Please note: This project is no longer active. The website is kept online for historic purposes only.
If you´re looking for a Linux driver for your Atheros WLAN device, you should continue here .

Ticket #1957 (new defect)

Opened 14 years ago

Last modified 13 years ago

Beacon miss and associated client disconnect

Reported by: petar@smokva.net Assigned to:
Priority: major Milestone:
Component: madwifi: driver Version: trunk
Keywords: beacon miss disconnect Cc:
Patch is attached: 0 Pending:

Description

I'm sorry for this duplicate but since the milestone of #848 has been reached and the patch there can't be applied any more, I had no choice than trying again.

This problem is reproducible -- 1 beacon miss, 1 disconnect -- and it happens only, when the client has to transmit lots of data.. something like `dd if=/dev/zero bs=1M | nc AccessPoint? port'.

This bug makes my ath card almost unusable for large file transfers (daily backups), so I'll be very happy to assist.

Kernel: 2.6.24

Madwifi: trunk (changeset 3677)

Hardware: IBM T42 with a Winstron CM9

Access Point: FreeBSD 7.0

BTW: FreeBSD 6.2 on the same device does not show any similar behaviour.


athstats on the T42:

$ athstats
4 beacon miss interrupts
235 tx management frames
3400 long on-chip tx retries
203 tx frames with no ack marked
59438 tx frames with short preamble
292 tx frames with an alternate rate
88 rx failed due to bad CRC
1 PHY errors
    1 CCK restart
63 periodic calibrations
1 rfgain value change
rssi of last ack: 43
rssi of last rcv: 42
6 switched default/rx antenna
Antenna profile:
[1] tx     3761 rx      416
[2] tx    55496 rx    57822

Change History

05/29/08 21:32:24 changed by petar@smokva.net

I forgot to mention that sometimes my T42 freezes when I try to transmit a large file.. and while being frozen, the caps-lock led blinks continuously.

06/03/08 19:58:27 changed by petar@smokva.net

The following command did the trick:

# iwpriv ath0 bmiss 50
# iwpriv --all
(...)
ath0      get_bmiss_ms:5120  
ath0      get_bmiss:50
(...)

I'm not sure about the values so if someone knows better thresholds, please let me know.

Another question would be if this is a bug or not?

P.S. The freezes I mentioned in my last post had nothing to do with the madwifi driver. It was a kernel oops caused by the bash. Sorry for the noise.

07/02/08 17:30:15 changed by lyingmonkey@gmx.de

I am experiencing something similar. The network consists of multiple APs, non-encrypted with a VPN on top of it. When a beacon is missed, a disconnect, and reauthentication follows. The state changes are beacon miss triggers the state transitions run -> scan -> auth -> run. All this takes only a couple of sec. and the VPN connection stays intact.

But whenever a beacon is received while authenticating, there is an additional auth -> scan transition and and a new scan is triggered. Beacon miss triggers the state transitions run -> scan -> auth (here a beacon is received)-> scan ... So the reathentication is delayed by several seconds and the VPN connection is gone.

Is there anything I can do about it?

(I am using a T40p with a 5211 chip)

07/06/08 21:29:09 changed by petar@smokva.net

I just checked out the latest revision (3756) and apparently the problem is gone. But that does also apply to 1/3 of my max. wireless performance.. while I could reach a stable 3.5 MB/s with 0.9.4.3382 (up and down) now it's only somewhere between 1.5 and 2 MB/s.

07/08/08 18:49:16 changed by mentor

  • summary changed from beacon miss causes client to reassociate to Beacon miss and associated client disconnect.