Home >Backend Development >PHP Tutorial >Are sessions in PHP safe? _php tips
Are sessions in PHP safe? _php tips
- WBOYOriginal
- 2016-05-16 19:59:361596browse
I have been doing PHP development for so long and I have never really paid attention to security issues. I always focus on completing the project. Recently I saw an article about security on the Internet. After reading it, I noticed that I had previously All the projects had big security vulnerabilities, so I picked one project and tested it, and found that it was easy to get caught. Here I will share a test example I wrote to illustrate how the session in PHP is unsafe and how to strengthen its security in the project.
Regarding the principle and mechanism of session, there are many good articles on the Internet to introduce it, and we can check it by ourselves. Let’s share examples for testing directly.
The example of this test is mainly a login page. After successful login, you can change the password. It is such a simple function.
The interface is as follows
First, use the function session_start() at the project entrance to open the session. In this way, when the client initiates a request, an identity identifier, namely SessionID, will be generated. It is saved on the client through a cookie. Each communication between the client and the server relies on this SessionID for identification.
After successful login, the user ID and user name will be stored in the session
$_SESSION[‘userid'] = 用户id $_SESSION[‘uname'] = 用户名
All future operations will check whether the user is logged in by judging whether $_SESSION['userid'] exists. The code is as follows:
if(isset($_SESSION['userid'])) return true;
The call to the password modification interface transmits data to the server through ajax post.
$.post("接口*******",
{
oldpass:oldpass,
newpass:newpass,
userid:uid,
},
function(data){
data = eval('(' +data+ ')');
$('.grant_info').html(infos[data.info]).show();
}
);
Note that I have written this code in the html page, so if you see the html code, you will know the interface address.
The interface for changing the password is implemented in this way. First, it is judged whether the user is logged in. If the user is logged in, the password modification operation will be performed.
The implementation idea of the test example is roughly as described above.
Using SessionID Attack
1. The first is to obtain the SessionID. Of course, there are many ways for attackers to obtain this ID. Due to my limited level, I will not introduce how to obtain it here. We can simulate it by first accessing this project normally, and then checking the SessionID through the browser to get a legal user ID. This ID can be seen in the request header
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Encoding: gzip, deflate Accept-Language: zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3 Connection: keep-alive Cookie: Hm_lvt_bf1154ec41057869fceed66e9b3af5e7=1450428827,1450678226,1450851291,1450851486; PHPSESSID=2eiq9hcpu3ksri4r587ckt9jt7; Host: ****** Referer: ****** User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:41.0) Gecko/20100101 Firefox/41.0
After getting the session ID, if the user logs in successfully, then the user's information will be in the session on the server side.
2. After obtaining the SessionID, if the attacker already knows the interface for changing the password, he can directly change the user's password. If the attacker has not yet obtained the interface address, he or she can find out the interface address by looking at the page code. You can use the following command
#curl --cookie "PHPSESSID=2eiq9hcpu3ksri4r587ckt9jt7" 页面地址
As we said above, in this example the ajax code is written in the html page, so the interface address can be viewed on this page
Part of the html code is as follows
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
……
var uid = $(".userid").val();
$.post("/User/User/modifypass_do",
{
oldpass:oldpass,
newpass:newpass,
userid:uid,
},
function(data){
data = eval('(' +data+ ')');
$('.grant_info').html(infos[data.info]).show();
}
);
……
<span><input type="password" name="oldpass" id="textfield_o" placeholder="原密码"></span>
<span><input type="password" name="newpass" id="textfield_n" placeholder="新密码"></span>
<span><input type="password" name="confirmpass" id="textfield_c" placeholder="确认密码"></span>
<input type="button" class="btn_ok" value="确认修改" />
3. After getting the interface, you can use curl to simulate post to send data to change the password
The command is as follows
# curl --cookie "PHPSESSID=2eiq9hcpu3ksri4r587ckt9jt7" -d oldpass=111111 -d newpass=000000 -d userid=用户id 接口地址
If this user is already logged in, the attacker can change the user's password by executing the above command.
Solution
For the above attacks, we can enhance its security by complicating the verification method. One way is to use the User-Agent item in the request header to enhance its security
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 Accept-Encoding: gzip, deflate Accept-Language: zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3 Connection: keep-alive Cookie: Hm_lvt_bf1154ec41057869fceed66e9b3af5e7=1450428827,1450678226,1450851291,1450851486; PHPSESSID=2eiq9hcpu3ksri4r587ckt9jt7; Host: ****** Referer: ****** User-Agent: Mozilla/5.0 (Windows NT 6.1; rv:41.0) Gecko/20100101 Firefox/41.0
At the beginning of the project, we just used the session_start() function to start the session. Now we can add this code below session_start()
$_SESSION[‘User_Agent'] = md5($_SERVER[‘HTTP_USER_AGENT']);
Then every time when judging whether to log in, add the following judgment conditions
If(isset($_SESSION[‘userid']) && $_SESSION[‘User_Agent'] == md5($_SERVER[‘HTTP_USER_AGENT'])){
return true;
}
This way you can avoid the simple attacks mentioned above.
Summary:
Of course, the actual attack is far from that simple. First, it is difficult to obtain the SessionID. Then, the code interacting with the server must be encrypted as much as possible to avoid the above situation. After we modify the code for the second time, we can increase the complexity of the attack, but it cannot eliminate the attack. There are many ways to attack. This is just a simple way and only provides one idea, but the principle is the same. In actual situations, the security of our code can be enhanced according to the actual situation.
Here I am just sharing the problems I encountered at work, and I hope everyone can learn more in depth.