Ethernet ports shall provide same (or defined) time at the same moment or there should be a way to set their offset to 0 or desired value. IDK reboot changes the value. See steps below
Reproduction steps:
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Ethernet ports shall provide same (or defined) time at the same moment or there should be a way to set their offset to 0 or desired value. IDK reboot changes the value. See steps below
Reproduction steps:
Hello Marek,
Update: I still do not have access to hardware, but I am working on getting some feedback for you soon.
Regards,
Nick
Hello Marek,
Note that each ICSSG port has its own IEP. Each IEP will be initialized at a different time during boot. That means there will be some time offset between each IEP counter value. We suspect that the reason you see a difference in latency measured is because the IEP counters are not in sync - you want user space to sync those clocks.
We see 0ns difference in latency if we modify the sendrec.c file to use a software timestamp.
Regards,
Nick
Hello Marek,
Just checking that your issue was actually resolved here.
FYI, the clocks can be synced with a command like:
phc2sys -m -q -s /dev/ptp2 -c /dev/ptp3 -O 0 -l 5 -R 10 -c /dev/ptp1
dev/ptp2 is master
/dev/ptp3 and /dev/ptp1 are slave (sync rate 10Hz)
Regards,
Nick
Hello
unfortunately this did not synchronize our PRUs.
runtest.sh must be put into the same directory as compareDirections.sh
it does one measurement before running phc2sys and then 10 measurements for 2 combinations of PTP devices ({ptp2, ptp4, ptp6} and {ptp3, ptp5, ptp7}) and then reboots
runtest.sh, logs with our measurements and output from phc2sys are attached.
Best regards,
Marek Vonka