Lower stored MAC state data

[adriansmares]

Summary

Closes #5127

Changes

• As part of the MACState message, add minimized versions of UplinkMessage, DownlinkMessage, Message, TxSettings and RxMetadata
  • The scope here is to ensure that fields which are not going to be available from the 'real' UplinkMessage/DownlinkMessage are not accidentally used (i.e. using zero values as if they were intentional)
• Use the minimized versions as part of the Network Server MAC state handling
• Fix LoRaWAN 1.1 AFCntDown sanity check
  • The old check did not take into account the fact that the frame header frame counter is 16 bit, and instead the full (32bit) frame counter should be used for checks

Testing

Unit testing.

Regressions

The code that actually uses these fields is unchanged in this PR - if a field was skipped in the minimal representation, it would result in a compile error. The only error prone operation is the conversion between the normal message and the minimal message, which I've double checked and also covered via existing unit tests.

Checklist

• Scope: The referenced issue is addressed, there are no unrelated changes.
• Compatibility: The changes are backwards compatible with existing API, storage, configuration and CLI, according to the compatibility commitments in README.md for the chosen target branch.
• Documentation: Relevant documentation is added or updated.
• Changelog: Significant features, behavior changes, deprecations and fixes are added to CHANGELOG.md.
• Commits: Commit messages follow guidelines in CONTRIBUTING.md, there are no fixup commits left.

[johanstokking]

I would also be fine with keep using the original messages but define a field mask for storage.

Code quality wise I don't think it's necessary to declare all fields reserved with the original message's field name as they diverge. If it's just to make sure that our future selves don't add fields with the same numbers and old binary data gets unmarshaled and lead to corrupted data, we can just declare the reservations in short form.

[adriansmares]

I've minimised the reserved by removing the comments and joining the statements.

Regarding the field mask - The field mask tooling does not support applying field masks to repeated fields - see TheThingsIndustries/protoc-gen-fieldmask#37 , and a lot of the dropped fields are in rx_metadata.

Offtopic: We probably should fix that before we add field masks to webhooks, since most of the filtering that I expect will be in the rx_metadata of ApplicationUp.