clockspeed: increase timeout for smtpclock
We're seeing greater than 1 second latencies in some venues. While
increasing the timeout decreases the clock accuracy, it also unblocks
logging.
This is a stop-gap fix, the correct fix is to switch to a persistent
time daemon that can handle jitter and skew detection.
Bug: 34455653
Change-Id: Ib548dc13e48a2f1b1518a3b7bcdb303f6fbea6e3
diff --git a/package/clockspeed/clockspeed-0005-wait-longer.patch b/package/clockspeed/clockspeed-0005-wait-longer.patch
new file mode 100644
index 0000000..d235402
--- /dev/null
+++ b/package/clockspeed/clockspeed-0005-wait-longer.patch
@@ -0,0 +1,11 @@
+--- a/sntpclock.c 2017-01-19 12:08:54.003237871 -0800
++++ b/sntpclock.c 2017-01-19 12:27:31.212751112 -0800
+@@ -202,7 +202,7 @@
+ strerr_die2sys(111,FATAL,"unable to send request: ");
+ FD_ZERO(&rfds);
+ FD_SET(s,&rfds);
+- tvselect.tv_sec = 1;
++ tvselect.tv_sec = 5;
+ tvselect.tv_usec = 0;
+ if (select(s + 1,&rfds,(fd_set *) 0,(fd_set *) 0,&tvselect) != 1) {
+ strerr_warn2(WARNING,"unable to read clock: timed out",0);
