Hey there,
I'm new at developing for the PRU and Linux in general. I'm using the BeagleBone Black (Linux beaglebone 3.8.13-bone47), which has the AM355x processor. I'm using the prussdrv.h library to interface between the PRU and the Linux host system.
In the linux host application, I'm using a thread to listen to events from the PRU:
void *threadFunction(void *value){
do {
int notimes = prussdrv_pru_wait_event (PRU_EVTOUT_1); // event 1 is sendt from PRU, when a sample is ready to be read
unsigned int counter_val = *(pru0DataMemory_int);
prussdrv_pru_clear_event (PRU_EVTOUT_1, PRU0_ARM_INTERRUPT);
} while (1);
}
I'm sending the event to the Host system from the PRU, in his way:
MOV R31.b0, PRU0_R31_VEC_VALID | PRU_EVTOUT_1
When the application is launched, it will wait at prussdrv_pru_wait_event(), but when the event is sent from the PRU the first time, it will continue executing, and somehow not clear the event at prussdrv_pru_clear_event(), causing these three lines to loop as fast as possible.
This problem was also observed by fung - see question from June 7, 2015 - who also has been following Derek Molloy's tutorials. Derek Molloy suggests: "For later work I avoided the use of interrupts and I polled memory on a thread" - Is this really the only solution to his problem?
