C2244 when trying to call the pow function from inline assembly

Question

I would like to call the pow function from inline assembly. The problem is i'm getting error C2244: 'pow' : unable to match function definition to an existing declaration. I'm new to assembly so this may be a trivial question but how do i resolve this? I guess it has something to do with the compiler not beeing able to properly resolve the overload of pow. The following code fragment is causing the error:

do_POW:
  // push first argument to the stack
  sub   esp, size value_type
  fld   qword ptr [ecx]
  fstp  qword ptr [esp]

  // push second argument to the stack
  sub   esp, size value_type
  fld   qword ptr [ecx - size value_type]
  fstp  qword ptr [esp]

  // call the pow function
  call  pow
  sub   ecx, size value_type
  fstp  qword ptr [ecx]
  add   esp, 2 * size value_type
  jmp loop_start 

Answer 1

I found the solution. For anyone having similar problems: Create a pointer to the pow function in the C++ code and use this pointer in the inline assembly code:

double (*pow_func)(double, double) = pow;
__asm 
{
  call pow_func
}

Answer 2

It's not finding the symbol pow because the actual name of the library function is decorated in some way. Each platform has a standard for how the names get decorated, which may vary with what calling convention is used. Many (most?) platforms use a single underbar prefix as the usual decoration, so I would try:

call _pow

Alternatively, you can write a simple program like this:

#include <math.h>

double foo(double x) {
    return pow(1.5, x);
}

and look at the generated assembly code to see what the decoration convention is on your platform.