git log implementation details

Hello,

I was wondering how the command "git log" is actually retrieving the
commit log for a given file behind the scene.
Is it by walking down the object tree and scanning each commit/tree
object? Or any cache/index used here?

Thank you,

Nicolas