RE: Packet length problems w/ WEP

From: Martin Whitlock (
Date: 2002-09-18 07:11:47 UTC

> The dmesg output you sent is not produced by 2002-09-12. Instead, it is
> from a somewhat older CVS snapshot.

Oops... I should have noticed this... Don't ask me how, but when I imported the 2002-09-12 release into my own cvs I somehow managed to update only _some_ of the files in the project. As a result, I ended up with half cvs snapshot and half 2002-09-12 release. Apparently this could do some nasty things with the driver. Funny though that I didn't get any compilation errors/warnings... Anyway, after doing a new import everything works great. Sorry to bother you all with this. (Btw, I hate cvs... ;-)

For interested I have answered Jouni's questions.


> Do you mean that the system crashed/rebooted immediately after getting
> that message? The dmesg log did show 'AP queue full' messages, but there
> was no reboot..

Sorry... Wasn't clear about that. I got like a hundred of those messages, and then a reboot without a warning.

> > wlan0: authentication: 00:05:4e:81:02:94 len=6, auth_alg=0,
> > auth_transaction=1,
> > status_code=0, fc=0x00b0
> > wlan0: AP trying to authenticate?
> Do you have other_ap_policy set to something else than 0? This message
> would indicate that the station 00:05:4e:81:02:94 was first noticed from
> a beacon frame and then it tried to authenticate. This could happen,
> e.g., with other_ap_policy=3 and station acting first in IBSS and then
> trying to associate with the AP.

Correct. other_ap_policy=2

> > wlan0: WEP decryption failed (SA=00:05:4e:81:01:fd)
> Is this related to the ping packet not getting through? What was the
> hwaddr of the device sending those packets? What card is in the station?
> What about the driver? Does the same test work with another AP?

This is definately related to the ping. The address belongs to the "pinging" client.


This archive was generated by hypermail 2.1.4.