Jetson AGX Some modules are abnormal after multiflash burning

We use( ./ nvmflash.sh)command multiple Jetson modules to burn together, prompting that several modules are abnormal after successful burning. Either keep printing alarms or get stuck。Serial port cannot be input。
This module was previously used( ./ flash. sh) commands burning is normal
alarms log:
WARNING: at platfom/drivers/mailbox/mail mach. c: 177
WARNING: at platfom/drivers/mailbox/mail mach. c: 161
stuck log.log (96.3 KB)

A custom board or devkit?

Which jetpack release is in use?

4.6

PCIe of exception jetson module is connected with X86. On the boot PCIe link, abnormal jetson modules can enter the system normally.
When the x86 device reboots, the PCIe of the jetson module will be reset. At this time, the serial port of the jetson module will print an abnormal log and cannot be used normally.
WARNING: at platfom/drivers/mailbox/mail mach. c: 177
WARNING: at platfom/drivers/mailbox/mail mach. c: 161

Sorry that I don’t know what you are talking about.

How about you directly put your comment in Chinese. I think you can tell more detail in such way.

启动异常的jetson板卡模块直接通过PCIE插到X86主板的PCIE上面,是可以正常进入系统,串口能输入。不插到x86上面PCIE没有link就会打印这个异常的log,无法正常进入系统,一段时间后模块会自动重启。

If you use flash.sh on the same module and it won’t have issue?

用./ nvmflash.sh烧写出现启动异常的模块,再用 flash.sh烧写,还是一样异常,只有模块pcie link才不会异常

Ok, so I can say this issue is purely PCIe. Not related to massflash, right?

Could you describe what change you’ve made to PCIe?

我没有修改PCIE,我之前烧的自己的软件都是用 flash.sh命令烧写都是正常的,只是这种方式效率比较低,改成了多个模块一起用nvmflash.sh烧写,还是烧写一样的软件,烧完之后启动就异常了,异常之后在用 flash.sh烧写还是异常的,只是这个异常模块在pcie link上时能正常。

I mean any kind of software configuration…

By default we didn’t enable PCIe ep…

PCIE在设备上还是要用的软件不能直接关掉。现在问题是同样的软件都是enable PCIE的,之前用 flash.sh是能正常进入系统,现在用nvmflash.sh烧写,就不行了,这是什么原因。有什么影响

At least tell me what you’ve changed in PCIe software. I am not asking you to disable something.

This is debug purpose.

之前用 flash.sh是能正常进入系统,现在用nvmflash.sh烧写,就不行了,这是什么原因。有什么影响

I don’t think this change is not able to recover. If such thing happened, I would consider to be hardware problem.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.