>-----Original Message----- >From: Dan Tucholski [mailto:[email protected]] >Sent: Monday, June 27, 2011 9:07 AM >To: [email protected] >Subject: [E1000-devel] Driver issue with I340-T2/hardware timestamping > >Hello, > >I am doing research with two of your intel I340-T2 gigabit network cards >and I am running into some technical issues possibly with the driver. >We >are trying to use the Precision Time Protocol daemon (PTPd) in order to >get two linux machines both using those networks cards to synchronize >their clocks through hardware timestamping. In all of your support for >these network cards, it says that hardware timestamping is possible, but >in actually doing the experiment, it seems as if the driver is not fast >enough to timestamp all packets within nanoseconds. Two packets always >need to be timestamped within nanoseconds of each other and the driver >only seems to allow stamping of one of them everytime and fails to give >the second a timestamp at all. When sleeps were added to the test code >to >increase the time between packets needing to be stamped, the driver >executed properly timestamping all packets. Just wanted to clarify this >was a driver issue and see if there was any other solution or further >support. > >Here are my specs: > >Intel Corporation 82580 Gigabit Network Connection >Intel Corporation Ethernet Server Adapter I340-T2 > >Driver Version: 3.0.22 > >Linux kernel 2.6.32-5-686 > >Thank you, >Dan Tucholski > > >------------------------------------------------------------------------ >------ >All of the data generated in your IT infrastructure is seriously >valuable. >Why? It contains a definitive record of application performance, >security >threats, fraudulent activity, and more. Splunk takes this data and makes >sense of it. IT sense. And common sense. >http://p.sf.net/sfu/splunk-d2d-c2 >_______________________________________________ >E1000-devel mailing list >[email protected] >https://lists.sourceforge.net/lists/listinfo/e1000-devel >To learn more about Intel® Ethernet, visit >http://communities.intel.com/community/wired
Hello Dan, I'll need to research a bit to see if I can find a reason for the behavior you're seeing. I'm not aware of anything off the top of my head specific to this part and the timestamp feature. This feature has been tested as working, but I'm not sure exactly of testing that was done. If you open an issue at sourceforge.net we'll have a place to keep track of data. Can you send me a full lspci -vvv of your system. Also, separately, a run of lspci with -t. Can I get a copy of your .config? How much delay did you need to add to get the second packet stamped with this part? What is your network connectivity? 1Gb? What brand of switches are you using? How are you doing the test, through switches or back to back? Thanks, Carolyn Carolyn Wyborny Linux Development LAN Access Division Intel Corporation ------------------------------------------------------------------------------ All of the data generated in your IT infrastructure is seriously valuable. Why? It contains a definitive record of application performance, security threats, fraudulent activity, and more. Splunk takes this data and makes sense of it. IT sense. And common sense. http://p.sf.net/sfu/splunk-d2d-c2 _______________________________________________ E1000-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/e1000-devel To learn more about Intel® Ethernet, visit http://communities.intel.com/community/wired
