Become a leader in the IoT community!
Join our community of embedded and IoT practitioners to contribute experience, learn new skills and collaborate with other developers with complementary skillsets.
Join our community of embedded and IoT practitioners to contribute experience, learn new skills and collaborate with other developers with complementary skillsets.
I am using FreeRTOS on an STM32F4-Nucleo board and `vTaskGetRunTimeStats()` to measure task CPU usage with the following code:
void configureTimerForRunTimeStats(void)
{
g_osRuntimeCounter = 0;
}
unsigned long getRunTimeCounterValue(void)
{
return g_osRuntimeCounter;
}
The `g_osRuntimeCounter` is incremented in a timer callback, and I can measure CPU usage for tasks and IDLE time. However, I need to track CPU usage for specific ISRs (CAN reception and USB-OTG), which are interrupt-based.
Context:
– The interrupts for CAN and USB-OTG are triggered frequently , which are 10ms and 5ms respectively . Understanding the impact of ISRs on task performance is important in this setup.
– The hardware timer used for `g_osRuntimeCounter` runs at a frequency of 1MHz which determines how accurately I can track time.
Challenges:
– I am unsure how to extend this setup to measure CPU usage inside ISRs, especially with interrupt nesting complicating things.
– Timer increments within ISRs may affect the accuracy of the measurement.
How can I accurately measure CPU usage for these ISRs while addressing nested interrupts? Should ISR CPU usage be tracked separately from task CPU usage?
CONTRIBUTE TO THIS THREAD