fix ts2phc to handle large NMEA delay (RHEL-23208)

Resolves: RHEL-23208
This commit is contained in:
Miroslav Lichvar 2024-01-30 09:49:16 +01:00
parent 2c55db4376
commit a5f9f599a2
2 changed files with 31 additions and 0 deletions

29
linuxptp-nmeadelay.patch Normal file
View File

@ -0,0 +1,29 @@
commit 0404c1924254c9162222dd5000a140165d21250c
Author: Miroslav Lichvar <mlichvar@redhat.com>
Date: Thu Jan 25 15:23:44 2024 +0100
ts2phc: Fix offset for NMEA delays over 0.5 seconds.
The current code of ts2phc assumes that the NMEA RMC message is received
within 0.5 seconds of the pulse the timestamp in the message is refering
to. However, with most receivers NMEA messages are referenced to the
previous pulse. This causes a 1-second error in the measured offset for
receivers with delays over 0.5 seconds.
Add a 0.5 second correction to map the whole interval between pulses to
the preceding pulse.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
diff --git a/ts2phc_nmea_pps_source.c b/ts2phc_nmea_pps_source.c
index 3a4267d..274b70a 100644
--- a/ts2phc_nmea_pps_source.c
+++ b/ts2phc_nmea_pps_source.c
@@ -191,6 +191,7 @@ static int ts2phc_nmea_pps_source_getppstime(struct ts2phc_pps_source *src,
return -1;
}
rmc = tmv_add(rmc, duration_since_rmc);
+ rmc = tmv_add(rmc, nanoseconds_to_tmv(500000000));
utc_time = tmv_to_nanoseconds(rmc);
utc_time /= (int64_t) 1000000000;
*ts = tmv_to_timespec(rmc);

View File

@ -29,6 +29,8 @@ Patch2: linuxptp-deprecated.patch
Patch3: linuxptp-ptpver.patch
# limit unicast message rate per address and grant duration
Patch4: linuxptp-ucastrate.patch
# fix ts2phc to handle large NMEA delay
Patch5: linuxptp-nmeadelay.patch
# check for EL-specific kernels with vclock support
Patch12: linuxptp-vclock.patch