Hi Richard,
Please find the answers inline to your questions.
On Mon, 25 Jul 2022 at 02:22, Richard Cochran <richardcoch...@gmail.com>
wrote:
> On Fri, Jun 03, 2022 at 05:00:24PM +0530, SyncMonk Technologies wrote:
>
> > @@ -1996,8 +1996,10 @@ static void handle_state_decision_event(struct
> clock *c)
> > if (c->sanity_check)
> > clockcheck_reset(c->sanity_check);
> > tsproc_reset(c->tsproc, 1);
> > - if (!tmv_is_zero(c->initial_delay))
> > + if (!tmv_is_zero(c->initial_delay) || (best &&
> > + port_delay_mechanism(best->port) ==
> DM_NO_MECHANISM)) {
>
> Can you please explain what this extra test accomplishes?
>
[Devasish]: This is to enable ts_proc to use t1 and t2 for
tsproc_update_offset. As of now if the initial delay is zero it returns an
error.
In the case of the NO_DELAY mechanism applications can have zero
initial delay as they are expecting no delay in the path or can
program a known delay. So to use both the cases this test has been applied.
Adding snippet from older discussion:
*"On Thu, Jun 02, 2022 at 12:33:21PM +0530, Devasish Dey wrote:> NO_DELAY
mechanism address when user or system does not have any delay> measurement
mechanism. The Newly introduced options for ts_proc gives users> clear
indication that delay measurement mechanism is not going to be used> in
this mode of operation.Even if we ignore the technical aspect of the
implementation, I thinkit would not be very user friendly to have to set
two differentoptions to get it working.> Currently we do not see any
available option to work with no_delay> mechanism. If we need to use the
same option as suggested we need to update> the existing behavior.You could
change the behavior of the initial_delay option to acceptzero as a valid
value, but that would break compatibility withexisting configuration files
that set it to 0 for "no value".I'd suggest to call tsproc_set_delay() in
handle_state_decision_event(),which already does that for the initial_delay
option. The conditioncould be extended like this: if
(!tmv_is_zero(c->initial_delay) || best_dm == DM_NO_MECHANISM)
tsproc_set_delay(c->tsproc, c->initial_delay);."*
Other suggestions have been accepted and a new patch has been raised.
Thanks,
Devasish Dey
Vice President
Engineering Product Development | SyncMonk Technologies Pvt Limited
+91-783-8079202
devasish....@syncmonk.net
www.syncmonk.net
367, KPC Layout, Bangalore-560035
_______________________________________________
Linuxptp-devel mailing list
Linuxptp-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-devel
