Fix #79 IMU interrupt not trigerring on STM32 #81
Conversation
src/imu/imu_cpp.h
Outdated
| #ifdef ARDUINO_ARCH_STM32 | ||
| // (#79) on stm32, increase the interrupt priority, otherwise it does not get called | ||
| // configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY is the maximum priority allowed for interrupts that needs to call FreeRTOS api | ||
| HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), 0, configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY); |
There was a problem hiding this comment.
I know this works, but does not make sense to me. See
as far as I understand it configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY is a interrupt priority, but in the HAL_NVIC_SetPriority call it is used as a subpriority.
Why not use HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), 0, 0); i.e. use the highest priority?
Or use HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY, 0);, but I'm not sure if this will fix the issue.
There was a problem hiding this comment.
You're right, I mixed up preempt priority & subpriority. It worked because the preempt priority was 0 (the highest).
HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY, 0); doesn't work, which seems logical since configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY's value is already the default preempt priority.
Why not use HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), 0, 0); i.e. use the highest priority?
Because using a priority higher than configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY will not allow the ISR to call FreeRTOS functions (FreeRTOSConfig_Default.h).
It looks like no FreeRTOS calls are made in the imu ISR, so it's ok to use a priority > configLIBRARY_MAX_SYSCALL_INTERRUPT_PRIORITY, but that is something to remember for future development.
There was a problem hiding this comment.
Currently on STM32 the whole imu interrupt (including imu_loop() ) runs in interrupt context. So probably will have issues as soon as SDCARD (uses separate task) is implemented for STM32.
On ESP32/RP2 the imu stuff runs as high priority task, which is notified by the imu interrupt. I could not get this to work on STM32, but I did not make a big effort on this...
Anyway - to move ahead and get at least the interrupt working again:
Let's go with HAL_NVIC_SetPriority(hal_get_irqn_from_pin(interrupt_pin), 0, 0), as this will trigger the imu interrupt for sure.
Maybe enclose everything (including IRQn_Type hal_get_irqn_from_pin(int pin)) with #ifdef MF_STM32_INTERRUPT_HACK instead of #ifdef ARDUINO_ARCH_STM32 and set #define MF_STM32_INTERRUPT_HACK 1 at the beginning of in hal_stm32_cpp.h with some comments. Makes it easier to find this hack back later.
There was a problem hiding this comment.
Maybe enclose everything (including IRQn_Type hal_get_irqn_from_pin(int pin)) with #ifdef MF_STM32_INTERRUPT_HACK instead of #ifdef ARDUINO_ARCH_STM32 and set #define MF_STM32_INTERRUPT_HACK 1 at the beginning of in hal_stm32_cpp.h with some comments. Makes it easier to find this hack back later.
This won't worlk - leave as is...
|
Super! Many thanks for your PRs! |
2 participants
This PR fixes issue #79 where the IMU interrupt was not getting triggered.
Now right after attachInterrupt, HAL_NVIC_SetPriority is called to set the interrupt priority to the maximum allowed while still allowing FreeRTOS API calls inside the ISR.