You mean The Post by AndrewZ?
I don't think ModemManager supports this modem in the given configuration (USB+RNDIS).
What you can probably do right now is remove MM, manually initiate connection from a terminal and read IP address(es), then manually configure the interface with the IP obtained from the modem.
I'm a novice linux user and this does not help much...
then your in way over your head, unless @AndrewZ will elaborate.
well for starters I could try to manually configure the interface, without automation. But how?
this sounds promising. It's bit over my head, so correct me if I'm wrong:https://gitlab.freedesktop.org/mobile-broadband/ModemManager/-/tags
I can't tell you here. I found information on the Internet, but I did not test it myself. I have OpenWRT and the modem works fine in it.
what did you do to get it working? I see some people have it working, but can't find the steps they took.
Hey, I've posted earlier what you need to do to get this to work.
Connectconnect using these AT commands:
AT+CGDCONT=1,"IP","mobile.o2.co.uk"
AT+CGACT=1,1
Then to get the IP address:
AT+CGPADDR=1
I then set the IP manually and the gateway to "x.x.x.1" with subnet of /24 and the connection works.
Please give it a try if you have the module and let me know how it works
In this repo @Prusa pr for FM350GL luci-proto and modeminfo, but it got bug on auth.
At first, I wrote my own script to automate the process, but the Prusa solution is much more convenient and elegant.
Yes I do this and I get IP address, but what next, step by step? It does not appear as network interface, I only get many virtual com ports, ttyUSB0...ttyUSB8... But nothing else, or I could not find.
The @Prusa 's luci-proto works for me, but I had to remove line 17 send "AT+CGPADDR=1^m" from the etc/gcom/fm350-connect.gcom file.
I also added a service /etc/init.d/fm350:
#!/bin/sh /etc/rc.common
START=99
start() {
echo "0e8d 7126" > /sys/bus/usb-serial/drivers/generic/new_id
echo "0e8d 7127" > /sys/bus/usb-serial/drivers/generic/new_id
}
DroZDi
March 11, 2024, 10:31pm
57
In my repo i put some step by step instruction plus some scripts.
One general remark:
DroZDi
March 14, 2024, 11:40am
58
Erratum:
DroZDi
March 17, 2024, 12:42pm
60
@daniel
After last clock fix on pcie2 modem fm350gl stopped to work in USB mode and now it appears like:
I wish to know if there is any chance to have a control on it (how we can revert such clock fix if not used - in my case before it was not used and in the slot modem fm350gl was working ok? otherwise I dont know of we can go with trying to fix the issue on pcie communication here. Probably we need a kmod_t7xx (https://cateee.net/lkddb/web-lkddb/MTK_T7XX.html ) but it seems now it is not yet available in openwrt?
============================================
Per now I found only way to come back as revert the patch:
in target/linux/mediatek/patches-6.x/252-clk-mediatek-mt7988-infracfg-fix-clocks-for-2nd-PCIe.patch (line 159):
and recompile the image.
daniel
March 17, 2024, 10:42pm
61
It's easy to package the driver to use the modem in PCIe mode which will be better and faster than using USB:
openwrt:main ← dangowrt:mtk-t7xx
opened 10:41PM - 17 Mar 24 UTC
Build and package driver for MediaTek PCIe 5G WWAN modem T7xx device available i… n Linux 6.1 and 6.6 as kmod-mtk-t7xx.
2 Likes
DroZDi
March 18, 2024, 2:54pm
63
Huston, Huston....
[ 12.686678] Call trace:
reboot
[ 14.527729] Call trace:
PANIC at PC : 0x000000004300490c
[ 74.711006] mtk_t7xx 0003:01:00.0: MD handshake timeout
This modem is a bit chimerical and needs a long time to reboot or setup...
daniel
March 18, 2024, 5:45pm
64
(Always) complete logs please, the interesting part just above is missing.
DroZDi
March 18, 2024, 7:02pm
65
I will try to compile to the kernel module and next load manually after the system starts but it fails too:
root@BPI-R4:/nvme/work# opkg install kmod-mtk-t7xx_6.6.22-1_aarch64_cortex-a53.i#1 ] SMP#0
DroZDi
March 20, 2024, 8:32am
66
I have replaced build_dir/.../t7xx directory with (newer one?) from https://github.com/torvalds/linux/blob/master/drivers/net/wwan/t7xx
compiled and .... the same.... KERNEL PANIC
@daniel any idea?
daniel
March 20, 2024, 11:30am
67
Using driver from Linux git HEAD would also have been the first thing I'd have tried. Sad that it doesn't solve it, so the driver is buggy or something with PCIe MMIO ranges is borked.
From first glance it looks like the driver is accessing an address in the PCIe MMIO range in 32-bit alignment (ffffffc084a1d004 ) but likely the SoC only supports 64-bit aligned (so only addresses ending on 0 or 8 will work) access there, hence the [ 294.051349] FSC = 0x21: alignment fault.
I've ordered the hardware and will have it in a week or two,
good is that modem is ready to work with commands to re-init 
