Implement recent spec changes regarding collab channel close outputs

[bitromortac]

Fixes #7541

• Implements option_shutdown_anysegwit, which lets peers close their channels to taproot scripts.
• Constrains the allowed shutdown scriptpubkeys to only segwit types.
• The dust limit is negotiated at channel opening. We currently reject dust limits that are lower than 546 sat, which leads to Dust limit too low when opening channels with latest lnd #7541

Fulfills spec changes:

• BOLT 2: option_shutdown_anysegwit (Feature 26/27) lightning/bolts#672
• Peers need to check each other's dust limit lightning/bolts#894

Related:

• Allowing dust limit below 546 sats lightning/bolts#905
• BOLT 2: peers need to check each other's dust limit lightning/bolts#696
• multi: replace DefaultDustLimit with accurate dust limit lightningnetwork/lnd#5781
• Configure dust in flight threshold ACINQ/eclair#1985

Todo:

• decide on a maximally accepted dust limit from the remote party
• should we keep our dust limit at 546 sat?
• decide on whether to force close upon dust limit violation at channel closure

• a further improvement in the security when forwarding is to apply a limit on the allowed total value of dust HTLCs

[SomberNight]

decide on a maximally accepted dust limit from the remote party

I guess 3000 sat as in current PR is unreasonable.
I would prefer an even lower number though.
Not sure what other implementations use; rust-lightning caps it at 660 sat.
Maybe use 1000 sat?

should we keep our dust limit at 546 sat?

Yes, let's keep that for now. We know from experience that others accept it during channel open, and I think ~500 sat outputs are hardly ever worth sweeping anyway (so not much point in lowering it IMO).

decide on whether to force close upon dust limit violation at channel closure

Hmm. Well I guess it's not critical, as the user can force-close manually if needed.
But yeah the many "dust limits" are becoming a bit confusing, seems to me we are not using the correct one either:

electrum/electrum/lnchannel.py

Line 1452 in 1ff9f99

dust_limit_sat=self.config[LOCAL].dust_limit_sat)

it's ~harmless though.

[bitromortac]

Agree with your other points, thanks.

Hmm. Well I guess it's not critical, as the user can force-close manually if needed. But yeah the many "dust limits" are becoming a bit confusing, seems to me we are not using the correct one either:

electrum/electrum/lnchannel.py

Line 1452 in 1ff9f99

dust_limit_sat=self.config[LOCAL].dust_limit_sat)

it's ~harmless though.

I think so, yes, very confusing 😕.

[SomberNight]

Looks good. Thanks!

EDIT: ah, one final thing; see comment below.