Comment 85 for bug 446657

(In reply to comment #1)
> The same problem on Fedora 13 with gnome-bluetooth-2.30.0-1.fc13.i686 on T61.
> Bastien, please, let us know if some other package might be causing this and
> you'd need more detailed version info, or if it is configurable somewhere.
>
> I am pretty sure that on Fedora 11 my bluetooth was off after reboot (and I had
> the option to turn it on with the applet on my panel). That's why I'd consider
> this a regression.

Definitely not a regression, if anything, it's a kernel one. The ibm-laptop (or whatever the kernel module is called for Thinkpads) module changed behaviour.

(In reply to comment #2)
> Funny thing: if I change in /etc/bluetooth/main.conf InitiallyPowered = true to
> false, and reboot, the applet will show that white x in red on its icon, to say
> that bluetooth is off. However, the bluetooth LED is still on, so in that case
> the applet shows incorrect information.

Nope. Check the output of "hciconfig hci0", you'll see that the device is enabled, just not powered. The hardware has no ways to know that...

> So maybe this is a bug in bluez (which owns that /etc/bluetooth/main.conf),
> that it does not honor the setting (and provides bad info to the applet as
> well)? But I don't see a component bluez in bugzilla to report it ...

This is a BIOS bug. gnome-bluetooth doesn't hold *any* information as to the state of Bluetooth, and bluetoothd (somewhat) remembers the state of the device. The problem is that when it's hard-blocked, eg. disabled in hardware, bluetoothd has no Bluetooth devices to remember the setting of.

With a good BIOS, the BIOS should remember the previous state, and use that.