C# P/Invoke supports UTF-16 and ASCII encodings out of the box. Mono supports UTF-8 but MS CLR does not. Here you can find typemaps for handling utf-8 strings portably.
I did not create a manual entry. There are examples and quite comprehensive documentation in the .i file. IMO it suffices if it is mentioned in the manual that also UTF-8 enconding is supported when passing strings to the native code and reader is referred to read the documentation in the utf8_strings.i file and the examples. This is to keep the manual from (further) bloat - UTF-8 strings are of interest only if you are using a native API that expects that encoding and I don't think they are all that common (not sure, though).