Inproper handling of ASCII representations of integers on 64 bit platforms allowed remote attackers to cause a denial of service or possibly execute arbitrary code. A buffer overflow allowed remote attackers to execute arbitrary code by sending a large client master key in SSL2 or a large session ID in SSL3. A buffer overflow when Kerberos is enabled allowed attackers to execute arbitrary code by sending a long master key. Note that this flaw did not affect any released version of 0.9.6 or 0.9.7 A flaw in the ASN1 library allowed remote attackers to cause a denial of service by sending invalid encodings. The use of assertions when detecting buffer overflow attacks allowed remote attackers to cause a denial of service (crash) by sending certain messages to cause OpenSSL to abort from a failed assertion, as demonstrated using SSLv2 CLIENT_MASTER_KEY messages, which were not properly handled in s2_srvr.c. sl3_get_record in s3_pkt.c did not perform a MAC computation if an incorrect block cipher padding was used, causing 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." The SSL and TLS components allowed 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 caused OpenSSL to leak information regarding the relationship between ciphertext and the associated plaintext, aka the "Klima-Pokorny-Rosa attack" RSA blinding was not enabled by default, which could allow local and remote attackers to obtain a 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). An integer overflow could allow remote attackers to cause a denial of service (crash) via an SSL client certificate with certain ASN.1 tag values. Incorrect tracking of the number of characters in certain ASN.1 inputs could allow remote attackers to cause a denial of service (crash) by sending an SSL client certificate that causes OpenSSL to read past the end of a buffer when the long form is used. Certain ASN.1 encodings that were rejected as invalid by the parser could trigger a bug in the deallocation of the corresponding data structure, corrupting the stack, leading to a crash. A flaw in OpenSSL 0.9.6k (only) would cause certain ASN.1 sequences to trigger a large recursion. On platforms such as Windows this large recursion cannot be handled correctly and so the bug causes OpenSSL to crash. A remote attacker could exploit this flaw if they can send arbitrary ASN.1 sequences which would cause OpenSSL to crash. This could be performed for example by sending a client certificate to a SSL/TLS enabled server which is configured to accept them. The Codenomicon TLS Test Tool uncovered a null-pointer assignment in the do_change_cipher_spec() function. A remote attacker could perform a carefully crafted SSL/TLS handshake against a server that used the OpenSSL library in such a way as to cause a crash. The Codenomicon TLS Test Tool found that some unknown message types were handled incorrectly, allowing a remote attacker to cause a denial of service (infinite loop). A flaw in SSL/TLS handshaking code when using Kerberos ciphersuites. A remote attacker could perform a carefully crafted SSL/TLS handshake against a server configured to use Kerberos ciphersuites in such a way as to cause OpenSSL to crash. Most applications have no ability to use Kerberos ciphersuites and will therefore be unaffected. The der_chop script created temporary files insecurely which could allow local users to overwrite files via a symlink attack on temporary files. Note that it is quite unlikely that a user would be using the redundant der_chop script, and this script was removed from the OpenSSL distribution. A deprecated option, SSL_OP_MISE_SSLV2_RSA_PADDING, could allow an attacker acting as a "man in the middle" to force a connection to downgrade to SSL 2.0 even if both parties support better protocols. Daniel Bleichenbacher discovered an attack on PKCS #1 v1.5 signatures where under certain circumstances it may be possible for an attacker to forge a PKCS #1 v1.5 signature that would be incorrectly verified by OpenSSL. During the parsing of certain invalid ASN.1 structures an error condition is mishandled. This can result in an infinite loop which consumes system memory Certain types of public key can take disproportionate amounts of time to process. This could be used by an attacker in a denial of service attack. A buffer overflow was discovered in the SSL_get_shared_ciphers() utility function. An attacker could send a list of ciphers to an application that uses this function and overrun a buffer. A flaw in the SSLv2 client code was discovered. When a client application used OpenSSL to create an SSLv2 connection to a malicious server, that server could cause the client to crash. Testing using the Codenomicon TLS test suite discovered a flaw in the handling of server name extension data in OpenSSL 0.9.8f and OpenSSL 0.9.8g. If OpenSSL has been compiled using the non-default TLS server name extensions, a remote attacker could send a carefully crafted packet to a server application using OpenSSL and cause it to crash. Testing using the Codenomicon TLS test suite discovered a flaw if the 'Server Key exchange message' is omitted from a TLS handshake in OpenSSL 0.9.8f and OpenSSL 0.9.8g. If a client connects to a malicious server with particular cipher suites, the server could cause the client to crash. The Google Security Team discovered several functions inside OpenSSL incorrectly checked the result after calling the EVP_VerifyFinal function, allowing a malformed signature to be treated as a good signature rather than as an error. This issue affected the signature checks on DSA and ECDSA keys used with SSL/TLS. One way to exploit this flaw would be for a remote attacker who is in control of a malicious server or who can use a 'man in the middle' attack to present a malformed SSL/TLS signature from a certificate chain to a vulnerable client, bypassing validation. The function ASN1_STRING_print_ex() when used to print a BMPString or UniversalString will crash with an invalid memory access if the encoded length of the string is illegal. Any OpenSSL application which prints out the contents of a certificate could be affected by this bug, including SSL servers, clients and S/MIME software. The function CMS_verify() does not correctly handle an error condition involving malformed signed attributes. This will cause an invalid set of signed attributes to appear valid and content digests will not be checked. When a malformed ASN1 structure is received it's contents are freed up and zeroed and an error condition returned. On a small number of platforms where sizeof(long) < sizeof(void *) (for example WIN64) this can cause an invalid memory access later resulting in a crash when some invalid structures are read, for example RSA public keys. Fix a NULL pointer dereference if a DTLS server recieved ChangeCipherSpec as first record. A remote attacker could use this flaw to cause a DTLS server to crash Implement RFC5746 to address vulnerabilities in SSL/TLS renegotiation. Fix denial of service flaw due in the DTLS implementation. A remote attacker could use this flaw to cause a DTLS server to crash. Fix denial of service flaws in the DTLS implementation. A remote attacker could use these flaws to cause a DTLS server to use excessive amounts of memory, or crash. A memory leak in the zlib_stateful_finish function in crypto/comp/c_zlib.c allows remote attackers to cause a denial of service via vectors that trigger incorrect calls to the CRYPTO_cleanup_all_ex_data function. It was discovered that OpenSSL did not always check the return value of the bn_wexpand() function. An attacker able to trigger a memory allocation failure in that function could cause an application using the OpenSSL library to crash or, possibly, execute arbitrary code A missing return value check flaw was discovered in OpenSSL, that could possibly cause OpenSSL to call a Kerberos library function with invalid arguments, resulting in a NULL pointer dereference crash in the MIT Kerberos library. In certain configurations, a remote attacker could use this flaw to crash a TLS/SSL server using OpenSSL by requesting Kerberos cipher suites during the TLS handshake In TLS connections, certain incorrectly formatted records can cause an OpenSSL client or server to crash due to a read attempt at NULL. A flaw in the handling of CMS structures containing OriginatorInfo was found which could lead to a write to invalid memory address or double free. CMS support is disabled by default in OpenSSL 0.9.8 versions. An invalid Return value check in pkey_rsa_verifyrecover was discovered. When verification recovery fails for RSA keys an uninitialised buffer with an undefined length is returned instead of an error code. This could lead to an information leak. A flaw in the OpenSSL TLS server extension code parsing which on affected servers can be exploited in a buffer overrun attack. Any OpenSSL based TLS server is vulnerable if it is multi-threaded and uses OpenSSL's internal caching mechanism. Servers that are multi-process and/or disable internal session caching are NOT affected. An error in OpenSSL's experimental J-PAKE implementation which could lead to successful validation by someone with no knowledge of the shared secret. The OpenSSL Team still consider the implementation of J-PAKE to be experimental and is not compiled by default. A flaw in the OpenSSL SSL/TLS server code where an old bug workaround allows malicious clients to modify the stored session cache ciphersuite. In some cases the ciphersuite can be downgraded to a weaker one on subsequent connections. This issue only affects OpenSSL based SSL/TLS server if it uses OpenSSL's internal caching mechanisms and the SSL_OP_NETSCAPE_REUSE_CIPHER_CHANGE_BUG flag (many applications enable this by using the SSL_OP_ALL option). Under certain circumstances OpenSSL's internal certificate verification routines can incorrectly accept a CRL whose nextUpdate field is in the past. Applications are only affected by the CRL checking vulnerability if they enable OpenSSL's internal CRL checking which is off by default. Applications which use their own custom CRL checking (such as Apache) are not affected. OpenSSL server code for ephemeral ECDH ciphersuites is not thread-safe, and furthermore can crash if a client violates the protocol by sending handshake messages in incorrect order. Only server-side applications that specifically support ephemeral ECDH ciphersuites are affected, and only if ephemeral ECDH ciphersuites are enabled in the configuration. OpenSSL was susceptable an extension of the Vaudenay padding oracle attack on CBC mode encryption which enables an efficient plaintext recovery attack against the OpenSSL implementation of DTLS by exploiting timing differences arising during decryption processing. If X509_V_FLAG_POLICY_CHECK is set in OpenSSL 0.9.8, then a policy check failure can lead to a double-free. The bug does not occur unless this flag is set. Users of OpenSSL 1.0.0 are not affected OpenSSL failed to clear the bytes used as block cipher padding in SSL 3.0 records which could leak the contents of memory in some circumstances. RFC 3779 data can be included in certificates, and if it is malformed, may trigger an assertion failure. This could be used in a denial-of-service attack. Builds of OpenSSL are only vulnerable if configured with "enable-rfc3779", which is not a default. Support for handshake restarts for server gated cryptograpy (SGC) can be used in a denial-of-service attack. A malicious TLS client can send an invalid set of GOST parameters which will cause the server to crash due to lack of error checking. This could be used in a denial-of-service attack. Only users of the OpenSSL GOST ENGINE are affected by this bug. A flaw in the fix to CVE-2011-4108 can be exploited in a denial of service attack. Only DTLS applications are affected. A weakness in the OpenSSL CMS and PKCS #7 code can be exploited using Bleichenbacher's attack on PKCS #1 v1.5 RSA padding also known as the million message attack (MMA). Only users of CMS, PKCS #7, or S/MIME decryption operations are affected, SSL/TLS applications are not affected by this issue. An integer underflow flaw, leading to a buffer over-read, was found in the way OpenSSL handled TLS 1.1, TLS 1.2, and DTLS (Datagram Transport Layer Security) application data record lengths when using a block cipher in CBC (cipher-block chaining) mode. A malicious TLS 1.1, TLS 1.2, or DTLS client or server could use this flaw to crash its connection peer. A weakness in the handling of CBC ciphersuites in SSL, TLS and DTLS which could lead to plaintext recovery by exploiting timing differences arising during MAC processing. A flaw in the OpenSSL handling of CBC ciphersuites in TLS 1.1 and TLS 1.2 on AES-NI supporting platforms can be exploited in a DoS attack. A flaw in the OpenSSL handling of OCSP response verification can be exploited in a denial of service attack.