What is a correct function input type (HDL coder)

I want to design my own 2D convolution function to implement in FPGA. I couldn't give vector like a function input. I tried to use few way but I got errors.

I created the code

function [out] = my_conv2(target,kernel)
    mult = target.*kernel;
    summ = sum(sum(mult));
    out = summ/9;
end

I want to use matrix 3x3, but I reshaped it to 1x9

But I got the errors.

I think I it's my mistake, because matlab offer me option with vector or matrix input

I can't believe that MATLAB not allow to use vector (or matrix) like a function input. Is it true?

Of course i can change my code. This way works, but it looks so silly... If I wanted conv 7x7 for example, I would use about 100 inputs...

function [out] = my_conv2(t1,t2,t3,t4,t5,t6,t7,t8,t9 ...
                        ,k1,k2,k3,k4,k5,k6,k7,k8,k9)
    mult1 = t1*k1
    mult2 = t2*k2
    mult3 = t3*k3
    mult4 = t4*k4
    mult5 = t5*k5
    mult6 = t6*k6
    mult7 = t7*k7
    mult8 = t8*k8
    mult9 = t9*k9
    % etc
    summ = mult1+mult2+mult3+mult4...
        +mult5+mult6+mult7+mult8+mult9;
    out = summ/9;
end

I know that "summ = mult1+mult2+mult3+mult4+mult5+mult6+mult7+mult8+mult9;" is really bad for real FPGA. I know about "persistent", I wrote such I because just want to try my concept.

I read some documents https://www.mathworks.com/help/fixedpoint/ug/propose-data-types-based-on-derived-ranges.html for example, but i can't belive... "input [7:0] IN [8:0];" is normal verilog code...

My main question: Can I use vector/matrix like a function input?

Answer 1

[shvlad]

I think I've fixed it. I just used "analize ranges using simulations", and it works.