Depends on the message length. If the length is 1 byte (2 bytes, 3 bytes, ...) - it is trivial to do an exhaustive search to find the message that maps into the known hash. The longer the message - the harder it gets.
I agree it's (quite) trivial for 1,2,3 bytes. Not doing the reverse of a hash, but simply having the possibility to calculate 256/65536/16777216 hashes and compare for a match. The good old and simple brute force attack.