Ethernet Frame Check Sequence Incorrect

Can you please attach your Wireshark capture file?

Please also see this post that discusses this issue:

    http://e2e.ti.com/support/embedded/bios/f/355/t/270257.aspx

Steve

To tell the truth, That post you mentioned is mine I posted.

and I don't understand this problem yet.

so I attached wireshark screenshot.

Please be more specific. Thanks a lot !!

I think this is happening because the driver is configured not to send/receive the frame check.  I did a quick test and noticed that it's always 0x0.

I see some code in the driver (in the NSP product, under ti/drv/omapl138/csl_emac.c) that checks this:

    /* Set the pass RX CRC mode and adjust max buffer accordingly */
    if (localDev.Config.ModeFlags & EMAC_CONFIG_MODEFLG_RXCRC)
    {
        if(corenum == localDev.Config.CoreNum) {
            CSL_FINST(EMAC_REGS->RXMBPENABLE, EMAC_RXMBPENABLE_RXPASSCRC, INCLUDE);
        }
        localDev.PktMTU = (pEMACConfig->PktMTU) + 4;
    }
    else
        localDev.PktMTU = (pEMACConfig->PktMTU);

I think you may need to configure the driver using the code in bold above and rebuild.  You may need to do it in the function HwPktOpen of ti/drv/omapl138/ethdriver.c.

You can find some more info on this in spruh77a.pdf "OMAP-L138 DSP+ARM Processor Technical Reference Manual", section "19.2.10.7 Receive Frame Classification"

Steve

Steve, thanks for your reply so much !!

I checked the file "ti/drv/omapl138/csl_emac.c".

The file is same with your mention. 

/* Set the pass RX CRC mode and adjust max buffer accordingly */
    if (localDev.Config.ModeFlags & EMAC_CONFIG_MODEFLG_RXCRC)
    {
        if(corenum == localDev.Config.CoreNum) {
            CSL_FINST(EMAC_REGS->RXMBPENABLE, EMAC_RXMBPENABLE_RXPASSCRC, INCLUDE);
        }
        localDev.PktMTU = (pEMACConfig->PktMTU) + 4;
    }
    else
        localDev.PktMTU = (pEMACConfig->PktMTU);

How can I do next step after checking this file? This job is very difficult. 

so I need your help a lot, sorry, but I will wait for your reply....

Good weekend ^^

I corrected HwPktOpen() function of ti/drv/omapl138/ethdriver.c like following and save file.

uint HwPktOpen( PDINFO *pi )

{

...

//    ecfg.ModeFlags = EMAC_CONFIG_MODEFLG_RXOFFLENBLOCK |
//                                  EMAC_CONFIG_MODEFLG_RMII;

      ecfg.ModeFlags = EMAC_CONFIG_MODEFLG_RXOFFLENBLOCK |
                                    EMAC_CONFIG_MODEFLG_RMII |
                                    EMAC_CONFIG_MODEFLG_RXCRC;

...

}

and compiled my project.

run and got the packets between board and PC by wireshark.

But I still got FCS Incorrect packet. 

How can I do next ?

I've been digging into this a little. I think what I told you before is the wrong flag.  I think that is for the receive side only.

I think the flag that matters is "EMAC_DSC_FLAG_PASSCRC".  From the document mentioned above:

19.2.5.4.11 Pass CRC (PASSCRC) Flag
This flag is set by the software application in the SOP packet descriptor before it adds the descriptor to the
transmit queue. Setting this bit indicates to the EMAC that the 4 byte Ethernet CRC is already present in
the packet data, and that the EMAC should not generate its own version of the CRC.
When the CRC flag is cleared, the EMAC generates and appends the 4-byte CRC. The buffer length and
packet length fields do not include the CRC bytes. When the CRC flag is set, the 4-byte CRC is supplied
by the software application and is already appended to the end of the packet data. The buffer length and
packet length fields include the CRC bytes, as they are part of the valid packet data. Note that this flag is
valid on SOP descriptors only.

This is new territory for me, but from what I can tell, this bit is not set, so the EMAC should be calculating and setting the CRC.  However, in Wireshark, I too see that the CRC is always 0.  I'm not sure why the EMAC computation of the CRC for all Ethernet frames is 0 ...?

One thing you could try is set the PASSCRC to true.  But this will require you to compute the check sum yourself, and ensure that the buffer length is correct (e.g. you must be sure to comply to the parts in bold above).

I *think* the part of the code to do this is in csl_emac.c, in the function emacEnqueueTx:

            if (pPkt->Flags & EMAC_PKT_FLAGS_SOP) {
                pDescThis->PktFlgLen = ((pPkt->Flags&
                                       (EMAC_PKT_FLAGS_SOP|EMAC_PKT_FLAGS_EOP))
                                       |pPkt->PktLength|EMAC_DSC_FLAG_OWNER|EMAC_DSC_FLAG_PASSCRC);
            }
            else {
                pDescThis->PktFlgLen = (pPkt->Flags&EMAC_PKT_FLAGS_EOP)
                                       |EMAC_DSC_FLAG_OWNER;
            }

However, I don't think the above is complete.  In addition to adding the above bolded code, you would need to compute the CRC, append it to the packet data, and update the buffer length.

Steve

Thanks steve !!

So sorry I am a beginner so I don't have a lot of knowledge about this part.

and I don't understand your mention especially last sentence "In addition to adding the above bolded code, you would need to compute the CRC, append it to the packet data, and update the buffer length".

I corrected the file csl_emac.c like above your mention.

But I will not compute the CRC, and update the buffer length. 

I need your help again.

I will wait for your reply. thanks!!

Thank you so much for a long time !!

I will be relieved at last...

Just for reference, this discussion also came up here, and referenced this answer in a different forum.

Could you please have a look at this question also?

http://e2e.ti.com/support/dsp/davinci_digital_media_processors/f/99/p/288506/1011979.aspx#1011979