Identifying mobile phones as modems

Alex Kanavin ak at sensi.org
Tue Apr 8 13:02:11 PDT 2008


2008/4/8, Daniel Qarras <dqarras at yahoo.com>:
>  > I think the best that hal can do is to mark all CA-42, DKU-2, etc
>  > devices (detected by interface class == 2 and interface subclass ==
>  > 2,
>  > no hardcoding of interface numbers or vendor/product ids) as generic
>  > v.25 modems, and let the applications figure out additional command
>  > sets, if they need them, by either asking the user, or sending AT
>  > commands to the device.
>  >
>  > Here's some discussion on DKU-2 phones which has a similar problem
>  > (unable to distinguish between GSM and CDMA):
>  > http://bugzilla.gnome.org/show_bug.cgi?id=525305
>  > http://bugs.freedesktop.org/show_bug.cgi?id=15346
> Thanks a lot for your help! I've attached a patch proposal, I think it
>  does what you suggest. I'm bit surprised if leaving vendor/product ids
>  is ok but then I fully don't know the finer logic involved here.
>
>  Anyway, if the patch looks sane perhaps it could be merged then?

Seems fine to me, if it's not going to clash with the probing feature
David Zeuthen plans to add. The only thing that I have a comment on is
this:
+      <!-- PHONE CONNECTOR CABLES -->
+
+      <!-- Nokia CA-42 style cables for GSM/CDMA phones -->

The cable can be anything, CA-42 or DKU-5 (which have a builtin
USB-to-serial chip, hence the price), DKU-2 or CA-53 (which are
completely passive, but still have a proprietary plug on the phone
side), or even a standard mini-USB cable like the one my Nokia N95 is
using. It's about the USB interface that the PC ultimately sees, in
this case CDC ACM. I suggest changing the comment to:

Communication Device Class Abstract Control Model (CDC ACM) modems,
typically provided by GSM/CDMA phones

-- 
Alexander


More information about the hal mailing list