The idea behind email-injection is that attacker inject line feed (LF) in the email headers and so he adds as many headers as he wants. Stripping those line feeds will protect you from this attack. For detailed info check http://www.phpsecure.info/v2/article/MailHeadersInject.en.php
The best practice is to rely on a well-written, frequently updated and widely-used code. For that I would suggest using PEAR_MAIL OR Zend_Mail
If you don't want to load those modules or you need to keep things very simple. You can extract the filtering functionality from those modules. Although I do recommend to use them and frequently update the library so that if new attack appears in future you will just need to update your library (Pear or Zend) and you are done.
This is the function that sanitize headers in Pear Mail package:
function _sanitizeHeaders(&$headers)
{
foreach ($headers as $key => $value) {
$headers[$key] =
preg_replace('=((<CR>|<LF>|0x0A/%0A|0x0D/%0D|\n|\r)S).*=i',
null, $value);
}
}
Zend_Mail uses different filter for email,name and other fields:
function _filterEmail($email)
{
$rule = array("
" => '',
"
" => '',
"" => '',
'"' => '',
',' => '',
'<' => '',
'>' => '',
);
return strtr($email, $rule);
}
function _filterName($name)
{
$rule = array("
" => '',
"
" => '',
"" => '',
'"' => "'",
'<' => '[',
'>' => ']',
);
return trim(strtr($name, $rule));
}
function _filterOther($data)
{
$rule = array("
" => '',
"
" => '',
"" => '',
);
return strtr($data, $rule);
}
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…