What causes TX2 ubuntu into Minimized mode?

[chjchoi]

At booting time My carrier board on debugging shows the intro image of nvidia (1920x1080p) but, after kernel stage the ubuntu is falling into minimized mode, so I can’t see any HDMI output.

I think there may be many candidates causing my carrier board into the mode, which one will be the most common case?
Why and On which conditions, the ubuntu falls into the minimized mode?

Thank you.

[chjchoi]

When TX2 is moved on the Reference board , there is no such problem. it shows gnome graphic shell well.

[linuxdev]

By “minimized”, do you mean text-only shell? Or are you referring to the text showing up which states “…This system has been minimized…”? If the latter, then this probably isn’t what you think it is. That latter statement is just a way of saying man pages and documentation has been left out to save on disk storage space. The “unminimize” comment just says that if you are ok to use the disk space, then man pages and documentation packages can be told to go ahead and install. Those are completely unrelated to whether or not a GUI shows up, and are also 100% not a bug or limitation on system function.

Incidentally, if you are interested in seeing what disk space you have, then a useful command is “df -H -T”. I added the “-H” for “human readable” units (besides just plain bytes), and the “-T” to show filesystem type. Many of the filesystems are “pseudo” filesystems in RAM and don’t count (the ones you are interested in are either ext4 or vfat…those are actual hard drive space).

[jchaves]

Hi chjchoi,

When I have worked in the past with custom carrier boards for the TX2 and the HDMI output is not working then it was due to lack of support in the device tree (i.e you would need a patch for the device tree). So I would say this is most likely your case as well. Is the carrier board your own design or did you purchased it from some hardware manufacturer? This has been a common issue in the past but to provide more guidance you would need to share more information of your design or ask for support from the carrier board manufacturer.

Also, I am sure that if you browse through the forums your will find some threads with useful information of similar cases in the past, let me share some links of threads that I consider with a similar issue as this one:

Enabling and display with HDMI on Custom Carrier Boards

Hi everyone, I want to enable HDMI on Custom Carrier Boards , I foun it‘s not working follow this steps: ====================================== Enabling HDMI on Custom Carrier Boards This information is specific for modifying the device tree file for your custom carrier board and does not apply when you use the Developer Kit carrier board. Download the example HDMI DTS file Download the 鈥渢egra234-dcb-p3701-0000-a02-p3737-0000-a01_hdmi.dtsi鈥� file from https://developer.nvidia.com/embedde…

HDMI Port for Custom Jetson Carrier Board

Hi for my drone I want to develop a custom Jetson carrier board with minimal required interfaces. For quick debugging I would like to add an HDMI port on the PCB. However, I’m not familiar with HDMI and I found some contradictions between guides. In most guides it’s written the impedance of the trace should be 100 Ohm. I found some pretty easy schematics examples [grafik] However in the Jetson Nano design guide there more required elements as these capacitors and Rs resistors, which I do not…

How to Enable HDMI on TX2 with custom carrier board

Hi WayneWWW, Thanks for your reply. Sorry, I didn’t describe “device” specifically. I am working on a project which is developed on a TX2 custom carrier board. The TX2 custom carrier board is the “device” I mentioned above. In the case above, I want the device could enable or disable the power of HDMI connector when it detects the HDMI cable is plugged-in or unplugged. Currently, the power of HDMI connector is always enabled. I know that HDMI_HPD is used to indicate the connecting status …

Regards,

Jafet Chaves,
Embedded SW Engineer at RidgeRun
Contact us: support@ridgerun.com
Developers wiki: https://developer.ridgerun.com/
Website: www.ridgerun.com

[chjchoi]

Thanks for your generous reply. I’m on developing my own carrier board and I’m now not familiar to Device Tree System. and I must update the dtsi files for my board.
I have only experienced the 2.6 kernel at about ten years ago when the device tree concept did not exist.
I will follow the information you gives ,
Thank you.

[DaneLLL]

Hi,
It seems like EDID of HDMI device is not correctly read. You may also check this part.

For customizing kernel and device tree, please refer to Kernel Customization in the developer guide

[linuxdev]

You might find this of interest:
https://forums.developer.nvidia.com/t/recovery-port-serial-on-custom-linux-install/220390/8

To expand on what @DaneLLL just mentioned, HDMI is “plug-n-play” because the GPU can query the monitor for its specs. That query uses an i2c protocol over the “DDC” wire of HDMI (the format of the data is EDID). The GPU itself powers that i2c query as a method of making it possible to query a monitor even when its power is off. An example of device tree causing a fail is if the power supplied to the GPU side of the i2c query was not enabled in device tree.

[chjchoi]

Thank. I have removed gpio expander chip which makes power enable signal…
But the power supply description of the devicetree file may not be exact.
I will check the power description of the dts/dtsi
Thank’s
vdd-hdmi-suppy was changed to &battery_reg

[chjchoi]

I have found the reason…

There was some mistake of making customized HDMI cable harness.
I think that if the System is not able to get EDID from the HDMI monitor than fall into CLI and minimized.

Thank you.