[Etherlab-users] Integration of 6631-0010 EC-Master 1.5.2
Matthias HAAS
matthias.haas at knapp.com
Mon Sep 20 18:29:54 CEST 2021
Hi everyone,
we are currently using EC-Master 1.5 with EoE on kernel 4.14.78-preempt-rt47 and the e1000e driver.
Now we have tried to update the EC-Master to 1.5.2 so that we can integrate the Beckhoff 6631-0010 Profinet Slave Module.
Version 1.5 works almost perfectly and I appreciate the work you have done.
Now with version 1.5.2 we have some problems with timing, I think....
we have seen errors about WC!=1 and lost datagrams, but when we look at Wireshark they are all there.
This error stacks up with the number of slaves on the bus.
Is it possible for the FSM to lose data? If so, how should I look for the leak?
What do I need for a slave, like the "Beckhoff 6631-0010 Profinet Slave" to configure my IO-like modules?
We also tried the igb driver for another chip on our motherboard, but it booted up the ram, so we crashed in about 15 minutes.
Our free ramsize at the time of the start was 1GB.
Does anyone know anything about a memory leak with this driver?
What is the best way to determine the bus cycle time, actually we have a frequency of 4000 on a 2Ghz CPU, is there some known value that works well?
We tried version 1.5.2 with the same setup and configuration and saw a lot more datagrams with Wireshark. Could this be a reason for the loss of datagrams because we can't process as much?
Thank you in advance
kind regrads,
Matthias Haas
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.etherlab.org/pipermail/etherlab-users/attachments/20210920/cf0d0ab8/attachment.htm>
More information about the Etherlab-users
mailing list