workflow 01-mpoller

本篇开始介绍 workflow 源码 从里往外 进行了解 ，作为一个异步事件驱动服务器引擎，无论如何底层都是对事件的管理 和封装，本质就是一个事件收集器 ，可以这么理解，将io 事件 进行 收集回调 上成应用 ，简称”套娃“， workflow: C++并行计算与异步网络引擎 ⭐⭐⭐ (gitee.com)

/*
  Copyright (c) 2019 Sogou, Inc.

  Licensed under the Apache License, Version 2.0 (the "License");
  you may not use this file except in compliance with the License.
  You may obtain a copy of the License at

      http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software
  distributed under the License is distributed on an "AS IS" BASIS,
  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  See the License for the specific language governing permissions and
  limitations under the License.

  Author: Xie Han (xiehan@sogou-inc.com)
*/

#ifndef _MPOLLER_H_
#define _MPOLLER_H_

#include <stddef.h>
#include "poller.h"



/*
* 1、poller ->>mpoller_t 管理
* 2、结构体的别名
*/
typedef struct __mpoller mpoller_t;

#ifdef __cplusplus
extern "C"
{
#endif

mpoller_t *mpoller_create(const struct poller_params *params, size_t nthreads);
int mpoller_start(mpoller_t *mpoller);
void mpoller_stop(mpoller_t *mpoller);
void mpoller_destroy(mpoller_t *mpoller);

#ifdef __cplusplus
}
#endif

struct __mpoller
{
    void **nodes_buf;
    unsigned int nthreads;
    poller_t *poller[1];
};

static inline int mpoller_add(const struct poller_data *data, int timeout,
                              mpoller_t *mpoller)
{
    unsigned int index = (unsigned int)data->fd % mpoller->nthreads;
    return poller_add(data, timeout, mpoller->poller[index]);
}

static inline int mpoller_del(int fd, mpoller_t *mpoller)
{
    unsigned int index = (unsigned int)fd % mpoller->nthreads;
    return poller_del(fd, mpoller->poller[index]);
}

static inline int mpoller_mod(const struct poller_data *data, int timeout,
                              mpoller_t *mpoller)
{
    unsigned int index = (unsigned int)data->fd % mpoller->nthreads;
    return poller_mod(data, timeout, mpoller->poller[index]);
}

static inline int mpoller_set_timeout(int fd, int timeout, mpoller_t *mpoller)
{
    unsigned int index = (unsigned int)fd % mpoller->nthreads;
    return poller_set_timeout(fd, timeout, mpoller->poller[index]);
}

static inline int mpoller_add_timer(const struct timespec *value, void *context,
                                    mpoller_t *mpoller)
{
    static unsigned int n = 0;
    unsigned int index = n++ % mpoller->nthreads;
    return poller_add_timer(value, context, mpoller->poller[index]);
}

#endif

/*
  Copyright (c) 2019 Sogou, Inc.

  Licensed under the Apache License, Version 2.0 (the "License");
  you may not use this file except in compliance with the License.
  You may obtain a copy of the License at

      http://www.apache.org/licenses/LICENSE-2.0

  Unless required by applicable law or agreed to in writing, software
  distributed under the License is distributed on an "AS IS" BASIS,
  WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  See the License for the specific language governing permissions and
  limitations under the License.

  Author: Xie Han (xiehan@sogou-inc.com)
*/

#include <stddef.h>
#include <stdlib.h>
#include "poller.h"
#include "mpoller.h"

extern poller_t *__poller_create(void **, const struct poller_params *);
extern void __poller_destroy(poller_t *);

static int __mpoller_create(const struct poller_params *params,
                            mpoller_t *mpoller)
{
    // 双指针  指向指针的指针  因为传入的mpoller 就是个指针  文件描述符的大小
    void **nodes_buf = (void **)calloc(params->max_open_files, sizeof (void *));
    unsigned int i;

    if (nodes_buf)
    {    
        // 根据线程数据 创建相对于的poller 传入参数
        for (i = 0; i < mpoller->nthreads; i++)
        {
            mpoller->poller[i] = __poller_create(nodes_buf, params);
            if (!mpoller->poller[i])
                break;
        }
        // 如果创建的i 和线程数相同的话 
        if (i == mpoller->nthreads)
        {    
            // nodes_buf 指定
            mpoller->nodes_buf = nodes_buf;
            return 0;
        }
        // 如果创建失败，则进行销毁poller 和释放  nodes_buf
        while (i > 0)
            __poller_destroy(mpoller->poller[--i]);

        free(nodes_buf);
    }

    return -1;
}

// poller 创建  需要考虑创建成功 或者 失败
mpoller_t *mpoller_create(const struct poller_params *params, size_t nthreads)
{
    mpoller_t *mpoller;
    size_t size;

    if (nthreads == 0)
        nthreads = 1;
    // 根据线程分配 结构体偏移 到  poller 并增加 nthreads 个指针大小内存
    // poller_t *poller[1] 扩增 poller 数组大小 
    size = offsetof(mpoller_t, poller) + nthreads * sizeof (void *);
    mpoller = (mpoller_t *)malloc(size);
    if (mpoller)
    {
        mpoller->nthreads = (unsigned int)nthreads;
        if (__mpoller_create(params, mpoller) >= 0)
            return mpoller;

        free(mpoller);
    }

    return NULL;
}

// poller 启动  启动成功 和启动失败的措施
int mpoller_start(mpoller_t *mpoller)
{
    size_t i;

    for (i = 0; i < mpoller->nthreads; i++)
    {
        if (poller_start(mpoller->poller[i]) < 0)
            break;
    }

    if (i == mpoller->nthreads)
        return 0;

    while (i > 0)
        poller_stop(mpoller->poller[--i]);

    return -1;
}

// poller 停止
void mpoller_stop(mpoller_t *mpoller)
{
    size_t i;

    for (i = 0; i < mpoller->nthreads; i++)
        poller_stop(mpoller->poller[i]);
}

// poller 销毁 并释放内存 和 mpoller 因为再创建的时候 mpoller_create 分配了该内存
void mpoller_destroy(mpoller_t *mpoller)
{
    size_t i;

    for (i = 0; i < mpoller->nthreads; i++)
        __poller_destroy(mpoller->poller[i]);

    free(mpoller->nodes_buf);
    free(mpoller);
}

这个mpoller 就是用来管理 epoll 创建，几个线程就创建几个 poller ,整体比较简单，其中要注意的是对

c 语言的使用，如下；extern 外部函数 说明在其他地方已经实现

extern poller_t *__poller_create(void **, const struct poller_params *);
extern void __poller_destroy(poller_t *);

分配 max_open_files 个文件描述符大小空间

void **nodes_buf = (void **)calloc(params->max_open_files, sizeof (void *));

根据线程分配 结构体偏移 到 poller 并增加 nthreads 个指针大小内存 poller_t *poller[1] 扩增 poller 数组大小 ， offsetof 的使用 用于 结构体变量的字节偏移

size = offsetof(mpoller_t, poller) + nthreads * sizeof (void *);

workflow 的代码非常值得我们学习和研究 ，你要了解 整个架构的是如何 设计，以及各个模块之间是如何配合的，希望我的理解可以对你有帮助，通过对开源的学习，可以提高自己能力，其余的会在其他短篇文章里面进行讲解，尽量更新。