RBF: Final part #4399
RBF: Final part #4399
Commits on Mar 5, 2021
-
df-rbf: init an rbf for real, using openchannel_bumpfee
Reorg a bit of the RBF code so we use the same codepaths for we-init vs they-init starts.
-
-
df: update reserve calculation
We can use the funding amount to derive the reserve requirement.
-
df-rbf: only permit the channel initiator to init RBF
We're *mostly* set up for both sides doing RBF, except that it reverses the callback flow (using the plugin vs RPC calls) and we're not currently smart enough to flip between them gracefully
-
df-rbf: separate first-time channel save from update (RBF)
We can't "first save" a channel twice; instead we split in two and just update the underlying channel on subsequent passes (RBFs)
-
rbf: consolidate failure paths, use "warnings"
We move over to the new "warning" paradigm, instead of using an "rbf_fail" message. Every failure is either a warning or an error; on warnings we hang up and reconnect later, effectively resetting the state.
-
rbf: clean up channel on error
Make sure we clean up unsaved channels appropriately on failure. We forget the peer/channel if it's unsaved!
-
df: move methods around so we can call them elsewhere
Will reuse these elsewhere soon
-
df: handle funding_locked + tx_signatures messages out of order
If we're doing an RBF, it's possible that the peer will send us a funding_locked, shutdown, or tx_signatures message. (We get tx_sigs out of order on a reconnect) This lets us gracefully handle a shutdown or funding_locked sent at any time (after first funding tx) as well.
-
listpeers: include info on channels that are in-process
Channels that we're in negotiation for, but don't have a commitment transaction saved for yet.
-
channel: move inflight + feerate methods to channel.c
Need these more accessible for next patch, which moves the next_feerate info into listpeers
-
listpeers: include feerate info for RBF-candidate channels
Changelog-Added: JSON-RPC: `listpeers` now includes 'last_feerate', 'next_feerate', 'initial_feerate' and 'next_fee_step' for channels in state DUALOPEND_AWAITING_LOCKIN fixup! listpeers: include feerate info for RBF-candidate channels
-
df-rbf: update rejection criteria for tx_sigs
since we might be in the middle of an RBF, update our checks to be more robust
-
df: cleanup error handling on lightningd side
Make existing methods understand how unsaved channels work, re-work errors so that we handle everything appropriately
-
-
-
-
Revert "df: turn off integration tests that depend on v2 funding"
This reverts commit 2e40a13.
-
df-memleak: expose memleak error and fix
We were getting a memleak error that the open_attempt isnt' being cleaned up in test_rbf_reconnect_tx_construct. I had some trouble reproducing it, so I removed the reliance on using `tmpctx` to clean it up and was more surgical about cleaning it up inline.
-
df: add (over zealous?) note about the usage of
psbt_has_req_fieldsRequested-In-Part-By: Rusty Russell @rustyrussell
-
connectd: clean up the channel stuffs when we get a reconnect
If they've disconnected/reconnected we need to terminate all the inflight stuff, plus go ahead and call 'disconnect' plugin trigger etc.
