[etherlab-dev] 1 datagram UNMATCHED! / working counter changes

Dipl.-Ing. (FH) Florian Pose fp at igh-essen.com
Mon Aug 3 14:05:25 CEST 2009


Am 03. August um 08:58 schrieb Daniel Bechter:
> BTW is there a parser for pcap sniffs which can perform some EtherCAT
> datagram analyzation? Or is there something that can analyze the frames
> within the debug ring?

Wireshark has an EtherCAT protocol dissector on board.

> What is meant by "unanswered datagram"? How does performance of the PC
> running the master play a role in this scenario?

I can't explain this, because there should be no unmatched datagrams
during operation, as long as frames are not trashed on the bus.

> Hope I am not missing the most important parts:
> Aug  3 10:27:13 SWES-Linux kernel:  00 00 0cU E]<7>[3250077.7103C <7A7T40d:
> U50Et: Received 07r40<7>[3250077.7107 a7E.t00R 307E.TA 00 00 00 07179G:
> sendin070B8 <0rE 0 rd 3C 00 00 00h7BC1 00 00 00D C0 91A]G      
> Aug  3 10:27:13 SWES-Linux kernel: [37BC0 00 00 00 00 00DD7 D 8CU[9  
> Aug  3 10:27:13 SWES-Linux kernel: 
> Aug  3 10:27:13 SWES-Linux kernel: <37Bh3 00 00 00 00 00 D C3 D7g>t  00    


Your logs look quite distorted... I can't read anything out of this.
This looks like you have another problem. Does this also happen if the
EtherCAT master or your application is unloaded?

> I used the 3 GHz P4 and set up a loop with 10 kHz (user space app without
> realtime patches applied).

It depends on the width of your process data (the bandwidth is the
limiting factor). But this is typically enough for most cases.

Best regards,
Florian Pose


More information about the Etherlab-dev mailing list