Electrically there is no harm in this. In fact, jumper of TX to RX (and possibly also CTS to RTS) is considered “loopback”, and is useful for testing purposes. The issue you would run into is if there is an application listening to the UART. For example, serial console.
Imagine if this is serial console, and everything output during a boot gets fed to the console’s input just as if you were sitting there typing nonsense in on the keyboard. The console could get rather confused.
Every once in awhile some nonsense might actually be a command. There might be some side effect of that command, and the side effect might be harmless, or it might be fatal (very unlikely). Just as an example, what happens if the nonsense is “
rm -Rf”? That’d be a problem. What if the command was “
ls”? Technically that would not be a problem, but perhaps the name of a file is also a command, e.g., if the “
ls” were in “
Was the UART being used as serial console? Was the UART associated with anything else?
If you look at the output of this:
ls -l /dev/ttyS* /dev/ttyTHS*
…then you will see permissions of all serial UARTs.
If the group of that UART is “
tty”, then this is a console. If the group of the UART is “
dialout”, then this is a general purpose UART and not being used as a console.
NOTE: In the naming convention
ttyTHS#, if the “
#” is the same, then it is the same UART. There are often legacy driver (the
ttyS# notation) and also “Tegra High Speed” (
ttyTHS#) drivers available for the UART…you just wouldn’t want to use both drivers at the same time.