Managing big messy complex code

Question

So, I'm planning on doing some conversion of about 800K (compressed) of C source to C++.  It's already in a nice C-OO style of coding.  That is:
struct xyz { ... };
void xyz_move(struct xyz* pxyz, ...other args...)
void xyz_add(struct xyz* pxyz, ...other args...)
Yet it's also written K&R style which died in the 80s...
void
xyz_something(pxyz, a, b, c, d)
struct xyz *pxyz,
int a, b
char *c
int d
...body here
So the conversion isn't that hard.  The problem is that it's totally and wickedly unorganized... i.e. one file has 3 C "classes" listed, as well as the prototypes for 2 of the function sets.  The other function set is in a different header, and all 3 are in their own implementation files, along with other random functions prototyped in a totally different header.
Short story: it's a mess.  I'd like to convert it to a nice 1-to-1 header/implementation relationship, but there's alot of code.  I'm wondering if anyone knows of any tools (or even simple refactoring techniques) to aid me.  I can do it manually, it will just take me a long time.
Note: during this conversion I will not be switching the stuff over to C++ types (i.e. std::list<xyz*> instead of struct xyz**).  That will be done later, so in all honesty I only need C refactoring tools....

Answer 1

Hmmm, yeah I guess I could collect code and things, but what I'd like to do, is take a struct, along with related prototypes into one header file, and cram the implementations on the function in a related file.  Bascially I'm trying to save myself a few days of random... "ok this is a XYZ function, copy, paste in xyh.c and xyz.h, delete from original"
There should be refactoring tools out there to do something similar... maybe I'll check the vim scripts...