Daniel Blueman wrote:
> Alan, David,
> 
> Thanks for your response - the problem observed in linux-2.6.5 is fixed in
> 2.6.6. Now I get a status code of -75 [1] and see no additional networking
> devices.

In drivers/usb/core/message.c::usb_string(), try making the retry
cases consider not just -EPIPE faults but also -EOVERFLOW.  Likewise
in the get_descriptor() logic.  (Some devices seem to have firmware
bugs that cause those failures, and are often amenable to retries.)

Also, this patch should help correct the "missing diagnostics with
CONFIG_USB_DEBUG during CDC Ethernet probe()" issue ... and help
show more precisely where the problem is.

- Dave