The oncore hardware provides a 'negative saw-tooth' the offset of the PPS signal

which is on the zero crossing of its internal clock, from its true value.  the
refclock_oncore.c driver applies this difference to correct the time of the
provided timestamp.  This value was being applied correctly for the 6 and 8 channel
receivers, but not for the 12 channel.  This has been corrected.

bk: 47296f44Tjq8QobLIpMYPmtB6T9H4A

diff --git a/ntpd/refclock_oncore.c b/ntpd/refclock_oncore.c
index c720add2ca87120fd7710e7b8766d9fc8301ee3b..8cffb0c538bfed1e087922b01a887e20dd39b750 100644 (file)
--- a/ntpd/refclock_oncore.c
+++ b/ntpd/refclock_oncore.c
@@ -2422,14 +2422,14 @@ oncore_msg_BnEnHn(
                        return;
 
                dt1 = instance->saw_tooth + instance->offset;    /* dt this time step */
-               instance->saw_tooth = (s_char) instance->BEHn[10]; /* update for next time Hn[10] */
+               instance->saw_tooth = (s_char) instance->BEHn[14]; /* update for next time Hn[14] */
                dt2 = instance->saw_tooth + instance->offset;    /* dt next time step */
        } else {
                if (instance->BEHn[21]) /* bad TRAIM */
                        return;
 
                dt1 = instance->saw_tooth + instance->offset;    /* dt this time step */
-               instance->saw_tooth = (s_char) instance->BEHn[25]; /* update for next time */
+               instance->saw_tooth = (s_char) instance->BEHn[25]; /* update for next time Bn[25], En[25] */
                dt2 = instance->saw_tooth + instance->offset;    /* dt next time step */
        }