1. Silberschatz, Galvin and Gagne  2002 8.1 Operating System Concepts 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 P 0 P 1 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.

2. Silberschatz, Galvin and Gagne  2002 8.2 Operating System Concepts Deadlock Characterization 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 {P 0, P 1, …, P n } 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.

3. Silberschatz, Galvin and Gagne  2002 8.3 Operating System Concepts System Model for OS Solution Resource types R 1, R 2,..., 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:  request  use  release

4. Silberschatz, Galvin and Gagne  2002 8.4 Operating System Concepts Resource-Allocation Graph A set of vertices V and a set of directed edges E V is partitioned into two types:  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 1  R j Assignment edge – directed edge R j  P i 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 PiPi PiPi RjRj RjRj

5. Silberschatz, Galvin and Gagne  2002 8.5 Operating System Concepts 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.

6. Silberschatz, Galvin and Gagne  2002 8.6 Operating System Concepts RAG with No Cycles (no deadlock)

7. Silberschatz, Galvin and Gagne  2002 8.7 Operating System Concepts RAG with a Deadlock

8. Silberschatz, Galvin and Gagne  2002 8.8 Operating System Concepts RAG with a Cycle but No Deadlock

9. Silberschatz, Galvin and Gagne  2002 8.9 Operating System Concepts Methods for Handling Deadlocks Deadlock prevention - Ensure that the system will never enter a deadlock state by constraining possible requests Deadlock avoidance - Stop the system from entering a deadlock state by examining current request and maximal requirement for process (declared in advance) Deadlock detection and recovery - Allow the system to enter a deadlock state and then recover. Deadlock acceptance - Ignore the problem and pretend that deadlocks never occur in the system; used by most operating systems, including UNIX.

10. Silberschatz, Galvin and Gagne  2002 8.10 Operating System Concepts Deadlock Prevention Mutual Exclusion – not required for sharable resources; must hold for nonsharable resources. 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. E.g. Network->Disk->Printer  Low resource utilization (early requests); starvation possible (always one resource unavailable). No Preemption - if a process 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. Circular Wait – impose a total ordering of all resource types, and require that each process requests resources in an increasing order of enumeration.  Least valuable claimed first

11. Silberschatz, Galvin and Gagne  2002 8.11 Operating System Concepts Deadlock Avoidance When a resource request is made  Pretend to allocate the resources  Use deadlock detection to check if the allocation and maximal future requirements lead to a deadlock state  If “yes”, delay the allocation Expensive running the algorithm so often Deadlock Detection Allow system to enter deadlock state Periodically run a deadlock detection algorithm Recover if deadlock is detected

12. Silberschatz, Galvin and Gagne  2002 8.12 Operating System Concepts Single Instance of Each Resource Type Maintain wait-for graph - a cycle implies deadlock (but not vice versa) Cycle detection is O(n 2 ) Resource-Allocation GraphCorresponding wait-for graph

13. Silberschatz, Galvin and Gagne  2002 8.13 Operating System Concepts Several Instances of a Resource Type A cycle in the wait-for graph does not imply deadlock Detection algorithm data structures  Available: A vector of length m indicates the number of available resources of each type.  Allocation: An n x m matrix defines the number of resources of each type currently allocated to each process.  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.  Work: A vector length m  Finish: A vector length n Algorithm is O(mn 2 )

14. Silberschatz, Galvin and Gagne  2002 8.14 Operating System Concepts Detection Algorithm Work = Available; //---Copy what is available now For i = 1,2, …, n, if Allocation i  0, then Finish[i] = false; //---P i needs more resources else Finish[i] = true; do { Find an index i such that (Finish[i] == false && Request[i]  Work) If i exists { //----P i can get its resources Work = Work + Allocation[i] Finish[i] = true } while (i exists); If there exists i such that Finish[i] == false the system is in deadlock state P i is deadlocked.

15. Silberschatz, Galvin and Gagne  2002 8.15 Operating System Concepts Example of Detection Algorithm Five processes P 0 through P 4 ; three resource types A (7 instances), B (2 instances), and C (6 instances). Snapshot at time T 0 : AllocationRequestAvailable A B C A B C A B C P 0 0 1 0 0 0 0 0 0 0 P 1 2 0 0 2 0 2 P 2 3 0 30 0 0 P 3 2 1 1 1 0 0 P 4 0 0 2 0 0 2 Sequence will result in Finish[i] = true for all i.

16. Silberschatz, Galvin and Gagne  2002 8.16 Operating System Concepts Example (Cont.) P 2 requests an additional instance of type C. Request A B C P 0 0 0 0 P 1 2 0 1 P 2 0 0 1 P 3 1 0 0 P 4 0 0 2 State of system?  Can reclaim resources held by process P 0, but insufficient resources to fulfill other processes; requests.  Deadlock exists, consisting of processes P 1, P 2, P 3, and P 4.

17. Silberschatz, Galvin and Gagne  2002 8.17 Operating System Concepts Deadlock Recovery Abort all deadlocked processes. Abort one process at a time until the deadlock cycle is eliminated. 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? Resource preemption  Selecting a victim – minimize cost.  Rollback – return to some safe state, restart process for that state.  Starvation – same process may always be picked as victim - include number of rollback in cost factor.