Advanced Operating Systems Spring 2009 Lecture 10 February

Advanced Operating Systems - Spring 2009 Lecture 10 – February 11, 2009 l Dan C. Marinescu l l TA: Chen Yu l l Email: dcm@cs. ucf. edu Office: HEC 439 B. Office hours: M, Wd 3 – 4: 30 PM. Email: yuchen@cs. ucf. edu Office: HEC 354. Office hours: M, Wd 1. 00 – 3: 00 PM. Warning: A student missing classes takes a serious risks as the problems at exams require understanding of fine points discussed in class and problem analyzed in class. 1

Last, Current, Next Lecture l Last time: l l Deadlocks Today l l Problem solving More about deadlocks Consumable resources Elements of queuing theory l l Arrival and service processes Little’s law Problem Next time: l M/M/1 systems 2

Problem l Write a procedure consisting of a single executable instruction that could cause the process calling this procedure to deadlock. 3

Problem l You have a system consisting of n processes and one nonsharable resource with m units. Show that a sufficient condition for the system to be deadlock-free is that the sum of the maximum amount of resources required by all processes is less than (m+n) units. 4

Problem l Two processes P 1 and P 2 share a resource with two units. Each process can be in one of five states: l l l The following transitions may occur: l l S 0 no unit S 1 no unit, request one unit S 2 hold one unit S 3 hold one unit request a second one S 4 hold two units T 1 request one unit T 2 acquire one unit T 3 release one unit Construct a state diagram of the system and identify the states Sij with i, j in the range (0, 4) when the system is deadlocked. 5

Problem l Prove that a necessary condition for deadlock is that a cycle exists in the resource allocation graph. 6

Deadlocks and graph theory l If a directed graph does not contain a cycle there exists a linear ordering of nodes such that when there exits a path from node i to node j then node i appears in this ordering before node j 7

Banker’s algorithm Multiple resource instances. l Each process must a priori claim maximum use. l When a process l l l requests a resource it may have to wait gets all its resources it must return them in a finite amount of time. 8

Data structures for banker’s algorithm n # of processes; m # of resources types. l Available: Vector of length m: l Available [j] = k there are k instances of resource type Rj available. Max: n x m matrix: l Max [i, j] = k Pi may request at most k instances of resource type Rj. l Allocation: n x m matrix: l l l Allocation[i, j] = k Pi is currently allocated k instances of Rj. Need: n x m matrix: l Need[i, j] = k Pi may need k more instances of Rj to complete its task. Need [i, j] = Max[i, j] – Allocation [i, j]. 9

Safety algorithm 1. Work and Finish are vectors of length m and n, respectively. Initialize: Work = Available Finish [i] = false for i = 0, 1, …, n- 1. 2. Find i such that: (a) Finish [i] = false (b) Needi Work If no such i exists, go to step 4. 3. Worki = Worki + Allocationi Finish[i] = true go to step 2. 4. If Finish [i] == true for all i, then the system is in a safe state. 10

Resource request algorithm for process Pi If request vector Requesti [j] = k then Pi wants k instances of resource type j (Rj. ) 1. If Requesti Needi go to step 2. Otherwise error (process has exceeded its maximum claim). 2. If Requesti Available, go to step 3. Otherwise Pi must wait (resources are not available). 3. Pretend to allocate requested resources to Pi by modifying the state as follows: Available = Available – Request; Allocationi = Allocationi + Requesti; Needi = Needi – Requesti; l If safe the resources are allocated to Pi l If unsafe Pi must wait, and the old resourceallocation state is restored 11

Example 5 processes P 0 through P 4; 3 resource types: A (10 instances), B (5 instances), and C (7 instances). l Snapshot at time T 0: Allocation Max Available ABC ABC P 0 010 753 332 l P 1 200 322 P 2 302 902 P 3 211 222 P 4 002 433 12

Example (cont’d) l The content of the matrix Need is defined to be Max – Allocation. P 0 P 1 P 2 P 3 P 4 l Need ABC 743 122 600 011 431 The system is in a safe state since the sequence < P 1, P 3, P 4, P 2, P 0> satisfies safety criteria. 13

Example: P 1 Request (1, 0, 2) Check that Request Available (that is, (1, 0, 2) (3, 3, 2) true. Allocation Need Available ABC ABC P 0 010 743 230 P 1 302 020 P 2 301 600 P 3 211 011 P 4 002 431 l Executing safety algorithm shows that sequence < P 1, P 3, P 4, P 0, P 2> satisfies safety requirement. l Can request for (3, 3, 0) by P 4 be granted? l Can request for (0, 2, 0) by P 0 be granted? l 14

Deadlock detection l Allow system to enter deadlock state l Detection algorithm l Recovery scheme 15

Single instance of each resource type l Maintain wait-for graph l Nodes are processes. l Pi Pj if Pi is waiting for Pj. l Periodically invoke an algorithm that searches for a cycle in the graph. If there is a cycle, there exists a deadlock. l An algorithm to detect a cycle in a graph requires an order of n 2 operations, where n is the number of vertices in the graph. 16

Resource allocation and wait-for graphs Resource-Allocation Graph Corresponding wait-for graph 17

Multiple instances of a resource type l Available: A vector of length m indicates the number of available resources of each type. l Allocation: An n x m matrix defines the number of resources of each type currently allocated to each process. l Request: An n x m matrix indicates the current request of each process. If Request [ij] = k, then process Pi is requesting k more instances of resource type. Rj. 18

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 Allocationi 0, then Finish[i] = false; otherwise, Finish[i] = true. 2. Find an index i such that both: (a) Finish[i] == false (b) Requesti Work If no such i exists, go to step 4. 19

Detection algorithm (cont’d) 3. Work = Work + Allocationi 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 Pi is deadlocked. The algorithm requires O(m x n 2) operations to detect whether the system is in deadlocked state. 20

Example Five processes P 0 through P 4; three resource types A (7 instances), B (2 instances), and C (6 instances). l Snapshot at time T 0: Allocation Request Available ABC ABC P 0 010 000 P 1 200 202 P 2 303 000 P 3 211 100 P 4 002 l Sequence <P 0, P 2, P 3, P 1, P 4> will result in Finish[i] = true for all i. l 21

Example (Cont’d) P 2 requests an additional instance of type C. Request ABC P 0 000 P 1 201 P 2 001 P 3 100 P 4 002 l State of system? l Can reclaim resources held by process P 0, but insufficient resources to fulfill other processes; requests. l Deadlock exists, consisting of processes P 1, P 2, P 3, and P 4. l 22

Detection algorithm l When, and how often, to invoke depends on: l How often a deadlock is likely to occur? l How many processes will need to be rolled back? l one for each disjoint cycle l 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. 23

Recovery from deadlock: process termination Abort all deadlocked processes. l Abort one process at a time until the deadlock cycle is eliminated. l In which order should we choose to abort? l Priority of the process. l How long process has computed, and how much longer to completion. l Resources the process has used. l Resources process needs to complete. l How many processes will need to be terminated. l Is process interactive or batch? l 24

Recovery from deadlock: resource preemption l Selecting a victim – minimize cost. l Rollback – return to some safe state, restart process for that state. l Starvation – same process may always be picked as victim, include number of rollback in cost factor. 25