We now *never* consider asking them anything, even if they are
capable (e.g. enabling full gossip stream). This aligns with
the latest spec proposal, where lack of `gossip_queries` means
"we don't have anything useful to say".
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's a u64, we should pass by copy. This is a big sweeping change,
but mainly mechanical (change one, compile, fix breakage, repeat).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This has the benefit of being shorter, as well as more reliable (you
will get a link error if we can't print it, not a runtime one!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We have various functions to convert to a string, rename them all so we can
count on fmt_X being the formatter for struct X, and make them all return
`char *`.
Sometimes they existed but were private, sometimes they had a
different name. Most take a pointer, but simple types pass by copy:
short_channel_id, amount_msat and amount_sat.
The following public functions changed:
1. psbt_to_b64 -> fmt_wally_psbt.
2. pubkey_to_hexstr -> fmt_pubkey.
3. short_channel_id_to_str -> fmt_short_channel_id (scid by copy now!)
4. fmt_signature -> fmt_secp256k1_ecdsa_signature
5. fmt_amount_sat/fmt_amount_msat pass copy not pointer, return non-const char *.
6. node_id_to_hexstr -> fmt_node_id
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we delete it the first time a channel before it is closed, we will
crash when we try to move it again:
```
$ lightning_gossipd: gossip_store: get delete entry offset 47411992/51608943 (version v23.11-378-gac2a386-modded)
0x1002544b send_backtrace
common/daemon.c:33
0x1003415f status_failed
common/status.c:221
0x10016ef3 gossip_store_get_with_hdr
gossipd/gossip_store.c:466
0x10016faf check_msg_type
gossipd/gossip_store.c:491
0x1001722b gossip_store_set_flag
gossipd/gossip_store.c:509
0x1001752b gossip_store_del
gossipd/gossip_store.c:561
0x10017f5b remove_channel
gossipd/gossmap_manage.c:299
0x100181cf kill_spent_channel
gossipd/gossmap_manage.c:1144
0x1001a7df gossmap_manage_new_block
gossipd/gossmap_manage.c:1183
0x10014673 new_blockheight
gossipd/gossipd.c:483
0x10014d73 recv_req
gossipd/gossipd.c:594
```
Reported-by: @microsatosi on Discord
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Normally, channels are marked dying, the 12 blocks later, removed.
But for local channels, we can access any spliced channel already, so
we remove them immediately from our local gossip. This left a hole in
our logic, if that channel was the last one keeping a
node_announcement alive.
Solution is to unify with the "moved node_announcement" path.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We make sure a node_announcement is preceeded by at least one channel_announcement,
but dying ones don't count (as they are not broadcast!).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We accept node_announcements on dying channels, but make sure we
set the dying flag it channels are alll dying.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We can update dying channels, though it seems weird! We accept gossip about them,
we just don't propagate it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This avoids us gossiping about nodes which don't have live channels.
Interstingly, we previously tested that we *did* gossip such node
announcements, and now we fix that test.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Instead of "new" and "load", we don't really need to "load" anything,
so do everything in gossip_store_new.
Have it do the compaction/rewrite, and collect the dying records
gossmap offsets are to the beginning of the message, whereas
the gossip_store uses the header offset. Convert the internals
of gossip_store to use gossmap-style uniformly, even where it's
a little less convenient.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We don't use the dying flag, and we can manually append the
addendum rather than having gossip_store_add present a
bizarre interface.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The gossip_store_load is now basically a noop, since gossmap
does that.
gossipd removes a pile of routines dealing with messages,
in favor of just handing them to gossmap_manage.
The stub gossmap_manage constructor is removed entirely.
We simplified behaviour around channel_announcements with
no channel update: we now add them to the store, and go
back to fix the timestamp later. This changes a test,
which explicitly tests for the old behaviour.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is a bit less optimal than before, where we had an ordered map of
channels and could easily serve "channels between scids 800000x and
900000x". We now iterate all of them.
The rest is fairly mechanical.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We add a temporary stub gossmap_manage constructor, which simply opens
the gossmap and doesn't do anything else.
Then seeker uses this, rather than routing.c, to probe.
We optimize our "get random node announcements" a bit by traversing a
random set of nodes directly, and seeing if we have no
node_announcement, then querying their first channel.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
If we've got more than 10000 pending channel_announcements,
complain and stop processing any more.
If this becomes a problem, we can limit individual peers.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It's an unnecessary round-trip, and can cause us to complain in CI, in
the case where the channel has been closed by the time we ask.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
At initialization, gossipd is supposed to send all the local
channel_updates and any node_announcement it knows, so lightningd
doesn't generate fresh ones unnecessarily.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is a fair amount of code, but much is taken from
the old routing.c, with the difference that this uses
common/gossmap instead of our own structures.
The interfaces are fairly clear:
1. gossmap_manage_new - allocator
2. gossmap_manage_channel_announcement
- handle new channel announcement msg
- if too early, keeps it in early map
- queues it, asks lightingd about UTXO.
3. gossmap_manage_handle_get_txout_reply
- handle response from lightningd for above.
4. gossmap_manage_channel_update
- handle channel_update message
- may have to wait on pending channel_announcement
5. gossmap_manage_node_announcement
- handle node_announcement msg
- may have to wait on pending channel_announcement
6. gossmap_manage_new_block
- see if early announces can now be processed.
7. gossmap_manage_channel_spent
- lightningd tells us UTXO is spent
- may prepare channel for closing in 12 blocks.
8. gossmap_manage_channel_dying
- gossip_store load tells us channel was spent earlier.
- like gossmap_manage_channel_spent, but maybe < 12.
9. gossmap_manage_get_gossmap
- gossmap accessor: seeker and queries will need this.
10. gossmap_manage_new_peer
- a new peer has connected, give them all our gossip.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We weakened this progressively over time, and gossip v1.5 makes spam
impossible by protocol, so we can wait until then.
Removing this code simplifies things a great deal!
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Removed: Protocol: we no longer ratelimit gossip messages by channel, making our code far simpler.
We never enabled it, because we seemed to be eliminating valid
channels. We discard zombie-marked records on loading.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Most nodes don't really care about exact timestamps on gossip filters,
so just keep a flag on whether we have anything in the gossip_store,
and use that to determine whether we ask peers for everything.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is easier for future callers, which don't have a convenient peer
structure: in particular, asynchronous processing of gossip for peers.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
It was an obscure dev command, as it never worked reliably.
It would be much easier to re-implement once this is done.
This turned out to reveal a tiny leak on
tests/test_gossip.py::test_gossip_store_load_amount_truncated where we
didn't immedately free chan_ann if it was dangling.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
gossip_store_del - takes a gossmap-style offset-of-msg not offset-of-hdr.
gossip_store_flag: set an arbitrary flag on a gossip_store hdr.
gossip_store_get_timestamp/gossip_store_set_timestamp: access gossip_store hdr.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
