aljaz/lightning
1
0
Fork 0
mirror of https://github.com/aljazceru/lightning.git synced 2025-12-20 15:44:21 +01:00
Code Issues Packages Projects Releases Wiki Activity

connectd: add counters to each peer connection.

Browse Source 
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>
This commit is contained in:
Rusty Russell
2022-07-18 21:42:27 +09:30
committed by neil saitug
parent 41b379ed89
commit d31420211a
12 changed files with 82 additions and 30 deletions
Download Patch File Download Diff File Expand all files Collapse all files

38
lightningd/peer_control.c
View File

@@ -1079,6 +1079,7 @@ tell_connectd:
subd_send_msg(ld->connectd,
take(towire_connectd_peer_connect_subd(NULL,
&channel->peer->id,
channel->peer->connectd_counter,
&channel->cid)));
subd_send_fd(ld->connectd, fds[1]);
return;
@@ -1089,6 +1090,7 @@ send_error:
/* Get connectd to send error and close. */
subd_send_msg(ld->connectd,
take(towire_connectd_peer_final_msg(NULL, &channel->peer->id,
channel->peer->connectd_counter,
error)));
}
@@ -1134,6 +1136,7 @@ static void peer_connected_hook_final(struct peer_connected_hook_payload *payloa
"dev_disconnect permfail");
subd_send_msg(ld->connectd,
take(towire_connectd_peer_final_msg(NULL, &peer->id,
peer->connectd_counter,
channel->error)));
}
return;
@@ -1157,6 +1160,7 @@ send_error:
/* Get connectd to send error and close. */
subd_send_msg(ld->connectd,
take(towire_connectd_peer_final_msg(NULL, &peer->id,
peer->connectd_counter,
error)));
}
@@ -1279,12 +1283,14 @@ void peer_connected(struct lightningd *ld, const u8 *msg)
u8 *their_features;
struct peer *peer;
struct peer_connected_hook_payload *hook_payload;
u64 connectd_counter;
hook_payload = tal(NULL, struct peer_connected_hook_payload);
hook_payload->ld = ld;
hook_payload->error = NULL;
if (!fromwire_connectd_peer_connected(hook_payload, msg,
&id, &hook_payload->addr,
&id, &connectd_counter,
&hook_payload->addr,
&hook_payload->remote_addr,
&hook_payload->incoming,
&their_features))
@@ -1298,6 +1304,14 @@ void peer_connected(struct lightningd *ld, const u8 *msg)
peer = new_peer(ld, 0, &id, &hook_payload->addr,
hook_payload->incoming);
/* We track this, because messages can race between connectd and us.
* For example, we could tell it to attach a subd, but it's actually
* already reconnected: we would tell it again when we read the
* "peer_connected" message, and it would get upset (plus, our first
* subd wouldn't die as expected. So we echo this back to connectd
* on peer commands, and it knows to ignore if it wrong. */
peer->connectd_counter = connectd_counter;
/* We mark peer in "connecting" state until hooks have passed. */
assert(peer->connected == PEER_DISCONNECTED);
peer->connected = PEER_CONNECTING;
@@ -1336,6 +1350,7 @@ void peer_spoke(struct lightningd *ld, const u8 *msg)
{
struct node_id id;
u16 msgtype;
u64 connectd_counter;
struct channel *channel;
struct channel_id channel_id;
struct peer *peer;
@@ -1343,17 +1358,13 @@ void peer_spoke(struct lightningd *ld, const u8 *msg)
u8 *error;
int fds[2];
if (!fromwire_connectd_peer_spoke(msg, &id, &msgtype, &channel_id))
if (!fromwire_connectd_peer_spoke(msg, &id, &connectd_counter, &msgtype, &channel_id))
fatal("Connectd gave bad CONNECTD_PEER_SPOKE message %s",
tal_hex(msg, msg));
/* We must know it, and it must be the right connectd_id */
peer = peer_by_id(ld, &id);
if (!peer) {
/* This race is possible, but I want to see it in CI. */
log_broken(ld->log, "Unknown active peer %s",
type_to_string(tmpctx, struct node_id, &id));
return;
}
assert(peer->connectd_counter == connectd_counter);
/* Do we know what channel they're talking about? */
channel = find_channel_by_id(peer, &channel_id);
@@ -1470,12 +1481,15 @@ send_error:
/* Get connectd to send error and close. */
subd_send_msg(ld->connectd,
take(towire_connectd_peer_final_msg(NULL, &peer->id,
peer->connectd_counter,
error)));
return;
tell_connectd:
subd_send_msg(ld->connectd,
take(towire_connectd_peer_connect_subd(NULL, &id, &channel_id)));
take(towire_connectd_peer_connect_subd(NULL, &id,
peer->connectd_counter,
&channel_id)));
subd_send_fd(ld->connectd, fds[1]);
}
@@ -1495,16 +1509,20 @@ static void destroy_disconnect_command(struct disconnect_command *dc)
void peer_disconnect_done(struct lightningd *ld, const u8 *msg)
{
struct node_id id;
u64 connectd_counter;
struct disconnect_command *i, *next;
struct peer *p;
if (!fromwire_connectd_peer_disconnect_done(msg, &id))
if (!fromwire_connectd_peer_disconnect_done(msg, &id, &connectd_counter))
fatal("Connectd gave bad PEER_DISCONNECT_DONE message %s",
tal_hex(msg, msg));
/* If we still have peer, it's disconnected now */
/* FIXME: We should keep peers until it tells us they're disconnected,
* and not free when no more channels. */
p = peer_by_id(ld, &id);
if (p) {
assert(p->connectd_counter == connectd_counter);
log_peer_debug(ld->log, &id, "peer_disconnect_done");
p->connected = PEER_DISCONNECTED;