Deposit contract assumes that controlAddress is unique for each validator

[JamesHinshelwood]

When calling depositTopup() or withdraw(), we look up the validator's BLS key based on the validator controlAddress. However, we do not enforce uniqueness of the controlAddress. So we should either:

1. Demand that each validator has a unique controlAddress.
2. Add a blsPubKey parameter to depositTopup and withdraw and look up the validator based on that instead (and this means we can probably remove the _stakerKeys map)

I would prefer 2 since it makes the deposit contract simpler and means validator pools can use a single wallet to control many validators.

[86667]

Agreed that 2. would be best, the _stakerKeys map is confusing.

I assume it's purpose was to link the caller (msg.sender) to the data acted upon (stakerKey), so we should still do this check. @DrZoltanFazekas is that correct, was there any other intended use for _stakerKeys?