-
Notifications
You must be signed in to change notification settings - Fork 922
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Splice: Interop Final (probably) with Eclair #8021
base: master
Are you sure you want to change the base?
Splice: Interop Final (probably) with Eclair #8021
Conversation
@remyers reported this fix made his splice interop test pass! 🎉🚀 |
e86bfbc
to
facfeeb
Compare
29d52d1
to
45ae772
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
very cool!
@@ -200,6 +200,7 @@ new_inflight(struct channel *channel, | |||
|
|||
inflight->i_am_initiator = i_am_initiator; | |||
inflight->force_sign_first = force_sign_first; | |||
inflight->is_locked = false; | |||
inflight->splice_locked_memonly = false; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
it'd be great to add this to wallet/tests/run-wallet.c
for the inflights
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added
@@ -3829,7 +3829,7 @@ static void splice_accepter(struct peer *peer, const u8 *inmsg) | |||
new_inflight->remote_funding = peer->splicing->remote_funding_pubkey; | |||
new_inflight->outpoint = outpoint; | |||
new_inflight->amnt = both_amount; | |||
new_inflight->psbt = tal_steal(new_inflight, ictx->current_psbt); | |||
new_inflight->psbt = clone_psbt(new_inflight, ictx->current_psbt); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
yeah tal_steal and psbt code aren't good friends
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
No more theft among psbt 🙅
|
||
if (!fromwire_splice_locked(msg, &chanid)) | ||
if (!fromwire_splice_locked(msg, &chanid, &splice_txid)) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
we're not checking it?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Added a check for it in e117ec4
e117ec4
to
e91b6b7
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
ACK e91b6b7
pending tests pass :)
c31f4cc
to
d991a45
Compare
a225d8c
to
c8eecaf
Compare
Interop testing with Eclair revealed an issue with remote funding key rotation. This searches for the funding output using the rotated remote funding pubkey instead of the furrent funding pubkey. Also update the variable name to be more clear which this represents. Changelog-Changed: Interop fixes for compatability with Eclair
This is needed to remember if a splice was locked and reconnect occurs mid `splice_locked` attempted so it can be resumed in reestablish.
This field should be serialized for communication between channeld and lightningd.
To support resuming `splice_locked` across channel reconnect, we need to pass inflight `is_locked` between lightningd and channeld. This implements that interface between channeld and lightningd so each inflight should have up to date `is_locked` values between restarts.
Use the inflight information to set a correct `locked_txid` value on restart. This is critical for handling interrupted `splice_locked` events that need to be resumed on reconnect/reestablish.
A new case where `splice_locked` must be sent again on reestablish. This handles the case where `splice_locked` did not complete locally or remotely and must be resumed.
The interaction betwen libwally and CLN’s memory management is tricky. Let’s dodge that problem and just clone the PSBTs. Clean up some unused PSBT / ictx code while we’re at it
Upscale user provided PSBTs to v2 and convert them back to user preference when returned.
Check that the peer sent the correct txid in their `splice_locked` message. We have to check this later on in `check_mutal_splice_locked` so we store the value in `splice_state`
PSBT changeset routines were using linearize_output which mutated the memory of the objects it was comparing. This commit fixes that and also cleans up the memory usage to be more clear and more guarentee there is no memory corruption. Changelog-None
Update splice flows to use the new `clone_psbt` method instead of stealing back and forth.
`bitcoin_tx_with_psbt` would somewhat opaquely steal the passed `psbt` value. This caused a bug where code made a `bitcoin_tx` using a psbt without realizing the value was stolen. Because the resulting `bitcoin_tx` was placed in tmpctx it was not immediately clear that using `psbt` afterwards was an error until the tmpctx was cleared — creating a valgrind backtrace far from the actual issue. Switching to the routine to using TAKES and adding documentation in the header, makes it explicitly clear which operation the user is doing — helping prevent future regressions of this kind. Changelog-None
A routine that audit’s and asserts PSBT memory to confirm it has a sane memory allocation hierarchy. Changelog-None
Default wally_tal_ctx to NULL, add extra asserts and tal_checks, and documentation explaning the usage of tal_wally_start/end. Changelog-None
An extra check to ensure the user doesn’t try to sign a splice that wasn’t finalized.
Cleaning up the memory hierarchy of PSBT usage in splicing and `psbt_finalize_input`
c8eecaf
to
4f23e3b
Compare
Rebased on master which will fix some broken tests... |
Interop testing with Eclair revealed an issue with remote funding key rotation.
This searches for the funding output using the rotated remote funding pubkey instead of the current funding pubkey.
Update the variable name to be more clear which this represents.
Implement new pending spec changes for
splice_locked
resuming in reestablish.Fixes #8030