Dell 5570 (Sierra) supported?

[Aleksander Morgado]

On Tue, Nov 25, 2014 at 9:07 PM, Dan Williams <dcbw at redhat.com> wrote:
> On Tue, 2014-11-25 at 11:37 -0600, Dan Williams wrote:
>> On Tue, 2014-11-25 at 18:27 +0100, Aleksander Morgado wrote:
>> > On Tue, Nov 25, 2014 at 3:31 PM, Manuel Lauss <manuel.lauss at gmail.com> wrote:
>> > >>> <debug> [000000.007862] [mm-plugin-manager.c:640] build_plugins_list(): (Plugin Manager) [cdc-wdm0] Found '2' plugins to try...
>> > >>> <debug> [000000.007863] [mm-plugin-manager.c:644] build_plugins_list(): (Plugin Manager) [cdc-wdm0]   Will try with plugin 'Novatel'
>> > >>> <debug> [000000.007864] [mm-plugin-manager.c:644] build_plugins_list(): (Plugin Manager) [cdc-wdm0]   Will try with plugin 'Generic'
>> > >>> <debug> [000000.007871] [mm-plugin.c:767] mm_plugin_supports_port(): (Novatel) [cdc-wdm0] probe required: 'at'
>> > >>> <debug> [000000.007873] [mm-port-probe.c:1353] mm_port_probe_run(): (usbmisc/cdc-wdm0) launching port probing: 'at'
>> > >>
>> > >>
>> > >> This doesn't look so good, though...  Why do we end up probing this as
>> > >> an 'at' port? No cdc-wdm0 device owned by the cdc_mbim driver should
>> > >> ever be probed as anything but MBIM.  I don't know how this device
>> > >> reacts, but there is a fair chance that some MBIM firmwares will barf at
>> > >> random AT commands bveing thrown at them.
>> > >
>> >
>> > The Novatel plugin is conflicting here it seems.
>>
>> Because "Dell":
>>
>>     static const gchar *subsystems[] = { "tty", "net", "usb", NULL };
>>     static const guint16 vendors[] = { 0x1410, /* Novatel */
>>                                        0x413c, /* Dell */
>>                                        0 };
>>
>> I'm not sure how to further restrict to Novatel-only devices though,
>> given that the vendor ID can be anything (Dell, HP, whatever).  I'm not
>> sure there's a single AT command that works on all the AT-only Novatel
>> modems, though we could certainly put together a list that covers 97% of
>> them at least.
>
> The case we care about for Novatel is AT-only devices.  For those, here
> is a list of commands we can use to identify them and some sample
> replies.  Built from the XU870 (Dell Wireless 5510), EX725, S720, E362.
>
> We should only do this probing if the following are true:
>
> 1) The device only has AT ports; eg it does *not* have any QMI, MBIM, or
> "net" ports
>
> 2) The device does *not* report LTE capabilities via ATI or GCAP
>
> Early Novatel QMI-capable devices (like the E362) support these commands
> (at$nw*) but of course we don't want to treat them as AT-only Novatel.
>
> =======
>
> ati1
> Manufacturer: Novatel Wireless Incorporated
> Manufacturer: NOVATEL WIRELESS INCORPORATED
>
> AT+GMI
> Novatel Wireless Incorporated
> NOVATEL WIRELESS INCORPORATED
>
> at$nwdegc   (GSM/UMTS only)
> (non-error response sufficient)
>
> =======
>
> If for some reason those don't work, here are some more esoteric ones
> that we could potentially use:
>
> AT$NWHFP?   (gsm only)
> AT$NWCSQ    (gsm only)
> AT$NWWDIS?  (gsm only)
> AT$NWERR?   (gsm only)
> AT$NWCFT?   (gsm only)
> AT$NWRAT?   (gsm only)
> AT$NWLTIME  (cdma only)
> at$nwrssi   (cdma only)
>
> Dan
>

Opened a new bugreport to track the issue:
https://bugs.freedesktop.org/show_bug.cgi?id=86713

-- 
Aleksander
https://aleksander.es