Nginx-rtmp之监听端口的管理

1. 概述

监听端口属于 server 虚拟主机，它是由 server{} 块下的 listen 配置项决定的。

每监听一个 TCP 端口，都将使用一个独立的 ngx_rtmp_conf_port_t 结构体来表示：

typedef struct {
    /* socket 地址家族 */
    int                     family;
    
    /* 监听端口 */
    in_port_t               port;
    
    /* 监听的端口下对应着的所有 ngx_rtmp_conf_addr_t 地址 */
    ngx_array_t             addrs;       /* array of ngx_rtmp_conf_addr_t */
} ngx_rtmp_conf_port_t;

一个端口，可能对应着多个地址（当主机上有多个 IP 地址时），该地址用 ngx_rtmp_conf_addr_t 表示：

typedef struct {
    struct sockaddr        *sockaddr;
    socklen_t               socklen;

    ngx_rtmp_conf_ctx_t    *ctx;

    unsigned                bind:1;
    unsigned                wildcard:1;
#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
    unsigned                ipv6only:2;
#endif
    unsigned                so_keepalive:2;
    unsigned                proxy_protocol:1;
#if (NGX_HAVE_KEEPALIVE_TUNABLE)
    int                     tcp_keepidle;
    int                     tcp_keepintvl;
    int                     tcp_keepcnt;
#endif
} ngx_rtmp_conf_addr_t;

在 ngx_rtmp_core_main_conf_t 结构体中，保存着 nginx.conf 中所有需要监听的端口：

typedef struct {
    ...
    /* 该数组保存着的元素是指向 ngx_rtmp_listen_t 结构体的指针，
     * 每个元素代表一个需要监听的端口 */
    ngx_array_t     listen;  /* ngx_rtmp_listen_t */
    ...
}ngx_rtmp_core_main_conf_t;

Nginx 每从 nginx.conf 中解析中一个 listen 配置项，就把该配置项的值组织成 ngx_rtmp_listen_t 结构体添加到
ngx_rtmp_core_main_conf_t 的成员 listen 数组中，如下：

typedef struct {
    u_char                  sockaddr[NGX_SOCKADDRLEN];
    socklen_t               socklen;

    /* server ctx */
    ngx_rtmp_conf_ctx_t    *ctx;

    unsigned                bind:1;
    unsigned                wildcard:1;
#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
    unsigned                ipv6only:2;
#endif
    unsigned                so_keepalive:2;
    unsigned                proxy_protocol:1;
#if (NGX_HAVE_KEEPALIVE_TUNABLE)
    int                     tcp_keepidle;
    int                     tcp_keepintvl;
    int                     tcp_keepcnt;
#endif
} ngx_rtmp_listen_t;

对同一个端口 1935，可以同时监听 127.0.0.1:8000、192.168.56.101:1935 这两个地址，当一台物理机器具备多个 IP
地址时这是很有用的。对应到 RTMP 的框架上，Nginx 是使用 ngx_rtmp_conf_addr_t 结构体来表示一个对应着具体地址
的监听端口，因此，一个 ngx_rtmp_conf_port_t 将会对应多个 ngx_rtmp_conf_addr_t，而 ngx_rtmp_conf_addr_t 是
以动态数组的形式保存在 ngx_rtmp_conf_port_t 中的 addrs 成员中的。

2. 监听端口的解析与管理

2.1 ngx_rtmp_core_listen：解析 listen 配置项

static char *ngx_rtmp_core_listen(ngx_conf_t *cf, ngx_command_t *cmd, void *conf)
{
    size_t                      len, off;
    in_port_t                   port;
    ngx_str_t                  *value;
    ngx_url_t                   u;
    ngx_uint_t                  i, m;
    ngx_module_t              **modules;
    struct sockaddr            *sa;
    ngx_rtmp_listen_t          *ls;
    struct sockaddr_in         *sin;
    ngx_rtmp_core_main_conf_t  *cmcf;
#if (NGX_HAVE_INET6)
    struct sockaddr_in6        *sin6;
#endif

    /* 指向 "listen xxx" 配置项，如 "listen 1935" */
    value = cf->args->elts;

    ngx_memzero(&u, sizeof(ngx_url_t));

    /* 将监听的端口赋给 u.url，如 "1935" */
    u.url = value[1];
    u.listen = 1;

    /* 解析该监听的端口 */
    if (ngx_parse_url(cf->pool, &u) != NGX_OK) {
        if (u.err) {
            ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                               "%s in "%V" of the "listen" directive",
                               u.err, &u.url);
        }

        return NGX_CONF_ERROR;
    }

    cmcf = ngx_rtmp_conf_get_module_main_conf(cf, ngx_rtmp_core_module);

    ls = cmcf->listen.elts;

    /* 遍历listen数组，若有元素，将待监听的地址与数组中的保存的地址进行比较，
     * 若有相同的，则返回错误，否则添加到listen数组中 */
    for (i = 0; i < cmcf->listen.nelts; i++) {

        sa = (struct sockaddr *) ls[i].sockaddr;

        if (sa->sa_family != u.family) {
            continue;
        }

        switch (sa->sa_family) {

#if (NGX_HAVE_INET6)
        case AF_INET6:
            off = offsetof(struct sockaddr_in6, sin6_addr);
            len = 16;
            sin6 = (struct sockaddr_in6 *) sa;
            port = sin6->sin6_port;
            break;
#endif

        default: /* AF_INET */
            off = offsetof(struct sockaddr_in, sin_addr);
            len = 4;
            sin = (struct sockaddr_in *) sa;
            port = sin->sin_port;
            break;
        }

        if (ngx_memcmp(ls[i].sockaddr + off, (u_char *) &u.sockaddr + off, len)
            != 0)
        {
            continue;
        }

        if (port != u.port) {
            continue;
        }

        ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                           "duplicate "%V" address and port pair", &u.url);
        return NGX_CONF_ERROR;
    }

    ls = ngx_array_push(&cmcf->listen);
    if (ls == NULL) {
        return NGX_CONF_ERROR;
    }

    ngx_memzero(ls, sizeof(ngx_rtmp_listen_t));

    ngx_memcpy(ls->sockaddr, (u_char *) &u.sockaddr, u.socklen);

    ls->socklen = u.socklen;
    ls->wildcard = u.wildcard;
    ls->ctx = cf->ctx;

#if (nginx_version >= 1009011)
    modules = cf->cycle->modules;
#else
    modules = ngx_modules;
#endif

    for (m = 0; modules[m]; m++) {
        if (modules[m]->type != NGX_RTMP_MODULE) {
            continue;
        }
    }

    /* 若监听端口后面还附加有其他条件，则进行处理，否则直接返回 */
    for (i = 2; i < cf->args->nelts; i++) {

        if (ngx_strcmp(value[i].data, "bind") == 0) {
            ls->bind = 1;
            continue;
        }

        if (ngx_strncmp(value[i].data, "ipv6only=o", 10) == 0) {
#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
            struct sockaddr  *sa;
            u_char            buf[NGX_SOCKADDR_STRLEN];

            sa = (struct sockaddr *) ls->sockaddr;

            if (sa->sa_family == AF_INET6) {

                if (ngx_strcmp(&value[i].data[10], "n") == 0) {
                    ls->ipv6only = 1;

                } else if (ngx_strcmp(&value[i].data[10], "ff") == 0) {
                    ls->ipv6only = 0;

                } else {
                    ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                                       "invalid ipv6only flags "%s"",
                                       &value[i].data[9]);
                    return NGX_CONF_ERROR;
                }

                ls->bind = 1;

            } else {
                len = ngx_sock_ntop(sa,
#if (nginx_version >= 1005003)
                                    ls->socklen,
#endif
                                    buf, NGX_SOCKADDR_STRLEN, 1);

                ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                                   "ipv6only is not supported "
                                   "on addr "%*s", ignored", len, buf);
            }

            continue;
#else
            ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                               "bind ipv6only is not supported "
                               "on this platform");
            return NGX_CONF_ERROR;
#endif
        }

        if (ngx_strncmp(value[i].data, "so_keepalive=", 13) == 0) {

            if (ngx_strcmp(&value[i].data[13], "on") == 0) {
                ls->so_keepalive = 1;

            } else if (ngx_strcmp(&value[i].data[13], "off") == 0) {
                ls->so_keepalive = 2;

            } else {

#if (NGX_HAVE_KEEPALIVE_TUNABLE)
                u_char     *p, *end;
                ngx_str_t   s;

                end = value[i].data + value[i].len;
                s.data = value[i].data + 13;

                p = ngx_strlchr(s.data, end, ':');
                if (p == NULL) {
                    p = end;
                }

                if (p > s.data) {
                    s.len = p - s.data;

                    ls->tcp_keepidle = ngx_parse_time(&s, 1);
                    if (ls->tcp_keepidle == (time_t) NGX_ERROR) {
                        goto invalid_so_keepalive;
                    }
                }

                s.data = (p < end) ? (p + 1) : end;

                p = ngx_strlchr(s.data, end, ':');
                if (p == NULL) {
                    p = end;
                }

                if (p > s.data) {
                    s.len = p - s.data;

                    ls->tcp_keepintvl = ngx_parse_time(&s, 1);
                    if (ls->tcp_keepintvl == (time_t) NGX_ERROR) {
                        goto invalid_so_keepalive;
                    }
                }

                s.data = (p < end) ? (p + 1) : end;

                if (s.data < end) {
                    s.len = end - s.data;

                    ls->tcp_keepcnt = ngx_atoi(s.data, s.len);
                    if (ls->tcp_keepcnt == NGX_ERROR) {
                        goto invalid_so_keepalive;
                    }
                }

                if (ls->tcp_keepidle == 0 && ls->tcp_keepintvl == 0
                    && ls->tcp_keepcnt == 0)
                {
                    goto invalid_so_keepalive;
                }

                ls->so_keepalive = 1;

#else

                ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                                   "the "so_keepalive" parameter accepts "
                                   "only "on" or "off" on this platform");
                return NGX_CONF_ERROR;

#endif
            }

            ls->bind = 1;

            continue;

#if (NGX_HAVE_KEEPALIVE_TUNABLE)
        invalid_so_keepalive:

            ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                               "invalid so_keepalive value: "%s"",
                               &value[i].data[13]);
            return NGX_CONF_ERROR;
#endif
        }

        if (ngx_strcmp(value[i].data, "proxy_protocol") == 0) {
            ls->proxy_protocol = 1;
            continue;
        }

        ngx_conf_log_error(NGX_LOG_EMERG, cf, 0,
                           "the invalid "%V" parameter", &value[i]);
        return NGX_CONF_ERROR;
    }

    return NGX_CONF_OK;
}

2.2 ngx_rtmp_add_ports: 将监听端口添加到 ports 数组中

static ngx_int_t ngx_rtmp_add_ports(ngx_conf_t *cf, ngx_array_t *ports,
    ngx_rtmp_listen_t *listen)
{
    in_port_t              p;
    ngx_uint_t             i;
    struct sockaddr       *sa;
    struct sockaddr_in    *sin;
    ngx_rtmp_conf_port_t  *port;
    ngx_rtmp_conf_addr_t  *addr;
#if (NGX_HAVE_INET6)
    struct sockaddr_in6   *sin6;
#endif

    sa = (struct sockaddr *) &listen->sockaddr;

    switch (sa->sa_family) {

#if (NGX_HAVE_INET6)
    case AF_INET6:
        sin6 = (struct sockaddr_in6 *) sa;
        p = sin6->sin6_port;
        break;
#endif

    default: /* AF_INET */
        sin = (struct sockaddr_in *) sa;
        p = sin->sin_port;
        break;
    }

    /* 首先比较 ports 数组中是否已经存在该待监听端口的，若已经存在，
     * 则不用再次添加了，直接跳到 found 下，否则添加一个新的端口 */
    port = ports->elts;
    for (i = 0; i < ports->nelts; i++) {
        if (p == port[i].port && sa->sa_family == port[i].family) {

            /* a port is already in the port list */

            port = &port[i];
            goto found;
        }
    }

    /* add a port to the port list */

    port = ngx_array_push(ports);
    if (port == NULL) {
        return NGX_ERROR;
    }

    port->family = sa->sa_family;
    port->port   = p;

    if (ngx_array_init(&port->addrs, cf->temp_pool, 2,
                       sizeof(ngx_rtmp_conf_addr_t))
        != NGX_OK)
    {
        return NGX_ERROR;
    }

found:
    
    /* 这里有可能一个端口对应几个 IP 地址 */
    addr = ngx_array_push(&port->addrs);
    if (addr == NULL) {
        return NGX_ERROR;
    }

    addr->sockaddr       = (struct sockaddr *) &listen->sockaddr;
    addr->socklen        = listen->socklen;
    addr->ctx            = listen->ctx;
    addr->bind           = listen->bind;
    addr->wildcard       = listen->wildcard;
    addr->so_keepalive   = listen->so_keepalive;
    addr->proxy_protocol = listen->proxy_protocol;
#if (NGX_HAVE_KEEPALIVE_TUNABLE)
    addr->tcp_keepidle   = listen->tcp_keepidle;
    addr->tcp_keepintvl  = listen->tcp_keepintvl;
    addr->tcp_keepcnt    = listen->tcp_keepcnt;
#endif
#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
    addr->ipv6only       = listen->ipv6only;
#endif

    return NGX_OK;
}

2.3 ngx_rtmp_optimize_servers

static char *ngx_rtmp_optimize_servers(ngx_conf_t *cf, ngx_array_t *ports)
{
    ngx_uint_t             i, p, last, bind_wildcard;
    ngx_listening_t       *ls;
    ngx_rtmp_port_t       *mport;
    ngx_rtmp_conf_port_t  *port;
    ngx_rtmp_conf_addr_t  *addr;

    port = ports->elts;
    for (p = 0; p < ports->nelts; p++) {

        ngx_sort(port[p].addrs.elts, (size_t) port[p].addrs.nelts,
                 sizeof(ngx_rtmp_conf_addr_t), ngx_rtmp_cmp_conf_addrs);

        addr = port[p].addrs.elts;
        last = port[p].addrs.nelts;

        /*
         * if there is the binding to the "*:port" then we need to bind()
         * to the "*:port" only and ignore the other bindings
         */

        if (addr[last - 1].wildcard) {
            addr[last - 1].bind = 1;
            bind_wildcard = 1;

        } else {
            bind_wildcard = 0;
        }

        i = 0;

        while (i < last) {

            if (bind_wildcard && !addr[i].bind) {
                i++;
                continue;
            }

            /* 创建并初始化 ngx_listening_t 结构体 */
            ls = ngx_create_listening(cf, addr[i].sockaddr, addr[i].socklen);
            if (ls == NULL) {
                return NGX_CONF_ERROR;
            }

            ls->addr_ntop = 1;
            ls->handler = ngx_rtmp_init_connection;
            ls->pool_size = 4096;

            /* TODO: error_log directive */
            ls->logp = &cf->cycle->new_log;
            ls->log.data = &ls->addr_text;
            ls->log.handler = ngx_accept_log_error;

            ls->keepalive = addr[i].so_keepalive;
#if (NGX_HAVE_KEEPALIVE_TUNABLE)
            ls->keepidle = addr[i].tcp_keepidle;
            ls->keepintvl = addr[i].tcp_keepintvl;
            ls->keepcnt = addr[i].tcp_keepcnt;
#endif

#if (NGX_HAVE_INET6 && defined IPV6_V6ONLY)
            ls->ipv6only = addr[i].ipv6only;
#endif

            mport = ngx_palloc(cf->pool, sizeof(ngx_rtmp_port_t));
            if (mport == NULL) {
                return NGX_CONF_ERROR;
            }

            ls->servers = mport;

            if (i == last - 1) {
                mport->naddrs = last;

            } else {
                mport->naddrs = 1;
                i = 0;
            }

            switch (ls->sockaddr->sa_family) {
#if (NGX_HAVE_INET6)
            case AF_INET6:
                if (ngx_rtmp_add_addrs6(cf, mport, addr) != NGX_OK) {
                    return NGX_CONF_ERROR;
                }
                break;
#endif
            default: /* AF_INET */
                if (ngx_rtmp_add_addrs(cf, mport, addr) != NGX_OK) {
                    return NGX_CONF_ERROR;
                }
                break;
            }

            addr++;
            last--;
        }
    }

    return NGX_CONF_OK;
}