Using array input and output for simulink S-function C code block

I'm having a tough time attempting to use a C code block that needs uint8 array input and output . I've tried starting from a number of legacy code examples but encounter crashes every time my C code actually tries to access the incoming memory. If someone can point me to a working example (I've already tried everything that looked relevant here) I'd appreciate it. In my latest attempt I gave up on using the legacy code functions and took code from this example Create a Basic C MEX S-Function - MATLAB & Simulink (mathworks.com) which at least runs without crashing, and started to change it to take uint8 in/out as below. However I've no idea where to find the equivalent of InputRealPtrsType for unit8_T . If I use InputPtrsType instead of InputRealPtrsType I get the error error C2100: you cannot dereference an operand of type const void`

static void mdlOutputs(SimStruct *S, int_T tid)
{
    int_T i;
    //  replced ssGetInputPortRealSignalPtrs(S,0) with ssGetInputPortSignalPtrs 
    InputRealPtrsType uPtrs = ssGetInputPortSignalPtrs(S,0);
    //    replaced ssGetOutputPortRealSignal(S,0) with ssGetOutputPortSignal
    real_T *y = ssGetOutputPortSignal(S,0);
    int_T width = ssGetOutputPortWidth(S,0);
    
    for (i=0; i<width; i++) {
    *y++ = 2.0 *(*uPtrs[i]);    
    }
}

Answer 1

[Gojo]

Hey Jeremy

I understand you wish to implement a C code block that accepts uint8 array input and perform some operations upon it.

I have tried out a simple example using "InputPtrsType" and it works as intended. If you are facing any issues upon its usage, I suggest you to share the model and code files to reproduce the issue.

Meanwhile you can refer to my model where I have modified the "mdlOutputs" as following:

static void mdlOutputs(SimStruct *S, int_T tid)
{
    InputPtrsType uPtrs = ssGetInputPortSignalPtrs(S, 0);
    uint8_T *y = (uint8_T *)ssGetOutputPortSignal(S, 0);
    int_T width = ssGetOutputPortWidth(S, 0);

    for (int_T i = 0; i < width; i++) {
        y[i] = 2 * (*(uint8_T *)uPtrs[i]);
    }
}

I am attaching "times_two" file for your reference. You can change the file extension from .txt to .c and build it using the following command:

mex times_two.c

I have created a simple Simulink model where I have used the constant block to create an input signal array and modified the Signal Attributes of the block to output uint8 data type. The S-function block is just used to double the values of the input signal.

I hope this helps!

[Jeremy]

Thanks 10^6 that did the trick. For completeness what was missing was setting input/output data types with ssSetInputPortDataType(S, 0, SS_UINT8) and ssSetOutputPortDataType(S, 0, SS_UINT8) and then using casting to (uint8_T *) for the pointer dereferences to avoid error C2100. One thing that does puzzle me is that the width is determined by using

int_T width = ssGetOutputPortWidth(S,0)

as opposed to checking the input port width ; maybe the output is assumed to be the same size as the input? What if I wanted output of a different size?

[Gojo]

Yes Jeremy, I assumed the output to be as the same size as the input and this is true for most of the models/blocks I work with. If you want the output to be of different size then surely you can use the input port width and perform the computations accordingly. Anyways, I was using it as a terminating condition for the loop as done by you in your code snippet.

Happy to know that your query is resolved!

[Jeremy]

It still puzzles me since the output port width is defined only as DYNAMICALLY_SIZED , so where is this size coming from? For an input I can understand that its size would be determined by the incoming size of the input but this is an output...