I have the same error as that poster, but a new topic was wished for.
My Plan is to get a AGX Xavier into the endpoint mode and get the virtual network between to Xavier.
I have done this multiple times with the old jetpack, so the steps are clear to me, but I do not get it up and running with the new jetpack. First as the original owner of this post the endpoint Xavier stuck and restarts.
And then I tried the patch you posted. This is where I am now.
The endpoint xavier is not restarting anymore, but the network is not working as well.
Not able to ping from both sides. The endpoint is just returning “Destination Host Unreachable”.
On the root Xavier the USB driver is stuck now and I cannot use the mouse or keyboard anymore… Also, the ethernet connection is dieing as well.
PS.: @WayneWWW like so? Now you can answer in two topics simultaneously…
This is not to make my work double. This is to prevent the noise from original owner.
Your usb issue seems not related to the PCIe issue but seems affecting you to test pcie issue, could you fix that first? Is that hardware problem or something else?
Have you validated the hardware correctness in jp4.x? If so, could you set up root port in jp4.x and endpoint in jp5.0.2 and see if they can ping each other under such situation?
Sorry to annoy you, but that’s the definition of a forum…
You look for similar things first, try to see if it is already solved, and do not start the same topic over again…
but you are the boss here.
Hmm I thing it is related as it appears only when I try to get eth1 up.
It´s been some time since I looked at the schematics, but aren´t there some tings shared between USB and pcie on the dev kit?
Yes both are flashed with jp5.0.2 and on both a update/upgrade was executed.
Have you validated the hardware correctness in jp4.x? If so, could you set up root port in jp4.x and endpoint in jp5.0.2 and see if they can ping each other under such situation?
So, i tried all permutations of jp4.6 and jp5.0.2
The connection works if the root system is jp4.6 with both ep systems.
If the root is jp5.0.2 it doesn’t work.
If ep is jp5.0.2 no ping or anything is possible and if the ep is powered off I got a kernel panic.
If ep is jp4.6 I could not log in as the USB was not working anymore.
(I scripted the network build up routine in a service)
Yes, the RP- 4.6, EP -5.0.2 case works fine with your patch applied only to the ep.
I do not have the unpatched ep system anymore and would need to rebuild the kernel if that is needed as well.
Set static IP
use netplan.io to set static ip address
→ sudo apt install netplan.io
→ Static IP Address Assignment
To configure your system to use static address assignment, create a netplan configuration in the file /etc/netplan/99_config.yaml. The example below assumes you are configuring your first Ethernet interface identified as eth0. Change the addresses, gateway4, and nameservers values to meet the requirements of your network.
I just reflashed to rp, because this seems to be the trouble maker.
I disabled the Network-manager service and tried it again without anything else connected.
Should it work without the static IP? As I cannot use a static IP in my companies network i left it away.
Still no pinging possible :(
Why does this issue matter with your company network…? Under this setup, only two jetson connected to each other and none of them is connecting with ethernet… So company network totally not involves here…
What I am talking about is set static IP to the two PCIe eth1 interface with netplan tool.
Hi,
just to clarify, I know the company network doesn’t matter, but I thought that’s where you were going with your netplan stuff. Both systems have monitors and everything so accessing them is not a problem.
My current steps are:
flash rp normally
flash ep in the right mode
build kernel on ep with your patch.
stop and disable network-manager.service on rp.
put the two together
boot ep and execute:
cd /sys/kernel/config/pci_ep/
mkdir functions/pci_epf_tvnet/func1
echo 16 > functions/pci_epf_tvnet/func1/msi_interrupts
ln -s functions/pci_epf_tvnet/func1 controllers/141a0000.pcie_ep/
echo 1 > controllers/141a0000.pcie_ep/start
boot rp
ep → ifconfig eth1 up
rp → ifconfig eth1 up
ep → ifconfig eth1 192.168.2.2
rp → ifconfig eth1 192.168.2.1
And after that, I still cannot ping. Now was my question should I replace the part at 9 and 11 with your netplan stuff or should it work like this? I will now look into netplan.io to set something from point 8 onwards…