Operating systems. Lecture 5. Deadlock: System Model. Deadlock: System Model. Process synchronization Deadlocks. Deadlock: System Model

Transcription

1 Lecture 5 Operating systems Process synchronization Deadlocks Deadlock: System Model Computer system: Processes (program in execution); Resources (CPU, memory space, files, I/O devices, on so on). Deadlock: System Model A process use a resource as: 3 Request: A process must request a resource before using it. If the resource is not available, the process enter a wait state. Use: The process holds and operates on the resource. Release: The process must release the resource after using it. System calls are request and release. Deadlock: System Model A deadlock state occurs when two or more processes are waiting for a event that can be caused only by one of the waiting processes. STOP Train Train STOP

2 Deadlock: Necessary Conditions A deadlock situation can arise if all of the following four conditions hold simultaneously in a system: Mutual exclusion: At least one resource must be held in a non-sharable mode; only one process at a time can use the resource.. If another process requests that resource, the requesting process must be delayed until the resource has been released. Deadlock: Necessary Conditions Hold and wait: There must be exist a process that is holding at least one resource and waiting to get additional resources that are currently being held by other processes. Deadlock: Necessary Conditions 3 No preemption: Resources that cannot be preempted; ; that is, a resource can be released only voluntary by the process holding it, after that process has completed its task. Only the process holding this resource can release it. 4 Deadlock: Necessary Conditions Circular wait: There must exist a set {P{ 0, P,.., P n } of waiting processes such that P 0 is waiting for a resource that is held by P, P is waiting for a resource that is held by P,,, P n- 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. For a deadlock to occur, all the four conditions must hold.

3 Deadlocks can be described more precisely in term of a directed graph, called a system resource-allocation graph. The nodes are processes Deadlock Characterization: Resource- Allocation Graph ( P = {P{, P,,P n }) And resource types ( R = {R{, R,,R n }) A directed edge from process P i to resource R j is denoted by P i -> R j (request edge); it signifies that process P i requested an instance of resource R j and is currently waiting for that resource. Directed edge from resource type R j to process P i is denoted by R j -> P i (assignment edge); it signifies that an instance of resource type R j has been allocated to process P i. Deadlock: Resource- Allocation Graph Deadlock: Resource- Allocation Graph P i is denoted as a circle and R j is denoted as a square. Each instance of R j is denoted as a dot in the square. R 3 Deadlock: Resource- Allocation Graph P = {P, P, P 3 } R = {, R, R 3, R 4 } E = {P, P R 3, P, R P, R P, R 3 P 3 } has instance, R, R 3, R 4 3 Process P and P are in waiting state. R 3 P P P 3 P P P 3 R R 4 R R 4 3

4 Deadlock: Resource- Allocation Graph If the graph has no cycle, no process is deadlocked. If the graph has a cycle, deadlock may exist. If each resource type involved in a cycle has only one instance, the cycle implies a deadlock: R 3 If each resource type involved in a cycle has more than one instance, the deadlock may or may not exist: Deadlock: Resource- Allocation Graph P P P P 3 R R 4 P With a cycle but no deadlock R P 3 P 4 Deadlock: Methods for Handling Principally, there are three methods for dealing with the deadlock problem: 3 We can use a protocol to ensure that the system will never enter a deadlock state. We can allow the system to enter a deadlock state and then recover. We can ignore the problem all together, and pretend that deadlocks never occur in the system. This solution is the one used by most OS, including UNIX. Deadlock: Methods for Handling To ensure deadlocks not to occur, deadlock prevention or deadlock avoidance schemes can be used. Deadlock prevention ensures that at least one of the four necessary conditions for deadlock cannot hold. Deadlock avoidance require in advance the information on resources that a process will use. 4

5 Deadlock: Prevention Deadlock: Prevention To deny one of the following: Mutual exclusion: it is not possible to prevent deadlocks by denying mutual-exclusion condition: some resources are intrinsically nonsharable (e.g. printer) Hold and wait: ensure when a process requests a resource, it does not hold any other resource. Methods: : () a process requests and gets all the resources before it begins execution, or () a process requests resources only when it has none. Disadvantages: : lower resource utilization and starvation. Deadlock: Prevention Deadlock: Prevention 3 No preemption: 4 Circular wait: preemption resources that are held by the processes which are waiting for other resources. define a total on the resource types. A process can requests a resources only in the increasing order of the resources. Assume T(R j ) is the order of R j, i m. After a process requests R j, it can only request the R j with T(R j ) > T(R i ). 5

6 Deadlock: Avoidance Deadlock: Avoidance The methods of deadlock avoidance require additional information about how resources are to be requested. The simplest model requires the maximum number of resources that a process uses in advance. The resource-allocation state is defined by the number of available, allocated,, and the maximum requested resources. Safe sequence A sequence of processes <P<, P,, P n > is a safe sequence for the current allocation state if, for each P i, the resources that P i can still request can be satisfied by currently available resources plus the resources held by all the P j, with j < i. Deadlock: Avoidance Safe sequence is such that: the first process can finish for sure there are enough unallocated resources to satisfy all of its claim If the first process releases its currently held resources, the second process can finish for sure (even if it asks all its claim) and so on. In this situation, if the resources that P i needs are not immediately available, then P i can wait until all P j have finished. When they have finished, P i can obtain all of its needed resources, complete task, return its allocated resources, and terminate. When P i terminates, P i + can obtain its needed resources, and so on. If no such safe sequence exists, then the system state is said to be unsafe. Safe State 6

7 Safe State: Example (One resource class only) total resources: unallocated: process holding max claims A 4 6 B 4 C 7 safe sequence: A,C,B Unsafe State: Example (One resource class only) total resources: unallocated: process holding max claims A 4 6 B 4 C 9 safe sequence: none! In addition to request edge (P i -> R j ) and assignment edge (R j -> P i ), define a new type of edge claim edge P i -> R j (represented by dashed line), denotes that P i may requests R j in the future. When P i request R j, the claim edge P i -> R j is converted to request edge.. A claim edge P i -> R j can be added to the graph only if all the edges associated to P i are claim edges. Deadlock: Resource- Graph Algorithm Allocation To avoid the deadlock, a request can be granted only if converting the request edge P i -> R j to an assignment edge R j -> P i does not produce a cycle in the graph. Deadlock: Resource- Graph Algorithm P P R P P R Allocation Unsafe state 7

8 The resource-allocation graph algorithm is not applicable to a resource-allocation system with multiple instances for each resource type. For this case, we will use the banker's algorithm. This algorithm could be used in a banking system to ensure that the bank never allocates its available cash such that it can no longer satisfy the needs of all its customers. Deadlock Avoidance: Banker's Algorithm When a new process enters the system, it must declare the maximum number of each resource type that it may need. This number may not exceed the total number of resources. When a user requests a set of resources, the system must determine whether the allocation of these resources will leave the system in a safe state. If it will, the resources are allocated; otherwise, the process must wait until some other process releases enough resources. Deadlock Avoidance: Banker's Algorithm Deadlock Avoidance: Banker's Algorithm Deadlock Avoidance: Banker's Algorithm Let n be the number of processes in the system and m be the number of resource types. We need the following data structures: Available: Ava[ m]. Ava[j] = k means k instances of R j are available. Max: Max[..n,..m]. Max[i, j] = k means P i requests at most k instances of R j. Allocation: Allo[..n,..m]. Allo[i, j] = k means P i holds k instances of R j. Need: Need[..n,..m]. Need[i, j] = k means P i may need k additional instances of R j. Need[i, j] = Max[i, j] - Allo[i, j]. For X[..n] and Y[..n], X Y if X[i] Y[i],, i n. X < Y if X Y and X Y. The i-th rows of Max, Allo, Need are denoted as Max i, Allo i, Need i, respectively. 8

9 Banker's Algorithm: Safety Algorithm The algorithm for finding out whether or not a system is in a safe state can be presented as:. Work[..m] := Ava; Finish[..n] := false;. Find i such both Finish[i] := false and Need i Work then step 3 else step 4; 3. Work := Work + Allo i ; Finish[i] := true; go to step ; 4. If Finish[i] = true for all i then the system is safe else unsafe; O(mn ) operation Banker's Algorithm: Resource- Request Algorithm Let Req i be the request vector for process P i. If Req i [j] = k, then process P i wants k instances of resource type R j. When a request for resources is made by process P i, the following actions are taken: 3 4 Banker's Algorithm: Resource- Request Algorithm If Req i Need i then go to step else error, since the process has exceeded is maximum claim. If Req i Ava then go to step 3 else P i must wait, since the resources are not available. Ava := Ava - Req i ; Allo i := Allo i Need i := Need i Req i ; + Req i ; Check the system by Safety Algorithm. If the system is safe then allocates Req i to P i else P i wait and the old resource-allocation state is restored. Five processes P 0,P,P,P 3,P 4. Three types of resources A (0 instances), B (5 instances), C (7 instances). At the time T 0. Banker's Algorithm: An Example The system is in a safe state. The sequence P,P 3,P 4,P,P 0 satisfies the safety criteria. P 0 P P P 3 P 4 Ava 3 3 Max Allo Need

10 Banker's Algorithm: An Example Banker's Algorithm: An Example Assume P gives a request of Req = (,0,).( We first check Req Ava : (,0,) (3,3,). The system enters a state of: P 0 P P P 3 P 4 Ava 3 3 Max Allo Need By safety algorithm, P,P 3,P 4,P,P 0 is a safety sequence and Req can be allocated. If P 4 requests (3,3,0)( ) at this state, the request cannot be granted since the resources are not available. If P 0 requests (0,,0),( even though the resources are available, the granting will make the system unsafe. Deadlock Detection Deadlock Detection: Single Instance Case If a system does not employ either a deadlock-prevention or a deadlock- avoidance, then a deadlock situation may occur. In this environment, the system must provide: An algorithm that examines the state of the system to determine whether a deadlock has occurred An algorithm to recover from a deadlock If all resources have only a single instance, then we can define a deadlock detection algorithm that uses a variant of the resource-allocation graph, called a wait-for graph. We obtain this graph from the resource- allocation graph by removing the nodes of type resource and collapsing the appropriate edges. 0

11 Deadlock Detection: Single Instance Case Deadlock Detection: Single Instance Case An edge from P i to P j in a wait-for graph implies that process P i is waiting for process P j to release a resource that P i needs. An edge P i -> P j exist in a wait-for graph iff the corresponding resource-allocation graph contains two edges P i -> R q and R q -> P i for some resource R q. (see example on the next slide). Resource-allocation graph P R P 5 R 3 R 4 P P 3 P 4 R 5 Corresponding wait-for graph P 5 P P P 3 P 4 Deadlock Detection: Single Instance Case Deadlock Recovery: Process Termination A deadlock exists in the system iff the wait-for graph contains a cycle. To detect deadlocks, the system needs to maintain the wait-graph and periodically to invoke an algorithm that searches for a cycle in the graph. An algorithm requires O(n ) operations, where n is the number of vertices in the graph. Abort all deadlocked processes. Simple method, high expense. Abort one process at a time until the deadlock cycle eliminated. Large overhead, since, after each process is aborted, deadlock-detection detection algorithm must be invoked to determine whether any process are still deadlocked.

12 Deadlock Recovery: Process Termination To choose a process to be aborted, the following should be considered: The priority of the process. The time the process has executed and will be executed. How many and what type of resources used/using. Resources needed further. Number of processes needed to be aborted. Whether the process is interactive or batch. Deadlock Recovery: Resource Preemption To eliminate deadlocks using resource preemption, we successively preempt some resources from processes and give these resources to other processes until the deadlock cycle is broken. If preemption is required to deal with deadlocks, then three issues need to be addressed: Deadlock Recovery: Resource Preemption Deadlock Recovery: Resource Preemption Selecting a victim: Which resources and which processes are to be preempted? As in process termination, we must determine the order of preemption to minimize cost. Cost includes such factors as: the number of resources a deadlock process is holding, and amount of time a deadlocked process has thus far consumed during its execution. Rollback: If we preempt a resource from a process, what should be done with that process? It cannot continue with its normal execution; it is missing some resource. We must roll back the process to some safe state, and restart it from that state. The simplest method is aborting that process and then restart it.

13 3 Deadlock Recovery: Resource Preemption Starvation: How do we ensure that starvation will not occur? That is, how we guarantee that resources will not always be preempted from the same process? Deadlock: Combined Approach A deadlock handling approach can be made by the combination of deadlock-prevention, deadlock-avoidance, avoidance, and detection- recovery. Deadlock: Summary A deadlock state occurs when two or more processes are waiting indefinitely for an event that can be caused only by one of the waiting processes. Three methods for handling deadlocks: ) To ensure the system never enters a deadlock state. ) Allow the system to enter the deadlock state and then recover. 3) Ignore the problem. Deadlock: Summary There are four necessary conditions for a deadlock. Deadlock-prevention ensures one of the conditions never occur. Deadlock-avoidance avoidance requires the information of the requested resources of each process in advance. Detection and recovery is used in the second approach. Termination of processes and preemption of resources can be used in deadlock recovery. 3