Docstoc

Processes - PowerPoint

Document Sample
Processes - PowerPoint Powered By Docstoc
					Outline
   Processes
   Threads
   Inter-process communication (IPC)
   Classical IPC problems
   Scheduling




                                        1
    Issues about IPC
   Communication between processes is
    indispensable.
        Interrupt is not preferred. Why?
        How about pipe?
    Basic issues
        How to pass information
        How to solve the competition for resources
         (mutual exclusion)
        How to guarantee the dependency
         (synchronization)

                                                      2
             Sharing Resources

                      Var in A   Var in B                            Spooler
                                                  The next
                                                                     directory
A: reads ininA       inA=7                      file to print           .
                                                                         .
                                                                         .
B: reads ininB       inA=7      inB=7      Process A
                                                                 4      abc      out = 4
B: Put file in slot                                              5    prog.c
7                     inA=7      inB=8
Update in8                                                      6    prog.n
A: Put file in slot                         Process              7               in = 7
7                     inA=8      inB=8         B                         .
Update in8
                                                                         .
                                             The next                    .
 A’s file overwrites B’s!!!                   free slot


                                                                                     3
    Race Conditions
   Two or more processes are reading or writing
    some shared data, the final result depends on
    who/when runs precisely
        Must be avoided
   Solution: mutual exclusion
        Critical region/section: the part of program where
         the shared resource is accessed
        No two processes in their critical regions at the
         same time



                                                         4
Conditions for Sharing Resource
   No two processes simultaneously inside
    their critical regions (mutual exclusion)
   No assumption about speeds or number
    of CPUs
   No process running outside its critical
    region may block other processes
   No process waits forever to enter its
    critical region (no deadlock)

                                           5
      Mutual Exclusion Using Critical Regions

              A enters                 A leaves
              critical region          critical region
Process A
                     B attempts
                     to enter      B enters B leaves
                     critical      critical critical
                     region        region   region
Process B
                     B blocked

             T1     T2            T3         T4


                         Time
                                                         6
Achieve Mutual Exclusion
   Busy waiting
   Sleep and wakeup
   Semaphores
   Mutexes
   Monitors
   Message passing
   Barriers

                           7
Busy Waiting – Disabling Interrupts

   Enter critical region  disable interrupts
     …  enable interrupts  leave
    critical region
       Simplest, intuitive
       Why disable interrupt? No process
        switching in critical region
   Problem: lose control of user processes
   Not work in the case of multiple CPUs
   Good for kernel.

                                            8
    Busy Waiting – Lock Variables
   Disable interrupt is not good.
        Any software solution?
   Lock variable: single, shared variable
        Before entering critical region, test and set
        Race condition on the lock variable
             After one process reads the lock but before it sets to 1,
              switching happens.
   Problem: test and set lock variable must be
    inseparable, i.e., atomic
        Ask for help from hardware

                                                                     9
      TSL: Help from Hardware
        Problem of lock variables
            Test, if 0, enter, else wait
            What if both process test 0 and enter?
        TSL instruction: TSL RX, LOCK
            Read, update and write-back
            All three operations are indivisible

Enter_region:                Leave_region:
  TSL REG, LOCK                MOV LOCK, #0
  CMP REG, #0                  RET
  JNE Enter_region
  RET
                                                      10
     Busy Waiting: Strict Alternation
// Initially turn=0
While (TRUE){                 While (TRUE){
  while (turn!=0); /*loop*/     while (turn!=1); /*loop*/
  critical_region();            critical_region();
  turn=1;                       turn=0;
  non_critical_region();        non_critical_region();
}                             }
      Process 0                     Process 1

                               A process may be blocked
Process 0 and 1                by another one not in its
use critical region            critical region! It requires
alternatively in a             strictly alternate in using
strict order!                  critical regions.
         Get Rid of Strict Alternation
        Peterson’s solution: combine taking turns with the
         idea of lock variables
        Two procedures
             enter_region and leave_region
        For each process
             Call enter_region before entering its critical region
             Call leave_region after leaving its critical region

Process 0:                               Process 1:
……                                       ……
Enter_region(0);                         Enter_region(1);
critical_part;                           critical_part;
Leave_region(0);                         Leave_region(1);
…..                                      …..                          12
      C Code of Peterson’s Solution
 int turn;          // shared variable. whose turn is it?
 int interested[2]; //shared variable. all values initially FALSE
 Void enter_region(int process) //parameter: 0 or 1
 {
    int other;           //local variable. number of the other process
    other=1-process; //the opposite of process
1 interested[process]=TRUE; //show that you are interested
2 turn=process;              //set flag
3 while (turn==process && interested[other]==TRUE);
 }


Void leave_region(int process) //process: who is leaving?
{
  interested[process]=FALSE //departure from critical region
}
void enter_region(int 0) //process=0           void enter_region(int 1) //process=1
{                                              {
    int other; //local variable                    int other; //local variable
    other=1-process=1;                             other=1-process=0;

1   interested[0]=TRUE;                       1    interested[1]=TRUE;

2   turn=0;                                   2    turn=1;

3   while (turn==0 && interested[1]==TRUE);
                                              3    while (turn==1 && interested[0]==TRUE);
}                                              }

…….                          ……..
P0:1                         interested[0]=TRUE
P1:1                         interested[1]=TRUE
P0:2                         turn=0
P1:2                         turn=1
P0:3                         While( turn==0 && interested[1]==TRUE) entering!
P1:3                         While( turn==1 && interested[0]==TRUE) busy waiting!
    Problems of Busy Waiting
   Waste CPU time
   Some processes may never get chance to run
        Priority inversion problem
        E.g., L in critical region, while H is busy waiting
   Solution: block instead of waiting
        sleep and wakeup operations
        sleep(): cause the caller to block, until another
         process wakes it up
        wakeup(pid): wake up the process specified by pid


                                                               15
     Producer-Consumer Problem
       Elements: producer, consumer, buffer
       Producer: put items into buffer
       Consumer: take items out of buffer
       Buffer: fixed size holder of items

producer

   items            buffer        items

                                     consumer
                                            16
#define N 100                                 // number of slots in the buffer
int count = 0;                                // number of items in the buffer

void producer(void){
  int item;

    while (TRUE){                             // repeat forever
      item = produce_item();                  // generate next item
      if (count == N) sleep();                // if buffer is full, go to sleep
      insert_item(item);                      // put item into buffer
      count = count+1;                        // increment count of items in buffer
      if (count == 1) wakeup(consumer);       // was buffer empty?
    }
}
                                 Switch to producer.
void consumer(void){              Lost-wakeup call!
  int item;

    while (TRUE){                             // repeat forever
      if (count == 0)    sleep();             // if buffer is empty, go to sleep
      item = remove_item();                   // take item out of buffer
      count = count-1;                        // decrement count of items in buffer
      if (count == N – 1) wakeup(producer);   // was buffer full?
      consume_item(item);                     // print item
    }
}
    Problem: Lost-Wakeup
   Race condition on the shared variable: count
        Buffer is empty and consumer reads the counter
         as value 0. But before consumer can make the
         sleep() call, CPU switches to producer. Producer
         inserts an item, sets counter to 1 and sends a
         wakeup call to consumer. However, consumer is
         not sleeping yet, the wakeup signal is lost!
        Now CPU switches back to consumer and it finally
         makes the sleep() call- consumer is asleep now.
        Producer will fill up the buffer sooner or later, and
         it goes to sleep as well.
        Both are sleeping!

                                                           18
Semaphores
   Two atomic operations: down and up
       Suppose s is a semaphore
       down(&s): s--; if s < 0 then sleep
       up(&s): s++; if s <= 0 then wake up one
        of the processes blocked by down
        operation on s.
            up operation never blocked!
       What’s the value of s now? Don’t know
            Can only be accessed by the two operations
             after initialization.

                                                          19
Atomic Operations
   Both down and up are as a single,
    indivisible atomic actions
   Value < 0 indicates some process
    sleeping
   Solve the lost-wakeup problem (why?)
   How to implement these atomic
    operations?
       Disable all interrupts
       Use TSL instruction

                                       20
    Use of Semaphores
   Mutual exclusion
        Semaphore s is initialized to 1.
        Call down(&s) before enter critical region
        Call up(&s) after leave critical region
        Can be used for buffer in Producer-Consumer
         problem
   Synchronization
        Guarantee certain event sequences do or do not
         occur
        E.g., the Producer stops running when the buffer
         is full.

                                                       21
           Producer-Consumer using Semaphores
#define N 100                 // number of slots in the buffer
semaphore mutual = 1;         // controls access to critical region
semaphore empty = N;          // counts empty buffer slots
semaphore full = 0;           // counts full buffer slots
void producer(void){
  int item;
  while (TRUE){
     item = produce_item();   // generate something to put in buffer
     down(&empty);            // decrement empty count
     down(&mutual);           // enter critical region
     insert_item(item);       // put new item in buffer
     up(&mutual);             // leave critical region
     up(&full);               // increment count of full slots
  }
}
void consumer(void){
  int item;
  while (TRUE){
     down(&full);             // decrement full count
     down(&mutual);           // enter critical region
     item = remove_item();    // take item from buffer
     up(&mutual);             // leave critical region
     up(&empty);              // increment count of empty slots
  }
}
    Mutexes: Binary Semaphores

   Achieve mutual exclusion
   Two states: unlocked (0) / locked (1)
   Two operations: mutex_lock(&m) and
    mutex_unlock(&m)
   For a process
        Call mutex_lock before entering critical region
        Call mutex_unlock after leaving critical region
   Can be implemented in user space with TSL
    instruction.

                                                           23
      Implementation of Mutex
mutex_lock:
  TSL REGISTER, MUTEX   | copy mutex to register and set mutex to 1
  CMP REGISTER, #0      | was mutex zero?
  JZE ok                | if it was zero, mutex was unlocked, so return
  CALL thread_yield     | mutex is busy; schedule another thread
  JMP mutex_lock        | try again latter
ok: RET                 | return to caller; critical region entered

mutex_unlock:
 MOV MUTEX, #0          | store a zero in mutex
 RET                    | return to caller



  Notice: thread_yield is used. Busy waiting does not
  work here. (Why?)
Problems with Mutual Exclusion
   Diskette and CD-Rom protected by
    mutexes
       Process A: got diskette, waiting CD-Rom
       Process B: got CD-Rom, waiting diskette
       A deadlock! Blocked forever
   Programming with semaphores or
    mutexes is tricky


                                                  25

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:17
posted:11/1/2011
language:English
pages:25