Re: Question about your reversion of the udev Huawei modem rule

[Dan Williams <dcbw@redhat.com>]

On Tue, 2010-01-05 at 22:36 +0100, Martin Pitt wrote:
> Hey Dan,
> 
> in [1] you reverted a previous commit [2] to fix the Huawei E1550 GSM
> modem in udev (using modem-modeswitch) with the rationale that "other
> manufacturers have completely different methods for killing their fake
> driver CD". However, the rule was tested on that actual hardware, and
> it worked fine.

Lets use the *correct* method for ejecting a device instead of using one
manufacturer's method with another manufacturer's device.  'rezero' is
*NOT* the correct method for ejecting Huawei devices.  It doesn't really
matter whether it works or not.

usb_modeswitch has working methods for ejecting the E1550 (at least they
certainly work for me).  Rezero is Option NV specific and can/will cause
undesirable results on other manufacturer's equipment.  ZTE devices
specifically can be affected by this and do not react well, confusing
both the device and the kernel USB stack in the process.

I should think it would be clear that we shouldn't be using 'rezero'
with devices for which its not intended, but instead to use the method
that all other Huawei devices use (usb_stor_huawei_e220_init from
drivers/usb/storage/initializers.c).

The E1550 has a new USB ID that's not found in unusual_devs.h; just add
the ID there, or start using usb_modeswitch.

Dan

> I'm interested in why this rule was bad, and what should be done
> instead?
> 
> Thanks,
> 
> Martin
> 
> [1] http://git.kernel.org/?p=linux/hotplug/udev.git;a=commitdiff;hŠ993fab2f218234f06743bcc127dda61a45711b
> [2] http://git.kernel.org/?p=linux/hotplug/udev.git;a=commit;hS842b53820a0f0b2b0ff5b28caf01e3f7e26f22