你已经回答了你自己的问题!：）

function getRealIpAddr() {
    if(!empty($_SERVER['HTTP_CLIENT_IP']))   //Check IP address from shared Internet
    {
        $IPaddress = $_SERVER['HTTP_CLIENT_IP'];
    }
    elseif (!empty($_SERVER['HTTP_X_FORWARDED_FOR']))   //To check IP address is passed from the proxy
    {
        $IPaddress = $_SERVER['HTTP_X_FORWARDED_FOR'];
    }
    else
    {
        $IPaddress = $_SERVER['REMOTE_ADDR'];
    }
    return $IPaddress;
}

源

虽然这篇文章很老了，但这个话题仍然值得关注。所以我在我的项目中使用了另一个解决方案。我在这里找到了其他的解决方案，要么不完整，要么太复杂，难以理解。

if (! function_exists('get_visitor_IP'))
{
    /**
     * Get the real IP address from visitors proxy. e.g. Cloudflare
     *
     * @return string IP
     */
    function get_visitor_IP()
    {
        // Get real visitor IP behind CloudFlare network
        if (isset($_SERVER["HTTP_CF_CONNECTING_IP"])) {
            $_SERVER['REMOTE_ADDR'] = $_SERVER["HTTP_CF_CONNECTING_IP"];
            $_SERVER['HTTP_CLIENT_IP'] = $_SERVER["HTTP_CF_CONNECTING_IP"];
        }

        // Sometimes the `HTTP_CLIENT_IP` can be used by proxy servers
        $ip = @$_SERVER['HTTP_CLIENT_IP'];
        if (filter_var($ip, FILTER_VALIDATE_IP)) {
           return $ip;
        }

        // Sometimes the `HTTP_X_FORWARDED_FOR` can contain more than IPs 
        $forward_ips = @$_SERVER['HTTP_X_FORWARDED_FOR'];
        if ($forward_ips) {
            $all_ips = explode(',', $forward_ips);

            foreach ($all_ips as $ip) {
                if (filter_var($ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE | FILTER_FLAG_NO_RES_RANGE)){
                    return $ip;
                }
            }
        }

        return $_SERVER['REMOTE_ADDR'];
    }
}

来自Symfony的Request类 https://github.com/symfony/symfony/blob/1bd125ec4a01220878b3dbc3ec3156b073996af9/src/Symfony/Component/HttpFoundation/Request.php

const HEADER_FORWARDED = 'forwarded';
const HEADER_CLIENT_IP = 'client_ip';
const HEADER_CLIENT_HOST = 'client_host';
const HEADER_CLIENT_PROTO = 'client_proto';
const HEADER_CLIENT_PORT = 'client_port';

/**
 * Names for headers that can be trusted when
 * using trusted proxies.
 *
 * The FORWARDED header is the standard as of rfc7239.
 *
 * The other headers are non-standard, but widely used
 * by popular reverse proxies (like Apache mod_proxy or Amazon EC2).
 */
protected static $trustedHeaders = array(
    self::HEADER_FORWARDED => 'FORWARDED',
    self::HEADER_CLIENT_IP => 'X_FORWARDED_FOR',
    self::HEADER_CLIENT_HOST => 'X_FORWARDED_HOST',
    self::HEADER_CLIENT_PROTO => 'X_FORWARDED_PROTO',
    self::HEADER_CLIENT_PORT => 'X_FORWARDED_PORT',
);

/**
 * Returns the client IP addresses.
 *
 * In the returned array the most trusted IP address is first, and the
 * least trusted one last. The "real" client IP address is the last one,
 * but this is also the least trusted one. Trusted proxies are stripped.
 *
 * Use this method carefully; you should use getClientIp() instead.
 *
 * @return array The client IP addresses
 *
 * @see getClientIp()
 */
public function getClientIps()
{
    $clientIps = array();
    $ip = $this->server->get('REMOTE_ADDR');
    if (!$this->isFromTrustedProxy()) {
        return array($ip);
    }
    if (self::$trustedHeaders[self::HEADER_FORWARDED] && $this->headers->has(self::$trustedHeaders[self::HEADER_FORWARDED])) {
        $forwardedHeader = $this->headers->get(self::$trustedHeaders[self::HEADER_FORWARDED]);
        preg_match_all('{(for)=("?\[?)([a-z0-9\.:_\-/]*)}', $forwardedHeader, $matches);
        $clientIps = $matches[3];
    } elseif (self::$trustedHeaders[self::HEADER_CLIENT_IP] && $this->headers->has(self::$trustedHeaders[self::HEADER_CLIENT_IP])) {
        $clientIps = array_map('trim', explode(',', $this->headers->get(self::$trustedHeaders[self::HEADER_CLIENT_IP])));
    }
    $clientIps[] = $ip; // Complete the IP chain with the IP the request actually came from
    $firstTrustedIp = null;
    foreach ($clientIps as $key => $clientIp) {
        // Remove port (unfortunately, it does happen)
        if (preg_match('{((?:\d+\.){3}\d+)\:\d+}', $clientIp, $match)) {
            $clientIps[$key] = $clientIp = $match[1];
        }
        if (!filter_var($clientIp, FILTER_VALIDATE_IP)) {
            unset($clientIps[$key]);
        }
        if (IpUtils::checkIp($clientIp, self::$trustedProxies)) {
            unset($clientIps[$key]);
            // Fallback to this when the client IP falls into the range of trusted proxies
            if (null ===  $firstTrustedIp) {
                $firstTrustedIp = $clientIp;
            }
        }
    }
    // Now the IP chain contains only untrusted proxies and the client IP
    return $clientIps ? array_reverse($clientIps) : array($firstTrustedIp);
}

我想出了这个函数，它不简单地返回IP地址，而是一个具有IP信息的数组。

// Example usage:
$info = ip_info();
if ( $info->proxy ) {
    echo 'Your IP is ' . $info->ip;
} else {
    echo 'Your IP is ' . $info->ip . ' and your proxy is ' . $info->proxy_ip;
}

函数如下:

/**
 * Retrieves the best guess of the client's actual IP address.
 * Takes into account numerous HTTP proxy headers due to variations
 * in how different ISPs handle IP addresses in headers between hops.
 *
 * @since 1.1.3
 *
 * @return object {
 *         IP Address details
 *
 *         string $ip The users IP address (might be spoofed, if $proxy is true)
 *         bool $proxy True, if a proxy was detected
 *         string $proxy_id The proxy-server IP address
 * }
 */
function ip_info() {
    $result = (object) array(
        'ip' => $_SERVER['REMOTE_ADDR'],
        'proxy' => false,
        'proxy_ip' => '',
    );

    /*
     * This code tries to bypass a proxy and get the actual IP address of
     * the visitor behind the proxy.
     * Warning: These values might be spoofed!
     */
    $ip_fields = array(
        'HTTP_CLIENT_IP',
        'HTTP_X_FORWARDED_FOR',
        'HTTP_X_FORWARDED',
        'HTTP_X_CLUSTER_CLIENT_IP',
        'HTTP_FORWARDED_FOR',
        'HTTP_FORWARDED',
        'REMOTE_ADDR',
    );
    foreach ( $ip_fields as $key ) {
        if ( array_key_exists( $key, $_SERVER ) === true ) {
            foreach ( explode( ',', $_SERVER[$key] ) as $ip ) {
                $ip = trim( $ip );

                if ( filter_var( $ip, FILTER_VALIDATE_IP, FILTER_FLAG_NO_PRIV_RANGE | FILTER_FLAG_NO_RES_RANGE ) !== false ) {
                    $forwarded = $ip;
                    break 2;
                }
            }
        }
    }

    // If we found a different IP address then REMOTE_ADDR then it's a proxy!
    if ( $forwarded != $result->ip ) {
        $result->proxy = true;
        $result->proxy_ip = $result->ip;
        $result->ip = $forwarded;
    }

    return $result;
}

就像之前有人说的，这里的关键在于你为什么想要存储用户的ip。

我将从我工作的注册系统中给出一个例子，当然解决方案只是为了在这个经常出现在我搜索中的旧讨论中贡献一些东西。

许多php注册库使用ip来限制/锁定基于用户ip的失败尝试。 看看这张表:

-- mysql
DROP TABLE IF EXISTS `attempts`;
CREATE TABLE `attempts` (
  `id` int(11) NOT NULL AUTO_INCREMENT,
  `ip` varchar(39) NOT NULL, /*<<=====*/
  `expiredate` datetime NOT NULL,
  PRIMARY KEY (`id`)
) ENGINE=InnoDB DEFAULT CHARSET=utf8;
 -- sqlite
...

然后，当用户尝试登录或任何与重置密码相关的服务时，在开始时调用一个函数:

public function isBlocked() {
      /*
       * used one of the above methods to capture user's ip!!!
       */
      $ip = $this->ip;
      // delete attempts from this ip with 'expiredate' in the past
      $this->deleteAttempts($ip, false);
      $query = $this->dbh->prepare("SELECT count(*) FROM {$this->token->get('table_attempts')} WHERE ip = ?");
      $query->execute(array($ip));
      $attempts = $query->fetchColumn();
      if ($attempts < intval($this->token->get('attempts_before_verify'))) {
         return "allow";
      }
      if ($attempts < intval($this->token->get('attempts_before_ban'))) {
         return "captcha";
      }
      return "block";
   }

例如，$this->令牌->get('attempts_before_ban') === 10和2个用户来相同的ips，就像前面的代码中头部可以被欺骗的情况一样，然后在5次尝试后，两者都被禁止! 更糟糕的是，如果所有用户都来自同一个代理，那么只有前10个用户将被记录，其余的用户将被禁止!

这里的关键是我们需要一个表尝试的唯一索引，我们可以从这样的组合中获得它:

 `ip` varchar(39) NOT NULL,
 `jwt_load varchar(100) NOT NULL

其中jwt_load来自一个遵循json web令牌技术的HTTP cookie，我们只存储加密的有效负载，应该包含每个用户的任意/唯一值。 当然，请求应该修改为:"SELECT count(*) FROM {$this->token->get('table_attempts')} WHERE ip = ?AND jwt_load = ?"并且该类还应该初始化一个私有的$jwt。

我想知道也许你应该在爆炸的HTTP_X_FORWARDED_FOR以相反的顺序迭代，因为我的经验是，用户的IP地址结束在逗号分隔的列表的末尾，所以从头的开始，你更有可能得到一个返回的代理的IP地址，这可能仍然允许会话劫持，因为许多用户可能通过该代理。