search

nxp-archive / openil_linuxptp

PTP IEEE 1588 stack for Linux
GNU General Public License v2.0
136 stars 60 forks source link

ptp4l: Master offset always increments and freq maxes out #9

Open PavanC-BRCM opened 4 years ago

PavanC-BRCM commented 4 years ago

Hi, I am new here, please excuse any error in posting. Please ask or inform if I am doing something that is not correct by the forum standards, I'd be glad to correct myself.

I wanted to get help in an issue I am facing with ptp4l.

I see a strange issue in my setup where the clocks are perfectly synchronizing when phc2sys with ptp4l is used, but ptp4l reports master offset that always increments and has frequency maxed out at -31000000. The path delay however is stable. With this, the slave is able to get sync with master's clock but I am worried about these values. I am not finding a clue as to where could be the problem. Snippets below.

ptp4l output on the slave: ptp4l[165701.384]: master offset -23587382 s2 freq -31000000 path delay 990 ptp4l[165702.384]: master offset -23591840 s2 freq -31000000 path delay 990 ptp4l[165703.384]: master offset -23596265 s2 freq -31000000 path delay 959 ptp4l[165704.384]: master offset -23600618 s2 freq -31000000 path delay 853 ptp4l[165705.384]: master offset -23605075 s2 freq -31000000 path delay 853 ptp4l[165706.384]: master offset -23609640 s2 freq -31000000 path delay 959 ptp4l[165707.384]: master offset -23614100 s2 freq -31000000 path delay 959 phc2sys output on the slave: phc2sys[165701.977]: CLOCK_REALTIME phc offset 82 s2 freq -19421 delay 25707 phc2sys[165702.978]: CLOCK_REALTIME phc offset -239 s2 freq -19717 delay 25551 phc2sys[165703.981]: CLOCK_REALTIME phc offset 82 s2 freq -19468 delay 26697 phc2sys[165704.983]: CLOCK_REALTIME phc offset 689 s2 freq -18837 delay 23866 phc2sys[165705.985]: CLOCK_REALTIME phc offset -844 s2 freq -20163 delay 25785 phc2sys[165706.987]: CLOCK_REALTIME phc offset -673 s2 freq -20245 delay 26847 phc2sys[165707.988]: CLOCK_REALTIME phc offset 441 s2 freq -19333 delay 25704

I think if the values are entirely wrong, the clocks should not have sync'ed correctly. Any pointers as to where to look for the issue will be highly appreciated.

Commands I have used are below: Master: (p5p1 is the interface) ptp4l -i p5p1 -m phc2sys -s CLOCK_REALTIME -c p5p1 -w -m

Slave: (p2p1 is the interface) ptp4l -i p2p1 -m -s phc2sys -s p2p1 -c CLOCK_REALTIME -w -m

Thanks in advance for any help.

Regards, Pavan

PavanC-BRCM commented 4 years ago

OK turns out I am not applying the frequency drift correctly. However still wondering how did the time sync work well...

vladimiroltean commented 4 years ago

What hardware are you working on?