aljaz/lightning
1
0
Fork 0
mirror of https://github.com/aljazceru/lightning.git synced 2025-12-29 03:44:30 +01:00
Code Issues Packages Projects Releases Wiki Activity

connect: randomize backoff a little.

Browse Source 
Since we now fixed the bug where nodes receiving a connection would
try to reconnect to the source IP/port of that connection, we now expose
an issue mentioned by other implementers: we can continually cross over
reconnections unless we add some fuzz.  One second should be sufficient.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This commit is contained in:
Rusty Russell
2018-08-09 09:58:30 +09:30
committed by Christian Decker
parent 4f1186c4b1
commit be7a27a765
1 changed files with 7 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
lightningd/connect_control.c
View File

@@ -5,6 +5,7 @@
#include <ccan/tal/str/str.h>
#include <common/features.h>
#include <common/memleak.h>
#include <common/pseudorand.h>
#include <common/timeout.h>
#include <common/wireaddr.h>
#include <connectd/gen_connect_wire.h>
@@ -229,7 +230,12 @@ void delay_then_reconnect(struct channel *channel, u32 seconds_delay,
log_debug(channel->log, "Will try reconnect in %u seconds",
seconds_delay);
notleak(new_reltimer(&ld->timers, d, time_from_sec(seconds_delay),
/* We fuzz the timer by up to 1 second, to avoid getting into
* simultanous-reconnect deadlocks with peer. */
notleak(new_reltimer(&ld->timers, d,
timerel_add(time_from_sec(seconds_delay),
time_from_usec(pseudorand(1000000))),
maybe_reconnect, d));
}