FPGA Intellectual Property
PCI Express*, Networking and Connectivity, Memory Interfaces, DSP IP, and Video IP
6343 Discussions

10g mac - avalon_st_tx_ready de-asserted in the middle of the packet transfer

Altera_Forum
Honored Contributor II
1,568 Views

Hi All, 

 

Does anybody know why avalon_st_tx_ready de-asserted in the middle of the packet transfer? 

Attached a capture of signal tap. 

 

I can see a missing data on the pc that receive this packet. the missing data is correlated with data transmitted while avalon_st_tx_ready de-asserted. 

 

Thanks
0 Kudos
12 Replies
Altera_Forum
Honored Contributor II
287 Views

That means the receiver cannot receive the data. If the data isn't held during this period then it will be lost.

0 Kudos
Altera_Forum
Honored Contributor II
287 Views

The data is transfered to the PC. 

Do you mean there is a problem with the PC's 10G card?
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

What is the ready latency parameter on the ethernet IP core?

0 Kudos
Altera_Forum
Honored Contributor II
287 Views

I am sending a 124Byte packet every 10 sec. I don't see how it can be a latency problem. 

What do you mean by pointing ready latency parameter? 

 

 

( "The ready latency on these interfaces is 0 and the MACexpects the empty signal to contain a valid value."----10-Gbps Ethernet MAC MegaCore Function User Guide----page 7-4)
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

If the ready latency = 0, what does the data change when ready is low? 

The length of the package from the (very low resolution) picture you sent appears to show 15 transactions, when a 124 byte package would require 16 transactions with tx_empty set when end of packet is asserted. 

Please post a better picture with more signals 

 

Next question - did you simulate this design before putting it in the FPGA?
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

better picture with more signals: 

 

At the top of the picture you can see the data comes in the fpga (from pc) 

At the bottom of the picture you can see the data transmited from the fpga to the output(to pc).
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

Without the code in the FPGA design, I dont know what you want me to do. 

The FPGA appears to be not working properly.
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

does anybody else has an idea what can be the problem.????? 

 

-- The problem can be seen also in simulation
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

If the problem is also in simulation, then it is clearly a problem with the fpga logic that generates the packets

0 Kudos
Altera_Forum
Honored Contributor II
287 Views

In Signal Tap : the packets are transmited by an external PC (was validate and works fine). 

In simuation : the same packets are transmited by a test bench. 

 

The fpga logic is ALTERA's IPs (MAC & PHY). no logic was added. Do you think altera has problem with this IP?
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

n Signal Tap : the packets are transmited by an external PC (was validate and works fine). 

In simuation : the same packets are transmited by a test bench. 

 

The fpga logic is ALTERA's IPs (MAC & PHY). no logic was added. Do you think altera has problem with this IP?
0 Kudos
Altera_Forum
Honored Contributor II
287 Views

I think it's highly unlikely a problem with altera IP - ethernet cores are very mature. 

You could try raising a ticket with them 

Without you posting the design - there is little we can do to help.
0 Kudos
Reply