Modified from Silberschatz, Galvin and Gagne & Stallings Lecture 10 Chapter 6: Process Synchronization

Modified from Silberschatz, Galvin and Gagne & Stallings

Lecture 10

Chapter 6: Process Synchronization

2CS 446/646 Principles of Computer Operating Systems

Chapter 6: Process Synchronization

Background

The Critical-Section Problem

Peterson’s Solution

Synchronization Hardware

Semaphores

Classic Problems of Synchronization

Monitors

Synchronization Examples

Atomic Transactions


Objectives

To introduce the critical-section problem,

whose solutions can be used to ensure the consistency of shared data

To present both software and hardware solutions of the critical-section problem

To introduce the concept of an atomic transaction

describe mechanisms to ensure atomicity


Background

Concurrent access to shared data may result in data inconsistency concurrency is a fundamental part of O/S design concurrency includes

communication among processes/threads sharing of, and competition for system resources cooperative processing of shared data synchronization of process/thread activities organized CPU scheduling solving deadlock and starvation problems

Maintaining data consistency requires mechanisms to ensure the orderly execution of cooperating processes


Concurrency can be viewed at different levels:

multiprogramming — interaction between multiple processes running on one CPU (pseudo-parallelism)

multithreading — interaction between multiple threads running in one process

multiprocessors — interaction between multiple CPUs running multiple processes/threads (real parallelism)

multicomputers — interaction between multiple computers running distributed processes/threads

the principles of concurrency are basically the same in all of these categories

Process Interaction & ConcurrencyS

oft

ware

vie

wH

ard

ware

vie

w


o processes unaware of each other they must use shared resources independently,

without interfering, and leave them intact for the others

P1

P2

resource

o processes indirectly aware of each other they work on common data and build some result

together via the data

P2

P1

data

o processes directly aware of each other they cooperate by communicating, e.g.,

exchanging messagesP2

P1

messages

Whether processes or threads: three basic interactions

Process Interaction & Concurrency


Multithreaded shopping diagram and possible outputs

> ./multi_shoppinggrabbing the salad...grabbing the milk...grabbing the apples...grabbing the butter...grabbing the cheese...>

> ./multi_shoppinggrabbing the milk...grabbing the butter...grabbing the salad...grabbing the cheese...grabbing the apples...>

Molay, B. (2002) UnderstandingUnix/Linux Programming (1st Edition).

Insignificant race condition in the shopping scenario there is a “race condition” if the outcome depends on the order of the execution

Race Condition


CPU

CPU

Insignificant race condition in the shopping scenario the outcome depends on the CPU scheduling or “interleaving” of the threads

(separately, each thread always does the same thing)


A

B

salad

apples

milk

butter

cheese

> ./multi_shoppinggrabbing the milk...grabbing the butter...grabbing the salad...grabbing the cheese...grabbing the apples...>

A salad

apples

B

milk

butter

cheese

Race Condition


Insignificant race condition in the shopping scenario the CPU switches from one process/thread to another, possibly on the basis of a

preemptive clock mechanism


A

B

salad

apples

milk

butter

cheese

CPU

thread A

thread B

salad

milk

apples

butter cheese

Thread view expanded in real execution time

Race Condition


char chin, chout;

void echo(){ do { chin = getchar(); chout = chin; putchar(chout); } while (...);}

A

char chin, chout;


B

> ./echoHello world!Hello world!

Single-threaded echo Multithreaded echo (lucky)


1

23

4

56

lucky

CPU

scheduling

Significant race conditions in I/O & variable sharing

Race Condition


char chin, chout;


A


Single-threaded echo

char chin, chout;


B


1

56

2

34

unlucky

CPU

scheduling

Multithreaded echo (unlucky)

> ./echoHello world!ee....

Race Condition


void echo(){ char chin, chout;

do { chin = getchar(); chout = chin; putchar(chout); } while (...);}

B

void echo(){ char chin, chout;

do { chin = getchar(); chout = chin; putchar(chout); } while (...);}

A


Single-threaded echo


1

56

2

34

unlucky

CPU

scheduling

Multithreaded echo (unlucky)

> ./echoHello world!eH....

changedto local

variables

Race Condition


Significant race conditions in I/O & variable sharing in this case, replacing the global variables with local variables did not solve the problem

we actually had two race conditions here: one race condition in the shared variables and the order of value assignment another race condition in the shared output stream:

• which thread is going to write to output first• this race persisted even after making the variables local to each thread

generally, problematic race conditions may occur whenever resources and/or data are shared by processes unaware of each other or processes indirectly aware of each other

Race Condition


Producer / Consumer Problem

Suppose that we wanted to provide a solution to the consumer-producer problem that fills all the buffers.

We can do so by having an integer count that keeps track of the number of full buffers.

Initially, count is set to 0.

incremented by the producer after it produces a new buffer,

decremented by the consumer after it consumes a buffer.

http://gaia.ecs.csus.edu/~zhangd/oscal/ProducerConsumer/ProducerConsumer.html

while (true) {

/* Produce an item */

while (((in = (in + 1) % BUFFER SIZE) == out)

; /* do nothing -- no free buffers */

buffer[in] = item;

in = (in + 1) % BUFFER SIZE;

}

while (true) {

while (in == out)

; // do nothing -- nothing to consume

// remove an item from the buffer

item = buffer[out];

out = (out + 1) % BUFFER SIZE;

return item;

}


Producer / Consumer while (true) {

/* produce an item and put in nextProduced */

while (count == BUFFER_SIZE)

; // do nothing

buffer [in] = nextProduced;

in = (in + 1) % BUFFER_SIZE;

count++;

}

while (true) {

while (count == 0)

; // do nothing

nextConsumed = buffer[out];

out = (out + 1) % BUFFER_SIZE;

count--;

/* consume the item in nextConsumed

}


Race Condition count++ could be implemented as

register1 = count register1 = register1 + 1 count = register1

count-- could be implemented as

register2 = count register2 = register2 - 1 count = register2

Consider this execution interleaving with “count = 5” initially:

S0: producer execute register1 = count {register1 = 5}S1: producer execute register1 = register1 + 1 {register1 = 6} S2: consumer execute register2 = count {register2 = 5} S3: consumer execute register2 = register2 - 1 {register2 = 4} S4: producer execute count = register1 {count = 6 } S5: consumer execute count = register2 {count = 4}


How to avoid race conditions? find a way to keep the instructions together this means actually. . . reverting from too much interleaving and going back to

“indivisible/atomic” blocks of execution!!

thread A

thread B

chin='H'

chin='e'

putchar('e')

chout='e' putchar('e')

(a) too much interleaving may create race conditions

(b) keeping “indivisible” blocks of execution avoids race conditions

thread A

thread B

chin='H'

chin='e'

putchar('H')

chout='e' putchar('e')

Race Condition


The “indivisible” execution blocks are critical regions a critical region is a section of code that may be executed by only one process or

thread at a time

B

Acommon critical region

B

A A’s critical region

B’s critical region

although it is not necessarily the same region of memory or section of program in both processes

but physically different or not, what matters is that these regions cannot be interleaved or executed in parallel (pseudo or real)

Critical Regions


enter critical region?

exit critical region

enter critical region?

exit critical region

critical regions can be protected from concurrent access by padding them with entrance and exit gates o a thread must try to check in, then it must check out

We need mutual exclusion from critical regions

void echo(){

char chin, chout; do {

chin = getchar(); chout = chin; putchar(chout);

} while (...);}

BA

void echo(){

char chin, chout; do {

chin = getchar(); chout = chin; putchar(chout);

} while (...);}

Critical Regions


Critical Section

do {

entry section

critical section

exit session

remainder section

} while (TRUE);


1. thread A reaches the gate to the critical region (CR) before B

2. thread A enters CR first, preventing B from entering (B is waiting or is blocked)

3. thread A exits CR; thread B can now enter

4. thread B enters CR

Desired effect: mutual exclusion from the critical region

critical regionB

A

B

A

B

A

B

A

HO

W i

s th

isac

hie

ved

??

Mutual Exclusion


Solution to Critical-Section Problem

1. Mutual Exclusion

o If process Pi is executing in its critical section,

o then no other processes can be executing in their critical sections

2. Progresso If no process is executing in its critical section and there exist some

processes that wish to enter their critical section,o then the selection of the processes that will enter the critical section

next cannot be postponed indefinitely

3. Bounded Waitingo A bound must exist on the number of times that other processes are

allowed to enter their critical sections after a process has made a request to enter its critical section and before that request is granted

Assume that each process executes at a nonzero speed

No assumption concerning relative speed of the N processes



Two process solution

Assume that the LOAD and STORE instructions are atomic;

that is, cannot be interrupted.

The two processes share two variables:

int turn;

Boolean flag[2]

The variable turn indicates whose turn it is to enter the critical section.

The flag array is used to indicate if a process is ready to enter the critical section.

flag[i] = true implies that process Pi is ready!


Algorithm for Process Pi

do {

flag[i] = TRUE;

turn = j;

while (flag[j] && turn == j);

critical section

flag[i] = FALSE;

remainder section

} while (TRUE);


1. A and B each have their own lock; an extra toggle is also masking either lock

2. A arrives first, sets its lock, pushes the mask to the other lock and may enter

3. then, B also sets its lock & pushes the mask, but must wait until A’s lock is reset

4. A exits the CR and resets its lock; B may now enter

critical regionB

A

B

A

B

A

B

A



1. A and B each have their own lock; an extra toggle is also masking either lock

2.1 A is interrupted between setting the lock & pushing the mask; B sets its lock

2.2 now, both A and B race to push the mask: whoever does it last will allow the other one inside CR

mutual exclusion holds!! (no bad race condition)

critical regionB

A

B

A

B

A

pushed last, allowing A

B

A pushed last, allowing B



Synchronization Hardware

Uniprocessors – could disable interrupts Currently running code would execute without preemption what guarantees that the user process is going to ever exit the critical

region? meanwhile, the CPU cannot interleave any other task

even unrelated to this race condition Generally too inefficient on multiprocessor systems Operating systems using this not broadly scalable

Modern machines provide special atomic hardware instructions

Atomic = non-interruptable

Either test memory word and set value

Or swap contents of two memory words


Solution to Critical-section Problem Using Locks

do {

acquire lock

critical section

release lock

remainder section

} while (TRUE);

Many systems provide hardware support for critical section code


1. thread A reaches CR and finds the lock at 0 and sets it in one shot, then enters

1.1’ even if B comes right behind A, it will find that the lock is already at 1

2. thread A exits CR, then resets lock to 0

3. thread B finds the lock at 0 and sets it to 1 in one shot, just before entering CR

critical regionB

A

B

A

B

A

B

A

Atomic lock


TestAndSet Instruction

Definition:

boolean TestAndSet (boolean *target)

{

boolean rv = *target;

*target = TRUE;

return rv:

}


Solution using TestAndSet

Shared boolean variable lock., initialized to false. Solution:

do {

while ( TestAndSet (&lock ))

; // do nothing

// critical section

lock = FALSE;

// remainder section

} while (TRUE);


Swap Instruction

Definition:

void Swap (boolean *a, boolean *b)

{

boolean temp = *a;

*a = *b;

*b = temp:

}


Solution using Swap

Shared Boolean variable lock initialized to FALSE; Each process has a local Boolean variable key

Solution:

do {

key = TRUE;

while ( key == TRUE)

Swap (&lock, &key );

// critical section

lock = FALSE;


} while (TRUE);


Bounded-waiting Mutual Exclusion with TestandSet()

do {

waiting[i] = TRUE;

key = TRUE;

while (waiting[i] && key)

key = TestAndSet(&lock);

waiting[i] = FALSE;

// critical section

j = (i + 1) % n;

while ((j != i) && !waiting[j])

j = (j + 1) % n;

if (j == i)

lock = FALSE;

else

waiting[j] = FALSE;


} while (TRUE);


Semaphore

Synchronization tool that does not require busy waiting

Semaphore S – integer variable

Two standard operations modify S: wait() and signal()

Less complicated

Can only be accessed via two indivisible (atomic) operations

wait (S) {

while S <= 0

; // no-op

S--;

}

signal (S) {

S++;

}


Semaphore as General Synchronization Tool

Counting semaphore integer value can range over an unrestricted domain

Binary semaphore integer value can range only between 0 and 1; can be simpler to implement Also known as mutex locks

Can implement a counting semaphore S as a binary semaphore

Provides mutual exclusion

Semaphore mutex; // initialized to 1

do {

wait (mutex);

// Critical Section

signal (mutex);


} while (TRUE);


value = 1 (“off”)no queue

signal signal signal signal

signal

value = 01 in queue

wait

value = 02 in queue

wait

. . .

value = 03 in queue

wait

value = 0 (“on”)no queue

wait

Binary Semaphore


value = +2no queue

value = –11 in queue

wait

0

. . .

value = –22 in queue

wait

00

value = 0no queue

wait

0

value = +1no queue

wait

0

signal signal signal signal

Counting Semaphore


Semaphore Implementation

Must guarantee that no two processes can execute wait () and signal () on the same semaphore at the same time

Thus, implementation becomes the critical section problem where the wait and signal code are placed in the critical section.

Could now have busy waiting in critical section implementation

But implementation code is short

Little busy waiting if critical section rarely occupied

Note that applications may spend lots of time in critical sections and therefore this is not a good solution.


in busy waiting, the PC is always looping (increment & jump back)

it can be preemptively interrupted but will loop again tightly whenever rescheduled tight polling Blocke

d

Ready

dispatch

event wait(block)

event occurs(unblock)

timeout

Running

when blocked, the process’s PC stalls after executing a “yield” call

either the process is only timed out, thus it is “Ready” to loop-and-yield again sparse polling

or it is truly “Blocked” and put in event queue condition waiting

Running

dispatch

voluntaryevent wait

(block)

event occurs(unblock)

voluntarytimeout

Blocked

Ready

Busy waiting


Semaphore Implementation with no Busy waiting

With each semaphore there is an associated waiting queue.

Each entry in a waiting queue has two data items:

value (of type integer)

pointer to next record in the list

Two operations:

Block

place the process invoking the operation on the appropriate waiting queue.

wakeup

remove one of processes in the waiting queue and place it in the ready queue.


Semaphore Implementation with no Busy waiting

Implementation of wait:

wait(semaphore *S) {

S->value--;

if (S->value < 0) {

add this process to S->list;

block();

}

}

Implementation of signal:

signal(semaphore *S) {

S->value++;

if (S->value <= 0) {

remove a process P from S->list;

wakeup(P);

}

}


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

P0 P1

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

Priority Inversion - Scheduling problem when lower-priority process holds a lock needed by higher-priority process

Documents

Modified from Silberschatz, Galvin and Gagne & Stallings Lecture 10 Chapter 6: Process Synchronization