On Wed, 2013-08-14 at 13:58 -0500, Peter Hyman wrote: > On 08/14/2013 01:58 PM, Dan Williams wrote: > > snip > > The 4G part of the 250U is a Beceem WiMAX chip, for which there are no > > drivers on Linux at this time. The 4G parts are used with a > > pseudo-ethernet interface, but since there aren't any kernel drivers, > > that's not going to work. The 4G parts do not use PPP. The 3G part, > > however, *does* use PPP over the serial ports. So with this > > information, it's not surprising that when you choose "4G Only" in > > Windows, the 3G part (eg, PPP over serial ports) isn't going to work > > on Linux. Dan > > Yes, but if we knew 1) to Identify that 4G has been set, and 2) the > command to toggle the device from 4G-3G, then this problem could be > identified by the driver and the unit switched back to 3G and a log > message printed. Then, the issue would be avoided, right? It's almost certainly not the domain of the driver to switch this mode, it's the responsibility of userland. Kernel drivers must only talk to the device, they are not supposed to do any kind of policy or switching. Also, it would have to be a choice to switch, since some users may already be using the out-of-tree Beceem WiMAX stack and drivers and we can't screw it up for them. It's possible that we could determine the current mode via the proprietary QCDM using the "modepref" tool from ModemManager sources: http://cgit.freedesktop.org/ModemManager/ModemManager/tree/libqcdm/tests but QCDM has no "WIMAX" option that I'm aware of, so I'd be quite curious to see what modepref reports when the device is in 4G-only mode. I've attached an x86_64 build of modepref, give it a shot on ttyUSB1, ttyUSB2, or ttyUSB3 and see what comes out: modepref /dev/ttyUSB1 --debug Thanks, Dan
Attachment:
modepref
Description: application/executable
