Let's say I have a file called query.sql with the following content in it:
SELECT * FROM `users` WHERE `id`!=".$q->Num($_POST['id'])."
And in my php-script, which has a html form with input named "id" in it, I do the following trick:
$sql=file_get_contents('query.sql');
$query= eval("return \"$sql\";");
//here follows something like $mysqli->query($query); and so on..
I am not concerned about sql-injections since I'm using prepared statements and $q->Num performs is_int check.
But is it safe to use eval such way?
As far as I understand, what is actually eval-ed here is "${_POST['id']}" and it evals to some string value the user entered. And this becomes dangerous only if I eval this string second time. While I eval string only once user's input is just string and can not be interpreted as php-code by compiler and no php-injection is possible.
UPDATE
Thank you for proposing different methodologies and stressing need to use prepared statements. But this not my question at all.
My question is all about php-injections. Is such use of eval bad? If yes, why?
There is no need to use eval - put in a token, and replace it:
// file query.sql
SELECT * FROM `users` WHERE `id`!="{id}";
//php
$sql = file_get_contents('query.sql');
$query = str_replace("{id}", $_POST['id'], $sql);
Update
No, it's not safe. Someone could edit your query.sql script to do anything you want. You may say "the app is internal only", or "i have permissions locked down" or whatever - but at the end of the day there are no guarantees.
The eval statement - although I would try to find a way without using eval - is not vulnerable for PHP injection because $sql is enclosed in double quotes "; One can not ending this quoting with a prepared variable in PHP.
I am not conserned about sql-injections since I'm using prepared statements
Aren't you? ;) You are!
Why do you add the $id to the query using the '.' operator (string manipulation)? If you really use the benefits from prepared statements I would expect something like a bindParam()
Note how prepared statements prevent from SQL injections: The SQL query syntax is been kept separate from arguments. So the server would
parse the SQL query
apply arguments
As the query has been already parsed before arguments will been applied, the query syntax cannot be manipulated by the arguments.
If you prepare a MySQL query that has been created using '.' and external inputs you are potentially vulnerable against SQL injections
What you are doing defeats the principals of prepared statements
Reference from this Answer.
https://stackoverflow.com/a/951868/627868
The main problems with eval() are:
Potential unsafe input. Passing an untrusted parameter is a way to
fail. It is often not a trivial task to make sure that a parameter (or
part of it) is fully trusted.
Trickyness. Using eval() makes code clever, therefore more difficult
to follow. To quote Brian Kernighan "Debugging is twice as hard as
writing the code in the first place. Therefore, if you write the code
as cleverly as possible, you are, by definition, not smart enough to
debug it"
Let's consider the following code:
$sql=file_get_contents('query.sql');
$query= eval("return \"$sql\";");
The danger points at this are:
if the file query.sql is modified from what you expect, then it could be used to execute any arbitrary code in your program.
the name of the file is shown hard-coded; if this isn't the case, then a malicious user could find a way to load an unexpected file (possibly even one from a remote site), again resulting in arbitrary code execution.
the only reason to use a file for this (rather than hard coding the SQL code directly in the program) would be because you want to use it as a config file. The problem here is that the syntax in this file is invalid for both SQL and PHP. Due to the way it's run in the eval(), it also requires that the syntax is exactly correct. Use the wrong quotes or miss one out, and it'll blow up. This is likely to result in brittle code, that fails badly rather than gracefully when the config is marginally incorrect.
There doesn't appear to be a direct SQL injection attack here, but that's really the least of your worries when it comes to eval().
I have personally worked in projects where code existed that worked pretty much exactly the way you've described. There were some very nasty bugs in the system as a direct result of this, and they have been difficult to rectify without wholesale rewrites. I would strongly recommend stepping back from this idea and using a sensible templating mechanism instead as recommended by others in the comments.
Related
P.S: I know it is a best practice to user prepared statements all the time, and it should be a habit to use prepared statements. So putting this aside, please just let me know technically what can go wrong if I don't write below scenario with prepare.
1. I am not getting any input from users or any other class file, and I won't in the future.
2. I am getting an input from a variable in the same PHP file (an array for example).
Example: ($myID will be a variable hardcoded in the same PHP file)
$myID=12;
$wpdb->query("UPDATE `$table_name` SET `your_column_1` = 1 WHERE `myTable`.`your_column_id` = $myID");
Tl;DR This is a very bad idea. You are introducing long-term risk to save a few seconds of coding effort. It is overwhelmingly likely that you will, sooner or later, introduce a SQL injection risk as your code and data evolve.
If you:
are extremely careful about validating your data,
are extremely careful about constructing your queries,
are absolutely certain that your data is safe and free of any user input (from any source at all, including forms, sensors, APIs, scraping websites, etc.), and
are absolutely certain that no one will ever modify your data or your code (or reuse your code inappropriately), including you,
Then we can say:
it would be safe to execute queries without prepared statements, and - and this is the important part -
you would be living in a fantasy world.
You can never assume safely that you can avoid preparing statements. It’s highly, highly likely that your code will break if you do. You could have bad data, overlook a problem, change the code yourself somehow, repurpose the code for something insecure, or make any number of other errors. You may end up doing harmless SQL injection yourself (with, say, a name or spelling like O’Brien) or you may get massively hacked...
...all to save a few seconds of coding time.
Short version: either (1) use prepared statements on every query that has any kind of variable information or (2) learn the hard way why that’s the rule.
This is a wordpress plugin and will only be used by me in admin panel. So the problem caused because of using "in" statement because it is hard to write the query like update column where color in ('black','white').
If you're developing for WordPress, have you considered using the wpdb API? It makes it pretty easy to add parameters to your SQL queries.
Example of using parameters for an IN( ) predicate:
$colors_array = ["black", "white"];
$placeholders = array_fill(0, count($colors_array), "%s");
$placeholder_list = implode(",", $placeholders);
$wpdb->query( $wpdb->prepare(
"
UPDATE $wpdb->stock
SET quantity = 327
WHERE color IN ($placeholder_list)
",
$colors_array
));
See https://codex.wordpress.org/Class_Reference/wpdb#Protect_Queries_Against_SQL_Injection_Attacks
I agree with Ed Cottrell's advice that you should not compromise on secure programming methods. Use the most secure method and use it consistently.
You don't have to waste time thinking about whether any given case is "safe enough" to skip using the secure method.
You don't have to worry if it's still safe after your PHP variables are no longer hard-coded.
You don't have to worry that someone will copy & paste your code as an example, and they use it in an unsafe way.
This question already has answers here:
Why is using a mysql prepared statement more secure than using the common escape functions?
(7 answers)
Closed 9 years ago.
Okay, I still don't really get it. I keep reading that in order to properly escape your MySQL queries, you need to use mysqli_prepare() and mysqli_bind_param().
I tried using this setup and, quite frankly, it's a little clunkier. I'm stuck passing variables by reference when I don't need to ever reference them again, and it's just more lines of code to accomplish the same task.
I guess I just don't get what the difference is between:
<?php
$sql = new \MySQLi(...);
$result = $sql->query('
UPDATE `table`
SET
`field` = "'.$sql->real_escape_string($_REQUEST[$field]).'";
');
?>
and
<?php
$sql = new \MySQLi(...);
$stmt = $sql->prepare('
UPDATE `table`
SET
`field` = ?;
');
$value = $_REQUEST[$field];
$stmt->bind_param('s', $value);
$stmt->execute();
$result = $stmt->get_result();
unset($value);
?>
other than more code.
I mean, did they implement this so that people wouldn't forget to escape values before sending them in a query? Or is it somehow faster?
Or should I use this method when I intend to use the same query repeatedly (since a mysqli_stmt can be reused) and use the traditional method in other cases?
What you are reading, that you need to use mysqli_prepare() and mysqli_bind_param() functions to "properly escape your MySQL queries" is wrong.
It is true that if you use mysqli_prepare() and mysqli_bind_param(), you needn't (and shouldn't) "escape" the values supplied as bind parameters. So, in that sense, there's some truth in what you are reading.
It's only when unsafe variables are included in the SQL text (the actual text of the query) that you need to "properly escape" the variables, usually by wrapping the variables in mysqli_real_escape_string() function calls.
(We note that it's possible to make of use of prepared statements and still include un-escaped variables in the SQL text, rather than passing the variable values as bind_parameters. That does sort of defeats the purpose of using prepared statements, but the point is, either way, you can write code that is vulnerable.
MySQL now supports "server side" prepared statements (if the option is enabled in the connection), and that's a performance optimization (in some cases) of repeated executions of identical SQL text. (This has been long supported in other databases, such as Oracle, where making use of prepared statements has been a familiar pattern for, like, since forever.)
Q: Did they implement [prepared statements] so that people wouldn't forget to escape values before sending them in a query?
A: Based on the number of examples of code vulnerable to SQL Injection when not using prepared statements, despite the documentation regarding mysql_real_escape_string() function, you'd think that certainly would be sufficient reason.
I think one big benefit is that when we're reading code, we can see a SQL statement as a single string literal, rather than a concatenation of a bunch of variables, with quotes and dots and calls to mysql_real_escape_string, which isn't too bad with a simple query, but with a more complex query, it is just overly cumbersome. The use of the ? placeholder makes for a more understandable SQL statement,... true, I need to look at other lines of code to figure out what value is getting stuffed there. (I think the Oracle style named parameters :fee, :fi, :fo, :fum is preferable to the positional ?, ?, ?, ? notation.) But having STATIC SQL text is what is really the benefit.
Q: Or is it somehow faster?
As I mentioned before, the use of server side prepared statements can be and advantage in terms of performance. It's not always the case that it's faster, but for repeated execution of the same statement, where the only difference is literal values (as in repeated inserts), it can provide a performance boost.
Q: Or should I use this method when I intend to use the same query repeatedly (since a mysqli_stmt can be reused) and use the traditional method in other cases?
That's up to you. My preference is for using STATIC SQL text. But this really comes from a long history of using Oracle, and using the same pattern with MySQL fits naturally. (Albeit, from Perl using the DBI interface, and Java using JDBC and MyBATIS, or other ORMs (Hibernate, Glassfish JPA, et al.)
Following the same pattern just feels natural in PHP; the introduction of mysqli_ and PDO are a welcome relief from the arcane (and abused) mysql_ interface.
Good code can be written following either pattern. But I challenge you to think ahead, about more complex SQL statements, and whether the choice to use mysqli_real_escape_string() and concatenating together a dynamic string to be executed, rather than using static SQL text and bind parameters, might make reading, and deciphering, the actual SQL being executed more complicated for the soul that finds themselves maintaining code they didn't write.
I think studies have shown that code is read ten times more than it is written, which is why we strive to produce readable, understandable code, even if that means more lines of code. (When each statement is doing a single identifiable thing, that's usually easier for me to understand than reading a jumble of concatenated function calls in one complicated statement.
I think it's less a question of the latter method being more secure per se than encouraging separation of logic. With prepared statements the SQL query is independent of the values we use. This means, for example, when we go back and change our query we don't have to concatenate a bunch of different values to a string, and maybe risk forgetting to escape our input. Makes for more maintainable code!
There are a couple main benefits I found that were well written:
The overhead of compiling and optimizing the statement is incurred
only once, although the statement is executed multiple times. Not
all optimization can be performed at the time the prepared statement
is compiled, for two reasons: the best plan may depend on the
specific values of the parameters, and the best plan may change as
tables and indexes change over time.
Prepared statements are resilient against SQL injection, because
parameter values, which are transmitted later using a different
protocol, need not be correctly escaped. If the original statement
template is not derived from external input, SQL injection cannot
occur.
On the other hand, if a query is executed only once, server-side prepared statements can be slower because of the additional round-trip to the server. Implementation limitations may also lead to performance penalties: some versions of MySQL did not cache results of prepared queries, and some DBMSs such as PostgreSQL do not perform additional query optimization during execution.
Source
I would like to add that mysqli_bind_param() has been removed as of PHP 5.4.0. You should use mysqli_stmt_bind_param()
Right now I am using htmlspecialchars(mysql_escape_string($value)), but is there a way to sanitize it with one statement rather than a nested statement?
Well there's no one function that handles both of them.
You can use prepared statements and html puffier class, maybe then the "look and feel" will be little bit better :)
mysql_real_escape_string has actually fallen out of favor lately.
It is now preferred to use PDO or mysqli. They both come with PHP by default. They use something called parameterized queries to access the database, rather than having you write the SQL command yourself. This means that you don't need to worry about escaping anymore, since the query and the variables are passed into the function separately.
You can learn more about PDO here:
http://net.tutsplus.com/tutorials/php/why-you-should-be-using-phps-pdo-for-database-access/
On a related note, it is conventional to store user-supplied input into the database "as it was written", rather than using htmlspecialchars. You should then use htmlspecialchars to escape the data wherever the it appears on the site. This is a convention recommended by OWASP.
This is because you need to escape different things depending on context. This string:
' <script src="http://example.org/malice.js"></script> ]}\\\\--
...will need to be treated differently if it is used as a parameter in JSON (the quotes and backslashes and ] and } need to be escaped), HTML (the quotes and <s need to be escaped), or written as a URL (almost everything needs to be escaped). If you need to spend time instructing your JavaScript to un-encode the HTML, then your code is going to be confusing quickly.
This approach also makes fixing bugs simpler: if your site has a bug where content isn't escaped properly on a single page, then you can update the page and everything is fixed. If your site has a bug where the data is getting stored in the database incorrectly, then you need to fix everything in the database (which will take much longer and harm more users).
I'm developing a simple PHP database application for internal use but would like to code it to best practices. Some of my pages are receiving integer values from GET requests and I'm just wondering how much validation and sanitation is really required.
Currently I'm using $num = filter_input(INPUT_GET, 'num', FILTER_VALIDATE_INT, $num_options); with specified min and max values. From here I'm exiting with an error message if $num == false
Is it necessary to also use $mysqli->real_escape_string($num);
Currently I am not bothering because I think it's quite hard to do SQL injection using an integer...
Thanks,
Kevin
UPDATE: To clarify the query I'm doing looks like this
$sql = "SELECT employeeID, concat(FirstName, ' ', LastName) as Name FROM employee WHERE employeeID='$num'";
I see your using mysqli, your best option for security is to look into Prepared Statements.
PHP mysqli Prepared Statements
It's a bit involved for an example, but the above link has indepth examples.
Once you get the hang of it though, and build your class. It's really only a normal sql query but instead of including your values you use ?
"SELECT * FROM account WHERE username = ? AND password = ?"
and you bind your values to the statement:
array("bradley", "Passw0rd");
The security comes from, as a short answer, is the fact you don't concat the values into the query string yourself. Making it less prone to sql injection.
Like many other PHP users you are taking escaping wrong. You taking it as a some sort of magic wand which makes some "evil characters" "safe".
This is wrong idea.
though prepared statements can be taken as a sort of such a magic wand, escaping is not a synonym for "SQL injection protection". It is a merely string syntax rule - no more, no less.
Is it necessary to also use $mysqli->real_escape_string($num);
It is irrelevant question.
To escape or not to escape decision have to be bound to SQL, not to the data source or any validations:
real_escape_string() have to be used for the sql strings, i.e. parts of the query enclosed in quotes. Have to be used unconditionally, despite of whatever previous manipulations.
For the any other part of the query real_escape_string() being completely useless.
An explanation:
Data validation rules can be changed.
While SQL building rules have to be explicit and unconditional. To make a developer never ask himself a question like this.
In fact, it's completely different matters: data validation and query building. Why keep in mind such details and build the query accordingly? Why not to build the query based on some set of general purpose rules, irrelevant of the data nature at all?
So, to your question again:
if you are adding your data to the query as is, without quotes, real_escape_string() going to be completely useless in this case, but casting/validation become essential.
if you are adding your data to the query using prepared statement, real_escape_string() going to be completely useless and even harmful.
if you are adding your data to the query in quotes - you ought to do real_escape_string() in this case.
it is also worth to mention that if you are adding your data to the query as a part of SQL language - as an identifier or an SQL keyword - real_escape_string() is completely useless too, as well as prepared statement. Whitelisting is your only friend here
I have a feedback form which will take a couple of user inputted fields along with a few fields generated by PHP functions like 'user-agent' and 'referer.'
My question is should these strings be sanitized before being inputted? I realize one could easily alter the user-agent and the referring page, but could it be possible for a visitor to add a SQL injection like string so when PHP pulls this info it potentially breaks my form?
For instance if a user changed their user-agent or referring page to include the string Robert'); DROP TABLE Students;--
The word "sanitize" is pretty ambiguous and and better to be avoided.
Speaking of a database interaction, there is no need to "sanitize" at all. Just use prepared statements.
What is even more important, the data source doesn't matter. It should never be a question, "should we properly handle the data from such and such source?". It's just illogical, if you think of it. Why making such a distinction? Why rely on such a vague judgement? Why not to have an established process that uniformly treats any data despite the source?
Not to mention it's just super simple to use prepared statements:
$stmt = $db->prepare("INSERT INTO log (user_agent, referrer) VALUES (?,?)");
$stmt->execute([$_SERVER['HTTP_USER_AGENT'],$_SERVER['HTTP_REFERER']]);
And it will not only make the code simpler yet secure, but also make it proof against human errors of all sorts.
Simple Answer: validate/sanitize/escape everything (like client-side data, for example) because everything could be modified and evil or contain unexpected characters that could break your query (like Col. Shrapnel explained).
to minimize risk you should also about using prepared statements instead of building SQL-strings on your own (Note: this doesn't mean you can leave out the checks).
First off all - I believe the best practice is to parametr-ise everything in the query including self generated values. For me it does not make the query (almost) bulletproof but it creates much nicer and readable queries.
When you use parameters and assign them later you use more explicit logic in your code and therefore it will function better in the long term.
Longer explanation can be found in the attached link:
How can I prevent SQL injection in PHP?
Always sanitize/filter any input from a browser.
Just assume all users are evil, and you should be fine.
A connection doesn't have to come from a browser - anyone can write their own HTTP requests with a telnet client. There are probably specialized tools for this as well, and they wouldn't be hard to create.