Mutual Exclusion -- Addendum

Mutual Exclusion -- Addendum

Mutual Exclusion in Critical Sections

RoadMap Today there are libraries that provide

application programmers with semaphores

Semaphores are used by programmers to provide for critical sections

Let’s first talk about semaphores and then the OS support for them.

What is a semaphore?A semaphore is an integer variable with the

following three operations.1. Initialize: You can initialize the semaphore to

any non-negative value.2. Decrement: The down operation

• Semaphore value > 0: Decrement by 1. • Semaphore value == 0: If value is 0, the process blocks

(gets put into queue). It is said to be sleeping on the semaphore..

3. Increment: The up operation.• Semaphore value == 0 and some processes are

sleeping on the semaphore, one is unblocked. Otherwise, value is incremented.

What is a Semaphore? Use down before entering a critical

section Use up after finishing with a critical

section i.e., Example: Assume S is initialized to 1. ……… down(S); //critical section up(S);

Using SemaphoresProcess P0

………………..down(S); //critical sectionup(S);………………..

Process P1

………………..down(S); //critical sectionup(S);………………..

Using SemaphoresProcess P0 Process P1

down(S); down(S); critical section critical sectionup(S); up(S); Initialize the semaphore variable, S, to 1

Why not zero?

Using SemaphoresProcess P0 Process P1

down(S); down(S); critical section critical sectionup(S); up(S); Now what would happen if P0 executes the

down operation? The semaphore S is currently 1. It becomes 0 and P0 enters the critical section

Using SemaphoresProcess P0 Process P1

down(S); down(S); critical section critical sectionup(S); up(S); Now what would happen if P1 executes the

down operation? The semaphore S is currently 0 P1 blocks

Using SemaphoresProcess P0 Process P1

down(S); down(S); critical section critical sectionup(S); up(S); Assume now that P0 is done with the critical section

It calls the up function P1 is unblocked If there was no process waiting to enter the

critical section the value of s would become one

Using Semaphores What happens if there are three processes:

P0,P1,P2 Assume P0 enters its critical section If P1 and P2 execute the down operation they

will block When P0 leaves the critical section then P1 is

unblocked allowing P1 to enter its critical section P2 is still blocked

What if P0 wants to enter its critical section again when P1 is in it?

Using Semaphores What if we want to allow 10 processes

to use a critical section? How would we initialize a semaphore, s?

Semaphore Binary Semaphore: Value is either 0 or 1

Often referred to as a mutex Counting Semaphore: Value ranges from

0 to N where N can be any integer number

Deadlock Deadlock – Two or more processes are

waiting indefinitely for an event that can be caused by only one of the waiting processes

Something to watch for

Deadlock Example: Let S and Q be two

semaphores initialized to oneProcess P0 Process P1

down(S); down(Q); down(Q); down(S); …… …. up(S); up(Q); up(Q); up(S);

Implementation 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

Implementation A semaphore can be defined as a C

struct along these lines:

typedef struct { int value; struct process *list; } semaphore

Implementation down() operation can be defined as

down(semaphore *S) { S->value--; if (S->value < 0) {

add this process to S->list;

block(); }

} The block() operation suspends the

process that invokes it.

Implementation up() operation can be defined asup(semaphore *S) {

S->value++; if (S->value <= 0) {

remove process P from S->list;

wakeup(P); }

} The wakeup() operation sends a signal

that represents an event that the invoking process is no longer in the critical section

Implementation BTW, the implementation just described

is how Linux implements semaphores The up and down operations represent

require access to a critical section which is the semaphore variable

Need hardware/OS support e.g., Hardware support e.g., TSL. Signals

Synchronization Hardware Any solution to the

critical section problem requires a lock

Process must acquire a lock before entering a critical section

Process must release the lock when it exists the critical section.

We will present a hardware solution

while (true) { acquire lock critical section release lock other}

Test and Lock Instruction (TSL)

Many computers have the following type of instruction: TSL REGISTER, LOCK

One use of this instruction is to provide a lock for a critical region such code that operations on a semaphore variable

Using the TSL Instruction

TSLenter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller TSL Register, Lock

Reads LOCK into register REGISTER Stores a nonzero value at the memory

location LOCK The operations of reading the word and

storing it are guaranteed to be indivisible

TSLenter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller TSL Register, Lock

The CPU executing the TSL instruction locks the memory bus to prohibit other CPUs from accessing memory until the TSL instruction is done

TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller CMP Register, #0

Compare register value with 0 JNE Enter_Region

If Register value is not 0 then go to the start of enter region

TSLenter_region: leave region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller Return to caller only if Register value is 0

TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller Before entering its critical region, a process calls enter_region

What if LOCK is 1? Busy wait until lock is 0

When leaving the critical section, a process calls leave_region

Using TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller Assume two processes: P0 and P1

LOCK is initialized to zero

Using TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller Assume that P0 wants to enter the critical section

It executes the TSL instruction. The register value is 0 which reflects the

value of LOCK LOCK is set to 1

Using TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller

Now P1 wants to enter the critical section; It executes the TSL instruction The register value is 1 which reflects the

value of LOCK P1 cannot enter the critical section It repeats the TSL instruction and

comparison operation until it can get into the critical section

Using TSL enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller P0 is done with the critical section

LOCK becomes 0

Using TSL Enter_region: leave_region: TSL Register, Lock move Lock, #0 CMP Register, #0 RET JNE enter_region RET //to caller The next time P1 executes the TSL instruction and comparison operation it finds that the register value (which reflects LOCK) is zero. It can now enter the critical section.

Implementing TSL Implementing atomic TSL instructions

on multiprocessors is not trivial. This is a subject for a computer

architecture course

Signals A signal is used in UNIX systems to

notify a process that a particular event has occurred A signal is generated by the occurrence of a

particular event A generated signal is delivered to a process Once delivered, the signal must be handled

A signal process is used to associate computation with a signal

Signals Example

<control> <C> is entered This causes an event to be generated to a

running process that is in the foreground When that process receives the event it

executes a signal handler which terminates the process

Signals Two possible handlers

Default signal handler: Provided by kernel User-defined signal handler: Provided by

user and overrides the default What if a process has multiple threads?

How a signal is handled depends on the type of signal. Options• Deliver the signal to the thread to which the signal

applies• Deliver the signal to every thread in the process• Deliver the signal to certain threads in the process• Assign a specific threa to receive all signals for the

process

Semaphore Implementation When the up is executed a blocked

process is woken up. This is done using signals

Semaphore operations are critical sections – use TSL.

Questions How are multiple processes prevented

from being in the critical section ?

Why different than disabling interrupts?

Which is better in a multicore system? Disabling interrupts or TSL test?

Question Assume that instead of a TSL instruction

there is an instruction to swap the contents of a register and memory word in a single indivisible action. Can you write a enter_region routine based

on this

Mars PathFinder Priority Inversion: Scheduling problem

when lower-priority process holds a lock needed by higher-priority process

Now back to the Mars Pathfinder problem High priority task was taking longer than

expected to complete its work The task was forced to wait for a shared

resource that was being held by a lower-priority process

Lower-priority process was being pre-empted by a medium priority process

Scheduler detected problem and would reset

Mars PathFinder The fix

The OS had a global variable to enable priority inheritance on all semaphores

It was off It was set to on and then everything worked

Summary Defined race condition Examined different solutions