Skip to content

Demand Paging

What is demand paging

  • A page is brought into memory only when it is needed
  • Page is needed when there is a reference to the page
  • Pure demand paging
    • Start a process with no page
    • Never bring a page into memory until it is reqired.

Note

  • Basically lazy loading

How to do demand paging

  • Swapper manipulates the entire process, while a pager is concerned with the individual pages of a process.

Hardware support

  • Page Table : a valid-invalid bit

    • 1 means page in memory, 0 means not
    • All bits are set to 0 initially.

  • Secondary memory is used.

Page Fault

  • First reference to a page will trap to OS, called page-fault trap

  • OS will look at the internal talbe in PCB to see if it is a valid reference

  • Get an empty frame
  • Swap the page from disk into the frame
  • Reset page table setting the valid-invalid bit to 1
  • Restart the instruction (PC -= 4)

Demand Paging Performance

  • Effective Access Time
\[ \text{EAT}=(1-p)\times ma+p\times pft \]

where \(ma\) is the memory access time (from memory to CPU register), \(pft\) is page fault time, and \(p\) is page fault rate.

Programs tend to have locality of reference (unless you use linked list for every thing), since a single page fault brings in 4KB of memory content, the page fault rate is reduced greatly.

  • Page fault time is dominated by disk access time