Debian testing wpa_supplicant 1.03 cannot roam between wifi APs without me killing wpa_supplicant (intel 6300 AGN)

[Marc MERLIN]

On Thu, Feb 14, 2013 at 11:28:21PM -0800, Marc MERLIN wrote:
> [resent with correct Email]
> 
> Howdy
> 
> First, the basics:
> 
> kernel: 3.7.4-amd64-preempt-20121226-fixwd
> debian-testing
> wpasupplicant-1.0-3+b2
> 
> iwlwifi 0000:03:00.0: loaded firmware version 9.221.4.1 build 25532
> iwldvm: Intel(R) Wireless WiFi Link AGN driver for Linux, in-tree:d
> iwldvm: Copyright(c) 2003-2012 Intel Corporation
> iwlwifi 0000:03:00.0: CONFIG_IWLWIFI_DEBUG enabled
> iwlwifi 0000:03:00.0: CONFIG_IWLWIFI_DEBUGFS disabled
> iwlwifi 0000:03:00.0: CONFIG_IWLWIFI_DEVICE_TRACING disabled
> iwlwifi 0000:03:00.0: CONFIG_IWLWIFI_DEVICE_TESTMODE enabled
> iwlwifi 0000:03:00.0: CONFIG_IWLWIFI_P2P enabled
> iwlwifi 0000:03:00.0: Detected Intel(R) Centrino(R) Ultimate-N 6300 AGN, REV=0x74
> iwlwifi 0000:03:00.0: L1 Enabled; Disabling L0S
> iwlwifi 0000:03:00.0: RF_KILL bit toggled to enable radio.
> iwlwifi 0000:03:00.0: device EEPROM VER=0x43a, CALIB=0x6
> iwlwifi 0000:03:00.0: Device SKU: 0x1F0
> iwlwifi 0000:03:00.0: Valid Tx ant: 0x7, Valid Rx ant: 0x7
> 
> I started with being connected ok to 2C:B0:5D:3C:7D:F1, and walked mostly out of range and close to 2c:b0:5d:a0:c8:a6
> 
> wpa_supplicant actually did connect to 2c:b0:5d:a0:c8:a6, then I had a few pings that went through the
> new AP, and then the pings stopped.
> Networking never recovered until after I ran kilall wpa_supplicant.
> This basically happens every time I roam :(
> 
> Debian runs it like so:
> root     14862  0.0  0.0  31024  2828 ?        S    20:59   0:00 /sbin/wpa_supplicant -u -s -O /var/run/wpa_supplicant
> 
> Should I be overriding what debian does and run it like so instead?
> /sbin/wpa_supplicant -u -onl80211 -O/var/run/wpa_supplicant -f /var/log/wpa_supplicant.log               
So, I tried that, and it didn't help:

Roaming works:
wlan0: SME: Trying to authenticate with 2c:b0:5d:a0:c8:a6 (SSID='magicnet-a' freq=5785 MHz)
wlan0: Trying to associate with 2c:b0:5d:a0:c8:a6 (SSID='magicnet-a' freq=5785 MHz)
wlan0: Associated with 2c:b0:5d:a0:c8:a6
wlan0: WPA: Key negotiation completed with 2c:b0:5d:a0:c8:a6 [PTK=CCMP GTK=CCMP]
wlan0: CTRL-EVENT-CONNECTED - Connection to 2c:b0:5d:a0:c8:a6 completed (reauth) [id=0 id_str=]


Feb 15 19:40:06 gandalfthegreat NetworkManager[18472]: <info> (wlan0): supplicant interface state: 4-way handshake -> completed
Feb 15 19:40:06 gandalfthegreat kernel: [388631.790948] cfg80211: Found new beacon on frequency: 5785 MHz (Ch 157) on phy0
Feb 15 19:40:09 gandalfthegreat NetworkManager[18472]: <info> (wlan0): roamed from BSSID 2C:B0:5D:3C:7D:F1 (magicnet-a) to 2C:B0:5D:A0:C8:A6 (magicnet-a)

I get 2 pings out after connecting to the new AP, and then the
connection hangs.

A while later wpasupplicant goes back to the old unusable AP:

wlan0: WPA: Group rekeying completed with 2c:b0:5d:a0:c8:a6 [GTK=CCMP]
wlan0: WPA: Group rekeying completed with 2c:b0:5d:a0:c8:a6 [GTK=CCMP]
wlan0: WPA: Group rekeying completed with 2c:b0:5d:a0:c8:a6 [GTK=CCMP]
wlan0: WPA: Group rekeying completed with 2c:b0:5d:a0:c8:a6 [GTK=CCMP]
wlan0: CTRL-EVENT-DISCONNECTED bssid=2c:b0:5d:a0:c8:a6 reason=2
wlan0: SME: Trying to authenticate with 2c:b0:5d:3c:7d:f1 (SSID='magicnet-a' freq=5745 MHz)
wlan0: Trying to associate with 2c:b0:5d:3c:7d:f1 (SSID='magicnet-a' freq=5745 MHz)
wlan0: Associated with 2c:b0:5d:3c:7d:f1
wlan0: WPA: Key negotiation completed with 2c:b0:5d:3c:7d:f1 [PTK=CCMP GTK=CCMP]
wlan0: CTRL-EVENT-CONNECTED - Connection to 2c:b0:5d:3c:7d:f1 completed (reauth) [id=0 id_str=]
Feb 15 19:41:00 gandalfthegreat kernel: [388685.212350] iwlwifi 0000:03:00.0: fail to flush all tx fifo queues
Feb 15 19:41:00 gandalfthegreat kernel: [388685.212462] wlan0: deauthenticated from 2c:b0:5d:a0:c8:a6 (Reason: 2)
(...)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.455482] wlan0: authenticate with 2c:b0:5d:3c:7d:f1
Feb 15 19:41:02 gandalfthegreat NetworkManager[18472]: <info> (wlan0): supplicant interface state: scanning -> authenticating
Feb 15 19:41:02 gandalfthegreat kernel: [388687.487733] wlan0: send auth to 2c:b0:5d:3c:7d:f1 (try 1/3)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.689936] wlan0: send auth to 2c:b0:5d:3c:7d:f1 (try 2/3)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.690640] wlan0: authenticated
Feb 15 19:41:02 gandalfthegreat NetworkManager[18472]: <info> (wlan0): supplicant interface state: authenticating -> associating
Feb 15 19:41:02 gandalfthegreat kernel: [388687.693975] wlan0: associate with 2c:b0:5d:3c:7d:f1 (try 1/3)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.897369] wlan0: associate with 2c:b0:5d:3c:7d:f1 (try 2/3)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.905992] wlan0: RX AssocResp from 2c:b0:5d:3c:7d:f1 (capab=0x11 status=0 aid=1)
Feb 15 19:41:02 gandalfthegreat kernel: [388687.909358] wlan0: associated

and then things don't work either because that AP is too far.

So why is roaming broken for me and how do I fix it?

Thanks,
Marc
-- 
"A mouse is a device used to point at the xterm you want to type in" - A.S.R.
Microsoft is to operating systems ....
                                      .... what McDonalds is to gourmet cooking
Home page: http://marc.merlins.org/