Remaining actions on FIDO2 implementation on main MCU

[limpkin]

1. if it makes sense, not use a hmac based prng anymore... create our own one to pass to the ecc functions

• hopefully it should be as simple as tweaking hmac_drbg.c :)

2. check for buffer overflows for ecc related functions:

• for example br_ec_keygen returns the size written into the private key buffer, we should check it with the size of our buffer and while() in case of errors

3. use logic_encryption_delete_context function to do house cleaning

4. removed unused functions:

void logic_encryption_sha256_hmac_init(uint8_t const *key, uint32_t klen);
void logic_encryption_sha256_hmac_update(uint8_t const * data, size_t len);
void logic_encryption_sha256_hmac_final(uint8_t *hmac_out);

void logic_encryption_sha512_init(void);
void logic_encryption_sha512_update(uint8_t const * data, size_t len);
void logic_encryption_sha512_final(uint8_t *hash);

5. answer remaining comments of FIDO2 merge, especially regarding logic_encryption_ecc256_derive_public_key()

[0x0ptr]

1. Why reinvent the wheel? What does it gain us?
2. Agree
3. What do you want to clean and when?
4. Yes
5. I think I answered this in the pull request. Let me know if there is anything other than that.

[limpkin]

1. right now it's generating pseudo random bits and not real random bits
2. any context that may have been initialized (ecc perhaps? nto sure...)
   thanks again for the great PR :)

[limpkin]

2 has been implemented
3 is currently not required given line 335 of logic_encryption
4 has been done
5 has been done

1 remaining

[limpkin]

closing as 1. is a nice to have but not a must