Our system design used two ORIN NX core boards, and we encountered a download exception

We use two NX core boards. When both NX core boards are installed on the seat, the download will fail. What is the reason? Most likely which IO affected the download process?

The NX device cannot be found during the boot up phase. Expectation can be solved.

Hello @liupeng21,

Would it be possible for you to share a bit more detail on the issue you are experiencing ?

  1. What are you trying to accomplish?
  2. What download is failing ?
  3. Could you describe your use case?

We just want to understand a bit better to make sure we are on the same page.

regards,
Andrew
Embedded Software Engineer at ProventusNova

We use a baseboard with two NX core boards on top;

Two NX core boards are connected via Ethernet SWITCH to exchange data within the LAN.
Now the question is:
If the core board A is powered on or is in recovery mode, the image download fails when the core board A is downloaded through USB, indicating that it fails in the boot-up phase.
The details are waiting for target to boot-up and then timeout.

It has been proven so far,
If core board A is powered on or in recovery mode, core board B fails to be downloaded. If I remove core board A, then core board B can be downloaded successfully at this time.
If the fault persists in separate power supply scenarios, remove the influence of power supplies.

Two logs of the download process are saved respectively, one is of the successful download process and the other is of the failure. Please help to confirm whether you can confirm the causes of the download failure.

I made a partial effort and found some problems, but couldn’t confirm what the deeper cause was. Please help analyze, rest
fail.txt (94.3 KB)
success.txt (125.0 KB)


Partial differences between successful and failed logs

Looking forward to your reply

抱歉用中文請問一下 你所謂的download到底是什麼意思? 是燒錄的意思嗎?
通常我們會用flash而不是用download

如果是flash失敗能請你能附上完整的host端加上device端log嗎?


What is the exact meaning of “download” in your comment? Is it “flash failure”?

If it is flash failure, could you attach full host side log and also device side log?

是的,是燒錄的意思

上面的附件是device端的Log;我没有host端的log,不知道怎么导出。

图片是失败时的截图。

你貼的東西就是host端log了 麻煩要貼log就全給文字檔.
我們需要看全部資訊, 不是只是這樣一段截圖

另外你有用過非VM的host做燒錄過嗎?

没有,我们都是在VM烧录的,我抓取一下Log给到您这边。

另外請問一下你這邊的敘述大概是什麼意思? 不太懂為何兩個board會影響對方
你是同時在燒兩片板子嗎

不是,两个板子,其中一个正常开机状态,然后烧录另外一个。
这两个中间通过网口交换机连接在一起,然后就是一些IO引脚控制MAX96712这个芯片,除此之外两个核心板没有直接连接的部分。

这也是我不理解的地方,一个开机,烧录另一个100%不行。但是当一个上电未开机时,另一个板子下载就没有问题。

1.txt (247.0 KB)

附件是整个下载过程的Log,帮忙看一下是什么导致的烧录失败。

從你前面UART serial console的log上來看

成功跟失敗的差別其實是在USB device mode沒有開起來
失敗的時候USB device mode直接沒有反應

你這應該是自己做的板子而不是NV devkit吧?

使我们自己做的板子,这个USB device mode是boot阶段自动检测状态的对吧,这个和外部硬件存在什么关联吗?

先根據我過去的經驗跟你提問一下…

請問你有幫自己的板子寫好device tree讓他有支援usb device mode嗎?
有很多用戶以為這件事情不用做, 或是巧合讓他們剛好能進device mode所以一直認為自己的設定是ok的.

不太確定你的自動檢測是什麼意思. 就是OTG的功能要正常使用. software設定如果沒有對的話也會有問題
由於你在使用VM進行燒錄, 又增加了一個變因

請問你是兩台的usb flash port同時都有接在VM host上嗎

那没有,只连接了一个在VM上烧录;

你的"那沒有"是指說你從來沒幫你的USB device tree做設定嗎?

指的是,每次只连接一个核心板去烧录,VM上只能看到一个英伟达设备端口;你说的设备树设定是什么意思?