Getting FTDI Driver to work with mmWave Radar

I have a AWR1443 which worked just fine on tx1 and tx2.
http://www.ti.com/product/AWR1443/toolssoftware

However xavier seams to not have a ftdi driver i tried installing the following from ftdi but it doesn’t work.
[url]https://www.ftdichip.com/Drivers/D2XX.htm[/url]
has anyone tried a ftdi serial chipset with anything else?
dmesg shows
[ 181.375467] usb 1-2-port4: attempt power cycle
[ 182.027001] usb 1-2.4: new full-speed USB device number 15 using tegra-xusb
[ 182.131412] usb 1-2.4: device descriptor read/64, error -32
[ 182.343389] usb 1-2.4: device descriptor read/64, error -32
[ 182.554950] usb 1-2.4: new full-speed USB device number 16 using tegra-xusb
[ 182.659359] usb 1-2.4: device descriptor read/64, error -32
[ 182.871329] usb 1-2.4: device descriptor read/64, error -32
[ 182.979328] usb 1-2-port4: unable to enumerate USB device
This usually shows up as 2 different serial ports.

The binary drivers published are generally for a desktop PC and the wrong architecture for arm64. If you have the module you can run the “file” command on that file to see what architecture it is. On the other hand, if the original module was overwritten by the x86 binary, you’ll have to find the original and put it back in place.

What do you see from this?

zcat /proc/config.gz | egrep FTDI

On my R31.0.1 it has this:

CONFIG_USB_SERIAL_FTDI_SIO=m

The USB messages you are seeing are errors, and not due to a lack of driver. I don’t know what “error -32” translates as. I’ve seen something similar on USB of another Ubuntu system when UEFI firmware was not correct. Normally FTDI serial devices do not need firmware, but USB might.

One suggestion is to test if this device, all by itself, has the same error upon plugin. Try first direct to the port, and then again to a HUB which has no other devices on it.

Note: I have plugged in an FTDI serial USB UART on my Xavier and it works.

I tried with no hub and got nothing in dmesg or lsusb. I’ll try zcat when i get to work in the morning but its not looking good.

FYI, when connecting the USB side of an FTDI serial USB cable dmesg should show something. Even devices with no driver should show a note that the device is seen. If this is the UART side (and not the USB side), then there would be no log message. A comparison log message of plugging the USB end into another Linux box would be a good debug step just to show the UART is functioning on the USB side. The corollary, that a HUB or some other USB device, when plugged into the same port on the Xavier, should show a dmesg as a verification that the Xavier port is not dead. If UART shows up on another host, and if a HUB or anything else shows up on the Xavier, then it is a mix of the two which is of issue.

well i to get CONFIG_USB_SERIAL_FTDI_SIO=m
Tried 3 other machines windows and ubuntu but no joy.
In researching this i see other people have had same problems and just when they seam to be getting somewhere the admins close threads and say solved. I have seen a disturbing trend for the past few years where companies are farming out there help desk and nothing getting resolved other than have you rebooted system. Maybe i’m to old but if you hire someone to do a job they should do more than click resolved and get paid. I would review all there answers and if there woefully lacking real solutions supervisor and the employee are fired. Rant off. At any rate i’ll keep beating my head against the wall.

well i to get CONFIG_USB_SERIAL_FTDI_SIO=m
Tried 3 other machines windows and ubuntu but no joy.
In researching this i see other people have had same problems and just when they seam to be getting somewhere the admins close threads and say solved. I have seen a disturbing trend for the past few years where companies are farming out there help desk and nothing getting resolved other than have you rebooted system. Maybe i’m to old but if you hire someone to do a job they should do more than click resolved and get paid. I would review all there answers and if there woefully lacking real solutions supervisor and the employee are fired. Rant off. At any rate i’ll keep beating my head against the wall.

On that port do you see any output from “dmesg --follow” for any other USB device as it is plugged in or unplugged? Which specific port is it?

Is it correct that the FTDI device does have some output when plugged/unplugged from another Linux system? This should narrow down the issue.

oddly it now only shows when i use the hub i bet the direct usb connection is now turned off till next reboot.
this is what i get when plugging into usb hub.

[66163.956908] usb 1-1.3: new full-speed USB device number 24 using tegra-xusb
[66163.957892] usb 1-1.3: Device not responding to setup address.
[66164.165781] usb 1-1.3: Device not responding to setup address.
[66164.372882] usb 1-1.3: device not accepting address 24, error -71
[66164.476874] usb 1-1.3: new full-speed USB device number 25 using tegra-xusb
[66164.477801] usb 1-1.3: Device not responding to setup address.
[66164.685728] usb 1-1.3: Device not responding to setup address.
[66164.892859] usb 1-1.3: device not accepting address 25, error -71
[66164.893291] usb 1-1-port3: attempt power cycle
[66165.544819] usb 1-1.3: new full-speed USB device number 26 using tegra-xusb
[66165.649139] usb 1-1.3: device descriptor read/64, error -32
[66165.861136] usb 1-1.3: device descriptor read/64, error -32
[66166.072782] usb 1-1.3: new full-speed USB device number 27 using tegra-xusb
[66166.177176] usb 1-1.3: device descriptor read/64, error -32
[66166.389166] usb 1-1.3: device descriptor read/64, error -32
[66166.497036] usb 1-1-port3: unable to enumerate USB device

It has a separate power source which i’ve tried 3 different ones but still nothing.
When running ROS example it updates the eeprom with chirp setting and it seems as though it got bricked, but your suppose to be able to recover with uniflash, however if the ftdi chipset doesnt show up as serial ports all bets are off.

Just tried connecting straight to the usb no hub with radar then a opencr. Did it a few times back and forth and the radar flat doesnt show anything but opencr connects normally. Could this be a bad ftdi chip, just a coincidence it went bad after running example code seams unlikely.

Was that HUB error showing only on the Jetson and the same HUB gives no error on the host PC (but still shows up upon connect)? If so, and given other devices do not show at all, I am thinking it is a hardware problem. Which specific port are you using?

You are correct that you should be able to re-flash regardless of a prior flash success/fail status.

No hub error showed up on pc with ubuntu and jetson with different hub. So hub isnt bad and there industrial not the $20 walmart specials. I even swapped them out for one that was in a embedded system. At any rate i have to cut my loses and buy another, but i dont want this to happen again.

To clarify, are you saying there is a HUB which does work on the Xavier? Is it just particular devices/HUBs which fail? Apparently everything you tried on the Xavier does work on some other system, and at least some HUB works on Xavier…is this correct?