chapter 7 - deadlocks - sduimada.sdu.dk/~jamik/dm510-16/material/chapter7.pdf · chapter 7 -...

1

CHAPTER 7 -DEADLOCKS

2

OBJECTIVESTo develop a description of deadlocks, which prevent setsof concurrent processes from completing their tasks

To present a number of different methods for preventingor avoiding deadlocks in a computer system

3 . 1

SYSTEM MODEL

SYSTEM MODELSystem consists of resources

Resource types R1, R2, … , Rm

CPU cycles, memory space, I/O devices

Each resource type Ri has Wi instances.

Each process utilizes a resource as follows:

request

use

release

3 . 24 . 1

DEADLOCK CHARACTERIZATION

DEADLOCK CHARACTERIZATIONDeadlock can arise if four conditions hold simultaneously.

Mutual exclusion

Hold and wait

No preemption

Circular wait

4 . 24 . 3

MUTUAL EXCLUSIONOnly one process at a time can use a resource

4 . 4

HOLD AND WAITA process holding at least one resource is waiting to acquire

additional resources held by other processes

4 . 5

NO PREEMPTIONA resource can be released only voluntarily by the process

holding it, a�er that process has completed its task

4 . 6

CIRCULAR WAITThere exists a set {P0 , P1 , … , Pn } of waiting processes such

that P0 is waiting for a resource that is held by P1 , P1 iswaiting for a resource that is held by P2 , … , Pn–1 is waiting

for a resource that is held by Pn, and Pn is waiting for aresource that is held by P0.

DEADLOCK WITH MUTEX LOCKS/ * t h r e a d o n e r u n s i n t h i s f u n c t i o n * / v o i d * d o _ w o r k _ o n e ( v o i d * p a r a m ) { p t h r e a d _ m u t e x _ l o c k ( & f i r s t m u t e x ) ; p t h r e a d _ m u t e x _ l o c k ( & s e c o n d m u t e x ) ; / * * * D o s o m e w o r k * / p t h r e a d _ m u t e x _ u n l o c k ( & s e c o n d m u t e x ) ; p t h r e a d _ m u t e x _ u n l o c k ( & f i r s t m u t e x ) ; p t h r e a d _ e x i t ( 0 ) ; } / * t h r e a d t w o r u n s i n t h i s f u n c t i o n * / v o i d * d o _ w o r k _ t w o ( v o i d * p a r a m ) { p t h r e a d _ m u t e x _ l o c k ( & s e c o n d m u t e x ) ; p t h r e a d _ m u t e x _ l o c k ( & f i r s t m u t e x ) ; / * * * D o s o m e w o r k * / p t h r e a d _ m u t e x _ u n l o c k ( & f i r s t m u t e x ) ; p t h r e a d _ m u t e x _ u n l o c k ( & s e c o n d m u t e x ) ; p t h r e a d _ e x i t ( 0 ) ; }

4 . 7

RESOURCE-ALLOCATION GRAPHA set of vertices V and a set of edges E.

V is partitioned into two types:

P = {P1 , P2 , … , Pn }, the set consisting of all theprocesses in the system

R = {R1 , R2 , … , Rn }, the set consisting of all resourcetypes in the system

request edge – directed edge Pi → Rj

assignment edge – directed edge Rj → Pi

4 . 84 . 9

LEGEND

Process

LEGEND

Resource Type with 3 instances

4 . 10

LEGEND

P2 requests instance of R

4 . 11

LEGEND

P2 is holding an instance of R

4 . 124 . 13

EXAMPLE RESOURCESP = { P1, P2, P3 }

R = { R1, R2, R3, R4 }

E = { P1 → R1, P2 → R3, R1 → P2, R2 → P2, R2 → p1, R3 → P3 }

RESOURCE ALLOCATION GRAPH

4 . 14

GRAPH WITH A DEADLOCK

4 . 15

WITH A CYCLE BUT NO DEADLOCK

4 . 16

BASIC FACTSIf graph contains no cycles → no deadlock

If graph contains a cycle →

if only one instance per resource type, then deadlock

if several instances per resource type, possibility ofdeadlock

4 . 175 . 1

METHODS FOR HANDLINGDEADLOCKS

METHODS FOR HANDLINGDEADLOCKS

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

2. Allow the system to enter a deadlock state and thenrecover

3. Ignore the problem and pretend that deadlocks neveroccur in the system; used by most operating systems

5 . 26 . 1

DEADLOCK PREVENTION

6 . 2

DEADLOCK PREVENTIONRestrain the ways request can be made

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

DEADLOCK PREVENTIONHold and Wait – multiple ways

1. must guarantee that whenever a process requests aresource, it does not hold any other resources.

2. Request all resources up front before execution. (all ornone)

6 . 3

DEADLOCK PREVENTIONNo Preemption

If a process that is holding some resources requests anotherresource that cannot be immediately allocated to it, then all

resources currently being held are released

Preempted resources are added to the list of resources forwhich the process is waiting

Process will be restarted only when it can regain its oldresources, as well as the new ones that it is requesting

6 . 46 . 5

DEADLOCK PREVENTIONCircular Wait

impose a total ordering of all resource types, and requirethat each process requests resources in an increasing order

of enumeration

DEADLOCK EXAMPLE WITH LOCKORDERING

v o i d t r a n s a c t i o n ( A c c o u n t f r o m , A c c o u n t t o , d o u b l e a m o u n t ) { m u t e x l o c k 1 , l o c k 2 ; l o c k 1 = g e t _ l o c k ( f r o m ) ; l o c k 2 = g e t _ l o c k ( t o ) ; a c q u i r e ( l o c k 1 ) ; a c q u i r e ( l o c k 2 ) ; w i t h d r a w ( f r o m , a m o u n t ) ; d e p o s i t ( t o , a m o u n t ) ; r e l e a s e ( l o c k 2 ) ; r e l e a s e ( l o c k 1 ) ; }

6 . 67 . 1

DEADLOCK AVOIDANCE

DEADLOCK AVOIDANCERequires that the system has some additional a priori

information available

Simplest and most useful model requires that each processdeclare the maximum number of resources of each type that

it may need

7 . 2

DEADLOCK AVOIDANCEThe deadlock-avoidance algorithm dynamically examines

the resource-allocation state to ensure that there can neverbe a circular-wait condition

Resource-allocation state is defined by the number ofavailable and allocated resources, and the maximum

demands of the processes

7 . 3

SAFE STATEWhen a process requests an available resource, system must

decide if immediate allocation leaves the system in a safestate

System is in safe state if there exists a safe sequence

<P1, P2, … , Pn>

of ALL the processes in the systems such that for each P~i ,the resources that Pi can still request can be satisfied by

currently available resources + resources held by all the Pj ,with j < i

7 . 4

SAFE STATEThat is:

If Pi resource needs are not immediately available, then Pican wait until all Pj have finished

When Pj is finished, Pi can obtain needed resources,execute, return allocated resources, and terminate

When Pi terminates, Pi+1 can obtain its needed resources,and so on

7 . 57 . 6

BASIC FACTSIf a system is in safe state → no deadlocks

If a system is in unsafe state → possibility of deadlock

Avoidance → ensure that a system will never enter anunsafe state.

SAFE, UNSAFE, DEADLOCK STATE

7 . 77 . 8

AVOIDANCE ALGORITHMSSingle instance of a resource type → Use a resource-

allocation graph

Multiple instances of a resource type → Use the banker’salgorithm

SCHEMEClaim edge Pi → Rj indicated that process Pi may requestresource Rj ; represented by a dashed line

Claim edge converts to request edge when a processrequests a resource

Request edge converted to an assignment edge when theresource is allocated to the process

When a resource is released by a process, assignmentedge reconverts to a claim edge

Resources must be claimed a priori in the system

7 . 9

RESOURCE-ALLOCATION GRAPH

7 . 10

UNSAFE STATE IN RESOURCE-ALLOCATION GRAPH

7 . 11

RESOURCE-ALLOCATION GRAPHALGORITHM

Suppose that process Pi requests a resource Rj

The request can be granted only if converting the requestedge to an assignment edge does not result in the formation

of a cycle in the resource allocation graph

Cycle detection is O(n2)

7 . 12

BANKER’S ALGORITHMMultiple instances

Each process must a priori claim maximum use

When a process requests a resource it may have to wait

When a process gets all its resources it must return themin a finite amount of time

7 . 137 . 14

BANKER’S ALGORITHM SETUPLet n = number of processes,

Let m = number of resources types.

DATA STRUCTURES NEEDEDAvailable: Vector of length m. If available[j] = k,there are k instances of resource type Rj available

Max: n x m matrix. If Max [i,j] = k, then process Pi mayrequest at most k instances of resource type R j

Allocation: n x m matrix. If Allocation[i,j] = kthen Pi is currently allocated k instances of Rj

Need: n x m matrix. If Need[i,j] = k, then Pi mayneed k more instances of Rj to complete its task

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

7 . 15

SAFETY ALGORITHM 1Let Work and Finish be vectors of length m and n,

respectively.

Initialize:

Work = Available

Finish[i] = false for i = 0,1,… ,n-1

7 . 16

SAFETY ALGORITHM 2Find an i such that both:

1. Finish[i] = false

2. Needi ≤ Work

If no such i exists, go to step 4

7 . 177 . 18

SAFETY ALGORITHM 3Work = Work + Allocationi

Finish[i] = true

go to step 2

7 . 19

SAFETY ALGORITHM 4If Finish [i] == true for all i, then the system is in a

safe state

RESOURCE-REQUEST ALGORITHMLet Requesti be request vector for for Process Pi

If Requesti[j] = k then process Pi wants k instances ofresource type Rj

When resource request made by Pi:

7 . 20

RESOURCE-REQUEST ALGORITHM1. If Requesti ≤ Needi go to step 2. Otherwise, raise error

condition, since process has exceeded its maximum claim

2. If Requesti ≤ Available, go to step 3. Otherwise Pi mustwait, since resources are not available

3. Pretend to allocate requested resources to Pi bymodifying the state as follows:

Available = Available – Request

Allocationi = Allocationi + Requesti

Needi = Needi – Requesti

7 . 217 . 22

RESOURCE-REQUEST ALGORITHMIf safe → the resources are allocated to Pi

If unsafe → Pi must wait, and the old resource-allocationstate is restored

EXAMPLE OF BANKER’S ALGORITHM5 processes P0 through P4

3 resource types:

A (10 instances)

B (5 instances)

C (7 instances)

7 . 23

EXAMPLE OF BANKER’S ALGORITHM

7 . 24

EXAMPLE OF BANKER’S ALGORITHM

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

7 . 257 . 26

EXAMPLE OF BANKER’S ALGORITHMNow a request from P1 comes for (1,0,2)

Check first that

Requesti ≤ Available

EXAMPLE OF BANKER’S ALGORITHM

<P1,P3,P4,P0,P2> satisfies safety criteria

7 . 277 . 28

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

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

8 . 1

DEADLOCK DETECTION

8 . 2

DEADLOCK DETECTIONAllow system to enter deadlock state

Detection algorithm

Recovery scheme

SINGLE INSTANCE OF EACHRESOURCE TYPE

Maintain wait-for graph

Nodes are processes

Pi → Pj if Pi is waiting for Pj

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

An algorithm to detect a cycle in a graph requires O(n2)operations, where n is the number of vertices in the graph

8 . 3

GRAPHS

8 . 4

SEVERAL INSTANCES OF ARESOURCE TYPE

Available: A vector of length m indicates the number ofavailable resources of each type

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

Request: An n x m matrix indicates the current request ofeach process. If Request[i][j] = k, then process Piis requesting k more instances of resource type Rj

8 . 5

DETECTION ALGORITHM 1Let Work and Finish be vectors of length m and n,

respectively

Initialize:

Work = Available

For i = 1,2, … , n, if Allocationi ≠ 0, thenFinish[i] = false; otherwise, Finish[i] =true

8 . 6

DETECTION ALGORITHM 2Find an index i such that both:

Finish[i] == false

Request i ≤ Work

If no such i exists, go to step 4

8 . 78 . 8

DETECTION ALGORITHM 3Work = Work + Allocationi

Finish[i] = true

go to step 2

DETECTION ALGORITHM 4If Finish[i] == false, for some i, 1 ≤ i ≤ n, then thesystem is in deadlock state.

Moreover, if Finish[i] == false, then Pi isdeadlocked

Algorithm requires an order of O(m x n2) operations todetect whether the system is in deadlocked state

8 . 9

EXAMPLEFive processes P0 through P4

Three resource types

A (7 instances)

B (2 instances)

C (6 instances)

8 . 10

EXAMPLESnapshot at time T0

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

8 . 11

EXAMPLEP2 requests an additional instance of type C

State of system?

8 . 12

DETECTION-ALGORITHM USAGEWhen, and how o�en, to invoke depends on:

How o�en a deadlock is likely to occur?

How many processes will need to be rolled back?

→ One for each disjoint cycle

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

8 . 139 . 1

RECOVERY FROM DEADLOCK

9 . 2

PROCESS TERMINATIONAbort all deadlocked processes

Abort one process at a time until the deadlock cycle iseliminated

PROCESS TERMINATIONIn which order should we choose to abort?

1. Priority of the process

2. How long process has computed, and how much longer tocompletion

3. Resources the process has used

4. Resources process needs to complete

5. How many processes will need to be terminated

6. Is process interactive or batch?

9 . 3

RESOURCE PREEMPTIONSelecting a victim – minimize cost

Rollback – return to some safe state, restart process forthat state

Starvation – same process may always be picked asvictim, include number of rollback in cost factor

9 . 410 . 1

QUESTIONS

10 . 2

BONUSExam question number 5: Deadlocks