Re: Passive Scanning Bugreport

On Sat, 2006-07-29 at 00:38 +0200, Thomas King wrote:
> Hi all,
> unfortunately, I figured out that my passive scanning implementation does not 
> work in all circumstances. 

This is caused by the timing. The default scan watchdog timer is (5 *
HZ). In your implementation, you set the passive scan dwell time to
200ms. For example, a card in SKU "ZZA" has 11 bg channels and 13 a
channels, totally 24 channels. To finish scanning all the channels, it
will take 24 * 200ms = 4800ms ~= 5 HZ. So the watchdog is likely to be
kicked off. If you set the dwell time to 120ms as it was originally, the
watchdog shouldn't be fired off. Or if you'd insist 200ms dwell time
gains better result, set the watchdog timer to something like 8~10 HZ.

> As far as I understand how the driver works, I believe the firmware is not 
> generating a scan completed event.

The firmware will generate a scan complete even either the scan is
really completed or the scan is aborted. So in the watchdog kicking off
case, you will still receive a scan complete event. See below,

watchdog kicks off
-> adapter_restart
   -> ipw_down
      -> ipw_deniit
         -> ipw_abort_scan
         -> wait up to 10ms for the SCANNING status, etc. changed

If you see the "Took xxms to de-init" debug message, you have already
received the scan completed event. I can always receive the event here
in my test environment, if you cannot get the event, maybe you want to
increase the wait time (10ms) a little bit longer.

Thanks,
-yi

-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys -- and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV