Re: [PATCH 0/2] *** SUBJECT HERE ***

[Johan Hovold <jhovold@gmail.com>]

On Thu, Jun 27, 2013 at 11:50:52PM +0200, Anders Hammarquist wrote:
> In a message of Wed, 26 Jun 2013 12:39:24 +0200, Johan Hovold writes:
> >On Wed, Jun 26, 2013 at 10:29:59AM +0200, Anders Hammarquist wrote:
> >> In a message of Tue, 25 Jun 2013 16:39:11 -0700, Greg KH writes:
> >> >> Indeed. I'd already had some (failed) thoughts about how to handle it
> >> >> nicely. Now I've had another think through, and I have something which
> >> >> deals with it and at least complains if TI_EXTRA_VID_PID_COUNT is changed
> >> >> without changing the initializer. Patch 2/2
> >> >
> >> >Why don't we just drop the extra id thing entirely?  The usb-serial
> >> >subsystem handles new device ids being added dynamically from sysfs for
> >> >a long time now.  Removing this module option would clean up the code a
> >> >lot, and prevent these errors from ever happening again.
> >> 
> >> Aha, yes, I'm all for that (had I only known I'd have done that to start
> >> with). I'll look in to it.
> >
> >I already have a few patches here (part of a larger 3.11 clean-up series)
> >which removes the vid/pid module parameters from all usb-serial modules
> >including ti_usb_3410_5052.
> >
> >I hope to be able to submit the whole series a later tonight, but here's
> >the ti_usb_3410_5052 part if anyone's interested.
> 
> I did a quick check of adding the device id though sysfs, and although
> it partly works, it doesn't find the correct firmware (it ends up trying
> to load 5052 firmware for a 3410 device. Looking at the code it seems
> (struct ti_device) td_is_3410 isn't set properly.)

Turns out that the drivers device-type detection has never worked with
the dynamic id interface (all devices were detected as 2-port devices).

I'm responding to this mail with a fix. Care to give it a try?

Thanks,
Johan