It appears that mcrypt_decrypt pads the *RETURN STRING* with nulls ('\0') to fill out to n * blocksize. For old C-programmers, like myself, it is easy to believe the string ends at the first null. In PHP it does not:
strlen("abc\0\0") returns 5 and *NOT* 3
strcmp("abc", "abc\0\0") returns -2 and *NOT* 0
I learned this lesson painfully when I passed a string returned from mycrypt_decrypt into a NuSoap message, which happily passed the nulls along to the receiver, who couldn't figure out what I was talking about.
My solution was:
<?php
$retval = mcrypt_decrypt( ...etc ...);
$retval = rtrim($retval, "\0"); // trim ONLY the nulls at the END
?>
mcrypt_decrypt
(PHP 4 >= 4.0.2, PHP 5, PHP 7 < 7.2.0, PECL mcrypt >= 1.0.0)
mcrypt_decrypt — Decrypts crypttext with given parameters
Ostrzeżenie
This function has been DEPRECATED as of PHP 7.1.0 and REMOVED as of PHP 7.2.0. Relying on this function is highly discouraged.
Opis
mcrypt_decrypt
( string
$cipher
, string $key
, string $data
, string $mode
[, string $iv
] ) : string
Decrypts the data and returns the unencrypted data.
Parametry
-
cipher -
Jedna ze stałych
MCRYPT_szyfr, lub nazwa algorytmu jako ciąg znaków. -
key -
The key with which the data was encrypted. If the provided key size is not supported by the cipher, the function will emit a warning and return
FALSE -
data -
The data that will be decrypted with the given
cipherandmode. If the size of the data is not n * blocksize, the data will be padded with '\0'. -
mode -
Jedna ze stałych
MCRYPT_MODE_nazwatrybu, lub jeden z następujących ciągów znaków: "ecb", "cbc", "cfb", "ofb", "nofb" lub "stream". -
iv -
Stosowany dla inicjalizacji trybów CBC, CFB, OFB, oraz - przy niektórych algorytmach - trybu STREAM. Nie podając IV, jeśli algorytm tego wymaga, spowoduje wygenerowanie ostrzeżenia i zwrócenie
FALSEprzez funkcję.
Zwracane wartości
Returns the decrypted data as a string lub FALSE w przypadku niepowodzenia.
Rejestr zmian
| Wersja | Opis |
|---|---|
| 5.6.0 |
Invalid key and iv sizes
are no longer accepted. mcrypt_decrypt() will now throw
a warning and return FALSE if the inputs are invalid. Previously keys and
IVs were padded with '\0' bytes to the next valid size.
|
add a note
User Contributed Notes 8 notes
eddiec at stararcher dot com ¶
18 years ago
gmmarco at outlook dot my ¶
6 years ago
Look at http://php.net/manual/en/function.openssl-decrypt.php for a replacement.
Supports PHP 5 >= 5.3.0, PHP 7
david at sickmiller dot com ¶
15 years ago
If you happen to be decrypting something encrypted in ColdFusion, you'll discover that its encrypt function apparently pads the plaintext with ASCII 4, the "end of transmission" character.
Building on eddiec's code, you can remove both nulls and EOTs with this:
<?php
$retval = mcrypt_decrypt( ...etc ...);
$retval = rtrim($retval, "\0\4"); // trim ONLY the nulls and EOTs at the END
?>
pnz dot r00t at yandex dot ru ¶
8 years ago
When i using this function i find some problem, with adding additional binary symbols in decode message.
$sDecrypt – «2433091»
$sDecrypt strlen – 16 (before trim)
$sDecrypt – «2433091»
$sDecrypt strlen – 7 (after trim)
At local PC this problem solved by using trim/trim, but if I send my encrypt message to sever, and try encrypt them, I see that my message is not 16 symbols, as it was in my local pc, it is 32 sybmols, after encode. And trim function not working. I look at symbols code and what I see
50
52
51
51
48
57
49
0
0
0
0
0
0
0
0
0
10
158
112
183
154
27
95
85
42
35
95
54
227
41
179
77
After “tabulation symbols” I get some junk symbols, and I solve this problems like this.
After mcrypt_decrypt I use this code
$plaintext_dec = substr($plaintext_dec, 0, strpos($plaintext_dec, "\0"));
I think it is not best resolution for this problem, but I solve my problem :)
beltrachi ¶
12 years ago
Caution, MCRYPT_RIJNDAEL_256 is not equivalent to AES_256.
The way to make RIJNDAEL be decrypted from AES with openssl is to use MCRYPT_RIJNDAEL_128 and padd the string to encrypt before encrypting with the follwing function:
<?php
function pkcs5_pad ($text, $blocksize) {
$pad = $blocksize - (strlen($text) % $blocksize);
return $text . str_repeat(chr($pad), $pad);
}
?>
On the decryption, the choosing of AES_256 or AES_128, etc. is based on the keysize used in the crypting. In my case it was a 128bit key so I used AES_128.
evangelion207 at hotmail dot com ¶
13 years ago
Be careful, sometimes mcrypt_decrypt return additional white spaces to the uncrypted string; use trim() for deleting them. I was like 2 hours searching the error and it was that..
Anonymous ¶
13 years ago
To remove PKCS7 padding:
<?php
$decrypted = mdecrypt_generic($td, base64_decode($enc_auth_token));
$dec_s = strlen($decrypted);
$padding = ord($decrypted[$dec_s-1]);
$decrypted = substr($decrypted, 0, -$padding);
?>
artaxerxes2 at iname dot com ¶
10 years ago
To decrypt data coming from MySQL's AES_ENCRYPT function:
<?php
function mysql_aes_key($key)
{
$new_key = str_repeat(chr(0), 16);
for($i=0,$len=strlen($key);$i<$len;$i++)
{
$new_key[$i%16] = $new_key[$i%16] ^ $key[$i];
}
return $new_key;
}
function aes_decrypt($encrypted,$key)
{
// if $encrypted is HEXed, then return it to binary
$encrypted = pack('H*',$encrypted);
$key = mysql_aes_key($key);
return rtrim(mcrypt_decrypt(MCRYPT_RIJNDAEL_128,$key,$encrypted,MCRYPT_MODE_ECB,''),"\x00..\x1F");
}
?>
adapted from the article "Replicating MySQL AES Encryption Methods With PHP" (dated 2012-05-20) found somewhere online.
