AES cleanup fixes pt 1 for #179 #210
Conversation
There was a problem hiding this comment.
This looks solid, thank you for addressing! I appreciate the coherent commits.
I have a few comments; the most important ones are about docs and naming in SBox -- the rest are mostly nitpicky.
One other top-level comments: please check off the boxes for the things you finished in #179!
| sbox : SBox | ||
| sbox = [ transform (GF28::inverse b) | b <- [0 .. 255] ] where | ||
| sboxTable : SBox | ||
| sboxTable = [ transform (GF28::inverse b) | b <- [0 .. 255] ] where |
There was a problem hiding this comment.
todo: Since this table is not used anywhere except the sbox function, let's make it private. Also, please update the docs on this to point to Table 4 (my bad) and mention that it's pulled out here for efficiency (e.g. this lets us compute the table once for every access in AES, instead of recomputing every time).
Same comment on the inverse below.
| // Equation 5.3. | ||
| transform b = GF28::add [b, (b >>> 4), (b >>> 5), (b >>> 6), (b >>> 7), c] | ||
| // The constant byte {01100011}. | ||
| c = 0x63 | ||
| sbox : GF28 -> GF28 |
There was a problem hiding this comment.
todo: this function needs docs. You can use the description that's currently on sboxTable.
Ditto for inverse.
Three of the tasks from issue 179, cleaning up small issues to do with the AES spec.