Allow ntpd to report the status of peers and sensors to syslog. This

happens when a SIGINFO is received, or when the majority of peers or
sensors is bad. The latter with a maximum of once per 24 hour.
ok henning@ ckuethe@ mbalmer@

src/usr.sbin/ntpd/ntp.c

@@ -1,4 +1,4 @@
-/*	$OpenBSD: ntp.c,v 1.99 2007/08/04 02:58:02 ckuethe Exp $ */
+/*	$OpenBSD: ntp.c,v 1.100 2007/10/15 06:59:31 otto Exp $ */
 
 /*
  * Copyright (c) 2003, 2004 Henning Brauer <henning@openbsd.org>
@@ -38,14 +38,18 @@
 #define	PFD_MAX		2
 
 volatile sig_atomic_t	 ntp_quit = 0;
+volatile sig_atomic_t	 ntp_report = 0;
 struct imsgbuf		*ibuf_main;
 struct ntpd_conf	*conf;
 u_int			 peer_cnt;
+u_int			 sensors_cnt;
+time_t			 lastreport;
 
 void	ntp_sighdlr(int);
 int	ntp_dispatch_imsg(void);
 void	peer_add(struct ntp_peer *);
 void	peer_remove(struct ntp_peer *);
+void	report_peers(int);
 
 void
 ntp_sighdlr(int sig)
@@ -55,6 +59,9 @@ ntp_sighdlr(int sig)
 	case SIGTERM:
 		ntp_quit = 1;
 		break;
+	case SIGINFO:
+		ntp_report = 1;
+		break;
 	}
 }
 
@@ -65,7 +72,6 @@ ntp_main(int pipe_prnt[2], struct ntpd_conf *nconf)
 	int			 hotplugfd, nullfd;
 	u_int			 pfd_elms = 0, idx2peer_elms = 0;
 	u_int			 listener_cnt, new_cnt, sent_cnt, trial_cnt;
-	u_int			 sensors_cnt = 0;
 	pid_t			 pid;
 	struct pollfd		*pfd = NULL;
 	struct passwd		*pw;
@@ -135,6 +141,7 @@ ntp_main(int pipe_prnt[2], struct ntpd_conf *nconf)
 
 	signal(SIGTERM, ntp_sighdlr);
 	signal(SIGINT, ntp_sighdlr);
+	signal(SIGINFO, ntp_sighdlr);
 	signal(SIGPIPE, SIG_IGN);
 	signal(SIGHUP, SIG_IGN);
 	signal(SIGCHLD, SIG_DFL);
@@ -173,6 +180,9 @@ ntp_main(int pipe_prnt[2], struct ntpd_conf *nconf)
 	TAILQ_FOREACH(p, &conf->ntp_peers, entry)
 		peer_cnt++;
 
+	/* wait 5 min before reporting first status to let things settle down */
+	lastreport = time(NULL) + (5 * 60) - REPORT_INTERVAL;
+
 	while (ntp_quit == 0) {
 		if (peer_cnt > idx2peer_elms) {
 			if ((newp = realloc(idx2peer, sizeof(void *) *
@@ -324,6 +334,8 @@ ntp_main(int pipe_prnt[2], struct ntpd_conf *nconf)
 			if (s->next <= getmonotime())
 				sensor_query(s);
 		}
+		report_peers(ntp_report);
+		ntp_report = 0;
 	}
 
 	msgbuf_write(&ibuf_main->w);
@@ -403,6 +415,9 @@ ntp_dispatch_imsg(void)
 					h->next = NULL;
 					npeer->addr = h;
 					npeer->addr_head.a = h;
+					npeer->addr_head.name =
+					    peer->addr_head.name;
+					npeer->addr_head.pool = 1;
 					client_peer_init(npeer);
 					npeer->state = STATE_DNS_DONE;
 					peer_add(npeer);
@@ -654,3 +669,59 @@ error_interval(void)
 	return (interval + r);
 }
 
+void
+report_peers(int always)
+{
+	time_t now;
+	u_int badpeers = 0;
+	u_int badsensors = 0;
+	struct ntp_peer *p;
+	struct ntp_sensor *s;
+
+	TAILQ_FOREACH(p, &conf->ntp_peers, entry) {
+		if (p->trustlevel < TRUSTLEVEL_BADPEER)
+			badpeers++;
+	}
+	TAILQ_FOREACH(s, &conf->ntp_sensors, entry) {
+		if (!s->update.good)
+			badsensors++;
+	}
+
+	now = time(NULL);
+	if (!always) {
+		if ((peer_cnt == 0 || badpeers == 0 || badpeers < peer_cnt / 2)
+		    && (sensors_cnt == 0 || badsensors == 0 ||
+		    badsensors < sensors_cnt / 2))
+			return;
+
+		if (lastreport + REPORT_INTERVAL > now)
+			return;
+	}
+	lastreport = now;
+	if (peer_cnt > 0) {
+		log_warnx("%u out of %u peers valid", peer_cnt - badpeers,
+		    peer_cnt);
+		TAILQ_FOREACH(p, &conf->ntp_peers, entry) {
+			if (p->trustlevel < TRUSTLEVEL_BADPEER) {
+				const char *a = "not resolved";
+				const char *pool = "";
+				if (p->addr)
+					a = log_sockaddr(
+					    (struct sockaddr *)&p->addr->ss);
+				if (p->addr_head.pool)
+					pool = "from pool ";
+				log_warnx("bad peer %s%s (%s)",  pool,
+				    p->addr_head.name, a);
+			}
+		}
+	}
+	if (sensors_cnt > 0) {
+		log_warnx("%u out of %u sensors valid",
+		    sensors_cnt - badsensors, sensors_cnt);
+		TAILQ_FOREACH(s, &conf->ntp_sensors, entry) {
+			if (!s->update.good)
+				log_warnx("bad sensor %s", s->device);
+		}
+	}
+}
+

src/usr.sbin/ntpd/ntpd.8

@@ -1,4 +1,4 @@
-.\" $OpenBSD: ntpd.8,v 1.20 2007/09/13 14:34:36 pyr Exp $
+.\" $OpenBSD: ntpd.8,v 1.21 2007/10/15 06:59:32 otto Exp $
 .\"
 .\" Copyright (c) 2003, 2004, 2006 Henning Brauer <henning@openbsd.org>
 .\"
@@ -14,7 +14,7 @@
 .\" AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT
 .\" OF OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 .\"
-.Dd $Mdocdate: September 13 2007 $
+.Dd $Mdocdate: October 15 2007 $
 .Dt NTPD 8
 .Os
 .Sh NAME
@@ -108,6 +108,12 @@ This option allows
 .Nm
 to send DEBUG priority messages to syslog.
 .El
+.Pp
+When
+.Nm
+receives a
+.Dv SIGINFO
+signal, it will write its peer and sensor status to syslog.
 .Sh FILES
 .Bl -tag -width "/var/db/ntpd.driftXXX" -compact
 .It Pa /etc/ntpd.conf

src/usr.sbin/ntpd/ntpd.h

@@ -1,4 +1,4 @@
-/*	$OpenBSD: ntpd.h,v 1.87 2007/09/13 14:34:36 pyr Exp $ */
+/*	$OpenBSD: ntpd.h,v 1.88 2007/10/15 06:59:32 otto Exp $ */
 
 /*
  * Copyright (c) 2003, 2004 Henning Brauer <henning@openbsd.org>
@@ -61,6 +61,7 @@
 #define	LOG_NEGLIGEE		32	/* negligible drift to not log (ms) */
 #define	FREQUENCY_SAMPLES	8	/* samples for est. of permanent drift */
 #define	MAX_FREQUENCY_ADJUST	128e-5	/* max correction per iteration */
+#define REPORT_INTERVAL		(24*60*60) /* interval between status reports */
 
 
 #define	SENSOR_DATA_MAXAGE	(15*60)