Boot error, need help

Hi All!

Today i restore my jetson tx-2 from another tx-2 board and get this error from dmesg

tegradc 15210000.nvdisplay: blank - powerdown
nvgpu: 17000000.gp10b nvgpu_gm20b_acr_wait_for_completion:1069 [ERR] flcn-0: ACR boot failed, err b
nvgpu: 17000000.gp10b gk20a_finalize_poweron:315 [ERR] ACR bootstrap failed

Hdmi display don’t working, but i can connect via ssh

What this error mean?

The blank powerdown is normal for inactivity. The other errors are probably a real error.

The rootfs of one release typically does not work on other releases (there are exceptions for bug fix releases). What release is the source rootfs, and what was flashed to the destination prior to the clone being put in place?

Hi, linuxdev.

Installed rootfs is L4T 32.1 with latest jetpack. Clone source rootfs is same. I was try flash with Nvidia SDK manager ang always get errors, but sometimes it was error on linux initialization start screens - after entering the login and password for the OS, tx2 hung. After that, i had to reboot it and get back into the same behavior.

On the source which was cloned can I assume you had performed the one-time login and setup of the login account? If this is the case, then the first boot on the unit restored to should not need first time boot content. Using the “-r” option (manually) in a flash.sh operation would not add that content back in, but some uses of the actual SDK Manager might.

When you clone you get both a “.img” and “.img.raw” file. What is the exact size of the “.img.raw” file, and is this the file you used (either one works, but I throw away the “.img” file and keep only the “.img.raw” since is the only one which can be edited or viewed)?

What was the exact “flash.sh” command you used to add the clone?

Are you sure the system hung, or is it just video failure? Can you still ping or ssh to the unit?

There is two cases:

  1. restore from cloned image (or image raw) - I can connect by ssh but have no video and get this errors when connect hdmi monitor
    tegradc 15210000.nvdisplay: blank - powerdown
    nvgpu: 17000000.gp10b nvgpu_gm20b_acr_wait_for_completion:1069 [ERR] flcn-0: ACR boot failed, err b
    nvgpu: 17000000.gp10b gk20a_finalize_poweron:315 [ERR] ACR bootstrap failed
  2. when restore with original nvidia jetpack system.img.raw - then i get hung after execution of the linux initialization start screen (after step “configuring hardware” or something like that) and i cant connect via ssh. I can switch to tty(1-6), login, but i do not know what to do next, GUI still hang (nothing happens there).

img - 16 467 836 708 bytes
img.raw - 30 064 771 072 bytes

sudo ./flash.sh -r -k APP -G backup.img jetson-tx2 mmcblk0p1 (https://elinux.org/Jetson/TX2_Cloning)

It is’n all system hang, hang gui only - no buttons work and there is no normal desktop

The image size is valid, so the source clone was probably correct. For the original system, did video ever work with that monitor/cable? Which L4T release is involved, and is it the same release on the source of the clone as it is on the destination of the clone?

Yes. It’s always works. But with clone it flashing with tegradc 15210000.nvdisplay: blank - powerdown/powerup. And with original image gui hung.

L4T 32.1 Jetpack 4.2 in all cases

This should work. The “blank” is not an error, it is a log of the display controller going to power savings state from inactivity or determination that there is no monitor. If the blank screen does not know there is a monitor, then it would be an issue, but it would be quite different from a GUI hang. Similar if a mouse/keyboard activity is not detected.

Via ssh or serial console, do you see any errors from:

sha1sum -c /etc/nv_tegra_release

Does unplug/replug of the monitor cable change anything?