Commit Graph

115 Commits

Author SHA1 Message Date
Christian Decker
d30f3f1a40 sphinx: Remove obsolete onion implementation 2016-10-16 16:40:54 +02:00
Christian Decker
7bb5c279a8 sphinx: Implemented sphinx onion routing
Implements a spec-compliant sphinx onion routing format. The format has
been cross-checked with the go implementation
cdecker/lightning-onion@b9e117e.
2016-10-16 16:40:43 +02:00
Rusty Russell
8a3c9908ce protocol: no signature in update_commit if receiver has no outputs.
So if there are no HTLCs, and the receiver can't spend anyway, don't
sign.  This has the added benefit that no two signed commitment
transactions will ever be identical (the revocation preimage changes).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:25:13 +09:30
Rusty Russell
de7fb4a83f test_protocol: restart support.
We keep a "database" for each side's persistent state.  Upon restart,
each side tells the other where it was up to, in terms of the number
of commit and revocation messages it receives.

Because only one update can be in flight at a time, we can tell w

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:45 +09:30
Rusty Russell
25d9962e81 test_protocol: add --verbose flag.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:45 +09:30
Rusty Russell
3e34d50339 test_protocol: fix --svg
Had bitrotted, and was completely broken with recent changes.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:45 +09:30
Rusty Russell
ba50776c47 test_protocol: fix up expected test output.
Our implementation now displays things differently, so output changes.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:44 +09:30
Rusty Russell
73df39e0c9 test_protocol: attach states to each HTLC, rather than using queues.
This is simpler for database representation, and also allows simple
bit-tests for what is happening to a HTLC (eg. am I committed to it?
Are you?  etc.)

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:44 +09:30
Rusty Russell
d54a59f010 test_protocol: remove support for multiple commits in flight.
Laolu pointed out that we can't have multiple in flight, since we can't
sign without knowing the next revocation preimage.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:44 +09:30
Rusty Russell
836eee048b test_protocol: rename struct funding to struct commit_tx.
This is really what it is.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-08-18 14:23:44 +09:30
Christian Decker
e37bd38fdd Switched from openssl to sodium for test_onion 2016-08-09 13:10:19 +09:30
Rusty Russell
c649f62e43 libsecp256k1: update.
Add Makefile target update-secp256k1, and run it.

The only API change is that len is now an IN-OUT parameter to serialization
functions.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-07-01 12:00:17 +09:30
Rusty Russell
69cb158edd base58, script, protobuf_convert: don't use temporary secp256k1 context.
We use libsecp256k1 to convert signatures to DER; we were creating a
temporary one, but we really should be handing the one we have in dstate
through.  This does that, everywhere.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-07-01 12:00:17 +09:30
Rusty Russell
889db659c5 test: test commands during commit phase.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-26 15:25:25 +09:30
Rusty Russell
a2ced6c16c test/test_protocol: add actual crossover case.
Not just add-htlc crossing over, but all the commit stages.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-18 06:55:28 +09:30
Rusty Russell
e46a2ed49a test/test_protocol: rename internally to match updated 02-wire-protocol.md
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 15:06:12 +09:30
Rusty Russell
06d2c19c41 test/test_protocol: make pretty(-ish) SVG diagrams.
Could make them much prettier, but they're functional.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 15:02:08 +09:30
Rusty Russell
126c91e1b1 test/test_protocol: don't allow multiple commits in flight by default.
Add "nocommitwait" for the two cases we test that, otherwise add
assertions that we never end up dealing with anything other than
the previous commit.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 15:01:10 +09:30
Rusty Russell
a77466aa21 test/test_protocol: Keep pointer to previous state.
Since our pre-change state is always the same as the previous step's
post-change state, we can simply keep a pointer, with a dummy empty
state for the initial one.

We could function-wrap it, but this change is even simpler.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:58:06 +09:30
Rusty Russell
4c1b50baab test/test_protocol: don't bother keeping acked (outgoing) queue.
We can simply set a boolean to indicate there are changes.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:56:08 +09:30
Rusty Russell
bc8cd0dce7 test/test_protocol: don't manually reapply changes.
This finished the previous patch by simply copying funding_next to funding
when we want to apply changes.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:55:58 +09:30
Rusty Russell
52db7ae0c4 test/test_protocol: keep cache of state with all changes applied.
This makes it easier to test for validity, though we still double-check
that a change doesn't overlap previous changes.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:51:25 +09:30
Rusty Russell
388dfc355e test/test_protocol: add fee support.
We simply record how many fee changes there are, rather than supporting
a particular level.

Fees are tricky: it's a noop to apply them when incoming, but we apply them
when they've been acked.  Unlike HTLC modifications, which are symmetric,
fee updates only apply when returning to the originating node.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:49:32 +09:30
Rusty Russell
056c219bc1 test/test_protocol: add checksync command to assert nodes are in sync.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:48:11 +09:30
Rusty Russell
7f90d183da test/test_protocol: simulator for the updated wire-protocol BOLT.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-05-17 13:47:44 +09:30
Rusty Russell
cb0cc80984 patch remove-timeout.patch 2016-03-15 17:07:12 +10:30
Rusty Russell
983000428f daemon: rename CMD_SEND_HTLC_UPDATE to CMD_SEND_HTLC_ADD
There are other updates than just adding an HTLC; make this explicit.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-03-08 10:42:15 +10:30
Rusty Russell
4f67b59c26 protocol: rename routefail to fail.
It's a generic "I couldn't complete this" failure.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-03-08 10:41:15 +10:30
Rusty Russell
40361e3dd8 state: explicit callback to check pkt_open_complete.
For now, this always succeeds, but include the hook for completeness.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-03-08 10:39:15 +10:30
Rusty Russell
a2f4000d22 state: cleanup after anchor has been transmitted.
If something goes wrong after we've broadcast the anchor tx, we need to use
the commit tx to spend it.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-03-08 10:38:15 +10:30
Rusty Russell
5ac1d5df00 state: include daemon/peer (only if !TEST_COVERAGE).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:49 +10:30
Rusty Russell
65cc6bbd50 pkt_err: make it variadic.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
79e1c9bb62 state: make bitcoin_commit() take a non-const peer.
We have to sign the commit at this stage, so easiest if peer isn't const
so we can sign it in-place.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
6a7a046f60 state: add peer_unwatch_close_timeout() call.
This tell us to disarm the INPUT_CLOSE_COMPLETE_TIMEOUT: either we hit
an error and are going to unilateral close, or we received their signature
successfully.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
144ab3bef4 state: make tx arguments const.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
ebf2bc57d8 state: add async anchor creation.
Actually generating the anchor transaction in my implementation
requires interaction with bitcoind, which we want to be async.  So add
a callback and a new state to wait for it.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
0db3c03ed1 state: fix BITCOIN_HTLC_TOTHEM_SPENT idata.
We use both union fields idata->btc and idata->htlc, which is clearly
wrong.  Have peer_tx_revealed_r_value return the HTLC it's talking
about.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
2b3d5b4049 state: remove all traces of struct state_effect.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
63ea6bfd7a state: add callbacks for adding/removing watches.
This lets us eliminate struct state_effect altogether (the next patch
removes the now-unused arguments).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
1a20e8094f state: return tx to be queued directly.
Instead of effect->broadcast_tx.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
e984df486d state: return Pkt to be queued directly.
Instead of effect->send_pkt.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:47 +10:30
Rusty Russell
7383da5f87 state: remove update_theirsig effect.
They get this from accept_pkt_update_accept() or accept_pkt_update_signature().

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
63cb0667f6 state: use callbacks for htlc management.
We only have one htlc in flight at a time, but sometimes it changes:
particularly when we are lowpriority and a highpriority request comes
in.  Handle this using a set of callbacks for htlc handling.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
f48adb097e state: use peer_unexpected_pkt() for an unexpected packet.
Instead of effect->in_error.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
4d9041e62a state: fix case where we replied to an error with an error.
Found by fixing bogus test in test_state_coverage.c

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
9013a7d872 state: set peer->state directly.
Instead of new_state effect.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
4c9a9f8982 state: return status of current command.
We temporarily move effect to the end of the arg list: we'll get rid
of it eventually.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
03268014b4 state: use PEER_BUSY when doing a command.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
14ede84d87 state: set peercond in peer directly.
Instead of stop_packets or stop_commands effects.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30
Rusty Russell
50a4b14a9a test/test_state_coverage.c: remove deferred packet handling.
We no longer defer packets.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
2016-01-22 06:41:46 +10:30