I dont know what the theory is supposed to be, but yes 7.Bc4 makes sense in terms of general principles - since ...h5 seems specifically designed against the g3 plan it seems possible to switch plans with simple development and try to make the move ..h5 look silly.

gxf6 Burn in the french IS interesting, but u have to accept playing some other variations as well...

Some (maybe not very good...) reasons for selecting the CK Larsen instead of french burn gxf6 variation:
-french exchange is less tempting to me than the "CK exchange"
-french tarrash
-Even if white plays 3.Nc3 Nf6 he can ruin my plans with 4.e5

-> Of course there are some downsides to the CK as well....

6...h5!? I suppose is consistently provocative! 
In fairness I recall Seirewan playing a game with this one time, and it does cause a bit of havoc - but I suspect someone will get back to us with the way to take advantage of the weaknesses behind the move (no piece developed, pawn potentially a weakness later)  but if you insist on playing this opening it could be one of the lines to look at!?

Is that not one of the lines where the Ng5 trick kills Black? 12.Ng5 what happens

Yes, well for a start I think 0-0-0 is far too adventurous in that position for Black - he must accept that he is slightly worse and keep his king safe for the time being. It is not a position though I must admit that I would particularly want to be playing as Black

6.c3 Bf5 7.Nf3 Nd7 8.g3 Nb6!? 9.Bg2 Qd7 10.0-0 Bh3!? (seems most logical)
Then 11.Bxh3 Qxh3 12.a4! intending to kick the Knight around as 12...a5 13.Qb3 is awkward. This seems very nice for White, but at least it is not a line where Black gets mated and he can grovell on....

Viking, seems strange to write in English to a fellow Norwegian, but for the benefit of other readers, here goes: 6. c3 Nd7 is very interesting. According to chessgames.com, the statistics after c3 are:

6...Bf5  119 games  51.3% - 28.6% - 20.2% 
6...Nd7  14 games  35.7% - 21.4% - 42.9% 
6...Qd5  10 games  50% - 40% - 10% 
6...h5  4 games 50% - 25% - 25% 
 
Basically, c3/Bg2 is a strong plan so alternatives to Bf5 are most welcome. Black's plus score after Nd7 is probably mainly down to Turov, cf http://www.chessgames.com/perl/chess.pl?pid=49029&playercomp=black&eco=B16. ;

I like Turov's Nd7 - b6 and sometimes to d5 to attack Bf4. Plus his Bf5/g4, Qd7 and Bh3 plan to exchange off white's g2 bishop, which is a strong piece both in attack and defense. Btw, did you notice that Turov also has played 6.Qd3 as white. 
 
 

Viking, Neil McDonald: Caro-Kann Main Line, p.120, discusses your 6...Qd5: "An unusual alternative for black is 6...Qd5!? The idea is to deter White from financhettoing his bishop after say Nf3 Bg4 (or perhaps 7...Rg8 8. g3 Qe4+!?). Instead 7. Qb3!? is an interesting riposte, e.g. 7...Qe4+ 8. Be3 Bh6 9. Kd2 Bxe3+ 10. fxe3 b6 11. Ne2 Be6 12. Ng3! Bxb3 13. Nxe4 Be6 14. Ng3 h5! 15. Be2 h4 16. Nh5 Rg8 17. Nf4 Bg4 18. Bd3 e5 19. Ne2 Ke7 20. Rhg1 Nd7 21. Be4 Rac8 22. Taf1 Rg5 and although White eventually won in Turov-Eliseev, St Petersburg 1998, I think that Black is doing okay here."