C# call to a C DLL is only partly functional - c#

I am learning C# from my C++/CLR background by rewriting a sample C++/CLR project in C#.
The project is a simple GUI (using Visual Studio/ Windows Forms) that performs calls to a DLL written in C (in fact, in NI LabWindows/CVI but this is just ANSI C with custom libraries). The DLL is not written by me and I cannot perform any changes to it because it is also used elsewhere.
The DLL contains functions to make an RFID device perform certain functions (like reading/writing RFID tag etc). In each of these functions, there is always a call to another function that performs writing to a log file. If the log file is not present, it is created with a certain header and then data is appended.
The problem is: the C++/CLR project works fine.
But, in the C# one, the functions work (the RFID tag is correctly written/read etc.) but there is no activity regarding the log file!
The declarations for DLL exports look like this (just one example, there are more of them, of course):
int __declspec(dllexport) __stdcall Magnetfeld_einschalten(char path_Logfile_RFID[300]);
int save_Logdatei(char path_Logdatei[], char Funktion[], char Mitteilung[]);
The save_Logdatei function is called during execution of Magnetfeld_einschalten like this:
save_Logdatei(path_Logfile_RFID, "Magnetfeld_einschalten", "OK");
In the C++/CLR project, I declared the function like this:
#ifdef __cplusplus
extern "C" {
#endif
int __declspec(dllexport) __stdcall Magnetfeld_einschalten(char path_Logfile_RFID[300]);
#ifdef __cplusplus
}
#endif
then a simple call to the function is working.
In the C# project, the declaration goes like:
[DllImport("MyDLL.dll", CallingConvention = CallingConvention.StdCall, EntryPoint = "Magnetfeld_einschalten", CharSet = CharSet.Ansi, ExactSpelling = false)]
private static extern int Magnetfeld_einschalten(string path_Logfile_RFID);
and, as I said, although the primary function is working (in this case, turning on the magnetic field of the RFID device), the logging is never done (so, the internal DLL call to save_Logdatei is not executing correctly).
The relevant code in the Form constructor is the following:
pathapp = Application.StartupPath;
pathlog = string.Format("{0}\\{1:yyyyMMdd}_RFID_Logdatei.dat", pathapp, DateTime.Now);
//The naming scheme for the log file.
//Normally, it's autogenerated when a `save_Logdatei' call is made.
Magnetfeld_einschalten(pathlog);
What am I missing? I have already tried using unsafe for the DLL method declaration - since there is a File pointer in save_Logdatei - but it didn't make any difference.
===================================EDIT==================================
Per David Heffernan's suggestion, i have tried to recreate the problem in an easy to test way. For this, i have created a very simple DLL ("test.dll") and I have stripped it completely from the custom CVI libaries, so it should be reproducible even without CVI. I have uploaded it here. In any case, the code of the DLL is:
#include <stdio.h>
int __declspec(dllexport) __stdcall Magnetfeld_einschalten(char path_Logfile_RFID[300]);
int save_Logdatei(char path_Logdatei[], char Funktion[], char Mitteilung[]);
int __declspec(dllexport) __stdcall Magnetfeld_einschalten(char path_Logfile_RFID[300])
{
save_Logdatei(path_Logfile_RFID, "Opening Magnet Field", "Success");
return 0;
}
int save_Logdatei(char path_Logdatei[], char Funktion[], char Mitteilung[])
{
FILE *fp; /* File-Pointer */
char line[700]; /* Zeilenbuffer */
char path[700];
sprintf(path,"%s\\20160212_RFID_Logdatei.dat",path_Logdatei);
fp = fopen (path, "a");
sprintf(line, "Just testing");
sprintf(line,"%s %s",line, Funktion);
sprintf(line,"%s %s",line, Mitteilung);
fprintf(fp,"%s\n",line);
fclose(fp);
return 0;
}
The C# code is also stripped down and the only thing i have added to the standard Forms project, is Button 1 (and the generated button click as can be seen). The code is this:
using System;
using System.Runtime.InteropServices;
using System.Windows.Forms;
namespace TestDLLCallCSharp
{
public partial class Form1 : Form
{
public int ret;
public string pathapp;
public string pathlog;
[DllImport("test", CallingConvention = CallingConvention.StdCall, EntryPoint = "Magnetfeld_einschalten", CharSet = CharSet.Ansi, ExactSpelling = false)]
private static extern int Magnetfeld_einschalten(string path_Logfile_RFID);
public Form1()
{
pathapp = #"C:\ProgramData\test";
pathlog = string.Format("{0}\\20160212_RFID_Logdatei.dat", pathapp);
InitializeComponent();
}
private void Form1_Load(object sender, EventArgs e)
{
}
private void button1_Click(object sender, EventArgs e)
{
ret = Magnetfeld_einschalten(pathlog);
}
}
}
As can be seen, I have avoided using an automatic naming scheme for the log file (normally i use the date) and in both the dll and the C# code, the log file is "20160212_RFID_Logdatei.dat". I have also avoided using the app path as the directory where to put the log file and instead I have opted for a folder named test i created in ProgramData
Again, no file is created at all

This looks like a simple typo in your calling code. Instead of:
ret = Magnetfeld_einschalten(pathlog);
you mean to write:
ret = Magnetfeld_einschalten(pathapp);
In the C# code, these two strings have the following values:
pathapp == "C:\ProgramData\\test"
pathlog == "C:\ProgramData\\test\\20160212_RFID_Logdatei.dat"
When you pass pathlog to the unmanaged code it then does the following:
sprintf(path,"%s\\20160212_RFID_Logdatei.dat",path_Logdatei);
which sets path to be
path == "C:\\ProgramData\\test\\20160212_RFID_Logdatei.dat\\20160212_RFID_Logdatei.dat"
In other words you are appending the file name to the path twice instead of once.

An extensive overview for P/Invoke in C# is given in Platform Invoke Tutorial - MSDN Library.
The problematic bit is you need to pass a fixed char array rather than the standard char*. This is covered in Default Marshalling for Strings.
The gist is, you need to construct a char[300] from your C# string and pass that rather than the string.
For this case, two ways are specified:
pass a StringBuilder instead of a string initialized to the specified length and with your data (I omitted non-essential parameters):
[DllImport("MyDLL.dll", ExactSpelling = true)]
private static extern int Magnetfeld_einschalten(
[MarshalAs(UnmanagedType.LPStr)] StringBuilder path_Logfile_RFID);
<...>
StringBuilder sb = new StringBuilder(pathlog,300);
int result = Magnetfeld_einschalten(sb);
In this case, the buffer is modifiable.
define a struct with the required format and manually convert your string to it:
[StructLayout(LayoutKind.Sequential, CharSet=CharSet.Ansi)]
struct Char300 {
[MarshalAs(UnmanagedType.ByValTStr,SizeConst=300)]String s;
}
[DllImport("MyDLL.dll")]
private static extern int Magnetfeld_einschalten(Char300 path_Logfile_RFID);
<...>
int result = Magnetfeld_einschalten(new Char300{s=pathlog});
You can define an explicit or implicit cast routine to make this more straightforward.
According to UnmanagedType docs, UnmanagedType.ByValTStr is only valid in structures so it appears to be impossible to get the best of both worlds.

The String is in Unicode format, convert it to byte[]
Encoding ec = Encoding.GetEncoding(System.Threading.Thread.CurrentThread.CurrentCulture.TextInfo.ANSICodePage);
byte[] bpathlog = ec.GetBytes(pathlog);
and change parameter type to byte[]
[DllImport("MyDLL.dll", CallingConvention = CallingConvention.StdCall, EntryPoint = "Magnetfeld_einschalten", CharSet = CharSet.Ansi, ExactSpelling = false)]
private static extern int Magnetfeld_einschalten(byte[] path_Logfile_RFID);
For me it is working
JSh

Related

Tracing source of access violation in unmanaged C# code

I am currently working on some C# code that talks to a C++ dll. This is not an area in which I - or anyone else at my company - has any experience. It's been an eye-opener to say the least.
After a lot of reading, trial and error, and frustration, I've managed to iron out most of the kinks and get something that's largely functional. However, from time to time, it still throws this at me ...
System.AccessViolationException: Attempted to read or write protected memory. This is often an indication that other memory is corrupt.
.. and then dies. This error only appears when I run the call on parallel threads - it's fine single threaded. This dll is supposed to be thread safe and we've good reason to believe it ought to be, if handled correctly.
The cause of this error is always a call to the same function:
[DllImport(DLL, SetLastError = true, CharSet = CharSet.Ansi)]
public static extern int QABatchWV_Close(IntPtr vi1);
I have the header file for the library, which defines this function as:
__declspec(dllimport) int __stdcall QABatchWV_Close(int);
From what I understand there are additional tools at my disposal like SafeHandle and MarshalAs. But, frankly, I'm unsure as to how to best deploy them in this situation.
This error tends to take several hours of use time to show up, so tweaking and hoping isn't going to be a productive approach here. Can anyone point me as to what I might be doing wrong in calling down to the C++ function?
Well, first of all you don't need setting Charset here, because there are no strings.
Second of all - function in cpp should be declared as exported not imported, so it should look like:
__declspec(dllimport) int __stdcall QABatchWV_Close(int);
Next, you should set calling convention in your C# code to stdcall:
[DllImport(DLL, SetLastError = true, CallingConvention=CallingConvention.Stdcall)]
Next you should have int instead of IntPtr in C# code. And I'm nearly sure that name of this function (in C++ dll) is mangled and it's not QABatchWV_Close but rather something like QABatchWV_Close#32. You should check it using "dll export viewer".
Have a look at the following code which I use to call a c (not c++) dll. I know it is not really an answer to your question, but perhaps you can use some of this going foreward.
Note the "CallingConvention"-specifier in the dll declaration and also the "FreeGlobal" in the "finally" part of the try catch.
public class csInterface
{
[DllImport(#"myDLL.dll", EntryPoint = "dllFunc", CallingConvention = CallingConvention.StdCall)]
private static extern void dllFunc(IntPtr inp, IntPtr outp);
public static int myDll(ref MyInput myInput, ref MyOutput myOutput)
{
int sizeIn, sizeOut;
IntPtr ptr_i = IntPtr.Zero, ptr_u = IntPtr.Zero;
sizeIn = Marshal.SizeOf(typeof(myInput));
sizeOut = Marshal.SizeOf(typeof(myOutput));
/* Calling C */
try
{
ptr_i = Marshal.AllocHGlobal(sizeIn);
ptr_u = Marshal.AllocHGlobal(sizeOut);
Marshal.StructureToPtr(myInput, ptr_i, true);
Marshal.StructureToPtr(myOutput, ptr_u, true);
dllFunc(ptr_i, ptr_u);
myOutput = (MyOutput)(Marshal.PtrToStructure(ptr_u, typeof(MyOutput)));
}
catch (Exception)
{
//Return something meaningful (or not)
return -999;
}
finally
{
//Free memory
Marshal.FreeHGlobal(ptr_i);
Marshal.FreeHGlobal(ptr_u);
}
//Return something to indicate it all went well
return 0;
}
}
In C# I declare my types
[StructLayout(LayoutKind.Sequential, CharSet = CharSet.Ansi)]
public struct MySubType
{
public int a;
public double b;
}
[StructLayout(LayoutKind.Sequential, CharSet = CharSet.Ansi)]
public struct MyInput
{
[MarshalAsAttribute(UnmanagedType.ByValTStr, SizeConst = 4)]
public string aString; //A string of length 3
public bool aBoolean;
public int anInt;
public char aChar;
public double aDouble;
[MarshalAs(UnmanagedType.ByValArray, ArraySubType = UnmanagedType.Struct, SizeConst = 12)]
public MySubType[] aSubType; //Array of struct of length 12
}
And something similar for the output.
Now in C (its probably the same or similar in c++) i declare my dll
__declspec(dllexport) void _stdcall dllFunc(MyCInput *myCInput, MyCOutput *myCOutput)
{
//Code
}
And the corresponding C types which obviously have to mirror the C# types exactly
typedef struct
{
int a;
double b;
} MyCSubType;
typedef struct
{
char aString[4];
int aBoolean; //This needs to be cast over to your C boolean type
int anInt;
char aChar;
double aDouble;
MyCSubType myCSubType[12];
} MyCType;
Now the types I have used in this example do not exactly match what I have used in my code, and i have not tested this code. So there may be typos and such, but the "principle" is ok.

P/Invoke in ASP.NET (Reading/writing text file from dll)

I have a C++ Win32 program in which I am writing and reading a text file. This C++ program generates a dll and I am referencing this dll in my ASP.NET web application.
Using P/Invoke, I am calling methods to read and write file from this dll.
The dll is working fine when I tested this out with P/invoke in WPF application.
The reference dll is in the bin/Debug folder for this WPF app, and the write method in dll when called generates a text file in the same folder.
Further, from the same folder, I can use the dll's read method to read the text file.
However, when I call the Dll methods from my ASP.NET web app, the genearted file goes to some other directory (most probably because the dll is loaded somewhere else to execute) and I am not able to locate where this generated file goes (without any error)
Similar to desktop application, is there some way that the fie will be written in bin folder itself, so that I can read from the bin folder itself?
Example code:
.cpp file
extern "C" D_API int Write1()
{
ofstream myfile;
myfile.open ("example.txt");
myfile << "Writing this to a file.\n";
myfile.close();
return 1;
}
extern "C" D_API char* Read1()
{
ifstream myReadFile;
myReadFile.open("test.txt");
char output[100];
if (myReadFile.is_open())
{
while (!myReadFile.eof())
{
myReadFile >> output;
}
}
return output;
}
C# .aspx.cs
[DllImport("Testing1.dll", EntryPoint = "fnTest", CallingConvention = CallingConvention.Cdecl, CharSet = CharSet.Ansi)]
public static extern int Write1();
[DllImport("Testing1.dll", EntryPoint = "ReadTest", CallingConvention = CallingConvention.Cdecl, CharSet = CharSet.Ansi)]
public static extern StringBuilder Read1();
Since you are using relative paths, the file will be relative to the working directory of the process at the point at which you call into the native code. This is a rather brittle arrangement as you have discovered.
I would solve the problem by adding an extra string parameter to the native code that specifies the full path of the file to use. You can generate this easily enough from your managed code I am sure.
Native code
extern "C" D_API int WriteTest(char *filename)
{
....
myfile.open(filename);
....
}
Managed code
[DllImport("Testing1.dll", CallingConvention = CallingConvention.Cdecl)]
public static extern int WriteTest();
The other point to make is that your function to read data is incorrect. It attempts to return a stack allocated buffer. You need to allocate a buffer in the managed code and then pass that to the native code. Perhaps something like this:
extern "C" D_API int ReadTest(char *filename, char* buffer, int len)
{
//read no more than len characters from filename into buffer
}
And on the managed side:
[DllImport("Testing1.dll", CallingConvention = CallingConvention.Cdecl)]
public static extern int ReadTest(string filename, StringBuilder buffer, int len);
....
StringBuilder buffer = new StringBuilder(100);
int retval = ReadTest(FullySpecifiedFileName, buffer, buffer.Capacity);

Unable to find an entry point when calling C++ dll in C#

I am trying to learn P/Invoke, so I created a simple dll in C++
KingFucs.h:
namespace KingFuncs
{
class KingFuncs
{
public:
static __declspec(dllexport) int GiveMeNumber(int i);
};
}
KingFuns.cpp:
#include "KingFuncs.h"
#include <stdexcept>
using namespace std;
namespace KingFuncs
{
int KingFuncs::GiveMeNumber(int i)
{
return i;
}
}
So it does compile, then I copied this dll into my WPF's debug folder, with code:
[DllImport("KingFuncDll.dll", EntryPoint = "GiveMeNumber", SetLastError = true, CharSet = CharSet.Ansi, ExactSpelling = true, CallingConvention = CallingConvention.StdCall)]
public static extern int GiveMeNumber(
int i
);
And calling it in button click:
private void Button_Click(object sender, RoutedEventArgs e)
{
int num = GiveMeNumber(123);
}
But it gives me exception:
Unable to find an entry point named 'GiveMeNumber' in DLL
'KingFuncDll.dll'.
Really.... what have I done wrong... It obviously able to find the DLL, otherwise would be another exception. But my method name is exactly the same.... I can't think of other reason.
You need to use extern "C" when you export your function so that you suppress C++ name mangling. And you also should not try to p/invoke to members of a class. Use free functions instead:
extern "C" {
__declspec(dllexport) int GiveMeNumber(int i)
{
return i;
}
}
On the managed side your DllImport attribute is all wrong. Don't use SetLastError which is for Win32 APIs only. Don't bother setting CharSet if there are not text parameters. No need for ExactSpelling. And the calling convention is presumably Cdecl.
[DllImport("KingFuncDll.dll", CallingConvention=CallingConvention.Cdecl)]
public static extern int GiveMeNumber(int i);
The problem is that you are declaring the C++ "function" inside a C++ class and are telling P/Invoke to use StdCall.
Try to declare a C++ function outside a class and and export it like you did. Then your code should work.
If you really must have a C++ function inside a class, take a look at CallingConvention.ThisCall. But then you are responsible for creating your unmanaged class instance and pass it as the first parameter of your P/Invoke call
The entry point name of a dll file is given in .exp file which is found in debug folder where other source files are present. If dumpbin doesn't work you can try this.

Convert managed String (C#) to LPCOLESTR (C++)

I wave a method in C++ that receives a parameter of the LPCOLESTR type. I'm accessing this method through C#, but I can't make the correct conversion between String and this type.
Let's say the method signinature in C++ is:
void Something(LPCOLESTR str)
In C#, I'm trying to call it (all reference issues to access the method through a DLL have been solved already):
String test = "Hello world";
Something(test);
But with no luck, of course. If anyone can help me, I'd be very glad. Thank you!
Code snippet:
As an example, here's my C++ portion of code, defined in the file MixedCodeCpp.h (CLR Class Library)
#include "windows.h"
#pragma once
using namespace System;
namespace MixedCodeCpp
{
public ref class MyClass
{
public:
HRESULT Something(LPCOLESTR str)
{
return S_OK;
}
};
}
And here's my code in C# (I've added a reference to the C++ project in the C# project, through Visual Studio):
StringBuilder sb = new StringBuilder();
sb.Append("Hello world");
MixedCodeCpp.MyClass cls = new MixedCodeCpp.MyClass();
cls.Something(sb);
The argument will appear as Char* on the C# side. That requires unsafe code, like this:
unsafe static void CallSomething(MyClass obj, string arg) {
IntPtr mem = Marshal.StringToCoTaskMemUni(arg);
try {
obj.Something((Char*)mem);
}
finally {
Marshal.FreeCoTaskMem(mem);
}
}
It makes very little sense to expose the LPCOLESTR to other managed code. This method really should accept a String^ and convert to wchar_t* internally.
Try StringBuilder instead of String thusly:
System.Text.StringBuilder test = new System.Text.StringBuilder ();
test.Append("Hello world");
Something(test);
I've used it that way in pinvoke to Win32 functions that required various string pointers as parameters. Not sure it will work with your API but it's worth a shot. Here's some MSDN info about the process. And here is another.
Here's an arbitrary sample of what your import statement and declaration ought to look like. (To be taken with a grain of salt.)
[DllImport(SomeLib.SomeName, CharSet = CharSet.Unicode, SetLastError = true)]
[return: MarshalAs(UnmanagedType.Bool)]
public static extern bool Something(StringBuilder pMyString);
StringBuilder str = new StringBuilder(MAX_PATH);
DWORD uSize;
bool b = Something(str);

C# & C++, runtime error when call C++ dll from C#

I have written a C++ wrapper DLL for C# to call. The DLL was tested and worked fine with my C++ test program.
now integrated with C#, I got runtime error and crashed. Cannot use debugger to see more details.
The C++ side has only one method:
#ifdef DLLWRAPPERWIN32_EXPORTS
#define DLLWRAPPERWIN32_API __declspec(dllexport)
#else
#define DLLWRAPPERWIN32_API __declspec(dllimport)
#endif
#include "NB_DPSM.h"
extern "C" {
DLLWRAPPERWIN32_API int WriteGenbenchDataWrapper(string fileNameToAnalyze,
string parameterFileName,
string baseNameToSaveData,
string logFileName,
string& message) ;
}
in the C# side, there is a definition,
[DllImport("..\\..\\thirdParty\\cogs\\DLLWrapperWin32.dll")]
public static extern int WriteGenbenchDataWrapper(string fileNameToAnalyze,
string parameterFileName,
string baseNameToSaveData,
string logFileName,
ref string message);
and a call:
string msg = "";
int returnVal = WriteGenbenchDataWrapper(rawDataFileName,
parameterFileName, outputBaseName, logFileName, ref msg);
I guess there must be something wrong with the last parameter of the function. string& in C++ should be ref string in C#?
EDIT:
Do we really need the extern "C"?
EDIT 2:
after I remove the extern "C from the dll, I got the EntryPointNotFoundException. When I look at the dll by using DLL Export Viewer, I found the function name is "int __cdecl WriteGenbenchDataWrapper(class std:: ..." Do I need to include the " __cdecl"?
There are a bunch of rules for marsheling with PInvoke.
For reference Marsheling between managaed & unmanaged
Focusing on the C# side first.
If you knew a reasonable size of the message up front you could use StringBuilder type and define that size, something like.
[DllImport("DLLWrapperWin32.dll")]
public static extern int WriteGenbenchDataWrapper(string fileNameToAnalyze,
string parameterFileName,
string baseNameToSaveData,
string logFileName,
StringBuilder message
int messageLength );
Impression from the name message (and other posts) indiciates you don't know the size up front, and you won't be passing a partial message to the function so maybe
[DllImport("DLLWrapperWin32.dll")]
public static extern int WriteGenbenchDataWrapper(in string fileNameToAnalyze,
in string parameterFileName,
in string baseNameToSaveData,
in string logFileName,
out string message );
Now on the C/C++ side - to match the second definition
extern "C" // if this is a C++ file to turn off name mangling for this function only
int WriteGenbenchDataWrapper( char * fileNameToAnalyze,
char * parameterFileName,
char * baseNameToSaveData,
char * logFileName,
char ** message ) {
string internalMessage;
SomeFunc( internalMessage ); // these functions won't have extern "C" applied
* message = (char *)::CoTaskMemAlloc(internalMessage.length()+1);
strcpy(* message, internalMessage.c_str());
}
Consideration of unicode/ansi strings is also important, refer to [MarshalAsAttribute(UnmanagedType.LPWSTR)]
For release mode you will want to remove your development path settings "..\..\thirdParty\cogs"
In your C++ code:
I've always needed the extern "C". C++ mangles function names if you don't (the mangling is needed to support function overloading). The extern "C" tells it not to do this.
I also will declare the functions as __stdcall. I believe you can tell C# which type of calling convention to use, but I think __stdcall is the default.
As far as passing a string object, I'm not sure about that, I stick to only using primitives for parameter passing, so I would use const char * and adjust accordingly in my C++ code.
Also, I try to avoid passing by reference. Rather, if I need to return several values, I'll set up a series of getters to handle this (a const char * returns as an IntPtr).
In your C# code:
I use String for the const char *, int for int, and so on. I believe Microsoft has a chart somewhere to tell you what should sub in for what.
When dealing with a returned string, you need to convert it to ANSI. This can be done with a call to Marshal.PtrToStringAnsi().
For Example:
In my C++ code:
extern "C" __declspec(dllexport) const char* __stdcall GetCompany(const char *In) {
return MyGetCompany(In); // Calls the real implementation
}
In my C# code:
[DllImport("TheDLL.dll", EntryPoint = "GetCompany")]
private static extern IntPtr privGetCompany(String In);
// Call this one, not the one above:
public String GetProvince(String In)
{
return Marshal.PtrToStringAnsi(privGetCompany(In));
}
One final note, if you're running on a 64-bit machine, the 'Any CPU' configuration will make a 64-bit C# executable, which will need a 64-bit DLL. If you only have a 32-bit DLL, you'll need to add a configuration (x86).
The error message you got indicates that your C# program is probably finding the DLL correctly and the function as well, so name mangling is not likely the problem. It sounds like calling convention issue or a problem with the parameter passing.

Categories

Resources