This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TM4C1294NCPDT issue: EMAC: ping OK, but socket can't be connected on some boards.

I have a board, which uses integrated EMAC. Now I found that with the exactly same software, on some boards, socket can't be connected, while one some other board it can. But ping is OK on both boards. I can't understand why. Is it hardware issue, or software issue?

I am using CCS6, tirtos_tivac_2_10_01_38 and its NDK. MAC address has been programmed using LM Flash Programmer.

Does anybody have any idea? Thanks.

  • Hi JB,

    >But ping is OK on both boards

    Ping is a unicast broadcast set by default in LWIP. Ping is no indication a unique TCP/IP address has been assigned to the socket on the network.

  • Thanks.
    So if ping is OK, can I make the conclusion that the HW is OK? If so, what makes some board fail to acquire a TCP/IP?
  • Think a unicast ping and or response does not use an IP address to connect to a socket. Unicast Ping uses all ones address 255.255.255.255 or all zero 0.0.0.0

    Believe the MAC address is more important in unicast ping (is like a broadcast) and the destination IP address is dropped.

    >can I make the conclusion that the HW is OK?

    Not sure but be sure to check the DHCP server on your network each MAC in ? has an unique IP address assigned in the lease. This is not Windows networking client that tests for duplicate IP, duplicate IP's on the network do very strange things.

  • I don't use DHCP but static IP instead. I have tried two cases, and both failed to connect a socket, and ping is OK in both cases:
    1. Connect the board with PC with a direct cable.
    2. Connect them using a hub/router.

    So I think the IP is unique in the network.
  • Try to use DHCP and see if the DHCP server has assigned a lease from the pool of addresses. Static IP are typically reserved for server/host on networks, sometimes a device but rarely yet never a client.
  • >Is it hardware issue, or software issue?
    That is debatable and DHCP client would help to uncover what is going on. 10 to 1 the failed board can not pull an DHCP IP address.
    WIndows ping uses the destination address to form a MAC frame with the client once ARP returns the MAC address.

    Have you typed (ARP -a) at Windows CMD prompt to see if a MAC frame was formed during unicast ping?

    >socket can't be connected

    What leads you to that conclusion?
  • BTW:
    Seems you have given us no details about the client: Network type IPV4/6 IP address, gateway address and the mask address.
  • This issue seems to be caused by illeagal/invalid MAC address written in the User0/User1. After a change a new MAC address, socket connection is OK.
  • So wouldn't you agree (arp -a) confirms that was the case and a DHCP assigned address confirms such true as well?

    In my opinion the question was answered more than once - this ain't my first rodeo with TCP!