Andreas Böhler <news@xxxxxxxxxxx> writes: > Unfortunately, it doesn't change anything. I applied the patch, > recompiled the kernel (it updated to 4.13.2 at the same time), rebooted > and upon connecting, I can see the 'cdc_mbim 1-6:1.0: nonzero urb status > received: -EPIPE' message again. Thanks for testing. Then I'll rule out that theory. I guess that's mostly good, since it didn't make much sense :-) > What makes it work for this session only is the following (also works > without your patch, though): > > * connect to /dev/ttyACM0 using a terminal emulator (PuTTY in my case) > * send AT+CFUN=15 to reset the modem > > Then, the device re-appears as /dev/cdc-wdm1 (instead of /dev/cdc-wdm0) > and connecting to the web is possible. If I then disable and re-enable > the modem (in this case, using NetworkManager/ModemManager), the nonzero > urb status message reappears. I can make it work again by resetting the > modem a second time (and stopping ModemManager before the reset, > starting it again after the reset). > > I can also provide you with Wireshark USB traces, if that helps? It might help. We are obviously interacting badly with the firmware. Seeing where and how it stops would be useful. What I don't understand is why the -EPIPE error is permanent. It is an indication from the firmware that there was no data to read (which I believe is a frmware bug, but whatever..). But the stall should be cleared on the next read. Unless the firmware decides to stop sending notifications too? Don't know why it would do that, though. I am just so lost here... Bjørn -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
