Hi,
[deadline: September 23]
Currently one can't use Bluetooth in Tails without going through
manual operations that 1. are not documented; 2. can only be performed
in the GNOME session (which implies that e.g. one can't use
a Bluetooth keyboard in the Greeter anyway). So in short, we support
Bluetooth usage neither officially, nor effectively (except for
technical users who want to use things like Bluetooth speakers and
will manage to make it work).
But Bluetooth is not fully disabled at the lower level, so our users
are exposed to security vulnerabilities in the Linux Bluetooth stack
such as BlueBorne (
https://www.armis.com/blueborne/), which forced me
to think about this topic urgently today.
I think this current trade-off is pretty bad for most Tails users: it
makes them vulnerable, without providing them anything useful
in exchange.
I propose we:
1. Short-term: fully disable Bluetooth (#14655, I'll probably get my
branch merged today so 3.2~rc1 is safe vs. BlueBorne, but we can
revert this change in 3.2 final if it's not OK). As explained
above, this should not impact UX much.
2. Mid-term: document how to re-enable Bluetooth in the GNOME session
and with Additional Software Packages (see commits 4fa1c46,
613b14c, 3d53d2c and 9320ca7 for inspiration).
3. Long-term: better support Bluetooth (e.g. #10801 + persistence of
Greeter settings).
Don't get me wrong! I'm convinced that at some point, we do need to
support Bluetooth properly: our current userbase might be dedicated
and motivated enough to deal with limitations like "I can't use my
Bluetooth speakers and there's no wired speakers around so well I'll
watch this video with my crappy laptop integrated speakers", but if we
want Tails to have a greater impact we should probably make it work
nicely for people who won't want to use a tool that breaks such common
use cases.
Cheers,
--
intrigeri
