On 07/12/2011 07:20 AM, Igor Smolyar wrote: > Hi all, > > I have some minor problem with large icmp packets on Intel 82580 Gigabit > Fiber NIC that use igb 3.0.22 driver. > > When mtu set to 1500 I can ping remote host with large icmp packets (larger > than 1500) > with no errors and fragmentation is working properly. > > When mtu set less than 1500 on interface (e.g 1200) and I am trying to send > icmp packet larger than mtu (e.g 1500) > then icmp reply for last fragment is being dropped. > > I noticed that each time icmp reply dropped following counters incremented > by one in ethtool statistic: > > rx_long_length_errors > rx_errors > rx_length_errors. > > > > Topology is quite simple, 82580 Gigabit Fiber NIC installed on gateway > > gateway host > ip 5.5.5.5 ----------------------------------------- ip 5.5.5.1 > mtu 1200 mtu 1500
This topology is invalid. When the MTU is set to 1200 it is expected that the maximum frame size received will not be greater than 1214. As such when the system with a 1200 MTU receives a frame larger than 1214 it will be dropped as an invalid packet and will be reported as an RX length error. Thanks, Alex ------------------------------------------------------------------------------ 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
