side channel attack deterrence in nonce generation in secp256k1 ecdsa signing

[psingh-io]

I have posted this question on decred support, re-positing it here for additional visibility since bcd uses decred`s implementation.

I have a question about the nonce generation (k-value) during the secp256k1 ecdsa signing process. The dcrd implementation implements RFC6979 for deterministic nonce gneration. The spec mentions that the implementations shall take defensive measures to deter side channel attacks to leak private key (section 4, last para):
https://datatracker.ietf.org/doc/html/rfc6979#page-15:~:text=Side%2Dchannel%20attacks%20are%20an%20important%20consideration

The dcrd nonce implementation appears to account for that by using the extra and version fields (https://github.com/decred/dcrd/blob/master/dcrec/secp256k1/nonce.go#L117). However, the signing operation passes nil for both the values(https://github.com/decred/dcrd/blob/master/dcrec/secp256k1/ecdsa/signature.go#L713).

Bitcion's libsecp256k1 uses secp256k1_context_randomize funciton to deter against the side channel attacks:
https://github.com/bitcoin-core/secp256k1/blob/master/include/secp256k1.h#L849

My questions are:

1. What is the purpose of extra and version fields in secp256k1.NonceRFC6979 function call?
2. Can the above fields be used as an additional deterrence for side-channel attacks?
3. Does inclusion of message hash in the nonce calculation offer any deterrence for side-channel attacks?
4. Is it a weakness from side channel attack perspective to not pass the extra fields from the secp256k1.NonceRFC6979 call? If not, why?