adi@rc:~$ cat /etc/nvidia/version-ubuntu-rootfs.txt
6.0.5.0-31732390
adi@rc:~$ sudo bash /etc/systemd/scripts/nv_aurix_check_fw.sh
[sudo] password for adi:
starting Aurix FW checking...
Checking arguments...
Checking Ping to AURIX
Ping to Aurix successful
Flashed UPDATE FW version is 1.48.01.00
Aurix Update : Main FW selected: AFW
Aurix Update : Alternate FW selected: NV
Rootfs FW version is 5.06.05.0, DRIVE-V6.0.5-P3710-AFW-Aurix-StepB-5.06.05
Response_From_Aurix
Flashed FW version is 5.06.05.00
Aurix is booted with Latest FW
Rootfs UPDATE FW version is 1.48.01.0
Flashed UPDATE FW version is 1.48.01.00
Aurix is flashed with Latest UPDATE FW
Shell>version
DDPX Aurix Serial Console
P3710-A05
with TLF35584 B/C-Step
SW Version 1.46.12
DRIVE-v6.x.x-P3710-NV-Aurix-IFW-StepB-1.46.12
Patch 0
* P3710-TS5 bringup firmware
* this binary does not support P3710 TS1 ~ TS3
TC397 Step BD
@jace.mogill1
From the syslog, looks like MAC addresses are missing from the Tegra:
Oct 21 03:08:50 tegra-ubuntu kernel: [ 1.817831] ether_get_mac_address_dtb: bad mac address at /chosen/nvidia,ether-mac0: NULL.
Oct 21 03:08:50 tegra-ubuntu kernel: [ 1.817843] nvethernet 2310000.ethernet: No MAC address in local DT!
Oct 21 03:08:50 tegra-ubuntu kernel: [ 1.817844] nvethernet 2310000.ethernet: failed to get MAC address
From the Tegra console please do this to confirm if MAC addresses are there or not.
adi@tegra-ubuntu:~$ sudo i2cdump -y -r 0x44-0x49 0 0x50
No size specified (using byte-data access)
0 1 2 3 4 5 6 7 8 9 a b c d e f 0123456789abcdef
40: 98 d9 85 2d b0 48 ???-?H
I’m not sure what you mean by “need to be re-flashed”, I’ve repeatedly tried flashing this machine with the SDKManager using 6.0.4 and 6.0.5, scripts you recommended, and the container, but re-flashing this machine is not fixing this problem.
From Aurix console put the board in recovery:
> tegrarecovery x1 on
> tegrareset x1
Then from another window on the host make sure two nvidia USB devices appear:
$ lsusb -d 0955:
flash with MAC addresses reprogrammed
$ cd ~/nvidia/nvidia_sdk/DRIVE_OS_6.0.5_SDK_Linux_DRIVE_AGX_ORIN_DEVKITS/DRIVEOS/drive-foundation
$ make/bind_partitions -b p3710-10-s05 linux
$ cd tools/flashtools/bootburn
$ ./bootburn.py -b p3710-10-s05 -B qspi --customer-data {Absolute Path to}/customer_data_orin.json
Once flashing is complete do the followng on the target again to confirm MAC addresses are present:
cat /proc/device-tree/chosen/nvidia,ether-mac0
cat /proc/device-tree/chosen/nvidia,ether-mac2
cat /proc/device-tree/chosen/nvidia,ether-mac4
cat /proc/device-tree/chosen/nvidia,ether-mac6
cat /proc/device-tree/chosen/nvidia,ether-mac8
Because the serial number does not appear on the Orin itself, only on the cardboard shipping box, I don’t know if the machine we’re working on now is S/N 1412422000100 or 1412422000069. How can I confirm the S/N for this hardware?
Besides reprogramming MAC address, please elaborate on what had been done on the system before the first time aware of network issues and help us figure out why the MAC address got erased.
Except flashing with sdkmanager or docker (following DRIVE OS 6.0.5 Installation Guide for NVIDIA Developers), had you ever run any commands manually (e.g. bootburn.py)? Have you ever run bootburn command with --clean option?
does this have anything to do with this MAC issue?
I never issued commands from the NVidia distributions that were not provided by you, I relied entirely on SDKmanager. The network on this machine was working until the SDKmanager was used to update it to DRIVE OS 6.0.4. The system came up with the 6.0.4 desktop, but no network.
The system date is now reporting Jan 1 1970, I presume without the network the OS install skipped NTP and has other problems. Does this system need to be re-flashed to fix the OS?
The network interface itself appears to work.
Do I need to start another ticket just to ask if I need to re-flash this machine because the OS did not install properly due to this problem?
I don’t know how the MAC addresses were wiped out previously.
After reprogramming MAC addresses, now please try if able to check and update the firmware with the method (confirm no “Ping to AURIX failed. Exit Aurix Updater”).
adi@tegra-ubuntu:~$ sudo bash /etc/systemd/scripts/nv_aurix_check_fw.sh
[sudo] password for adi:
starting Aurix FW checking...
Checking arguments...
Checking Ping to AURIX
Ping to Aurix successful
Flashed UPDATE FW version is 1.48.01.00
Aurix Update : Main FW selected: AFW
Aurix Update : Alternate FW selected: NV
Rootfs FW version is 5.06.05.0, DRIVE-V6.0.5-P3710-AFW-Aurix-StepB-5.06.05
Response_From_Aurix
Flashed FW version is 5.06.05.00
Aurix is booted with Latest FW
Rootfs UPDATE FW version is 1.48.01.0
Flashed UPDATE FW version is 1.48.01.00
Aurix is flashed with Latest UPDATE FW
adi@tegra-ubuntu:~$ sudo /bin/bash /etc/systemd/scripts/nv_aurix_check_fw.sh -auto_update
starting Aurix FW checking...
Checking arguments...
Checking Ping to AURIX
Ping to Aurix successful
Flashed UPDATE FW version is 1.48.01.00
Aurix Update : Main FW selected: AFW
Aurix Update : Alternate FW selected: NV
Rootfs FW version is 5.06.05.0, DRIVE-V6.0.5-P3710-AFW-Aurix-StepB-5.06.05
Response_From_Aurix
Flashed FW version is 5.06.05.00
Aurix is booted with Latest FW
Rootfs UPDATE FW version is 1.48.01.0
Flashed UPDATE FW version is 1.48.01.00
Aurix is flashed with Latest UPDATE FW
adi@tegra-ubuntu:~$
The network issue seems to be fixed by reprogramming MAC addresses (still don’t know what’s done previously wiped them out though).
You can use flash to it with docker again. If no network issue, everything should be good.