I would say 10. ..0-0? instead of ?!. The modern Petroff (3. d4) is a very fun opening for both sides. It gives each the most winning chances, in my opinion.
Petroff revisited
Sort:
This variation is really fun to play with white against the petroff because basically the only thing black can do is to defend and white gets all the fun attacking. 
True, but I would play 6. ...d5 as Black to grab some space. Also, trading pieces immediately as White in the opening can dampen the tactical possibilities down the line. Still, it's nice food for thought.
The idea with 6. d5 is good but I think its a bit to slow here.
White can for example try to take advantage of this immediatly by playing 7. c4!?
Defending the pawn with 7...c6 is no good because white can play 8. cxd5 Qxd5 9. Qxd5 with a comfortable advantage thanks to blacks isolated d-pawn.
7... Be6 seems to run into problem after 8. Ng5! when white manages to grab the bishop pair.
5. d5? is actually also a mistake because white can play 6. Qe2! which threatnes to win a pawn and black cant defend it in a good way. Black doesnt have much options besides going into the main lines in these positions.
Ok. But what about 5...f5?! or the quiet 5...Nf6? Even 5..Qe7?! looks acceptable.
5... Nf6 is terrible, its exactly like playing 1. e4 e5 2. Nf3 Nf6 3. Nxe5 d6 4. Nf3 Nxe4 5. d4 Nf6?!
5... Qe7?? is actually loosing immediately because of 6. Nd5!! When black moves the queen whites gonna play Qe2 winning a piece.
5... f5? has the same problem as 5... d5?
Cymantex, although it takes longer, what's wrong with black trying to castle queenside? It would seem to avoid attacks. Black also has the option of developing the bishop to g6.
Trying to castle queenside is actually blacks best side line, many grandmasters has used that strategy in the past. Although it seems like white can still get a pretty decent advantage and therefore most grandmasters stoped playing the idea with long castling. I recomend that you watch GM Ronen Har-Zvi's guide on chessfm to learn more about this.
Yeah white probably still has the edge because of his easier development and open lines (like the d file) while black struglles a little more with development. maybe black could in the long run try for ...d5 to gain space but even if it could could succeed he still wouldn't really have any targets to aim at. But I figure long castling avoids more attacks than short castling does.
I don't see 5...Nf6 as being "terrible". In the Classical Petroff it is common for Black to play the Knight back to f6.
...Nf6 is poor because black is moving a piece three times and thus getting behind in development by two pieces for no reason.
Yes, but Black can still play a solid position. 5...Nf6 6. d4 d5 7. Be2 Be7 8. 0-0 0-0
Doesn't look fun. He faces risk for attack and so he will need a tough defense, which will probably give him no more than a draw.
True, it is a drawish opening, especially that line.
The line with Nf6 maybe is a draw with the best play from the black side but white will definitely have a clear advantage thanks to all the tempis black has given white. White could probably use all those tempis to still play for long castling and therefore black doesn't really gain anything by playing this way.
Here is a recent top level Petroff game played between Vugar Gashimov and Boris Gelfand. I do not take credit for the annotations. Those come from our friends at chessbase:)