Difference between revisions of "Sezioni Critiche in C"

From Sistemi Operativi
Jump to navigation Jump to search
m
 
(One intermediate revision by one other user not shown)
Line 28: Line 28:
 
     /* enter */
 
     /* enter */
 
     need[id] = 1;
 
     need[id] = 1;
 +
    __sync_synchronize();
 
     while (need[1 - id]) {
 
     while (need[1 - id]) {
 +
      __sync_synchronize();
 
       if (turn != id) {
 
       if (turn != id) {
 
         need[id] = 0;
 
         need[id] = 0;
Line 35: Line 37:
 
         need[id] = 1;
 
         need[id] = 1;
 
       }
 
       }
 +
      __sync_synchronize();
 
     }
 
     }
 
     /* /enter */
 
     /* /enter */
Line 53: Line 56:
 
   pthread_t t[NPROC];
 
   pthread_t t[NPROC];
 
   srand(time(NULL));
 
   srand(time(NULL));
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_create(&t[i], NULL, codet, (void *) i);
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_join(t[i], NULL);
 +
  printf("%ld %d\n",val,NPROC*MAX);
 +
}
 +
</syntaxhighlight>
 +
 +
== Peterson ==
 +
Proviamo con Peterson
 +
<syntaxhighlight lang=C>
 +
#include<stdio.h>
 +
#include<stdint.h>
 +
#include<pthread.h>
 +
#include<unistd.h>
 +
 +
#define MAX 4000000
 +
#define STEP 1000000
 +
#define NPROC 2
 +
 +
volatile long val = 0;
 +
volatile long need[NPROC];
 +
volatile long turn;
 +
 +
void *codet(void *arg) {
 +
  uintptr_t id = (uintptr_t) arg;
 +
  int count;
 +
  for (count=0; count  < MAX; count++) {
 +
    /* enter */
 +
    need[id] = 1;
 +
    turn = 1 - id;
 +
    __sync_synchronize();
 +
    while (need[1 - id] && turn != id)
 +
      ;
 +
    /* /enter */
 +
    val = val + 1;
 +
    if (count % STEP == 0)
 +
      printf("%d %ld\n",id,val);
 +
    /* exit */
 +
    need[id] = 0;
 +
    /* /exit */
 +
  }
 +
}
 +
 +
int main(int argc, char *argv[]) {
 +
  uintptr_t i;
 +
  pthread_t t[NPROC];
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_create(&t[i], NULL, codet, (void *) i);
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_join(t[i], NULL);
 +
  printf("%d %d\n",val,NPROC*MAX);
 +
}
 +
</syntaxhighlight>
 +
La funzione del compilatore C "__sync_synchronize" (che nella macchine Intel genera l'istuzione assembler "__asm__ __volatile__("mfence");") &egrave; necessaria per x86 multicore perch&eacute; altrimenti le cache dei processori potrebbero non essere allineate.
 +
 +
== Peterson per N processi ==
 +
<syntaxhighlight lang=C>
 +
#include<stdio.h>
 +
#include<stdint.h>
 +
#include<pthread.h>
 +
#include<unistd.h>
 +
 +
#define MAX 4000000
 +
#define NPROC 4
 +
 +
volatile long val = 0;
 +
volatile long turn;
 +
volatile long stage[NPROC];
 +
volatile long last[NPROC];
 +
 +
void *codet(void *arg) {
 +
  uintptr_t id = (uintptr_t) arg;
 +
  int count;
 +
  int j, k;
 +
  for (count=0; count  < MAX; count++) {
 +
    /* enter */
 +
    for (j=0; j<NPROC; j++) {
 +
      stage[id] = j+1; last[j] = id;
 +
      __sync_synchronize();
 +
      for (k=0; k<NPROC; k++) {
 +
        if (id != k)
 +
          while (stage[k] >= stage[id] && last[j] == id)
 +
            ;
 +
      }
 +
    }
 +
    /* /enter */
 +
    val = val + 1;
 +
    if ((count & ((1<<20) - 1)) == 0)
 +
      printf("%d %ld\n",id,val);
 +
    /* exit */
 +
    stage[id] = 0;
 +
    /* /exit */
 +
  }
 +
}
 +
 +
int main(int argc, char *argv[]) {
 +
  uintptr_t i;
 +
  pthread_t t[NPROC];
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_create(&t[i], NULL, codet, (void *) i);
 +
  for (i=0; i<NPROC; i++)
 +
    pthread_join(t[i], NULL);
 +
  printf("%d %d\n",val,NPROC*MAX);
 +
}
 +
</syntaxhighlight>
 +
 +
== Test&Set ==
 +
 +
Ora proviamo con test&set (che viene fornito dal compilatore C)
 +
<syntaxhighlight lang=C>
 +
#include<stdio.h>
 +
#include<stdint.h>
 +
#include<pthread.h>
 +
#include<unistd.h>
 +
#include<sched.h>
 +
 +
#define MAX 4000000
 +
#define STEP 1000000
 +
#define NPROC 4
 +
 +
volatile long val = 0;
 +
volatile long lock = 0;
 +
 +
void *codet(void *arg) {
 +
  uintptr_t id = (uintptr_t) arg;
 +
  int count;
 +
  for (count=0; count  < MAX; count++) {
 +
    /* enter */
 +
    while (__sync_lock_test_and_set(&lock,1))
 +
      /*sched_yield()*/;
 +
    /* /enter */
 +
    val = val + 1;
 +
    if (count % STEP == 0)
 +
      printf("%d %ld\n",id,val);
 +
    /* exit */
 +
    __sync_lock_release(&lock);
 +
    /* /exit */
 +
  }
 +
}
 +
 +
int main(int argc, char *argv[]) {
 +
  uintptr_t i;
 +
  pthread_t t[NPROC];
 
   for (i=0; i<NPROC; i++)
 
   for (i=0; i<NPROC; i++)
 
     pthread_create(&t[i], NULL, codet, (void *) i);
 
     pthread_create(&t[i], NULL, codet, (void *) i);
Line 60: Line 207:
 
}
 
}
 
</syntaxhighlight>
 
</syntaxhighlight>
 +
questo funziona bene!
 +
(se si decommenta la chiamata (di system call) sched_yield() l'esecuzione &egrave; molto pi&ugrave veloce, perch&eacute;?)
 +
 +
== Test&Set fra processi ==
 +
 +
Ecco il T&S:
 +
<syntaxhighlight lang=C>
 +
#include<stdio.h>
 +
#include<fcntl.h>
 +
#include<stdlib.h>
 +
#include<unistd.h>
 +
#include<sched.h>
 +
#include <sys/mman.h>
 +
 +
#define MAX 400000000
 +
#define STEP 1000000
 +
 +
volatile int *shared;
 +
#define val shared[0]
 +
#define lock shared[1]
 +
 +
void *create_shmem(char *name, off_t length) {
 +
  int shmfd = shm_open(name, O_CREAT|O_RDWR, 0600);
 +
  ftruncate(shmfd,length);
 +
  return mmap(NULL, length, PROT_READ|PROT_WRITE, MAP_SHARED, shmfd, 0);
 +
}
 +
 +
int destroy_shmem(char *name) {
 +
  return shm_unlink(name);
 +
}
 +
 +
int main(int argc, char *argv[]) {
 +
  shared = create_shmem("test", 3*sizeof(int));
 +
  int count;
 +
  for (count = 0; count < MAX; count ++) {
 +
    /* enter */
 +
    while (__sync_lock_test_and_set(&lock,1))
 +
      sched_yield();
 +
    /* /enter */
 +
    val = val + 1;
 +
    if (count % STEP == 0)
 +
      printf("%ld\n",val);
 +
    /* exit */
 +
    __sync_lock_release(&lock);
 +
    /* /exit */
 +
  }
 +
  printf("%d\n",val);
 +
 +
  destroy_shmem("test");
 +
}
 +
</syntaxhighlight>
 +
 +
* shm_create e shm_unlink sono definite nella libreria librt
 +
* lanciate questo programma da piu' terminali virtuali
 +
* se bloccate l'esecuzione puo' rimane definita l'area di memoria condivisa. se quando avete "ucciso" il processo lock valeva 1, il programma non funzionera' piu'. In GNU-Linux le aree di memoria condivisa sono visibili nel file system nella directory "/dev/shm", quindi occorre cancellare lo pseudo-file "/dev/shm/test".

Latest revision as of 08:24, 17 October 2017

Tutti questi esempi hanno NPROC thread che svolgono un loop MAX volte sommando 1 ad ogni iterazione ad una variabile globale condivisa val.

Se tutto va bene il risultato finale deve essere NPROC*MAX.

Dekker

#include<stdio.h>
#include<stdint.h>
#include<stdlib.h>
#include<sched.h>
#include<pthread.h>
#include<unistd.h>

#define MAX 400
#define STEP 10
#define NPROC 2

volatile long val = 0;
int need[NPROC];
int turn;

void *codet(void *arg) {
  uintptr_t id = (uintptr_t) arg;
  int count;
  long tmp;
  for (count=0; count  < MAX; count++) {
    /* enter */
    need[id] = 1;
    __sync_synchronize();
    while (need[1 - id]) {
      __sync_synchronize();
      if (turn != id) {
        need[id] = 0;
        while (turn != id)
          sched_yield();
        need[id] = 1;
      }
      __sync_synchronize();
    }
    /* /enter */
    tmp = val;
    //usleep(rand() % 2);
    val = tmp + 1;
    if (count % STEP == 0)
      printf("%d %ld\n",id,val);
    /* exit */
    turn = 1 - id;
    need[id] = 0;
    /* /exit */
  }
}

int main(int argc, char *argv[]) {
  uintptr_t i;
  pthread_t t[NPROC];
  srand(time(NULL));
  for (i=0; i<NPROC; i++)
    pthread_create(&t[i], NULL, codet, (void *) i);
  for (i=0; i<NPROC; i++)
    pthread_join(t[i], NULL);
  printf("%ld %d\n",val,NPROC*MAX);
}

Peterson

Proviamo con Peterson

#include<stdio.h>
#include<stdint.h>
#include<pthread.h>
#include<unistd.h>

#define MAX 4000000
#define STEP 1000000
#define NPROC 2

volatile long val = 0;
volatile long need[NPROC];
volatile long turn;

void *codet(void *arg) {
  uintptr_t id = (uintptr_t) arg;
  int count;
  for (count=0; count  < MAX; count++) {
    /* enter */
    need[id] = 1;
    turn = 1 - id;
    __sync_synchronize();
    while (need[1 - id] && turn != id)
      ;
    /* /enter */
    val = val + 1;
    if (count % STEP == 0)
      printf("%d %ld\n",id,val);
    /* exit */
    need[id] = 0;
    /* /exit */
  }
}

int main(int argc, char *argv[]) {
  uintptr_t i;
  pthread_t t[NPROC];
  for (i=0; i<NPROC; i++)
    pthread_create(&t[i], NULL, codet, (void *) i);
  for (i=0; i<NPROC; i++)
    pthread_join(t[i], NULL);
  printf("%d %d\n",val,NPROC*MAX);
}

La funzione del compilatore C "__sync_synchronize" (che nella macchine Intel genera l'istuzione assembler "__asm__ __volatile__("mfence");") è necessaria per x86 multicore perché altrimenti le cache dei processori potrebbero non essere allineate.

Peterson per N processi

#include<stdio.h>
#include<stdint.h>
#include<pthread.h>
#include<unistd.h>

#define MAX 4000000
#define NPROC 4

volatile long val = 0;
volatile long turn;
volatile long stage[NPROC];
volatile long last[NPROC];

void *codet(void *arg) {
  uintptr_t id = (uintptr_t) arg;
  int count;
  int j, k;
  for (count=0; count  < MAX; count++) {
    /* enter */
    for (j=0; j<NPROC; j++) {
      stage[id] = j+1; last[j] = id;
      __sync_synchronize();
      for (k=0; k<NPROC; k++) {
        if (id != k)
          while (stage[k] >= stage[id] && last[j] == id)
            ;
      }
    }
    /* /enter */
    val = val + 1;
    if ((count & ((1<<20) - 1)) == 0)
      printf("%d %ld\n",id,val);
    /* exit */
    stage[id] = 0;
    /* /exit */
  }
}

int main(int argc, char *argv[]) {
  uintptr_t i;
  pthread_t t[NPROC];
  for (i=0; i<NPROC; i++)
    pthread_create(&t[i], NULL, codet, (void *) i);
  for (i=0; i<NPROC; i++)
    pthread_join(t[i], NULL);
  printf("%d %d\n",val,NPROC*MAX);
}

Test&Set

Ora proviamo con test&set (che viene fornito dal compilatore C)

#include<stdio.h>
#include<stdint.h>
#include<pthread.h>
#include<unistd.h>
#include<sched.h>

#define MAX 4000000
#define STEP 1000000
#define NPROC 4

volatile long val = 0;
volatile long lock = 0;

void *codet(void *arg) {
  uintptr_t id = (uintptr_t) arg;
  int count;
  for (count=0; count  < MAX; count++) {
    /* enter */
    while (__sync_lock_test_and_set(&lock,1))
      /*sched_yield()*/;
    /* /enter */
    val = val + 1;
    if (count % STEP == 0)
      printf("%d %ld\n",id,val);
    /* exit */
    __sync_lock_release(&lock);
    /* /exit */
  }
}

int main(int argc, char *argv[]) {
  uintptr_t i;
  pthread_t t[NPROC];
  for (i=0; i<NPROC; i++)
    pthread_create(&t[i], NULL, codet, (void *) i);
  for (i=0; i<NPROC; i++)
    pthread_join(t[i], NULL);
  printf("%d %d\n",val,NPROC*MAX);
}

questo funziona bene! (se si decommenta la chiamata (di system call) sched_yield() l'esecuzione è molto pi&ugrave veloce, perché?)

Test&Set fra processi

Ecco il T&S:

#include<stdio.h>
#include<fcntl.h>
#include<stdlib.h>
#include<unistd.h>
#include<sched.h>
#include <sys/mman.h>

#define MAX 400000000
#define STEP 1000000

volatile int *shared;
#define val shared[0]
#define lock shared[1]

void *create_shmem(char *name, off_t length) {
  int shmfd = shm_open(name, O_CREAT|O_RDWR, 0600);
  ftruncate(shmfd,length);
  return mmap(NULL, length, PROT_READ|PROT_WRITE, MAP_SHARED, shmfd, 0);
}

int destroy_shmem(char *name) {
  return shm_unlink(name);
}

int main(int argc, char *argv[]) {
  shared = create_shmem("test", 3*sizeof(int));
  int count;
  for (count = 0; count < MAX; count ++) {
    /* enter */
    while (__sync_lock_test_and_set(&lock,1))
      sched_yield();
    /* /enter */
    val = val + 1;
    if (count % STEP == 0)
      printf("%ld\n",val);
    /* exit */
    __sync_lock_release(&lock);
    /* /exit */
  }
  printf("%d\n",val);

  destroy_shmem("test");
}
  • shm_create e shm_unlink sono definite nella libreria librt
  • lanciate questo programma da piu' terminali virtuali
  • se bloccate l'esecuzione puo' rimane definita l'area di memoria condivisa. se quando avete "ucciso" il processo lock valeva 1, il programma non funzionera' piu'. In GNU-Linux le aree di memoria condivisa sono visibili nel file system nella directory "/dev/shm", quindi occorre cancellare lo pseudo-file "/dev/shm/test".