[Bug 234578] Support for Sierra Wireless EM7455 modem

From: <bugzilla-noreply_at_freebsd.org>
Date: Thu, 16 Sep 2021 07:35:33 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=234578

Maurice J Niessen <du@hohlhupe.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |du@hohlhupe.de

--- Comment #11 from Maurice J Niessen <du@hohlhupe.de> ---
As Qingjun Wei already mentioned, you are discussing different issues here and
you are losing focus...

The device 1199:9079 is not handled by u3g.ko to give us support (like
/dev/ttyU*, /dev/cuaU*) for sending AT commands to it. AFAIK other methods like
MBIM, QMI and RMNET* aren’t supported in FreeBSD at all.

Sending “magic packets” to specific OEM versions of 1199:9079 (and yes, mine
says “customer: lenovo-laptop” as well) to overcome the “low power mode (LPM)/
FCC AUTH” issue is something else. BUT for the “magic” of a challenge/response
authentication and setting FCC AUTH=0, it’s necessary to send AT commands to
our lovely 1199:9079.
Dean already mentioned the “description” in form of a python script how to do
that “magic”. Within the years, the link has changed. It can be found here now:

https://github.com/bkerler/edl/blob/master/modem/sierrakeygen.py

Patching the firmware of 1199:9079 with a black-boxed nvu-file from a
Russian-speaking technical forum nvu-file to make it less specific, is
something that I would neither do, nor recommend. This might be a workaround,
but like Ed Maste I want to know how to reproduce the nvu-file/fixes on my own
- generally. I love the fact, that my 1199:9079 is not able to produce high
power signals (or even signals at all, so that ue0 gets an IP address for
example) without “permission”. Sure… it’s more an annoying “feature”, but much
more important it’s not a bug of FreeBSD and nothing which needs to be
discussed in this topic/thread here (no offense).

Yes, I know how to patch and build my own kernel… BUT you are speaking of a
single line of “code” (more harmless definitions than real code) in two files
for years now. I would love to see those lines within FreeBSD 13 sources in the
nearer future.

-- 
You are receiving this mail because:
You are the assignee for the bug.