Jetpack 6 Update

[xychelsea]

The configuration files and DTB trees need to be updated to work with Jetpack 6. Specifically, I am using the ReServer Industrial Orin NX 16GB.

[Pillar1989]

We are working on this and expect an update in the coming weeks. @xychelsea
cc @bigbearishappy

[bhaskara-res]

@bigbearishappy, @Pillar1989 could you please share any updates on timelines ?

[bigbearishappy]

@bhaskara-res Our base work on orin nano and orin nx is finished. We are now working on different carrier board, which include recompter J401, recomputer industrial J201 and reserver industrial J401 ...

[bigbearishappy]

@xychelsea @bhaskara-res
We have uploaded the code of seeed's carrier board to the branch r36.3.0. You can now try it by following the README.

[thomaskleiven]

First, thanks for your support on this topic.

We've followed all the steps on the r36.3.0 branch and successfully flashed our reComputer Industrial J4012. However, it appears that some configuration files and DTB trees are not correctly mapped. Specifically, we need /dev/ttyTHS0 to be available, as outlined in this NVIDIA forum issue.

Here's the output from sudo dmesg | grep tty. We would expect to see logs related to ttyTHS0, similar to the behavior observed in Jetpack 5.x.

sudo dmesg | grep tty
[    0.000000] Kernel command line: root=PARTUUID=3675fdfb-22bf-4708-877b-27007408e511 rw rootwait rootfstype=ext4 mminit_loglevel=4 console=ttyTCU0,115200 firmware_class.path=/etc/firmware fbcon=map:0 net.ifnames=0 nospectre_bhb video=efifb:off console=tty0 bl_prof_dataptr=2031616@0x471E10000 bl_prof_ro_ptr=65536@0x471E00000 
[    0.000329] printk: console [tty0] enabled
[    0.064691] 31d0000.serial: ttyAMA0 at MMIO 0x31d0000 (irq = 117, base_baud = 0) is a SBSA
[    0.154771] printk: console [ttyTCU0] enabled
[    1.659897] printk: console [tty0]: printing thread started
[    1.660634] printk: console [ttyTCU0]: printing thread started
[    1.736486] 3100000.serial: ttyTHS1 at MMIO 0x3100000 (irq = 112, base_baud = 0) is a TEGRA_UART
[    1.737445] 3140000.serial: ttyTHS2 at MMIO 0x3140000 (irq = 113, base_baud = 0) is a TEGRA_UART

Is this expected or do the DTB trees need to be updated?

[bigbearishappy]

First, thanks for your support on this topic.

We've followed all the steps on the r36.3.0 branch and successfully flashed our reComputer Industrial J4012. However, it appears that some configuration files and DTB trees are not correctly mapped. Specifically, we need /dev/ttyTHS0 to be available, as outlined in this NVIDIA forum issue.

Here's the output from sudo dmesg | grep tty. We would expect to see logs related to ttyTHS0, similar to the behavior observed in Jetpack 5.x.

sudo dmesg | grep tty
[    0.000000] Kernel command line: root=PARTUUID=3675fdfb-22bf-4708-877b-27007408e511 rw rootwait rootfstype=ext4 mminit_loglevel=4 console=ttyTCU0,115200 firmware_class.path=/etc/firmware fbcon=map:0 net.ifnames=0 nospectre_bhb video=efifb:off console=tty0 bl_prof_dataptr=2031616@0x471E10000 bl_prof_ro_ptr=65536@0x471E00000 
[    0.000329] printk: console [tty0] enabled
[    0.064691] 31d0000.serial: ttyAMA0 at MMIO 0x31d0000 (irq = 117, base_baud = 0) is a SBSA
[    0.154771] printk: console [ttyTCU0] enabled
[    1.659897] printk: console [tty0]: printing thread started
[    1.660634] printk: console [ttyTCU0]: printing thread started
[    1.736486] 3100000.serial: ttyTHS1 at MMIO 0x3100000 (irq = 112, base_baud = 0) is a TEGRA_UART
[    1.737445] 3140000.serial: ttyTHS2 at MMIO 0x3140000 (irq = 113, base_baud = 0) is a TEGRA_UART

Is this expected or do the DTB trees need to be updated?

Actually, the ttyTHS0 in JP5 has been changed to ttyTHS1 in JP6, you can have a try with it~

[thomaskleiven]

We are using pin 1 (Data-) and pin 2 (Data+) of the COM 232/422/485 - DB9 J1 connector to communicate with various hardware over the RS485 protocol, as referenced in the Seeed Studio documentation.

Under Jetpack 5, communication over the /dev/ttyTHS0 port worked very well. After upgrading to Jetpack 6, the /dev/ttyTHS0 port appears to be disabled (as you mentioned). We have attempted to use the /dev/ttyTHS1 and /dev/ttyTHS2 ports, but we are not receiving any signals through these ports.

Is there a way to re-enable or configure the /dev/ttyTHS0 port in Jetpack 6, or do you know any alternative solutions for RS485 communication on Jetpack 6?

[thomaskleiven]

When using the new image provided in #5, it works as expected. I'm not entirely sure what differs in this image compared to when we followed the steps on the r36.3.0 branch, I guess we must have done something wrong there.

As you mentioned, the memory address 0x3100000 now maps to /dev/ttyTHS1, whereas it mapped to /dev/ttyTHS0 on Jetpack 5. So, we'll be using /dev/ttyTHS1 going forward.

For others experiencing similar issues, I recommend using the images provided in the Seeed documentation.

Thanks for the support!

[bigbearishappy]

@thomaskleiven Thank you for your feedback.
The code on branch r36.3.0 is actually the same with the source code which generate the image you download from our wiki.
We would recommend you to use the code on r36.3.0 if you need to add custom code for your device.

[bigbearishappy]

Sine the code for JP6 is uploaded. Closing this issue. If you have any further problem. Please create a new issue~