Page Faults (7+8)

Question 1

Page fault

Answer 1

when page for given process (to be mapped) to phys address but the given page is not currently loaded into a physical space.

Process page virtual address may be accessible to the process in the virtual address space but the memory page/mapping for this page has not yet been added to the process page tables

Thus page still needs to loaded from secondary storage

Question 2

What detects page faults and what handles the exceptions

Answer 2

MMU - detects

OS-kernel level applications do error handling

Question 3

What happens when the OS handles a page fault

Answer 3

OS tries to make page accessible by mapping to phys mem otherwise, if required page cannot be accessed - process is terminated and illegal mem error is thrown

Question 4

what is a trap (OS context)

Answer 4

exception or fault where OS temporarily switches to kernel mode, performs fault handling action before returning control back to originating process

Question 5

Examples of OS traps

Answer 5

break points, division by zero, invalid memory access

Question 6

Steps that OS takes to handle a page fault

Answer 6

1. When MMU references a page table - either page is already in phys mem (accessed via temp mem cache used by page table to access frequent pages) or not in mem e.g. MMU tried to do mapping but there is no accessible page - OS elicits a trap
2. Trap
3. if trap - missing page pulled from secondary storage and mapped
4. loaded into main mem address
   Once loaded OS exits kernel control (released from trap) and process continues

Question 7

Three different page faults - name them

Answer 7

compulsory
capacity
policy

Question 8

compulsory faults

Answer 8

most common
required pages not paged into mem
To avoid this - pre-fetching

Question 9

what is pre fetching

Answer 9

frequently used pages are loaded int cache memory

Question 10

capacity faults

Answer 10

page is scheduled but main mem is full of other pages. Another page must be swapped out

Question 11

policy faults

Answer 11

page prematurely swapped out because of replacement algorithm

Question 12

Pages that are high priority should be swapped into main mem before low priority pages? T/F

Answer 12

True

Question 13

What does the scheduling of page resources

Answer 13

page replacement policy

Question 14

what are the 3 core implementations in the MMU that page fault handling makes use of

Answer 14

mapping, page tables. page replacement policy

Question 15

What kind of mem mapping does MMU use?

Answer 15

Assume that the MMU uses fully associative memory mapping

Question 16

Assume that the MMU uses fully associative memory mapping - what does this mean

Answer 16

any page in VMem can be translated to any frame is phys mem - excluding those dedicated to kernel operations

Question 17

How does the MMU locate the page and load it to main mem

Answer 17

Done by using Page Table Base Register
Tells MMU whether page is in main mem or not
1. MMU looked in Translation Lookaside Buffer (not in TLB - flagged as TLB miss)
2. MMU checks page table to see if already in main mem
Else: loaded from secondary storage

Question 18

what is the TLB

Answer 18

Translation Lookaside Buffer

special mem cache for storing frequently used pages