Please note: This project is no longer active.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 #1139 (new defect)

Opened 15 years ago

Last modified 15 years ago

client reassoc causes master kernel panic with WPA, bridging

Reported by: msmith@cbnco.com Assigned to:
Priority: major Milestone:
Component: madwifi: other Version: trunk
Keywords: Cc: msmith@cbnco.com
Patch is attached: 0 Pending:

Description

Hi,

I'm using madwifi r2070 and kernel 2.6.18.3. The master runs hostapd 0.5.7, the client runs wpa_supplicant 0.5.7. They do WPA2-PSK with CCMP. Four-address headers are turned on at both ends. On the client, ath0 and eth0 are bridged together. (No bridging on the master.)

I can trigger a couple of different kernel BUGs and panics on the master by forcing a disassociation and reassociation on the client (e.g. by "wpa_cli reassoc"), and then pinging through the master to a machine on the client's eth0. I'll attach four stack traces:

  1. failed at madwifi/net80211/ieee80211_power.c:184
  2. failed at mm/slab.c:595
  3. failed at mm/slab.c:595 again
  4. failed at madwifi/net80211/ieee80211_power.c:184 again

In the fourth stack trace, right before the crash it complains:

ath_rate_sample: no rates for 00:04:23:94:65:44?

That MAC address isn't anywhere on the network that I know of, so I think it is confused.

I don't think I saw this problem with r2002. I'm more sure I didn't see it with r1931.

Thanks, Mike

Attachments

bug-on-master.txt (2.2 kB) - added by msmith@cbnco.com on 02/06/07 21:07:21.
stack trace from madwifi/net80211/ieee80211_power.c:184
bug2-on-master.txt (3.8 kB) - added by msmith@cbnco.com on 02/06/07 21:08:24.
stack trace from mm/slab.c:595
bug3-on-master.txt (4.0 kB) - added by msmith@cbnco.com on 02/06/07 21:09:11.
slightly different stack trace from mm/slab.c:595
bug4-on-master.txt (3.9 kB) - added by msmith@cbnco.com on 02/06/07 21:09:47.
another trace from madwifi/net80211/ieee80211_power.c:184

Change History

02/06/07 21:07:21 changed by msmith@cbnco.com

  • attachment bug-on-master.txt added.

stack trace from madwifi/net80211/ieee80211_power.c:184

02/06/07 21:08:24 changed by msmith@cbnco.com

  • attachment bug2-on-master.txt added.

stack trace from mm/slab.c:595

02/06/07 21:09:11 changed by msmith@cbnco.com

  • attachment bug3-on-master.txt added.

slightly different stack trace from mm/slab.c:595

02/06/07 21:09:47 changed by msmith@cbnco.com

  • attachment bug4-on-master.txt added.

another trace from madwifi/net80211/ieee80211_power.c:184

02/06/07 22:35:11 changed by kelmo

PRelated to #162 ?

02/07/07 14:55:18 changed by xmxwx@asn.pl

Mike, can you tell whether the same problem occurs without four-address headers (iwpriv ath* wds 0)?

02/07/07 15:33:10 changed by xmxwx@asn.pl

If so, then it could be related to #575

02/13/07 15:07:15 changed by msmith@cbnco.com

Sorry, I've been away. If I turn off four-address headers, I think the bridging setup (ath0 bridged to eth0 on the client) won't work.

I didn't have the problem with bridging turned off on the client, even with four-address headers turned on. I also only have the problem with WPA enabled. I'll double check later today, but this is what I remember.

02/14/07 19:03:50 changed by msmith@cbnco.com

Thanks Michal. Your patch from #575 fixes my problem, so I think it is related. Maybe someone can close this bug as a dupe of #575.