PPC: failed with error -12

Live forum: http://rt2x00.serialmonkey.com/viewtopic.php?t=3753


07-05-2007 21:35:58

EDIT Just read should be debug=2 . Will try again. ooops
EDI2 Still nothing in debug with the above setting.

Dongle Belkin f5d7050 rev3.
Kernel 2.6.18-1.2798 PPC
Fedora 6

I'm getting the error on plugging in the dongle,

Device Descriptor not matching
rt73: probe of 1-1@1.0 failed with error -12

I'm using the same dongle on my Ubuntu i386 box and it works fine with the rt73 module. Guessing its another problem with the PPC kernel /

There is nothing in /var/log/debug before you ask ) .

I built the module with make debug and inserted it with debug=31.

Any ideas?



10-05-2007 04:42:45

Hi there.

Are you plugging the dongle straight into the PC or through a hub ?


10-05-2007 20:09:42

Please retry with a current CVS.


11-05-2007 00:24:26

It was plugged directly into the back on the laptop. I'll try again with the latest CVS.

Thanks for replying, be back soon )


11-05-2007 01:01:10

K just tried latest CVS and I get the same thing.

Am I doomed by using a PowerBook?

Any ideas welcome,



11-05-2007 09:33:06

It seems your device isn't in the list of supported devices. We might have to add it.
Please show us the following command output[code39j8fvru]# lsusb[/code39j8fvru]


11-05-2007 11:48:35

I checked to make sure and my FCC ID matches a v3 dongle which on your hardware page has been tested

Belkin F5D7050 Ver 3 Ubuntu Feisty Fawn 7.04 managed

Right, my output is

idVendor 0x050d Belkin Components
idProduct 0x705a



11-05-2007 12:04:37

I wrote this because of the [i2n53qrxo]'Device Descriptor not matching'[/i2n53qrxo] error message, but it is supported indeed.

Error -12 is supposed be a [i2n53qrxo]'Out of memory error'[/i2n53qrxo]...
I suggest you read the TESTING file and post here everything it asks you to.


13-05-2007 10:28:58

Attached is modinfo and debug. I downloaded the latest CVS about ten mins ago.


EDIT btw, I realise there is nothing about the error in the debug log. I noted this in my first post but I tried again just incase I had done something wrong. Both times I inserted the module with debug=2 and once I tried it with debug=31 and nothing was ever entered into the debug log (afaik).

EDIT2 Oh an I've got about 700mb RAM free


15-05-2007 22:05:08

I've got a feeling I've run out of luck. Thanks for your help once again, all the best.


16-05-2007 13:44:02

Sorry jonface, there is just nothing useful in the logs you posted, and i can't tell from the little info i have where you problem comes from.


01-07-2007 18:14:43

I've the same problem, and the driver (latest cvs as of today) compiled with debug and inserted with "debug=2" option on the 2.6.22-7-powerpc "gutsy" ubuntu kernel gives me

[code25lqkiyj]Jul 1 19:07:28 localhost kernel: [ 313.309403] rtusb init ====>
Jul 1 19:07:28 localhost kernel: [ 313.312731] usbcore: registered new interface driver rt73
Jul 1 19:07:50 localhost kernel: [ 335.971096] usb 2-1: new full speed USB device using ohci_hcd and address 2
Jul 1 19:07:51 localhost NetworkManager: <debug> [1183309671.307854] nm_hal_device_added(): New device added (hal udi is '/org/freedesktop/Hal/devices/usb_d
Jul 1 19:07:51 localhost kernel: [ 336.414303] usb 2-1: configuration #1 chosen from 1 choice
Jul 1 19:07:51 localhost kernel: [ 336.416213] Device Descriptor not matching
Jul 1 19:07:51 localhost kernel: [ 336.416226] rt73: probe of 2-1:1.0 failed with error -12[/code25lqkiyj]

now, i'm trying to figure why is failing here. I've put an additional print of idVendor and idProduct at line 2104 of rtmp_main.c, and now when i insert the usb device, it prints out the following ids

[code25lqkiyj]Jul 1 19:30:25 localhost kernel: [ 369.171343] rtusb init ====>
Jul 1 19:30:25 localhost kernel: [ 369.171520] usbcore: registered new interface driver rt73
Jul 1 19:30:33 localhost kernel: [ 377.143219] usb 2-1: new full speed USB device using ohci_hcd and address 2
Jul 1 19:30:34 localhost NetworkManager: <debug> [1183311034.073368] nm_hal_device_added(): New device added (hal udi is '/org/freedesktop/Hal/devices/usb_d
Jul 1 19:30:34 localhost kernel: [ 377.589424] usb 2-1: configuration #1 chosen from 1 choice
Jul 1 19:30:34 localhost kernel: [ 377.591331] azazel_debug idVendor = 0x7d10000, idProduct = 0x3c030000 [/code25lqkiyj]

now, i've tried exactly the same code on a i386 that is running a debian 2.6.21, and it prints out different ids, respectively idVendor = 0x7d1 and idProduct = 0x3c03 which are the correct ids configured on the header rtmp_def.h

is it still a problem with big endian? please help me. I can do every test you guys need

i've tried also the new rt2x00 driver. It loads well and handles the dongle, but the scan results always empty.


02-07-2007 21:31:18

It doesn't look like an endianess problem, bytes would be in reverse order if it was the case. Have you tried the "ugly way" modifying "rtmp_def.h" accordingly? Of course, that would only be a temporary workaround.



06-07-2007 13:24:35

This specific problem can be fixed with relative ease (which I will do within the next 24 hours). However, don't be too hopeful in getting the legacy to work with ppc.

You should try rt2x00.

EDIT Try sticking in ">>16" after le32_to_cpu(xxxxx) on line 2108 and 2019 in rtmp_main.c.

it should end up looking like
if (le32_to_cpu(dev->descriptor.idVendor)>>16 == rtusb_usb_id[i].idVendor &&
le32_to_cpu(dev->descriptor.idProduct)>>16 == rtusb_usb_id[i].idProduct)


16-07-2008 01:23:55

I posted my solution here