redis源码笔记

hanling

作者在bio.c的头注释中对设计进行了详细的介绍
/* Background I/O service for Redis.
   这个文件是redis后台IO服务的实现
*
* This file implements operations that we need to perform in the background.
* Currently there is only a single operation, that is a background close(2)
* system call. This is needed as when the process is the last owner of a
* reference to a file closing it means unlinking it, and the deletion of the
* file is slow, blocking the server.
   这个文件负责我们需要在后台执行的操作。现在redis的版本中只有一类的操作，后台的close 系统调用。
   为了避免一个文件最后的owner在执行close操作带来的unlink使得阻塞server，将这类操作用单独的后台线程来执行
*
* In the future we'll either continue implementing new things we need or
* we'll switch to libeio. However there are probably long term uses for this
* file as we may want to put here Redis specific background tasks (for instance
* it is not impossible that we'll need a non blocking FLUSHDB/FLUSHALL
* implementation).
*
* DESIGN
* ------
*
* The design is trivial, we have a structure representing a job to perform
* and a different thread and job queue for every job type.
* Every thread wait for new jobs in its queue, and process every job
* sequentially.
   每种作业类型一个queue。每个线程在它的queue里等待新的job到来。并且按照FIFO的顺序处理作业。
*
* Jobs of the same type are guaranteed to be processed from the least
* recently inserted to the most recently inserted (older jobs processed
* first).
*
* Currently there is no way for the creator of the job to be notified about
* the completion of the operation, this will only be added when/if needed.
   作业完成后，其creator无法得到通知。
*/
　　现在的两类作业类型：1.close 2.aof_fsync



/* Background job opcodes */
#define REDIS_BIO_CLOSE_FILE    0 /* Deferred close(2) syscall. */
#define REDIS_BIO_AOF_FSYNC     1 /* Deferred AOF fsync. */


  1 #include "redis.h"
  2 #include "bio.h"
  3
    //使用互斥量+条件变量，作为线程的保护条件
  4 static pthread_mutex_t bio_mutex[REDIS_BIO_NUM_OPS];
  5 static pthread_cond_t bio_condvar[REDIS_BIO_NUM_OPS];
    //两类作业的队列
  6 static list *bio_jobs[REDIS_BIO_NUM_OPS];
  7 /* The following array is used to hold the number of pending jobs for every
  8  * OP type. This allows us to export the bioPendingJobsOfType() API that is
  9  * useful when the main thread wants to perform some operation that may involve
10  * objects shared with the background thread. The main thread will just wait
11  * that there are no longer jobs of this type to be executed before performing
12  * the sensible operation. This data is also useful for reporting. */
13 static unsigned long long bio_pending[REDIS_BIO_NUM_OPS];
14
15 /* This structure represents a background Job. It is only used locally to this
16  * file as the API deos not expose the internals at all. */
17 struct bio_job {
18     time_t time; /* Time at which the job was created. */
19     /* Job specific arguments pointers. If we need to pass more than three
20      * arguments we can just pass a pointer to a structure or alike. */
21     void *arg1, *arg2, *arg3;
22 };
23
24 void *bioProcessBackgroundJobs(void *arg);
25
26 /* Make sure we have enough stack to perform all the things we do in the
27  * main thread. */
28 #define REDIS_THREAD_STACK_SIZE (1024*1024*4)
29
30 /* Initialize the background system, spawning the thread. */
31 void bioInit(void) {
32     pthread_attr_t attr;
33     pthread_t thread;
34     size_t stacksize;
35     int j;
36
37     /* Initialization of state vars and objects */
38     for (j = 0; j < REDIS_BIO_NUM_OPS; j++) {
39         pthread_mutex_init(&bio_mutex[j],NULL);
40         pthread_cond_init(&bio_condvar[j],NULL);
41         bio_jobs[j] = listCreate();
42         bio_pending[j] = 0;
43     }
44
45     /* Set the stack size as by default it may be small in some system */
46     pthread_attr_init(&attr);
47     pthread_attr_getstacksize(&attr,&stacksize);
48     if (!stacksize) stacksize = 1; /* The world is full of Solaris Fixes */
49     while (stacksize < REDIS_THREAD_STACK_SIZE) stacksize *= 2;
50     pthread_attr_setstacksize(&attr, stacksize);
51
52     /* Ready to spawn our threads. We use the single argument the thread
53      * function accepts in order to pass the job ID the thread is
54      * responsible of. */
55     for (j = 0; j < REDIS_BIO_NUM_OPS; j++) {
56         void *arg = (void*)(unsigned long) j;
57         if (pthread_create(&thread,&attr,bioProcessBackgroundJobs,arg) != 0) {
58             redisLog(REDIS_WARNING,"Fatal: Can't initialize Background Jobs.");
59             exit(1);
60         }
61     }
62 }
63
64 void bioCreateBackgroundJob(int type, void *arg1, void *arg2, void *arg3) {
65     struct bio_job *job = zmalloc(sizeof(*job));
66
67     job->time = time(NULL);
68     job->arg1 = arg1;
69     job->arg2 = arg2;
70     job->arg3 = arg3;
71     pthread_mutex_lock(&bio_mutex[type]);
72     listAddNodeTail(bio_jobs[type],job);
73     bio_pending[type]++;
74     pthread_cond_signal(&bio_condvar[type]);
75     pthread_mutex_unlock(&bio_mutex[type]);
76 }
77
78 void *bioProcessBackgroundJobs(void *arg) {
79     struct bio_job *job;
80     unsigned long type = (unsigned long) arg;
81
82     pthread_detach(pthread_self());
83     pthread_mutex_lock(&bio_mutex[type]);
84     while(1) {
85         listNode *ln;
86
87         /* The loop always starts with the lock hold. */
88         if (listLength(bio_jobs[type]) == 0) {
89             pthread_cond_wait(&bio_condvar[type],&bio_mutex[type]);
90             continue;
91         }
92         /* Pop the job from the queue. */
93         ln = listFirst(bio_jobs[type]);
94         job = ln->value;
95         /* It is now possible to unlock the background system as we know have
96          * a stand alone job structure to process.*/
97         pthread_mutex_unlock(&bio_mutex[type]);
98
99         /* Process the job accordingly to its type. */
100         if (type == REDIS_BIO_CLOSE_FILE) {
101             close((long)job->arg1);
102         } else if (type == REDIS_BIO_AOF_FSYNC) {
103             aof_fsync((long)job->arg1);
104         } else {
105             redisPanic("Wrong job type in bioProcessBackgroundJobs().");
106         }
107         zfree(job);
108
109         /* Lock again before reiterating the loop, if there are no longer
110          * jobs to process we'll block again in pthread_cond_wait(). */
111         pthread_mutex_lock(&bio_mutex[type]);
112         listDelNode(bio_jobs[type],ln);
113         bio_pending[type]--;
114     }
115 }
116
117 /* Return the number of pending jobs of the specified type. */
118 unsigned long long bioPendingJobsOfType(int type) {
119     unsigned long long val;
120     pthread_mutex_lock(&bio_mutex[type]);
121     val = bio_pending[type];
122     pthread_mutex_unlock(&bio_mutex[type]);
123     return val;
124 }