UC3M

Let us suppose that the memory address are represented internally by the processor using 32 bits (4 bytes). We can then store the address of d1 (the number 100) in a different memory location occupying 4 bytes because it is a memory address. The following figure shows this situation with the memory addresses stored in position 200.

The memory address of any variable in C may be obtained using the & operand followed by the a variable name. In the previous figure, the expression &d1 returns the value 100.

The value returned by the & operator depends on the position of its operand, and therefore is not under the control of the programmer. The memory address stored as a regular data in the address 200 is known generically as a “pointer” because its value “points to” the location of d1. Another way of saying the same is: address 200 contains a pointer to d1.

If position 200 contains a pointer to the variable d1, the fields of this data structure can be accessed through an “indirection”. The data stored in position 200 is taken and its value (the number 100) is now interpreted as a memory address. We access to that direction and in there we may access to the fields of d1. We just accessed to d1 indirectly, or through an “indirection”.

The indirection can be applied multiple times in the same access in what is known as “multiple indirection”. Following with the previous example, we may now store the address of the pointer (that is, the value 200) in another memory location, for example, position 300. In this position we now have the address of the address of d1. Or analogously, in position 300 we have a pointer to a pointer to d1. As in the previous case, we may access the fields in d1 but this time using a double indirection. It is possible to build multiple indirections with an arbitrary number of levels.