need help to understand a piece of code

I am struck in understanding a piece of code.
It is the implementation of a function called UX_location.
the code for it is

/* Procedure: 0x000446C4 - 0x000446CF
* Argument size: 0
* Local size: 0
* Save regs size: 0
*/

UX_location()
{

r1 = 0x6c000;
asm("ba,a 0x00044514");
}



I am giving below a part of the code that calls this function

RUshmdetach()
{
/* unknown */ void Vfffffffc;



if(*RUshmsegatt != -1) {
if(UXsh_detach( *RUshmaddr) != -20) {
goto L0002478c;
}
UX_location("/apx/ecpc21/OFC18L21.0/1apx10/rdbu/gen/omp/RUitdetach.c", 39);
UX_elog("RUshmdetach() failure: RUshmaddr = %d, RUshmtblptr = %d", *RUshmaddr, *RUshmtblptr);
Vfffffffc = 0;
} else {
L0002478c:
*RUshmsegatt = -1;
Vfffffffc = 1;
}
r24 = Vfffffffc;
return(r3);
}



This is output of the REC .
I am trying to reverse a ELF 32-bit MSB executable SPARC Version 1, dynamically linked,not stripped code.


Any one can lead me into this?

I am giving below a part of the code that calls this function

RUshmdetach()
{
/* unknown */ void Vfffffffc;

I suspect you have, or this was written on, an old piece of crap compiler. Something like the one talked about here:

C: A reference Manual, Fifth Edition
Page 87, 4.4.1 Default Type Specifiers

Originally, C allowed the type specifier in a variable declaration or function to
be omitted, in which chase it defaulted to int. This is considered bad programming style in
modern C, and in fact C99 treats it as an error. Older compilers did not implement the
void type, so rationale behind omitting the type specifier on function definitions was to
indicate to human readers that the default function did not really return a value (although the com-
piler had to assume that it did).

Example

Code:

In pre-Standard C, it was common to see function definitions like this:
/* Sort v[0] ... vpn-1] into increasing order. */
sort(v, n)
    int v[], n;
{
    ...
}

The modern, Standard C style is to declare those functions with the void type:

Code:

/* Sort v[0] ... v[n-1] into increasing order. */
void sort(int v[], int n)
{
    ...
}

Example
When using a compiler that does not implement void, it is much nicer to define void your-
self and then use it explicitly than to omit the type specifier entirely:

Code:

/* Make "void" be a synonym for "int". */
typedef int void;

At least one compiler we know of actually reserves the identifier void, but does not imple-
ment it. For that compiler, the preprocessor definition

Code:

#define void int

is one of the few cases in which using a reserved word as a macro name is justified.

As per the above, I suspect the code, or you, whatever, was written for a compiler which did not implement void, or, that it was implemented as illustrated above.

Because otherwise, you cannot declare a void variable instance. In otherwords, this code isn't ANSI C. At least that's my take on it.

Quzah.

> I am trying to reverse a ELF 32-bit MSB executable SPARC Version 1
Yeah thought as much.
That code is so horrible it could only come from either a shrouder or a decompiler.

> Any one can lead me into this?
You're on your own I'm afraid.
You need "good to excellent" skills in both C and the assembly of the target processor to have any real chance of achieving what you want.

Thanks guys