[Linuxptp-users] PTP Clock completely wrong on an Intel 82579L

[Keller, Jacob E]

-----Original Message-----
Sent: Wednesday, February 22, 2017 1:20 PM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on a Intel 82579L
for out-of-tree drivers. Still waiting approval.
I am not so sure about a driver bug, all those intel chips are rather
similiar. Will try to read out the hardware PTP clock directly.

Ok that should still hit the right person. Driver bug, because each hardware runs at different clock rates and the likely problem is that the software is using the wrong values to take that into account.

Thanks,
Jake

[Norbert Lange]

ptp doesnt work on any intel hardware, this scientific proof was based
on samplesize 2. Sent another mail to that list.
Just out of curiosity, what hardware is this tested on, and verified
to work as master?

Kind regards,
Norbert

Post by Keller, Jacob E

-----Original Message-----
Sent: Wednesday, February 22, 2017 1:20 PM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on a Intel 82579L
for out-of-tree drivers. Still waiting approval.
I am not so sure about a driver bug, all those intel chips are rather
similiar. Will try to read out the hardware PTP clock directly.

Ok that should still hit the right person. Driver bug, because each hardware runs at different clock rates and the likely problem is that the software is using the wrong values to take that into account.
Thanks,
Jake

[Ledda William EXT]

Post by Norbert Lange
ptp doesnt work on any intel hardware

Probably with the e1000 driver. I have an Intel i350 with the igbx driver. I'm using it since 4 years as both master and slave and I have never had any issue.


-----Original Message-----
From: Norbert Lange [mailto:***@gmail.com]
Sent: 23 February 2017 10:19
To: Keller, Jacob E
Cc: linuxptp-***@lists.sourceforge.net
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on an Intel 82579L

ptp doesnt work on any intel hardware, this scientific proof was based on samplesize 2. Sent another mail to that list.
Just out of curiosity, what hardware is this tested on, and verified to work as master?

Kind regards,
Norbert

Post by Norbert Lange

-----Original Message-----
Sent: Wednesday, February 22, 2017 1:20 PM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on a Intel 82579L
for out-of-tree drivers. Still waiting approval.
I am not so sure about a driver bug, all those intel chips are rather
similiar. Will try to read out the hardware PTP clock directly.

Ok that should still hit the right person. Driver bug, because each hardware runs at different clock rates and the likely problem is that the software is using the wrong values to take that into account.
Thanks,
Jake

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most engaging tech sites, SlashDot.org! http://sdm.link/slashdot _______________________________________________
Linuxptp-users mailing list
Linuxptp-***@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linuxptp-users

[Richard Cochran]

Post by Norbert Lange
ptp doesnt work on any intel hardware, this scientific proof was based
on samplesize 2. Sent another mail to that list.
Just out of curiosity, what hardware is this tested on, and verified
to work as master?

See:

http://linuxptp.sourceforge.net/

5.3 Driver Support Matrix


Thanks,
Richard

[Norbert Lange]

Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.

Kind regards,
Norbert

Post by Richard Cochran

Post by Norbert Lange
ptp doesnt work on any intel hardware, this scientific proof was based
on samplesize 2. Sent another mail to that list.
Just out of curiosity, what hardware is this tested on, and verified
to work as master?

http://linuxptp.sourceforge.net/
5.3 Driver Support Matrix
Thanks,
Richard

[Richard Cochran]

Post by Norbert Lange
Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.

Please take a closer look. Neither the 82579 nor the i354 are
mentioned on that page.

Thanks,
Richard

[Keller, Jacob E]

-----Original Message-----
Sent: Thursday, February 23, 2017 2:20 AM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on an Intel 82579L
Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.
Kind regards,
Norbert

The igb driver is likely bugged, it should be maintaining a separate "real counter" and the internal hardware counter wrapping should be hidden from the user.

Most likely in all these cases you have hardware the driver recognizes, and assumes a certain hardware frequency, but the software was not updated for the new frequency rate, so it is assuming a different internal rate which results in these problems.

Hopefully one of the e1000e maintainers will be able to give more insight on the intel-wired-lan list.

Thanks,
Jake

[Richard Cochran]

Post by Norbert Lange
Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.

Looks like you aren't the only one with this issue:

https://www.spinics.net/lists/kernel/msg2446767.html

So it does look like a bug where driver the driver selects the wrong
clock values.

Cheers,
Richard

[Keller, Jacob E]

-----Original Message-----
Sent: Sunday, February 26, 2017 2:53 AM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on an Intel 82579L

Post by Norbert Lange
Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.

https://www.spinics.net/lists/kernel/msg2446767.html
So it does look like a bug where driver the driver selects the wrong
clock values.
Cheers,
Richard

Yep that's what it sounded like. Glad that appears resolved now.

Regards,
Jake

[Norbert Lange]

Good,

any estimate when this ends up in Linux 4.9?

Also, is the other issue [1] looked at?
Just want to know if I need to poke somewhere / someone, the mailing
list seems to more of a patch tracker.

Kind Regards,
Norbert Lange

[1] - http://lists.osuosl.org/pipermail/intel-wired-lan/Week-of-Mon-20170220/008171.html

Post by Keller, Jacob E

-----Original Message-----
Sent: Sunday, February 26, 2017 2:53 AM
Subject: Re: [Linuxptp-users] PTP Clock completely wrong on an Intel 82579L

Post by Norbert Lange
Yeah, I seen that matrix, and both NICs are on that list. Still have
some rather serious issues, one is completely useless (82579LM, e1000e
driver) and the other (i354, igb driver) apparently wraps the timer
very 18 minutes, losing their master role.

https://www.spinics.net/lists/kernel/msg2446767.html
So it does look like a bug where driver the driver selects the wrong
clock values.
Cheers,
Richard

Yep that's what it sounded like. Glad that appears resolved now.
Regards,
Jake