |
| sean-k-mooney | https://github.com/openstack/keystone/commit/bbbbd10e482a67b53187a93daa58f8bee770852f | 17:27 |
| sean-k-mooney | amoralej: so ya htat why it works upstream and fails with keystone form antelope | 17:27 |
| opendevreview | Merged openstack/cyborg master: Add Intel XXV710 25GbE backplane NIC to known NIC list https://review.opendev.org/c/openstack/cyborg/+/997782 | 18:10 |
| jgilaber | sean-k-mooney, thanks I saw the patch and added it to my review list, but have not really looked much | 18:19 |
| sean-k-mooney | jgilaber: no worries i am more higlighing it in case it impact your other work and jsut incase you found it interesting | 18:42 |
| sean-k-mooney | its proably more impactfult ot chandan actully | 18:42 |
| jgilaber | yes I don't think it'll conflict with my patches | 18:43 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!