f3b0ca69d3
trivial: fix a spelling error in a comment.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-12 17:30:37 +02:00
d4e761138e
bugfix: use a sensible test to detect a new master.
...
The code to detect a new master used pointer equality using a stale
pointer within the clock instance. Instead, the clock needs to remember
the identity of the foreign master in order to correctly detect a
change of master.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-12 17:30:36 +02:00
9edd319f11
Use clock_adjtime from glibc if available.
...
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-09-06 08:23:21 +02:00
7a50ead1d0
bugfix: correct wrong type for the cumulativeScaledRateOffset field.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-04 09:21:02 +02:00
8f5bd0edda
Guard against divide by zero.
...
If a buggy driver or hardware delivers bogus time stamps, then we might
crash with a divide by zero exception.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-04 08:16:53 +02:00
35170cf6e4
Introduce a macro for the constant 2^41.
...
We are going to need this more than once for working with the
cumulativeScaledRateOffset fields.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-04 07:37:10 +02:00
8cc3433b8b
Add an implementation specific management code.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:28 +02:00
5ca99377b6
Be tolerant of wrong a OUI in the follow up information.
...
Certain AVB bridges send their own OUI instead of the offical one.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:27 +02:00
8a4b2ab9b8
Remove unnecessary wait state from frequency estimator.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:27 +02:00
843661dfa2
Calculate the master/local rate ratio in two ways.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
e5edeacb9a
Extract the follow up info and pass it to the clock.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
94dd186cdf
Add a clock method to receive the follow up information TLV.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
f38338827b
Provide the clock with the estimated neighbor rate ratio.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
5e40b5ec01
Estimate the neighbor rate ratio from a slaved port.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
f86997185e
Give the frequency estimation interval to the port as well.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
c06f3adc84
Differentiate between the E2E and P2P delay timeout intervals.
...
We have one timer used for both delay request mechanisms, and we ought
to set the message interval accordingly.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:07 +02:00
ace0d71a80
Adjust the peer delay request message interval field.
...
IEEE 802.1AS-2011 specifies using the current logMinPdelayReqInterval
for this field.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:06 +02:00
db36fb52ca
Reset the master/local frequency estimator when changing masters.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-09-02 14:40:06 +02:00
1bb8ebbdc5
Invert the frequency ratio estimation.
...
In 802.1AS-2011 the ratio is defined as master/local, so we should follow
suit.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 17:28:14 +02:00
829c785414
Rework the frequency ratio estimator to use the tmv functions.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 17:28:14 +02:00
95598b4fde
Provide a conversion function from tmv to double.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 17:28:14 +02:00
845e947bcb
Correct a comment about the frequency ratio estimation.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 17:28:14 +02:00
1925f07b3f
Fix copy-pasto in config file parsing of the logging level.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 17:28:14 +02:00
6330764538
phc2sys: add options for number of PHC readings and update rate
2012-08-31 17:14:55 +02:00
bf21c7c6a9
phc2sys: improve servo start up
2012-08-31 17:14:55 +02:00
41ffd8dfb8
Add an acknowledgment in the readme of the hardware donors.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-31 16:55:45 +02:00
28ac1cb1b1
Bring the readme up to date.
...
- all of the original goals have been implemented
- add gPTP to the list of supported features
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-30 14:55:35 +02:00
d42914c335
Improve stability of PHC reading in phc2sys.
...
Make five PHC readings in sequence and pick the quickest one.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-08-30 14:48:32 +02:00
2fcf5cbb4f
Add non-PPS mode to phc2sys.
...
As some PHC hardware/drivers don't provide PPS, it may be useful to keep the
system clock synchronized to PHC via clock_gettime(). While the precision
of the clock is only in microsecond range, the error seems to be quite stable.
The -d parameter now can be omitted if -s is provided.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-08-30 14:47:01 +02:00
be87b05f52
Revert "Add non-PPS mode to phc2sys."
...
This reverts commit 998093b6f7
2012-08-30 14:45:50 +02:00
0a5caa8476
Revert "Improve stability of PHC reading in phc2sys."
...
This reverts commit f08d264e1d
2012-08-30 14:44:46 +02:00
f08d264e1d
Improve stability of PHC reading in phc2sys.
...
Make five PHC readings in sequence and pick the quickest one.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-08-29 15:29:56 +02:00
998093b6f7
Add non-PPS mode to phc2sys.
...
As some PHC hardware/drivers don't provide PPS, it may be useful to keep the
system clock synchronized to PHC via clock_gettime(). While the precision
of the clock is only in microsecond range, the error seems to be quite stable.
The -d parameter now can be omitted if -s is provided.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-08-29 15:29:43 +02:00
afd2a50b5b
pmc: add a simple interactive help command.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-28 19:16:08 +02:00
81648e81ed
Always set the clock class to 255 when slave only mode is configured.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-28 18:47:45 +02:00
852d587fd3
Drop the reference to best when freeing the foreign masters.
...
This fixes the following issue reported by valgrind, which occurs
after a port disable/initialize subsequent to having entered slave
mode.
==10651== Invalid read of size 4
==10651== at 0x804E6E2: fc_clear (port.c:175)
==10651== by 0x805132F: port_event (port.c:1352)
==10651== by 0x804B383: clock_poll (clock.c:597)
==10651== by 0x80498AE: main (ptp4l.c:278)
==10651== Address 0x41cba60 is 16 bytes inside a block of size 60 free'd
==10651== at 0x4023B6A: free (vg_replace_malloc.c:366)
==10651== by 0x804EB09: free_foreign_masters (port.c:287)
==10651== by 0x804FB14: port_disable (port.c:722)
==10651== by 0x8051228: port_dispatch (port.c:1298)
==10651== by 0x804B3C6: clock_poll (clock.c:602)
==10651== by 0x80498AE: main (ptp4l.c:278)
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-28 07:51:04 +02:00
2040933d74
Configure logging facility sooner.
...
Error messages from initialization were missing.
Signed-off-by: Miroslav Lichvar <mlichvar@redhat.com>
2012-08-27 21:20:08 +02:00
afbec38d6e
Release all resources when quitting the main loop.
...
This will allow running under valgrind to detect memory leaks
and the like.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 21:11:14 +02:00
40dc2e13ef
Make the clock release method a public function.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 21:09:59 +02:00
577d34facf
Provide a method to release the message cache.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 21:09:10 +02:00
889c707ba1
Break out of the main loop on Control-C.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 21:07:38 +02:00
2b819b6b76
uds: fix bug in file descriptor array opening.
...
By not touching the event descriptor, uds is leaving the value as zero,
resulting in polling on stdin.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 21:05:34 +02:00
3936316468
Correct the TAI-UTC offset when it is reasonable to do so.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-27 18:57:43 +02:00
1bfb6fd217
Warn if the master's time properties are suspicious.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 20:36:43 +02:00
d377229ab4
Update the hard coded TAI-UTC offset.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 18:48:54 +02:00
ce745572b4
Use zero for the UDS port number.
...
Using 0xffff looks dumb in the log and in the pmc output.
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 15:09:15 +02:00
420d48ce02
Use the ingress port identity for clock management replies.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 14:39:57 +02:00
9da511d6f6
Provide a method to obtain a port's identity.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 14:39:57 +02:00
b9c4fbc11d
Clean up the tlv and pmc include directives.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 14:39:57 +02:00
529206112d
Move shared layer 2 global declarations into an appropriate header file.
...
Signed-off-by: Richard Cochran <richardcochran@gmail.com>
2012-08-26 14:39:57 +02:00
Richard Cochran
Miroslav Lichvar