cato, 1 month ago

>Timer period shadow register and actual value register both have the value 3746

>Change the shadow register to 0, the debugger confirms

>Initiate the transfer to the actual register, the debugger confirms

>Disable peripheral

>System reset

>Back at the beginning of main()

>Check registers

>Both are 3746 again

why are you like this

lambda, 1 month ago

@cato does infineon have errata sheets? Or do they not bother because they'd be longer than the reference manual?

cato, 1 month ago

I completely forgot that I have my bootloader on this controller and with every system reset, it executes

And I think the bootloader uses the timer too, and while it does stop it before switching to the main code, it does not reset it 🤦‍♂️

cato, 1 month ago

I knew it was gonna be something dumb but not THIS dumb omg

cato, 1 month ago

okay so this isn't the full story, I just spent like another hour desperately trying to figure out why my timer refused to generate any interrupts when the code was started via the bootloader

well, you see, I wrote the code I'm woprking on now before I wrote the bootloader. Now that I have the bootloader, I'm coming back to it. I did a little thing in the bootloader before jumping to the application code to ensure nothing would mess up the process.

Can you guess what it was?

salkinium, 1 month ago

@cato Ah, a classic. On some STM32, the system ROM bootloader doesn't set the stack pointer back to what is in your application vector table, so your application has a stack overflow during boot.

WooShell, 1 month ago

@cato This smells incredibly like a "couldn't be bothered to make it right" thing.
"We'll just footnote it in the datasheet, it'll be fine."

cato, 1 month ago

@WooShell well it isn't footnoted in the datasheet