Re: [Linuxptp-users] ptp4l and multiple network devices
PTP IEEE 1588 stack for Linux
Brought to you by:
rcochran
|
From: Vadim B. <va...@gm...> - 2017-08-04 19:06:47
|
It appears that it was finding another master clock through the wireless interface. After I disabled it, it looks better: ptp4l[805.793]: master offset 161 s2 freq +4647 path delay 16058 Thank for the help! On Fri, Aug 4, 2017 at 11:37 AM, Keller, Jacob E <jac...@in...> wrote: > How many devices are you running ptp4l on? 2? 3? If only 2, it seems like > you have an extra instance of ptp4l on the network somehow, > > > > ptp4l[1186.939]: port 2: new foreign master 003053.fffe.20597e-1 > ptp4l[1187.460]: port 1: new foreign master 003053.fffe.20597f-1 > > > > Not really sure what that is about. > > > > The line “ptp4l[1204.949]: failed to step clock: Invalid argument” > probably indicates a driver bug since the clock step might be returning an > error code as if the feature was not supported.... > > > > Thanks, > > Jake > > > > *From:* Vadim Butakov [mailto:va...@gm...] > *Sent**:* Friday, August 04, 2017 11:33 AM > *To:* lin...@li... > *Subject:* Re: [Linuxptp-users] ptp4l and multiple network devices > > > > Thanks for the reply > > The "ptp4l -i enp6s0 -i enp7s0 -m -S" command indeed launches fine, but > doesn't provide the expected result: > ptp4l[1185.776]: port 1: INITIALIZING to LISTENING on INIT_COMPLETE > ptp4l[1185.777]: port 2: INITIALIZING to LISTENING on INIT_COMPLETE > ptp4l[1185.777]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE > ptp4l[1186.939]: port 2: new foreign master 003053.fffe.20597e-1 > ptp4l[1187.460]: port 1: new foreign master 003053.fffe.20597f-1 > ptp4l[1190.539]: selected best master clock 003053.fffe.20597e > ptp4l[1190.539]: foreign master not using PTP timescale > ptp4l[1190.539]: running in a temporal vortex > ptp4l[1190.539]: port 2: LISTENING to UNCALIBRATED on RS_SLAVE > ptp4l[1191.443]: selected best master clock 003053.fffe.20597e > ptp4l[1191.443]: port 1: LISTENING to PRE_MASTER on RS_MASTER > ptp4l[1191.444]: foreign master not using PTP timescale > ptp4l[1191.444]: running in a temporal vortex > ptp4l[1192.348]: master offset 1501870118887912631 s0 freq +100000000 > path delay 13227463 > ptp4l[1195.048]: master offset 1501870118588155041 s0 freq +100000000 > path delay 13227463 > ptp4l[1195.444]: port 1: PRE_MASTER to MASTER on > QUALIFICATION_TIMEOUT_EXPIRES > ptp4l[1195.948]: master offset 1501870118491188556 s0 freq +100000000 > path delay 10117670 > ptp4l[1204.948]: master offset 1501870121056683138 s1 freq +100000000 > path delay 28161309 > ptp4l[1204.949]: failed to step clock: Invalid argument > ptp4l[1205.848]: master offset 1501870120956815970 s2 freq +100000000 > path delay 28161309 > ptp4l[1205.848]: port 2: UNCALIBRATED to SLAVE on MASTER_CLOCK_SELECTED > > etc. > > I assume there might be something wrong with my PC or network config, but > don't know where to start. My linux kernel is 4.4.0-83-generic (I don't > know if it matters). > > Cheers > > > > > > On Thu, Aug 3, 2017 at 9:52 PM, Richard Cochran <ric...@gm...> > wrote: > > On Thu, Aug 03, 2017 at 01:12:48PM -0700, Vadim Butakov wrote: > > First of all, thanks for ptp4l. You did an awesome job! > > I'm using two PTP cameras connected to two ethernet network devices > > (non-PTP) on my PC, and I'm trying to sync them by using the software > PTP. > > Can you tell us why you want "slave only" mode on the PC's second > interface? > > > Can you please give me a hint on what is the proper way to do it? > > Start one instance of ptp4l which uses both interfaces: > > ptp4l -i enp6s0 -i enp7s0 -m -S > > The master/slave port roles will be selected automatically using the > protocol. > > HTH, > Richard > > > |