16:20:42 <nomad> Is the Intel E830 supported? Specifically the E830-CQDA2  dual-QSFP100G card?
16:21:25 <jbk> not yet
16:21:31 <jbk> i'm actively working on it
16:22:00 <jbk> however, the NIC is _insanely_ complicated
16:22:40 <jbk> it's a NIC, a switch, a CPU, and a floor wax
16:25:38 <nomad> jbk: So... long story short, the person who swore he was selling us a supported card... well. Anyway, can I be a pest and ask how long you think it'll be before it is supported at a production level?
16:26:44 <jbk> no idea.. i'm hoping it'll soon be to the point where it can actually start passing traffic to be able to really kick the tires
16:28:53 <nomad> I've got a host that I haven't put into production yet because I want to get the network issues resolved before making it so that I can't. Would it be helpful to you to have me run some tests?
16:29:57 <jbk> it'd certainly be nice to get as much testing as possible..
16:32:06 <nomad> I should warn you, this is the first time I've dealt with a QSFP so I'm not sure my Juniper switch is properly configured. (I had planned to spend some quality time with their support people after I got the card installed. So much for that plan for now.)
16:32:31 <nomad> but I'm happy to install any T-patch you want and see what happens. You'll probably need to give me very specific and clear instructions :)
16:43:24 <jbk> though that does remind me... i've not received any reviews on my i40e fixes
17:20:06 <danmcd> Sorry.  I haven't had the ability to test it in anger yet (and I know *just* the way to do it), but it's Triton release week @jbk.
17:23:29 <jbk> np, it's been like 2-3 weeks w/o anyone biting, so figure it's probably time to ask again...
17:29:33 <sommerfeld> jbk: noticed a couple things in a first look..
17:37:47 <sommerfeld> jbk: if you really want functional overload, take a look at the intel IPU's (mt evans/E2100 I think).  
17:39:08 <jbk> the problem is, even if you don't care about having a full blown network switch, or a CPU where you can run LLDP and all sorts of other things, you still have to manage those aspects
17:39:31 <jbk> (e.g. multicast and promiscuous mode require programming the switch's NIC filters on the 'port')
17:40:01 <LXGHTNXNG> a floor wax?
17:40:03 <jbk> and it also suffers from Intel's rather unfortunate (to put it kindly) DMA limitations for LSO
17:41:01 <jbk> which are more poorly documented in each successive model NIC
17:42:56 <sommerfeld> "let's throw every single functional unit that might 'help' virtualize a system onto the NIC.  compression! encryption!  NVME target mode so we can make your globe-spanning cloud storage stack look like local flash!"
17:44:25 <LXGHTNXNG> <joke>let's make illumos run on the card!</joke valence="completely unserious">
17:45:06 <jbk> (those DMA limitations are the reason for that i40e changes btw)
17:48:22 <jbk> LXGHTNXNG: it's a reference to an old Saturday Night Live sketch which was a TV commercial spoof for a product that was both a dessert topping and a floor wax
17:51:36 <LXGHTNXNG> thanks, now I don't want to drink my teea.
18:32:31 * nomad feels old
18:54:02 <LXGHTNXNG> same
20:09:01 <gitomat> [illumos-gate] 18097 sys/pci.h: add few missing PCI power management 1.2 constants -- Toomas Soome <tsoome⊙mc>