…or, in fact, the cc_slave core.
I’ve been looking into the code, and noticed that a hardware serial rx interrupt initiates a rather long chain of calls:
in ReUART.cpp
-
ISRon an USART0_RX_vect vector calls_rx_complete_irq() -
_rx_complete_irq()callsuart_byte_recv() -
uart_byte_recv()callscc_parse()fromcore.c, and, if the incoming message is complete, -
cc_parse()callsparcer(), also incore.c
I think this last step is a bit too much to be processed in an interrupt, so what I’ve done is I’ve included a simple flag into cc_handle_t:
typedef struct cc_handle_t {
void (*response_cb)(void *arg);
void (*events_cb)(void *arg);
int comm_state, msg_state;
cc_msg_t *msg_rx, *msg_tx;
int device_id;
int msg_rx_ok; // <-- ADDED
} cc_handle_t;
that in cc_parser I set to 1 if a message is received, and let the code finish processing the interrupt without calling the parser():
// crc
case 6:
if (crc8(msg->header, CC_MSG_HEADER_SIZE + msg->data_size) == byte)
{
//parser(handle); // commented this out, we'll do this later
handle->msg_rx_ok = 1; // ADDED for quicker exit out of an interrupt
msg_ok = 1;
}
and then I process the received message with parcer() in cc_process(), basically in the loop() of a main sketch, right before we get to checking the actuators’ states!
void cc_process(void)
{
// ADDED BEGIN
// see if we have an incoming message to process
cc_handle_t *handle = &g_cc_handle;
if (handle->msg_rx_ok) {
handle->msg_rx_ok = 0;
parser(handle);
}
// ADDED END
// process each actuator going through all assignments
// data update messages will be queued and sent in the next frame
cc_actuators_process(g_cc_handle.events_cb);
}
It seems to work smoothly. This even has allowed me to do some more heavy coding in the parcer(), like introducing new events and calling callbacks on those with output to LCD - for debugging purposes et al, which in case of the original code was not possible, e.g.:
raise_event(handle, CC_EV_HANDSHAKE_INITIATED, 0);
...
raise_event(handle, CC_EV_HANDSHAKE_SUCCESSFULL, 0);
...
raise_event(handle, CC_EV_DEBUG, 0);
@jon, if you guys can also look into this, if it makes sense for the future incarnations of the code?
Cheers!
P.S. Ooh, now that I’m on a break form my band, I’m having so much fun digging this code!
P.P.S. But I can’t, for the life of me, get the library running on Arduino Mega. I hope this little change helps me debugging it.
One look at the pinout on the Arduino site, et voila! RTFM, kids!