On 8/12/16 1:10 , Ian Collins wrote: > On 08/12/16 07:43 PM, Kilian Ries wrote: >> Hi Ian, >> are you running the nics standalone or in bonding mode? >> I had the same or a similar problem a while ago: >> >> https://www.mail-archive.com/[email protected]/msg02842.html >> >> Unfortunately we didn't find a solution for the problem ... >> > > It's just a single NIC which (I assume) is directly connected to a > switch. There's an identical box next to it in the rack running hyperv > which works fine. >
I think these are likely related problems, which is getting close to what Garrett said. One of the things we should probably look at is updating the 5719/5720 to leverage the newer virtualization bits instead of the legacy format that we're using. Unfortunately we don't really have these handy, so we're not in a great position to develop against this right now. I'm not sure how straightforward a change that'd end up being. Killian, sorry I had kind of dropped the ball on this. There is probably a bit more debugging we can do on the driver here that'll require a bit of custom stuff to go better dig into. Unfortunately I can't really guarantee much of a prod-safe way on those next steps. Ian, Killian, are any of these in a place where we could maybe put together something which dumps some of the register space and you might be able to run it? I can try putting together a one off module / command or an updated platform / ioctl. Though again, testing this stuff is a bit hard for me to do given hw. Robert ------------------------------------------- smartos-discuss Archives: https://www.listbox.com/member/archive/184463/=now RSS Feed: https://www.listbox.com/member/archive/rss/184463/25769125-55cfbc00 Modify Your Subscription: https://www.listbox.com/member/?member_id=25769125&id_secret=25769125-7688e9fb Powered by Listbox: http://www.listbox.com
