is there any special approach to using (char *) 0 in C?

Question

I understand that (char *) 0 is a char pointer which points an address at 0. So there is no memory address, out of bound. Often, I hav seen people using it instead of NULL, so why don't they just simply use NULL? Or, I would like to know, whether there is any special approach to it.

Answer 1

A null pointer is strictly speaking no valid address encoding:

An integer constant expression with the value 0, or such an expression cast to type void *, is called a null pointer constant.66) If a null pointer constant is converted to a pointer type, the resulting pointer, called a null pointer, is guaranteed to compare unequal to a pointer to any object or function.

Thus I try to avoid the phrase "address 0" when it comes to null pointers, because this is the encoding and a null pointer need not have all bits zero. An architecture could - for instance - use an unused bit in the pointer variables to tag it as invalid. For comparision e.g. bitops would be used instead of compare.

The most common encoding is "all bits zero", though so I will mostly concentrate on this. This encoding is used because most CPUs have special (and shorter/faster) instructions to store a bit value of 0x0 into a memory cell and compare with/test for 0x0. Using a special bitpattern also makes comparing two pointers straight forward, using a single compare. The version with setting a dedicated bit would require additional instructions.

Also a value which is at the edge of the address range avoids gaps. This is true for "all bits 1", too, but with less CPU support and address-width dependency.

---

Note that most architectures very well do have a valid physical address 0, so there is in fact a problem with this encoding for a null pointer without address translation. Luckily this is typically a special address and not used by normal application code. With virtual memory addressing, it is quite easy just to make address 0x0 invalid, so this is out of discussion.

On some systems, address 0x0 is used for code (PC-boot code/BIOS, embedded systems) or data-accesses (RAM/peripheral hardware registers). Strictly speaking you cannot convert the integer-constant 0 to a pointer without invoking the implicit conversion to a null pointer. So, if your system does not use the 0x0 encoding, there is a problem accessing address 0x0 with standard C syntax only. (FYI: this is one reason other languages use a special keyword for a null pointer constant, e.g pascal NIL, Python None, C++ nullptr.)

The conversion of an integer to a pointer is implementation defined anyway, as is the null pointer encoding. Thus the conversion of a 0 to a pointer is well in-line with the common copy-only integer->pointer encoding without modification of the representation.

A problem arises when dereferencing such a pointer. Dereferencing a null pointer is undefined behaviour (UB). As both encodings are identical, there is no way for the CPU to detect if this is a valid access or a null pointer dereferencing. Even worse: the compiler is free to exploit this UB and generate arbitrary code (or none at all) if it detects something like char c = *((char *)0);.

For a true variable (i.e. not const qualified) pointer, this is normally not much of a problem, as that would require dynamic code analysis and in C no code for runtime-checks is generated by the compiler, but for the constant expression above, it could very well cause problems.

Answer 2

A null pointer is a special pointer value that is known not to point anywhere. What this means that no other valid pointer, to any other variable or array cell or anything else, will ever compare equal to a null pointer.

The most straightforward way to get a null pointer in your program is by using the predefined constant NULL, which is defined for you by several standard header files, including , , and . To initialize a pointer to a null pointer, you might use code like

#include <stdio.h>

int *ip = NULL;

and to test it for a null pointer before inspecting the value pointed to you might use code like

if(ip != NULL)
    printf("%d\n", *ip);

It is also possible to refer to the null pointer by using a constant 0, and you will see some code that sets null pointers by simply doing

char *ip = 0;

Answer 3

Short answer: Whether you say

char *p1 = NULL;

or

char *p2 = (char *)0;

or

char *p3 = 0;

they are all pretty much equivalent. They're all legal, they'll all work, they'll all initialize the pointer variables to null pointer values. The differences between them have to do with style, not correctness.

Whoever wrote the code you saw may just have had a different notion about what constitutes good style.

(Now, if you're dealing with pointers of type other than char *, using (char *)0 would be poor, and should at least generate warnings, and under suitably obscure circumstances might actually cause problems.)

Answer 4

NULL represent that the pointer isnt pointing to any address, NULL is a macro for (void *)0

so you can use (char *)0, it would still point to nothing and it'll be easier to explain the code