* Format benchmarking code

* Update `*_less_then_signers` bench to scale signers in storage

Since we end up reading from this storage vec we want to make sure it is populated with up to
`MAX_SIGNERS` for the purposes of benchmarking.

* Change benchmarking for session handler to only use two benches

We add an extra storage read each time, but we can simplify the benchmarking a bit

* Always update session weight

If the signer length was more than the total signers we wouldn't
update the session weight, even though we would still do a few
storage operations.

* Remove `expect` by using `match` statement

* Remove unused `new_session_not_adding_new_signer` bench

* Update the name for one of the benches

* Update benchmark results

* Add missing parameter to base bench