Could you please provide the mapping of failure events to the respective CMU instances (CMU0, CMU1, and CMU2)?
The interrupt documentation lists a total of seven interrupts, but it is not clear which failure events are handled by each CMU instance and their corresponding failure actions.
Please see attached imagessystem clock monitoring.pngsystem clock monitoring.png
Interruptmapping.pngInterruptmapping.png
What is the exact meaning of the "Reset Reaction Interrupt"?
Does this interrupt get generated prior to a destructive reset, allowing software intervention before the reset occurs?The clock monitoring latency can be configured from 1 µs to 1 ms.
Does selecting a lower latency effectively reduce the debounce/filtering time and make clock failure detection more sensitive or faster?
Hi@WadkarY
1.Could you please provide the mapping of failure events to the respective CMU instances (CMU0, CMU1, and CMU2)?
As can be seen, only CMU_FC_0，CMU_FM_1,CMU_FM_2 is configurable as an interrupt; the others should be treated by default as sources of a destructive reset, rather than as standard CMU interrupts.
CMU_FC_0->CMU0
CMU_FM_1->CMU1
CMU_FM_2->CMU2
CMU_FC_3->CORE_CLK_FAIL CMU reset reaction interrupt
CMU_FC_4->AIPS_PLAT_CLK_FAIL CMU reset reaction interrupt
CMU_FC_5->HSE_CLK_FAIL CMU reset reaction interrupt
CMU_FC_6->CM7_CORE_CLK_FAIL CMU reset reaction interrupt
2.Does this interrupt get generated prior to a destructive reset, allowing software intervention before the reset occurs?
No， expect destructive reset behavior unless the MC_RGM/DCM destructive-reset interrupt bypass is deliberately configured
for example:
1.CMU_FC_4 detects AIPS_PLAT_CLK frequency exceeding the threshold
↓
Destructive Reset asserted + IRQ 215 issued simultaneously
↓ Almost instantaneously
MCU reset → Restart from Reset Vector
↓
Software reads MC_RGM.DES[AIPS_PLAT_CLK_FAIL] to identify the reset cause
3.Does selecting a lower latency effectively reduce the debounce/filtering time and make clock failure detection more sensitive or faster?
This parameter is related to REF_CNT.
•Higher values of RCCR[REF_CNT] results in longer measurement window, leading to better accuracy in monitored clock check.
•Lower values of RCCR[REF_CNT] results in shorter measurement window, leading to faster FHH and FLL event response, but higher inaccuracy in reported result.