On 5/16/23 3:56 AM, Erez wrote:
> On Tue, 16 May 2023 at 00:29, Kishen Maloor wrote:
>
>> In a setup with multiple gPTP domains, the Common Mean Link Delay Service
>> (CMLDS) (IEEE 1588/16.6.3) performs link delay measurements in a single
>> domain and must (somehow) convey those to other domains
A Boundary Clock might choose to switch a PHC when jbod is active,
and the transition will be based on the port's phc_index. However,
when using the slave event monitoring + free_running the phc_index
is -1, thus causing a fault in the transition.
Hence, avoid switching the PHC in case phc_index i
A Boundary Clock might choose to switch a PHC when jbod is active,
and the transition will be based on the port's phc_index. However,
when using the slave event monitoring + free_running the phc_index
is -1, thus causing a fault in the transition.
Hence, avoid switching the PHC in case phc_index i
Our use case is using the Slave Event Monitoring feature alongside free_running
so to implement an external servo - on a BC with jbod.
Alternatively, we saw that a similar commit
(fd94d9af9f82e410574a05e672e3f92b7c165655) was merged recently to include a
check for a negative phc_index. If it is
On Tue, May 16, 2023 at 01:34:56PM +, Eyal Itkin wrote:
> Our use case is using the Slave Event Monitoring feature alongside
> free_running so to implement an external servo - on a BC with jbod.
I would encourage you to contribute the servo. After all, the code is
modular with PI, linear reg
On Tue, 16 May 2023 at 00:29, Kishen Maloor wrote:
> In a setup with multiple gPTP domains, the Common Mean Link Delay Service
> (CMLDS) (IEEE 1588/16.6.3) performs link delay measurements in a single
> domain and must (somehow) convey those to other domains. IEEE 1588 does not
> specify this int