data shared between RT and non-RT programs

[Le Chaudron Nautique <vanvan@xxxxxxxxxxxxxxxxxxxx>]

Hello,

I'm setting up a system on a raspberry Pi where 2 isolated cpus are each 
dedicated to monitoring an encoder (and compare the current values to 
the tresholds "enc_target_1" and "enc_target_2"). These tresholds will 
be written at a stage where nothing moves, thus I have no RT constraints 
while writing them.

The two other "housekeeping" cpus  will also run some programs : 
external communication, setting of enc_target_x values, launching of the 
RT-sequence on the dedicated cpus and display of the value of the 
encoder "enc_current_value" (for information only, no time constraint)

What would be the more efficient/safe way to communicate the current 
encoder value to the no-RT cpus ?

If the speed of the encoder is near the limits of the RT-cpus, reading 
the "enc_current_value" from the non-RT cpus won't be possible ?

I thought to have the RT cpus write a variable not too often (every 
0,5s, or every 50 steps on the encoder) to leave enough time for the 
non-RT to read it (the non-RT reading being preempted by the RT writing 
if it is too slow ?).

Is it correct ? Or is there a more "standardized" or more efficient way 
to do it ?

Thanks,

Mathieu

--