Posted on December 03, 2003 at 16:46We have some very old test software that we’ve used for years. I’m experiencing problems with our recent micro change from the ST 168 to the ST269. When using the intrinsic divide routines, like _divu32, inside interrupts, the micro resets. In the interrupt, I use the capture compare values to compute an incoming frequency. The problem is repeatable although not every execution of the divide causes a reset. Often several hundred interrupts are required before a reset.
The code used an old v5.1 r0 version of the tasking compiler. The same software image used on the ST168 does not have reset problems.
Any suggestions on what I should try next? We have not yet upgraded our In-Circuit Emulator to support the ST269. I suppose there could be a divide by zero or a stack overflow but why would there be a difference between the 168 and 269? Besides, I check for a zero in the denominator before attempting the divide.
