multi-precision

Sorry for not submitting a pull request, I'm working on a bignum library and I saw your implementation of multiplication so I just want to point out that you can simplify it by doing the multiply and the add in one loop into the result bn as follow:

typedef unsigned int           u32;
typedef unsigned long long int u64;
void _bn_add_mul_u32(u32 *out, u32 *a, size_t a_len, u32 b)
{

Seen with @arnetheduck.

for uint128, stint properly generates add + adc (see status-im/nim-stint#10)

But for uint256+, the way carry is done will loop multiple time on the low half
https://github.com/status-im/nim-stint/blob/9993b9dca4a2dbf76bc432252ad4e8db944ae831/stint/private/uint_addsub.nim#L17-L29

Important for WASM codesize optimization.

Current CMAC implementation assumens a block cipher with 16 bytes block size.

This can be adjusted to take the cipher block size during initialization