LabVIEW
Sinc Filter Implementation using DLL
Solved!
unsigned char is an unsigned byte in LabVIEW. double is a 64 byte floating point number. The asterisk means it's a pointer. This could be a single byte passed by reference or an array data pointer. Which it is can not be determined from the C syntax of the declaration alone, as C syntax is rather ambigous here. That is something that needs to be elaborated over in the documentation to the function. If it is a skalar passed as reference you are done but if it is a pointer to an array you have some more detective work to do.
You have to find out if the function expects data to read in that parameter, or if it will write something into that array. For an input parameter to the function you are also done now, because whatever array or string you will create in LabVIEW to pass to the function, will be enough. If the function however returns some information in that array the average LabVIEW programmer is tempeted to just pass in an empty array. After all a VI returning a string or array doesn't require anything special, it simply creates whatever string or array is needed automatically. Big failure!!!!!!
The C DLL has no way to create such an array in a way that LabVIEW could deal with. In fact if you would do a DLL function that creates such a memory block you could not interface it to other applications without following the exact protocol defined by this particular DLL developer, including using a very specifically defined memory manager function, documented in the funciton description, to deallocate that pointer later, or your program would create a memory leak otherwise.
Instead the usual way is that the caller allocates the necessary memory (being very careful to make sure to allocate for the maximum possible requirement of the DLL function) then calls the function with that buffer and then deallocates the memory afterwards with whatever method matches the used allocator. In LabVIEW this means you would use the Initialize Array function to allocate an array of the correct size, pass it as Array Data Pointer to the DLL and then use the resulting array or string like any other LabVIEW data value. No explicit deallocation is required here, since the original array created with Initialize Array is a fully managed LabVIEW data type already and will be taken care of by LabVIEW.
I tried to give you the general discription of what to watch out for. Because in about 97% of the cases where people get an 1097 error or the Call Library Node doesn't seem to do what they want, it's because they forgot to allocate buffers for output parameters. In your C code if you declare your two variables with
unsigned char data[MAX_DATA_BYTES];
double yout[5000000+512];
you are allocating implicetedly memory on the heap as static memory. LabVIEW being a fully managed environment doesn't have something like static memory for variables, but you still need to allocate those buffers somehow. The Initialize Array node is the method to allocate a buffer to hold an array of a certain size, so that you can pass it to the Call Library Node.
Nathan already explained to you in detail about your specific issues. I was hoping to create a somewhat more generic explanation, so that other people coming across this might see the point they have to be careful about. I'm not sure I have succeeded though.
There are many discrepancies between what you do in your VI and what your text file shows is the C code.
1) the order of parameters is VERY different.
2) You use int64 in the LabVIEW diagram, but int in the C code. Under Windows 32 bit and 64 bit (and also Linux), an int is ALWAYS a 32 bit value.
3) You read in a 1D array of strings and pass it to the Call Library Node as native data. This passes a LabVIEW array handle of LabVIEW string handles. A VERY big difference to the byte array pointer you expect in the C code.
You need to convert the strings into a byte and then configure the Call Library Node to pass this as an array of unsigned 8-bit integers, passed as Array Data Pointer.
