> It is evident from the history of this bug that the maintainers don't > agree with this as an approach. >
I wonder how that is evident given that we (maintainers) still keep the bug open upstream and discuss about it from time to time... https://bugs.freedesktop.org/show_bug.cgi?id=85007 I personally acknowledge the problem, but so far I've failed to find time to think of what the proper solution would be... A full whitelist is something we don't really consider, given the large amount of devices coming out all over the world, and the lack of time we already have. Asking the user... well, what would we ask? Should we ask "is this a modem" for every TTY we find? Is it better to annoy thousands of users each time a TTY is found instead of blacklist for all of them? We try to keep the blacklist in stable releases updated and stables happening each 2-3 months. Plus, there isn't always a "user" to ask when a modem is plugged in, ModemManager (as NetworkManager) don't really require a GUI to work and lots of headless systems out there use it. One thing to maybe consider is whether we should just assume that it is a modem if: * We have more than one TTY exposed by the same USB device (are there non-modem devices out there that would usually do this?). * We have a TTY and a NET port exposed by the same USB device. Or in other words, don't autoprobe a TTY if the USB device only has that one single TTY... Less and less mobile broadband modems fall in that last case, and definitely none of the recent modems I've seen would fall into the single-TTY category (except for maybe "professional" ruggerized devices and such but sometimes they're not even USB but pure RS232, and those aren't probed automatically anyway). Maintaining a whitelist of single-TTY modems may be feasible... Blindly requiring a udev tag for a TTY to be probed during automatic scans will break all non-QMI non-MBIM modems :/ So yeah, totally open to discussion here. -- Aleksander
