After connecting 100,000 peers with one channel each (not all at
once!), we see various places where we exhibit O(N^2) behaviour.
Fix these by keeping a map of id->peer instead of a simple
linked-list.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We used to tell connectd to remember our connect delay, and hand it
back (increased if necessary).
Instead, simply record when we last tried to connect. If it was less
than 10 minutes ago, double delay (up to 5 minutes max), otherwise
reset delay to 1 second.
This covers all scenarios: whether we reconnect then immediately
disconnect, or never successfully connect, it doesn't matter.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Fixes: #5453
Give them time to process any final messages! If there's a reconnect,
then we need to clean them up immediately of course.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This allows us to detect when lightningd hasn't seen our latest
disconnect/reconnect; in particular, we would hit the following pattern:
1. lightningd says to connect a subd.
2. connectd disconnects and reconnects.
3. connectd reads message, connects subd.
4. lightningd reads disconnect and reconnect, sends msg to connect to subd again.
5. connectd asserts because subd is alreacy connected.
This way connectd can tell if lightningd is talking about the previous
connection, and ignoere it.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Before this patch:
1. connectd says it's connected (peer_connected)
2. we tell connectd we want to talk about each channel (peer_make_active)
3. connectd gives us an fd for each channel, and we connect it to a subd (peer_active)
4. OR, connectd says it sent something about a channel we didn't tell it about, with an fd (peer_active)
Now:
1. connectd says it's connected (peer_connected)
2. we start all appropriate subds and tell connectd to what channels/fds (peer_connect_subd).
3. if connectd says it sent something about a channel we didn't tell it about, we either tell
it to hang up (peer_final_msg), or connect a new opening daemon (peer_connect_subd).
This is the minimal-size patch, which is why we create socket pairs in
so many places to use the existing functions. Many cleanups are
possible, since the new flow is so simple.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
First, connectd tells us the peer has connected, and we call the connected hook,
and if it says it's fine, we are actually connected and we fire off notifications.
Of course, we could be disconnected while in the connected hook, and that would
mean we tell people about a connection which is no longer current.
Make this clear with a tristate: if we're not marked disconnected by
the time the hooks finish, we're good. It also gives us a cleaner
"connect" command return when we connected but disconnected before
processing.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is redundant now, since connectd only sends us this once we tell
it it's OK, but that's changing, so clean up now. This means that
connectd will be able to make *unsolicited* closes, if it needs to.
We share logic with peer_please_disconnect.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We have them split over common/param.c, common/json.c,
common/json_helpers.c, common/json_tok.c and common/json_stream.c.
Change that to:
* common/json_parse (all the json_to_xxx routines)
* common/json_parse_simple (simplest the json parsing routines, for cli too)
* common/json_stream (all the json_add_xxx routines)
* common/json_param (all the param and param_xxx routines)
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Activate means a specific thing now (connectd said something), so avoid
confusing it with this function.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Either because lightningd tells us it wants to talk, or because the peer
says something about a channel.
We also introduce a behavior change: we disconnect after a failed open.
We might want to modify this later, but we it's a side-effect of openingd
not holding onto idle connections.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We currently intuit this by whether there's a subdaemon owning it.
But we're about to change the rules and allow connectd to hold idle
connections, so we need an explicit flag.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Instead of doing this weird chaining, just call them all at once and
use a reference counter.
To make it simpler, we return the subd_req so we can hang a destructor
off it which decrements after the request is complete.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is neater than what we had before, and slightly more general.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Changed: JSON_RPC: `sendcustommsg` now works with any connected peer, even when shutting down a channel.
We now let gossipd do it.
This also means there's nothing left in 'struct per_peer_state' to
send across the wire (the fds are sent separately), so that gets
removed from wire messages too.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Enable non-dev builds to send custom messages.
Preserves 'dev-' for compat-enabled builds.
Changelog-Changed: JSON-RPC: moved dev-sendcustommsg to sendcustommsg
This means remembering the connection direction. We also use the address to try
to reconnect, which we shouldn't bother with if they connect to us.
For peers from the database, we currently always save the addr: we shouldn't really
do this if they connected to us, since it's not useful for reconnecting (we don't
show the addr in JSON reply to listpeers unless we're connected, so it's only an
internal issue). This is left for future work.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is in line with the warnings draft, where all-zeroes in a
channel_id is no longer special (i.e. it will be ignored).
But gossipd would send these if it got upset with us, so it's best
practice to ignore them for now anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Changelog-Added: Protocol: we treat error messages from peer which refer to "all channels" as warnings, not errors.
Since this all stays in dualopend/dual_open_control, we can hold
onto the openchannel_signed command to wait for a response here locally.
Previously we were splitting across the channeld/openingd boundary.
`openchannel_signed` commands hang out across the openingd/channeld
boundary -- we don't return until we've successfully broadcast the
transaction (or timed out waiting for them to send a tx_sigs back).
It's almost always "their_features" and "our_features" respectively, so
make those names clear.
Suggested-by: @cdecker
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is needed to fully implement handling of blockheight disagreements
between us and payee.
If payee believes the blockheight is higher than ours, then `pay`
should wait for our node to achieve that blockheight.
Changelog-Add: Implement `waitblockheight` to wait for a specific blockheight.
We had a separate logbook for each peer, and copy log entries above
the printable log level into the master logbook. This didn't always
work well, since we didn't dump it on crash for example.
Keep a single global logbook instead, and remove this infrastructure.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This is mainly an internal-only change, especially since we don't
offer any globalfeatures.
However, LND (as of next release) will offer global features, and also
expect option_static_remotekey to be a *global* feature. So we send
our (merged) feature bitset as both global and local in init, and fold
those bitsets together when we get an init msg.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
The spec says to close the channel if they send us an error, but we
need to be more lenient to preserve channels with other
implementations.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Encapsulating the peer state was a win for lightningd; not surprisingly,
it's even more of a win for the other daemons, especially as we want
to add a little gossip information.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
These are always handed to subdaemons as a set, so group them. This makes
it easier to add an fd (in the next patch).
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
I tried to just do gossipd, but it was uncontainable, so this ended up being
a complete sweep.
We didn't get much space saving in gossipd, even though we should save
24 bytes per node.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
We always have an addr entry in the db (though it may be an ephemeral socket
the peer connected in from). We don't have to handle a NULL address.
While we're there, simplify new_peer not to take the features args;
the caller who cares immediately updates the features anyway.
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Rusty Russell
Michael Schmoock
niftynei
Christian Decker
ZmnSCPxj