[PATCH] Ignore Huawei E169/E620 USB GPRS modem

[Dan Williams]

On Thu, 2008-10-09 at 12:50 +0100, Matthew Garrett wrote:
> On Thu, Oct 09, 2008 at 01:27:59PM +0200, Danny Kukawka wrote:
> > On Donnerstag, 9. Oktober 2008, Matthew Garrett wrote:
> > > Yeah, and the kernel drivers are being extended to do that. I think we
> > > should probably remove the existing Huawei entry from
> > > 20-broken-usb-sticks.fdi.
> > 
> > Since which kernel version should this work?
> 
> Well, my argument here is that the device *is* a USB mass storage 
> device. HAL shouldn't be the one responsible for telling desktops to 
> ignore it.

Novatel U727, Option iCON 225, various Sierra devices...

Here's the right way to do this:

1) Make the USB serial drivers (cdc-acm, hso, sierra, novatel, etc)
_only_ match the real character device by using stricter matching of the
interface class, interface sub-class, and protocol

2) Add a small usb-storage driver that claims the device on insert, and
based on defaults and/or module parameters, do the magic to reset the
device into TTY mode instead of storage mode in that small usb-storage
driver

3) Add an entry to unusual-devs for that small usb-storage driver

Please see:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=0585e4dfe5670d3ccb77bf86551a657699e9e52e
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=32fe5e393455d87db4988af03915634304870fb4

For 'hso'-driven devices, you would basically do the operations that the
'rezero' tool does in the usb-storage driver.

This stuff can be done (badly) with udev, but that's really the wrong
place to do it.  Fix the drivers and the system the right way, don't do
a bad solution in HAL that won't really work.

Dan