[rt2x00-users] rmmod rt2800pci broken since compat-wireless-2011-04-21 - CPU load is 1

[Gertjan van Wingerde]

Hi Andreas,

On 05/08/11 06:35, Andreas Hartmann wrote:
> Hello Gertjan,
> 

I see that the rt2x00 users mailing list was stripped from our email conversation.
Please keep it in as there may be other people on the list that are able to help
out and debug this.

> meanwhile, I did some more testing. I tried the same problem with my
> atheros AR9285 as hostapd. The problem is the same - rekeying of PTK
> doesn't work. Seems to be a problem of hostapd.
> 

OK. You better take that issue to the linux-wireless mailing list then, as that
particular issue seems to be more related to hostapd. Please include the hostapd
trace you provided in a previous message to illustrate the issue.

> What do you think about this?
> 
> 
> Another test with rt2800pci showed this error in messages:
> 
> May  8 06:08:36 dualc kernel: [ 2268.969512] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:08:36 dualc kernel: [ 2268.973201] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:08:36 dualc kernel: [ 2268.973301] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:08:36 dualc kernel: [ 2268.977310] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:08:36 dualc kernel: [ 2268.977442] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.818481] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.822332] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.822460] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.870765] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.870857] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.874714] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.874811] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.881911] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:01 dualc kernel: [ 2293.968070] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> May  8 06:09:02 dualc kernel: [ 2294.931837] phy0 -> rt2800pci_txdone:
> Warning - Got TX status for an empty queue 2, dropping
> 

OK. These seem to be new instances of the race condition I talked about.
I'm auditing the queue management code right now, but nothing did stand
out yet. I'll continue to look at this, but hopefully others on the
mailing list can help in this effort as well, as my time is fairly
limited, especially on weekdays.

> 
> It came up during association:
> 
> 1304827833.633061: association request: STA=48:5d:60:3e:a3:18
> capab_info=0x431 listen_interval=10
> 1304827833.633078: WMM IE - hexdump(len=7): 00 50 f2 02 00 01 00
> 1304827833.633091: Validating WMM IE: OUI 00:50:f2  OUI type 2  OUI
> sub-type 0  version 1  QoS info 0x0
> 1304827833.633114:   new AID 1
> 1304827833.633123: HT: STA 48:5d:60:3e:a3:18 HT Capabilities Info: 0x11ce
> 1304827833.633133: update_sta_ht STA 48:5d:60:3e:a3:18 - no greenfield,
> num of non-gf stations 1
> 1304827833.633144: hostapd_ht_operation_update current operation mode=0x0
> 1304827833.633165: hostapd_ht_operation_update new operation mode=0x4
> changes=1
> 1304827833.633268: nl80211: Set beacon (beacon_set=1)
> 1304827833.637366: wlan1: STA 48:5d:60:3e:a3:18 IEEE 802.11: association
> OK (aid 1)
> 1304827836.649574: Data/PS-poll frame from not associated STA
> 48:5d:60:3e:a3:18
> 1304827836.649618: Association Response to the STA has already been
> sent, but no TX status yet known - ignore Class 3 frame issue with
> 48:5d:60:3e:a3:18
> 1304827843.655563: mgmt::disassoc
> 1304827843.655600: disassocation: STA=48:5d:60:3e:a3:18 reason_code=3
> 1304827843.655627: wlan1: AP-STA-DISCONNECTED 48:5d:60:3e:a3:18
> 

---
Gertjan