STM32 ISR Count Doesn't match the real data - Is this STM32 interrupt handling bug?

[frankqt]

Posted on August 15, 2011 at 05:52

Hi,

I am doing a very simple test, enable EXT11 interrupt. In the ISR, I toggle a GPIO pin. I connect the scope to both EXT11 line and the GPIO I am toggling in the ISR.

They somehow do not match. It seems ISR is skipping 2 beats out of 8 (at this in my test case). 

Has anybody had a similar issue? What I could be doing wrong? The whole thing is so simple that it should just work, unless there is some major bug in interrupt handling part of STM32. 

(even bloody forum post took several tries, come on ST get a grip)

#stm32f2

[Tesla DeLorean]

Posted on August 15, 2011 at 21:50

(even bloody forum post took several tries, come on ST get a grip)

The forums a pretty awful creation from Microsoft, you've succeeded in posting the same question 3 times. Certain error messages indicate failures after the post was accepted (the Share Point one?), others occur when you've been logged off after a timeout. I can't say I've lost interrupts. If the toggle rate is too high you won't be able to service them quickly enough or you'll violate setup/hold times. The core does not latch multiple events. You must also dismiss interrupts several cycles before exiting an interrupt service routine as there are pipelining issues writing to peripheral blocks. Suggest you clear the interrupt, then increment your counter.