Jetson nano Black Screen

Hi
I facing with Black screen issue on Jackson nano. Issue happen when i plug power supply and booting up.
Black screen issue don’t happen when reboot, it is sometime issue.
Do anyone can help me for debug black screen issue?
I would like attach log file.
varlog.tar.gz (2.5 MB)
Thanks!

First, I think you should dump the log just after the error happened and only share that kernel log. It is kind of wasting the time of other people to check log like yours. You shared a kernel log that is lasting for 4 days. Most of them are not needed. We only need the time when you hit error.

If you don’t know how to dump log when monitor is gone, use the serial console from uart. It will also share more detail.

https://elinux.org/Jetson/General_debug

Second, one suspicious point here is your monitor seems has problem. The kernel log says it is using fallback edid.

Oct 11 02:00:14 vsm kernel: [ 1.259800] tegradc tegradc.0: hdmi: edid read failed
Oct 11 02:00:14 vsm kernel: [ 1.264914] tegradc tegradc.0: hdmi: using fallback edid
Oct 11 02:00:14 vsm kernel: [ 1.265001] tegradc tegradc.0: blank - powerdown

Generally, our driver will read the edid from your monitor and use this info to select the proper mode for your monitor. However, in your case, it is using a hardcoded edid we prepared in driver. This mode may not work on your monitor. That is why you would see some unstable case here.

Please share more detail about your monitor. Is it really a pure HDMI monitor ? Or you are using some adapter here?

Also, please clarify that this issue happens on DP or HDMI. If this only happens to one kind of monitor, please remove the other one. Otherwise all the display logs would mix up and make it harder to debug. I saw you have some DP activity in kernel log.

Thanks.

Thank you for your response.
Please share more detail about your monitor. Is it really a pure HDMI monitor ? Or you are using some adapter here?
my monitor is DP. However, HDMI is always active due to HW issue. Issue happen on DP and don’t happen on HDMI. When i attach HDMI, it display normal.
Could you please guide me for disable hdmi for checking?
Thanks so much

What do you mean you need my guidance to disable HDMI?

I am just talking about just removing the hdmi cable.

If you are using custom carrier board, you should reveal such info from the beginning.

What do you mean you need my guidance to disable HDMI?
yes, I need guide for disable hdmi by command or something because hdmi cable don’t attach but it is alway active(connected status) due HW issue.
My board is custom carrier board, I’m sorry because i don’t say it from the beginning.

You need to modify the device tree, disable the corresponding tegradc.

if modify the device tree, then difficult for confirm that issue relate to hdmi that always connected status because issue is sometime.
Do you see abnormal from DP ? (issue happen on Oct 14, log is given by my QA team, not me. So, sorry about this inconvenience.

Hdmi always connected issue is obviously a hardware problem… which I have nothing to comment here.

Fix that first and check the DP issue again. Dump only the error log to us.

I see log below:
Oct 14 16:28:58 vsm kernel: [ 1.119449] tegradc tegradc.1: disp1 connected to head1->/host1x/sor
Oct 14 16:28:58 vsm kernel: [ 1.119538] tegradc tegradc.1: No lt-data, using default setting
Oct 14 16:28:58 vsm kernel: [ 1.119570] tegradc tegradc.1: No hpd-gpio in DT
Oct 14 16:28:58 vsm kernel: [ 1.119606] tegradc tegradc.1: DT parsed successfully
Oct 14 16:28:58 vsm kernel: [ 1.119658] tegradc tegradc.1: Display dc.ffffff800d500000 registered with id=1
Oct 14 16:28:58 vsm kernel: [ 1.121319] tegradc tegradc.1: dpd enable lookup fail:-19
Oct 14 16:28:58 vsm kernel: [ 1.122660] edid invalid
This log is abnormal, right?

Do as what I said first and come back with full log.

yes, i will sure do it.
nvidia have display debug log that need to turn on for checking more deeply or only kernel log is enough?

You can dump

  1. dmesg
  2. /var/log/Xorg.0.log.

Only need that log when error happens.
If we need something else to check, I will tell you.

Also, try to use same monitor on the devkit and see if devkit is able to hit same issue too. This could show if this is hardware issue.

Hi WayneWWW,

Issue can not reproduce until now when i disable hdmi.
I confirmed again with my QA team and define issue time as well as log error:
Oct 14 16:28:58 vsm kernel: [ 3.836979] dp lt: CE not done
Oct 14 16:28:58 vsm kernel: [ 3.836981] dp lt: CE retry limit 5 reached
Oct 14 16:28:58 vsm kernel: [ 3.838391] tegradc tegradc.1: DP: no prod_c_rbr prod settings node in device tree
Oct 14 16:28:58 vsm kernel: [ 3.838415] dp lt: switching from state 3 (channel equalization) to state 4 (link training fail/disable)
Oct 14 16:28:58 vsm kernel: [ 3.838646] dp lt: state 4 (link training fail/disable), pending_lt_evt 0
Oct 14 16:28:58 vsm kernel: [ 3.838647] dp lt: NULL state handler in state 4
Oct 14 16:28:58 vsm kernel: [ 4.854898] extcon-disp-state extcon:disp-state: cable 44 state 1
blackscreen_log.txt (131.7 KB)

Could you please explain this log?

Thank you

I cannot understand what is your debug logic here. Please just reply my question here.

  1. Is this issue still reproducible on your board now or not?

  2. If it can, then what did you mean “Issue can not reproduce until now when i disable hdmi.”.

  3. There is a process called link training is Displayport spec. It fails 5 times in your board so it gave out that error. Please also try the same monitor on NV developer kit and see if it can reproduce same issue. But I am not sure why I still see HDMI log in your dmesg now. What are you doing exactly? I already told to disable HDMI to check the rest issue. Is there any problem to do this?

  1. Is this issue still reproducible on your board now or not?
    As i said, issue is sometime, not always and it is difficult for reproduce. We is still continue testing. We test on more device, not one.
  2. If it can, then what did you mean “Issue can not reproduce until now when i disable hdmi.”.
    I try to disable hdmi for testing on device that issue happen and issue can not reproduce until now but we can not conclude hdmi issue because issue is sometime and it is difficult for reproduce.
  3. Log that i provide to you is old log(do not disable hdmi) but it is log when issue happen. I don’t think hdmi enable then issue can not check from log. If issue happen on DP, then log will show even hdmi is enable, right? If both DP and HDMI is enable, then display should show on both, not only hdmi.
    From log i see error log:
    Oct 14 16:28:58 vsm kernel: [ 3.836979] dp lt: CE not done
    It can be retry more time but finally it should be “dp lt: CE done” so, i would like ask you for your recommend.

Hi,

Maybe you can try to use different function provided on the forum to clean up your comment. Honestly, it is not easy to read your comment.

you can use such block to quote comment from other people. And reply your own answer in non-block area.

Actually, the HDMI should not be related to your issue. But it is silly to see that a board which does not have HDMI port still tries to output HDMI signal. Also, this makes the kernel log hard to read. So I asked you to disable it.

My first recommendation is:

If this issue only happens sometimes, then maybe you can try to do the tuning.

And the second one is please try your monitor on Nano developer kit too.
Also, please make sure you use a common CEA mode like 1080p or 720p to test. Make sure this base case can pass your test first.