TLS vulnerabilities

OpenSSH portable 4.1 on SUSE Linux, and possibly other platforms and versions, and possibly under limited configurations, allows remote attackers to determine valid usernames via timing discrepancies in which responses take longer for valid usernames than invalid ones, as demonstrated by sshtime. NOTE: as of 20061014, it appears that this issue is dependent on the use of manually-set passwords that causes delays when processing /etc/shadow due to an increased number of rounds.

packet.c in ssh in OpenSSH allows remote attackers to cause a denial of service (crash) by sending an invalid protocol sequence with USERAUTH_SUCCESS before NEWKEYS, which causes newkeys[mode] to be NULL.

The get_server_hello function in the SSLv2 client code in OpenSSL 0.9.7 before 0.9.7l, 0.9.8 before 0.9.8d, and earlier versions allows remote servers to cause a denial of service (client crash) via unknown vectors that trigger a null pointer dereference.

Buffer overflow in the SSL_get_shared_ciphers function in OpenSSL 0.9.7 before 0.9.7l, 0.9.8 before 0.9.8d, and earlier versions has unspecified impact and remote attack vectors involving a long list of ciphers.

OpenSSL 0.9.7 before 0.9.7l, 0.9.8 before 0.9.8d, and earlier versions allows attackers to cause a denial of service (CPU consumption) via parasitic public keys with large (1) "public exponent" or (2) "public modulus" values in X.509 certificates that require extra time to process when using RSA signature verification.

OpenSSL 0.9.7 before 0.9.7l and 0.9.8 before 0.9.8d allows remote attackers to cause a denial of service (infinite loop and memory consumption) via malformed ASN.1 structures that trigger an improperly handled error condition.

Unspecified vulnerability in portable OpenSSH before 4.4, when running on some platforms, allows remote attackers to determine the validity of usernames via unknown vectors involving a GSSAPI "authentication abort."

Signal handler race condition in OpenSSH before 4.4 allows remote attackers to cause a denial of service (crash), and possibly execute arbitrary code if GSSAPI authentication is enabled, via unspecified vectors that lead to a double-free.

sshd in OpenSSH before 4.4, when using the version 1 SSH protocol, allows remote attackers to cause a denial of service (CPU consumption) via an SSH packet that contains duplicate blocks, which is not properly handled by the CRC compensation attack detector.

OpenSSL before 0.9.7, 0.9.7 before 0.9.7k, and 0.9.8 before 0.9.8c, when using an RSA key with exponent 3, removes PKCS-1 padding before generating a hash, which allows remote attackers to forge a PKCS #1 v1.5 signature that is signed by that RSA key and prevents OpenSSL from correctly verifying X.509 and other certificates that use PKCS #1.

Heap-based buffer overflow in cURL and libcURL 7.15.0 through 7.15.2 allows remote attackers to execute arbitrary commands via a TFTP URL (tftp://) with a valid hostname and a long path.

OpenSSH on FreeBSD 5.3 and 5.4, when used with OpenPAM, does not properly handle when a forked child process terminates during PAM authentication, which allows remote attackers to cause a denial of service (client connection refusal) by connecting multiple times to the SSH server, waiting for the password prompt, then disconnecting.

scp in OpenSSH 4.2p1 allows attackers to execute arbitrary commands via filenames that contain shell metacharacters or spaces, which are expanded twice.

Multiple off-by-one errors in the cURL library (libcurl) 7.11.2 through 7.15.0 allow local users to trigger a buffer overflow and cause a denial of service or bypass PHP security restrictions via certain URLs that (1) are malformed in a way that prevents a terminating null byte from being added to either a hostname or path buffer, or (2) contain a "?" separator in the hostname portion, which causes a "/" to be prepended to the resulting string.

The SSL/TLS server implementation in OpenSSL 0.9.7 before 0.9.7h and 0.9.8 before 0.9.8a, when using the SSL_OP_MSIE_SSLV2_RSA_PADDING option, disables a verification step that is required for preventing protocol version rollback attacks, which allows remote attackers to force a client and server to use a weaker protocol than needed via a man-in-the-middle attack.

Stack-based buffer overflow in the ntlm_output function in http-ntlm.c for (1) wget 1.10, (2) curl 7.13.2, and (3) libcurl 7.13.2, and other products that use libcurl, when NTLM authentication is enabled, allows remote servers to execute arbitrary code via a long NTLM username.

The default configuration on OpenSSL before 0.9.8 uses MD5 for creating message digests instead of a more cryptographically strong algorithm, which makes it easier for remote attackers to forge certificates with a valid certificate authority signature.

sshd in OpenSSH before 4.2, when GSSAPIDelegateCredentials is enabled, allows GSSAPI credentials to be delegated to clients who log in using non-GSSAPI methods, which could cause those credentials to be exposed to untrusted users or hosts.

OpenSSH 4.0, and other versions before 4.2, does not properly handle dynamic port forwarding ("-D" option) when a listen address is not provided, which may cause OpenSSH to enable the GatewayPorts functionality.

SSH, as implemented in OpenSSH before 4.0 and possibly other implementations, stores hostnames, IP addresses, and keys in plaintext in the known_hosts file, which makes it easier for an attacker that has compromised an SSH user's account to generate a list of additional targets that are more likely to have the same password or key.

The design of Advanced Encryption Standard (AES), aka Rijndael, allows remote attackers to recover AES keys via timing attacks on S-box lookups, which are difficult to perform in constant time in AES implementations.

Multiple stack-based buffer overflows in libcURL and cURL 7.12.1, and possibly other versions, allow remote malicious web servers to execute arbitrary code via base64 encoded replies that exceed the intended buffer lengths when decoded, which is not properly handled by (1) the Curl_input_ntlm function in http_ntlm.c during NTLM authentication or (2) the Curl_krb_kauth and krb4_auth functions in krb4.c during Kerberos authentication.

The der_chop script in the openssl package in Trustix Secure Linux 1.5 through 2.1 and other operating systems allows local users to overwrite files via a symlink attack on temporary files.

sshd in OpenSSH 3.5p1, when PermitRootLogin is disabled, immediately closes the TCP connection after a root login attempt with the correct password, but leaves the connection open after an attempt with an incorrect password, which makes it easier for remote attackers to guess the password by observing the connection state, a different vulnerability than CVE-2003-0190. NOTE: it could be argued that in most environments, this does not cross privilege boundaries without requiring leverage of a separate vulnerability.

sshd.c in OpenSSH 3.6.1p2 and 3.7.1p2 and possibly other versions, when using privilege separation, does not properly signal the non-privileged process when a session has been terminated after exceeding the LoginGraceTime setting, which leaves the connection open and allows remote attackers to cause a denial of service (connection consumption).

The SSL/TLS handshaking code in OpenSSL 0.9.7a, 0.9.7b, and 0.9.7c, when using Kerberos ciphersuites, does not properly check the length of Kerberos tickets during a handshake, which allows remote attackers to cause a denial of service (crash) via a crafted SSL/TLS handshake that causes an out-of-bounds read.

OpenSSL 0.9.6 before 0.9.6d does not properly handle unknown message types, which allows remote attackers to cause a denial of service (infinite loop), as demonstrated using the Codenomicon TLS Test Tool.

The do_change_cipher_spec function in OpenSSL 0.9.6c to 0.9.6k, and 0.9.7a to 0.9.7c, allows remote attackers to cause a denial of service (crash) via a crafted SSL/TLS handshake that triggers a null dereference.

The default configuration for OpenSSH enables AllowTcpForwarding, which could allow remote authenticated users to perform a port bounce, when configured with an anonymous access program such as AnonCVS.

Directory traversal vulnerability in scp for OpenSSH before 3.4p1 allows remote malicious servers to overwrite arbitrary files. NOTE: this may be a rediscovery of CVE-2000-0992.

sshd in OpenSSH 3.6.1p2 and earlier, when PermitRootLogin is disabled and using PAM keyboard-interactive authentication, does not insert a delay after a root login attempt with the correct password, which makes it easier for remote attackers to use timing differences to determine if the password step of a multi-step authentication is successful, a different vulnerability than CVE-2003-0190.

OpenSSL 0.9.6k allows remote attackers to cause a denial of service (crash via large recursion) via malformed ASN.1 sequences.

The PAM conversation function in OpenSSH 3.7.1 and 3.7.1p1 interprets an array of structures as an array of pointers, which allows attackers to modify the stack and possibly gain privileges.

The SSH1 PAM challenge response authentication in OpenSSH 3.7.1 and 3.7.1p1, when Privilege Separation is disabled, does not check the result of the authentication attempt, which can allow remote attackers to gain privileges.

Double free vulnerability in OpenSSL 0.9.7 allows remote attackers to cause a denial of service (crash) and possibly execute arbitrary code via an SSL client certificate with a certain invalid ASN.1 encoding.

OpenSSL 0.9.6 and 0.9.7 does not properly track the number of characters in certain ASN.1 inputs, which allows remote attackers to cause a denial of service (crash) via an SSL client certificate that causes OpenSSL to read past the end of a buffer when the long form is used.

Integer overflow in OpenSSL 0.9.6 and 0.9.7 allows remote attackers to cause a denial of service (crash) via an SSL client certificate with certain ASN.1 tag values.

OpenSSL 0.9.6e uses assertions when detecting buffer overflow attacks instead of less severe mechanisms, which allows remote attackers to cause a denial of service (crash) via certain messages that cause OpenSSL to abort from a failed assertion, as demonstrated using SSLv2 CLIENT_MASTER_KEY messages, which are not properly handled in s2_srvr.c.

Multiple "buffer management errors" in OpenSSH before 3.7.1 may allow attackers to cause a denial of service or execute arbitrary code using (1) buffer_init in buffer.c, (2) buffer_free in buffer.c, or (3) a separate function in channels.c, a different vulnerability than CVE-2003-0693.

"Memory bugs" in OpenSSH 3.7.1 and earlier, with unknown impact, a different set of vulnerabilities than CVE-2003-0693 and CVE-2003-0695.

A "buffer management error" in buffer_append_space of buffer.c for OpenSSH before 3.7 may allow remote attackers to execute arbitrary code by causing an incorrect amount of memory to be freed and corrupting the heap, a different vulnerability than CVE-2003-0695.

OpenSSH 3.6.1 and earlier, when restricting host access by numeric IP addresses and with VerifyReverseMapping disabled, allows remote attackers to bypass "from=" and "user@host" address restrictions by connecting to a host from a system whose reverse DNS hostname contains the numeric IP address.

OpenSSH-portable (OpenSSH) 3.6.1p1 and earlier with PAM support enabled immediately sends an error message when a user does not exist, which allows remote attackers to determine valid usernames via a timing attack.

OpenSSL does not use RSA blinding by default, which allows local and remote attackers to obtain the server's private key by determining factors using timing differences on (1) the number of extra reductions during Montgomery reduction, and (2) the use of different integer multiplication algorithms ("Karatsuba" and normal).

The SSL and TLS components for OpenSSL 0.9.6i and earlier, 0.9.7, and 0.9.7a allow remote attackers to perform an unauthorized RSA private key operation via a modified Bleichenbacher attack that uses a large number of SSL or TLS connections using PKCS #1 v1.5 padding that cause OpenSSL to leak information regarding the relationship between ciphertext and the associated plaintext, aka the "Klima-Pokorny-Rosa attack."

ssl3_get_record in s3_pkt.c for OpenSSL before 0.9.7a and 0.9.6 before 0.9.6i does not perform a MAC computation if an incorrect block cipher padding is used, which causes an information leak (timing discrepancy) that may make it easier to launch cryptographic attacks that rely on distinguishing between padding and MAC verification errors, possibly leading to extraction of the original plaintext, aka the "Vaudenay timing attack."

sshd in OpenSSH 3.2.2, when using YP with netgroups and under certain conditions, may allow users to successfully authenticate and log in with another user's password.

The ASN1 library in OpenSSL 0.9.6d and earlier, and 0.9.7-beta2 and earlier, allows remote attackers to cause a denial of service via invalid encodings.

Buffer overflow in OpenSSL 0.9.7 before 0.9.7-beta3, with Kerberos enabled, allows attackers to execute arbitrary code via a long master key.

Buffer overflows in OpenSSL 0.9.6d and earlier, and 0.9.7-beta2 and earlier, allow remote attackers to execute arbitrary code via (1) a large client master key in SSL2 or (2) a large session ID in SSL3.