OpenSSH 5.4 released
From: | Damien Miller <djm-AT-cvs.openbsd.org> | |
To: | lwn-AT-lwn.net | |
Subject: | Announce: OpenSSH 5.4 released | |
Date: | Sun, 7 Mar 2010 19:05:58 -0700 (MST) | |
Message-ID: | <201003080205.o2825wof025763@cvs.openbsd.org> |
OpenSSH 5.4 has just been released. It will be available from the mirrors listed at http://www.openssh.com/ shortly. OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0 implementation and includes sftp client and server support. Once again, we would like to thank the OpenSSH community for their continued support of the project, especially those who contributed code or patches, reported bugs, tested snapshots or donated to the project. More information on donations may be found at: http://www.openssh.com/donations.html This is a major feature and bugfix release. Changes since OpenSSH 5.3 ========================= Features: * After a transition period of about 10 years, this release disables SSH protocol 1 by default. Clients and servers that need to use the legacy protocol must explicitly enable it in ssh_config / sshd_config or on the command-line. * Remove the libsectok/OpenSC-based smartcard code and add support for PKCS#11 tokens. This support is automatically enabled on all platforms that support dlopen(3) and was inspired by patches written by Alon Bar-Lev. Details in the ssh(1) and ssh-add(1) manpages. * Add support for certificate authentication of users and hosts using a new, minimal OpenSSH certificate format (not X.509). Certificates contain a public key, identity information and some validity constraints and are signed with a standard SSH public key using ssh-keygen(1). CA keys may be marked as trusted in authorized_keys or via a TrustedUserCAKeys option in sshd_config(5) (for user authentication), or in known_hosts (for host authentication). Documentation for certificate support may be found in ssh-keygen(1), sshd(8) and ssh(1) and a description of the protocol extensions in PROTOCOL.certkeys. * Added a 'netcat mode' to ssh(1): "ssh -W host:port ..." This connects stdio on the client to a single port forward on the server. This allows, for example, using ssh as a ProxyCommand to route connections via intermediate servers. bz#1618 * Add the ability to revoke keys in sshd(8) and ssh(1). User keys may be revoked using a new sshd_config(5) option "RevokedKeys". Host keys are revoked through known_hosts (details in the sshd(8) man page). Revoked keys cannot be used for user or host authentication and will trigger a warning if used. * Rewrite the ssh(1) multiplexing support to support non-blocking operation of the mux master, improve the resilience of the master to malformed messages sent to it by the slave and add support for requesting port- forwardings via the multiplex protocol. The new stdio-to-local forward mode ("ssh -W host:port ...") is also supported. The revised multiplexing protocol is documented in the file PROTOCOL.mux in the source distribution. * Add a 'read-only' mode to sftp-server(8) that disables open in write mode and all other fs-modifying protocol methods. bz#430 * Allow setting an explicit umask on the sftp-server(8) commandline to override whatever default the user has. bz#1229 * Many improvements to the sftp(1) client, many of which were implemented by Carlos Silva through the Google Summer of Code program: - Support the "-h" (human-readable units) flag for ls - Implement tab-completion of commands, local and remote filenames - Support most of scp(1)'s commandline arguments in sftp(1), as a first step towards making sftp(1) a drop-in replacement for scp(1). Note that the rarely-used "-P sftp_server_path" option has been moved to "-D sftp_server_path" to make way for "-P port" to match scp(1). - Add recursive transfer support for get/put and on the commandline * New RSA keys will be generated with a public exponent of RSA_F4 == (2**16)+1 == 65537 instead of the previous value 35. * Passphrase-protected SSH protocol 2 private keys are now protected with AES-128 instead of 3DES. This applied to newly-generated keys as well as keys that are reencrypted (e.g. by changing their passphrase). Bugfixes: * Hold authentication debug messages until after successful authentication. Fixes a minor information leak of environment variables specified in authorized_keys if an attacker happens to know the public key in use. * When using ChrootDirectory, make sure we test for the existence of the user's shell inside the chroot and not outside (bz#1679) * Cache user and group name lookups in sftp-server using user_from_[ug]id(3) to improve performance on hosts where these operations are slow (e.g. NIS or LDAP). bz#1495 * Fix problem that prevented passphrase reading from being interrupted in some circumstances; bz#1590 * Ignore and log any Protocol 1 keys where the claimed size is not equal to the actual size. * Make HostBased authentication work with a ProxyCommand. bz#1569 * Avoid run-time failures when specifying hostkeys via a relative path by prepending the current working directory in these cases. bz#1290 * Do not prompt for a passphrase if we fail to open a keyfile, and log the reason why the open failed to debug. bz#1693 * Document that the PubkeyAuthentication directive is allowed in a sshd_config(5) Match block. bz#1577 * When converting keys, truncate key comments at 72 chars as per RFC4716. bz#1630 * Do not allow logins if /etc/nologin exists but is not readable by the user logging in. * Output a debug log if sshd(8) can't open an existing authorized_keys. bz#1694 * Quell tc[gs]etattr warnings when forcing a tty (ssh -tt), since we usually don't actually have a tty to read/set; bz#1686 * Prevent sftp from crashing when given a "-" without a command. Also, allow whitespace to follow a "-". bz#1691 * After sshd receives a SIGHUP, ignore subsequent HUPs while sshd re-execs itself. Prevents two HUPs in quick succession from resulting in sshd dying. bz#1692 * Clarify in sshd_config(5) that StrictModes does not apply to ChrootDirectory. Permissions and ownership are always checked when chrooting. bz#1532 * Set close-on-exec on various descriptors so they don't get leaked to child processes. bz#1643 * Fix very rare race condition in x11/agent channel allocation: don't read after the end of the select read/write fdset and make sure a reused FD is not touched before the pre-handlers are called. * Fix incorrect exit status when multiplexing and channel ID 0 is recycled. bz#1570 * Fail with an error when an attempt is made to connect to a server with ForceCommand=internal-sftp with a shell session (i.e. not a subsystem session). Avoids stuck client when attempting to ssh to such a service. bz#1606: * Warn but do not fail if stat()ing the subsystem binary fails. This helps with chrootdirectory+forcecommand=sftp-server and restricted shells. bz #1599 * Change "Connecting to host..." message to "Connected to host." and delay it until after the sftp protocol connection has been established. Avoids confusing sequence of messages when the underlying ssh connection experiences problems. bz#1588 * Use the HostKeyAlias rather than the hostname specified on the commandline when prompting for passwords. bz#1039 * Correct off-by-one in percent_expand(): we would fatal() when trying to expand EXPAND_MAX_KEYS, allowing only EXPAND_MAX_KEYS-1 to actually work. Note that nothing in OpenSSH actually uses close to this limit at present. bz#1607 * Fix passing of empty options from scp(1) and sftp(1) to the underlying ssh(1). Also add support for the stop option "--". * Fix an incorrect magic number and typo in PROTOCOL; bz#1688 * Don't escape backslashes when displaying the SSH2 banner. bz#1533 * Don't unnecessarily dup() the in and out fds for sftp-server. bz#1566 * Force use of the correct hash function for random-art signature display as it was inheriting the wrong one when bubblebabble signatures were activated. bz#1611 * Do not fall back to adding keys without constraints (ssh-add -c / -t ...) when the agent refuses the constrained add request. bz#1612 * Fix a race condition in ssh-agent that could result in a wedged or spinning agent. bz#1633 * Flush stdio before exec() to ensure that everying (motd in particular) has made it out before the streams go away. bz#1596 * Set FD_CLOEXEC on in/out sockets in sshd(8). bz#1706 Portable OpenSSH Bugfixes: * Use system's kerberos principal name on AIX if it's available. bz#1583 * Disable OOM-killing of the listening sshd on Linux. bz#1470 * Use pkg-config for opensc config if it's available. bz#1160 * Unbreak Redhat spec to allow building without askpass. bz#1677 * If PidFile is set in sshd_config, use it in SMF init file. bz#1628 * Print error and usage() when ssh-rand-helper is passed command- line arguments as none are supported. bz#1568 * Add missing setsockopt() to set IPV6_V6ONLY for local forwarding with GatwayPorts=yes. bz#1648 * Make GNOME 2 askpass dialog desktop-modal. bz#1645 * If SELinux is enabled set the security context to "sftpd_t" before running the internal sftp server. bz#1637 * Correctly check libselinux for necessary SELinux functions; bz#1713 * Unbreak builds on Redhat using the supplied openssh.spec; bz#1731 * Fix incorrect privilege dropping order on AIX that prevented chroot operation; bz#1567 * Call aix_setauthdb/aix_restoredb at the correct times on AIX to prevent authentication failure; bz#1710 Checksums: ========== - SHA1 (openssh-5.4.tar.gz) = 1776832d902f7b4c7863afd41a5ec7a14efe95d6 - SHA1 (openssh-5.4p1.tar.gz) = 2a3042372f08afb1415ceaec8178213276a36302 Reporting Bugs: =============== - Please read http://www.openssh.com/report.html Security bugs should be reported directly to openssh@openssh.com OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and Ben Lindstrom.
Posted Mar 8, 2010 15:10 UTC (Mon)
by nix (subscriber, #2304)
[Link]
Posted Mar 8, 2010 16:57 UTC (Mon)
by aaron (guest, #282)
[Link] (13 responses)
Please, O pufferfish, strap on a jetpack!
Posted Mar 8, 2010 17:04 UTC (Mon)
by alex (subscriber, #1355)
[Link] (4 responses)
Posted Mar 9, 2010 1:51 UTC (Tue)
by BrucePerens (guest, #2510)
[Link] (3 responses)
Posted Mar 9, 2010 4:20 UTC (Tue)
by djm (subscriber, #11651)
[Link] (2 responses)
Posted Mar 9, 2010 4:51 UTC (Tue)
by BrucePerens (guest, #2510)
[Link] (1 responses)
Posted Mar 9, 2010 11:04 UTC (Tue)
by djm (subscriber, #11651)
[Link]
This limit is enough for a path with 5 seconds latency at your DSL speed.
If you are on Internet2 and want to move files between continental USA and
Posted Mar 9, 2010 4:24 UTC (Tue)
by djm (subscriber, #11651)
[Link] (2 responses)
The other component of the HPN patches that people sometimes ask for is the
Posted Mar 9, 2010 11:51 UTC (Tue)
by tialaramex (subscriber, #21167)
[Link]
No-one is asking for this to be the default of course, but why can't we have it as an option without hacking the code?
Posted Mar 9, 2010 14:45 UTC (Tue)
by andikleen (guest, #39006)
[Link]
And yes multi threading your application is typically intrusive, but
Posted Mar 9, 2010 13:49 UTC (Tue)
by andikleen (guest, #39006)
[Link] (4 responses)
I wonder if it's related to being developed on OpenBSD which is not exactly known for SMP scalability?
Perhaps the distros will use it some day at least, even if the maintainers can't get out of the 70ies.
Posted Mar 9, 2010 21:12 UTC (Tue)
by djm (subscriber, #11651)
[Link] (3 responses)
Posted Mar 9, 2010 22:25 UTC (Tue)
by djm (subscriber, #11651)
[Link]
Posted Mar 11, 2010 2:48 UTC (Thu)
by martinfick (subscriber, #4455)
[Link] (1 responses)
Posted Mar 11, 2010 5:26 UTC (Thu)
by dlang (guest, #313)
[Link]
on very modern system I can transfer data twice as fast over local gigE networks with HTTP or FTP than I can via SSH
Posted Mar 8, 2010 17:27 UTC (Mon)
by lkundrak (subscriber, #43452)
[Link] (6 responses)
Posted Mar 8, 2010 17:36 UTC (Mon)
by ewan (guest, #5533)
[Link]
Posted Mar 8, 2010 20:30 UTC (Mon)
by drag (guest, #31333)
[Link] (4 responses)
For most uses of Linux killing off sshd would have a similar effect of killing up 'getty' on a local
Posted Mar 9, 2010 1:43 UTC (Tue)
by BrucePerens (guest, #2510)
[Link]
Posted Mar 9, 2010 23:26 UTC (Tue)
by herodiade (guest, #52755)
[Link] (2 responses)
Also note protection is on the _listening_ sshd (that's not a whole lot a code ; not the user's child processes for instance).
With regard to the new certificate format, does anyone here knows why X.509 doesn't fit well for SSH?
Someone maintains a patchset to include support for X.509 certificate here http://roumenpetrov.info/openssh/ ; I can see some benefit with that from an end-user perspective (mostly, reusing existing tools for pki management, certificates revocations lists, etc).
Posted Mar 9, 2010 23:56 UTC (Tue)
by djm (subscriber, #11651)
[Link] (1 responses)
The OpenSSH certificate format uses existing SSH signature formats and wire
Posted Mar 12, 2010 13:53 UTC (Fri)
by BSchuller (guest, #64336)
[Link]
X.509 may be complex, it is widely used, and many
Best regards,
Posted Mar 9, 2010 13:50 UTC (Tue)
by andikleen (guest, #39006)
[Link]
But then see other comment for a serious misfeature.
OpenSSH 5.4 released
protocol 1 by default'; i.e., you can still turn it back on.
Dang! No HPN-SSH!
(See http://www.psc.edu/networking/projects/hpn-ssh )
It's truly amazing how they help transfers over higher-latency links (i.e. any distance over a mile.) They also help local transfers a fair bit.
Unfortunately, for now, if you can't patch your own SSHd, you have to use commercial file-transfer software, a Riverbed, or RSA-SSHd.
Dang! No HPN-SSH!
auditing code that goes from statically allocated to dynamically allocated
buffers?
So, is the alternative to set the static buffers to 2MB?
Dang! No HPN-SSH!
Dang! No HPN-SSH!
Dang! No HPN-SSH!
Dang! No HPN-SSH!
Put differently, if your one-way path latency is 100ms then unmodified
OpenSSH's window size should only start restricting performance if your
transfer rate is ~160Mbit/s.
Europe at gigabit speeds, then you might still want the HPN patches.
Dang! No HPN-SSH!
which should have obviated the need for the HPN patches for all but the
highest BDP links. You should benchmark your connection to see if it
actually benefits from the HPN patches, which are quite intrusive.
ability to select a null cipher/MAC. We are not planning on implementing
that ever.
Dang! No HPN-SSH!
Dang! No HPN-SSH!
which is needed for fast enough links because a single core cannot do full performance otherwise.
it's also very needed if it is CPU time intensive and
you want to keep up with modern systems.
Dang! No HPN-SSH!
Dang! No HPN-SSH!
conditions in security software. AES on my 3 year old desktop is >800Mbit/s
on a single core and RC4 is 2x faster again, so I don't think crypto
performance is a massive problem.
err, s/prefer/prefer not/
Dang! No HPN-SSH!
Dang! No HPN-SSH!
Dang! No HPN-SSH!
OpenSSH 5.4 released
* Disable OOM-killing of the listening sshd on Linux. bz#1470
I'm wondering what sense does this make. Turning off a feature that would kill sshd once it goes mad and eats up all the memory and kill everything else instead? Or do they assume the oomkiller to be that stupid?
Oftentimes it is that stupid, and can make a bad situation considerably worse by cutting off any hope you might have had of SSHing in and fixing the actual problem.
OpenSSH 5.4 released
OpenSSH 5.4 released
ram is much much less then having other applications flip out and use all your RAM and having
OOM kill off sshd.
console. OOM Killer operating in this manner is effectively like having the Linux kernel perform
it's own denial service attack on userland.
OpenSSH 5.4 released
OpenSSH 5.4 released
OpenSSH 5.4 released
semantics. Consider the number of ASN.1 related bugs that OpenSSL and other
implementations have suffered from - the fact that nobody gets this right is
a good signal that it is overly complex. Unfortunately, key/cert parsing and
validation is by necessity in the critical pre-authentication attack surface
of sshd, so bugs there are particularly nasty and could be used to write
worms.
encoding primitives, so all of that code is reused. Also, because these
certificates are not tied into such a heavyweight and hierarchical model of
identity, the the semantics of certificates and the workflow of creating
them is much simpler.
OpenSSH 5.4 released
X.509 based infrastructures exist. Therefore, introducing a
home-grown version of X.509 requiring admins to effectively
run their own CA is hard to understand.
I think it would be a very good move if X.509 could be
supported in OpenSSH. Admins who think the security risk is
too high can always opt-out.
Bernd.
Tab completion in ssh