It's the setup I'm using now.
If not, you're subject to a "black bag" job that could easily bypass all your protections, e.g. a physical key logger that would pick up your password/pass phrase that unlocks your disk encryption so you can use it.
Otherwise, do you surf the net from it? I don't take all your precautions, but I do run my browsers in a dedicated VM that has limited access to the rest of my infrastructure.
Pretty much. It's a single user MacBook Air, and I carry it with me at all times, never leaving it out or in the open, or even at home. For suspicious individuals, the CIA can obtain a warrant without notifying the individual, and search his/her apartment/home at convenient times when away, so I never leave it at home, either.
It's pretty much the only computing device I have that has highly sensitive data on it. I don't really care if they take my Kindle Fire; the only stuff on that is books and personal email.
It doesn't really matter, because I've also set up a randomizing encryption passcode generator that only works with a synthetic amino acid solution that corresponds to my genome that I carry in a vial around my neck. If under duress, all I have to do is break the vial, and the system will be impenetrable. I also have another sample of the synthesis in my home, but government people can only access it with a warrant -- which they probably won't have.
I'm not at all sure Linux is the solution; the only plausible culprit for whom that would be a barrier is a non-state actor doing a modern variety of "Will no one rid me of this turbulent priest" (http://en.wikipedia.org/wiki/Thomas_Becket). But the reported sophistication of the attack makes that less likely that previously speculated, and I assume that "Linux" is not a high enough barrier to a state actor.