Description of problem: I've got a Huawei E3372 hostless modem, which worked fine in F22. With F24 it's stuck in Storage-Mode after plugging in. No udev/systemd activity shown in system-journal. Switching manually via usb_modeswitch works. Matching config-data is present. So my guess is to blame the usb_modeswitch udev-rules. Likely the same issue here (E3372 + E1800): https://bugs.launchpad.net/ubuntu/+source/usb-modeswitch/+bug/1562431 Version-Release number of selected component (if applicable): usb_modeswitch-2.3.0-1.fc24 Additional info: My old "classic" broadband modem Huawei E169 works fine. There is a patch included in usb_modeswitch 2.4.0, which improves systemd-integration. The commit and the forum post, which are linked in the changelog, look like a similar issue. http://www.draisberghof.de/usb_modeswitch/ChangeLog So suggestion is to try 2.4.0 in F24 testing (probably without karma based stable).
/usr/lib/udev/rules.d/40-usb_modeswitch.rules # Huawei E3372 ATTR{idVendor}=="12d1", ATTR{idProduct}=="14fe", RUN+="usb_modeswitch '%b/%k'"
To be precise: Yes, the specific rule is present and worked fine in F22 with systemd 219. But there is a regression with later systemd versions as shipped with F24. This is analyzed and confirmed by the developer in the upstream forum post (based on Archlinux with systemd 221): http://www.draisberghof.de/usb_modeswitch/bb/viewtopic.php?p=16777 /usr/lib/udev/usb_modeswitch was changed in 2.4.0 to fix this. So my suggestion was, that we could try 2.4.0, which is supposed to fix the regression.
I tried 2.4.0 from F25, but sadly no change in behavior. So my bug is most likely another issue.
(In reply to Peter Kovář from comment #1) > /usr/lib/udev/rules.d/40-usb_modeswitch.rules > > # Huawei E3372 > ATTR{idVendor}=="12d1", ATTR{idProduct}=="14fe", RUN+="usb_modeswitch > '%b/%k'" Now it's solved. Your advice was good, but I had to change the Id (my fault, as I have the correct Id in systemd-journal): # Huawei E3372 ATTR{idVendor}=="12d1", ATTR{idProduct}=="1f01", RUN+="usb_modeswitch '%b/%k'" I wonder why this worked out of box in F22...