ITFN 2601 Introduction to Operating Systems

1
ITFN 2601Introduction to Operating Systems

• Lecture 6
• Deadlock
• Deadlock Recovery

2
Agenda

• What is a Resource?
• Deadlock Defined
• Conditions for Deadlock
• Resource Allocation Graphs
• Dealing with Deadlock

3
Resources

• An object granted
• Hardware (disk, cpu)
• Piece of Information (memory, record in DB)
• Preemptable non-preemptable

4
Utilizing a Resource

• Request resource
• Use Resource
• Release resource
• Use a semaphore/mutex to get/release lock on
  resource

5
Deadlock Defined

• From our childhood
• Ive got the ball and want the bat
• Youve got the bat and want the ball
• A set of processes is deadlocked if each process
  in the set is waiting for an event that only
  another process in the set can cause.

6
Conditions for Deadlock

• Mutual exclusion of resource
• Hold and Wait (processes with a resource can
  request other resources)
• Non-preemptable resources
• Circular wait (chain of processes in wait)
• See RAGs (next slide)

7
Resource Allocation Graphs

• Figure 3-3

D
A
S
T
U
B
R
C
Process Aholdingresource R
Process Brequestingresource S
Circularity Deadlock
8
Dealing with Deadlock

• Ostrich Algorithm (today)
• Detection Recovery (today)
• Dynamic Avoidance (next lecture)
• Prevention (next lecture)

9
Ostrich Algorithm

• Stick your head in the sand and pretend there is
  no problem at all
• What is the typical use of the system?
• What is the probability of deadlock?
• Do the costs associated with dealing w/ deadlock
  outweigh the benefits?

10
Detection Recovery

• One resource of each type
• Multiple resource of each type
• Preemption
• Rollback
• Killing Processes

11
Detection (1 resource)

• Detect cycle in RAG
• Use a directed graph
• Utilize an optimized cycle detection algorithm
  speed essential!

12
Detection (multiple resource)

• Forget about it

13
OK, Now What?

• Once weve detected deadlock
• Handle it
• Preemption
• Rollback
• Killing Processes

14
Preemption

• Not always possible
• Highly dependant upon the resource
• Typically associated w/ hardware resources
• Some examples
• Suspending a large print job
• CD-burner

15
Rollback

• Save the state of each process periodically
  (checkpoint)
• Restore the process w/ the needed resource to an
  earlier checkpoint (before it acquired the
  resource)
• Then the resource is now available
• Drawback rolled back process loses work

16
Killing Processes

• Again, not always an option
• DB transactions cannot always be killed midway
  through!
• Kill process to free up resource

17
Summary

• Resources can be preeptable nonpreemptable
• Deadlock can cause processes to halt (stop making
  progress)
• We can detect deadlock
• RAGs are quite useful
• Ostrich algorithm quite popular