We have the notations $L[A]$ for the inner-model constructible relative to some $A$, and the notation $M[G]$ for a generic extention of the model $M$. Do they coincide? That is, if we look at the constructible inner model $L$ and we have some $G$ which is a generic subset of some forcing notion $P\in L$, is the generic extention the same as the model constructed relative to $G$? (I think it is but I want to make sure...)
In particular, do we have a Condensation Lemma for $L[G]$?
Yes, the two things coincide. Simply note that if $G\subseteq L$, then $L[G]$ is the smallest model of $\sf ZFC$ such that $G$ is an element of it.
Generally speaking, $L[G]$ is the smallest model such that $G\cap L[G]\in L[G]$, but if $G\subseteq L$, then we get this automatically.
You can also think about this as $M[G]$ being a model for a language augmented by a predicate, and we interpret the names from $M$ to define this model. This is the same as adding a predicate to the language and considering the definable subsets in the case of forcing over $L$, simply because every name will appear at some point and we will interpret it "correctly".