Make NTP run continuouslybaserock/liw/S8704-always-ntpd

With these changes:

* NTP is run at boot time, once, to set the system clock.
  This allows the rest of the boot to continue as soon as time is set,
  which should happen as soon as there is networking. This is
  necessary because there's parts of the system that need to wait
  for the system clock to be set.

* NTP is then also started as a daemon, which runs continuously,
  until stopped by the user, to keep system clock synchronised.

The two systemd units need to be separate, otherwise we can't have
network.target depend on a unit that finishes quickly: the daemon
unit doesn't exit until there's an error.

1 files changed, 0 insertions, 31 deletions

diff --git a/scripts/ntpd-set.sh b/scripts/ntpd-set.sh
deleted file mode 100755
index d6fd8f25b..000000000
--- a/scripts/ntpd-set.sh
+++ /dev/null
@@ -1,31 +0,0 @@
-#!/bin/sh
-max_attempts=4
-
-# This script takes a list of ntp servers and passes them to ntpd to set the
-# system time. If a /etc/ntpd.conf file exists, the servers there are used,
-# if not, some default values are passed
-set_time() {
-    # -q flag makes ntpd exit after setting the time once
-    ntpd -q -n -p "$1"
-}
-
-check_time() {
-    for attempt in $(seq "$max_attempts"); do
-        for arg ; do
-            echo $arg
-            if set_time "$arg" ; then
-                return 0
-            fi
-        done
-        sleep 2
-    done
-    return 1
-}
-
-if [ -f /etc/ntpd.conf ]; then
-    server_list=`cat /etc/ntpd.conf | sed ':a;N;$!ba;s/\n/ /g;s/server//g'`
-    check_time $server_list
-else
-    # Use a default list if there's no config
-    check_time 0.pool.ntp.org 1.pool.ntp.org 2.pool.ntp.org 3.pool.ntp.org
-fi