Movatterモバイル変換

PROCESSSYNCHRONIZATIONINOPERATING SYSTEMS1ByRITU RANJAN SHRIVASTWAEmail: ranjanbhai.latest@gmail.com

TABLE OF CONTENTS• What is Process Synchronization and why it is needed• The Critical Section Problem• Peterson’s Solution• Synchronization Hardware• Semaphores• Applications of Semaphores• Classical Problems of Synchronizations• Synchronization Examples• Monitors• Atomic Transactions• References2

WHAT IS PROCESSSYNCHRONIZATION?• Several Processes run in an Operating System• Some of them share resources due to which problems likedata inconsistency may arise• For Example: One process changing the data in amemory location where another process is trying to readthe data from the same memory location. It is possiblethat the data read by the second process will beerroneous3

WHAT IS PROCESSSYNCHRONIZATION?• PRODUCER CONSUMER PROBLEM(or Bounded-Buffer Problem)• Problem: To ensure that the Producer should not addDATA when the Buffer is full and the Consumer shouldnot take data when the Buffer is empty4BufferProducer(process A)Consumer(process B)

WHAT IS PROCESSSYNCHRONIZATION?• SOLUTION PRODUCER CONSUMER PROBLEM• Using Semaphores (In computer science, particularly in operating systems,a semaphore is a variable or abstract data type that is used for controlling access, bymultiple processes, to a common resource in a concurrent system such asa multiprogramming operating system.)• Using Monitors (In concurrent programming, a monitor is a synchronizationconstruct that allows threads to have both mutual exclusion and the ability to wait(block) for a certain condition to become true)• Atomic Transactions (Atomic read-modify-write access to shared variables isavoided, as each of the two Count variables is updated only by a single thread. Also,these variables stay incremented all the time; the relation remains correct when theirvalues wrap around on an integer overflow)5

RACE CONDITION6What if T5 happened before T4??

CRITICAL SECTION PROBLEM• A section of code, common to n cooperatingprocesses, in which the processes may beaccessing common variables.• A Critical Section Environment contains:• Entry Section Code requesting entry into the critical section.• Critical Section Code in which only one process can execute atany one time.• Exit Section The end of the critical section, releasing or allowingothers in.• Remainder Section Rest of the code AFTER the critical section.7

CRITICAL SECTION PROBLEM• The critical section must ENFORCE ALL THREEof the following rules:• Mutual Exclusion: No more than one process canexecute in its critical section at one time.• Progress: If no one is in the critical section andsomeone wants in, then those processes not in theirremainder section must be able to decide in a finitetime who should go in.• Bounded Wait: All requesters must eventually be letinto the critical section8

CRITICAL SECTION PROBLEM9ENTRY SECTIONCRITICAL SECTIONEXIT SECTIONREMAINDER SECTION

PETERSON’S SOLUTION• To handle the problem of Critical Section (CS), Peterson gavean algorithm with a bounded waiting• Suppose there are N processes (P1, P2, … PN) and each ofthem at some point need to enter the Critical Section• A FLAG[] array of size N is maintained which is by default falseand whenever a process need to enter the critical section it hasto set its flag as true, i.e. suppose Pi wants to enter so it will setFLAG[i]=TRUE• There is another variable called TURN which indicates theprocess number which is currently to enter into the CS. Theprocess that enters into the CS while exiting would change theTURN to another number from among the list of readyprocesses10

PETERSON’S SOLUTION11• If turn is 2 (say) then P2 enters the CS and while exitingsets the turn as 3 and thus P3 breaks out of wait loopFLAG[i] = false

Synchronization Hardware• Problems of Critical Section are also solvable by hardware• Uniprocessor systems disables interrupts while a Process Pi isusing the CS but it is a great disadvantage in multiprocessorsystems• Some systems provide a lock functionality where a Processacquires a lock while entering the CS and releases the lockafter leaving it. Thus another process trying to enter CS cannotenter as the entry is locked. It can only do so if it is free byacquiring the lock itself• Another advanced approach is the Atomic Instructions (Non-Interruptible instructions).12

MUTEX LOCKS• As the synchronization hardware solution is not easy toimplement fro everyone, a strict software approach calledMutex Locks was introduced. In this approach, in theentry section of code, a LOCK is acquired over the criticalresources modified and used inside critical section, and inthe exit section that LOCK is released.• As the resource is locked while a process executes itscritical section hence no other process can access it13

SEMAPHORES• A more robust alternative to simple mutual exclusions is touse semaphores, which are integer variables for whichonly two ( atomic ) operations are defined, the wait andsignal operations, as shown in the following figure• As given here, these are not atomic as written in "macrocode". We define these operations, however, to be atomic(Protected by a hardware lock.)14WAIT ( S ):while ( S <= 0 );S = S - 1;SIGNAL ( S ):S = S + 1;

SEMAPHORES• FORMAT:wait ( mutex ); // Mutual exclusion: mutex init to 1.CRITICAL SECTIONsignal( mutex );REMAINDER• Note that not only must the variable-changing steps ( S-- andS++ ) be indivisible, it is also necessary that for the waitoperation when the test proves false that there be nointerruptions before S gets decremented. It IS okay, however,for the busy loop to be interrupted when the test is true, whichprevents the system from hanging forever15

SEMAPHORES• PROPERTIES1. Simple2. Works with many processes3. Can have many different critical sections with differentsemaphores4. Each critical section has unique access semaphores5. Can permit multiple processes into the critical section atonce, if desirable.16

SEMAPHORES• TYPESSemaphores are mainly of two types:1. Binary SemaphoreIt is a special form of semaphore used for implementing mutualexclusion, hence it is often called Mutex. A binary semaphore isinitialized to 1 and only takes the value 0 and 1 during execution ofa program.2. Counting SemaphoresThese are used to implement bounded concurrency.17

SEMAPHORES• Semaphores can be used to force synchronization (precedence ) if the preceding process does a signal at theend, and the follower does wait at beginning. For example,here we want P1 to execute before P2.P1: P2:statement 1; wait ( synch );signal ( synch ); statement 2;• To PREVENT looping, we redefine the semaphore structure as:typedef struct {int value;struct process *list; //linked list of PTBL waiting on S} SEMAPHORE;18

SEMAPHORES• It's critical that these be atomic - in uniprocessors we candisable interrupts, but in multiprocessors othermechanisms for atomicity are needed.19SEMAPHORE s;wait(s) {s.value = s.value - 1;if ( s.value < 0 ) {add this process to s.L;block;}}SEMAPHORE s;signal(s) {s.value = s.value + 1;if ( s.value <= 0 ) {remove a process P from s.L;wakeup(P);}}

SEMAPHORES• DEADLOCK• One important problem that can arise when using semaphores to blockprocesses waiting for a limited resource is the problem of deadlocks, whichoccur when multiple processes are blocked, each waiting for a resource thatcan only be freed by one of the other ( blocked ) processes, as illustrated inthe following example.20

SEMAPHORES• STARVATION• Another problem to consider is that of starvation, inwhich one or more processes gets blocked forever, andnever get a chance to take their turn in the critical section.For example, in the semaphores above, we did notspecify the algorithms for adding processes to the waitingqueue in the semaphore in the wait( ) call, or selectingone to be removed from the queue in the signal( ) call. Ifthe method chosen is a FIFO queue, then every processwill eventually get their turn, but if a LIFO queue isimplemented instead, then the first process to startwaiting could starve.21

THE BOUNDED-BUFFER PROBLEM• This is a generalization of the producer-consumerproblem wherein access is controlled to a shared group ofbuffers of a limited size.• In this solution, the two counting semaphores "full" and"empty" keep track of the current number of full and emptybuffers respectively ( and initialized to 0 and Nrespectively. ) The binary semaphore mutex controlsaccess to the critical section. The producer and consumerprocesses are nearly identical - One can think of theproducer as producing full buffers, and the consumerproducing empty buffers.22

THE BOUNDED-BUFFER PROBLEM23PRODUCERCONSUMER

THE READERS-WRITERS PROBLEM• THE READERS/WRITERS PROBLEM:• This is the same as the Producer / Consumer problem except - we nowcan have many concurrent readers and one exclusive writer.• Locks: are shared (for the readers) and exclusive (for the writer).• Two possible ( contradictory ) guidelines can be used:• No reader is kept waiting unless a writer holds the lock (the readers haveprecedence).• If a writer is waiting for access, no new reader gains access (writer hasprecedence).• ( NOTE: starvation can occur on either of these rules if they are followedrigorously.)24

THE READERS-WRITERS PROBLEM25Reader:do {wait( mutex ); /* Allow 1 reader in entry*/readcount = readcount + 1;if readcount == 1 then wait(wrt ); /* 1st reader locks writer */signal( mutex );/* reading is performed */wait( mutex );readcount = readcount - 1;if readcount == 0 then signal(wrt ); /*last reader frees writer */signal( mutex );} while(TRUE);Writer:do {wait( wrt );/* writing is performed */signal( wrt );} while(TRUE);THE READERS/WRITERS PROBLEM:BINARY_SEMAPHORE wrt = 1;BINARY_SEMAPHORE mutex = 1;int readcount = 0;

DINING PHILOSOPHERS PROBLEM26• Five philosophers sitting at a round diningtable to eat• Each one has got a plate and one chopstick atthe right side of each plate• To start eating each of them need a plate andtwo chopsticks• Thus, clearly there will be a deadlock andstarvation because of the limited number ofchopsticks• This can be solved by either:• Allow only 4 philosophers to be hungry ata time• Allow pickup only if both chopsticks areavailable. ( Done in critical section )• Odd # philosopher always picks up leftchopstick 1st, even # philosopher alwayspicks up right chopstick 1st

DINING PHILOSOPHERS PROBLEM• One possible solution, as shown in the following code section, is touse a set of five semaphores ( chopsticks[ 5 ] ), and to have eachhungry philosopher first wait on their left chopstick ( chopsticks[ i ] ),and then wait on their right chopstick ( chopsticks[ ( i + 1 ) % 5 ] )27

SYNCHRONIZATION EXAMPLES• SYNCHRONIZATION IN WINDOWS• SYNCHRONIZATION IN LINUX28

MONITORS• High-level synchronization construct that allows the safesharing of an abstract data type among concurrentprocesses.• Semaphores can be very useful for solving concurrencyproblems, but only if programmers use themproperly. If even one process fails to abide by the properuse of semaphores, either accidentally or deliberately,then the whole system breaks down.29

MONITORS• A monitor is essentially aclass, in which all data isprivate, and with the specialrestriction that only onemethod within any givenmonitor object may beactive at the same time.• An additional restriction isthat monitor methods mayonly access the shared datawithin the monitor and anydata passed to them asparameters, i.e. they cannotaccess any data external tothe monitor.30

MONITORS• In order to fully realize the potential of monitors, weneed to introduce one additional new data type,known as a condition.• A variable of type condition has only two legaloperations, wait and signal. I.e. if X was definedas type condition, then legal operations would beX.wait( ) and X.signal( )• The wait operation blocks a process until someother process calls signal, and adds the blockedprocess onto a list associated with that condition.• The signal process does nothing if there are noprocesses waiting on that condition. Otherwise itwakes up exactly one process from thecondition's list of waiting processes. ( Contrastthis with counting semaphores, which alwaysaffect the semaphore on a signal call. )31

ATOMIC TRANSACTIONS• Database operations frequently need to carry out atomictransactions, in which the entire transaction must eithercomplete or not occur at all.• The classic example is a transfer of funds, which involveswithdrawing funds from one account and depositing them intoanother - Either both halves of the transaction must complete,or neither must complete.• Operating Systems can be viewed as having many of the sameneeds and problems as databases, in that an OS can be saidto manage a small database of process-related information. Assuch, OSs can benefit from emulating some of the techniquesoriginally developed for databases.32

ATOMIC TRANSACTIONS• FUNDAMENTAL PRINCIPLES – A C I D• Atomicity – to outside world, transaction happensindivisibly• Consistency – transaction preserves system invariants• Isolated – transactions do not interfere with each other• Durable – once a transaction “commits,” the changes arepermanent33

REFERENCES• https://www.cs.uic.edu/~jbell/CourseNotes/OperatingSystems/5_Synchronization.html• http://www.studytonight.com/operating-system/process-synchronization34

Movatterモバイル変換

Change Language

Process synchronization in Operating Systems

Embed presentation

Recommended

More Related Content

What's hot

Viewers also liked

Similar to Process synchronization in Operating Systems

Recently uploaded

Process synchronization in Operating Systems