=========================
Potentially-incompatible changes
* sshd(8): The default set of ciphers and MACs has been altered to
remove unsafe algorithms. In particular, CBC ciphers and arcfour*
are disabled by default.
The full set of algorithms remains available if configured
explicitly via the Ciphers and MACs sshd_config options.
* sshd(8): Support for tcpwrappers/libwrap has been removed.
* OpenSSH 6.5 and 6.6 have a bug that causes ~0.2% of connections
using the curve25519-sha256@libssh.org KEX exchange method to fail
when connecting with something that implements the specification
correctly. OpenSSH 6.7 disables this KEX method when speaking to
one of the affected versions.
New Features
* Major internal refactoring to begin to make part of OpenSSH usable
as a library. So far the wire parsing, key handling and KRL code
has been refactored. Please note that we do not consider the API
stable yet, nor do we offer the library in separable form.
* ssh(1), sshd(8): Add support for Unix domain socket forwarding.
A remote TCP port may be forwarded to a local Unix domain socket
and vice versa or both ends may be a Unix domain socket.
* ssh(1), ssh-keygen(1): Add support for SSHFP DNS records for
ED25519 key types.
* sftp(1): Allow resumption of interrupted uploads.
* ssh(1): When rekeying, skip file/DNS lookups of the hostkey if it
is the same as the one sent during initial key exchange; bz#2154
* sshd(8): Allow explicit ::1 and 127.0.0.1 forwarding bind
addresses when GatewayPorts=no; allows client to choose address
family; bz#2222
* sshd(8): Add a sshd_config PermitUserRC option to control whether
~/.ssh/rc is executed, mirroring the no-user-rc authorized_keys
option; bz#2160
* ssh(1): Add a %C escape sequence for LocalCommand and ControlPath
that expands to a unique identifer based on a hash of the tuple of
(local host, remote user, hostname, port). Helps avoid exceeding
miserly pathname limits for Unix domain sockets in multiplexing
control paths; bz#2220
* sshd(8): Make the "Too many authentication failures" message
include the user, source address, port and protocol in a format
similar to the authentication success / failure messages; bz#2199
* Added unit and fuzz tests for refactored code. These are run
automatically in portable OpenSSH via the "make tests" target.
Bugfixes
* sshd(8): Fix remote forwarding with the same listen port but
different listen address.
* ssh(1): Fix inverted test that caused PKCS#11 keys that were
explicitly listed in ssh_config or on the commandline not to be
preferred.
* ssh-keygen(1): Fix bug in KRL generation: multiple consecutive
revoked certificate serial number ranges could be serialised to an
invalid format. Readers of a broken KRL caused by this bug will
fail closed, so no should-have-been-revoked key will be accepted.
* ssh(1): Reflect stdio-forward ("ssh -W host:port ...") failures in
exit status. Previously we were always returning 0; bz#2255
* ssh(1), ssh-keygen(1): Make Ed25519 keys' title fit properly in the
randomart border; bz#2247
* ssh-agent(1): Only cleanup agent socket in the main agent process
and not in any subprocesses it may have started (e.g. forked
askpass). Fixes agent sockets being zapped when askpass processes
fatal(); bz#2236
* ssh-add(1): Make stdout line-buffered; saves partial output getting
lost when ssh-add fatal()s part-way through (e.g. when listing keys
from an agent that supports key types that ssh-add doesn't);
bz#2234
* ssh-keygen(1): When hashing or removing hosts, don't choke on
@revoked markers and don't remove @cert-authority markers; bz#2241
* ssh(1): Don't fatal when hostname canonicalisation fails and a
ProxyCommand is in use; continue and allow the ProxyCommand to
connect anyway (e.g. to a host with a name outside the DNS behind
a bastion)
* scp(1): When copying local->remote fails during read, don't send
uninitialised heap to the remote end.
* sftp(1): Fix fatal "el_insertstr failed" errors when tab-completing
filenames with a single quote char somewhere in the string;
bz#2238
* ssh-keyscan(1): Scan for Ed25519 keys by default.
* ssh(1): When using VerifyHostKeyDNS with a DNSSEC resolver, down-
convert any certificate keys to plain keys and attempt SSHFP
resolution. Prevents a server from skipping SSHFP lookup and
forcing a new-hostkey dialog by offering only certificate keys.
* sshd(8): Avoid crash at exit via NULL pointer reference; bz#2225
* Fix some strict-alignment errors.
Portable OpenSSH
* Portable OpenSSH now supports building against libressl-portable.
* Portable OpenSSH now requires openssl 0.9.8f or greater. Older
versions are no longer supported.
* In the OpenSSL version check, allow fix version upgrades (but not
downgrades. Debian bug #748150.
* sshd(8): On Cygwin, determine privilege separation user at runtime,
since it may need to be a domain account.
* sshd(8): Don't attempt to use vhangup on Linux. It doesn't work for
non-root users, and for them it just messes up the tty settings.
* Use CLOCK_BOOTTIME in preference to CLOCK_MONOTONIC when it is
available. It considers time spent suspended, thereby ensuring
timeouts (e.g. for expiring agent keys) fire correctly. bz#2228
* Add support for ed25519 to opensshd.init init script.
* sftp-server(8): On platforms that support it, use prctl() to
prevent sftp-server from accessing /proc/self/{mem,maps}
Changes since OpenSSH 6.5
=========================
This is primarily a bugfix release.
Security:
* sshd(8): when using environment passing with a sshd_config(5)
AcceptEnv pattern with a wildcard. OpenSSH prior to 6.6 could be
tricked into accepting any enviornment variable that contains the
characters before the wildcard character.
New / changed features:
* ssh(1), sshd(8): this release removes the J-PAKE authentication code.
This code was experimental, never enabled and had been unmaintained
for some time.
* ssh(1): when processing Match blocks, skip 'exec' clauses other clauses
predicates failed to match.
* ssh(1): if hostname canonicalisation is enabled and results in the
destination hostname being changed, then re-parse ssh_config(5) files
using the new destination hostname. This gives 'Host' and 'Match'
directives that use the expanded hostname a chance to be applied.
Bugfixes:
* ssh(1): avoid spurious "getsockname failed: Bad file descriptor" in
ssh -W. bz#2200, debian#738692
* sshd(8): allow the shutdown(2) syscall in seccomp-bpf and systrace
sandbox modes, as it is reachable if the connection is terminated
during the pre-auth phase.
* ssh(1), sshd(8): fix unsigned overflow that in SSH protocol 1 bignum
parsing. Minimum key length checks render this bug unexploitable to
compromise SSH 1 sessions.
* sshd_config(5): clarify behaviour of a keyword that appears in
multiple matching Match blocks. bz#2184
* ssh(1): avoid unnecessary hostname lookups when canonicalisation is
disabled. bz#2205
* sshd(8): avoid sandbox violation crashes in GSSAPI code by caching
the supported list of GSSAPI mechanism OIDs before entering the
sandbox. bz#2107
* ssh(1): fix possible crashes in SOCKS4 parsing caused by assumption
that the SOCKS username is nul-terminated.
* ssh(1): fix regression for UsePrivilegedPort=yes when BindAddress is
not specified.
* ssh(1), sshd(8): fix memory leak in ECDSA signature verification.
* ssh(1): fix matching of 'Host' directives in ssh_config(5) files
to be case-insensitive again (regression in 6.5).
Portable OpenSSH:
* sshd(8): don't fatal if the FreeBSD Capsicum is offered by the
system headers and libc but is not supported by the kernel.
* Fix build using the HP-UX compiler.
Changes since OpenSSH 6.4
=========================
This is a feature-focused release.
New features:
* ssh(1), sshd(8): Add support for key exchange using elliptic-curve
Diffie Hellman in Daniel Bernstein's Curve25519. This key exchange
method is the default when both the client and server support it.
* ssh(1), sshd(8): Add support for Ed25519 as a public key type.
Ed25519 is a elliptic curve signature scheme that offers
better security than ECDSA and DSA and good performance. It may be
used for both user and host keys.
* Add a new private key format that uses a bcrypt KDF to better
protect keys at rest. This format is used unconditionally for
Ed25519 keys, but may be requested when generating or saving
existing keys of other types via the -o ssh-keygen(1) option.
We intend to make the new format the default in the near future.
Details of the new format are in the PROTOCOL.key file.
* ssh(1), sshd(8): Add a new transport cipher
"chacha20-poly1305@openssh.com" that combines Daniel Bernstein's
ChaCha20 stream cipher and Poly1305 MAC to build an authenticated
encryption mode. Details are in the PROTOCOL.chacha20poly1305 file.
* ssh(1), sshd(8): Refuse RSA keys from old proprietary clients and
servers that use the obsolete RSA+MD5 signature scheme. It will
still be possible to connect with these clients/servers but only
DSA keys will be accepted, and OpenSSH will refuse connection
entirely in a future release.
* ssh(1), sshd(8): Refuse old proprietary clients and servers that
use a weaker key exchange hash calculation.
* ssh(1): Increase the size of the Diffie-Hellman groups requested
for each symmetric key size. New values from NIST Special
Publication 800-57 with the upper limit specified by RFC4419.
* ssh(1), ssh-agent(1): Support PKCS#11 tokens that only provide
X.509 certs instead of raw public keys (requested as bz#1908).
* ssh(1): Add a ssh_config(5) "Match" keyword that allows
conditional configuration to be applied by matching on hostname,
user and result of arbitrary commands.
* ssh(1): Add support for client-side hostname canonicalisation
using a set of DNS suffixes and rules in ssh_config(5). This
allows unqualified names to be canonicalised to fully-qualified
domain names to eliminate ambiguity when looking up keys in
known_hosts or checking host certificate names.
* sftp-server(8): Add the ability to whitelist and/or blacklist sftp
protocol requests by name.
* sftp-server(8): Add a sftp "fsync@openssh.com" to support calling
fsync(2) on an open file handle.
* sshd(8): Add a ssh_config(5) PermitTTY to disallow TTY allocation,
mirroring the longstanding no-pty authorized_keys option.
* ssh(1): Add a ssh_config ProxyUseFDPass option that supports the
use of ProxyCommands that establish a connection and then pass a
connected file descriptor back to ssh(1). This allows the
ProxyCommand to exit rather than staying around to transfer data.
Bugfixes:
* ssh(1), sshd(8): Fix potential stack exhaustion caused by nested
certificates.
* ssh(1): bz#1211: make BindAddress work with UsePrivilegedPort.
* sftp(1): bz#2137: fix the progress meter for resumed transfer.
* ssh-add(1): bz#2187: do not request smartcard PIN when removing
keys from ssh-agent.
* sshd(8): bz#2139: fix re-exec fallback when original sshd binary
cannot be executed.
* ssh-keygen(1): Make relative-specified certificate expiry times
relative to current time and not the validity start time.
* sshd(8): bz#2161: fix AuthorizedKeysCommand inside a Match block.
* sftp(1): bz#2129: symlinking a file would incorrectly canonicalise
the target path.
* ssh-agent(1): bz#2175: fix a use-after-free in the PKCS#11 agent
helper executable.
* sshd(8): Improve logging of sessions to include the user name,
remote host and port, the session type (shell, command, etc.) and
allocated TTY (if any).
* sshd(8): bz#1297: tell the client (via a debug message) when
their preferred listen address has been overridden by the
server's GatewayPorts setting.
* sshd(8): bz#2162: include report port in bad protocol banner
message.
* sftp(1): bz#2163: fix memory leak in error path in do_readdir().
* sftp(1): bz#2171: don't leak file descriptor on error.
* sshd(8): Include the local address and port in "Connection from
..." message (only shown at loglevel>=verbose).
Portable OpenSSH:
* Please note that this is the last version of Portable OpenSSH that
will support versions of OpenSSL prior to 0.9.6. Support (i.e.
SSH_OLD_EVP) will be removed following the 6.5p1 release.
* Portable OpenSSH will attempt compile and link as a Position
Independent Executable on Linux, OS X and OpenBSD on recent gcc-
like compilers. Other platforms and older/other compilers may
request this using the --with-pie configure flag.
* A number of other toolchain-related hardening options are used
automatically if available, including -ftrapv to abort on signed
integer overflow and options to write-protect dynamic linking
information. The use of these options may be disabled using the
--without-hardening configure flag.
* If the toolchain supports it, one of the -fstack-protector-strong,
-fstack-protector-all or -fstack-protector compilation flag are
used to add guards to mitigate attacks based on stack overflows.
The use of these options may be disabled using the
--without-stackprotect configure option.
* sshd(8): Add support for pre-authentication sandboxing using the
Capsicum API introduced in FreeBSD 10.
* Switch to a ChaCha20-based arc4random() PRNG for platforms that do
not provide their own.
* sshd(8): bz#2156: restore Linux oom_adj setting when handling
SIGHUP to maintain behaviour over retart.
* sshd(8): bz#2032: use local username in krb5_kuserok check rather
than full client name which may be of form user@REALM.
* ssh(1), sshd(8): Test for both the presence of ECC NID numbers in
OpenSSL and that they actually work. Fedora (at least) has
NID_secp521r1 that doesn't work.
* bz#2173: use pkg-config --libs to include correct -L location for
libedit.
Upstream condensed log:
Major changes between OpenSSL 1.0.1i and OpenSSL 1.0.1j [15 Oct 2014]
o Fix for CVE-2014-3513
o Fix for CVE-2014-3567
o Mitigation for CVE-2014-3566 (SSL protocol vulnerability)
o Fix for CVE-2014-3568
Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
*) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
SRP code can be overrun an internal buffer. Add sanity check that
g, A, B < N to SRP code.
Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
Group for discovering this issue.
(CVE-2014-3512)
[Steve Henson]
*) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
TLS 1.0 instead of higher protocol versions when the ClientHello message
is badly fragmented. This allows a man-in-the-middle attacker to force a
downgrade to TLS 1.0 even if both the server and the client support a
higher protocol version, by modifying the client's TLS records.
Thanks to David Benjamin and Adam Langley (Google) for discovering and
researching this issue.
(CVE-2014-3511)
[David Benjamin]
*) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
to a denial of service attack. A malicious server can crash the client
with a null pointer dereference (read) by specifying an anonymous (EC)DH
ciphersuite and sending carefully crafted handshake messages.
Thanks to Felix Gröbert (Google) for discovering and researching this
issue.
(CVE-2014-3510)
[Emilia Käsper]
*) By sending carefully crafted DTLS packets an attacker could cause openssl
to leak memory. This can be exploited through a Denial of Service attack.
Thanks to Adam Langley for discovering and researching this issue.
(CVE-2014-3507)
[Adam Langley]
*) An attacker can force openssl to consume large amounts of memory whilst
processing DTLS handshake messages. This can be exploited through a
Denial of Service attack.
Thanks to Adam Langley for discovering and researching this issue.
(CVE-2014-3506)
[Adam Langley]
*) An attacker can force an error condition which causes openssl to crash
whilst processing DTLS packets due to memory being freed twice. This
can be exploited through a Denial of Service attack.
Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
this issue.
(CVE-2014-3505)
[Adam Langley]
*) If a multithreaded client connects to a malicious server using a resumed
session and the server sends an ec point format extension it could write
up to 255 bytes to freed memory.
Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
issue.
(CVE-2014-3509)
[Gabor Tyukasz]
*) A malicious server can crash an OpenSSL client with a null pointer
dereference (read) by specifying an SRP ciphersuite even though it was not
properly negotiated with the client. This can be exploited through a
Denial of Service attack.
Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
discovering and researching this issue.
(CVE-2014-5139)
[Steve Henson]
*) A flaw in OBJ_obj2txt may cause pretty printing functions such as
X509_name_oneline, X509_name_print_ex et al. to leak some information
from the stack. Applications may be affected if they echo pretty printing
output to the attacker.
Thanks to Ivan Fratric (Google) for discovering this issue.
(CVE-2014-3508)
[Emilia Käsper, and Steve Henson]
*) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
for corner cases. (Certain input points at infinity could lead to
bogus results, with non-infinity inputs mapped to infinity too.)
[Bodo Moeller]
Changes between 1.0.1h and 1.0.1i [6 Aug 2014]
*) Fix SRP buffer overrun vulnerability. Invalid parameters passed to the
SRP code can be overrun an internal buffer. Add sanity check that
g, A, B < N to SRP code.
Thanks to Sean Devlin and Watson Ladd of Cryptography Services, NCC
Group for discovering this issue.
(CVE-2014-3512)
[Steve Henson]
*) A flaw in the OpenSSL SSL/TLS server code causes the server to negotiate
TLS 1.0 instead of higher protocol versions when the ClientHello message
is badly fragmented. This allows a man-in-the-middle attacker to force a
downgrade to TLS 1.0 even if both the server and the client support a
higher protocol version, by modifying the client's TLS records.
Thanks to David Benjamin and Adam Langley (Google) for discovering and
researching this issue.
(CVE-2014-3511)
[David Benjamin]
*) OpenSSL DTLS clients enabling anonymous (EC)DH ciphersuites are subject
to a denial of service attack. A malicious server can crash the client
with a null pointer dereference (read) by specifying an anonymous (EC)DH
ciphersuite and sending carefully crafted handshake messages.
Thanks to Felix Gröbert (Google) for discovering and researching this
issue.
(CVE-2014-3510)
[Emilia Käsper]
*) By sending carefully crafted DTLS packets an attacker could cause openssl
to leak memory. This can be exploited through a Denial of Service attack.
Thanks to Adam Langley for discovering and researching this issue.
(CVE-2014-3507)
[Adam Langley]
*) An attacker can force openssl to consume large amounts of memory whilst
processing DTLS handshake messages. This can be exploited through a
Denial of Service attack.
Thanks to Adam Langley for discovering and researching this issue.
(CVE-2014-3506)
[Adam Langley]
*) An attacker can force an error condition which causes openssl to crash
whilst processing DTLS packets due to memory being freed twice. This
can be exploited through a Denial of Service attack.
Thanks to Adam Langley and Wan-Teh Chang for discovering and researching
this issue.
(CVE-2014-3505)
[Adam Langley]
*) If a multithreaded client connects to a malicious server using a resumed
session and the server sends an ec point format extension it could write
up to 255 bytes to freed memory.
Thanks to Gabor Tyukasz (LogMeIn Inc) for discovering and researching this
issue.
(CVE-2014-3509)
[Gabor Tyukasz]
*) A malicious server can crash an OpenSSL client with a null pointer
dereference (read) by specifying an SRP ciphersuite even though it was not
properly negotiated with the client. This can be exploited through a
Denial of Service attack.
Thanks to Joonas Kuorilehto and Riku Hietamäki (Codenomicon) for
discovering and researching this issue.
(CVE-2014-5139)
[Steve Henson]
*) A flaw in OBJ_obj2txt may cause pretty printing functions such as
X509_name_oneline, X509_name_print_ex et al. to leak some information
from the stack. Applications may be affected if they echo pretty printing
output to the attacker.
Thanks to Ivan Fratric (Google) for discovering this issue.
(CVE-2014-3508)
[Emilia Käsper, and Steve Henson]
*) Fix ec_GFp_simple_points_make_affine (thus, EC_POINTs_mul etc.)
for corner cases. (Certain input points at infinity could lead to
bogus results, with non-infinity inputs mapped to infinity too.)
[Bodo Moeller]
_______________
Changes between 1.0.1g and 1.0.1h [5 Jun 2014]
*) Fix for SSL/TLS MITM flaw. An attacker using a carefully crafted
handshake can force the use of weak keying material in OpenSSL
SSL/TLS clients and servers.
Thanks to KIKUCHI Masashi (Lepidum Co. Ltd.) for discovering and
researching this issue. (CVE-2014-0224)
[KIKUCHI Masashi, Steve Henson]
*) Fix DTLS recursion flaw. By sending an invalid DTLS handshake to an
OpenSSL DTLS client the code can be made to recurse eventually crashing
in a DoS attack.
Thanks to Imre Rad (Search-Lab Ltd.) for discovering this issue.
(CVE-2014-0221)
[Imre Rad, Steve Henson]
*) Fix DTLS invalid fragment vulnerability. A buffer overrun attack can
be triggered by sending invalid DTLS fragments to an OpenSSL DTLS
client or server. This is potentially exploitable to run arbitrary
code on a vulnerable client or server.
Thanks to Jüri Aedla for reporting this issue. (CVE-2014-0195)
[Jüri Aedla, Steve Henson]
*) Fix bug in TLS code where clients enable anonymous ECDH ciphersuites
are subject to a denial of service attack.
Thanks to Felix Gröbert and Ivan Fratric at Google for discovering
this issue. (CVE-2014-3470)
[Felix Gröbert, Ivan Fratric, Steve Henson]
*) Harmonize version and its documentation. -f flag is used to display
compilation flags.
[mancha <mancha1@zoho.com>]
*) Fix eckey_priv_encode so it immediately returns an error upon a failure
in i2d_ECPrivateKey.
[mancha <mancha1@zoho.com>]
*) Fix some double frees. These are not thought to be exploitable.
[mancha <mancha1@zoho.com>]
Fix for the attack described in the paper "Recovering OpenSSL
ECDSA Nonces Using the FLUSH+RELOAD Cache Side-channel Attack"
by Yuval Yarom and Naomi Benger. Details can be obtained from:
http://eprint.iacr.org/2014/140
joerg