无论您做什么，请确保不要信任从客户端发送的数据。$_SERVER['REMOTE_ADDR']包含连接方的真实IP地址。这是你能找到的最可靠的值。

然而，它们可能在代理服务器的后面，在这种情况下，代理可能已经设置了$_SERVER['HTTP_X_FORWARDED_FOR']，但这个值很容易被欺骗。例如，它可以由没有代理的人设置，或者IP可以是来自代理后面LAN的内部IP。

这意味着如果您要保存$_SERVER['HTTP_X_FORWARDED_FOR']，请确保您也保存$_SERVER['REMOTE_ADDR']值。例如，将两个值保存在数据库的不同字段中。

如果要将IP保存为字符串到数据库，请确保至少有45个字符的空间。IPv6将继续存在，这些地址比旧的IPv4地址更大。

(请注意，IPv6通常最多使用39个字符，但IPv4地址也有一个特殊的IPv6符号，其完整形式最多可达45个字符。所以如果你知道你在做什么，你可以使用39个字符，但如果你只是想设置并忘记它，请使用45)。

我最喜欢的解决方案是Zend Framework 2使用的方式。它还考虑$_SERVER属性HTTP_X_FORWARDED_FOR, HTTP_CLIENT_IP, REMOTE_ADDR，但是它声明了一个类来设置一些可信代理，并且它返回一个IP地址而不是一个数组。我认为这是最接近它的解决方案:

class RemoteAddress
{
    /**
     * Whether to use proxy addresses or not.
     *
     * As default this setting is disabled - IP address is mostly needed to increase
     * security. HTTP_* are not reliable since can easily be spoofed. It can be enabled
     * just for more flexibility, but if user uses proxy to connect to trusted services
     * it's his/her own risk, only reliable field for IP address is $_SERVER['REMOTE_ADDR'].
     *
     * @var bool
     */
    protected $useProxy = false;

    /**
     * List of trusted proxy IP addresses
     *
     * @var array
     */
    protected $trustedProxies = array();

    /**
     * HTTP header to introspect for proxies
     *
     * @var string
     */
    protected $proxyHeader = 'HTTP_X_FORWARDED_FOR';

    // [...]

    /**
     * Returns client IP address.
     *
     * @return string IP address.
     */
    public function getIpAddress()
    {
        $ip = $this->getIpAddressFromProxy();
        if ($ip) {
            return $ip;
        }

        // direct IP address
        if (isset($_SERVER['REMOTE_ADDR'])) {
            return $_SERVER['REMOTE_ADDR'];
        }

        return '';
    }

    /**
     * Attempt to get the IP address for a proxied client
     *
     * @see http://tools.ietf.org/html/draft-ietf-appsawg-http-forwarded-10#section-5.2
     * @return false|string
     */
    protected function getIpAddressFromProxy()
    {
        if (!$this->useProxy
            || (isset($_SERVER['REMOTE_ADDR']) && !in_array($_SERVER['REMOTE_ADDR'], $this->trustedProxies))
        ) {
            return false;
        }

        $header = $this->proxyHeader;
        if (!isset($_SERVER[$header]) || empty($_SERVER[$header])) {
            return false;
        }

        // Extract IPs
        $ips = explode(',', $_SERVER[$header]);
        // trim, so we can compare against trusted proxies properly
        $ips = array_map('trim', $ips);
        // remove trusted proxy IPs
        $ips = array_diff($ips, $this->trustedProxies);

        // Any left?
        if (empty($ips)) {
            return false;
        }

        // Since we've removed any known, trusted proxy servers, the right-most
        // address represents the first IP we do not know about -- i.e., we do
        // not know if it is a proxy server, or a client. As such, we treat it
        // as the originating IP.
        // @see http://en.wikipedia.org/wiki/X-Forwarded-For
        $ip = array_pop($ips);
        return $ip;
    }

    // [...]
}

在这里查看完整代码: https://raw.githubusercontent.com/zendframework/zend-http/master/src/PhpEnvironment/RemoteAddress.php

下面是一段代码，可以通过检查各种源来选择一个有效的IP。

首先，它检查'REMOTE_ADDR'是否是一个公共IP(并且不是您信任的反向代理之一)，然后遍历其中一个HTTP头，直到找到一个公共IP并返回它。(PHP 5.2 +)

只要反向代理是可信的，或者服务器与客户端直接连接，它就应该是可靠的。

//Get client's IP or null if nothing looks valid
function ip_get($allow_private = false)
{
  //Place your trusted proxy server IPs here.
  $proxy_ip = ['127.0.0.1'];

  //The header to look for (Make sure to pick the one that your trusted reverse proxy is sending or else you can get spoofed)
  $header = 'HTTP_X_FORWARDED_FOR'; //HTTP_CLIENT_IP, HTTP_X_FORWARDED, HTTP_FORWARDED_FOR, HTTP_FORWARDED

  //If 'REMOTE_ADDR' seems to be a valid client IP, use it.
  if(ip_check($_SERVER['REMOTE_ADDR'], $allow_private, $proxy_ip)) return $_SERVER['REMOTE_ADDR'];

  if(isset($_SERVER[$header]))
  {
    //Split comma separated values [1] in the header and traverse the proxy chain backwards.
    //[1] https://en.wikipedia.org/wiki/X-Forwarded-For#Format
    $chain = array_reverse(preg_split('/\s*,\s*/', $_SERVER[$header]));
    foreach($chain as $ip) if(ip_check($ip, $allow_private, $proxy_ip)) return $ip;
  }

   return null;
}

//Check for valid IP. If 'allow_private' flag is set to truthy, it allows private IP ranges as valid client IP as well. (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16)
//Pass your trusted reverse proxy IPs as $proxy_ip to exclude them from being valid.
function ip_check($ip, $allow_private = false, $proxy_ip = [])
{
  if(!is_string($ip) || is_array($proxy_ip) && in_array($ip, $proxy_ip)) return false;
  $filter_flag = FILTER_FLAG_NO_RES_RANGE;

  if(!$allow_private)
  {
    //Disallow loopback IP range which doesn't get filtered via 'FILTER_FLAG_NO_PRIV_RANGE' [1]
    //[1] https://www.php.net/manual/en/filter.filters.validate.php
    if(preg_match('/^127\.$/', $ip)) return false;
    $filter_flag |= FILTER_FLAG_NO_PRIV_RANGE;
  }

  return filter_var($ip, FILTER_VALIDATE_IP, $filter_flag) !== false;
}

答案是使用$_SERVER变量。例如，$_SERVER["REMOTE_ADDR"]将返回客户端的IP地址。

这个函数很紧凑，可以在任何地方使用。但是!

别忘了这一点!在这类函数或代码块中，不能保证记录用户的真实IP地址，因为有些用户可以使用代理或其他安全网关来实现不可见或无法跟踪

PHP函数:

function GetIP()
{
    if ( getenv("HTTP_CLIENT_IP") ) {
        $ip = getenv("HTTP_CLIENT_IP");
    } elseif ( getenv("HTTP_X_FORWARDED_FOR") ) {
        $ip = getenv("HTTP_X_FORWARDED_FOR");
        if ( strstr($ip, ',') ) {
            $tmp = explode(',', $ip);
            $ip = trim($tmp[0]);
        }
    } else {
        $ip = getenv("REMOTE_ADDR");
    }
    return $ip;
}

用法:

$IP = GetIP();或直接GetIP();

正如所有其他人所说，你可以使用$_SERVER['REMOTE_ADDR'];获取客户端IP地址。

另外，如果你需要一个用户的更多信息，你可以使用这个:

<?php
    $ip = '0.0.0.0';
    $ip = $_SERVER['REMOTE_ADDR'];
    $clientDetails = json_decode(file_get_contents("http://ipinfo.io/$ip/json"));
    echo "You're logged in from: <b>" . $clientDetails->country . "</b>";
?>

客户端更具体的信息放在$clientDetails中。 $clientDetails->PostalCode/hostname/region/loc…

我用的是ipinfo。IO来获取额外的信息。