When I use our customized v1.1 carrier board to flash Orin image to NVME, it can flash successfully.
But it fails by using v1.2 carrier board. from the log, the ssh connection can’t be setup successfully.
HW engineer has compared schematic and HW components of these two version for the USB download portion, he can’t find the different.
Do you have any idea about this ?
Environment :
1.Orin32G module
2.L4T : 36.4.3
3.Desktop : intel i7 x86_64/Ubuntu 20.04
Attached are the UART and flash logs for these two versions.
It seems you are using custom carrier board and they are designed by you.
From the log you shared, ssh does not be established through USB interface(3550000.usb) to be recognized by host.
Do you have any HW modification for USB between v1.1 and v1.2 board?
1.Yes, both are our custom carrier boards, and both can flash to internal eMMC.
But only v1.1 can flash to NVME from slot of PCIe C4 controller.
2.HW engineer said the only different between these two versions is the USB layout distance of v1.2 is longer then v1.1, but it is under the USB 2.0 design spec.
We also have another custom carrier board for Orin, the USB layout distance is shorter than v1.1, but it also can’t flash to NVME. This carrier board is designed after v1.2 version, and is reference from v1.2.
3.The flash command to eMMC is
./flash.sh jetson-agx-orin-devkit mmcblk0p1
and to MVNE is
./tools/kernel_flash/l4t_initrd_flash.sh --external-device nvme0n1p1 -p “-c ./bootloader/generic/cfg/flash_t234_qspi.xml” -c ./tools/kernel_flash/flash_l4t_t234_nvme.xml --showlogs --network usb0 jetson-agx-orin-devkit external
Yes, I can flash to internal eMMC succefully on both v1.1 and v1.2 carrier boards.
The size of NVME is 250GB(TM8FPL250G0C127), and I don’t modify the flash_l4t_t234_nvme.xml , the expected size of rootfs should be 59.6GB (1/1562MB /61.2GB /59.6GB/APP/msftdata).
Yes, the result is the same, only v1.1 can be flashed to NVME.
Could you request your HW engineer to measure and compare the USB signal on both v1.1 and v1.2 to find any differences?
From the current log, it looks more relating to USB deice mode.