Novatel 551L + qmi

Wed Sep 17 15:04:25 PDT 2014

Thanks Dan! I was wondering if there exists a modem manager version that
works on openwrt/embedded systems ? I will start another thread on that to
keep things seperate.

Rgds,
Gopa.

On Wed, Sep 17, 2014 at 11:48 AM, Dan Williams <dcbw at redhat.com> wrote:

> On Wed, 2014-09-17 at 09:08 -0700, Gopakumar Choorakkot Edakkunni wrote:
> > Thanks again for the response Dan.
> >
> > >> If you are connected to LTE and run out of LTE coverage but EVDO is
> > available, the data connection will simply terminate, the modem will
> > eventually register on EVDO, and you will need to manually re-start the
> > data connection.
> >
> > From ModemManager code, I am assuming that reading the $NWQMISTATUS would
> > faithfully report such conditions ? Like if I keep polling for
> > $NWQMISTATUS, I would see status change from connected to not-connected
> > (whatever their equivalent values are) if this LTE-->EVDO change happens
> ?
>
> Right, the "novatel-lte" plugin in ModemManager polls $NWQMISTATUS every
> 5 seconds (because the modem doesn't report unsolicited $NWQMISTATUS I
> believe) and if it notices a disconnect it should indicate that to D-Bus
> clients by terminating the bearer.  It's then up to the connection
> manager to apply some kind of policy and request a reconnection when the
> modem enters the REGISTERED state.  But this plugin is only for the E362
> PCI-E minicard device, not the 551L.
>
> ModemManager drives the 551L with QMI, and since QMI is more completely
> implemented than the $NWQMI AT commands, network disconnections are
> indicated immediately via QMI instead of 5 seconds (worst case) with
> $NWQMI.
>
> Dan
>
> > Rgds,
> > Gopa.
> >
> >
> > On Wed, Sep 17, 2014 at 7:21 AM, Dan Williams <dcbw at redhat.com> wrote:
> >
> > > On Tue, 2014-09-16 at 12:39 -0700, Gopakumar Choorakkot Edakkunni
> wrote:
> > > > Hi Dan,
> > > >
> > > > Thanks a lot for the response. So I guess my firmware is old, I will
> try
> > > > upgrading to see if it work with just QMI without needing any
> > > $NWQMICONNECT.
> > > >
> > > > >> If eHRPD is not available on the network or is not enabled in the
> > > modem,
> > > > then you'll have to disconnect to move between 3G and 4G.
> > > >
> > > > When you say "you'll have to disconnect", do you mean to say user
> > > > intervention is required - should the user click "reconnect" on the
> modem
> > > > manager etc.. ? Or will the modem firmware do that internally
> without any
> > > > user intervention required (yes, the user might see a connection
> getting
> > > > reset and re-established, but thats still better than user having to
> > > click
> > > > "reconnect")
> > >
> > > Again, this is all if eHRPD is not available in the network, or somehow
> > > not enabled/supported in the modem.  eHRPD is the protocol that allows
> > > roaming between 3GPP2 (CDMA/EVDO) and 3GPP (LTE) RANs.
> > >
> > > So, if eHRPD is not available/working:
> > >
> > > If you are connected to EVDO, and LTE becomes available, the modem will
> > > *not* hand the data connection from EVDO to LTE.  You must terminate
> the
> > > data connection manually, wait for the modem to up-register to LTE, and
> > > then manually re-start the data connection.
> > >
> > > If you are connected to LTE and run out of LTE coverage but EVDO is
> > > available, the data connection will simply terminate, the modem will
> > > eventually register on EVDO, and you will need to manually re-start the
> > > data connection.
> > >
> > > Verizon had this hand-off issue for quite a while after deploying their
> > > LTE network in 2011, though they seem to have largely solved it by
> 2013.
> > > However, some of these initial launch devices (of which the USB551L is
> > > one) may not have firmware updates that fix all eHRPD issues, because
> > > they are long since out of production and unfortunately the
> > > manufacturers don't really care about old hardware.
> > >
> > > On the user side, a connection manager/modem manager would usually
> > > handle the reconnection without user interaction if that policy was
> > > desired.
> > >
> > > > Also when in 3G mode, eventually the packet encapsulation is ppp /
> slip I
> > > > guess - so does the firmware inside the modem initiate the ppp/slip
> when
> > > in
> > > > 3G mode, because from linux side all we see is a wwan0 interface.
> > >
> > > AFAIK yes, that's basically what happens in the firmware.  It
> > > decapsulates the IP packets from the over-the-air protocol (PPP for
> > > CDMA/EVDO, SNDCP for GPRS, IP for LTE) and re-encapsulates them into
> > > ethernet packets for the wwan0 interface.
> > >
> > > Dan
> > >
> > > > Rgds,
> > > > Gopa.
> > > >
> > > >
> > > > On Tue, Sep 16, 2014 at 11:47 AM, Dan Williams <dcbw at redhat.com>
> wrote:
> > > >
> > > > > On Mon, 2014-09-15 at 19:29 -0700, Gopakumar Choorakkot Edakkunni
> > > wrote:
> > > > > > Hi,
> > > > > >
> > > > > > First of all I tried that modem on my Ubuntu 14.04 with the
> > > modemmanager
> > > > > > GUI, I ask the GUI to scan for networks and it fails all the
> time.
> > > But
> > > > > > thats ok - my real requirement was to get it running on my
> openwrt
> > > box.
> > > > > > Plugged it in, saw that the option1 driver takes control of
> > > ttyUSB[0-3]
> > > > > and
> > > > > > qmi_wwan creates a wwan0 interface. After that running a dhcp
> client
> > > over
> > > > > > wwan0 dint come back with any IP, so reading through the
> modemmanager
> > > > > code,
> > > > > > I found that we have to send an AT+$NWQMICONNECT over ttyUSB0. I
> did
> > > that
> > > > > > and presto!, wwan0 gets an IP (thanks modemmanager developers
> !).  A
> > > > > couple
> > > > > > of questions
> > > > >
> > > > > The device runs optimally with QMI, but does work with
> $NWQMICONNECT as
> > > > > a fallback.  However, QMI works much better after the firmware
> update
> > > to
> > > > > 1.41.
> > > > >
> > > > > $NWQMICONNECT likely just calls the QMI WDS data session functions
> > > > > anyway.  What you are probably missing is using qmicli or
> ModemManager
> > > > > to start the WDS data session with the QMI protocol, and *then*
> you can
> > > > > run DHCP on the network port.  You cannot just run DHCP without
> first
> > > > > starting the data session via QMI or $NWQMICONNECT.
> > > > >
> > > > > > 1. Does anyone know "internally" what is the modem doing on
> getting
> > > an
> > > > > > AT+$NWQMICONNECT ? Does it start a ppp session if the available
> band
> > > is
> > > > > > only 3G ? Once started in 3G and later a 4G-LTE connection is
> > > available,
> > > > > > will the modem internally drop the ppp session and switch the IP
> on
> > > > > 4G-LTE
> > > > > > (and vice versa) ?
> > > > >
> > > > > Handoff between 3G and 4G is kinda dicey with these first-gen
> > > > > Qualcomm-based Verizon devices (USB551L, Pantech UML290).  Part of
> the
> > > > > issue is eHRPD which is required for handoff between 3G and 4G.  If
> > > > > eHRPD is not available on the network or is not enabled in the
> modem,
> > > > > then you'll have to disconnect to move between 3G and 4G.
> > > > >
> > > > > > 2. Other than the AT+$NWQMICONNECT, do we need to send anything
> else
> > > to
> > > > > the
> > > > > > modem to make it work using all available bands based on
> availability
> > > > > (3G /
> > > > > > 4G) or do we have to specify anything to force the modem to work
> in
> > > > > "these
> > > > > > specific modes" ? I dint find anything like that in the modem
> manager
> > > > > code,
> > > > > > just thought of asking anyways.
> > > > >
> > > > > If you've left the technology preference, system selection
> preference,
> > > > > and band preferences alone, then the modem will automatically
> search
> > > for
> > > > > and connect to the best available technology (eg, LTE) and network.
> > > But
> > > > > the device only supports one Verizon LTE band (eg, Band 13), so
> you're
> > > > > pretty limited on the LTE side.
> > > > >
> > > > > Dan
> > > > >
> > > > >
> > > > _______________________________________________
> > > > libqmi-devel mailing list
> > > > libqmi-devel at lists.freedesktop.org
> > > > http://lists.freedesktop.org/mailman/listinfo/libqmi-devel
> > >
> > >
> > >
> > _______________________________________________
> > libqmi-devel mailing list
> > libqmi-devel at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/libqmi-devel
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/libqmi-devel/attachments/20140917/0f2fa888/attachment.html>