Hi all, since hid2hci was moved to udev it bluetooth doesn't work anymore with Logitech dinovo. Now it is installed udev-151-r4,but I tried it with udev 164 which is in the instable tree with the same result Installing bluez 4.39 which has hid2hci incorporated bluetooth works. With newer versions the switch from hid to hci does not work anymore because the incorporated hid2hci does not exist anymore. A manual "/lib64/udev/hid2hci --method=logitech-hid --devpath=/dev/hidraw0" on newer version of bluez like 4.75 ends with "error: could not find '/dev/hidraw0". ls /dev/hidraw gives hidraw0 hidraw1, so the device exists. Thanks Reproducible: Always Steps to Reproduce: 1.connect a dinonvo bluetooth keyboard 2.install bluez-4.75 and udev151-r4 3./lib64/udev/hid2hci --method=logitech-hid --devpath=/dev/hidraw0 Expected Results: logitech swithes from hid to hci.
Chances are you've mixed up udev rules, cause if I'm reading the sources correctly, that udev version of hid2hci expects /sys path, not /dev.
Dear Rafal, maybe my description was a littlebit confuse. Fact is that in bluez newer than 4.39 the switching from hid to hci doesn't work anymore. I tried it with various combinations of udev 151 and 164 and bluez 4.69 and 4.75. So I think that or the udev rule or the hid2hci does not work properly. Do you have any idea what I can do? Thanks Roland
It would be nice, if you could check what exactly happens with your hardware, when you connect that keyboard, that is kernel logs and messages from 'udevadm monitor'. You need to see, if your hardware is covered by the rules shipped with udev - it may be something as simple as wrong product id. Don't forget to write the exact versions of udev and bluez you've used in your tests - those rules were modified a few times between March and July.
Created attachment 254331 [details] Log with versions of udev and bluez for udevadm monitor disconecting and connecting keyboard
Hi Rafal, excuse me that I write late. I attached you the requested file. I hope you can help me. Tanks Roland
Could you see if hid2hci from udev works if given correct /sys path ? You could check check it using 'udevadm info' on the hidraw device, just be sure to prefix it with /sys (output skips it). The values printed suggest that it should be covered by standard rules. Any chance that you've added some of your own, that circumvent it (or have some stale laying about) ?
