So I found an instance where the interrupt took a little over 200us extra to respond… but I don’t see anything during that time that would have prevented it from being scheduled in the normal 5-10us. The trace was 20sec long, so I can’t post it to the forum, any ideas???
mario, I emailed that trace to you, 7zip got it down to 900K so i couldn’t post it
Looking at the trace it seems you are getting hit by the SMI interrupt, even interrupt 0 is being delayed.
Contact your board manufacturer to check if SMI can be turned off (sometimes disabling USB emulation is enough)
That’s what I’ve also started to suspect. Thank you for the help!