We have an issue with EM7565. After reducing MTU size on the wwan0 interface it stops working after some time.
mmcli -b 1 reports that mtu in the network is 1430
Observed behavior:
If we set gsm.mtu in the network manager connection to a smaller value, for example, 1300 and reboot, we can’t receive packets anymore (mmcli -b 1 and ip -s -s link show dev wwan0 show that RX bytes doesn’t change, but TX grows).
RX: bytes packets errors dropped missed mcast
1299304 3216 345698 0 0 0
RX errors: length crc frame fifo overrun
0 0 0 0 1
TX: bytes packets errors dropped carrier collsns
1121548 7267 0 0 0 0
TX errors: aborted fifo window heartbeat transns
0 0 0 0 0
Statistics | start date: 2025-01-15T19:03:12Z
| duration: 1499
| bytes rx: 1303220 - constant
| bytes tx: 1120963 - increases
| attempts: 1
| total-duration: 1499
| total-bytes rx: 1303220 - constant
| total-bytes tx: 1120963 - increases
If we leave gsm.mtu = auto, then after it connects to network we change mtu with ip link set dev wwan0 mtu 1300 it keeps working until a packet larger than defined MTU arrives - can be simulated by sending ping to the device:
ping cellular-ip -M do -s 1300 -O
Then if we check ip -s -s link show dev wwan0 output, we can see that overrun in RX becomes 1, and each received packet after that falls in errors.
On the current setup we have Debian 11 with qcerrial and qmi_wwan drivers.
In the area where device operates reducing MTU is necessary to avoid packet loss while communicating with certain services.
We also have devices with Debian 9 and GobiNet driver in the same area, where it works well with reduced MTU, unfortunately, we can’t continue using the GobiNet driver.
mmcli -m 0
----------------------------------
General | path: /org/freedesktop/ModemManager1/Modem/0
| device id: 21ddc41241c8d2fafa61d6a29fa3dccc12f2b590
----------------------------------
Hardware | manufacturer: Sierra Wireless, Incorporated
| model: EM7565
| firmware revision: SWI9X50C_01.14.22.00 ba8ef1 jenkins 2022/11/11 09:59:48
| carrier config: default
| h/w revision: 1.0
| supported: gsm-umts, lte
| current: gsm-umts, lte
| equipment id: 353533102845297
----------------------------------
System | device: /sys/devices/pci0000:00/0000:00:15.0/usb2/2-6
| physdev: /sys/devices/pci0000:00/0000:00:15.0/usb2/2-6
| drivers: qcserial, qmi_wwan
| plugin: sierra
| primary port: cdc-wdm0
| ports: cdc-wdm0 (qmi), ttyUSB0 (ignored), ttyUSB1 (gps),
| ttyUSB2 (at), wwan0 (net)
----------------------------------
Status | lock: sim-pin2
| unlock retries: sim-pin (3), sim-puk (10), sim-pin2 (3), sim-puk2 (10)
| state: connected
| power state: on
| access tech: lte
| signal quality: 100% (recent)
Following the suggestion from this issue: MC7430 struck , we tried switching modem to USB 2 mode, which seems to be resolving the issue and modem keeps working with smaller MTU, however, RX overrun packets and errors are significant (10-50%), same as described in the issue, and with so many packets lost it doesn’t seem like a good way of solving this.
We also upgraded unit to the latest firmware SWI9X50C_01.14.22.00.
Please advise.