problem with huawei e355 with xubuntu 14.04 and modemmanager 1.0.0
aleksander at aleksander.es
Thu May 29 01:06:37 PDT 2014
On Thu, May 29, 2014 at 5:51 AM, som <drsoumalya-mm at yahoo.co.in> wrote:
> On Tuesday, 22 April 2014 9:36 PM, Dan Williams <dcbw at redhat.com> wrote:
>>On Tue, 2014-04-22 at 16:50 +0200, Aleksander Morgado wrote:
>>> On Tue, Apr 22, 2014 at 4:42 PM, Dan Williams <dcbw at redhat.com> wrote:
>>> >> commit 1f84eab4ad73bcb7d779cba65291fe62909e373f
>>> >> Author: Bjørn Mork <bjorn at mork.no>
>>> >> Date: Wed Feb 27 03:08:58 2013 +0000
>>> >> net: cdc_ncm: tag Huawei devices (e.g. E5331) with FLAG_WWAN
>>> >> Tag all Huawei NCM devices as WWAN modems, as we don't know of any which
>>> >> are not. This is necessary for userspace clients to know that the device
>>> >> requires further setup on e.g. an AT-capable serial ports before
>>> >> connectivity is available.
>>> >> The device is truly a WWAN modem, but again it doesn't require any
>>> >> ModemManager setup. Dan, can we maybe update NetworkManager to
>>> >> initially ignore all WWAN flagged devices but only if ModemManager
>>> >> takes them afterwards in a modem object? In other words, if MM doesn't
>>> >> end up taking the interface (hum.. how to notify that?) then NM can
>>> >> try to expose it as a USB net interface...
>>> > So yeah, we can do this. But I'd also like to pair it with a
>>> > ModemManager property for "detecting" that would get flipped on/off when
>>> > MM was in the process of detecting a new modem. That would give us two
>>> > things:
>>> > 1) NetworkManager would know when the process was done and could then
>>> > let the WWAN tagged device be an ethernet device
>>> > 2) User Interfaces could show an indicator when MM was working, so that
>>> > where detecting the modem might take a bit of time, the user isn't left
>>> > wondering what's happening, because they plugged in the modem 10 seconds
>>> > ago and it still hasn't shown up in the menu.
>>> > What do you think?
>>> I think that I have suggested that "MM is probing" in the past myself,
>>> and don't recall why it was not done at the end. Probably missed it in
>>> my TODOs.
>>> But anyway, what if we have multiple probing sequences for multiple
>>> devices? The "detecting" property not sure it can be fit in a boolean
>>> value... or maybe yes to simplify it. At the end I'm not sure there
>>> will be many WWAN-flagged net devices which are not handled by MM.
>>I think such a property has value even if there are WWAN-flagged devices
>>not handled by MM, just for the UI notification aspect, so users are
>>informed about what's going on, that the device is still being set up.
>>That way users don't think that something is wrong and just pull the
>>device and try again. That's the UI aspect.
>>To be more concrete, in GNOME Shell, I would modify the indicator menu
>>to show a Mobile Broadband section whenever either (a) a detected modem
>>is present, or (b) whenever the MM "detecting" property was TRUE. When
>>MM was just detecting, but no modem was found, it could show a spinner
>>in the Mobile Broadband section to indicate that something is happening.
>>The second, *separate* aspect would be that NM would use that property
>>to know when to present the WWAN-tagged device as a plain ethernet
>>interface. Otherwise we would have to do random timeouts in NM, like
>>"if ModemManager doesn't export a new modem that uses this port within
>>30 seconds, make it plain ethernet", which is somewhat fragile.
> what is the status of this huawei problem?
I just opened a bugreport to keep track of it:
More information about the ModemManager-devel