Re: iwn not working since 4.99.73

[Robert Elz <kre%munnari.OZ.AU@localhost>]

    Date:        Fri, 7 Nov 2008 10:30:29 +1030
    From:        Brett Lymn <blymn%baesystems.com.au@localhost>
    Message-ID:  <20081107000029.GD18608%baea.com.au@localhost>

  | I knew the PR existed, thanks for logging one.  I will close it off
  | when I have a chance.

Thanks.

  | I did consider that but I am thinking that the ieee802.11 layer needs
  | to be told that the scan has stopped too.

It looked to me as if "scan stopped" is a more generic event that seems
to be handled via the notify_intr() routine (that's where is_scanning
gets cleared), so I was just assuming that anything relating to the
scan being finished would be handled that way.   The SCAN_ABORT isn't
the only thing that's going to cause a scan to end (but of course, I
really have no idea whether an explicit abort would result in an
interrupt notification or not.)

  | At the time I was more
  | interested in getting people back up and running.

Yes, I understood that.   Only after it was fixed did I even contemplate
trying to have the best of both worlds...

  | Unfortunately, it didn't have any positive effect for me but busted
  | people who were not doing scans for APs.

Ah, in that case, just removing it sounds like the right solution.
You might just want to try the conditional version with WPA though
(I have no WPA type access points, so I cannot) just in case you ended
up being bitten by a variant of the same problem, aborting the scan when
it is in progress might be the right thing to do, but only if it is
in progress, if already aborted it might be just as bad for the firmware
as in the "never scan" case I am seeing.

(Oh, I can test WPA in the "wep or plain only" type config, no PSK or
anything fancy like that, if that will help - in fact, I'd rather like
to set this system up to operate that way in the long term.)

kre