No sign of life after successful flash using SDK Manager

Please provide the following info (tick the boxes after creating this topic):
Software Version
DRIVE OS 6.0.10.0
DRIVE OS 6.0.8.1
DRIVE OS 6.0.6
DRIVE OS 6.0.5
DRIVE OS 6.0.4 (rev. 1)
DRIVE OS 6.0.4 SDK
other

Target Operating System
Linux
QNX
other

Hardware Platform
DRIVE AGX Orin Developer Kit (940-63710-0010-300)
DRIVE AGX Orin Developer Kit (940-63710-0010-200)
DRIVE AGX Orin Developer Kit (940-63710-0010-100)
DRIVE AGX Orin Developer Kit (940-63710-0010-D00)
DRIVE AGX Orin Developer Kit (940-63710-0010-C00)
DRIVE AGX Orin Developer Kit (not sure its number)
other

SDK Manager Version
2.2.0
2.1.0
other

Host Machine Version
native Ubuntu Linux 20.04 Host installed with SDK Manager
native Ubuntu Linux 20.04 Host installed with DRIVE OS Docker Containers
native Ubuntu Linux 18.04 Host installed with DRIVE OS Docker Containers
other

Issue Description
Hi,

I wanted to update the DevKit to version 6.0.6 using SDK Manager. I had the “USB-C flashing” and the “DEBUG USB Micro-B” connected to the host machine. I tried first without putting the DevKit into recovery mode but the process was failing, and it only succeeded after switching to recovery mode using the MCU console.
After that (The fan and the lights were still on), I tried to access the MCU console to disable recovery mode but it was not responding, so I waited for about 10 minutes and tried again with no success. In the end, I decided to power cycle the DevKit. I switched off the power switch and unplugged the power cord. I then connected the power cord and switched on the power switch but the DevKit showed no sign of life since then (neither the fan nor any of the lights). After trying different sockets and power cords, I tried flashing the DevKit using SDK Manager again (as should be seen in the attached logs). I also tried to flash using Docker but the flash.sh /dev/ACM1 p3710 is failing with error (same error with all ttyACM[0-3] ports):
ERROR: Serial port Provided is not found /dev/ttYACM1. Check the AURIX_PORT env variable and reconnect the serial port and reboot host and target. Aborting..

When connecting the a DevKit to a laptop via the “USB-C flashing” nothing is detected by lsusb, unlike connecting via the “DEBUG (USB Micro-B)” port, where the /dev/ttyACM[0-3] ports appear in the system but there is no response (pressing Enter) on any of them when trying to access using minicom -D /dev/ttyACM1 (tried with the Hardware Flow Control Enabled and Disabled). The output of lsusb includes:
Bus 001 Device 007: ID 0955:7045 NVIDIA Corp. Tegra On-Platform Operator.

Reading through the “NVIDIA DRIVE OS Linux SDK Developer
Guide”, I reached the section “Flashing AURIX from Windows Host with Infineon Memtool”. I followed the instructions but I couldn’t find a way to download the hex files, so I found them in the Docker container. However, the “Aurix jumper” didn’t seem to easily reachable to put the Aurix into recovery mode, so I decided to add this post before going further.

I think issue is similar to Can not power on,after flash successfully. I followed all the suggestions and I had the same results.

Logs
driveinstaller.zip (935.8 KB)
sdkm.zip (214.1 KB)
logs.zip (138.3 KB)

Following are photos of the power connection, the power button, the USB connections and the output of the lsusb command:




Thanks.

Dear @mohannad.jooriah1,
Make sure you have connection from host to debug port and left type C USB port on target before flashing/accessing tegra or aurix
Check if any serial ports are detected at /dev/ttyACM*.
Check accessing all /dev/ttyACM ports using minicom. By default, /dev/ttyACM1 if for Aurix and /dev/ttyACM0 if for tegra.
Check changing connector cables.
If aurix is accessible, put the board in recovery mode from aurix(i.e tegrarecovery x1 on & tegrareset x1), you should see an additional NVIDIA device in lsusb on host.
Keep the board in normal mode(not recovery mode) before accessing/flashing tegra.
Use 6.0.6 docker container to flash to avoid host system state related issues.
Also, check running aurixreset and tegrareset on aurix console to see if it improves the situation.
Please let us know your observations.

Hello @SivaRamaKrishnaNV

Thank you for your response.

  • I made sure that the host is connected to the target’s debug port left USB-C port (as can be seen in the attached image).
  • I tried accessing all the /dev/ttyACM ports using minicom (with the HW flow control enabled7disabled), and none of them shows any response when pressing Enter within the minicom window.
  • I changed both connector cables and got the same results.
  • The Aurix is not accessible, so I cannot run any Aurix console commands.
  • I tried to flash using SDK Manager, but it failed. I attach the logs of two consecutive tries.
  • I used the 6.0.6 docker container to flash using the command `./flash.sh /dev/ttyACM[0-1] p3710, and it also failed. I attache the logs of flashing using ports ttyACM0 and ttyACM1.

Logs

Thanks.

I see the connections are good. If aurix can’t be opened, the flashing is expected to fail.

So did you see flashing successful message first time and you could not put the board in normal mode back? But lsusb indicates the board is normal mode now.

Is the board in working condition(which has 6.0.4) before trying to flash 6.0.6?

Yes, this is exactly what happened. I saw the flashing successful message in SDK Manager, but I was not able to access the aurix to activate the normal mode again.

Yes, the board was in perfect working condition (running 6.0.4) before trying to flash 6.0.6.

Dear @mohannad.jooriah1,
Could you share output of ls /dev/ttyACM*? I am assuming only one board is connected to host and followed minicom settings at To configure Minicom | NVIDIA Docs

Dear @SivaRamaKrishnaNV ,

Please find in the attached video, the output of the requested command and the confirmation of minicom configurations. When minicom was “connected” to the ports /dev/ttyACM0 and /dev/ttyACM1, I was pressing Enter but it was not showing any response. The setup that I used is shown in the attached image where only one board is connected to the host.

Dear @mohannad.jooriah1,
Connection between left type C port and host is missing. Check this connection as well.

Dear @SivaRamaKrishnaNV ,

I repeated the tests after connecting the left type C port to the host (as in the first attached image) and the results are still the same.

I also double checked the power cord using a multimeter, and it seems to be fine (second attached image). Neither the lights, not the fan is working with any power cord.

Thanks