UNIT 3/LECTURE 1

 Processes

•          Process Concept

•          Process Scheduling

•          Operation on Processes

•          Cooperating Processes

•          Interprocess Communication

Process Concept

•          An operating system executes a variety of programs:

–         Batch system – jobs

–         Time-shared systems – user programs or tasks

•          Textbook uses the terms job and process almost interchangeably.

•          Process – a program in execution; process execution must progress in sequential fashion.

•          A process includes:

–         program counter

–         stack

–         data section

Process State

•          As a process executes, it changes state

–         new:  The process is being created.

–         running:  Instructions are being executed.

–         waiting:  The process is waiting for some event to occur.

–         ready:  The process is waiting to be assigned to a process.

–         terminated:  The process has finished execution.

Diagram of Process State

Process Control Block (PCB)

Information associated with each process.

•          Process state

•          Program counter

•          CPU registers

•          CPU scheduling information

•          Memory-management information

•          Accounting information

•          I/O status information

Process Control Block (PCB)

CPU Switch From Process to Process

 Process Scheduling Queues

•          Job queue – set of all processes in the system.

•          Ready queue – set of all processes residing in main memory,
ready and waiting to execute.

•          Device queues – set of processes waiting for an I/O device.

•          Process migration between the various queues.

Ready Queue And Various I/O Device Queues

UNIT 3/LECTURE 2

Representation of Process Scheduling

Schedulers

•          Long-term scheduler (or job scheduler) – selects which processes should be brought into the ready queue.

•          Short-term scheduler (or CPU scheduler) – selects which process should be executed next and allocates CPU.

Addition of Medium Term Scheduling

•          Short-term scheduler is invoked very frequently (milliseconds)Þ (must be fast).

•          Long-term scheduler is invoked very infrequently (seconds, minutes) Þ (may be slow).

•          The long-term scheduler controls the degree of multiprogramming.

•          Processes can be described as either:

–         I/O-bound process – spends more time doing I/O than computations, many short CPU bursts.

–         CPU-bound process – spends more time doing computations; few very long CPU bursts.

Context Switch

•          When CPU switches to another process, the system must save the state of the old process and load the saved state for the new process.

•          Context-switch time is overhead; the system does no useful work while switching.

•          Time dependent on hardware support.

Process Creation

•          Parent process creates children processes, which, in turn create other processes, forming a tree of processes.

•          Resource sharing

–         Parent and children share all resources.

–         Children share subset of parent’s resources.

–         Parent and child share no resources.

•          Execution

–         Parent and children execute concurrently.

–         Parent waits until children terminate.

•          Address space

–         Child duplicate of parent.

–         Child has a program loaded into it.

•          UNIX examples

–         fork system call creates new process

–         execve system call used after a fork to replace the process’ memory space with a new program.

Process Termination

•          Process executes last statement and asks the operating system to decide it (exit).

–         Output data from child to parent (via wait).

–         Process’ resources are deallocated by operating system.

•          Parent may terminate execution of children processes (abort).

–         Child has exceeded allocated resources.

–         Task assigned to child is no longer required.

–         Parent is exiting.

T        Operating system does not allow child to continue if its parent terminates.

T        Cascading termination.

Cooperating Processes

•          Independent process cannot affect or be affected by the execution of another process.

•          Cooperating process can affect or be affected by the execution of another process

•          Advantages of process cooperation

–         Information sharing

–         Computation speed-up

–         Modularity

–         Convenience

Inter Process Communication

                                                           Race Condition

UNIT 3/LECTURE 3

Semaphore

·         Synchronization tool that does not require busy waiting

·         Semaphore S – integer variable

·         Two standard operations modify S: wait() and signal()

1.       Originally called P() and V()

·         Less complicated

·         Can only be accessed via two indivisible (atomic) operations

wait (S) {

                                    while S <= 0 ; // no-op

                                     S--;

                                }

signal (S) {

                                        S++;

                                  }

Semaphore as General Synchronization Tool

·         Counting semaphore – integer value can range over an unrestricted domain

·         Binary semaphore – integer value can range only between 0 and 1; can be simpler to implement

l          Also known as mutex locks

·         Can implement a counting semaphore S as a binary semaphore

·         Provides mutual exclusion

Semaphore S;    //  initialized to 1

         wait (S);

                              Critical Section

                            signal (S);

Semaphore Implementation

·         Must guarantee that no two processes can execute wait () and signal () on the same semaphore at the same time

·         Thus, implementation becomes the critical section problem where the wait and signal code are placed in the crtical section.

1.       Could now have busy waiting in critical section implementation

4  But implementation code is short

4  Little busy waiting if critical section rarely occupied

·         Note that applications may spend lots of time in critical sections and therefore this is not a good solution.

Semaphore Implementation with no Busy waiting

·         With each semaphore there is an associated waiting queue. Each entry in a waiting queue has two data items:

1.  value (of type integer)
2.  pointer to next record in the list

·         Two operations:

1.  block – place the process invoking the operation on the      appropriate waiting queue.

                    2.  wakeup – remove one of processes in the waiting queue and place it in the ready queue.

·         Implementation of wait:

                        wait (S){

                                      value--;

                                          if (value < 0) {

                                                  add this process to waiting queue

                                                   block();  }

                                      }

·         Implementation of signal:

                        Signal (S){

                                         value++;

                                                 if (value <= 0) {

                                                       remove a process P from the waiting queue

                                                      wakeup(P);  }

                                          }

–                 

–                 

Deadlock and Starvation

·         Deadlock – two or more processes are waiting indefinitely for an event that can be caused by only one of the waiting processes

·         Let S and Q be two semaphores initialized to 1

                                 P₀                                                    P₁

                            wait (S);                                             wait (Q);

                            wait (Q);                                            wait (S);

                        .                       .

                        .                       .

                        .                       .

                             signal  (S);                                           signal (Q);

                             signal (Q);                                           signal (S);

·         Starvation  – indefinite blocking.  A process may never be removed from the semaphore queue in which it is suspended.

\

UNIT 3/LECTURE 4

–          

UNIT 3/LECTURE 5

Sockets

·         A socket is defined as an endpoint for communication

·         Concatenation of IP address and port

·         The socket 161.25.19.8:1625 refers to port 1625 on host 161.25.19.8

·         Communication consists between a pair of sockets

Socket Communication

Remote Procedure Calls

·         Remote procedure call (RPC) abstracts procedure calls between processes on networked systems.

·         Stubs – client-side proxy for the actual procedure on the server.

·         The client-side stub locates the server and marshalls the parameters.

·         The server-side stub receives this message, unpacks the marshalled parameters, and peforms the procedure on the server.

Execution of RPC

Remote Method Invocation

·         Remote Method Invocation (RMI) is a Java mechanism similar to RPCs.

·         RMI allows a Java program on one machine to invoke a method on a remote object.

UNIT 3/LECTURE 6

CPU Scheduling

·         Basic Concepts

·         Scheduling Criteria

·         Scheduling Algorithms

·         Multiple-Processor Scheduling

·         Real-Time Scheduling

·         Thread Scheduling

·         Operating Systems Examples

·         Java Thread Scheduling

·         Algorithm Evaluation

Basic Concepts

·         Maximum CPU utilization obtained with multiprogramming

·         CPU–I/O Burst Cycle – Process execution consists of a cycle of CPU execution and I/O wait

·         CPU burst distribution

Alternating Sequence of CPU And I/O Bursts

CPU Scheduler

·         Selects from among the processes in memory that are ready to execute, and allocates the CPU to one of them

·         CPU scheduling decisions may take place when a process:

                  1. Switches from running to waiting state

                  2.Switches from running to ready state

                  3.Switches from waiting to ready

                  4.Terminates

·         Scheduling under 1 and 4 is nonpreemptive

·         All other scheduling is preemptive

Dispatcher

·         Dispatcher module gives control of the CPU to the process selected by the short-term scheduler; this involves:

1.       switching context

2.       switching to user mode

3.       jumping to the proper location in the user program to restart that program

·         Dispatch latency – time it takes for the dispatcher to stop one process and start another running

Scheduling Criteria

·         CPU utilization – keep the CPU as busy as possible

·         Throughput – # of processes that complete their execution per time unit

·         Turnaround time – amount of time to execute a particular process

·         Waiting time – amount of time a process has been waiting in the ready queue

·         Response time – amount of time it takes from when a request was submitted until the first response is produced, not output  (for time-sharing environment)

Optimization Criteria

·         Max CPU utilization

·         Max throughput

·         Min turnaround time

·         Min waiting time

·         Min response time

First-Come, First-Served (FCFS) Scheduling

                      Process          Burst Time           

                         P₁                     24

                         P₂                      3

                         P₃                                      3

Suppose that the processes arrive in the order: P₁ , P₂ , P_{3 .}The Gantt Chart for the schedule is:

·         Waiting time for P₁  = 0; P₂  = 24; P₃ = 27

·         Average waiting time:  (0 + 24 + 27)/3 = 17

Suppose that the processes arrive in the order

                         P₂ , P₃ , P₁

The Gantt chart for the schedule is:

·         Waiting time for P₁ = 6; P₂ = 0_; P₃ = 3

·         Average waiting time:   (6 + 0 + 3)/3 = 3

·         Much better than previous case

·         Convoy effect short process behind long process

Shortest-Job-First (SJF) Scheduling

·         Associate with each process the length of its next CPU burst.  Use these lengths to schedule the process with the shortest time

·         Two schemes:

1.       nonpreemptive – once CPU given to the process it cannot be preempted until completes its CPU burst

2.       preemptive – if a new process arrives with CPU burst length less than remaining time of current executing process, preempt.  This scheme is know as the
Shortest-Remaining-Time-First (SRTF)

·         SJF is optimal – gives minimum average waiting time for a given set of processes

Example of Non-Preemptive SJF

                       Process         Arrival Time          Burst Time

                        P₁                     0.0                              7

                         P₂                                  2.0                              4

                         P₃                    4.0                              1

                         P₄                    5.0                              4

·         SJF (non-preemptive)

·         Average waiting time = (0 + 6 + 3 + 7)/4  = 4

Example of Preemptive SJF

                     Process      Arrival Time          Burst Time

                        P₁                   0.0                           7

                         P₂                               2.0                           4

                         P₃                  4.0                           1

                         P₄                  5.0                           4

·         SJF (preemptive)

·         Average waiting time = (9 + 1 + 0 +2)/4 = 3

Determining Length of Next CPU Burst

·         Can only estimate the length

·         Can be done by using the length of previous CPU bursts, using exponential averaging

Examples of Exponential Averaging

·         a =0     then  t_n+1 = t_n    --    Recent history does not count

·         a =1    then   t_n+1 = a t_n       --   Only the actual last CPU burst counts

·         If we expand the formula, we get:

                                  t_n+1 = a t_n+(1 - a)a t_n -1 + …

                                               +(1 - a )^j a t_{n -j} + …

                                               +(1 - a )^{n +1} t₀

·         Since both a and (1 - a) are less than or equal to 1, each successive term has less weight than its predecessor

Priority Scheduling

·         A priority number (integer) is associated with each process

·         The CPU is allocated to the process with the highest priority (smallest integer º highest priority)

1.       Preemptive

2.       nonpreemptive

·         SJF is a priority scheduling where priority is the predicted next CPU burst time

·         Problem º Starvation – low priority processes may never execute

·         Solution º Aging – as time progresses increase the priority of the process

UNIT 3/LECTURE 7

Round Robin (RR)

·         Each process gets a small unit of CPU time (time quantum), usually 10-100 milliseconds.  After this time has elapsed, the process is preempted and added to the end of the ready queue.

·         If there are n processes in the ready queue and the time quantum is q, then each process gets 1/n of the CPU time in chunks of at most q time units at once.  No process waits more than (n-1)q time units.

·         Performance

                       1.  q large Þ FIFO

                       2.  q small Þ q must be large with respect to context switch, otherwise overhead is too high

Example of RR with Time Quantum = 20

                             Process                Burst Time

                                  P₁                                             53

                                  P₂                                            17

                                  P₃                                            68

                                  P₄                                            24

The Gantt chart is:

·         Typically, higher average turnaround than SJF, but better response

Time Quantum and Context Switch Time

Multilevel Queue

·         Ready queue is partitioned into separate queues:
           foreground (interactive)
           background (batch)

·         Each queue has its own scheduling algorithm

1.       foreground – RR

2.       background – FCFS

·         Scheduling must be done between the queues

            1. Fixed priority scheduling; (i.e., serve all from foreground then from background).   

                 Possibility of starvation.

                    2. Time slice – each queue gets a certain amount of CPU time which it can schedule amongst

                       its processes; i.e., 80% to foreground in RR

3.       20% to background in FCFS

Multilevel Feedback Queue

·         A process can move between the various queues; aging can be implemented this way

·         Multilevel-feedback-queue scheduler defined by the following parameters:

1.       number of queues

2.       scheduling algorithms for each queue

3.       method used to determine when to upgrade a process

4.       method used to determine when to demote a process

5.       method used to determine which queue a process will enter when that process needs service

Example of Multilevel Feedback Queue

·         Three queues:

1.       Q₀ – RR with time quantum 8 milliseconds

2.       Q₁ – RR time quantum 16 milliseconds

3.       Q₂ – FCFS

·         Scheduling

1.       A new job enters queue Q₀ which is served FCFS. When it gains CPU, job receives 8 milliseconds.  If it does not finish in 8 milliseconds, job is moved to queue Q₁.

2.       At Q₁ job is again served FCFS and receives 16 additional milliseconds.  If it still does not complete, it is preempted and moved to queue Q₂.

Multiple-Processor Scheduling

·         CPU scheduling more complex when multiple CPUs are available

·         Homogeneous processors within a multiprocessor

·         Load sharing

·         Asymmetric multiprocessing – only one processor accesses the system data structures, alleviating the need for data sharing

Real-Time Scheduling

·         Hard real-time systems – required to complete a critical task within a guaranteed amount of time

·         Soft real-time computing – requires that critical processes receive priority over less fortunate ones

                                                 Figure:-  Dispatch Latency

UNIT 3/LECTURE 8

Deadlocks

·         The Deadlock Problem

·         System Model

·         Deadlock Characterization

·         Methods for Handling Deadlocks

·         Deadlock Prevention

·         Deadlock Avoidance

·         Deadlock Detection

·         Recovery from Deadlock

The Deadlock Problem

·         A set of blocked processes each holding a resource and waiting to acquire a resource held by another process in the set.

·         Example

1.       System has 2 disk drives.

2.       P₁ and P₂ each hold one disk drive and each needs another one.

·         Example

3.       semaphores A and B, initialized to 1

                                                     P₀                                P₁

                                                  wait (A);                    wait(B)

                                              wait (B);                        wait(A)

Bridge Crossing Example

·         Traffic only in one direction.

·         Each section of a bridge can be viewed as a resource.

·         If a deadlock occurs, it can be resolved if one car backs up (preempt resources and rollback).

·         Several cars may have to be backed up if a deadlock occurs.

·         Starvation is possible.

System Model

·         Resource types R₁, R₂, . . ., R_m

                                   CPU cycles, memory space, I/O devices

·         Each resource type R_i has W_i instances.

·         Each process utilizes a resource as follows:

1.       request

2.       use

3.       release

Deadlock can arise if four conditions hold simultaneously.

·         Mutual exclusion:  only one process at a time can use a resource.

·         Hold and wait:  a process holding at least one resource is waiting to acquire additional resources held by other processes.

·         No preemption:  a resource can be released only voluntarily by the process holding it, after that process has completed its task.

·         Circular wait:  there exists a set {P₀, P₁, …, P₀} of waiting processes such that P₀ is waiting for a resource that is held by P₁, P₁ is waiting for a resource that is held by    P₂, …, P_n–1 is waiting for a resource that is held by P_n, and P₀ is waiting for a resource that is held by P₀.

Resource-Allocation Graph

A set of vertices V and a set of edges E.

·         V is partitioned into two types:

1.       P = {P₁, P₂, …, P_n}, the set consisting of all the processes in the system.

2.       R = {R₁, R₂, …, R_m}, the set consisting of all resource types in the system.

·         request edge – directed edge P₁ ® R_j

·         assignment edge – directed edge R_j ® P_i

4  Process

4  Resource Type with 4 instances

4  P_i requests instance of R_j

4  P_i is holding an instance of R_j

Example of a Resource Allocation Graph

Resource Allocation Graph With A Deadlock

Graph With A Cycle But No Deadlock

Basic Facts

4  If graph contains no cycles Þ no deadlock.

4  If graph contains a cycle Þ

1.       if only one instance per resource type, then deadlock.

2.       if several instances per resource type, possibility of deadlock.

Methods for Handling Deadlocks

4  Ensure that the system will never enter a deadlock state.

4  Allow the system to enter a deadlock state and then recover.

4  Ignore the problem and pretend that deadlocks never occur in the system; used by most operating systems, including UNIX.

Deadlock Prevention

Restrain the ways request can be made

4  Mutual Exclusion – not required for sharable resources; must hold for nonsharable resources.

4  Hold and Wait – must guarantee that whenever a process requests a resource, it does not hold any other resources.

1.       Require process to request and be allocated all its resources before it begins execution, or allow process to request resources only when the process has none.

2.       Low resource utilization; starvation possible.

4  No Preemption –

1.       If a process that is holding some resources requests another resource that cannot be immediately allocated to it, then all resources currently being held are released.

2.       Preempted resources are added to the list of resources for which the process is waiting.

3.       Process will be restarted only when it can regain its old resources, as well as the new ones that it is requesting.

4  Circular Wait – impose a total ordering of all resource types, and require that each process requests resources in an increasing order of enumeration.

UNIT 3/LECTURE 9

Deadlock Avoidance

Requires that the system has some additional a priori information available.

4  Simplest and most useful model requires that each process declare the maximum number of resources of each type that it may need.

4  The deadlock-avoidance algorithm dynamically examines the resource-allocation state to ensure that there can never be a circular-wait condition.

4  Resource-allocation state is defined by the number of available and allocated resources, and the maximum demands of the processes.

Safe State

4  When a process requests an available resource, system must decide if immediate allocation leaves the system in a safe state.

4  System is in safe state if there exists a sequence <P₁, P₂, …, P_n> of ALL the  processes  is the systems such that  for each P_i, the resources that P_i can still request can be satisfied by currently available resources + resources held by all the P_j, with j < i.

4  That is:

1.       If P_i resource needs are not immediately available, then P_i can wait until all P_j have finished.

2.       When P_j is finished, P_i can obtain needed resources, execute, return allocated resources, and terminate.

3.       When P_i terminates, P_{i +1} can obtain its needed resources, and so on.

Basic Facts

4  If a system is in safe state Þ no deadlocks.

4  If a system is in unsafe state Þ possibility of deadlock.

4  Avoidance Þ ensure that a system will never enter an unsafe state.

Safe, Unsafe , Deadlock State

Avoidance algorithms

4  Single instance of a resource type.  Use a resource-allocation graph

4  Multiple instances of a resource type.  Use the banker’s algorithm

Resource-Allocation Graph Scheme

4  Claim edge P_i ® R_j indicated that process P_j may request resource R_j; represented by a dashed line.

4  Claim edge converts to request edge when a process requests a resource.

4  Request edge converted to an assignment edge when the  resource is allocated to the process.

4  When a resource is released by a process, assignment edge reconverts to a claim edge.

4  Resources must be claimed a priori in the system.

Resource-Allocation Graph

Unsafe State In Resource-Allocation Graph

Resource-Allocation Graph Algorithm

4  Suppose that process P_i requests a resource R_j

4  The request can be granted only if converting the request edge to an assignment edge does not result in the formation of a cycle in the resource allocation graph

Banker’s Algorithm

4  Multiple instances.

4  Each process must a priori claim maximum use.

4  When a process requests a resource it may have to wait. 

4  When a process gets all its resources it must return them in a finite amount of time.

Data Structures for the Banker’s Algorithm

Let n = number of processes, and m = number of resources types

4  Available:  Vector of length m. If available [j] = k, there are k instances of resource type R_j  available.

4  Max: n x m matrix.  If Max [i,j] = k, then process P_i may request at most k instances of resource type R_j.

4  Allocation:  n x m matrix.  If Allocation[i,j] = k then P_i is currently allocated k instances of R_j.

4  Need:  n x m matrix. If Need[i,j] = k, then P_i may need k more instances of R_j to complete its task.

                         Need [i,j] = Max[i,j] – Allocation [i,j].

Safety Algorithm

1.       Let Work and Finish be vectors of length m and n, respectively.  Initialize:

              Work = Available

              Finish [i] = false for i = 0, 1, …, n- 1.

      2.   Find and i such that both:

              (a) Finish [i] = false

              (b) Need_i £ Work

                    If no such i exists, go to step 4.

       3.   Work = Work + Allocation_i
             Finish[i] = true
             go to step 2.

      4.  If Finish [i] == true for all i, then the system is in a safe state.

Resource-Request Algorithm for Process P_i

Request = request vector for process P_i.  If Request_i [j] = k then process P_i wants k instances of resource type R_j

1.   If Request_i £ Need_i go to step 2.  Otherwise, raise error condition, since process has exceeded its 

     maximum claim.

         2. If Request_i 1£ Available, go to step 3.  Otherwise Pi  must wait, since resources are not available.

         3. Pretend to allocate requested resources to P_i by modifying the state as follows:

                        Available = Available  – Request;

                        Allocation_i = Allocation_i + Request_i;

                        Need_i = Need_i – Request_i;

a.       If safe Þ the resources are allocated to Pi.

b.       If unsafe Þ Pi must wait, and the old resource-allocation state is restored

Example of Banker’s Algorithm

4  5 processes P₀  through P₄;

3 resource types:

                      A (10 instances),  B (5instances), and C (7 instances).

4  Snapshot at time T₀:

                                    Allocation        Max     Available

                                       A B C             A B C         A B C

                             P₀                  0 1 0           7 5 3          3 3 2

                             P₁                  2 0 0           3 2 2 

                             P₂         3 0 2            9 0 2

                             P₃         2 1 1            2 2 2

                       P₄           0 0 2          4 3 3 

4  The content of the matrix Need is defined to be Max – Allocation.

                                     Need

                                    A B C

                         P₀               7 4 3

                         P₁               1 2 2

                         P₂        6 0 0

                         P₃        0 1 1

                         P₄        4 3 1

4  The system is in a safe state since the sequence < P₁, P₃, P₄, P₂, P₀> satisfies safety criteria.

Example:  P₁ Request (1,0,2)

4  Check that Request £ Available (that is, (1,0,2) £ (3,3,2) Þ true.

4   

                                    Allocation        Need    Available

                                    A B C   A B C   A B C

                        P₀         0 1 0     7 4 3     2 3 0

                        P₁         3 0 2     0 2 0    

                        P₂         3 0 1     6 0 0

                        P₃         2 1 1     0 1 1

                        P₄         0 0 2     4 3 1

4  Executing safety algorithm shows that sequence < P₁, P₃, P₄, P₀, P₂> satisfies safety requirement.

4  Can request for (3,3,0) by P₄ be granted?

4  Can request for (0,2,0) by P₀ be granted?

UNIT 3/LECTURE 10

Deadlock Detection

4  Allow system to enter deadlock state

4  Detection algorithm

4  Recovery scheme

Single Instance of Each Resource Type

4  Maintain wait-for graph

a.       Nodes are processes.

b.       P_i ® P_j   if P_i is waiting for P_j.

4  Periodically invoke an algorithm that searches for a cycle in the graph. If there is a cycle, there exists a deadlock.

4  An algorithm to detect a cycle in a graph requires an order of n² operations, where n is the number of vertices in the graph.

Resource-Allocation Graph and Wait-for Graph

                          Resource-Allocation Graph                   Corresponding wait-for graph

Several Instances of a Resource Type

4  Available:  A vector of length m indicates the number of available resources of each type.

4  Allocation:  An n x m matrix defines the number of resources of each type currently allocated to each process.

4  Request:  An n x m matrix indicates the current request  of each process.  If Request [i_j] = k, then process P_i is requesting k more instances of resource type. R_j.

Detection Algorithm

         1. Let Work and Finish be vectors of length m and n, respectively Initialize:

                  (a) Work = Available

                    (b) For i = 1,2, …, n, if Allocation_i ¹ 0, then
                          Finish[i] = false;otherwise, Finish[i] = true.

         2. Find an index i such that both:

                  (a) Finish[i] == false

                 (b) Request_i £ Work

                    If no such i exists, go to step 4.

           3. Work = Work + Allocation_i
               Finish[i] = true
               go to step 2.

           4. If Finish[i] == false, for some i, 1 £ i £  n, then the system is in deadlock state.

               Moreover, if Finish[i] == false, then P_i is deadlocked.

Example of Detection Algorithm

4  Five processes P₀ through P₄; three resource types
A (7 instances), B (2 instances), and C (6 instances).

4  Snapshot at time T₀:

                                     Allocation            Request                     Available

                                        A B C                  A B C                         A B C

                        P₀              0 1 0                  0 0 0                          0 0 0

                        P₁              2 0 0                  2 0 2

                        P₂               3 0 3                 0 0 0

                        P₃              2 1 1                 1 0 0

                        P₄              0 0 2                  0 0 2

4  Sequence <P₀, P₂, P₃, P₁, P₄> will result in Finish[i] = true for all i.

4  P₂ requests an additional instance of type C.

                                    Request

                                    A B C

                         P₀        0 0 0

                         P₁        2 0 1

                        P₂         0 0 1

                        P₃         1 0 0

                        P₄         0 0 2

4  State of system?

1.       Can reclaim resources held by process P₀, but insufficient resources to fulfill other processes; requests.

2.       Deadlock exists, consisting of processes P₁,  P₂, P₃, and P₄.

Detection-Algorithm Usage

·         When, and how often, to invoke depends on:

1.       How often a deadlock is likely to occur?

2.       How many processes will need to be rolled back?

4  one for each disjoint cycle

·         If detection algorithm is invoked arbitrarily, there may be many cycles in the resource graph and so we would not be able to tell which of the many deadlocked processes “caused” the deadlock.

Recovery from Deadlock:  Process Termination

4  Abort all deadlocked processes.

4  Abort one process at a time until the deadlock cycle is eliminated.

4  In which order should we choose to abort?

·         Priority of the process.

·         How long process has computed, and how much longer to completion.

·         Resources the process has used.

·         Resources process needs to complete.

·         How many processes will need to be terminated.

·         Is process interactive or batch?

Recovery from Deadlock: Resource Preemption

4  Selecting a victim – minimize cost.

4  Rollback – return to some safe state, restart process for that state.

4  Starvation –  same process may always be picked as victim, include number of rollback in cost factor.