Good morning

When the computer starts up *without* the usb bluetooth adapter inserted, the [executable] rc.bluetooth is called but nothing further happens; no bluetooth services at all are started or left running. As far as I can tell from checking the script (-d /sys/class/bluetooth), it is because /sys/class/bluetooth directory does not exist at that moment.

Inserting the bluetooth adapter causes the /sys/class/bluetooth directory to be created (I presume from udev intervention ?).

Removing it *does not* appear to cause the directory to be deleted - why is that by the way ?

Call the rc.bluetooth script manually and now everything starts up and appears to leave all the services running (because the sys/class/bluetooth directory is present) even if the adapter is removed.

From then on, inserting and reinserting the bluetooth adapter works just fine.

The solution then, to the bluetooth services not starting up at boot time, is to ensure the adapter is in the machine at boot time.

Ideally, however, it would be better that the bluetooth services would start at boot and the action of plugging in the adapter at some point later would be picked up. This is not the case it would appear.

It would appear that the /sys/class/bluetooth directory is fundamental to this 'issue' and because the action of removing the adapter appears to not cause the directory to be removed (and the services can still run), is it fundamental for the rc.bluetooth script to check for its existence at all ?

Thank you for comments, info, insight etc.

Peace and best wishes

ok - looking at the matter further, in the /etc/rc.d/rc.bluetooth, the key line as mentioned earlier appears to be:
"if [ -d /sys/class/bluetooth ]; then..."

All the services are called within this condition block. With the fact that /sys/class/bluetooth does not exist at boot time (unless bluetooth devices are detected at boot), no services will be started, meaning that when a USB adapter is plugged in for instance (with the associated created on the bluetooth dir in /sys/class), the rc.bluetooth script will have to be run again but manually before it will all work.

From that time on, all will be fine until a reboot.

I took the opportunity to comment out the above 'if' condition and just leave the contained lines to run. It seemed to sort everything out.

Therefore I am left asking for confirmation if that 'if' condition is required or not; is there some fundamental reason for checking for the bluetooth directory before the other lines of the script (the bluetooth services startup) are actioned.

If anyone can help or point to where to ask why this rc script was set up with this check, please let me know.

It does not appear that starting the bluetooth services without the existence of that directory at startup causes any issues, *but* I am not sure !

2 members found this post helpful.