I ran
sudo ./flash.sh -k DTB jetson-nano-qspi-sd mmcblk0p1
Now it sounds like I need to flash -k RP1 as well. Let me do that and get back to you
Just FYI… I ran
sudo ./flash.sh -k RP1 jetson-nano-qspi-sd mmcblk0p1
sudo ./flash.sh -k DTB jetson-nano-qspi-sd mmcblk0p1
then rebooted. On reboot, mtd1, mtd2 and /sys/firmware/fdt all showed the new dtb.
Hello, Sorry for the delay getting back to this, had some office space issues I had to get sorted before I could try more.
I beleive that something is working now because stuff is breaking. Now when I flash the DTB and RP1 sections the Jetson will not boot. It is stuck on the NVIDIA screen. If I copy back the original .dtb into kernel/dtb and reflash the jetson will boot.
I also tried with the default Excel sheet before making any changes and I see the same. Is there a way to debug this before I get access to bootloader messages (Don’t have a way to buy the hardware right now)?
There are 9 copies of the device tree that I have found. 7 are on the SD card and 2 are on a 4MB flash on the SOM. It was a pain to finally realize the ones that matter are the ones of the 4mb flash.
However, even though I am now confident that the correct DTB is loaded and the DTB is configured with the DTSI files as per the documentation, we still can’t actually use the pinmux we configured. I agree that the process is broken.
As for not booting, I only got stuck at the nvidia screen when I tried to use a DTB for the 4GB nano and carrier board to try and boot a 2GB nano and carrier board. I believe the SOC is trying to look for hardware that isn’t available per the device tree. Make sure you are compiling the correct device tree for your test setup.
The excel sheet says “B01 SODIMM” on it. Which is the board I ordered. I was able to boot it successfully with the original “40 Pin Header” excel sheet. Does this mean we can’t use the “B01 SODIMM” pinmux for our devboards?
Disregard my comment about compiling the correct device tree. It looks like you are using the correct file - I went back and reread your previous posts.
I do believe you are using the correct files.
So you gave me something to look at. It seems that using the “B01 SODIMM” pinmux might have been part of the issue. If I copy over the exact same configuration (only the available PINS) to the “40 Pin Header” excel sheet and export that one I can boot. I see the results (to a point) that I expect.
So I am not sure why this is the case, but it does allow me to boot.
Thank you gtj as well because your guidance helped me figure out how to flash the device tree in the first place.
I am seeing another issue now but it is related to the way I expect the Pins to work. But I think the original post is answered for now.
I wish I could mark 2 responses as the solution because it really did take 2 steps to figure out why my pinmux wasn’t being used.
Thank you both for your help, I do have a follow up issue that I posted here: Jetson Nano DevBoard not pulling output pin to 0
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.