Programming Forums > Application Development > Assembly

Klarre · Nov 9th, 2005, 12:06 PM

How do I call a function with an array as argument?
I have been written this code, but it does'nt work of course...

#include <iostream>

void printArray(float m[])
{
	std::cout << m[0] << std::endl;
	std::cout << m[1] << std::endl;
	std::cout << m[2] << std::endl;
}

int main()
{
	float m1 = 3, m2 = 2, m3 = 9;

	_asm
	{
		push m3 ; How should I prepare the floats?
		push m2
		push m1
		call printArray
	}

	system("pause");
	return 0;
}

/Klarre

Polyphemus_ · Nov 9th, 2005, 12:30 PM

You create an array somewhere, and then you push the address of the array on the stack, IIRC.

DaWei · Nov 9th, 2005, 1:03 PM

You may either push the address of the array onto the stack, or you may push the individual elements. In either case, the function needs to be aware of extent (and order). Pushing the address allows the contents of the original array to be modified from inside the function, pushing the individual values does not.

If you wish to deal with floats, your function must understand their construction internally, or use the/a floating-point processor.

Klarre · Nov 9th, 2005, 1:11 PM

Quote:

Originally Posted by Polyphemus_

You create an array somewhere, and then you push the address of the array on the stack, IIRC.

I can't get it work. Would you mind modifying the code I posted? That would be worth alot! But you are only allowed to change the code in the asm scope, cause I need the arguments to be stored as three seperated floats, and not as an array.

/Klarre

Klarre · Nov 9th, 2005, 1:25 PM

Quote:

Originally Posted by DaWei

Pushing the address allows the contents of the original array to be modified from inside the function, pushing the individual values does not.

Ok. I wish to push the individuals. But how do I do this? What should I change in the code, to make it work?

/Klarre

DaWei · Nov 9th, 2005, 2:18 PM

You say you can only change the asm part, but you have printArray set up to accept an address. You can't do both, so make a decision.

Klarre · Nov 9th, 2005, 2:48 PM

OK. It seems like it is not possible to do as I want. Therefor I post the actual problem I am trying to solve.
I am working on a lookAt function to an OpenGL application. The function will work as gluLookAt, with the same arguments. I did the implementation several months ago using C++, and now wants to convert it to assembly code.

(Toggle Plain Text)

void Camera::lookAt(Vector3d eye, Vector3d center, Vector3d up)
{
	// Creates an move/rotation matrix
	Vector3d f(center.mX - eye.mX, center.mY - eye.mY, center.mZ - eye.mZ);
	Vector3d s = (f.normalize()).cross(up.normalize());
	Vector3d u = (s.normalize()).cross(f);

	GLfloat m[] = {s.mX,  u.mX, -f.mX, 0,
	                    s.mY,  u.mY, -f.mY, 0,
	                    s.mZ,  u.mZ, -f.mZ, 0,
	                         0,       0,       0, 1};
	glMultMatrixf(m);
	glTranslatef(-eye.mX, -eye.mY, -eye.mZ);
}

I have implemented the cross and normalize functions. So the only thing left is to make a call to void glMultMatrixf(const GLfloat *m). My idea to do this is to store the matrix (m[]) like

(Toggle Plain Text)

fx dd ?
fy dd ?
fz dd ?
sx dd ?
sy dd ?
sz dd ?
ux dd ?
uy dd ?
uz dd ?

...and then make a call to glMultMatrixf with these variables. But how will this call look like?

/Klarre

DaWei · Nov 9th, 2005, 3:59 PM

I'm going to ignore the question, per se. If you are deciding to move from a high-level language implementation to an assembly-code implementation, you have to understand a number of different things. You have to understand how the high-level language maps its abstraction to machine code. You have to understand what capabilities machine code has at its own level. You have to understand how to shift your mental paradigm from one to the other. I can't know where you stand there. It isn't a big help that you SEEM to think that a call like, "void printArray(float m[])," is passing an array to a function. If you don't know what that's passing, and if you should decide to think it's a "pointer", which it also isn't, exactly, well, it just ain't so. You might want to read the small blurb referenced in my sig as "Pointer Basics", with particular attention to "What is NOT a pointer."

I'm not trying to discourage you unduly. There is just little point in following this up, if you aren't prepared. Please post back with additional elucidation.

Klarre · Nov 9th, 2005, 4:30 PM

Thansk for your answears DaWei. As you wrote on your website that "An array is not a pointer" I am wondering if this is bad or not correct written code then. It works, and give me correct output though.

(Toggle Plain Text)

#include <iostream>

void printArray(const float* m) // Changed from float m[]
{
	std::cout << m[0] << std::endl;
	std::cout << m[1] << std::endl;
	std::cout << m[2] << std::endl;
}

int main()
{
	float m[] = {1.0f, 5.3f, 7.0f};
	float* x = m;

	_asm
	{
		push x
		call printArray
	}

	system("pause");
	return 0;
}

/Klarre

DaWei · Nov 9th, 2005, 4:51 PM

"x" is a pointer. "m" is the address of the first element of the array. "x" consequently points to "m". When you use the address of the array, "float []", the compiler knows the difference (one level of indirection) and generates code accordingly. It's the compiler's desire to make life easier for you that leads to confusion. We all know we can't assign a collection of objects to an array using the "=", but we all write, occassionally, "char myString [] = "abcdefg";" Again, the compiler is spoiling us. The compiler generates the array, element by element, as we'd have to do, if we did it at run-time. You can see why I call such things, "Confustoids."

Nov 9th, 2005, 12:06 PM	#1
Klarre Game engine designer Join Date: May 2005 Location: Sweden Posts: 297 Rep Power: 4	[x86] Arrays to functions How do I call a function with an array as argument? I have been written this code, but it does'nt work of course... (Toggle Plain Text) #include <iostream> void printArray(float m[]) { std::cout << m[0] << std::endl; std::cout << m[1] << std::endl; std::cout << m[2] << std::endl; } int main() { float m1 = 3, m2 = 2, m3 = 9; _asm { push m3 ; How should I prepare the floats? push m2 push m1 call printArray } system("pause"); return 0; } #include <iostream> void printArray(float m[]) { std::cout << m[0] << std::endl; std::cout << m[1] << std::endl; std::cout << m[2] << std::endl; } int main() { float m1 = 3, m2 = 2, m3 = 9; _asm { push m3 ; How should I prepare the floats? push m2 push m1 call printArray } system("pause"); return 0; } /Klarre

Nov 9th, 2005, 1:03 PM	#3
DaWei Resident Grouch Join Date: Jun 2005 Posts: 6,453 Rep Power: 10	You may either push the address of the array onto the stack, or you may push the individual elements. In either case, the function needs to be aware of extent (and order). Pushing the address allows the contents of the original array to be modified from inside the function, pushing the individual values does not. If you wish to deal with floats, your function must understand their construction internally, or use the/a floating-point processor. __________________ Abstraction doesn't make it impossible to write bad code; it makes it possible to write superior code. Contributor's Corner: Grumpy on C++ Exceptions DaWei on Pointers

Nov 9th, 2005, 2:18 PM	#6
DaWei Resident Grouch Join Date: Jun 2005 Posts: 6,453 Rep Power: 10	You say you can only change the asm part, but you have printArray set up to accept an address. You can't do both, so make a decision. __________________ Abstraction doesn't make it impossible to write bad code; it makes it possible to write superior code. Contributor's Corner: Grumpy on C++ Exceptions DaWei on Pointers

Nov 9th, 2005, 2:48 PM	#7
Klarre Game engine designer Join Date: May 2005 Location: Sweden Posts: 297 Rep Power: 4	OK. It seems like it is not possible to do as I want. Therefor I post the actual problem I am trying to solve. I am working on a lookAt function to an OpenGL application. The function will work as gluLookAt, with the same arguments. I did the implementation several months ago using C++, and now wants to convert it to assembly code. (Toggle Plain Text) void Camera::lookAt(Vector3d eye, Vector3d center, Vector3d up) { // Creates an move/rotation matrix Vector3d f(center.mX - eye.mX, center.mY - eye.mY, center.mZ - eye.mZ); Vector3d s = (f.normalize()).cross(up.normalize()); Vector3d u = (s.normalize()).cross(f); GLfloat m[] = {s.mX, u.mX, -f.mX, 0, s.mY, u.mY, -f.mY, 0, s.mZ, u.mZ, -f.mZ, 0, 0, 0, 0, 1}; glMultMatrixf(m); glTranslatef(-eye.mX, -eye.mY, -eye.mZ); } void Camera::lookAt(Vector3d eye, Vector3d center, Vector3d up) { // Creates an move/rotation matrix Vector3d f(center.mX - eye.mX, center.mY - eye.mY, center.mZ - eye.mZ); Vector3d s = (f.normalize()).cross(up.normalize()); Vector3d u = (s.normalize()).cross(f); GLfloat m[] = {s.mX, u.mX, -f.mX, 0, s.mY, u.mY, -f.mY, 0, s.mZ, u.mZ, -f.mZ, 0, 0, 0, 0, 1}; glMultMatrixf(m); glTranslatef(-eye.mX, -eye.mY, -eye.mZ); } I have implemented the cross and normalize functions. So the only thing left is to make a call to void glMultMatrixf(const GLfloat *m). My idea to do this is to store the matrix (m[]) like (Toggle Plain Text) fx dd ? fy dd ? fz dd ? sx dd ? sy dd ? sz dd ? ux dd ? uy dd ? uz dd ? fx dd ? fy dd ? fz dd ? sx dd ? sy dd ? sz dd ? ux dd ? uy dd ? uz dd ? ...and then make a call to glMultMatrixf with these variables. But how will this call look like? /Klarre

Nov 9th, 2005, 3:59 PM	#8
DaWei Resident Grouch Join Date: Jun 2005 Posts: 6,453 Rep Power: 10	I'm going to ignore the question, per se. If you are deciding to move from a high-level language implementation to an assembly-code implementation, you have to understand a number of different things. You have to understand how the high-level language maps its abstraction to machine code. You have to understand what capabilities machine code has at its own level. You have to understand how to shift your mental paradigm from one to the other. I can't know where you stand there. It isn't a big help that you SEEM to think that a call like, "void printArray(float m[])," is passing an array to a function. If you don't know what that's passing, and if you should decide to think it's a "pointer", which it also isn't, exactly, well, it just ain't so. You might want to read the small blurb referenced in my sig as "Pointer Basics", with particular attention to "What is NOT a pointer." I'm not trying to discourage you unduly. There is just little point in following this up, if you aren't prepared. Please post back with additional elucidation. __________________ Abstraction doesn't make it impossible to write bad code; it makes it possible to write superior code. Contributor's Corner: Grumpy on C++ Exceptions DaWei on Pointers

Nov 9th, 2005, 12:30 PM	#2
Polyphemus_ Expert Programmer Join Date: Aug 2005 Location: Rotterdam, the Netherlands Posts: 942 Rep Power: 4	You create an array somewhere, and then you push the address of the array on the stack, IIRC.

Nov 9th, 2005, 4:30 PM	#9
Klarre Game engine designer Join Date: May 2005 Location: Sweden Posts: 297 Rep Power: 4	Thansk for your answears DaWei. As you wrote on your website that "An array is not a pointer" I am wondering if this is bad or not correct written code then. It works, and give me correct output though. (Toggle Plain Text) #include <iostream> void printArray(const float* m) // Changed from float m[] { std::cout << m[0] << std::endl; std::cout << m[1] << std::endl; std::cout << m[2] << std::endl; } int main() { float m[] = {1.0f, 5.3f, 7.0f}; float* x = m; _asm { push x call printArray } system("pause"); return 0; } #include <iostream> void printArray(const float* m) // Changed from float m[] { std::cout << m[0] << std::endl; std::cout << m[1] << std::endl; std::cout << m[2] << std::endl; } int main() { float m[] = {1.0f, 5.3f, 7.0f}; float* x = m; _asm { push x call printArray } system("pause"); return 0; } /Klarre

Nov 9th, 2005, 4:51 PM	#10
DaWei Resident Grouch Join Date: Jun 2005 Posts: 6,453 Rep Power: 10	"x" is a pointer. "m" is the address of the first element of the array. "x" consequently points to "m". When you use the address of the array, "float []", the compiler knows the difference (one level of indirection) and generates code accordingly. It's the compiler's desire to make life easier for you that leads to confusion. We all know we can't assign a collection of objects to an array using the "=", but we all write, occassionally, "char myString [] = "abcdefg";" Again, the compiler is spoiling us. The compiler generates the array, element by element, as we'd have to do, if we did it at run-time. You can see why I call such things, "Confustoids." __________________ Abstraction doesn't make it impossible to write bad code; it makes it possible to write superior code. Contributor's Corner: Grumpy on C++ Exceptions DaWei on Pointers

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)

Thread Tools
Show Printable Version Email this Page
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Programming Forums > Application Development > Assembly

[x86] Arrays to functions