Hello
I'm building an app that sends pulses to stepper motors drivers L6474. I built it around
an STM32F429VE, the L6474 libs delivered as example with the x-nucleo board for L6474.
I'm using CubeMx 5.0.1 and cube F4 V1.23.
It's a C++ app.
How it works?
Basically the app send a pwm pulse to each of the 4 drivers using HAL_TIM_PWM_Start_IT() function.
It catches the timers interrupts to introduce a new pwm pulse while managing acceleration / deceleration.
For testing purpose, the apps runs forever N pulses forward and N pulses backward.
I firstly built the app without OS and it seems to work.
Then I added FreeRTOS with a single task that does the same operation. Its stack memory is widely set to 2048.
Since this moment, I experiments Hard faults after some forward / backward movements. (not always the same amount).
I read this page https://www.freertos.org/Debugging-Hard-Faults-On-Cortex-M-Microcontrollers.html#arm_cortex_imprecise_faults
and its related. CubeMx is managing the freeRTOS initialization and regarding the freeRTOS web page, it seems to be
correctly done.
TIM7 is used as a systickTimer.
I managed to stop the app just before it crashes. Generally (but not always), it happens in HAL_TIM_IRQHandler(&htim7) called by TIM7_IRQHandler().
What I definitely don't understand is on the screenshot: htim7.Instance points to 0x40001400 which is correct but htim->Instance is null
(htim is the parameter of HAL_TIM_IRQHandler()). It should be the same because htim is htim7...
Do you have any explanation on what may happen and more generally about the crash.
I suspect a classic freeRTOS newbie configuration mistake...
Among things, I did try to raise priorityID of timers interrupt. In the map file htim7 is not far after from an array of struct used to manage motors. I instumented the code to check ids everywhere in case of memory overwrite. Everything is unsuccessfull.
Thanks
Julien
