[SOLVED] Usb wifi TL-WN422G saturates my dmesg!

Mikro · 07-07-2010, 07:20 PM

Hi, I am trying to use a tp-link TL-WN422G under slackware 13.1 kernel 2.6.34 using compat-wireless for ath9k_htc with its ar9271 firmware.

The card is being recognized, iwlist wlan4 scan returns the AP's around, it seems to associate to the AP using wep (no network response, although, maybe it is not working), but the thing is it keeps logging

[dmesg]

Code:

[31101.200112] usb 2-3: link qh0-00ff/f25ed880 start 0 [2/0 us]
[31101.200297] usb 2-3: unlink qh0-00ff/f25ed880 start 0 [2/0 us]

every few mseconds. I have tried to disable the debug output using modprobe ath9k_htc debug=0x00000000 but it doesn't make any difference.
Also tried disabling leds and used nohwcrypt flags.

/var/log/messages says

Code:

Jul  8 02:15:54 darkstar kernel: [32358.813632] usb 2-3: firmware: requesting ar9271.fw
Jul  8 02:15:55 darkstar kernel: [32359.108041] usb 2-3: ath9k_htc: Transferred FW: ar9271.fw, size: 51280
Jul  8 02:15:55 darkstar kernel: [32359.375343] usb 2-3: ath9k_htc: HTC initialized with 33 credits
Jul  8 02:15:56 darkstar kernel: [32360.098588] cfg80211: Calling CRDA for country: CN
Jul  8 02:15:56 darkstar kernel: [32360.104067] usb 2-3: ath9k_htc: USB layer initialized
Jul  8 02:15:56 darkstar kernel: [32360.104082] usbcore: registered new interface driver ath9k_hif_usb
Jul  8 02:15:56 darkstar kernel: [32360.111654] udev: renamed network interface wlan1 to wlan4
Jul  8 02:15:56 darkstar kernel: [32360.135858] cfg80211: Current regulatory domain intersected:
Jul  8 02:15:56 darkstar kernel: [32360.135860]     (start_freq - end_freq @ bandwidth), (max_antenna_gain, max_eirp)
Jul  8 02:15:56 darkstar kernel: [32360.135862]     (2402000 KHz - 2472000 KHz @ 40000 KHz), (N/A, 2000 mBm)
Jul  8 02:15:56 darkstar kernel: [32360.135864]     (2457000 KHz - 2482000 KHz @ 20000 KHz), (N/A, 2000 mBm)
Jul  8 02:15:56 darkstar kernel: [32360.135866]     (2474000 KHz - 2482000 KHz @ 8000 KHz), (N/A, 2000 mBm)
Jul  8 02:15:56 darkstar kernel: [32360.554050] ADDRCONF(NETDEV_UP): wlan4: link is not ready
Jul  8 02:15:59 darkstar logger: /etc/rc.d/rc.inet1:  /sbin/iwconfig wlan4 nick darkstar
Jul  8 02:15:59 darkstar logger: /etc/rc.d/rc.inet1:  /sbin/iwconfig wlan4 essid "any"

Does anyone have any idea of what this means, what could be wrong or how I could disable debug information?

Thanks in advance

**EDIT**

I think this could be related to the usb suspend/wakeup features, but AFAIK there should be a /sys/bus/usb/devices/2-3/power/level file for it to support it... and there isn't.

Any ideas?

Dån

SCerovec · 07-08-2010, 06:12 AM

Why did You switch to the unofficial kernel?
Did You try build a module from mad-wifi (is there any?)
Are You familiar with C structures?

I just recently made an totally different kernel module (build from outside the kernel tree) stop spamming my "dmesg".

I just found and commented out the part of code that reports some sort of calibration debug info to the system logs.

recompiled and it was *away*.
I found the offending piece of code by grepping the fragments of the flood messages.

Could You provide a dozen of the "boring" lines so we could conclude how a pattern would look like?

Stevan C.

Mikro · 07-08-2010, 09:30 AM

I'm afraid I don't understand very well your post, but

Quote:

Originally Posted by SCerovec

Why did You switch to the unofficial kernel?

I like to configure my own kernels; gives me headaches like this, but I think it is worth...

Quote:

Did You try build a module from mad-wifi (is there any?)

I think the only one is the compat-wireless ath9k_htc

Quote:

Are You familiar with C structures?

Yep

The message is coming from drivers/usb/host/ehci-sched.c, functions qh_link_periodic and qh_unlink_periodic. I haven't studied the code, but it looks like it is just a debug informational message. I've disabled USB_DEBUG (USB verbose debug messages) in the kernel config and recompiled.

Dån

Mikro · 07-08-2010, 09:40 AM

Ok, it is now solved. Everything looks fine, no more debug messages... It was the USB_DEBUG=y kernel option (when I first turned it on, I thought the debug information was not going to be so verbose, but what the hell, I am not going to debug USB gadgets right now, so USB_DEBUG=n). I will look if it was really an information message and not error reporting and post if I find anything new.

Thanks SCerovec, asking me about the kernel made me look in the config...

Dån