Chapter 7 Deadlocks

Page 2 Chapter 7: Deadlocks System Model Deadlock Characterization Methods for Handling Deadlocks Deadlock Prevention Deadlock Avoidance Deadlock Detection Recovery from Deadlock

Page 3 System Model A system consists of a finite number of resources to be distributed among competing processes. Resources are partitioned into several types, each of which consists of multiple identical instances. CPU cycles, memory space, files, I/O devices, object locks §7.1

Page 4 System Model Each process utilizes a resource as following sequence: request If the request cannot be granted immediately, then the requesting process must wait until it can acquire the resource. use The process can operate on the resource. release The process releases the resource.

Page 5 Deadlock Characterization 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 {P 0, P 1, …, P 0 } of waiting processes such that P 0 is waiting for a resource that is held by P 1, P 1 is waiting for a resource that is held by P 2, …, P n–1 is waiting for a resource that is held by P n, and P n is waiting for a resource that is held by P 0. Deadlock can arise if four conditions hold simultaneously. §7.2

Page 6 Resource-Allocation Graph V is partitioned into two types of nodes: P = {P 1, P 2, …, P n }, the set consisting of all the processes in the system. R = {R 1, R 2, …, R m }, the set consisting of all resource types in the system. Request edge – directed edge P i  R j Assignment edge – directed edge R j  P i A set of vertices V and a set of edges E. §7.2.2

Page 7 Resource-Allocation Graph Process Resource Type with 4 instances P i requests instance of R j P i is holding an instance of R j A request edge points to only the square R j, whereas an assignment edge must also designate one of the dots in the square. PiPi PiPi RjRj RjRj

Page 8 Resource Allocation Graph Example:

Page 9 Resource Allocation Graph Deadlocked:

Page 10 Resource Allocation Graph With Cycle but No Deadlock:

Page 11 Basic Facts If 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 of deadlock.

Page 12 Methods for Handling Deadlocks Use a protocol to ensure that the system will never enter a deadlock state. §7.3 Allow the system to enter a deadlock state and then recover. The system can provide an algorithm that examines the state of the system to determine whether a deadlock has occurred, and an algorithm to recover from the deadlock. Deadlock Prevention: (§8.4) a set of methods for ensuring that at least one of the necessary conditions cannot hold. Deadlock avoidance: (§8.5) give OS information about processes and resources for the decision of satisfying or delaying the requests. Deadlock Detection: (§8.5) Deadlock Recovery: (§8.7) Ignore the problem and pretend that deadlocks never occur in the system.

Page 13 Methods for Handling Deadlocks If no prevention, no avoidance, no detection and no recovery schemes, then the system may reach a deadlock state and eventually need to be restarted manually. Although does not seem to be viable, it is used in most OS, including UNIX. In many systems, deadlocks occur infrequently; thus it is cheaper to use this method. JVM does nothing to manage deadlocks, it is up to the application developer to write programs that are deadlock-free.

Page 14 Deadlock Prevention By ensuring that at least one of the conditions (§8.2) cannot hold, we can prevent the occurrence of a deadlock.(§8.2) 1.Mutual Exclusion – not required for sharable resources, but must hold for nonsharable resources (printers, synchronized method) §7.4

Page 15 Deadlock Prevention 2.Hold and Wait – must guarantee that whenever a process requests a resource, it does not hold any other resources. 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… release all the resources that is is currently allocated before requesting additional resources. Low resource utilization; starvation possible.

Page 16 Deadlock Prevention 2.Hold and Wait – must guarantee that whenever a process requests a resource, it does not hold any other resources. 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… release all the resources that is is currently allocated before requesting additional resources. Low resource utilization; starvation possible.

Page 17 Deadlock Prevention 2.Hold and Wait – must guarantee that whenever a process requests a resource, it does not hold any other resources. 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… release all the resources that is is currently allocated before requesting additional resources. Low resource utilization; starvation possible.

Page 18 Deadlock Prevention 2.Hold and Wait – must guarantee that whenever a process requests a resource, it does not hold any other resources. 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… release all the resources that is currently allocated before requesting additional resources. Low resource utilization; starvation possible.

Page 19 Deadlock Prevention (Cont.) No Preemption – 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. Preempted resources are added to the list of resources for which the process is waiting. Process will be restarted only when it can regain its old resources, as well as the new ones that it is requesting. Often applied to resources whose state can be easily saved and restored later, such as CPU registers and memory space. It cannot be applied to resources such as printers and tape drives.

Page 20 Deadlock Prevention (Cont.) Circular Wait – impose a total ordering of all resource types, and require that each process requests resources in an increasing order of enumeration. Let R={R 1,R 2, …, R m } be the set of resource types. F:R→N F(type drive) =1 F(disk drive) = 5 F(printer) = 12 A process owning resource type R i can continue to request type R j iff F(R j )>F(R i ) or Whenever a process requests an instance of resource type R j, it has released any resources R i such that F(R i ) >=F(R j ). Should be defined according to the normal order of usage of the resources in a system.

Page 21 Proof by contradiction Assume circular wait exists in {P 0, P 1,..., P n }, where P i is waiting for a resource R i, which is held by process P i+1. F(R i ) F(R 0 ) F(R 0 ) < F(R 0 ) Impossible

Page 22 Deadlock Avoidance Simplest and most useful model requires that each process declare the maximum number of resources of each type that it may need. The deadlock-avoidance algorithm dynamically examines the resource- allocation state to ensure that there can never be a circular-wait condition. Resource-allocation state is defined by the number of available and allocated resources, and the maximum demands of the processes. Requires that the system has some additional a priori information available. §7.5

Page 23 Safe State A state is safe if the system can allocate resources to each process in some order and still avoid a deadlock. Safe sequence : if, for each P i, the resources that P i can still request can be safisfied by the currently available resources plus the resources held by all the P j with j < i. If no such sequence exists, then the system is said to be unsafe. §7.5.1

Page 24 Safe State A safe state is not a deadlock state. A deadlock state is an unsafe state. Not all unsafe states are deadlocks. An unsafe state may lead to a deadlock. In unsafe state, OS cannot prevent processes from requesting resources and cause deadlock: the behavior of the processes controls unsafe state.

Page 25 Safe State Example: 12 tape drives, 3 processes. max needs current holding P P1 4 2 P2 9 2 Safe for If P2 request one more, system go into unsafe state.

Page 26 Basic Facts If 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 an unsafe state.

Page 27 Resource-Allocation Graph Algorithm Claim edge P i  R j indicated that process P i may request resource R j at some time in the future. It is represented by a dashed line. Claim edge converts to request edge when a process requests a resource. When the resource is allocated to the process, assignment edge is added. When a resource is released by a process, assignment edge reconverts to a claim edge. P1P1 R1R1 §7.5.2

Page 28 Safe State A request can be granted only if converting the request edge to assignment edge does not result in the formation of a cycle in the resource- allocation graph. If no cycle exists, then the allocation of the resource will leave the system in a safe state. Otherwise, process have to wait.

Page 29 Banker’s Algorithm (From SHAY) Multiple 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 them in a finite amount of time. §7.5.3

Page 30 Safe Process 1,2, or 4 could request its maximum number of pages and finish For example: process 1 finish Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 4 pages are unallocated Example 1

Page 31 Now any process could request pages up to its maximum and finish Suppose process 3 finish Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 6 pages are unallocated Example 1

Page 32 Any one can finish. However, a sequence of requests that allowed all processes to finish does not indicate a safe state. Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 7 pages are unallocated Example 1

Page 33 If process 3 requests the 4 available pages, and each process subsequently requests 1 more page, the system is deadlocked However, it doesn't necessarily occur.....A unsafe state doesn't mean that the deadlock will occur inevitably. Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 0 pages are unallocated Example 1

Page 34 Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 1 pages are unallocated Example 2

Page 35 Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 1 pages are unallocated Process Process Process Process pages in the system 2 pages are unallocated Example 2

Page 36 If each process asserts its claim, deadlock will occur Unsafe ! Current Number of pages allocated Maximum Number of pages needed Process Process Process Process pages in the system 1 pages are unallocated Process Process Process Process pages in the system 2 pages are unallocated Process Process Process Process pages in the system 4 pages are unallocated Example 2

Page 37 Banker’s Algorithm Safe state. If request 1 =(1,0,2) §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 38 Banker’s Algorithm Safe state. If request 1 =(1,0,2) Still safe with §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 39 Banker’s Algorithm Safe state. Request 4 =(3,3,0) §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 40 Banker’s Algorithm Safe state. Request 4 =(3,3,0) Cannot be granted. No resources. §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 41 Banker’s Algorithm Safe state. Request 0 =(0,2,0) §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 42 Banker’s Algorithm Safe state. Request 0 =(0,2,0) Will result in unsafe state. §7.5.3 P P P P P Allocation Max Available A B C A B C A B C P P P P P Need A B C

Page 43 Deadlock Detection Allow system to enter deadlock state System should provide: An algorithm that examines the state of the system to determine whether a deadlock has occurred. An algorithm to recover from the deadlock. It requires overhead that includes maintaining necessary information executing the detection algorithm potential losses when recovering §7.6

Page 44 Single Instance of Each Resource Type Maintain wait-for graph: obtained from the resource-allocation graph by removing the nodes of type resource and collapsing the appropriate edges. Nodes are processes. P i  P j if P i is waiting for P j. to release a resource that P i needs. Needs to maintain the wait-for graph and periodically invoke an algorithm that searches for a cycle in the graph. An algorithm to detect a cycle in a graph requires an order of O(n 2 ) operations, where n is the number of vertices in the graph. §7.6.1

Page 45 Resource-Allocation Graph And Wait-for Graph Resource-Allocation GraphCorresponding wait-for graph

Page 46 Detection-Algorithm Usage When, and how often, to invoke depends on: How often a deadlock is likely to occur? How many processes will need to be rolled back? one for each disjoint cycle If deadlocks occur frequently, then the detection algorithm should be invoked frequently. In the extreme, we could invoke the deadlock-detection algorithm every time a request for allocation cannot be granted immediately. Expensive

Page 47 Detection-Algorithm Usage Less expensive alternative:invoke the algorithm at less frequent intervals. 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.

Page 48 Several Instances of a Resource Type Not deadlocked. If P2 make request (0, 0, 1), Deadlocked between P 1, P 2, P 3, P 4. §7.6.2 P P P P P Allocation Request Available A B C A B C A B C

Page 49 Several Instances of a Resource Type Not deadlocked. If P2 make request (0, 0, 1), Deadlocked between P 1, P 2, P 3, P 4. §7.6.2 P P P P P Allocation Request Available A B C A B C A B C

Page 50 Recovery from Deadlock: Process Termination Abort all deadlocked processes. Expensive to discard and re-compute partial computation. Abort one process at a time until the deadlock cycle is eliminated. overhead for invoking deadlock- detection algorithm after each process is aborted. §7.7

Page 51 Recovery from Deadlock: Process Termination We should abort those processes that will incur the minimum cost when terminated. 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?

Page 52 Recovery from Deadlock: Resource Preemption Successively preempt some resources from processes and give these resources to other processes until the deadlock cycle is broken. Three matters addressed: Selecting a victim – Which resources and which processes are to be preempted? We must determine the order to minimize cost. Factors include: Number of resources a deadlock process is holding the amount of time a deadlocked process has consumed Rollback – return to some safe state, restart process fro that state. Requires the system to keep more information about the state of all the running processes. Starvation – same process may always be picked as victim. Should include number of rollback in cost factor.