After moving to 2.6.10 from 2.4.x, trying to use firmware upload in kernel, I've got the following:
Feb 27 03:16:33 lapcio kernel: speedtch_test_sequence failed on URB147: -32 Feb 27 03:16:33 lapcio kernel: speedtch_test_sequence failed on URB148: -32 Feb 27 03:16:33 lapcio kernel: speedtch_test_sequence failed on URB149: -32 Feb 27 03:16:33 lapcio kernel: speedtch_test_sequence failed on URB150: -32 Feb 27 03:16:33 lapcio kernel: SpeedTouch: Failed to start ADSL synchronisation: -32 I've seen other people already reporting something like this, with no useful solutions, so I've decided to try to resolve it myself. First of all it appears that it only happens if I connect the modem via usb 2.0 hub. I never had any problems connecting it directly to my laptop. To make the whole thing more misterious, it sometimes works ok via hub - like every 5-10th try. It appears that the check that the driver performs at the beginning to see if the firmware is loaded is the reason. After plugging in the device it always returns -110 (ETIMEDOUT) on ep0in, but it also (sometimes, see above) makes all subsequent requests to ep0 fail with EPIPE (STALL condition). I haven't seen windows driver doing anything like that after plugging in - it goes directly to firmware upload stuff. Attached patch solved the problem for me (by resetting the usb device before firmware upload), modem starts up properly every time. I'm not sure what is the real reason - is it endpoint 0 in the device really getting STALLed, or is it some EHCI hcd/hub implementation weirdness. It's hard to say without digging very deep into ehci implementation ;) Best Regards, Przemyslaw -- Binary/unsupported file stripped by Listar -- -- Type: text/x-diff -- File: speedtch-reset.patch Liste de diffusion modem ALCATEL SpeedTouch USB Pour se d�sinscrire : mailto:[EMAIL PROTECTED]
