Can C code user function be used in Simulink S-Function

I have a large simulation in Simulink that uses S-functions written in c-code. For readability I would like to define a user function and call it from with an S method.

for example, see snippet below. Is this allowable in an S-Function? Is there anything I should be aware of or look out for?

static void UserFnc( real_T u, real_T v, real_T * x);      // delcare 
....
....
static void mdlOutputs(SimStruct *S, int_T tid)
{
   ..........
    UserFnc( real_T u, real_T v, real_T * x);              // use here
    .........
    .........
    UserFnc( real_T u, real_T v, real_T * x);              // and use here
    .........
}

    ........
    ........
static void UserFnc( real_T u, real_T v, real_T * x)       // define here
{
    // * User code here */
}

/*=============================*
 * Required S-function trailer *
 *=============================*/

#ifdef  MATLAB_MEX_FILE    /* Is this file being compiled as a MEX-file? */
#include "simulink.c"      /* MEX-file interface mechanism */
#else
#include "cg_sfun.h"       /* Code generation registration function */
#endif

Answer 1

[Walter Roberson]

Yes, see coder.ceval()

https://www.mathworks.com/help/simulink/slref/coder.ceval.html

On the other hand, your code structure looks to me to be more like the implementation of a Level-1 S Function, and the above is not the way to invoke a complete S-function.

[Brian Tremaine]

Just to be clear, can a user C function ** only ** be called from a Level-2 S-Function and in the manner described by your link? So, a Level-1 S-Function can ** not ** include an embedded user c-code function?

Thanks!

[Walter Roberson]

https://www.mathworks.com/help/simulink/slref/sfunction.html

The S-Function block provides access to S-functions from a block diagram. The S-function named as the S-function name parameter can be a Level-2 MATLAB® or a Level-1 or Level-2 C MEX S-function

We thus see that a C-implemented Level 1 S function is permitted.

The code you posted does not include any MATLAB, so it does not look like you are embedding C within a Level-1 S Function, it appears you are implementing the Level-1 S function as C. Which is not a problem. If you wanted to mix MATLAB and C in the same Level-1 S function, I would have to think about that more; I am not sure if it can be done.

[Brian Tremaine]

My S-Function is apparently a Level-1 and it is true it is implemented in C. What I am trying to do is the following. My mdlOutputs procedure processes data involving the current states and the input values. The computation is on the order of 300+ lines of code. I would like to process the data using a reference to a struct, for example: ComputeXY( sv), where sv is a pointer to my struct. The procedure ComputeXY is quite large and can be called in several locations in mdlOutputs, (from C switch statements), but it is only called once each execution of mdlOutputs. I have a second procedure, call it ComputeCntl(sv), that is also quite large that I would like to call before I update the output values and return from mdlOutputs.

Ordinarily for other projects my computation in mdlOutputs is only a small number of lines of code and I don't have a need to do the computation in a separate C function.

Thanks,

[Walter Roberson]

MATLAB has no problem with you creating separately linked C files if appropriate, or putting everything into one file if appropriate.

But if you are working on "current states" then you should be looking into whether you should be switching to Level 2 S Function. The major difference between Level 1 and Level 2 is the treatment of state vectors (and the fact that Level 2 allows you to access and control other blocks .)