1 Announcements The fixing the bug part of Lab 4’s assignment 2 is now considered extra credit. Comments for the code should be on the parts you wrote for the assignment.

2 Concurrency: Deadlock and Starvation Chapter 6

3

4 Deadlock Permanent blocking of a set of processes that either compete for system resources or communicate with each other No efficient solution Involve conflicting needs for resources by two or more processes

5 Deadlock Example P1 P2.. Get A Get B.. Get B Get A.. Release A Release B Release B Release A

6

7 Deadlock Example P1 P2.. Get A Get B.. Release A Get A.. Get B Release B Release B Release A

8

9 Deadlock Environment Mutual exclusion –Only one process may use a resource at a time Hold-and-wait –A process may hold allocated resources while awaiting assignment of others No preemption –No resource can be forcibly removed form a process holding it

10 Condition for Deadlock Circular wait –A closed chain of processes exists, such that each process holds at least one resource needed by the next process in the chain

11 Deadlock Prevention Mutual Exclusion –Must be supported by the operating system Hold and Wait –Require a process request all of its required resources at one time

12 Deadlock Prevention No Preemption –Process must release resource and request again –Operating system may preempt a process to require it releases its resources Circular Wait –Define a linear ordering of resource types

13 Deadlock Avoidance A decision is made dynamically whether the current resource allocation request will, if granted, potentially lead to a deadlock Requires knowledge of future process request

14 Two Approaches to Deadlock Avoidance Do not start a process if its demands might lead to deadlock Do not grant an incremental resource request to a process if this allocation might lead to deadlock

15 Determination of a Safe State Initial State

16 Determination of a Safe State P2 Runs to Completion

17 Determination of a Safe State P1 Runs to Completion

18 Determination of a Safe State P3 Runs to Completion

19 Resource Allocation Denial Referred to as the banker’s algorithm State of the system is the current allocation of resources to process Safe state is where there is at least one sequence that does not result in deadlock Unsafe state is a state that will lead to deadlock if no resources released and rest requested.

20 Reusable Resources Used by only one process at a time and not depleted by that use Processes obtain resources that they later release for reuse by other processes Processors, I/O channels, main and secondary memory, devices, and data structures such as files, databases, and semaphores Deadlock occurs if each process holds one resource and requests the other

21 Consumable Resources Created (produced) and destroyed (consumed) Interrupts, signals, messages, and information in I/O buffers Deadlock may occur if a Receive message is blocking May take a rare combination of events to cause deadlock

22 Example of Reusable Deadlock Deadlock occurs if receive is blocking P1... Receive(P2); Send(P2, M1); P2... Receive(P1); Send(P1, M2);

23 Deadlock Avoidance Summary Maximum resource requirement must be stated in advance Processes under consideration must be independent; no synchronization requirements There must be a fixed number of resources to allocate No process may exit while holding resources

24 Sites on Deadlock

25 Deadlock Detection

26 Strategies once Deadlock Detected Abort all deadlocked processes Back up each deadlocked process to some previously defined checkpoint, and restart all process –Original deadlock may occur Successively abort deadlocked processes until deadlock no longer exists Successively preempt resources until deadlock no longer exists

27 Selection Criteria Deadlocked Processes Least amount of processor time consumed so far Least number of lines of output produced so far Most estimated time remaining Least total resources allocated so far Lowest priority

28 Strengths and Weaknesses of the Strategies

29 Real World Solutions Integrated Deadlock Strategy –Group resources into different classes and enforce a linear ordering strategy. Prevention (avoiding circular waits) Within classes can still have multiple access.

30 Example Classes Swappable space : Virtual Memory –Must make all requests at the same time. (Avoiding the Hold-and-Wait problem) Process Resources : Assignable devices (ex: tape drives and files) –Know ahead of time (avoidance) –Linear ordering (circular wait nullified) Main Memory : –Preemption -> swap out to virtual memory

31 Example Classes Internal Resources : I/O channels –Resource ordering.

32 Dining Philosophers Problem

33 Dining Philosophers Problem

34 Dining Philosophers Problem

35 Dining Philosophers Problem

36 Dining Philosophers Problem

37 Deadlock in OS Most Operating Systems assume deadlock won’t occur. –Within the kernel processes. Make all kernel level resource allocations deadlock free. –Get all. –Preemption. –Linear ordering. Side Note: Interrupts can play havok in OS design. How does the kernel determine what is a resource to be concerned about?

38 UNIX Concurrency Mechanisms Shared memory –Block of memory that both processes have access to. Similar to a pipe, with fewer built in controls. Pipes –Sending messages but a fixed byte size to store the message. Messages (msgsnd and msgrcv) –Able to store messages in a queue. –Blocks on send when queue full and on a read when queue empty.

39 UNIX Concurrency Mechanisms Semaphores (semWait and semSignal) –Can be managed in sets semctl : Used to initialize all semaphores in the set to the same value at the same time. sem_op : Pass in a set of individual operations to be performed in order. Signals –Very similar to an interrupt, but all signals have the same priority level. –Type determines what happens when received. SIGKILL : Kill the process SIGCHLD : Inform process on the death of a child

40 Linux Kernel Concurrency Mechanisms Has a set of atomic operations. (atomic_t) –Operations execute without interruption and without interference. –atomic_dec_and_test : Subtract a 1 from the variable and return 1 if equals 0. –atomic_int_and_test : Add a 1 to the variable and return 1 if equals 0. –atomic_read : Gets the value of the atomic variable. –atomic_set : Sets the value of the atomic variable to the specified value.

41 Linux Kernel Concurrency Mechanisms Spinlocks –Very similar to semaphores but designed for shorts waits. –Doesn’t put the thread to sleep while waiting for the spinlock to release. –Can be done with interrupts enabled or disabled. Semaphores –Puts the thread to sleep if semaphore unavailable. –Can be done with interrupts enabled or disable.

42 Linux Kernel Concurrency Mechanisms Used when code is reordered to optimize processor efficiency. a = 1; b = 1;

43 Solaris Thread Synchronization Primitives Mutual exclusion (mutex) locks –Basically a binary semaphore. Semaphores –Basic counting semaphore. Condition variables –Allows a thread to wait until a specific condition occurs. –Used with a mutex lock.

44 Condition Example mutex_enter(&m) while (some_condition) { cv_wait(&cv,&m); } mutex_exit(&m)

45

46 Extra Slides

47 Example of Deadlock

48 Another Example of Deadlock Space is available for allocation of 200Kbytes, and the following sequence of events occur Deadlock occurs if both processes progress to their second request P1... Request 80 Kbytes; Request 60 Kbytes; P2... Request 70 Kbytes; Request 80 Kbytes;

49

50 Possibility of Deadlock Mutual Exclusion No preemption Hold and wait

51 Existence of Deadlock Mutual Exclusion No preemption Hold and wait Circular wait

52 Determination of an Unsafe State

53 Determination of an Unsafe State

54 Resource Allocation Graphs Directed graph that depicts a state of the system of resources and processes

55 Resource Allocation Graphs

56 Banker’s Algorithm

57 Banker’s Algorithm

58

59 Linux Atomic Operations

60 Linux Atomic Operations

61 Linux Kernel Concurrency Mechanisms Spinlocks –Used for protecting a critical section

62

63