Peter,

Do you know at which revision to sys/dev/ic/wi.c and crew the problems
began?

Eventually I will try to reproduce this bug and fix it, but I offer
these debugging suggestions in the mean time.

If you add "options WI_DEBUG" and "options IEEE80211_DEBUG" to your
kernel, then you can turn on debug messages by writing 1 (or 2, for more
messages) to each of wi_debug and ieee80211_debug. I use the 'write'
command in ddb to do this. During the pauses, abnormal behavior might
stand out in the debug messages.

Try to find out what is happening on the air, before, during, and after
the pauses.  If you can, monitor the pausing host from a different
host using, e.g., tcpdump -ne -D ieee802_11 -i wi0 . This might provide
some clues.

Dave

On Sun, Nov 03, 2002 at 03:12:25AM -0600, Peter Seebach wrote:
> In message <mtufzujh2na.fsf@contents-vnder-pressvre.mit.edu>, "Nathan J. Willia
> ms" writes:
> >So, I've been reading this thread with some curiosity, because I've
> >been using my wi0 interface on an iBook all along, tracking -current,
> >and haven't seen this problem. I'm using a D-Link DWL1000 access point.
> 
> What weirds me out is that it's quite repeatable on two totally
> different wi0 cards, but only when using an access point - and
> wasn't true, say, two months ago.
> 
> -s

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Engineering from the Right Brain
                        Urbana, IL * (217) 278-3933