Re: Suggestions for miniPCI wireless card for an accesspoint on OpenBSD - 2022q4
On 2022-11-24, Mikolaj Kucharski wrote: > Hi, > > I'm using for few years now on OpenBSD accesspoint (mediaopt hostap) > based on following miniPCI card: > > # dmesg | grep -e ^ath > athn0 at pci4 dev 0 function 0 "Atheros AR928X" rev 0x01: apic 5 int 16 > athn0: AR9280 rev 2 (2T2R), ROM rev 22, address 04:f0:21:45:6a:c4 > > I don't remember where I bought it, but I think it is one of those, or > compatibile: > > https://www.pcengines.ch/wle200nx.htm > > If you would build today an accesspoint, on hardware with miniPCI, what > would you choose, for OpenBSD? You probably won't find anything really better than that for hostap. Most recent wlan development on OpenBSD has been for client mode only. If you want something standalone (non OpenBSD) take a look at tp-link's omada range. Their java+mongo management software (unifi clone) won't run on OpenBSD but the APs have a web interface. -- Please keep replies on the mailing list.
Re: PC Engines APU alternative for OpenBSD - 2022h2
Hello, > > Seeing recent issues with buggy BIOSes I wanted to avoid mini pc hunting > on Aliexpress :/ > I would be happy to have more choices to hunt for. But there aren't many. Qotom mini PCs are not bad. I bought one 5 years ago and it is still OK. I ordered another one as a spare recently and wait for it to arrive: Qotom Mini PC 5* I225-V 2.5G Lan Celeron J4105 AES-NI Quad core Pfsense Firewall Router Mini PC Q730G5 -- Regards Maksim Rodin
ifconfig vs lladdr
Maybe I'm stupid, but the way we proceeded with disks was with making euids integral part of mount(8). Why not do the same with lladdr and ifconfig (and lower parts) (possibly with an option, though 60a44ce81465 is rather recognizable as a MAC) ?
Re: ifconfig vs lladdr
Marc Espie wrote: > Maybe I'm stupid, but the way we proceeded with disks was with > making euids integral part of mount(8). > > Why not do the same with lladdr and ifconfig (and lower parts) > (possibly with an option, though 60a44ce81465 is rather recognizable > as a MAC) ? lower parts? there are a lot of 'lower parts' Do you mean route, bgpd, ospfd, iked, pfctl, pf.conf, unbound, nfsd, and 150+ other places in the tree that currently take a 'if#' ? So you mean allowing all of them to use a non-interface name, but also some "new identifier", probably a MAC, probably not anything else because that would either requires 'storage' or intentiaonl assignment. The problem is smaller, because most of those places already take either an if# or an assigned IP address, so the main problem is the initial assignment of an IP address, after which IP identification is enough. To put it simply, the problem is only the initial assignment, which is what the idea being assigned is trying to do.
