Fundamental doubt

[João M. S. Silva]

Thanks for the explanation.

On 08/31/2015 07:59 PM, Dan Williams wrote:
> Not quite correct.  ModemManager does not use D-Bus to talk to the
> modem; it uses D-Bus to talk to client applications.  To talk to the

So, when I use mmcli (client application), I am using the DBus to talk 
to the ModemManager daemon, which then connects directly to the HSDPA 
modem through /dev/ttyUSB*?

> modem, MM uses the protocol that the modem supports (eg AT-over-serial,
> AT-over-WDM, MBIM, QMI, or QCDM-over-serial) to speak with the kernel
> driver, which then formats the data correctly for the modem itself.

But in the case of the SIM800L module, does the Linux kernel recognize 
it? Does it have to? Why? I am talking directly with the modem. There is 
no hardware recognition message in the kernel.

In the case of the HSDPA modem, the kernel recognizes the hardware and 
creates /dev/ttyUSB*, if I understand correctly.

> So your SIMCom device will have serial lines that a kernel driver knows
> how to control, and the kernel exposes a device node (like ttyACM0 or
> ttyUSB0 or cdc-wdm0), which is what ModemManager then uses to control
> the device based on what requests client applications make via MM's
> D-Bus API.

Hum... so I guess the kernels I'm using don't recognize this hardware, 
so no nodes are created.

> The only caveat here is that many modems use proprietary AT commands for
> some of their functionality, and if there is no support in ModemManager
> for those commands some functions may not be available.  MM has a
> SimTech/SimCom plugin already, but for other SimCom devices that appear
> to support different commands than the 800L.

So, I could eventually use MM but it currently does not support SIM800L 
commands, right?

How do I make Linux recognize SIM800L?

-- 
João M. S. Silva