ZC702 EMACPS example

[quadros1]

I am writing an Ethernet driver for a network stack product on ZC702 board for use with RTXC Quadros RTOS

I began by using the EMACPS example code that is provided in the SDK. I have it all built. it runs, and I get an interrupt as expected on the first packet out.

but the interrupt handler then reports HRESP NOT OK error along with other errors. I am running your example code 99.99% as is.

has that example ever been used on your ZC702 board?

what might be wrong that causes HRESP NOT OK error?

I am downloading the code after uBOOT starts, and before Linux is loaded. I am using latest DS-5 release. it appears to me that there may be some board setup code missing somewhere for support of Ethernet controller..

if that EMACPS example application code is not for ZC702, is there a better example of bare-metal support for the Ethernet subsystem on the ZC702 board?

bill

****************************

Bill Dittmann    <")))><

Chief Engineer

Quadros Systems, Inc.

13850 Gulf Freeway, Suite #122

Houston, TX 77034

 
bill.dittmann@quadros.com

ofc: 832 351 2830 x103

cel:  281 450 6836

****************************

[austin]

Bill,

Are you still having this problem?  Did anyone get back to you?

If not, reply back here.

[quadros1]

i have not yet received any reply.

i am getting DMA like errors and not sure what is wrong.  is the sample code supposed to work on the ZC702 board?

i can be reached at 832 351 2830 x103.

[quadros1]

i have not yet received any reply.



i am getting DMA like errors and not sure what is wrong. is the sample code supposed to work on the ZC702 board?


i can be reached at 832 351 2830 x103.

[harald.walter]

Hi,

I am facing the same issue. After transmission of the first frame the TX_STATUS.tx_corr_ahb_err flag is set. The TX descriptor also returns in word 1 with bit 27 set to one (Transmit frame corruption due to AHB error).

All subsequent trials to send a frame fail because the TX descriptor Used flag is never set back to one by the controller.

However, I am not using the example code.

What can cause a HRESP error?

Regards,

Harald

[bwiec]

What HRESP error are you getting?

http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.ddi0229c/ch06s05s02.html

[siktap]

This was a known issue where the Ethernet Interrupt DMA example appears to hang on user 702 board during transmit. If I step into the codes, the program does not return when it runs into this:

       /* Start transmit */

       XEmacPs_Transmit(EmacPsInstancePtr);

So the application hangs.But alternatively, I add the peripheral tests example applications which includes the Ethernet DMA Interrupt example, then it passes the test.

You can follow the following steps to run this test:

Right click on "Project Explorer" -> New -> Projects -> Xilinx Application Project -> Next -> name it for example "peripheral_test" -> Next -> Choose "Peripheral Tests"  -> Finish

You can see the Ethernet dma test also passes. It looks to me that we need to include the interrupt setup example from the peripheral tests and run the function prior to the Ethernet example to get the system working.

This issue is fixed in 2013.3

[saurabhnair]

Hello siktap,

I ran the peripheral test program on the hardware using the "Run on Hardware" option, and it runs successfully.

After running the peripheral test code, I ran the Ethernet Example code "XEmacPs_example_intr_dma.c" and yet it seems to hang.

The peripheral test seems to use the same functions as the example code and yet this hanging :D

Any suggestions? (I'm using Xilinx 14.6 if you want to know)

Thanks

Saurabh

[siktap]

Yes thats right. That is the issue here. That is beacuse the Ethernet Example code has few files missing.

Hence currently if your peripheral tests are passing for Ethernet, it should take care for now.

[saurabhnair]

Thanks for the quick reply siktap.

May I ask which files are missing, because I guess they must be present for the peripheral test to work.

Again thanks for the quick response.

Saurabh

[siktap]

This is basically a known bug that was reported and fixed. Something was broken in the drivers chain. This is fixed in 2013.3