Re: [Linuxptp-users] can't get linux PTP to run

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

FYI, the PTP4l does not directly modify the system (kernel) time, except in the case of Software timestamping. Each ptp device actually represents its own clock. You should use phc2sys application to modify the system time in addition. The reason for this design is because the clock on the adapter is completely disconnected from the system time in the kernel (The source is completely different). 

What version of the kernel are you using? If you don't have v3.5 with the get_ts_info ethtool support, you need to ensure that the /dev/ptp0 device is actually the device for the port you are using. Intel's Ethernet adapters create one ptp device per port due to being unable to represent multiple ports using the same clock.

If you have a multi-port adapter that could be part of why you are seeing some issues.

Do those negative adjust values go away over time? As in does it stabilize?

- Jake

> -----Original Message-----
> From: Tim...@de... [mailto:Tim-
> Ben...@de...]
> Sent: Tuesday, March 26, 2013 7:51 AM
> To: lin...@li...
> Subject: Re: [Linuxptp-users] can't get linux PTP to run
> 
> Hi,
> 
> ok it seems that the driver works now (down ask me what I have
> changed, don't know). I can now use HW stamping.
> 
> But the time does not change at all. If I start the master clock it shows:
> 	ptp4l -f ptp.cfg -p /dev/ptp0 -i eth2
> 	ptp4l[7470.955]: selected /dev/ptp0 as PTP clock
> 	ptp4l[7470.955]: failed to read out the clock frequency
> adjustment: Operation not supported
> 	ptp4l[7470.955]: port 1: get_ts_info not supported
> 	ptp4l[7470.958]: port 1: INITIALIZING to LISTENING on INITIALIZE
> 	ptp4l[7470.958]: port 0: INITIALIZING to LISTENING on INITIALIZE
> 	ptp4l[7476.958]: port 1: LISTENING to MASTER on
> ANNOUNCE_RECEIPT_TIMEOUT_EXPIRES
> 
> And if I start the slave it shows:
> 	ptp4l -f ptp.cfg -p /dev/ptp0 -i eth2 -s
> 	ptp4l[8394.096]: selected /dev/ptp0 as PTP clock
> 	ptp4l[8394.096]: failed to read out the clock frequency
> adjustment: Operation not supported
> 	ptp4l[8394.096]: port 1: get_ts_info not supported
> 	ptp4l[8394.098]: port 1: INITIALIZING to LISTENING on INITIALIZE
> 	ptp4l[8394.098]: port 0: INITIALIZING to LISTENING on INITIALIZE
> 	ptp4l[8395.985]: port 1: new foreign master 745798.fffe.000021-1
> 	ptp4l[8399.987]: selected best master clock 745798.fffe.000021
> 	ptp4l[8399.987]: port 1: LISTENING to UNCALIBRATED on
> RS_SLAVE
> 	ptp4l[8401.429]: negative path delay      -5520
> 	ptp4l[8401.429]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8401.429]: t2 - t3 = -427837600
> 	ptp4l[8401.429]: t4 - t1 = +427826560
> 	ptp4l[8401.429]: c1          0
> 	ptp4l[8401.429]: c2          0
> 	ptp4l[8401.429]: c3          0
> 	ptp4l[8402.001]: master offset -1637602062061189501 s0 freq
> +0 path delay     -5520
> 	ptp4l[8402.750]: negative path delay     -10180
> 	ptp4l[8402.750]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8402.750]: t2 - t3 = -748785600
> 	ptp4l[8402.750]: t4 - t1 = +748765240
> 	ptp4l[8402.750]: c1          0
> 	ptp4l[8402.750]: c2          0
> 	ptp4l[8402.750]: c3          0
> 	ptp4l[8403.002]: master offset -1637602062061158011 s1 freq
> +31485 path delay     -7850
> 	ptp4l[8404.002]: master offset -1637602062061128851 s2 freq -
> 32767999 path delay     -7850
> 	ptp4l[8404.002]: port 1: UNCALIBRATED to SLAVE on
> MASTER_CLOCK_SELECTED
> 	ptp4l[8404.216]: negative path delay      -2380
> 	ptp4l[8404.216]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8404.216]: t2 - t3 = -213503440
> 	ptp4l[8404.216]: t4 - t1 = +213498680
> 	ptp4l[8404.216]: c1          0
> 	ptp4l[8404.216]: c2          0
> 	ptp4l[8404.216]: c3          0
> 	ptp4l[8405.003]: master offset -1637602062061101475 s2 freq -
> 32767999 path delay     -6026
> 	ptp4l[8405.072]: negative path delay       -280
> 	ptp4l[8405.072]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8405.072]: t2 - t3 =  -69388320
> 	ptp4l[8405.072]: t4 - t1 =  +69387760
> 	ptp4l[8405.072]: c1          0
> 	ptp4l[8405.072]: c2          0
> 	ptp4l[8405.072]: c3          0
> 	ptp4l[8405.778]: negative path delay     -10560
> 	ptp4l[8405.778]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8405.778]: t2 - t3 = -774949920
> 	ptp4l[8405.778]: t4 - t1 = +774928800
> 	ptp4l[8405.778]: c1          0
> 	ptp4l[8405.778]: c2          0
> 	ptp4l[8405.778]: c3          0
> 	ptp4l[8406.003]: master offset -1637602062061072517 s2 freq -
> 32767999 path delay     -5784
> 	ptp4l[8406.374]: negative path delay      -4680
> 	ptp4l[8406.374]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8406.374]: t2 - t3 = -370912120
> 	ptp4l[8406.374]: t4 - t1 = +370902760
> 	ptp4l[8406.374]: c1          0
> 	ptp4l[8406.374]: c2          0
> 	ptp4l[8406.374]: c3          0
> 	ptp4l[8406.885]: negative path delay     -12120
> 	ptp4l[8406.885]: path_delay = (t2 - t3) + (t4 - t1)
> 	ptp4l[8406.885]: t2 - t3 = -881060360
> 	ptp4l[8406.885]: t4 - t1 = +881036120
> 	ptp4l[8406.885]: c1          0
> 	ptp4l[8406.885]: c2          0
> 	ptp4l[8406.885]: c3          0
> 	ptp4l[8407.004]: master offset -1637602062061042610 s2 freq -
> 32767999 path delay     -6531
> 
> My config is:
> [global]
> #
> # Port Data Set
> #
> logAnnounceInterval     1
> logSyncInterval         0
> logMinDelayReqInterval  0
> logMinPdelayReqInterval 0
> announceReceiptTimeout  3
> delayAsymmetry          0
> fault_reset_interval    4
> #
> # Run time options
> #
> assume_two_step         0
> logging_level           6
> path_trace_enabled      1
> follow_up_info          0
> tx_timestamp_retries    100
> use_syslog              0
> verbose                 1
> summary_interval        0
> #
> # Servo Options
> #
> pi_proportional_const   0.0
> pi_integral_const       0.0
> pi_offset_const         0.0
> clock_servo             pi
> #
> # Default interface options
> #
> network_transport       UDPv4
> delay_mechanism         E2E
> time_stamping           hardware
> 
> If I check the time (both date and hwclock) on both PCs, one says (for
> example) 10:12:11 and the other says 08:45:00. Even after an hour
> running ptp4l the time difference is the same.
> 
> I think it got something to do with the clock frequency adjustment.
> 
> I'm getting closer to success but what am I doing wrong now?
> 
> Thanks,
> Benny
> 
> 
> -----Ursprüngliche Nachricht-----
> Von: Richard Cochran [mailto:ric...@gm...]
> Gesendet: Montag, 25. März 2013 19:20
> An: Rapp, Tim-Benjamin
> Cc: lin...@li...
> Betreff: Re: [Linuxptp-users] can't get linux PTP to run
> 
> On Mon, Mar 25, 2013 at 08:11:12AM +0000, Tim-
> Ben...@de... wrote:
> > Hi,
> >
> > well im using the modified driver from sourceforge (see post of Jake).
> > Therfore I only can use software timetamping, because HW-
> timespamps are not supported. If i dont use HW timestamping then I
> dont need to pick an /dev/ptp (I dont have one).
> 
> Looking at the code, in e1000e-2.3.2.tar.gz, and in its README, HW time
> stamps *are* supported, but you need to add flag when building.
> 
>    IEEE 1588 Precision Time Protocol (PTP) Hardware Clock (PHC)
>   ------------------------------------------------------------
>   Support for the IEEE 1588 Precision Time Protocol (PTP) Hardware Clock
> (PHC)
>   is disabled by default in this out-of-tree driver even if it is enabled for
>   the in-kernel driver.
> 
>   The feature is available only on a subset of devices supported by the
>   driver, and can only be enabled on 3.0 and newer kernels that also
> have
>   the PTP_1588_CLOCK support compiled in statically or as a module. To
> enable
>   the feature when compiling the driver, add 'CFLAGS_EXTRA=-
> DE1000E_PTP' to
>   the command line.
> 
> > After your information I changed to using a config file. Did I do it like
> you ment? Adding the tx_timestamp_retries? I tried several different
> numbers. Always the same error.
> >
> > The command I use is:
> > 	ptp4l -i eht2 -f /etc/ptp4l.cfg
> >
> > the config is:
> > 	[global]
> > 	#
> > 	logging_level		6
> > 	tx_timestamp_retries	100
> > 	use_syslog			0
> > 	verbose			1
> > 	time_stamping		software
> 
> When using SW time stamping, the tx_timestamp_retries has no effect,
> since the time stamp is taken during the 'send' call.
> 
> I can only suggest to recheck if:
> 
> - you are really using the out-of-tree driver
> - that eth2 is really, truely serviced by the e1000e driver
> 
> There is a Linux test program for time stamping in
> 
>   Documentation/networking/timestamping
> 
> Can you test your interface with that program like so:
> 
>   timestamping eth2 SOF_TIMESTAMPING_TX_SOFTWARE
> SOF_TIMESTAMPING_RX_SOFTWARE SOF_TIMESTAMPING_SOFTWARE
> 
> and post the results?
> 
> Thanks,
> Richard
> 
> 
> ------------------------------------------------------------------------------
> Own the Future-Intel&reg; Level Up Game Demo Contest 2013
> Rise to greatness in Intel's independent game demo contest.
> Compete for recognition, cash, and the chance to get your game
> on Steam. $5K grand prize plus 10 genre and skill prizes.
> Submit your demo by 6/6/13. http://p.sf.net/sfu/intel_levelupd2d
> _______________________________________________
> Linuxptp-users mailing list
> Lin...@li...
> https://lists.sourceforge.net/lists/listinfo/linuxptp-users

Re: [Linuxptp-users] can't get linux PTP to run

PTP IEEE 1588 stack for Linux

Re: [Linuxptp-users] can't get linux PTP to run