Memory Segment

In the memory segment, available memory is divided into segments. The memory segment format looks like the following listing.

Listing: Sample MEMORY segment

MEMORY { 

segment_1 (RWX): ORIGIN = 0x8000, LENGTH = 0x1000 

segment_2 (RWX): ORIGIN = AFTER(segment_1), LENGTH = 0

segment_3 (RWX): ORIGIN = 0x4000, LENGTH = 0x1000, INITVAL = 0xABCD 

 data (RW) : ORIGIN = 0x2000, LENGTH = 0x0000 

 #segment_name (RW) : ORIGIN = memory address, LENGTH = segment 

 #length 

 #and so on... 

 }

The first memory segment definition ( segment_1) can be broken down as follows:

the ( RWX) portion of the segment definition pertains to the ELF access permission of the segment. The (RWX) flags imply r ead, w rite, and e x ecute access.
ORIGIN represents the start address of the memory segment (in this case 0x8000).
LENGTH represents the size of the memory segment (in this case 0x1000).
INITVAL represents the link-time initialization value to be used for watermarking a memory segment . For any ` INITVAL = (expression)' the `(expression)' is treated as a word value.
Example

Assume for above example that there is a program section of length 0xF00 words and it is placed in segment_ 3
```
      section{

      program_section_0x0F00;

     }>segment_ 3  
```
Then, the resulting memory map will have ( 0x1000-0xF00=0x100) words at the end of segment_3 that will be initialized with the pattern specified in INITVAL = 0xABCD .

Memory segments with RWX attributes are placed into P: memory while RW attributes are placed into X: memory.

If you cannot predict how much space a segment will occupy, you can use the function AFTER and LENGTH = 0 (unlimited length) to fill in the unknown values.