i2c bootloader on stm32h523 64pin

[PR.10]

1.
AN5959 for STM32H523 has a note PB8/PB9 Not available for LQFP48, UFQFN48, WLCSP39.
LQFP64 is not mentioned there but it has only PB8 pin and no PB9 pin. That makes it unusable by boot loader. So why it is not included in the note?
My question is if the PB8 pin is still taken over by boot loader (maybe all packages use the same one)?
Is there a reliable rule for partial missing IO and boot loader? 
What am I missing here?

2. 
Another potential problem is that some pins for smaller packages are remapped which is noted in AN5959, but AN2606 does not always mention the alternative pin so it is dangerous to rely on that document alone. And not obvious you have to check others like AN5959 AN5711 to be sure.

[TDK]

PB9 isn't on the LQFP64 chip so that interface is not available on the LQFP64. It should be in the note.

PB8 will be listened to, but won't ever be "taken over" since the bus can't possibly show activity as PB9 is going to read as logic 0 always. Can't use a bus that doesn't exist to talk to the chip. Same bootloader for all "STM32H523xx/533xx devices" chips.

I recommend following AN2606 and the datasheet. You shouldn't need other documents here, though I don't see any incorrect information in them.

[PR.10]

1. Thanks for the confirming that. 

2. in AN5959  table 5 for H523
note 4. PC9 is replaced by PB4 on LQFP48, UFQFN48 and WLCSP39.
and the same in AN5711 
refers to PB4 as pin used by boot loader. 

However I originally used 48pin package and when I went over all pins of H523 using AN2606 I did not notice PB4 mentioned in relevant section. page ~279. I did not occur to me to look for more pins elsewhere. That is why I called it dangerous.
Why is PB4 pin not there then?