Re: [Linuxptp-users] Which distributions have native ptp support?

[Gabe B. <Gab...@jd...>]

Thanks,

I may consider filing a bug if it works out of the box in another distribution.  I did try compiling the igb driver, but it gave so many compile errors when I tried to use the CFLAGS_EXTRA=-DIGB_PTP.  After fixing a static check for a version greater than the 2.6 kernel redhat supplies in the igb code, I got a ton of compile errors.  I decided not to pursue all the missing PTP related constants... (driver compiles fine without defining IGB_PTP).

Gabe


> -----Original Message-----
> From: Vick, Matthew [mailto:mat...@in...]
> Sent: Tuesday, October 29, 2013 9:08 AM
> To: Ledda William EXT; Gabe Black; Keller, Jacob E
> Cc: lin...@li...
> Subject: Re: [Linuxptp-users] Which distributions have native ptp
> support?
> 
> Gabe,
> 
> To chime in as well, if you continue to have problems I would recommend
> filing a Bugzilla with Red Hat. They own the version of igb and ptp4l
> included in their kernel and it could be something else in their stack
> interfering specifically with your configuration.
> 
> Another option for you to try is the latest version of igb from
> SourceForge (5.0.6) and compiling with "CFLAGS_EXTRA=-DIGB_PTP make" to
> see if that works for you.
> 
> I'm not sure what all the VM changes (other than obviously introducing
> some delay around when the app can run), but I would hope the device
> continues to operate correctly.
> 
> Cheers,
> Matthew
> 
> Matthew Vick
> Linux Development
> Networking Division
> Intel Corporation
> 
> 
> On 10/29/13, 1:11 AM, "Ledda William EXT" <Wil...@it...>
> wrote:
> 
> >Dear all,
> >Have you check the firewall? Is it enabled? The only problem that I
> had
> >(initially) was related to the firewall. After disabling the firewall
> >everything works!
> >
> >I worked without problem with the version released with RH 6.4
> >(linuxptp-0-0.6.20121114gite6bbbb.el6.x86_64) but I have also work
> with
> >1.2 and now I'm working with 1.3 without any problem.
> >How do you have compiled version 1.3? Do you have disabled the one
> step
> >option or do you have define HWTSTAMP_TX_ONESTEP_SYNC symbol? Base
> >kernel
> >2.36.32 haven't HWTSTAMP_TX_ONESTEP_SYNC in linux/net_tstamp.h, so in
> >order to compile, you should define this symbol or change the sk.c in
> >order to don't check for one step clocks.
> >
> >My syetem configuration:
> >
> >$ uname -a
> >Linux fc-vitro-tcn.codac.iter.org 2.6.32-358.2.1.el6.x86_64 #1 SMP Wed
> >Feb 20 12:17:37 EST 2013 x86_64 x86_64 x86_64 GNU/Linux
> >
> >$ ethtool -i eth1
> >driver: igb
> >version: 4.0.1-k
> >firmware-version: 1.61, 0x090f8000
> >bus-info: 0000:10:00.1
> >supports-statistics: yes
> >supports-test: yes
> >supports-eeprom-access: yes
> >supports-register-dump: yes
> >supports-priv-flags: no
> >
> >$ ethtool -T eth1
> >Time stamping parameters for eth1:
> >Capabilities:
> >	hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
> >	hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
> >	hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
> >PTP Hardware Clock: 1
> >Hardware Transmit Timestamp Modes:
> >	off                   (HWTSTAMP_TX_OFF)
> >	on                    (HWTSTAMP_TX_ON)
> >Hardware Receive Filter Modes:
> >	none                  (HWTSTAMP_FILTER_NONE)
> >	all                   (HWTSTAMP_FILTER_ALL)
> >
> >
> >-----Original Message-----
> >From: Gabe Black [mailto:Gab...@jd...]
> >Sent: 29 October 2013 01:18
> >To: Gabe Black; Keller, Jacob E
> >Cc: lin...@li...
> >Subject: Re: [Linuxptp-users] Which distributions have native ptp
> support?
> >
> >I downloaded ptp4l and compiled and ran it.  The behavior is the same,
> >except now poll times out.  I have increased the timeout and even
> >modified the code to see what packet was getting transmitted to verify
> >things.
> >
> >I verified that in wireshark I see the delay req message go out, and
> >the delay response as well.
> >
> >Looking at the code and reading Documentation/networking/timestamping
> >it looks like the code is trying to get the transmit timestamp of the
> >packet which is expected to be found in the socket error queue.
> >
> >So this leads me to believe that the timestamp is simply not making it
> >in to the MSG_ERRQUEUE...  Again, I have the default RH6.4
> >kernel/install and the igb driver seems to have support for it as
> shown below:
> >
> >ethtool -T eth3
> >Time stamping parameters for eth3:
> >Capabilities:
> >        hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
> >        hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
> >        hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
> >PTP Hardware Clock: 1
> >Hardware Transmit Timestamp Modes:
> >        off                   (HWTSTAMP_TX_OFF)
> >        on                    (HWTSTAMP_TX_ON)
> >Hardware Receive Filter Modes:
> >        none                  (HWTSTAMP_FILTER_NONE)
> >        all                   (HWTSTAMP_FILTER_ALL)
> >
> >ethtool -I eth3
> >driver: igb
> >version: 4.0.1-k
> >
> >Since RH6.4 still is on 2.6.32 kernel, I'm guessing they had to do a
> >bunch of work to back-port the stuff to get things like "ethtool -T"
> to
> >work.  Probably still buggy...
> >
> >Anyway, not sure what William did to get it to work...  I think I am
> >going to just switch to a different distribution that has it supported
> >out of the box as well.  Apparently on this thread Fedora 19 and
> >OpenSuse have it.
> >
> >Oh, I am running in a virtualized environment using DirectPath I/O
> >(a.k.a
> >"passthrough") to assign the I350 driver directly to the VM.  Maybe
> >that matters; but I'm fairly certain that DirectPath presents the
> >hardware registers/configuration in all its glory to the VM, thus
> >letting the igb-driver operate the device.  The software is the same,
> >so we should be getting timestamps...
> >
> >Oh well, on to the next distribution!
> >
> >Thanks!
> >Gabe
> >
> >
> >
> >> -----Original Message-----
> >> From: Gabe Black [mailto:Gab...@jd...]
> >> Sent: Monday, October 28, 2013 5:19 PM
> >> To: Keller, Jacob E
> >> Cc: lin...@li...
> >> Subject: Re: [Linuxptp-users] Which distributions have native ptp
> >> support?
> >>
> >> I am using the version that comes with a fresh install of the RH 6.4
> >> release:
> >>
> >> rpm -qa | grep ptp
> >> linuxptp-0-0.6.20121114gite6bbbb.el6.x86_64
> >>
> >> ptp4l doesn't appear to have a version option (in this release) and
> >> the readme file in /usr/shar/doc/linuxptp-0/README.org doesn't
> either.
> >>
> >> At any rate, I will get the latest version and try that and report.
> >>
> >> Thank you for the feedback!
> >>
> >> > -----Original Message-----
> >> > From: Keller, Jacob E [mailto:jac...@in...]
> >> > Sent: Monday, October 28, 2013 4:37 PM
> >> > To: Gabe Black
> >> > Cc: Ledda William EXT; lin...@li...
> >> > Subject: Re: [Linuxptp-users] Which distributions have native ptp
> >> > support?
> >> >
> >> > What version of linuxPTP are you using?
> >> >
> >> > It appears you aren't using the 1.3 version as we moved to a
> method
> >> > for obtaining the Tx timestamp that uses the poll() call. That
> >> > might fix your issue.
> >> >
> >> > The other option would be to increase the tx_timestamp_timeout
> >> > value, but I think moving to the newest Linux PTP should be a fix.
> >> >
> >> > Regards,
> >> > Jake
> >> >
> >> > On Mon, 2013-10-28 at 21:35 +0000, Gabe Black wrote:
> >> > > Nm, it was right.  We have two subnets that will route to the
> >> > meinberg, and they both work...  So that means I still  I don't
> >> > know what it is...
> >> > >
> >> > > strace it shows:
> >> > > sendto(11,
> >> > >
> >> >
> >>
> "\1\2\0,\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\2406\237\377\376\30*\213\0\1
> >> \
> >> > > 0\0"..., 44, 0, {sa_family=AF_INET, sin_port=htons(319),
> >> > > sin_addr=inet_addr("224.0.1.129")}, 16) = \
> >> > > 44
> >> > > Followed by a bunch of:
> >> > >
> >> > > recvmsg(11, 0x7fff67459250, MSG_ERRQUEUE) = -1 EAGAIN (Resource
> >> > temporarily unavailable)
> >> > > nanosleep({0, 1000}, NULL)              = 0
> >> > > recvmsg(11, 0x7fff67459250, MSG_ERRQUEUE) = -1 EAGAIN (Resource
> >> > temporarily unavailable)
> >> > > nanosleep({0, 1000}, NULL)              = 0
> >> > > recvmsg(11, 0x7fff67459250, MSG_ERRQUEUE) = -1 EAGAIN (Resource
> >> > temporarily unavailable)
> >> > > nanosleep({0, 1000}, NULL)              = 0
> >> > > recvmsg(11, 0x7fff67459250, MSG_ERRQUEUE) = -1 EAGAIN (Resource
> >> > > temporarily unavailable) ...
> >> > >
> >> > >
> >> > >
> >> > > > -----Original Message-----
> >> > > > From: Gabe Black
> >> > > > Sent: Monday, October 28, 2013 2:20 PM
> >> > > > To: Gabe Black; 'Keller, Jacob E'
> >> > > > Cc: 'Ledda William EXT'; 'linuxptp-
> us...@li...'
> >> > > > Subject: RE: [Linuxptp-users] Which distributions have native
> >> > > > ptp support?
> >> > > >
> >> > > > Oh sheesh.. I had a typo and was on the wrong subnet...
> >> > > >
> >> > > > It is working now!
> >> > > >
> >> > > > > -----Original Message-----
> >> > > > > From: Gabe Black
> >> > > > > Sent: Monday, October 28, 2013 1:18 PM
> >> > > > > To: 'Keller, Jacob E'
> >> > > > > Cc: Ledda William EXT; lin...@li...
> >> > > > > Subject: RE: [Linuxptp-users] Which distributions have
> native
> >> > > > > ptp support?
> >> > > > >
> >> > > > > > -----Original Message-----
> >> > > > > > From: Keller, Jacob E [mailto:jac...@in...]
> You
> >> > > > > > shouldn't have to, however, I would suggest disabling EEE
> >> > > > > > support, as this has been known to cause issues on the
> i350
> >> > > > > >
> >> > > > > > ethtool --set-eee device eee off
> >> > > > > >
> >> > > > > > This should fix your issue, if not please me know.
> >> > > > >
> >> > > > > Thank you for the reply.
> >> > > > >
> >> > > > > That option does not appear to be available with the stock
> >> > > > > igb driver of RH6.4
> >> > > > >
> >> > > > > [root@ network-scripts]# ethtool -i eth3
> >> > > > > driver: igb
> >> > > > > version: 4.0.1-k
> >> > > > > firmware-version: 0.1470, 0x05fc8000 ...
> >> > > > > [root@ network-scripts]# ethtool --show-eee eth3 Cannot get
> >> > > > > EEE
> >> > > > > settings: Operation not supported
> >> > > > >
> >> > > > > Is there another way to disable EEE? Or would not having
> that
> >> > > > > option in ethtool mean it is off?
> >> > > > >
> >> > > > > Gabe
> >> > > > >
> >> > > > > > On Mon, 2013-10-28 at 15:29 +0000, Gabe Black wrote:
> >> > > > > > > Hi William,
> >> > > > > > >
> >> > > > > > > I installed redhat 6.4 and have the intel i350 (with igb
> >> > > > > > > driver)
> >> > > > > and
> >> > > > > > am having trouble getting it to work with a meinberg
> master.
> >> > > > > > >
> >> > > > > > > The output I get is:
> >> > > > > > >
> >> > > > > > > ptp4l[246980.523]: selected /dev/ptp1 as PTP clock
> >> > > > > > > ptp4l[246980.525]: port 1: INITIALIZING to LISTENING on
> >> > > > INITIALIZE
> >> > > > > > > ptp4l[246980.526]: port 0: INITIALIZING to LISTENING on
> >> > > > INITIALIZE
> >> > > > > > > ptp4l[246980.773]: port 1: new foreign master
> >> > > > > > > 00606e.fffe.7c230e-
> >> > > > 1
> >> > > > > > > ptp4l[246985.014]: selected best master clock
> >> > > > > > > 00606e.fffe.7c230e
> >> > > > > > > ptp4l[246985.014]: foreign master not using PTP
> timescale
> >> > > > > > > ptp4l[246985.014]: running in a temporal vortex
> >> > > > > > > ptp4l[246985.014]: port 1: LISTENING to UNCALIBRATED on
> >> > > > > > > RS_SLAVE
> >> > > > > > > ptp4l[246985.989]: recvmsg tx timestamp failed: Resource
> >> > > > > temporarily
> >> > > > > > > unavailable
> >> > > > > > > ptp4l[246985.989]: port 1: send delay request failed
> >> > > > > > > ptp4l[246985.989]: port 1: UNCALIBRATED to FAULTY on
> >> > > > > > > FAULT_DETECTED
> >> > > > > > >
> >> > > > > > > Did you have to do anything differently on the stock
> >> > > > > > > RH6.4
> >> > > > install
> >> > > > > > > to
> >> > > > > > get it to work?
> >> > > > > > >
> >> > > > > > > Thanks,
> >> > > > > > > Gabe
> >> > > > > > >
> >> > > > > > > > -----Original Message-----
> >> > > > > > > > From: Ledda William EXT
> [mailto:Wil...@it...]
> >> > > > > > > > Sent: Monday, October 21, 2013 1:02 AM
> >> > > > > > > > To: Richard Cochran; Gabe Black
> >> > > > > > > > Cc: lin...@li...
> >> > > > > > > > Subject: RE: [Linuxptp-users] Which distributions have
> >> > > > > > > > native ptp support?
> >> > > > > > > >
> >> > > > > > > > I'm currently using RH 6.4 with Intel i350 (igb
> driver)
> >> > with
> >> > > > > > success
> >> > > > > > > > (no need to recompile the kernel). About RH 6.5 I know
> >> > > > > > > > that there will be many improvements, more eth driver
> >> > > > > > > > supported, and it should include version 1.3 of
> >> > > > > > > > linuxptp
> >> package.
> >> > > > > > > >
> >> > > > > > > > William
> >> > > > > > > >
> >> > >
> >> >
> >>
> >> --------------------------------------------------------------------
> -
> >> -
> >> -
> >> -------
> >> Android is increasing in popularity, but the open development
> >> platform that developers love is also attractive to malware
> creators.
> >> Download this white paper to learn more about secure code signing
> >> practices that can help keep Android apps secure.
> >>
> http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.
> >> c
> >> l
> >> ktrk
> >> _______________________________________________
> >> Linuxptp-users mailing list
> >> Lin...@li...
> >> https://lists.sourceforge.net/lists/listinfo/linuxptp-users
> >
> >----------------------------------------------------------------------
> -
> >---
> >----
> >Android is increasing in popularity, but the open development platform
> >that developers love is also attractive to malware creators. Download
> >this white paper to learn more about secure code signing practices
> that
> >can help keep Android apps secure.
> >http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.c
> l
> >ktr
> >k
> >_______________________________________________
> >Linuxptp-users mailing list
> >Lin...@li...
> >https://lists.sourceforge.net/lists/listinfo/linuxptp-users
> >
> >----------------------------------------------------------------------
> -
> >---
> >----
> >Android is increasing in popularity, but the open development platform
> >that developers love is also attractive to malware creators. Download
> >this white paper to learn more about secure code signing practices
> that
> >can help keep Android apps secure.
> >http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.c
> l
> >ktr
> >k
> >_______________________________________________
> >Linuxptp-users mailing list
> >Lin...@li...
> >https://lists.sourceforge.net/lists/listinfo/linuxptp-users