生产者消费者问题--进阶
生产者消费者问题（英语：Producer-consumer problem），也称有限缓冲问题（英语：Bounded-buffer problem），是一个多线程同步问题的经典案例。该问题描述了两个共享固定大小缓冲区的线程——即所谓的“生产者”和“消费者”——在实际运行时会发生的问题。生产者的主要作用是生成一定量的数据放到缓冲区中，然后重复此过程。与此同时，消费者也在缓冲区消耗这些数据。该问题的关键就是要保证生产者不会在缓冲区满时加入数据，消费者也不会在缓冲区中空时消耗数据
一种实现如下：
#include <stdio.h> #include <unistd.h> #include <pthread.h> #include <string.h> #define MAX 5 //缓冲区的的大小 pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; pthread_cond_t cond = PTHREAD_COND_INITIALIZER; typedef struct{ char buffer[MAX]; int count; }Buffer; Buffer share = {"", 0}; char ch = 'A'; void *producer(void *arg) { printf("Producer : starting \n"); while(ch != 'K') { pthread_mutex_lock(&mutex); if(share.count != MAX) { share.buffer[share.count++] = ch++; printf("Producer: put char[%c]\n", ch-1); if(share.count == MAX) { printf("Producer: signaling full\n"); pthread_cond_signal(&cond);//若果缓存满了发送信号 } } pthread_mutex_unlock(&mutex); } sleep(1); printf("Produce: Exiting \n"); pthread_exit(NULL); } void *consumer(void *junk) { int i; printf("Consumer : starting\n"); while (ch != 'K') { pthread_mutex_lock(&mutex); printf("\t Consumer : Waiting\n"); while(share.count != MAX){ pthread_cond_wait(&cond, &mutex); //条件不成立释放锁. printf("Consumer wating for FULL signal\n"); } printf("Consumer : getting buffer :: "); for(i = 0; share.buffer[i] && share.count;++i, share.count--) putchar(share.buffer[i]); putchar('\n'); pthread_mutex_unlock(&mutex); } } int main() { pthread_t read, write; pthread_create(&read, NULL, (void *) consumer, NULL); pthread_create(&write, NULL, (void *)producer, NULL); pthread_join(read, NULL); pthread_join(write, NULL); return 0; }
修改consumer的代码：
void *consumer(void *junk) { int i; printf("Consumer : starting\n"); while (ch != 'K') { pthread_mutex_lock(&mutex); printf("\t Consumer : Waiting\n"); //while(share.count != MAX){ pthread_cond_wait(&cond, &mutex); //条件不成立释放锁. printf("Consumer wating for FULL signal\n"); //} printf("Consumer : getting buffer :: "); for(i = 0; share.buffer[i] && share.count;++i, share.count--) putchar(share.buffer[i]); putchar('\n'); pthread_mutex_unlock(&mutex); } }
编译运行会有两种结果:
一种是：
Consumer : starting Consumer : Waiting Producer : starting Producer: put char[A] Producer: put char[B] Producer: put char[C] Producer: put char[D] Producer: put char[E] Producer: signaling full Consumer wating for FULL signal Consumer : getting buffer :: ABCDE Consumer : Waiting Producer: put char[F] Producer: put char[G] Producer: put char[H] Producer: put char[I] Producer: put char[J] Producer: signaling full Consumer wating for FULL signal Consumer : getting buffer :: FGHIJ Produce: Exiting
另一种是：
Producer : starting Producer: put char[A] Producer: put char[B] Producer: put char[C] Producer: put char[D] Producer: put char[E] Producer: signaling full Consumer : starting Consumer : Waiting
可以看出来第二种是先执行了生产者，生产者填充满buffer之后，发送条件消息，但是此时consumer还没有执行，也并没有等待条件
然后生产者释放锁，接着消费者获取锁，然后等待条件，由于缓冲已经满，
if(share.count == MAX)
生产者不会进入发送消息的代码,所以消费者一直等待条件．
而上面一种的结果是因为，消费这先执行了，进入等待条件，所以没有这个问题，但是问题在于我们不能保证pthread_cond_wait()一定是先与pthread_cond_signal()执行的．
但是通过加while()代码，就不会出现等待条件变量的问题，可以直接执行下面的代码．
这个生产者消费者的解决方案是，同时只有一个生产者和消费者，是1:1的关系，生产者需填满缓冲区，才让消费者来取数据．下面一种实现是只要缓冲区不为空，消费者就来取数据，为空，就等待，缓冲区只要不满，生产者就生产数据，否则等待，支持n:n的关系．
修改上面的代码:
#include <stdio.h> #include <unistd.h> #include <pthread.h> #include <string.h> #define P_COUNT 5 //producer NO. #define C_COUNT 5 //NO. #define MAX 5 //缓冲区的的大小 pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; //锁住缓冲区 /*队列满的时候，阻塞生产这线程，队列空时阻塞消费者线程*/ pthread_cond_t notFull = PTHREAD_COND_INITIALIZER; // pthread_cond_t notEmpty = PTHREAD_COND_INITIALIZER; typedef struct{ char buffer[MAX]; int count; }Buffer; Buffer share = {"", 0}; char ch = 'A'; void *producer(void *arg) { int id = *(int *)arg; printf("[%d] Producer : starting \n", id); //while(ch != 'K') while(1) { pthread_mutex_lock(&mutex); while(share.count == MAX){ pthread_cond_wait(¬Full, &mutex); printf("[%d] producer wating for not full signal\n", id); } share.buffer[share.count++] = ch; //字符不能无限加超出范围了 printf("[%d] Producer: put [%d] char[%c]\n", id, share.count-1, ch ); pthread_cond_signal(¬Empty); pthread_mutex_unlock(&mutex); sleep(1); } sleep(1); printf("Produce: Exiting \n"); } void *consumer(void *junk) { int id = *(int *)junk; printf("\t[%d] Consumer : starting\n", id); //while (ch != 'K') while (1) { pthread_mutex_lock(&mutex); printf("\t [%d] Consumer : Waiting\n", id); while(share.count == 0){ pthread_cond_wait(¬Empty, &mutex); //条件不成立释放锁. printf("\t[%d] Consumer wating for not empty signal\n", id); } //消费者取数据应该怎么取呢,这里是从数组的最右边开始取，也就是先取最新的数据 printf("\t[%d] Consumer : getting buffer[%d] :: [%c] \n", id, share.count, share.buffer[share.count-1]); share.count--; pthread_cond_signal(¬Full); pthread_mutex_unlock(&mutex); sleep(1); } } int main() { int i; pthread_t t_read[C_COUNT], t_write[P_COUNT]; int *pId=(int *)malloc(sizeof(int)*P_COUNT); int *cId=(int *)malloc(sizeof(int)*C_COUNT); for(i = 0; i < P_COUNT; ++i){ pId[i] = i; pthread_create(&t_write[i], NULL, (void *)producer, (void *)&pId[i]); } for(i = 0; i < C_COUNT; ++i){ cId[i] = i; pthread_create(&t_read[i], NULL, (void *) consumer, (void *)&cId[i]); } for(i = 0; i < P_COUNT; ++i){ pthread_join(t_read[i], NULL); } for(i = 0; i < C_COUNT; ++i){ pthread_join(t_write[i], NULL); } pthread_mutex_destroy(&mutex); pthread_cond_destroy(¬Full); pthread_cond_destroy(¬Empty); return 0; }
这里用到等待队列来实现同步，下面这种是采用信号量的方式来实现同步
#include <unistd.h> #include <sys/types.h> #include <pthread.h> #include <semaphore.h> #include <stdlib.h> #include <stdio.h> #include <errno.h> #include <string.h> #define ERR_EXIT(m) \ do \ { \ perror(m); \ exit(EXIT_FAILURE); \ } while(0) #define CONSUMERS_COUNT 1 #define PRODUCERS_COUNT 1 #define BUFFSIZE 10 int g_buffer[BUFFSIZE]; unsigned short in = 0; unsigned short out = 0; unsigned short produce_id = 0; //产品数不断累加 unsigned short consume_id = 0; sem_t g_sem_full; sem_t g_sem_empty; pthread_mutex_t g_mutex; pthread_t g_thread[CONSUMERS_COUNT + PRODUCERS_COUNT]; void *consume(void *arg) { int i; int num = *(int*)arg; while (1) { printf("%d wait buffer not empty\n", num); sem_wait(&g_sem_empty); pthread_mutex_lock(&g_mutex); for (i = 0; i < BUFFSIZE; i++) { printf("%02d ", i); if (g_buffer[i] == -1) printf("%s", "null"); else printf("%d", g_buffer[i]); if (i == out) printf("\t<--consume"); printf("\n"); } consume_id = g_buffer[out]; printf("%d begin consume product %d\n", num, consume_id); g_buffer[out] = -1; out = (out + 1) % BUFFSIZE; printf("out == %d \n", out); printf("%d end consume product %d\n", num, consume_id); pthread_mutex_unlock(&g_mutex); sem_post(&g_sem_full); sleep(1); } return NULL; } void *produce(void *arg) { int num = *(int*)arg; int i; while (1) { printf("%d wait buffer not full\n", num); sem_wait(&g_sem_full); pthread_mutex_lock(&g_mutex); for (i = 0; i < BUFFSIZE; i++) { printf("%02d ", i); if (g_buffer[i] == -1) printf("%s", "null"); else printf("%d", g_buffer[i]); if (i == in) printf("\t<--produce"); printf("\n"); } printf("%d begin produce product %d\n", num, produce_id); g_buffer[in] = produce_id; in = (in + 1) % BUFFSIZE; printf("in == %d \n", in); printf("%d end produce product %d\n", num, produce_id++); pthread_mutex_unlock(&g_mutex); sem_post(&g_sem_empty); sleep(5); } return NULL; } int main(void) { int i; for (i = 0; i < BUFFSIZE; i++) g_buffer[i] = -1; sem_init(&g_sem_full, 0, BUFFSIZE); sem_init(&g_sem_empty, 0, 0); pthread_mutex_init(&g_mutex, NULL); for (i = 0; i < CONSUMERS_COUNT; i++) pthread_create(&g_thread[i], NULL, consume, (void *)&i); for (i = 0; i < PRODUCERS_COUNT; i++) pthread_create(&g_thread[CONSUMERS_COUNT + i], NULL, produce, (void *)&i); for (i = 0; i < CONSUMERS_COUNT + PRODUCERS_COUNT; i++) pthread_join(g_thread[i], NULL); sem_destroy(&g_sem_full); sem_destroy(&g_sem_empty); pthread_mutex_destroy(&g_mutex); return 0; }
consumer每次消费完会sleep(1), producer每次生产完会sleep(5),消费者等待的时间更长，故消费者会经常阻塞在sem_wait(&g_sem_empty) 上面，因为缓冲区经常为空，可以将PRODUCTORS_COUNT 改成5，即有5个生产者线程和1个消费者线程，而且生产者睡眠时间还是消费者的5倍，从动态输出可以看出，基本上就动态平衡了，即5个生产者一下子生产了5份东西，消费者1s消费1份，刚好在生产者继续生产前消费完．
修改代码：
#define CONSUMERS_COUNT 1 #define PRODUCERS_COUNT 5
使用Posix信号量可模拟互斥量和条件变量，而且通常更有优势。
当函数sem_wait()和sem_post()用于线程内时，两个调用间的区域就是所要保护的临界区代码；当用于线程间时，则与条件变量等效。
此外，信号量还可用作资源计数器，即初始化信号量的值作为某个资源当前可用的数量，使用时递减释放时递增。这样，原先一些保存队列状态的变量都不再需要。
最后，内核会记录信号的存在，不会将信号丢失；而唤醒条件变量时若没有线程在等待该条件变量，信号将被丢失。
有时候缓冲区也可以是没有限制大小的，修改第一份程序为缓冲区不限制大小的代码:
#include <stdio.h> #include <unistd.h> #include <pthread.h> #include <string.h> #define P_COUNT 1 //producer NO. #define C_COUNT 2 //NO. #define MAX 5 //缓冲区的的大小 pthread_mutex_t mutex = PTHREAD_MUTEX_INITIALIZER; pthread_cond_t cond = PTHREAD_COND_INITIALIZER; typedef struct{ char buffer[MAX]; int count; }Buffer; Buffer share = {"", 0}; char ch = 'A'; void *producer(void *arg) { printf("Producer : starting \n"); while(1) { pthread_mutex_lock(&mutex); share.count++; printf("Producer: put charc count[%d]\n", share.count); pthread_cond_signal(&cond);//若果缓存满了发送信号 pthread_mutex_unlock(&mutex); sleep(1); } sleep(1); printf("Produce: Exiting \n"); pthread_exit(NULL); } void *consumer(void *junk) { printf("Consumer : starting\n"); while (1) { pthread_mutex_lock(&mutex); printf("\t Consumer : Waiting\n"); while(share.count == 0){ pthread_cond_wait(&cond, &mutex); } share.count--; printf("Consumer : getting buffer count[%d] \n", share.count); pthread_mutex_unlock(&mutex); sleep(1); } } int main() { int i; pthread_t t_read[C_COUNT], t_write[P_COUNT]; for(i = 0; i < P_COUNT; ++i){ pthread_create(&t_write[i], NULL, (void *)producer, NULL); } for(i = 0; i < C_COUNT; ++i){ pthread_create(&t_read[i], NULL, (void *) consumer, NULL); } for(i = 0; i < P_COUNT; ++i){ pthread_join(t_read[i], NULL); } for(i = 0; i < C_COUNT; ++i){ pthread_join(t_write[i], NULL); } pthread_mutex_destroy(&mutex); pthread_cond_destroy(&cond); return 0; }