[review] dcbw/huawei-at-dhcp: use AT^DHCP response if available for NDISDUP-capable devices

[Thomas Schäfer]

Am Mittwoch, 19. August 2015, 12:52:49 schrieb Dan Williams:
> Hi,
> 
> I had at least one report of a HiSense device that had successful
> AT^NDISSTATQRY reply but DHCP didn't work.  There's been discussion on
> this in the past [1] with no real conclusion that I can see, but maybe
> ^DHCP works here instead.  So I just went ahead and did that.  It'll
> need more testing with Huawei modems to make sure we can actually rely
> on the results (we know we can't for QMI-based devices) but at least the
> patch can be reviewed.
> 

Sorry for waiting so long.

Could you say, what I should test now? 
A special mm-version/branch?
The results of AT^DHCP and real dhcp on the interface? 


> [1]
> http://lists.freedesktop.org/archives/modemmanager-devel/2014-October/00157
> 6.html


This problem was caused by the kernel-driver. It is solved - in 4.2.
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/drivers/net/usb/huawei_cdc_ncm.c?id=4a0e3e989d66bb7204b163d9cfaa7fa96d0f2023

This patch did the wonder for at least two devices I know/own:

E3372 alias Telekom Speedstick V
E5786s-32a alias Telekom Speedbox LTE mini II.

The first one needs mm/nm to work - or just 

echo -en 'at^ndisdup=1,1,"internet.telekom"\r' > /dev/ttyUSB0
dhcpcd wwp0s29f7u4i2 

The second should work with nm, but it doesn't at the moment, because nm 
doesn't list the interface, but 

dhcpcd wwp0s29f7u4i1

works.


Regards,
Thomas