ipw2100 + WPA-PSK link dies if not used
Sat Dec 25 14:34:07 PST 2004
Jouni Malinen wrote:
> Have you configured the AP to do Group Key handshakes (broadcast
> rekeying, or something similar in configuration) frequently? If not,
> there should not be wpa_supplicant-related events happening that often
> and this would be something that would be better handled by ipw2100
> developers/mailing lists.
I'm not a WLAN expert - this is actually my first hands-on contat
with WLANs (though I know pretty much about other kinds of networks).
These are all relevant tunable parameters in Performance tab of my AP
Beacon interval: 100 (msec, range:1~1000, default:100)
RTS Threshold : 2432 (range: 256~2432, default:2432)
Fragmentation : 2346 (range: 256~2346, default:2346, even number only)
DTIM interval : 3 (range: 1~255, default:3)
I did not really bother to figure out what they really mean and
relied on the defaults. I only concentrated on the thing that I'm more
familliar of which is security and I guess that WPA-PSK is the best
possible solution for my little home network (although static WEP keys
would probably be easier to set up in Linux). There are a few other WLAN
networks in my range and I don't want anyone uninvited to get into my
> Which DI-514 revision (h/w ver) is that (A/B/C) and which firmware
It's revision B with firmware 1.04 (I upgraded it right after
purchase). There's not much to tweak with the WLAN part of this device
as it's also a router and most settings regard the router part, which I
don't use anyway. It was simply the cheapest D-Link available at my
dealer's and that's why I bought it. Speed is of no concern as long as
it's faster than my ISDN Internet connection connected through an old
good Cisco 1603 to my AP.
Those pings really help - I've been connected for more than 2 hours
now which would otherwise be impossible, unless I rebooted to Window$
(where of course it works perfectly).
I think of adjusting the ping interval in order to get more clues.
One second works for sure and one minute is obviously not enough (though
I'm not 100% sure - would have to check it again to make sure).
If this can be of any help to wpa_supplicant team, I can launch it in
the foreground again with more verbose debugging. Unless there's another
trivial cause of my problem and it's not related to wpa_supplicant at all...
szymon at juraszczyk.com
More information about the Hostap