• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 30 subscribers
  • Views 914 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

TM4C1294KCPDT: LMFlash Ethernet Update Bug - Magic Packet Sent on Wrong Interface

Bryce Hillam81
Prodigy 50 points
Part Number: TM4C1294KCPDT
Other Parts Discussed in Thread: EK-TM4C1294XL

I'm using LMFlash to perform an Ethernet firmware update on a custom board which uses the TM4C1294KCPDT microcontroller. This issue also affects our custom boards which use the LM3S9D96, as well as the evaluation board EK-TM4C1294XL. When I click Program in LMFlash, the status bar indicates "Attempting to connect..." and never progresses. This issue also affects the eflash utility (although I don't know if you can call it a bug in that case, since eflash has no option to select the local interface to use).

I have observed (using Wireshark) that when more than one network interface is present on the system, LMFlash may send the magic packet over an interface different to the one it was told to use. I would speculate that it uses whichever adapter enumerates first in a call to the Winsock function gethostbyname(), since this is what eflash does.

A workaround is available - by disabling all network adapters in the Windows control panel (except the interface we are using), LMFlash will send the magic packet to the correct interface.

I have also observed that by sending the magic packet myself (using scapy) whilst LMFlash is in the "Attempting to connect..." state, I can trigger the target to send the BOOTP request. Once this is done LMFlash sends a BOOTP reply and the rest of the firmware update process completes successfully.

We require a fix for this bug since it complicates our support requirements (less technically-minded people need to be able to operate the software reliably).

  • 0
    TI__Guru**** 191296 points
    Hi Bryce,
    Thanks for bringing the issue to our attention. Unfortunately, there is no plan to update the LM flash programmer in the near term. Your workaround to disable the rest of the network interfaces will be a good hint to other community members facing the same issue in the future. Thanks for the suggestion.

    I suppose both network interfaces are connected to the same LAN with the same subnet. Is this correct?
  • 0 in reply to Charles Tsai
    Prodigy 50 points

    Hi Charles,

    The network interfaces are on different LANs.

    Are you able to confirm that our diagnosis of the issue is correct? We can only speculate without access to the source code for LMFlash. I would also appreciate if you were able to give some kind of timeline for the next scheduled update, if any.

  • 0 in reply to Bryce Hillam81
    TI__Guru**** 191296 points
    Hi Bryce,

    I think your diagnosis is correct. I think the problem is due to the network interfaces on different LAN. Unfortunately, as I mentioned before, there is no plan to update the LM flash programmer. However, there is a eflash.exe (ethernet flash downloader) with source code that you might want to take a look. The source code is under <TivaWare_Installation>/tools/eflash.