Fei Su
01/30/2013 9:58 PM
post98914
|
Hi ,
I'm using bpf to send and receive the EtherCat frame,but there are something strange than when I send a broadcast
ECAT frame and this frame is less then 60 bytes,I can't receive the frame form the slave, but I add a command 'nop' to
fill the frame size over 60 bytes ,I can receive from slave. I am using QNX6.5 and the ethernet card rtl8139. And also
when I use rtl8169,there is no problem like this.
|
|
|
Armin Steinhoff
01/31/2013 3:34 AM
post98919
|
Hi,
EtherCAT is based on STANDARD Ethernet media ... so you shouldn't use
packets smaller than the minimum packet size.
--Armin
http://www.steinhoff-automation.com
Fei Su wrote:
> Hi ,
> I'm using bpf to send and receive the EtherCat frame,but there are something strange than when I send a broadcast
ECAT frame and this frame is less then 60 bytes,I can't receive the frame form the slave, but I add a command 'nop' to
fill the frame size over 60 bytes ,I can receive from slave. I am using QNX6.5 and the ethernet card rtl8139. And also
when I use rtl8169,there is no problem like this.
>
>
>
> _______________________________________________
>
> General
> http://community.qnx.com/sf/go/post98914
> To cancel your subscription to this discussion, please e-mail general-networking-unsubscribe@community.qnx.com
>
|
|
|
Fei Su
01/31/2013 3:49 AM
post98920
|
The driver or mac will padding for me or even I am padding it myself ,the problem is still occur.
|
|
|
Dave Brown
01/31/2013 9:42 AM
post98935
|
This is the Ethernet minimum . It could be that one driver is padding out the packet, while the other is not. Some of
the drivers maintain specific statistics for short packets being dropped.
Dave
-----Original Message-----
From: Fei Su [mailto:community-noreply@qnx.com]
Sent: January-30-13 9:58 PM
To: general-networking
Subject: EtherCat Problem
Hi ,
I'm using bpf to send and receive the EtherCat frame,but there are something strange than when I send a broadcast
ECAT frame and this frame is less then 60 bytes,I can't receive the frame form the slave, but I add a command 'nop' to
fill the frame size over 60 bytes ,I can receive from slave. I am using QNX6.5 and the ethernet card rtl8139. And also
when I use rtl8169,there is no problem like this.
_______________________________________________
General
http://community.qnx.com/sf/go/post98914
To cancel your subscription to this discussion, please e-mail general-networking-unsubscribe@community.qnx.com
|
|
|
Hugh Brown
01/31/2013 10:07 AM
post98942
|
I have checked the rtl driver and it is padding the packets, as I can ping
with 10-byte packets. There might be a different version of the chip that
doesn't pad the packets.
On 2013-01-31 9:42 AM, "Dave Brown (QNX)" <community-noreply@qnx.com>
wrote:
>This is the Ethernet minimum . It could be that one driver is padding out
>the packet, while the other is not. Some of the drivers maintain specific
>statistics for short packets being dropped.
>
>Dave
>
>-----Original Message-----
>From: Fei Su [mailto:community-noreply@qnx.com]
>Sent: January-30-13 9:58 PM
>To: general-networking
>Subject: EtherCat Problem
>
>Hi ,
> I'm using bpf to send and receive the EtherCat frame,but there are
>something strange than when I send a broadcast ECAT frame and this frame
>is less then 60 bytes,I can't receive the frame form the slave, but I add
>a command 'nop' to fill the frame size over 60 bytes ,I can receive from
>slave. I am using QNX6.5 and the ethernet card rtl8139. And also when I
>use rtl8169,there is no problem like this.
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post98914
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
>
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post98935
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
|
|
|
Fei Su
01/31/2013 10:15 AM
post98944
|
I am pretty sure that it will padding ,but I'm off duty,so I can try the ping command tomorrow.
|
|
|
Hugh Brown
01/31/2013 10:16 AM
post98946
|
Actually, please try a "ping -s10 ip.address" and see if that works.
On 2013-01-31 10:15 AM, "Fei Su" <community-noreply@qnx.com> wrote:
>I am pretty sure that it will padding ,but I'm off duty,so I can try the
>ping command tomorrow.
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post98944
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
|
|
|
Fei Su
01/31/2013 10:19 AM
post98947
|
Ok thx I will try it tomorrow.
|
|
|
Fei Su
02/01/2013 1:16 AM
post98980
|
hello I have tried the command in qnx and wince ,and the result is not the same.
this is capture by the wireshark on qnx:
|
|
|
Hugh Brown
02/01/2013 8:05 AM
post98989
|
What ping command were you using when you captured this? You also didn't
post the output from 'pci -v'.
On 2013-02-01 1:16 AM, "Fei Su" <community-noreply@qnx.com> wrote:
>hello I have tried the command in qnx and wince ,and the result is not
>the same.
>
>this is capture by the wireshark on qnx:
>
>
>
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post98980
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
|
|
|
Fei Su
02/01/2013 1:17 AM
post98981
|
hello I have tried the command in qnx and wince ,and the result is not the same.
this is capture by the wireshark on wince:
|
|
|
Fei Su
02/01/2013 1:25 AM
post98982
|
Because the frame from qnx is fcs error ,and the wireshark put the fsc to me ,so we can see 4 bytes more than the frame
from wince. But It seems like the 4byte are not fcs but padding.I guess that if we send frame less than 64byte(include
fcs) the driver will send the 4 padding data instead of the fsc.
|
|
|
Fei Su
02/01/2013 3:30 AM
post98983
|
Hi the conclusion above is wrong. Just now, I use the fpga to analyse the frame ,found that qnx pad the frame to 68
bytes (include the fcs). The fsc error is mentioned by wireshark just because the impc protocl. So please help me to
check why qnx padding more 4 bytes?
|
|
|
Hugh Brown
02/01/2013 8:27 AM
post98990
|
I think that I found the problem. Please try the attached driver.
On 2013-02-01 3:30 AM, "Fei Su" <community-noreply@qnx.com> wrote:
>Hi the conclusion above is wrong. Just now, I use the fpga to analyse
>the frame ,found that qnx pad the frame to 68 bytes (include the fcs).
>The fsc error is mentioned by wireshark just because the impc protocl. So
>please help me to check why qnx padding more 4 bytes?
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post98983
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
|
|
|
Fei Su
02/01/2013 12:11 PM
post98999
|
Thx ,i will try it next monday .
|
|
|
Fei Su
02/02/2013 4:59 AM
post99020
|
My problem is solved! thank for ur help. Would u mind telling me what's wrong in the driver?
|
|
|
Hugh Brown
02/02/2013 2:52 PM
post99021
|
The driver was padding packets to 64 bytes instead of 60 bytes. Glad it is
working for you.
On 2013-02-02 4:59 AM, "Fei Su" <community-noreply@qnx.com> wrote:
>My problem is solved! thank for ur help. Would u mind telling me what's
>wrong in the driver?
>
>
>
>_______________________________________________
>
>General
>http://community.qnx.com/sf/go/post99020
>To cancel your subscription to this discussion, please e-mail
>general-networking-unsubscribe@community.qnx.com
|
|
|
Fei Su
01/31/2013 10:16 AM
post98945
|
I am pretty sure that it will padding ,but I'm off duty,so I can try the ping command tomorrow.
|
|
|
|