EDIT this was not really a bug, I was looking at the wrong bits. But other things were clarified.
Or is it a datasheet bug?
SYSCFG0 bit 3 RSTPINCFG reads as 1 on the ATmega4809 Xplained Pro (0xF5) which means that the pin (PF6) is used as RESET.
However Studio says that it is set as I/O (0??) pin and in fact the reset button of the Xplained Pro doesn't work.
Any ideas?
Check if you have the latest ATmega ddev support (tools/device pack manager). I think I remember that there was an initial error in the xml...
When I updated Studio a few days ago it also updated about 70 packs, anyone in particular I should look check?
http://packs.download.atmel.com
.
ATmega 1.2.209 (2018-02-19) - Corrected reset pin configurations and changed PORTMUX signal names for ATmega4809, ATmega4808, ATmega3209 and ATmega3208. Corrected TWI0 pins on mega324PB.
Look like I have already that pack, most of the stuff is 14 June but a couple of files have the 19 Feb stamp. I will try and delete that pack and reinstall again.
Uninstalled that pack (it was correct anyway), restarted Studio, reinstalled the pack, restarted Studio and read the fuses.
Still the same as the screen shot above, bit 3 RSTPINCFG reads as 1 but the screen shows it as GPIO and not reset. Edit I'm getting bits mixed up, sorry.
so bit 3 is not programmed and the reset pin is set as I/O as Studio correctly shows.
However as I said above the reset pin is NOT working on the board so maybe the screen shot is correct?? Do you have an ATmega4809 Xplained Pro to check this with, if not you can ask Martin to send you one..... 
I was going to change the fuse but then I had all the dire warnings of impending disasters if I do that unless I have a programmer with 12V, so I left it alone.
EDBG has a firmware feature where it will ignore to write bits that would make the kit in a 'bad state'. Might be that you are seeing?
If the data sheet is correct at the part except for the question point, even if you rewrite the fuse, 12 V programmer will not be necessary.
It seems slightly different from the tiny series in that respect.
I can confirm that the RESET pin works when programmed, and that on the chip I am currently using the SYSCFG0 has a value of 0xE5 when RESET is unprogrammed (GPIO mode) and 0xED when it is programmed. I think this is also consistent with the data sheet.
Edited #6 above, my eyes were going funny and was looking at the wrong bit!
even if you rewrite the fuse, 12 V programmer will not be necessary.
This is what I get when trying to program the reset pin as a reset pin.
Is this not correct? I really want F6 to be a reset pin but I don't want to kill UPDI by doing so. UPDI works with F6 as an I/O
I would like you to take that part as a problem as well.
I can not find an explanation that requires 12V on the datasheet.
I think you can safely ignore the warning. I changed the setting with a SDK600 and reset worked, then changed it back and it could be programmed with UPDI again. (Although the SDK600 can successfully change the fuses, it doesn't not work as a programmer, at least when the device is off board. Odd behaviour.)
I can confirm that F6 can be switched back and forth between GPIO and RESET mode with the ATmega4809 Xplained Pro and the sky does not fall in.
UPDI continues to work in either modes, so I now have a functioning reset, thank you.
At least, even if the bug was non existing, something else was sorted out.
