In here different byte padding standards are listed as ANSI X.923, ISO 101.26, PKCS7, ISO/IEC 7816-4 and zero padding. All of them seems to apply padding to the last byte blocks. Odds are rare but what happens if the actual data is an integer product of the block size and last bytes conflict with a padding scheme.
For example, for PKCS7 let last 8 bytes be
FF FF FF FF FF FF FF 01
01 may actually be original data or padding. How do we know this? How do we avoid a data loss? (It seems trivial but image formats may become corrupted, because original data size changes.)