implementing usbip on Jetson

is there something like linux-tools-4.9.140-tegra available that would add usbip support somehow?
It seems that support of it can be built into kernel

gunzip -c /proc/config.gz | grep USBIP
# CONFIG_USBIP_CORE is not set

And other two of the three?

modprobe usbip-core

modprobe usbip-host

modprobe vhci-hcd

At device I can see:

/usr/src/linux-headers-4.9.140-tegra-linux_x86_64/kernel-4.9/drivers/usb$ ls
atm     chipidea  common  dwc2  early   host   isp1760  Makefile  mon   phy            serial   typec  wusbcore
c67x00  class     core    dwc3  gadget  image  Kconfig  misc      musb  renesas_usbhs  storage  usbip



I am just wondering if the only way to get the module built is to re-compile the kernel


I haven’t worked with CONFIG_USBIP_CORE, but no reason it wouldn’t be like any other driver. If those can be built as a module, then this would be a simple case of building modules and copying to the right place…if not, then you’d have to update the kernel Image and make sure “uname -r” remains constant via the CONFIG_LOCALVERSION. Did you run into a specific error while building or installing the changes?

no, no specific issues or errors yet

I just temporarily do not have x86_64 and speculating about attempting flashing Jetson from a remote x86_64 host [ from Google Cloud [GCP] in a way the Xavier will be connected in recovery mode to Jetson nano device and the flashing will use usbip bridge between nano-to GCP.
Yet I am at state of theoretical investigation of the issue. Thank you for your inputs.

Do you like to connect the Xavier to Nano to update the system image?
Current the flash script not support on ARM system yet.

Hi, and thank you for your response.
The answer is yes and no.
Yes I want to connect Xavier to nano when Xavier is in recovery mode, then I will try to execute Jetpack flash script at remote x86_64 computer to that Xavier device will be mapped from nano using usbip bridge.
It is merely a hypothesis though that I am trying to check if it works or not.
As you can see the flash script will be executed remotely with that approach and both the remote computer and nano having the usbip module loaded into the kernel might with low though probability allow to flash Xavier connected to nano from remote x86_64 workstsation in case the Xavier will be mapped and presented at lsusb at the remote workstation via means of usbip server running at the nano device and usbip client running at the 64_86 workstation.
But probably it will be over-complicated and easier is to flash directly from x86_64.

Cool idea, hope to hear how it works if you get a chance to try it. If the executable on the remote system detects the Nano port as its own (including holding ownership…similar to use of a VM if the VM does not hold ownership), then it should work.