Hi Richard,
Am 2021-03-06 22:32, schrieb Richard Cochran:
Dear linuxptp developers and users,
The time has come for another release. Starting today, a "code
freeze" is in effect, where only bug fixes will be applied. In one
week's time I'll release v3.2 of the software.
did I miss
Hi Luigi,
there is no reasoning, why you need cmake in this RFC, just the
diffs. Therefore, what is the additional value here compared to
standard make.
-michael
___
Linuxptp-devel mailing list
Linuxptp-devel@lists.sourceforge.net
Am 2021-02-20 07:26, schrieb Yangbo Lu:
IEEE 1588-2019 specified new UInteger4 type minorVersionPTP field in
header,
and minorVersionNumber data in portDS. It has the value 1 for IEEE
1588-2019,
and has the value 0 for IEEE 1588-2008.
This patch is to make versionNumber and minorVersionNumber
Am 2021-02-11 11:17, schrieb Karthikkumar V:
Example Use-Case
This is needed in the cases where T-SC/T-BC Slave might want to listen
only on PRC clockCLass and
anything beyond that might not be acceptible and would want to go to
holdover (with SyncE backup or
internal oscillator).
Hi all,
Am 2021-01-26 11:00, schrieb Miroslav Lichvar:
In port_open(), don't assume that UDS ports always have to have a zero
number. Check the transport directly to make make the code cleaner.
What about creating a helper? Something like port_is_uds(struct *port)?
That way the actual
Am 2021-01-05 04:17, schrieb Richard Cochran:
On Tue, Jan 05, 2021 at 11:20:09AM +0900, Cliff Spradlin wrote:
This sounds good to me. However, cid2str() uses static storage and
there are several log messages that log information about multiple
ports.
You mean 2+ ports appear in the same
mode?
-michael
Rodney
-Original Message-
From: Michael Walle
Sent: Tuesday, March 24, 2020 3:38 AM
To: Richard Cochran
Cc: Rodney Cummings ; Erik Hons
; linuxptp-devel@lists.sourceforge.net
Subject: [EXTERNAL] Re: [Linuxptp-devel] [RFC V2] Add IEEE
802.1AS-2011
time-aware bridge sup
Am 2020-03-20 18:32, schrieb Richard Cochran:
On Tue, Mar 17, 2020 at 04:41:14PM +, Rodney Cummings wrote:
Recommendation: Keep to 802.1AS-2011 bridge for the upcoming patches.
Okay.
Recommendation: After 802.1AS-2011 bridge is in the Linuxptp master,
NI runs the UNH-IOL Conformance
Hi all,
Am 2019-09-26 16:53, schrieb Erik Hons:
Hi Yangbo,
> > May I have your suggestion here? To maintain gPTP time in software,
> > I just copied kernel timecounter code into linuxptp for usage.
>
> Why? That sounds wrong.
Regarding to physical clock adjustment, that's confusing. This
Am 2018-09-03 13:58, schrieb Richard Cochran:
On Mon, Sep 03, 2018 at 01:44:19PM +0200, Michael Walle wrote:
Sending event messages to sockets which are already disconnected, eg.
a
client connecting over UDS, subscribing to an event and disconnects.
At the
moment, the subscriber is only
Am 2018-09-05 13:27, schrieb Richard Cochran:
On Wed, Sep 05, 2018 at 08:44:18AM +, Y.b. Lu wrote:
Currently linuxptp supports 802.1AS end-station, but not supports
time-aware bridge.
We will have switch devices supporting PHC APIs and SO_TIMESTAMPING.
Could I just run P2P TC config
Am 2018-08-18 16:04, schrieb Richard Cochran:
On Mon, Aug 06, 2018 at 05:29:09PM +0200, Michael Walle wrote:
Would an implemention (like execute some shell scripts on specific
events)
inside linuxptp be accepted at all?
No. Even though it seems like the easy way, in the end it will prove
Am 2018-08-18 15:50, schrieb Richard Cochran:
On Tue, Jul 31, 2018 at 10:46:50AM +0200, Michael Walle wrote:
If the uds transport returns an error, the subscriber is automatically
removed.
Can you please explain:
a) the problem, and
Sending event messages to sockets which are already
Am 17. August 2018 18:07:30 MESZ schrieb Richard Cochran
:
>On Tue, Aug 14, 2018 at 02:56:44PM +0200, Michael Walle wrote:
>> AVnu specifies a measurement method to remotely measure the clock
>> quality. Basically, this method sends Sync and/or Follow-up messages
>on
>> a
-Quality-Measurement-1.0_Approved-for-Public-Release.pdf
Signed-off-by: Michael Walle
---
config.c | 3 +++
fd.h | 3 ++-
p2p_tc.c | 4
port.c | 71 +++---
port.h | 17 ++
port_private.h
[please keep my mail address on CC]
Am 2018-08-06 16:53, schrieb Richard Cochran:
On Mon, Aug 06, 2018 at 09:27:18AM +0200, Stephan Gatzka wrote:
Hi!
> I came up with the following solutions to this problem:
> (1) Poll the state using the unix domain socket and PTP messages
> (2) Extend
Hi,
There is the need to inform an external program about the current state
of PTP. Like for example in 802.1Qbv, if PTP looses sync, we need to
switch to an emergency state (eg. unconditionally open all gates), or if
we regain sync, the Qbv scheduler might be reloaded.
I came up with the
While the pre_send() converts the timestamp inside the packet, the
post_recv() just copies (and converts) it into a special field inside
the message structure. But it doesn't convert the timestamp inside the
packet. Fix this.
Signed-off-by: Michael Walle
---
msg.c | 10 ++
1 file
If the uds transport returns an error, the subscriber is automatically
removed.
Signed-off-by: Michael Walle
---
clock.c | 11 ---
1 file changed, 8 insertions(+), 3 deletions(-)
diff --git a/clock.c b/clock.c
index faf2dea..feda257 100644
--- a/clock.c
+++ b/clock.c
@@ -237,9 +237,9
Am 2018-07-17 17:32, schrieb Richard Cochran:
On Tue, Jul 17, 2018 at 11:26:14AM +0200, Michael Walle wrote:
ping?
Sorry, I couldn't understand what you are trying to say.
Should I propose a fix? Ie. there have to be two versions ot t1 and
t2. One
for the offset and one for the delay
Am 2018-07-09 17:29, schrieb Michael Walle:
Hi,
at least in case of a free-running P2P TC (I don't know if this is
valid for the BC case), the tsproc mixes different measurements of the
SYNC and the PDELAY packets:
(1) tsproc_down_ts() in clock_synchronize() passes the timestamps of
the SYNC
Am 2018-07-10 16:49, schrieb Richard Cochran:
On Tue, Jul 10, 2018 at 11:55:44AM +0200, Michael Walle wrote:
These functions are just wrappers. Make it possible for the compiler
to
optimize the calls when compiled with optimization.
Are you sure?
Compiled with -O? I've looked
These functions are just wrappers. Make it possible for the compiler to
optimize the calls when compiled with optimization.
Signed-off-by: Michael Walle
---
msg.c | 12
msg.h | 11 +--
2 files changed, 9 insertions(+), 14 deletions(-)
diff --git a/msg.c b/msg.c
index
the tlv_list and tlv_count.
Signed-off-by: Michael Walle
---
msg.c | 18 --
1 file changed, 12 insertions(+), 6 deletions(-)
diff --git a/msg.c b/msg.c
index 6c20c97..d5b651f 100644
--- a/msg.c
+++ b/msg.c
@@ -174,6 +174,16 @@ static struct tlv_extra *msg_tlv_prepare(struct
The field is redundant with the length tlv_list. Replace it with a
function msg_tlv_count() instead. This iterates over the tlv_list. The
computational overhead should be small, because the lists are very short
and the tlv_count is only used in management paths (yet).
Signed-off-by: Michael Walle
[forgot to reply all]
Am 2018-07-10 06:38, schrieb Richard Cochran:
On Thu, Jul 05, 2018 at 09:46:26AM +0200, Michael Walle wrote:
@@ -193,6 +193,8 @@ static int suffix_post_recv(struct ptp_message
*msg, int len)
if (!ptr)
return 0;
+ msg_tlv_recycle(msg
Hi,
at least in case of a free-running P2P TC (I don't know if this is valid
for the BC case), the tsproc mixes different measurements of the SYNC
and the PDELAY packets:
(1) tsproc_down_ts() in clock_synchronize() passes the timestamps of
the SYNC packet; i guess this is to calculate
The field is redundant with the length tlv_list. Replace it with a
function msg_tlv_count() instead. This iterates over the tlv_list. The
computational overhead should be small, because the lists are very short
and the tlv_count is only used in management paths (yet).
Signed-off-by: Michael Walle
the tlv_list and tlv_count.
Signed-off-by: Michael Walle
---
msg.c | 21 +++--
msg.h | 7 +++
2 files changed, 22 insertions(+), 6 deletions(-)
diff --git a/msg.c b/msg.c
index 6c20c97..e408ea4 100644
--- a/msg.c
+++ b/msg.c
@@ -193,6 +193,8 @@ static int suffix_post_recv
Am 2018-07-03 07:21, schrieb Richard Cochran:
On Mon, Jun 25, 2018 at 05:06:49PM +0200, Michael Walle wrote:
Since commit 7fe69e7ba0c3bdce88d3a89d5ed43d0b382ef022 management
messages are dropped in some cases because the tlv_count isn't 1.
Further analysis shows that this is the case when
Am 2018-07-03 07:51, schrieb Richard Cochran:
On Tue, Jul 03, 2018 at 07:45:04AM +0200, Michael Walle wrote:
Ok. I'll look into that later. But still, don't you think its a better
idea to duplicate the message instead of encode/decode it?
I don't know. Is that what you have done in 'variant
Am 3. Juli 2018 07:21:37 MESZ schrieb Richard Cochran
:
>On Mon, Jun 25, 2018 at 05:06:49PM +0200, Michael Walle wrote:
>> Since commit 7fe69e7ba0c3bdce88d3a89d5ed43d0b382ef022 management
>> messages are dropped in some cases because the tlv_count isn't 1.
>> F
If the message is ignored, the tlv_extra isn't freed. Fix this.
Signed-off-by: Michael Walle
---
clock.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/clock.c b/clock.c
index 7bbb848..79ef917 100644
--- a/clock.c
+++ b/clock.c
@@ -440,6 +440,7 @@ static int clock_management_fill_response
and - in the forwarding case - there is already one TLV
in the list, which results in tlv_count being 2 and thus dropped in
subsequent code.
This patch fixes this behaviour by copying the message to avoid the
network byte order to host byte order cycle.
Signed-off-by: Michael Walle
---
This is the second
recreate in msg_post_recv().
Signed-off-by: Michael Walle
---
This is the first (not preferred) variant to fix this. It isn't that
invasive as the second one.
clock.c | 1 +
msg.c | 18 --
msg.h | 7 +++
3 files changed, 20 insertions(+), 6 deletions(-)
diff --git a/clock.c
Consistenly use the pid_eq() and cid_eq() helper functions. Move the
functions into utils.h (making them inline functions) because they are
used in multiple source files.
Signed-off-by: Michael Walle
---
clock.c| 13 -
phc2sys.c | 5 ++---
port.c | 18
36 matches
Mail list logo
