Linux thread and synchronization

Linux multithreading 1. thread Overview a thread is a basic scheduling unit in a process, also known as a lightweight process. Threads are concurrent multi-path execution paths in the shared memory space. they share the resources of a process, such as file descriptions and signal processing. Therefore, context switching overhead is greatly reduced .... Information

Linux multithreading

 

1. thread overview

A thread is a basic scheduling unit in a process, also known as a lightweight process. Threads are concurrent multi-path execution paths in the shared memory space. they share the resources of a process, such as file descriptions and signal processing. Therefore, context switching overhead is greatly reduced. A process can have multiple threads.

There are multiple thread control tables and stack registers, but one user address space is shared.

 

2. thread implementation

 

Thread creation pthread_create ()

Required header file # include

Function prototype int pthread_create (pthread_t * thread, pthread_attr_t * attr,

Void * (* start_routine) (void *), void * arg ))

Thread: specifies the thread id.

Attr: thread attribute settings

Start_routine: the starting address of the thread function.

Arg: parameters passed to start_routine

Function return value success: 0 error:-1

 

The thread exits pthread_exit ();

Required header file # include

Function prototype void pthread_exit (void * retval)

Function input value retval: the return value of the pthread_exit () caller's thread, which can be retrieved and obtained by other functions such as pthread_join.

 

Wait for the thread to exit and release the resource pthread_join ()

Required header file # include

Function prototype int pthread_join (pthread_t th, void ** thread_return ))

Function input value

Th: ID of the waiting thread

Thread_return: user-defined pointer, used to store the return value of the waiting thread (not NULL)

Function return value success: 0 error:-1

 

Code example

1. # include

2. # include

3. # include

4.

5./* thread 1 */

6. void thread1 ()

7 .{

8. int I = 0;

9.

10. while (1)

11 .{

12. printf ("thread1: % d \ n", I );

13. if (I> 3)

14. pthread_exit (0 );

15. I ++;

16. sleep (1 );

17 .}

18 .}

19.

20./* thread 2 */

21. void thread2 ()

22 .{

23. int I = 0;

24.

25. while (1)

26 .{

27. printf ("thread2: % d \ n", I );

28. if (I> 5)

29. pthread_exit (0 );

30. I ++;

31. sleep (1 );

32 .}

33 .}

34.

35. int main ()

36 .{

37. pthread_t t1, t2;

38.

39./* create thread */

40. pthread_create (& t1, NULL, (void *) thread1, NULL );

41. pthread_create (& t2, NULL, (void *) thread2, NULL );

42./* wait for the thread to exit */

43. pthread_join (t1, NULL );

44. pthread_join (t2, NULL );

45. return 0;

46 .}

3. synchronization and Mutex

<1> mutex lock

The operations of mutex lock mainly include the following steps.

• Mutex lock initialization: pthread_mutex_init

• Mutex lock: pthread_mutex_lock

• Mutex lock: pthread_mutex_trylock

• Mutex lock: pthread_mutex_unlock

• Remove mutex: pthread_mutex_destroy

 

1. # include

2. # include

3. # include

4.

5. int I = 0;/* shared variable */

6. pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER;/* mutex lock */

7.

8. void thread1 ()

9 .{

10. int ret;

11. while (1)

12 .{

13.

14.

15. ret = pthread_mutex_trylock (& mutex);/* Determine the lock */

16.

17. if (ret! = EBUSY)

18 .{

19. pthread_mutex_lock (& mutex);/* Lock */

20. printf ("This is thread1: % d \ n", I );

21. I ++;

22. pthread_mutex_unlock (& mutex);/* unlock */

23 .}

24. sleep (1 );

25 .}

26 .}

27.

28. void thread2 ()

29. {int ret;

30. while (1)

31 .{

32.

33. ret = pthread_mutex_trylock (& mutex );

34. if (ret! = EBUSY)

35 .{

36. pthread_mutex_lock (& mutex );

37. printf ("This is thread2: % d \ n", I );

38. I ++;

39. pthread_mutex_unlock (& mutex );

40 .}

41. sleep (1 );

42 .}

43 .}

44. int main ()

45 .{

46. pthread_t t1, t2;

47. pthread_mutex_init (& mutex, NULL );

48. pthread_create (& t1, NULL, (void *) thread1, NULL );

49. pthread_create (& t2, NULL, (void *) thread2, NULL );

50.

51. pthread_join (t1, NULL );

52. pthread_join (t2, NULL );

53.

54. pthread_mutex_destroy (& mutex );

55. return 0;

56 .}

<2> semaphores

Two threads that are not synchronized

 

1. # include

2. # include

3. # include

4.

5. int I = 0;

6. void thread1 ()

7 .{

8.

9. while (1)

10 .{

11. printf ("This is thread1: % d \ n", I );

12. I ++;

13. sleep (1 );

14 .}

15 .}

16.

17.

18. void thread2 ()

19 .{

20.

21. while (1)

22 .{

23. printf ("This is thread2: % d \ n", I );

24. I ++;

25. sleep (1 );

26 .}

27 .}

28.

29. int main ()

30 .{

31. pthread_t t1, t2;

32.

33. pthread_create (& t1, NULL, (void *) thread1, NULL );

34. pthread_create (& t2, NULL, (void *) thread2, NULL );

35.

36. pthread_join (t1, NULL );

37. pthread_join (t2, NULL );

38.

39. return 0;

40 .}

The execution result is as follows:

This is thread1: 0

This is thread2: 1

This is thread2: 2

This is thread1: 3

This is thread2: 4

This is thread1: 4

This is thread2: 6

This is thread1: 7

......

We can see that:

1. thread 2 does not have to be executed after thread 1. if Thread 2 is required to be executed after thread 1, it is called synchronization.

2. thread 1 and thread 2 may read shared variable I at the same time. if only one thread needs to read I at a time, it becomes mutually exclusive.

 

Use of semaphores

• Sem_init is used to create a semaphore and initialize its value.

• Sem_wait and sem_trywait are equivalent to P operations. both of them can reduce the semaphore value by one. The difference between them is that if the semaphore is smaller than zero, sem_wait will block the process, and sem_trywait will return immediately.

• Sem_post is equivalent to the V operation. It adds the semaphore value and sends a signal to wake up the waiting process.

• Sem_getvalue is used to obtain the semaphore value.

• Sem_destroy is used to delete semaphores

 

 

Code

1. # include

2. # include

3. # include

4. # include

5.

6.

7. int I = 0;

8. sem_t sem1, sem2;

9.

10.

11. void thread1 ()

12 .{

13.

14. while (1)

15 .{

16. sem_wait (& sem1 );

17. printf ("This is thread1: % d \ n", I );

18. I ++;

19. sleep (3);/* thread 1 sleep for 3 s to observe that thread 2 will be executed after 3 s output */

20. sem_post (& sem2 );

21.

22 .}

23 .}

24.

25.

26. void thread2 ()

27 .{

28.

29. while (1)

30 .{

31. sem_wait (& sem2 );

32. printf ("This is thread2: % d \ n", I );

33. I ++;

34. sem_post (& sem1 );

35. sleep (1 );

36 .}

37 .}

38.

39. int main ()

40 .{

41. pthread_t t1, t2;

42.

43.

44.

45. sem_init (& sem1,);/* initialize semaphores sem1 */

46. sem_init (& sem2, 0, 0 );

47.

48. pthread_create (& t1, NULL, (void *) thread1, NULL );

49. pthread_create (& t2, NULL, (void *) thread2, NULL );

50.

51. pthread_join (t1, NULL );

52. pthread_join (t2, NULL );

53.

54. return 0;

55 .}

 

 

 

1. # include

2. # include

3. # include

4. # include

5.

6.

7. int I = 0;

8. sem_t sem;

9.

10.

11. void thread1 ()

12 .{

13.

14. while (1)

15 .{

16. sem_wait (& sem );

17. printf ("This is thread1: % d \ n", I );

18. I ++;

19. sleep (3);/* thread 1 sleep for 3 s to observe that thread 2 will be executed after 3 s output */

20. sem_post (& sem );

21.

22 .}

23 .}

24.

25.

26. void thread2 ()

27 .{

28.

29. while (1)

30 .{

31. sem_wait (& sem );

32. printf ("This is thread2: % d \ n", I );

33. I ++;

34. sem_post (& sem );

35. sleep (1 );

36 .}

37 .}

38.

39. int main ()

40 .{

41. pthread_t t1, t2;

42.

43.

44.

45. sem_init (& sem, 0, 1);/* initialize semaphores sem */

46.

47. pthread_create (& t1, NULL, (void *) thread1, NULL );

48. pthread_create (& t2, NULL, (void *) thread2, NULL );

49.

50. pthread_join (t1, NULL );

51. pthread_join (t2, NULL );

52.

53. return 0;

54 .}