Orin Nx 16GB restoring/reflashing fail - "Error: Unable to partprobe /dev/nvme0n1"

Robotics & Edge Computing Jetson & Embedded Systems Jetson Orin NX
1

Hi,

  1. We had flashed our Orin NX 16GB SOM with Jetson version r36.3.0 and delivered the unit to the customer.

  2. Later due to repeated abnormal power on and power off, UEFI assertion crash issue happened and unit was returned back to us as it was not booting.

  3. We flashed the QSPI internal memory with the updated TOS.img and later and it was booted successfully.

  4. Now, we want to do a complete re-flashing/restoring to NVMe memory for testing purpose.
    But to our surprise, we see some NVME related error occuring during the middle of the flashing/restoring process.
    Pls find the attached log files along with serial console log.

Kindly let us know, why this is happening, suddenly.

Thanks,
Nagesh R

**
21_Jan_serial_log_restore_fail.txt (125.2 KB)
21_Jan_Restore_NVME_Error_fail_log.txt (99.9 KB)
Flash_Error_Log_PCIE_EP.txt (271.8 KB)
**

4

Hi,

Could you try r36.4.0 to cross check the issue whether exist?

Thanks

5

We were able to successfully flash the latest version 36.4.3
The UEFI crash issue is solved now.

1)Our next step to reflash the unit as END Point device . But after adding ODMDATA line in the conf file, and reflashing, we observed the flash failed in the middle causing similar NVME related error as per my first post.

2)Any idea what might be the root cause?
We have customised carrier board.

  1. Is there any specific place where ODMDATA line has to be added inside conf file?

Please find the PCIe Endpoint device configuring flash log attached for your reference.
22_Jan_PCIE_Flash_error_log.txt (621.1 KB)

4)Also as per our hardware design team ,they have connected NVME to the PCIex4 lane, is this causing flashing issue and it is not able to recognise NVME??

5)As per the website:
https://docs.nvidia.com/jetson/archives/r36.4.3/DeveloperGuide/SD/Communications/PcieEndpointMode.html#flashing-pcie-as-endpoint-on-a-jetson-orin-nx-nano-series-system

we need to add this line in CONF file:

image
image872×165 26 KB

  1. Just want to know, what is the meaning of

hsio-uphy-config-41

where can we get documentation for it.

  1. Do you need our carrier board design schematics to cross check?
    or is there some thing, we can modify in ODMDATA itself which matches our PCIE design so that flashing happens?

Thanks

6

Just want to update the root cause for flash failing with CONF file updated with ODMDATA for configuring it as EndPoint Device mode:

  1. We found out the actual root cause for this issue.
  2. The issue was with the hardware design, where they have routed this NVMe on PCIe x4 lane, instead of PCIe x1 or x2 lane.
  3. lane x2 was used to connect the SBC from GPU, which is basically for only ROOT PORT mode , hence flash fails when you try to configure it for EndPoint device in .CONF file.

image
image1354×475 45.9 KB

  1. We reconfirmed by flashing just normal Jetson Linux with default CONF file( without any change by adding ODMDATA line for EP device mode in CONF file ) for older version r36.3.0 and newer version r36.4.3, both flash fine.

Thanks for the support.

7

Wanted to know, where do we get documentation for this below ODMDATA configuration.

image
image872×165 26 KB

what exactly HSIO-UPHY-CONFIG-1 mean?

8

Hi,

Please refer to this doc first

Thanks

1 Like
10

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