wire_sync_write() adds length, but we already have it, so use write_all.
sync_crypto_read() handed an on-stack buffer to cryptomsg_decrypt_header,
which expected a tal() pointer, so use the known length instead.
sync_crypto_read() also failed to read the tag; add that in (no
overflow possible as 16 is an int, len is a u16).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The peer is woken up every 30 seconds to deliver the backlog of
messages. Additionally I added the normal message queue to be able to
send non-gossip message to the peer.
Turns out we want to permute transactions for the wallet too, so we
use void ** rather than assume we're shuffling htlc ** (and do inputs,
too!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This object is basically the embodyment of BOLT #2. Each HTLC already
knows its own state; this moves them between states and keeps them
consistent.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's currently written to produce "local" commit-txs, but of course we
need to produce remote ones too, for signing.
Thus instead of using "remote" and "local" we use "other" and "self",
and indicate with a single "side" flag which we're generating (because
that changes how HTLCs are interpreted).
This also adds to the tests: generate the remote view of the commit_tx
and make sure it matches!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We were using the remote per_commitment_point instead of the local
per_commitment_point to generate the remotekey for the local transaction.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's awkward to handle them differently. But this change means we
need to expose them to the generated code.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We used to have a permutation map; this reintroduces a variant which
uses the htlc pointers directly.
We need this because we have to send the htlc-tx signatures in output
order as part of the protocol: without two-stage HTLCs we only needed
to wire them up in the unilateral spend case so we simply brute-forced
the ordering.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Moved the broadcast functionality to broadcast.[ch]. So far this
includes only the enqueuing side of broadcasts, the dequeuing and
actual push to the peer is daemon dependent. This also adds the
broadcast_state to the routing_state and the last broadcast index to
the peer for the legacy daemon.
This used to be part of `lightningd_state` which is being split up for
the various subdaemons. The main change is the addition of the `struct
routing_state` in `routing.h` and the addition of `rstate` in `struct
lightningd_state` for backwards compatibility.
We can't run them in parallel, but we can at least have 'make check'
run them all.
Developers should be running "make check-source && make check".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The problem with wire headers not being generated in time before stuff
depended on it turns out to be related with inclusion order of
sub-makefiles. The inclusions must preceed the use of
LIGHTNINGD_HEADERS since they append to that variable.
Rusty Russell
Christian Decker