Chapter 4

Entering 32-bit Protected Mode

It would be nice to continue working in the 16-bit real mode with which we have now
become much better aquainted, but in order to make fuller use of the CPU, and to
better understand how developments of CPU architecures can benefit modern operating
systems, namely memory protection in hardware, then we must press on into 32-bit
protected mode.
The main differences in 32-bit protected mode are:

• Registers are extended to 32 bits, with their full capacity being accessed by pre-
  fixing aneto the register name, for example:mov ebx, 0x274fe8fe


• For convenience, there are two additional general purpose segment registers,fs
  andgs.


• 32-bit memory offsets are available, so an offset can reference a whopping 4 GB
  of memory (0xffffffff).


• The CPU supports a more sophisticated --- though slightly more complex ---
  means of memory segmentation, which offers two big advantages:
  
  
  • Code in one segment can be prohibited from executing code in a more priv-
    ilidged segment, so you can protect your kernel code from user applications
  
  
  • The CPU can implementvirtual memoryfor user processes, such thatpages
    (i.e. fixed-sized chunks) of a process’s memory can be swapped transparently
    between the disk and memory on an as-needed basis. This ensure main
    memory is used efficiently, in that code or data that is rarely executed
    needn’t hog valuable memory.
  
  
  
  


• Interrupt handling is also more sophisticated.
  [?]
  The most difficult part about switching the CPU from 16-bit real mode into 32-bit
  protected mode is that we must prepare a complex data structure in memory called the
  global descriptor table(GDT), which defines memory segments and their protected-mode
  attributes. Once we have defined the GDT, we can use a special instruction to load it

30