Hi All,
Summarising, for future reference...
I received some six responses. Overall the feedback was a little
disappointing. Three responses suggested that it would be easier/less
time consuming/more stable to simply connect a consumer access point
device via Ethernet. Of course, I wouldn't learn as much by doing this
:-(. The background to this seems to be mostly issues with the
configuration and stability of drivers e.g. ath and ral.
At least a couple of the respondents are successfully using ALIX boards,
including the desired 2D13. None of the responses related to the
specific wireless devices that I asked about. Some of those mentioned as
having been used included the AR5212 and AR5413 (with ath) and the
RT2561C (ral).
A couple of responses indicated that OpenBSD doesn't support 802.11n. I
got my initial information from the athn manual page. It begins:
> ...
> NAME
> athn - Atheros IEEE 802.11a/g/n wireless network device
> ...
> The athn driver provides support for a wide variety of
> Atheros 802.11n devices ...
Which I incorrectly took to mean that "n" networking was supported...
However, further down in the same man page, under caveats, it states:
> ...
> The athn driver does not support any of the 802.11n capabilities
> offered by the adapters. Additional work is required in
> ieee80211(9) before those features can be supported.
> ...
That should teach me (yet again) to read the whole man page :-)
Cato Auestad provided a very helpful link to a description of his
working (ral based) OpenBSD configuration:
http://bleakgadfly.com/notes/openbsd_wifi.html
There he mentions that support from the hostap daemon - hostapd(8) - is
also necessary for such a configuration. Something else that I hadn't
realised.
So, based on the feedback, it looks like while this might be a fun
project, it could be hard to create a stable "production" access point.
Thanks for all the info.
