Default IP not available error while trying to install Jetson SDK components

Yes. After the boot sequence (nvidia logo then the boot log), it just loads to a black screen.

The monitor natively supports hdmi and is connected to the jetson using a hdmi cable only.

Hi,

Are you sure you have hdmi connected with the log you shared to us?

From the log, I don’t see any activity from HDMI. Could you try to hotplug your HDMI cable and check if any new log coming out from your dmesg?

there was no hdmi inserted during the boot. I had the output with “edid invalid” after puggingit it in.

Could you plug it, boot up and share me the dmesg? BTW, just attach it as a text file. Do not copy paste it…

Here is the boot log with the monitor plugged in. still
a black screen after the boot sequence

boot log with monitor (127.9 KB)

Hi,

The mode from the HDMI seems a little weird. What is the default mode of this monitor?

[ 5.308840] tegradc 15200000.nvdisplay: hdmi: tmds rate:154000K prod-settingm

BTW, I am not quite sure why there is a length limitation in your log. Are you able to prevent that truncation happened?

I don’t know if its the default mode (or how to determine it). It’s a dell 2408wfp if that helps.

as for the truncating. My mistake, minicom did not have text wrapping on, here is the fixed file.

boot log fixed (121.5 KB)

Do you have other kind of monitor that you can test?

This is with another dell monitor (same model though)

boot log monitor2 (150.6 KB)

I’ll have to try other monitors tomorrow. Hope this helps for now

My purpose here is just switch to new kind of monitor instead of “same model”. The actual problem here is the mode we see from your monitor seems not not common one. 154000k may indicate it is 1920x1200 or something else.

Please try with a monitor with 1080p mode first.

I figured, but I don’t have access to any other model of monitor atm :D

Will try to find a 1080p monitor.

Your first monitor has an error. For reference, the DDC wire uses i2c protocol, and is what provides the EDID. In most cases, the i2c is powered by the host, not the monitor, and allows plug-n-play query of what monitor is attached even if the monitor is not powered.

Note this log line occurs only for the first monitor, and initially has an error, then seems to at least partially correct (this is basically a failure case with some sort of marginal signal as my guess):

[0002.305] E> I2C: Timeout while polling for transfer complete. Last value 0x0.
[0002.306] E> I2C: Could not write 0 bytes to slave: 0x00a0 with repeat start .
[0002.308] E> I2C_DEV: Failed to send register address 0x00000000.             
[0002.313] E> I2C_DEV: Could not read 128 registers of size 1 from slave 0xa0 .
[0002.323] E> could not read edid   
...
[    5.617402] tegradc 15200000.nvdisplay: hdmi: get YCC quant from EDID.

Now here is the second monitor of the same model, but it always works:

[0002.314] I> edid read success                                                 
[0002.326] I> edid read success   
...
[    4.665577] tegradc 15200000.nvdisplay: hdmi: get YCC quant from EDID.

Notice that any cable detect event will cause this to be read. Cable detect can include initial i2c power up during boot. Does your second monitor of the same type work? The logs don’t really clarify, but I suspect the second monitor does work, or at least gets further along earlier in boot. If it does work, then it is like a cable issue on the other monitor (keep in mind that the same cable can work on one monitor, but fail on another since signal quality is a chain of components). Check different cables as well.

Both monitors and their respective hdmi cables are in use by other devices.

I will try testing it with other monitors/cables when I can get them.

@linuxdev @WayneWWW

I have managed to solve the issue. The owner of the jetson gave up and decided to just proceed with Ubunto 20.04 and Noetic.

Reinstalled Jetson 5.0.2. Loaded it up to test a last test and noticed that the device had an ssd instead of just an emmc like the rest of the jetson in our lab. In addition, the ssd had an installation of ubuntu as well.

I have been attempting the flash to install to the default emmc. I would guess that having the same OS (ubuntu 18) installed into 2 different drives somehow caused the issues (and might be why I was able to install ubuntu 20 without issue)

wiped everything on both drives.

Popped into sdkmanager, select 4.6.2, installed to the ssd instead of the emmc.

Installed without any issues and is now operating just fine.

Thanks again for the help.

I don’t think I helped anything here.

The reason that the GUI is missing is still unknown to me. EMMC is the basic case that should always have things to work. It is not reasonable that something would work fine just because it is installed on SSD but not EMMC.

One possibility may be the disk space is running out of when you flashed with emmc. However, this is AGX Xavier which at least has 32GB disk space. I don’t think that is the reason.

I still appreciate the time you put in.

You do have a point, it doesn’t make sense as to why it’s behaving this way. And even if it was a size issue, why was it able to install 5.0.2 without any problems?

Unfortunately, the jetson is no longer in my hands for further testing.

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