From: Jouni Malinen (jkmaline_at_cc.hut.fi)
Date: 2002-04-03 12:15:34 UTC
On Mon, Apr 01, 2002 at 01:49:59PM -0800, Adam Megacz wrote:
> Hey, can anybody tell me what this means? I'm using that latest driver
> from http://people.ssh.com/jkm/Prism2/ on a Debian (kernel 2.2.18)
> system with an Addtron card (Prism2 chipset) in a PCI adapter stolen
> from a D-LINK DWL-650. Yes, I know this is sort of a Frankenstein
> setup, and I probably shouldn't be doing this. The client is an Apple
> Powerbook Titanium G4 using the integrated adapter.
Well, assuming it is working, I se no reason in not using such a setup ;-). At least this issue you are asking has nothing to do with the hardware on AP.
> wlan0: dropped received packet from 00:30:65:03:6e:f1 with no ToDS flag (type=0x02, subtype=0x04)
> wlan0: RX status=0x0000 (port=0, type=0, fcserr=0) silence=0 signal=0 rate=20 rxflow=0
> FC type=2:4 dur=0x0102 seq=0x1500 data_len=0
These are nullfunc frames (data frames, with no data). However, I'm not fully sure, why some cards send these without ToDS flag. At least Lucent firmware seems to generate these every now and then. Previously, the Host AP driver just silently dropped these frames in any case if ToDS flag was missing. Today I modified the driver to reject these frames since it apparently helped speed up re-association of a station using Lucent WaveLAN when the AP was restarted (i.e., it lost all information about the station).
Nullfunc frames can be used among other things to indicate the station is using power saving mode (or to indicate that it ends using PS mode). They are also used with CF-polling. However, I haven't found any description of how they are used without ToDS flag, but sent to the AP anyway. IEEE 802.11 does not seem to describe this behavior (or at least it is hidden somewhere, if it is described ;-). If anyone on the list happens to have a pointer to proper documentation for these, please let me know.
Your log messages indicated that all these nullfunc frames were received with zero signal level.. I have a proper signal level also on these frames (at least with Compaq WL100). Other than that, these messages about dropped nullfunc frames (type=0x02, subtype=0x04) can be ignored quite safely. I might skip that message with nullfunc frames completely in the future versions of the driver.
-- Jouni Malinen PGP id EFC895FA