multi: make legacy feature bits compulsory

[ProofOfKeags]

Change Description

This PR makes some legacy feature bits compulsory. Closes #8048

• var_onion_optin
• gossip_queries
• option_data_loss_protect
• option_static_remote_key

Steps to Test

Steps for reviewers to follow to test the change.

Pull Request Checklist

Testing

• Your PR passes all CI checks.
• Tests covering the positive and negative (error paths) are included.
• Bug fixes contain tests triggering the bug to prevent regressions.

Code Style and Documentation

• The change obeys the Code Documentation and Commenting guidelines, and lines wrap at 80.
• Commits follow the Ideal Git Commit Structure.
• Any new logging statements use an appropriate subsystem and logging level.
• Any new lncli commands have appropriate tags in the comments for the rpc in the proto file.
• There is a change description in the release notes, or [skip ci] in the commit message for small changes.

📝 Please see our Contribution Guidelines for further guidance.

[guggero]

Did a first pass, looks pretty good!
The linter/formatter CI step is failing due to a missing make fmt run and the itests fail because of an itest case that needs to be removed. Other than that I think this is quite close already.

[lightninglabs-deploy]

@ProofOfKeags, remember to re-request review from reviewers when ready

[ellemouton]

Looks good so far!

My main comment is regarding removing any logic that switches on LegacyPayload.

Something I noticed in testing btw is the following:
If we have Alice running with a required bit and Bob with no knowledge of that bit, then if Alice tries to connect to Bob, the connection fails (which we expect) but if Bob tries to connect to Alice then that connection succeeds.... I think this might be incorrect behaviour. Not relevant to this PR though - just mentioning.

Just for future: it's generally good practice to make sure the CI is green before re-requesting review to help reduce the number of rounds :) There is currently a panic in the unit tests and the itests are all failing

[ellemouton]

Did some testing with Alice running this PR and Bob running master with --protocol.legacy.onion set and found that if Bob creates an invoice the Alice will refuse to pay it due to the face that Bob doesnt support the TLV onion option. This is expected but I think that means that we can remove a bunch of logic that switches on route.Hop.LegacyPayload:

lnd/routing/route/route.go

Lines 133 to 136 in 708bd05

	// LegacyPayload if true, then this signals that this node doesn't
	// understand the new TLV payload, so we must instead use the legacy
	// payload.
	LegacyPayload bool

[ProofOfKeags]

Paging @Roasbeef to weigh in on ripping this code out.

[Roasbeef]

Have a slight pref for just kinda leaving it in, and also making sure the legacy protocol cfg stuff still works as well to enable, this way less sweeping itest changes are needed.

So I'm thinking:

• flip to required
• wait
• observe
• ???
• "enough time passes"
• rip out the code at the same time we start to repurpose the feature bits

[guggero]

Two questions around interpreting @Roasbeef's comment re leaving the command line option or not.

Other than that LGTM 🎉

[ellemouton]

LGTM! 🔥

[ProofOfKeags]

SNED 🚀

[RandyMcMillan]

ACK 08c18a3

(post merge)