Portable OpenSSH 10.6p2 was released on 2026-11-05. It will be available\nfrom the mirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nRecently the OpenSSH team have received a large number of security\nbug reports, many of which are findings from AI models or made with\nAI assistance. While many AI reports are determined not to have\nsecurity impact when considered in the context of a realistic\nthreat model, we very much welcome these reports, especially when\ncombined with human triage, analysis, test-cases and particularly\nwhen accompanied by proposed fixes.\n\n** We have seen a number of cases where a security bug identified\n** by AI tools is subsequently independently discovered by a\n** different researcher. This suggests that adversaries who do not\n** report bugs to OSS projects are likely to be able to discover\n** these bugs too. Given this, the OpenSSH team will, for now, be\n** making more frequent releases to get bugfixes into users' hands\n** more quickly rather than batching them until the next planned\n** release.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n-------------------------\n\n * scp(1): begin deprecating the -R flag, which is used to perform a\n remote-to-remote copy by executing scp on a remote host. This\n option is a fragile optimisation that is difficult to use because\n it requires credentials on the remote host. It also creates\n security risks if the shell quoting rules on the remote system\n where the copy is performed differ from the client's expectations.\n\n From OpenSSH 10.6, this option will continue to work but will\n cause a deprecation warning to be emitted to standard error. In\n a future release, the option will be ignored and will leave in\n place the default remote-to-remote copy behaviour (copy via the\n host running scp).\n\n * sshd(8): support for platforms that do not allow file descriptor\n passing and that also require root privilege for PTY allocation\n will be removed in a future release. Affected platforms are known\n to include SCO OpenServer 5 and QNX 6 but may include other\n similarly old operating systems. This deprecation can be avoided\n if the user community for these platforms is able to assist us in\n building alternatives, such as avoiding the need for root in PTY\n allocation.\n\nChanges since OpenSSH 10.6p1\n============================\n\nThis release contains a critical security workaround that affects\nLinux systems that use the musl libc and PAM disabled.\n\nThis release also includes a number of bugfixes that had already\nbeen committed on the portable OpenSSH V_10_6 stable branch.\n\nSecurity\n========\n\n * sshd(8): avoid password authentication bypass that is exploitable\n through sshd(8) that affects some Linux systems when built and run\n with PAM support disabled. We consider the root cause of this bug\n to be outside sshd, but it is exploitable via sshd and can be\n avoided there.\n\n This bug was reported by Özgün Kültekin of Wakefield\n (wakefield.co)\n\nBugfixes\n--------\n\n * sshd(8), ssh(1): enable the post-quantum hybrid signature scheme\n ssh-mldsa44-ed25519 as an allowed CA signature algorithm.\n\n * ssh(1), sshd(8): fix <1 second CPU spin if ChannelTimeout was in\n use.\n\nPortability\n-----------\n\n * sshd(8): fix compilation on (old) platforms that set SKIP_PRIVDROP\n\n * ssh(1), sshd(8): fix detection of Azerbaijani locales\n\n * sshd(8), ssh-agent(1): fix creation of agent sockets on platforms\n without the openat(2) syscall.\n\n * ssh(1), ssh-add(1), ssh-keygen(1): fix some strange behaviour if\n SIGINT/SIGQUIT was received while reading a passphrase.\n\n * sshd(8): allow restart_syscall(2) in seccomp sandbox; bz4036\n\nChecksums:\n==========\n\n- SHA1 (openssh-10.6p2.tar.gz) = c91a0b7860c80f29db874044af4d1a6f906231d9\n- SHA256 (openssh-10.6p2.tar.gz) = muRIcgkM8vBrPPn/FH5/FvMJZxNAdj7S+ZJIOYrinVs=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.com\nOpenSSH 10.6 was released on 2026-10-06. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nRecently the OpenSSH team have received a large number of security\nbug reports, many of which are findings from AI models or made with\nAI assistance. While many AI reports are determined not to have\nsecurity impact when considered in the context of a realistic\nthreat model, we very much welcome these reports, especially when\ncombined with human triage, analysis, test-cases and particularly\nwhen accompanied by proposed fixes.\n\n** We have seen a number of cases where a security bug identified\n** by AI tools is subsequently independently discovered by a\n** different researcher. This suggests that adversaries who do not\n** report bugs to OSS projects are likely to be able to discover\n** these bugs too. Given this, the OpenSSH team will, for now, be\n** making more frequent releases to get bugfixes into users' hands\n** more quickly rather than batching them until the next planned\n** release.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n-------------------------\n\n * scp(1): begin deprecating the -R flag, which is used to perform a\n remote-to-remote copy by executing scp on a remote host. This\n option is a fragile optimisation that is difficult to use because\n it requires credentials on the remote host. It also creates\n security risks if the shell quoting rules on the remote system\n where the copy is performed differ from the client's expectations.\n\n From OpenSSH 10.6, this option will continue to work but will\n cause a deprecation warning to be emitted to standard error. In\n a future release, the option will be ignored and will leave in\n place the default remote-to-remote copy behaviour (copy via the\n host running scp).\n\n * sshd(8): support for platforms that do not allow file descriptor\n passing and that also require root privilege for PTY allocation\n will be removed in a future release. Affected platforms are known\n to include SCO OpenServer 5 and QNX 6 but may include other\n similarly old operating systems. This deprecation can be avoided\n if the user community for these platforms is able to assist us in\n building alternatives, such as avoiding the need for root in PTY\n allocation.\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh(1), sshd(8): compression will be less effective as a result\n of the change noted below in the "Security" section\n\n * ssh(1): destination usernames entered on the commandline are now\n more strictly checked and will refuse usernames that include\n backslash and dollar symbols. Usernames that are specified by\n the "User" directive in configuration files are not subject to\n these restrictions. The motivation for this change is mentioned\n below in the "Security" section.\n\n * sshd(8): on platforms that do not support file descriptor passing\n and that require root for PTY allocation, the GatewayPorts and\n StreamLocalForwarding options are forcibly disabled. The\n motivation for this change is discussed below in the "Security"\n section.\n\nChanges since OpenSSH 10.5\n==========================\n\nThis release contains a number of security fixes, several new\nfeatures and some small bugfixes.\n\nSecurity\n========\n\n * sftp(1): more strictly validate paths returned from the server to\n avoid some cases where a server could return paths that could\n manipulate a recursive copy operation into writing outside its\n target directory. Report and patch from Junghoon Cho.\n\n * sshd(8): when GSSAPIAuthentication is in use, only store GSSAPI\n credentials when authentication has succeeded. Avoids a situation\n where credentials from a failed GSSAPIAuthentication attempt may\n persist and be made inappropriately available if another\n authentication subsequently succeeds. Issue report and patch from\n Moritz Theile.\n\n * sshd(8): reset GSSAPIAuthentication before authentication, avoiding\n state from one authentication attempt being confused with that of\n a later attempt. Report and feedback from Moritz Theile.\n\n * sshd(8), ssh(1): disable LZ77 dictionary coder to mitigate the\n side-channel leaks described in "Crossing the Streams: SSH\n Plaintext Recovery via a Common Compression Context in\n Multiplexed Channels" by Fabian Bäumer and Marcus Brinkmann,\n preprint https://arxiv.org/abs/2609.07709 (2026)\n\n A chosen-plaintext attack method exists which makes use of\n dictionary-based compression to recover secrets from one channel\n by interacting with the SSH session's shared compression\n dictionary through another channel.\n\n Attacker-controlled input can recognizably reflect into the\n total length of transmitted ciphertexts by virtue of LZ77\n replacing repeated strings with back-references into the SSH\n session's encoder search buffer, which is shared across all\n channels. For this reason, the documentation already recommended\n against enabling compression for connections that share trusted\n and untrusted traffic.\n\n This change will reduce the effectiveness of the Compression\n option. Users are encouraged to use application-level compression\n over the SSH protocol where possible, as this will typically be\n more effective and will be completely immune to this type of\n attack.\n\n * ssh(1): disallow '$' and '\\' characters in usernames entered on\n the command-line to avoid usernames from untrusted sources\n yielding injection in shell context via ProxyCommand, Match exec,\n etc. Usernames specified via the configuration files are not\n subject this this control. Reported by SecBuddyF KeenLab Tencent\n (CodeBuddy Security)\n\n We continue to recommend against directly exposing ssh(1) and\n other tools' command-lines to untrusted input. Mitigations such\n as this can not be absolute given the variety of shells and user\n configurations in use.\n\n * ssh-keygen(1): correct handling of Daylight Saving Time when\n converting dates. Previous handling could cause errors of\n up to +/- 1 hour (unless you are in the Antarctica/Troll\n timezone, where the error could be +/- 2 hours). These errors\n could result in creation of certificates with incorrect expiry\n times. bz4004; from Khush Patel\n\n * sshd(8), ssh(1): ensure that compressed payloads don't inflate\n past the maximum supported packet length. Reported by Oleh Konko.\n\n * sshd(8): fully honor the authorized_keys "restrict" keyword,\n which was not being properly applied to tunnel forwarding\n (PermitTunnel, disabled by default). This is a separate problem\n to the one fixed in openssh-10.5.\n\n * sshd(8): correctly handle some options that accept "none". Some\n options, including AuthorizedPrincipalsFile, were documented as\n accepting "none" as a way to disable them; however, when\n overridden by an sshd_config(5) Match keyword, this argument was\n being incorrectly interpreted as a literal file.\n With Chris Rohlf in collaboration with Claude and Anthropic Research\n\n * sshd(8): On OS X SDK >= 27, sandboxing is no longer supported\n as the API we depended upon has been removed and no obvious\n alternative provided.\n\n * sshd(8): on platforms that do not support file descriptor passing\n and that require root for PTY allocation, the post-authentication\n sshd-session process retains root privilege, whereas on other\n platforms this process runs with the privilege of the logged-in\n user. When sshd-session was run with elevated privilege, it could\n perform certain actions as root and circumvent controls that\n would normally have applied to the user, such as making unix\n domain socket connections or binding (via -R forwarding) to low-\n numbered TCP ports.\n\n For this reason, this release disables the GatewayPorts and\n StreamLocalForwarding options and support for these (few)\n platforms will be removed in future if no alternatives to\n requiring privilege in the post-authentication process are found.\n Affected platforms include QNX 6, SCO OpenServer 5 and builds\n that were made with the --disable-fd-passing configure option.\n\n This problem was reported separately by sn0x-sharma and by Dark\n River.\n\nNew features\n------------\n\n * All: enable hybrid post-quantum ssh-mldsa44-ed25519 signature\n algorithm. Note that this no longer uses the "@openssh.com"\n vendor extension suffix that the previous experimental\n implementation used. Keys generated with the previous experimental\n support must be regenerated and/or removed.\n\n * sshd(8): Add the WarnWeakCrypto option to sshd_config(5). This\n option was previously available for the client only. This option\n is enabled by default and will log when the client uses a key\n agreement scheme that is not post-quantum safe.\n\n * ssh-keygen(1), ssh-add(1): preserve user-verification (PIN or\n biometric) requirement for resident keys loaded from a FIDO\n token, by checking the credential's credProtect policy.\n GHPR701 from Savely Krasovsky\n\n * ssh(1): include local and remote version strings in the ~I\n connection information display.\n\n * ssh-add(1): add a -P flag to skip PIN entry for FIDO and PKCS#11\n tokens that do not require it.\n\n * sftp(1): add '-p' flag for mkdir/lmkdir to create directories as\n required. This flag has similar ergonomics to mkdir(1), and\n previously-existing directories do not cause an error.\n\n * ssh-keygen(1): add a "hexdump" key export mode that dumps the SSH\n wire-formatted key blob in hex format. Useful when writing\n documentation, tests, etc. E.g. `ssh-keygen -em hexdump -f /key`\n\n * ssh(1), sshd(8): ChannelTimeout now accepts timeouts with\n fractional seconds.\n\n * sshd(8): allow specification of the location of $SSH_AUTH_SOCK\n used for agent forwarding using a new AgentSocketPath option.\n This supports both using a user-specific path, such as the\n default of a subdirectory of $HOME ("user:.ssh/agent"), and\n the previous approach of allowing agent forwarding sockets to be\n located in a shared directory (e.g. "shared:/tmp"). Sockets\n created in shared directories will be created inside a\n subdirectory with a randomised name. bz3860\n\n * ssh-agent(1): allow specification of agent socket directories\n using a -A flag. It accepts "user:" and "shared:" directory\n styles similar to the sshd AgentSocketPath option.\n\n * sshd(8): account for public key authentication "key ok" tests\n separately to auth attempts. Add a `PubkeyAuthOptions max-pk-ok:nnn`\n option to allow a number of PK_OK tests (asking whether the\n server might accept a given public key) that do not count against\n MaxAuthTries, defaulting to 6 attempts. After these attempts are\n exhausted, further attempts count as failed authentications\n against MaxAuthTries. Practically, this allows more keys on disk\n or held in ssh-agent to be checked for use before the server\n disconnects.\n\n * ssh(1), sshd(8): extend the existing TCPKeepAlive option to also\n support setting keepalives on sockets created for forwarding\n connections. Previously this option controlled keepalives on the\n connection socket only. TCPKeepAlive "yes" or "transport" enables\n keepalives on the connection socket. "TCPKeepAlive all" additionally\n enables them for forwarding sockets. bz3921\n\nBugfixes\n--------\n\n * sshd(8), ssh(1): fix configuration matching on more Turkic\n languages which have disjoint dotted and dotless i/I characters,\n specifically Azerbaijani and Crimean Tatar. bz3991\n\n * ssh(1), sshd(8): don't attempt to set TCP_NODELAY on non-IP/IPv6\n sockets. Eliminates some noise in debug logs.\n\n * ssh(1): fix case for ssh -G option output; bz4005\n\n * ssh(1), sshd(8): Fix ChannelTimeout specificity; previously a\n more specific channel type (e.g. "session:shell") could clobber\n a user-specified ChannelTimeout if it was less specific (e.g.\n "session"). Also, in some cases, the debug messages were\n printing 0 instead of the effective timeout. bz3994\n\n * sftp(1): avoid NULL dereference crash in some circumstances when\n a server fails a stat/lstat operation. GHPR707\n\n * sshd(8): close a race condition where a SIGTERM/SIGQUIT would be\n ignored if it was received by the server while it was processing a\n SIGHUP restart request. bz3981\n\n * sshd(8), ssh(1): check key and CA signature types during key\n parsing against allowlists (PubkeyAcceptedAlgorithms, etc) as\n early as possible. This reduces the attack surface presented by\n disabled algorithms. Suggested by and with extensive feedback\n from Chris Rohlf in collaboration with Claude and Anthropic\n Research.\n\n * All: switch the fallback implementation of the ed25519 signature\n algorithm used when libcrypto is disabled from SUPERCOP ed25519\n to libsodium. The libsodium implementation includes a number of\n strictness checks over the original reference implementation we\n have used to this point and a more ergonomic API.\n\n * ssh-add(1), ssh(1), ssh-keygen(1): fix spin on password entry when\n the program attempting to read a password was started in a\n background process group, with no TTY and with certain signals\n ignored. bz3995\n\n * scp(1): disallow nul byte in received scp -O filename. This was\n not reachable in normal operation. Reported by Chua Wei Xun.\n\n * ssh-keygen(1), ssh(1), sshd(8): implement a maximum number of KDF\n rounds that will be accepted when writing an OpenSSH-format\n private key or when loading one. This limit is set quite high\n (1M), but ensures that a service that is passed a bad key with a\n ridiculously high number of rounds will eventually complete\n parsing it.\n\n * ssh-keygen(1): bump the default number of KDF rounds from 24 to\n 32 (this is a linear increase, not like bcrypt(3) which is\n exponential).\n\n * ssh(1): make StreamLocalBindMask properly respect Host/Match\n blocks and make it first-match-wins as documented. bz4013\n\n * sshd(8): make StreamLocalBindMask properly first-match-wins.\n\nPortability\n-----------\n\n * All: remove the NetBSD BROKEN_READ_COMPARISON workaround. This\n appears to be no longer required and caused pre-auth CPU spinning.\n\n * sshd(8): don't link sshd against libselinux when SELinux support\n is enabled (note: this library is still linked for the sshd-auth\n and sshd-session helper binaries).\n\n * sshd(8): allow madvise(..., MADV_DONTNEED_LOCKED) in the seccomp\n sandbox; needed by GrapheneOS' hardened allocator. bz4001\n\n * sshd(8): restrict mremap(2) flags accepted by the seccomp\n sandbox. Only MREMAP_MAYMOVE is now accepted as other flags may\n have some utility in attack chains. Reported by: Mohammad Hossein\n Abedini.\n\n * sshd(8): allow PAMServiceName in Match (regressed in 10.4).\n During the refactor of server option parsing, the ability to set\n PAMServiceName in Match blocks was accidentally disabled.\n\n * sshd(8): the --disable-fd-passing configure option has been\n removed.\n\nChecksums:\n==========\n\n - SHA1 (openssh-10.6.tar.gz) = 096e9cd60cd08e0acacfe16b615144f8c0463da5\n - SHA256 (openssh-10.6.tar.gz) = 0lP9h9/QH3k667vqbxKhkkusc3CN5iKUiDNI9fsR06A=\n - SHA1 (openssh-10.6p1.tar.gz) = e6a34f5625e20172b214c50ebd4e0dcb4fd18013\n - SHA256 (openssh-10.6p1.tar.gz) = qdyVZd/+hkD2TYY80poyvEo9vewFZqf8RMXW7nZ9Xzk=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.5 was released on 2026-08-11. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nRecently the OpenSSH team have received a large number of security\nbug reports, many of which are findings from AI models or made with\nAI assistance. While many AI reports are determined not to have\nsecurity impact when considered in the context of a realistic\nthreat model, we very much welcome these reports, especially when\ncombined with human triage, analysis, test-cases and particularly\nwhen accompanied by proposed fixes.\n\nWe have seen a number of cases where a security bug identified by\nAI tools is subsequently independently discovered by a different\nresearcher. This suggests that adversaries who do not report bugs\nto OSS projects are likely to be able to discover these bugs too.\nGiven this, the OpenSSH team will, for now, be making more frequent\nreleases to get bugfixes into users' hands more quickly rather than\nbatching them until the next planned release.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n--------------------------------\n\n * Portable OpenSSH now requires ECC (Elliptic Curve Cryptography)\n support in libcrypto, including support for the NISTP521 curve.\n ECC is included in the default build configurations of all\n versions of all libcrypto implementations currently supported by\n OpenSSH, including LibreSSL, OpenSSL, BoringSSL and AWS LC.\n The --without-openssl build configuration is not affected.\n\nChanges since OpenSSH 10.4\n==========================\n\nThis release contains a number of security fixes and small bugfixes.\n\nSecurity\n========\n\n * ssh-agent(1): fix an interaction between agent locking and the\n session-bind@openssh.com extension that is used to identify\n forwarded agents. These binding requests were refused when the\n agent was locked, with the result that operations that were\n intended to be limited to local use only could be performed\n remotely, including the ability to add PKCS#11 tokens and make\n use of keys that had destination restrictions applied.\n Reported by sn0x-sharma\n\n * ssh(1): avoid potential realloc use-after-free in the client if a\n remote forwarding is added via the local session multiplexing\n socket while a remote forwarding open request is pending with the\n server. Report and fix from Brian Mingus of Cognatory\n\n * sshd(8): make the authorized_keys "restrict" keyword apply\n correctly to tunnel forwarding too (which is administratively\n disabled by default). Reported by Erichen, Institute of Computing\n Technology, Chinese Academy of Sciences\n \nNew features\n------------\n\n * ssh-keygen(1): add ability to set or clear the touch-required and\n verify-required flags on FIDO private keys when resetting a\n private key's passphrase.\n\n * ssh(1): tweak ordering of certificates tried during pubkey\n authentication to prefer FIDO keys that do not require user\n presence (touch) first, and FIDO keys that require user\n verification via PIN or biometrics last. This effectively tries\n low-friction authenticators before higher friction ones.\n\n * ssh(1): add a "ssh -Z user@host" mode that prints the keys that\n will be tried for public key authentication in the order that\n they will be used.\n \n * sshd(8) use setproctitle(3) to identify sshd-session when its\n acting as a post-authentication monitor.\n \nBugfixes\n--------\n\n * ssh-keyscan(1): make reading the server banner a non-blocking\n operation to prevent a stuck server from blocking a many-host\n keyscan from proceeding.\n\n * sshd(8): use sshpkt_fatal() instead of plain fatal() for errors\n in the packet code as this provides context of the failing peer\n (address, port, user, etc).\n\n * sshd(8): when signing hostkey proofs for a client UpdateHostKeys\n request, allow each hostkey to perform at most one signature\n operation.\n \n * sshd(8) fix GSSAPI option names, that were broken during a\n servconf.c refactoring in openssh-10.4; bz3974.\n \n * ssh-keygen(1): pass back errors from ed25519 key generation, which\n theoretically can fail. GHPR702.\n\n * sshd(8): move check of public key type against allowed algorithms\n to before parsing of the key sent by the peer. This removes at\n least some key parsing and verification paths from the pre-auth\n attack surface. Suggested by Christopher Paul Rohlf of Anthropic.\n \n * ssh-keygen(1): fix double frees (impossible to reach outside of a\n test harness), and also use freezero where possible. From\n Christopher Paul Rohlf at Anthropic.\n\n * sshd(8): fix ChannelTimeout and RekeyLimit not being applied in\n sshd_config Match blocks.\n\n * sshd(8): in sshd config dump mode, write all directives in mixed\n case for consistency\n \nPortability\n-----------\n\n * sshd(8): re-allow PAMServiceName inside a Match block, which\n was incorrectly disabled during a refactoring in openssh-10.4.\n bz3987\n\nChecksums:\n==========\n\n - SHA1 (openssh-10.5.tar.gz) = 273163972f623bb9bffef9fd75a85c74d3b22633\n - SHA256 (openssh-10.5.tar.gz) = 9Zhp0C/mDWNLmnatq68mtbqOUhvbyYcffH04dsHUwZQ=\n\n - SHA1 (openssh-10.5p1.tar.gz) = 3067e2af7c526b31c7e94bc3ed38904ef791eb2c\n - SHA256 (openssh-10.5p1.tar.gz) = 1E0oqDnqna+WnMaRUP3lmRCys5Nh2tgaO9bL0ZIY2xE=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.4 was released on 2026-07-06. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n--------------------------------\n\n * sshd(8): configuration dump mode ("sshd -G") now writes directives\n in mixed case (e.g. "PubkeyAuthentication") whereas previously it\n emitted only lower-case names.\n\n * sshd(8): on Linux systems with the seccomp sandbox enabled,\n failures to enable SECCOMP or NO_NEW_PRIVS are now fatal.\n Previously sshd(8) would log the error but continue operation,\n to support systems that lacked these features. Now systems that\n lack these should instead disable the sandbox at configure time.\n\n * ssh(1), sshd(8): make the transport protocol stricter by\n disconnecting if the peer sends non-KEX messages during a post-\n authentication key re-exchange. Previously a malicious peer could\n continue sending non-key exchange messages without penalty. These\n would be buffered, causing memory to be wasted up until the\n connection terminated or the server/client hit a memory limit.\n Implementations that do not restrict messages sent during key\n exchange as per RFC4253 section 7.1 may be disconnected.\n Reported by Marko Jevtic.\n\nChanges since OpenSSH 10.3\n==========================\n\nThis release contains a number of security fixes as well as general\nbugfixes and a couple of new features.\n\nSecurity\n========\n\n * sftp(1): when downloading files on the command-line using\n "sftp host:/path .", a malicious server could cause the file to\n be downloaded to an unexpected location. This issue was identified\n by the Swival Security Scanner.\n\n * scp(1): when copying files between two remote destinations, do\n not allow a malicious server to write files to the parent\n directory of the intended target directory. This issue was\n identified by the Swival Security Scanner.\n\n * sshd(8): when using the "internal-sftp" SFTP server implementation\n (this is not the default), long command lines were previously\n truncated silently after the 9th argument. If a security-relevant\n option was in the 10th or later position, it would be discarded.\n Reported by Steve Caffrey.\n\n * sshd(8): add a documentation note to mention that the\n GSSAPIStrictAcceptorCheck option is ineffective when the server\n is joined to a Windows Active Directory. Reported by Yarin Aharoni\n of Safebreach.\n\n * sshd(8): DisableForwarding=yes didn't override PermitTunnel=yes\n as it was documented to do. Note that PermitTunnel is not enabled\n by default. Reported independently by Huzaifa Sidhpurwala of\n Redhat and Marko Jevtic.\n\n * sshd(8): avoid a potential pre-authentication denial of service\n when GSSAPIAuthentication was enabled (this feature is off by\n default). This was not mitigated by MaxAuthTries, but would be\n penalised by PerSourcePenalties. This was reported by Manfred\n Kaiser of the milCERT AT (Austrian Ministry of Defence).\n\n * sshd(8): fix a number of cases where the minimum authentication\n delay was not being enforced. Reported by the Orange Cyberdefense\n Vulnerability Team.\n\n * ssh(1): fix a possible client-side use-after-free if the server\n changes its host key during a key reexchange. This was reported by\n Zhenpeng (Leo) Lin of Depthfirst.\n\nNew features\n------------\n\n * All: add experimental support for a composite post-quantum\n signature scheme that combines ML-DSA 44 and Ed25519 as specified\n in draft-miller-sshm-mldsa44-ed25519-composite-sigs.\n\n This scheme is not enabled by default. To use it, you'll need\n to add it to HostKeyAlgorithms, PubkeyAcceptedAlgorithms, etc.\n Keys may be generated using "ssh-keygen -t mldsa44-ed25519".\n\n * ssh(1), sshd(8): replace the wildcard pattern matcher with an\n implementation based on an NFA. This avoids exponential worst-case\n behaviour for the old implementation.\n\nBugfixes\n--------\n\n * ssh-agent(1): fix incorrect reply to "query" SSH_AGENTC_EXTENSION\n requests. bz3967\n\n * sshd(8): avoid sending observably different messages for valid vs\n invalid users in GSSAPIAuthentication (disabled by default).\n\n * ssh(1), sshd(8): fix several bugs that incorrectly\n classified bulk traffic as interactive. bz3972, bz3958\n\n * ssh-keygen(1), ssh-add(1): skip unsupported key types when\n downloading resident keys from a FIDO token. Previously, downloads\n would abort when one was encountered. GHPR657\n\n * ssh(1): fix a potential use-after-free on an error path if\n cipher_init() fails.\n\n * sshd(8): perform stricter encoding and validation of transport\n state passed between sshd privilege separation subprocesses. This\n somewhat further hardens the server against attacks on sshd-auth\n or sshd-session subprocesses.\n\n * ssh-agent(1): avoid possible runtime denial of service by\n enforcing some limits on the length of usernames in key use\n constraints.\n\n * sftp(1): fix two separate one-byte out-of-bounds reads, in\n SSH2_FXP_REALPATH and batch command processing.\n\n * sftp-server(8): disallow use of the copy-data extension to read\n and write to the same inode simultaneously.\n\n * ssh(1), sshd(8): avoid strlen(NULL) crash if an X11 channel was\n created before the x11-req SSH_MSG_CHANNEL_REQUEST was sent.\n GHPR679\n\n * sftp(1), scp(1): avoid a situation where sftp_download() could get\n stuck in a loop if a broken server repeatedly returned zero length\n while reading a file.\n\n * ssh(1): avoid leaking DNS0x20 case-randomised names into names\n canonicalised using CanonicalizePermittedCNAMEs. bz3966\n\n * sftp-server(8): avoid truncation of pathnames passed to lstat()\n during SSH_FXP_REALPATH handling on systems where PATH_MAX is not\n the actual max. GHPR688\n\n * ssh(1), sshd(8): correct arming of poll(2) event masks for some\n socket-type channels. GHPR660\n\n * sshd(8): major refactor of sshd_config parsing and management\n code, to allow for more exact serialisation/deserialisation across\n privilege separation boundaries.\n\n * ssh-add(1): open connection to the agent only after getopt()\n processing has completed, to give options like "-v" a chance to\n display debug information about this operation.\n\n * crypto code: fix bounds checking when signing messages of length\n greater than will fit in a size_t. In OpenSSH, message sizes are\n bounded by SSHBUF_SIZE_MAX so this was unreachable.\n\n * crypto code: add signature malleability and pubkey validity checks\n to ed25519 verification. SSH doesn't depend on these properties\n\n * crypto code: fix ECDSA order check for curves with cofactor != 1.\n All supported EC curves have cofactor 1, so this was\n unreachable.\n\n * sshd(8): differentiate between execution failures and a subsystem\n that was not found when logging why a subsystem failed to start.\n GHPR637\n\n * All: use safer idioms for timegm(3) and mktime(3) error detection.\n\n * ssh(1), sshd(8): avoid accepting invalid cipher or MAC lists in\n config files or command-line arguments. This could cause runtime\n failures later.\n\n * ssh(1): fix NULL deref crash during pubkey auth when using a PEM\n style private key with no corresponding .pub key adjacent to it.\n\n * sshd(8): don't print an error message when trying to load a host\n private key when PKCS#11 keys are in use, as these don't need the\n private half on the filesystem. GHPR664\n\n * All: don't use deprecated ERR_load_crypto_strings(). GHPR650\n\n * ssh(1): properly report errors during configuration default\n setting. GHPR649\n\n * ssh(1): use correct directive name (Match instead of Host) in\n error message. bz3968\n\n * sftp(1): fix "ls -ln" which was not correctly showing numeric\n UID/GIDs but rather user and group names. bz3953\n\n * sshd(8): avoid possible NULL dereference if an allocation fails\n during config parsing. bz3948\n\n * All: fix ineffective guards against loading overly large public\n keys in several places. bz3969 and bz3970\n\n * sftp(1): ensure file descriptors used by sftp to communicate to\n its ssh(1) subprocess don't leak into executed subprocesses (e.g.\n via "!"). GHPR693\n\nPortability\n-----------\n\n * Sync fmt_scaled.c with OpenBSD upstream, picking up an exactness\n fix for large exponents (GHPR671)\n\n * sshd(8): remove duplicate sandbox entry for clock_gettime64.\n\n * ssh(1), sshd(8): use correct IPTOS_DSCP_VA value if not provided\n by the system headers.\n\n * Sync getrrsetbyname.c with OpenBSD upstream, picking up robustness\n fixes.\n\n * Disable replacements in openbsd-compat for strvisx(3) and\n stravis(3), as these are unused in OpenSSH\n\n * Avoid fortify warnings on Android bz3954\n\n * Fix a number of memory leaks on error paths in the portability\n code. GHPR681\n\n * Revise the README.privsep documentation to reflect sshd's recent\n switch to a multi-binary model.\n\nChecksums:\n==========\n\n - SHA1 (openssh-10.4.tar.gz) = 8502b516230865e229d55045bb4ccc67aeae905b\n - SHA256 (openssh-10.4.tar.gz) = qUVI+wMg4mVpiQbXvfMVkF3Zuyy2JrS+ZjN764mtyGE=\n\n - SHA1 (openssh-10.4p1.tar.gz) = ae8650a71cc52dbbd049519cee276ae6d65c2c4d\n - SHA256 (openssh-10.4p1.tar.gz) = 72Am3SrqjVYFljjV0yYpAsiSzrqfiDlYNeDQbT+2Mjg=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.3 was released on 2026-04-02. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh(1), sshd(8): remove bug compatibility for implementations\n that don't support rekeying. If such an implementation tries to\n interoperate with OpenSSH, it will now eventually fail when the\n transport needs rekeying.\n\n * sshd(8): prior to this release, a certificate that had an empty\n principals section would be treated as matching any principal\n (i.e. as a wildcard) when used via authorized_keys principals=""\n option. This was intentional, but created a surprising and\n potentially risky situation if a CA accidentally issued a\n certificate with an empty principals section: instead of being\n useless as one might expect, it could be used to authenticate as\n any user who trusted the CA via authorized_keys. [Note that this\n condition did not apply to CAs trusted via the sshd_config(5)\n TrustedUserCAKeys option.]\n\n This release treats an empty principals section as never matching\n any principal, and also fixes interpretation of wildcard\n characters in certificate principals. Now they are consistently\n implemented for host certificates and not supported for user\n certificates.\n\n * ssh(1): the -J and equivalent -oProxyJump="..." options now\n validate user and host names for ProxyJump/-J options passed\n via the command-line (no such validation is performed for this\n option in configuration files). This prevents shell injection in\n situations where these were directly exposed to adversarial\n input, which would have been a terrible idea to begin with.\n Reported by rabbit.\n\nChanges since OpenSSH 10.2\n==========================\n\nThis release contains some relatively minor security fixes as well\nas a number of feature improvements and general bugfixes.\n\nSecurity\n========\n\n * ssh(1): validation of shell metacharacters in user names supplied\n on the command-line was performed too late to prevent some\n situations where they could be expanded from %-tokens in\n ssh_config. For certain configurations, such as those that use a\n "%u" token in a "Match exec" block, an attacker who can control\n the user name passed to ssh(1) could potentially execute arbitrary\n shell commands. Reported by Florian Kohnhäuser.\n\n We continue to recommend against directly exposing ssh(1) and\n other tools' command-lines to untrusted input. Mitigations such\n as this can not be absolute given the variety of shells and user\n configurations in use.\n\n * sshd(8): when matching an authorized_keys principals="" option\n against a list of principals in a certificate, an incorrect\n algorithm was used that could allow inappropriate matching in\n cases where a principal name in the certificate contains a\n comma character. Exploitation of the condition requires an\n authorized_keys principals="" option that lists more than one\n principal *and* a CA that will issue a certificate that encodes\n more than one of these principal names separated by a comma\n (typical CAs strongly constrain which principal names they will\n place in a certificate). This condition only applies to user-\n trusted CA keys in authorized_keys, the main certificate\n authentication path (TrustedUserCAKeys/AuthorizedPrincipalsFile)\n is not affected. Reported by Vladimir Tokarev.\n\n * scp(1): when downloading files as root in legacy (-O) mode and\n without the -p (preserve modes) flag set, scp did not clear\n setuid/setgid bits from downloaded files as one might typically\n expect. This bug dates back to the original Berkeley rcp program.\n Reported by Christos Papakonstantinou of Cantina and Spearbit.\n\n * sshd(8): fix incomplete application of PubkeyAcceptedAlgorithms\n and HostbasedAcceptedAlgorithms with regard to ECDSA keys.\n Previously if one of these directives contains any ECDSA algorithm\n name (say "ecdsa-sha2-nistp384"), then any other ECDSA algorithm\n would be accepted in its place regardless of whether it was\n listed or not. Reported by Christos Papakonstantinou of Cantina\n and Spearbit.\n\n * ssh(1): connection multiplexing confirmation (requested using\n "ControlMaster ask/autoask") was not being tested for proxy mode\n multiplexing sessions (i.e. "ssh -O proxy ..."). Reported by\n Michalis Vasileiadis.\n\nNew features\n------------\n\n * ssh(1), sshd(8): support IANA-assigned codepoints for SSH agent\n forwarding, as per draft-ietf-sshm-ssh-agent. Support for the new\n names is advertised via the EXT_INFO message. If a server offers\n support for the new names, then they are used preferentially.\n\n Support for the pre-standardisation "@openssh.com" extensions for\n agent forwarding remains supported.\n\n * ssh-agent(1): implement support for draft-ietf-sshm-ssh-agent\n "query" extension.\n\n * ssh-add(1): support querying the protocol extensions via the\n agent "query" extension with a new -Q flag.\n\n * ssh(1): support multiple files in a ssh_config RevokedHostKeys\n directive. bz3918\n\n * sshd(8): support multiple files in a sshd_config RevokedKeys\n directive bz3918\n\n * ssh(1): add a ~I escape option that shows information about the\n current SSH connection.\n\n * ssh(1): add an "ssh -Oconninfo user@host" multiplexing command\n that shows connection information, similar to the ~I escapechar.\n\n * ssh(1): add an "ssh -O channels user@host" multiplexing command to\n get a running mux process to show information about what channels\n are currently open.\n\n * sshd(8): add 'invaliduser' penalty to PerSourcePenalties, which is\n applied to login attempts for usernames that do not match real\n accounts. Defaults to 5s to match 'authfail' but allows\n administrators to block such attempts for longer if desired.\n\n * sshd(8): add a GSSAPIDelegateCredentials option for the server,\n controlling whether it accepts delegated credentials offered by\n the client. This option mirrors the same option in ssh_config.\n GHPR614\n\n * ssh(1), sshd(8): support the VA DSCP codepoint in the IPQoS\n directive.\n\n * sshd(8): convert PerSourcePenalties to using floating point time,\n allowing penalties to be less than a second. This is useful if you\n need to penalise things you expect to occur at >=1 QPS.\n\n * ssh-keygen(1): support writing ED25519 keys in PKCS8 format.\n GHPR570\n\n * Support the ed25519 signature scheme via libcrypto.\n\nBugfixes\n--------\n\n * sshd(8): make IPQoS first-match-wins in sshd_config, like other\n configuration directives. bz3924\n\n * sshd(8): fix potential crash when MaxStartups is using a single\n argument (i.e. not using the MaxStartps x:y:z form) to a value\n below 10. bz3941\n\n * sshd(8): fix a potential hang during key exchange if needed DH\n group values were missing from /etc/moduli.\n\n * ssh-agent(1): fix return values from extensions to be correct wrt\n draft-ietf-sshm-ssh-agent: extension requests should indicate\n failure using SSH_AGENT_EXTENSION_FAILURE rather than the generic\n SSH_AGENT_FAILURE error code. This allows the client to discern\n between "the request failed" and "the agent doesn't support this\n extension".\n\n * ssh(1): use fmprintf for showing challenge-response name and info\n to preserve UTF-8 characters where appropriate. Prompted by GitHub\n PR#452.\n\n * scp(1): when uploading a directory using sftp/sftp (e.g. during a\n recursive transfer), don't clobber the remote directory\n permissions unless either we created the directory during the\n transfer or the -p flag was set. bz3925\n\n * All: implement missing pieces of FIDO/webauthn signature support,\n mostly related to certificate handling and enable acceptance of this\n signature format by default. bz3748 GHPR624 GHPR625\n\n * sshd_config(5): make it clear that DenyUsers/DenyGroups overrides\n AllowUsers/AllowGroups. Previously we specified the order in which\n the directives are processed but it was ambiguous as to what\n happened if both matched.\n\n * ssh(1): don't try to match certificates held in an agent to\n private keys. This matching is done to support certificates that\n were loaded without their private key material, but is\n unnecessary for agent-hosted certificate which always have\n private key material available in the agent. Worse, this matching\n would mess up the request sent to the agent in such a way as to\n break usage of these keys when the key usage was restricted in\n the agent. bz3752\n\n * sftp(1): if editline has been switched to vi mode (i.e. via "bind\n -v" in .editrc), setup a keybinding so that command mode can be\n entered.\n\n * ssh(1), sshd(8): improve performance of keying the sntrup761 key\n agreement algorithm.\n\n * ssh(1), sshd(8): enforce maximum packet/block limit during\n pre-authentication phase.\n\n * sftp(1): don't misuse the sftp limits extension's open-handles\n field. This value is supposed to be the number of handles a\n server will allow to be opened and not a number of outstanding\n read/write requests that can be sent during an upload/download.\n\n * sshd(8): don't crash at connection time if the main sshd_config\n lacks any subsystem directive but one is defined in a Match block.\n bz3906\n\n * sshd_config(5): add a warning next to the ForceCommand directive\n that forcing a command doesn't automatically disable forwarding.\n\n * sshd_config(5): add a warning that TOKENS are replaced without\n filtering or escaping and that it's the administrator's\n responsibility to ensure they are used safely in context.\n\n * scp(1): correctly quote filenames in verbose output for local->\n local copies. bz3900\n\n * sshd(8): don't mess up the PerSourceNetBlockSize IPv6 mask if\n sscanf didn't decode it. GHPR598\n\n * ssh-add(1): when loading FIDO2 resident keys, set the comment to\n the FIDO application string. This matches the behaviour of\n ssh-keygen -K. GHPR608\n\n * sshd(8): don't strnvis() log messages that are going to be logged\n by sshd-auth via its parent sshd-session process, as the parent\n will also run them though strnvis(). Prevents double-escaping of\n non-printing characters in some log messages. bz3896\n\n * ssh-agent(1): escape SSH_AUTH_SOCK paths that are sent to the\n shell as setenv commands. Unbreaks ssh-agent for home directory\n paths that contain whitespace. bz3884\n\n * All: Remove unnecessary checks for ECDSA public key validity.\n\n * sshd(8): activate UnusedConnectionTimeout only after the last\n channel has closed. Previously UnusedConnectionTimeout could fire\n early after a ChannelTimeout. This was not a problem for the\n OpenSSH client because it terminates once all channels have\n closed but could cause problems for other clients (e.g. API\n clients) that do things differently. bz3827\n\n * All: fix PKCS#11 key PIN entry problems introduced in\n openssh-10.1/10.2. bz3879\n\n * scp(1): when using the SFTP protocol for transfers, fix implicit\n destination path selection when source path ends with "..". bz3871\n\n * sftp(1): when tab-completing a filename, ensure that the completed\n string does not end up mid-way through a multibyte character, as\n this will cause a fatal() later on. GHPR#587\n\n * ssh-keygen(1): fix crash at exit (visible via ssh-keygen -D) when\n multiple keys loaded.\n\n * scp(1)/sftp(1): correctly display bandwidths >2GBps in the\n progress meter.\n\nPortability\n-----------\n\n * sshd(8): fix condition introduced in openssh 10.2p1 stable branch\n here a PAM module that changed the requested username between\n SSH_MSG_USERAUTH_REQUEST messages during authentication could\n confuse the PAM stack and let it proceed with a different\n understanding of the active username than the rest of sshd.\n Reported by Mike Damm.\n\n * sshd(8): immediately report interactive instructions to clients\n when using keyboard-interactive authentication with PAM. bz2876\n\n * sshd(8): fix duplicate PAM messages under some situations.\n\n * sshd(8): don't leak PAM handle on repeat invocations. bz3882\n\n * All: support linking libcrypto implementations (e.g. BoringSSL)\n that require libstdc++.\n\n * sshd(8): fix ut_type for btmp records, correctly using\n LOGIN_PROCESS and USER_PROCESS.\n\n * sshd(8): allow uname(3) in the seccomp sandbox. This is needed by\n zlib-ng on RISC-V platforms.\n\n * All: remove remaining OpenSSL_add_all_algorithms() calls.\n We already have OPENSSL_init_crypto() in the compat layer.\n Prompted by github PR#606\n\n * All: fix builds on older Mac OS wrt nfds_t.\n\n * mdoc2man: several improvements including better support for Dl\n and Ns inside Ic.\n\nChecksums:\n==========\n\n - SHA1 (openssh-10.3.tar.gz) = 854863c04cd28242d73ac6c3ee9c37fa756f1a2f\n - SHA256 (openssh-10.3.tar.gz) = aCU5P47rM+m4N8/i2JOHMOMhafMYqBhvQQSnPXczN5M=\n\n - SHA1 (openssh-10.3p1.tar.gz) = 9c78838ec07af14aff54f3755ac56ce6812452a9\n - SHA256 (openssh-10.3p1.tar.gz) = VmgqNruS3PS08Bb9jsjnQFm3mo3iXBXWcNcx59GORfQ=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.2 was released on 2025-10-10. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation warning\n--------------------------\n\n * A future release of OpenSSH will deprecate support for SHA1 SSHFP\n records due to weaknesses in the SHA1 hash function. SHA1 SSHFP\n DNS records will be ignored and ssh-keygen -r will generate only\n SHA256 SSHFP records.\n\n The SHA256 hash algorithm, which has no known weaknesses, has\n been supported for SSHFP records since OpenSSH 6.1, released in\n 2012.\n\nChanges since OpenSSH 10.1\n==========================\n\nThis is a bugfix release, primarily to fix a problem that rendered\nssh(1) unusable when ControlPersist was enabled.\n\nBugfixes\n--------\n\n * ssh(1): fix mishandling of terminal connections when\n ControlPersist was active that rendered the session unusable.\n bz3872\n\n * ssh-keygen(1): fix download of keys from PKCS#11 tokens.\n\n * ssh-keygen(1): fix CA signing operations when the CA key is held\n in a ssh-agent(1). bz3877\n\n\nPortability\n-----------\n\n * All: support platforms without mmap(2), e.g. WASM builds such as\n https://hterm.org\n\n * All: fix builds on FreeBSD for missing fnctl.h include.\n\n * All: fix builds on MacOS <10.12 Sierra, which lacks\n clock_gettime(3)\n\n * sshd(8): don't PAM_RHOST if the remote host is the "UNKNOWN"\n placeholder name. Avoids potential hangs in some PAM modules as\n they try to resolve it. Note, sshd(8) only uses the "UNKNOWN"\n name when the connection is not on an IPv4 or IPv6 socket.\n\nChecksums:\n==========\n\nSHA1 (openssh-10.2.tar.gz) = 6fcda8004bad0fb0eaee60e8308f91b605ad0dce\nSHA256 (openssh-10.2.tar.gz) = y0rCEdrVc4OJRZLg0u3F0frAgz87ydeTktCk3rQfVj8=\n\nSHA1 (openssh-10.2p1.tar.gz) = c34efade16109f065ec8c834f237bcedd8d7ef5c\nSHA256 (openssh-10.2p1.tar.gz) = zMQsBBmTeVkmP6Hb0W2vwYxWuYTANWLSk3zlamD3mLI=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.1 was released on 2025-10-06. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation warning\n--------------------------\n\n * A future release of OpenSSH will deprecate support for SHA1 SSHFP\n records due to weaknesses in the SHA1 hash function. SHA1 SSHFP\n DNS records will be ignored and ssh-keygen -r will generate only\n SHA256 SSHFP records.\n\n The SHA256 hash algorithm, which has no known weaknesses, has\n been supported for SSHFP records since OpenSSH 6.1, released in\n 2012.\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh(1): add a warning when the connection negotiates a non-post\n quantum key agreement algorithm.\n\n This warning has been added due to the risk of "store now, decrypt\n later" attacks. More details at https://openssh.com/pq.html\n\n This warning may be controlled via a new WarnWeakCrypto ssh_config\n option, defaulting to on. This option is likely to control\n additional weak crypto warnings in the future.\n\n * ssh(1), sshd(8): major changes to handling of DSCP marking/IPQoS\n\n In both client and server the default DSCP (a.k.a IPQoS) values\n were revised and the way these values are used during runtime has\n changed.\n\n Interactive traffic is now assigned to the EF (Expedited\n Forwarding) class by default. This provides more appropriate\n packet prioritisation information for the intermediate network,\n such as wireless media (cf. RFC 8325). Non-interactive traffic\n will now use the operating system default DSCP marking. Both the\n interactive and non-interactive DSCP values may be overridden via\n the IPQoS keyword, described in ssh_config(5) and sshd_config(5).\n\n The appropriate DSCP marking is now automatically selected and\n updated as needed over the course of a connection's lifetime.\n ssh(1) and sshd(8) will switch between the interactive and\n non-interactive IPQoS values depending on the type of SSH\n channels open at the time. For example, if an sftp session is\n using the connection alongside a shell session, then the non-\n interactive value will be used for the duration of the sftp. A\n connection which contains only interactive sessions is marked EF.\n\n * ssh(1), sshd(8): deprecate support for IPv4 type-of-service (ToS)\n keywords in the IPQoS configuration directive.\n\n Type of Service (ToS) was deprecated in the late nineties and\n replaced with the Differentiated Services architecture, which\n has significant advantages for operators because it offers more\n granularity.\n\n OpenSSH switched its default IPQoS from ToS to DSCP values in\n 2018 (openssh-7.7).\n\n IPQoS configurations with 'lowdelay', 'reliability', or\n 'throughput' will be ignored and will instead use the system\n default QoS settings. Additionally, a debug message will be logged\n about the deprecation with a suggestion to use DSCP QoS instead.\n\n * ssh-add(1): when adding certificates to an agent, set the expiry\n to the certificate expiry time plus a short (5 min) grace period.\n\n This will cause the agent to automatically remove certificates\n shortly after they expire. A new ssh-add -N option disables this\n behaviour.\n\n * All: remove experimental support for XMSS keys. This was never\n enabled by default. We expect to implement a new post-quantum\n signature scheme in the near future.\n\n * ssh-agent(1), sshd(8): move agent listener sockets from /tmp to\n under ~/.ssh/agent for both ssh-agent(1) and forwarded sockets\n in sshd(8).\n\n This ensures processes that have restricted filesystem access\n that includes /tmp do not ambiently have the ability to use keys\n in an agent.\n\n Moving the default directory has the consequence that the OS will\n no longer clean up stale agent sockets, so ssh-agent now gains\n this ability.\n\n To support $HOME on NFS, the socket path includes a truncated\n hash of the hostname. ssh-agent will, by default, only clean up\n sockets from the same hostname.\n\n ssh-agent(1) gains some new flags: -U suppresses the automatic\n cleanup of stale sockets when it starts. -u forces a cleanup\n without keeping a running agent, -uu forces a cleanup that ignores\n the hostname. -T makes ssh-agent put the socket back in /tmp.\n\nChanges since OpenSSH 10.0\n==========================\n\nThis release contains a minor security fix as well as a number of\nfeature improvements and bugfixes.\n\nSecurity\n========\n\n* ssh(1): disallow control characters in usernames passed via the\n commandline or expanded using %-sequences from the configuration\n file, and disallow \\0 characters in ssh:// URIs.\n\n If an ssh(1) commandline was constructed using usernames or URIs\n obtained from an untrusted source, and if a ProxyCommand that uses\n the %r expansion was configured, then it may be possible for an\n attacker to inject shell expressions that may be executed when the\n proxy command is started.\n\n We strongly recommend against using untrusted inputs to construct\n ssh(1) commandlines.\n\n This change also relaxes the validity checks in one small way:\n usernames supplied via the configuration file as literals (i.e.\n that have no % expansion characters) are not subject to these\n validity checks. This allows usernames that contain arbitrary\n characters to be used, but only via configuration files. This is\n done on the basis that ssh's configuration is trusted.\n\n This issue was reported by David Leadbeater.\n\nNew features\n------------\n\n * ssh(1), sshd(8): add SIGINFO handlers to log active channel and\n session information.\n\n * sshd(8): when refusing a certificate for user authentication, log\n enough information to identify the certificate in addition to the\n reason why it was being denied. Makes debugging certificate\n authorisation problems a bit easier.\n\n * ssh(1), ssh-agent(1): support ed25519 keys hosted on PKCS#11\n tokens.\n\n * ssh(1): add an ssh_config(5) RefuseConnection option that, when\n encountered while processing an active section in a\n configuration, terminates ssh(1) with an error message that\n contains the argument to the option.\n\n This may be useful for expressing reminders or warnings in config\n files, for example:\n\n Match host foo\n RefuseConnection "foo is deprecated, use splork instead"\n\n * sshd(8): make the X11 display number check relative to\n X11DisplayOffset. This will allow people to use X11DisplayOffset\n to configure much higher port ranges if they really want, while\n not changing the default behaviour.\n\n * unit tests: the unit test framework now includes some basic\n benchmarking capabilities. Run with "make UNITTEST_BENCHMARK=yes"\n on OpenBSD or "make unit-bench" on Portable OpenSSH.\n\nBugfixes\n--------\n\n * sshd(8): fix mistracking of MaxStartups process exits in some\n situations. At worst, this could cause all MaxStartups slots to\n fill and sshd to refuse new connections.\n\n * ssh(1): fix delay on X client startup when ObscureKeystrokeTiming\n is enabled. bz#3820\n\n * sshd(8): increase the maximum size of the supported configuration\n from 256KB to 4MB, which ought to be enough for anybody. Fail\n early and visibly when this limit is breached. bz3808\n\n * sftp(1): during sftp uploads, avoid a condition where a failed\n write could be ignored if a subsequent write succeeded. This is\n unlikely but technically possible because sftp servers are\n allowed to reorder requests.\n\n * sshd(8): avoid a race condition when the sshd-auth process exits\n that could cause a spurious error message to be logged.\n\n * sshd(8): log at level INFO when PerSourcePenalties actually\n blocks access to a source address range. Previously this was\n logged at level VERBOSE, which hid enforcement actions under\n default config settings.\n\n * sshd(8): GssStrictAcceptor was missing from sshd -T output; fix\n\n * sshd(8): Make the MaxStartups and PerSourceNetBlockSize options\n first-match-wins as advertised. bz3859\n\n * ssh(1): fix an incorrect return value check in the local forward\n cancellation path that would cause failed cancellations not to be\n logged.\n\n * sshd(8): make "Match !final" not trigger a second parsing pass\n of ssh_config (unless hostname canonicalisation or a separate\n "Match final" does). bz3843\n\n * ssh(1): better debug diagnostics when loading keys. Will now list\n key fingerprint and algorithm (not just algorithm number) as well\n as making it explicit which keys didn't load.\n\n * All: fix a number of memory leaks found by LeakSanitizer,\n Coverity and manual inspection.\n\n * sshd(8): Output the current name for PermitRootLogin's\n "prohibit-password" in sshd -T instead of its deprecated alias\n "without-password". bz#3788\n\n * ssh(1): make writing known_hosts lines more atomic by writing\n the entire line in one operation and using unbuffered stdio.\n\n Usually writes to this file are serialised on the "Are you sure\n you want to continue connecting?" prompt, but if host key\n checking is disabled and connections were being made with high\n concurrency then interleaved writes might have been possible.\n\nPortability\n-----------\n\n * sshd(8): check the username didn't change during the PAM\n transactions.\n\n PAM modules can change the user during their execution, but\n this is not supported by sshd(8). If such a case was incorrectly\n configured by the system administrator, then sshd(8) could end up\n using a different username to the one authorised by PAM.\n\n * sshd(8): don't log audit messages with UNKNOWN hostname to avoid\n slow DNS lookups in the audit subsystem.\n\n * All: when making a copy of struct passwd, ensure struct fields are\n non-NULL. Android libc can return NULL pw_gecos, for example.\n\n * All: Remove status bits from OpenSSL >=3 version check.\n\n * sshd(8), ssh(1): Use SSH_TUN_COMPAT_AF on FreeBSD. Otherwise tun\n forwarding from other OSes fails as soon as the first IPv6 message\n is sent by the other side (which is usually a Router Solicitation\n ICMPv6 message which is sent as soon as the interface is up).\n\n * ssh(1), ssh-agent(8): check for nlist function presence before\n attempting to use it instead of relying on the presence of the\n nlist.h header. Mac OS X, for example, has the header but not\n the function in the 64bit libraries.\n\n * All: fill in missing system header files.\n\n Create replacement header files inside openbsd-compat for common\n headers that are missing on a given platform. Usually these are\n just empty, but in some cases they'll include the equivalent file.\n This avoids having to wrap those includes in '#ifdef HAVE_FOO_H'\n and reduces the diff between Portable OpenSSH and OpenBSD.\n\n * sshd(8): handle futex_time64 properly in seccomp sandbox\n Previously we only allowed __NR_futex, but some 32-bit systems\n apparently support __NR_futex_time64. We had support for this\n in the sandbox, but because of a macro error only __NR_futex was\n allowlisted.\n\n * Add contrib/gnome-ssh-askpass4 for GNOME 40+ using the GCR API.\n\n * sshd(8): let ga_init() fail gracefully if getgrouplist does.\n Apparently getgrouplist() can fail on OSX when passed a\n non-existent group name. Other platforms seem to return a group\n list consisting of the numeric gid passed to the function. bz3848\n\n * ssh-agent(1): exit 0 from SIGTERM under systemd socket-activation,\n preventing a graceful shutdown of an agent via systemd from\n incorrectly marking the service as "failed".\n\n * build: wrap some autoconf macros in AC_CACHE_CHECK.\n\n This allows skipping/overriding the OSSH_CHECK_CFLAG_COMPILE and\n OSSH_CHECK_CFLAG_LINK macros used to discover supported compiler\n or linker flags. E.g.\n\n $ ./configure ossh_cv_cflag__fzero_call_used_regs_used=no\n \n checking if cc supports compile flag -fzero-call-used-regs=used\n and linking succeeds... (cached) no\n\nChecksums:\n==========\n\nSHA1 (openssh-10.1.tar.gz) = 8eef44a945a9a9a5a99213ab0d57e35b7ba60e75\nSHA256 (openssh-10.1.tar.gz) = j9ymvhvdGMeAvh1oTI2YmOAwv7Ao70gbGPyPedgQsBU=\n\nSHA1 (openssh-10.1p1.tar.gz) = 7fd17b99d1beffb47cd380d64079e920bb0bd91f\nSHA256 (openssh-10.1p1.tar.gz) = ufx6K4JXlGem8vQ+SoHI4d/aYU3bT5slWq/XAgu/B1g=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 10.0 was released on 2025-04-09. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n--------------------------------\n\n * This release removes support for the weak DSA signature\n algorithm, completing the deprecation process that began in\n 2015 (when DSA was disabled by default) and repeatedly warned\n over the last 12 months.\n\n * scp(1), sftp(1): pass "ControlMaster no" to ssh when invoked by\n scp & sftp. This disables implicit session creation by these\n tools when ControlMaster was set to yes/auto by configuration,\n which some users found surprising. This change will not prevent\n scp/sftp from using an existing multiplexing session if one had\n already been created. GHPR557\n\n * This release has the version number 10.0 and announces itself\n as "SSH-2.0-OpenSSH_10.0". Software that naively matches\n versions using patterns like "OpenSSH_1*" may be confused by\n this.\n\n * sshd(8): this release removes the code responsible for the\n user authentication phase of the protocol from the per-\n connection sshd-session binary to a new sshd-auth binary.\n Splitting this code into a separate binary ensures that the\n crucial pre-authentication attack surface has an entirely\n disjoint address space from the code used for the rest of the\n connection. It also yields a small runtime memory saving as the\n authentication code will be unloaded after the authentication\n phase completes. This change should be largely invisible to\n users, though some log messages may now come from "sshd-auth"\n instead of "sshd-session". Downstream distributors of OpenSSH\n will need to package the sshd-auth binary.\n\n * sshd(8): this release disables finite field (a.k.a modp)\n Diffie-Hellman key exchange in sshd by default. Specifically,\n this removes the "diffie-hellman-group*" and\n "diffie-hellman-group-exchange-*" methods from the default\n KEXAlgorithms list. The client is unchanged and continues to\n support these methods by default. Finite field Diffie Hellman\n is slow and computationally expensive for the same security\n level as Elliptic Curve DH or PQ key agreement while offering\n no redeeming advantages. ECDH has been specified for the SSH\n protocol for 15 years and some form of ECDH has been the default\n key exchange in OpenSSH for the last 14 years.\n\n * sshd(8): this release removes the implicit fallback to compiled-\n in groups for Diffie-Hellman Group Exchange KEX when the moduli\n file exists but does not contain moduli within the client-\n requested range. The fallback behaviour remains for the case\n where the moduli file does not exist at all. This allows\n administrators more explicit control over which DH groups will\n be selected, but can lead to connection failures if the moduli\n file is edited incorrectly. bz#2793\n\nChanges since OpenSSH 9.9\n=========================\n\nThis release contains a minor security fix as well as a number of\nfeature improvements and bugfixes.\n\nSecurity\n========\n\n* sshd(8): fix the DisableForwarding directive, which was failing\n to disable X11 forwarding and agent forwarding as documented.\n X11 forwarding is disabled by default in the server and agent\n forwarding is off by default in the client.\n\nNew features\n------------\n\n * ssh(1): the hybrid post-quantum algorithm mlkem768x25519-sha256\n is now used by default for key agreement. This algorithm is\n considered to be safe against attack by quantum computers,\n is guaranteed to be no less strong than the popular\n curve25519-sha256 algorithm, has been standardised by NIST\n and is considerably faster than the previous default.\n\n * ssh(1): prefer AES-GCM to AES-CTR mode when selecting a cipher\n for the connection. The default cipher preference list is now\n Chacha20/Poly1305, AES-GCM (128/256) followed by AES-CTR\n (128/192/256).\n\n * ssh(1): add %-token and environment variable expansion to the\n ssh_config SetEnv directive.\n\n * ssh(1): allow %-token and environment variable expansion in\n the ssh_config User directive, with the exception of %r and %C\n which would be self-referential. bz#3477\n\n * ssh(1), sshd(8): add "Match version" support to ssh_config and\n sshd_config. Allows matching on the local version of OpenSSH,\n e.g. "Match version OpenSSH_10.*".\n\n * ssh(1): add support for "Match sessiontype" to ssh_config.\n Allows matching on the type of session initially requested,\n either "shell" for interactive sessions, "exec" for command\n execution sessions, "subsystem" for subsystem requests, such as\n sftp, or "none" for transport/forwarding-only sessions.\n\n * ssh(1): add support for "Match command ..." support to\n ssh_config, allowing matching on the remote command as specified\n on the command-line.\n\n * ssh(1): allow 'Match tagged ""' and 'Match command ""' to match\n empty tag and command values respectively.\n\n * sshd(8): allow glob(3) patterns to be used in sshd_config\n AuthorizedKeysFile and AuthorizedPrincipalsFile directives.\n bz2755\n\n * sshd(1): support the VersionAddendum in the client, mirroring\n the option of the same name in the server; bz2745\n\n * ssh-agent(1): the agent will now delete all loaded keys when\n signaled with SIGUSR1. This allows deletion of keys without\n having access to $SSH_AUTH_SOCK.\n\n * Portable OpenSSH, ssh-agent(1): support systemd-style socket\n activation in ssh-agent using the LISTEN_PID/LISTEN_FDS\n mechanism. Activated when these environment variables are set,\n the agent is started with the -d or -D option and no socket path\n is set. GHPR502\n\n * ssh-keygen(1): support FIDO tokens that return no attestation\n data, e.g. recent WinHello. GHPR542\n\n * ssh-agent(1): add a "-Owebsafe-allow=..." option to allow the\n default FIDO application ID allow-list to be overridden.\n\n * Add a work-in-progress tool to verify FIDO attestation blobs\n that ssh-keygen can optionally write when enrolling FIDO keys.\n This tool is available under regress/misc/ssh-verify-attestation\n for experimentation but is not installed by "make install".\n\n * ssh-keygen(1): allow "-" as output file for moduli screening.\n GHPR393\n\nBugfixes\n--------\n\n * sshd(8): remove assumption that the sshd_config and any configs\n it includes can fit in a (possibly enlarged) socket buffer.\n Previously it was possible to create a sufficiently large\n configuration that could cause sshd to fail to accept any\n connection. sshd(8) will now actively manage sending its config\n to the sshd-session sub-process.\n\n * ssh(1): don't start the ObscureKeystrokeTiming mitigations if\n there has been traffic on a X11 forwarding channel recently.\n Should fix X11 forwarding performance problems when this setting\n is enabled. bz3655\n\n * ssh(1): prohibit the comma character in hostnames accepted, but\n allow an underscore as the first character in a hostname.\n\n * sftp(1): set high-water when resuming a "put". Prevents bogus\n "server reordered acks" debug message.\n\n * ssh(1), sshd(8): fix regression in openssh-9.8, which would fail\n to accept "Match criteria=argument" as well as the documented\n "Match criteria argument" syntax in ssh_config and sshd_config.\n bz3739\n\n * sftp(1), ssh(1): fix a number possible NULL dereference bugs,\n including Coverity CIDs 405019 and 477813.\n\n * sshd(8): fix PerSourcePenalty incorrectly using "crash" penalty\n when LoginGraceTime was exceeded. bz3797\n\n * sshd(8): fix "Match invalid-user" from incorrectly being\n activated in initial configuration pass when no other predicates\n were present on the match line\n\n * sshd(8): fix debug logging of user specific delay. GHPR#552\n\n * sshd(8): improve debug logging across sub-process boundaries.\n Previously some log messages were lost early in the sshd-auth and\n sshd-session processes' life.\n\n * ssh(1): require control-escape character sequences passed via\n the '-e ^x' command-line to be exactly two characters long. Avoids\n one byte out-of-bounds read if ssh is invoked as "ssh -e^ ..."\n GHPR368\n\n * ssh(1), sshd(8): prevent integer overflow in x11 port handling.\n These are theoretically possible if the admin misconfigured\n X11DisplayOffset or the user misconfigures their own $DISPLAY,\n but don't happen in normal operation. bz#3730\n\n * ssh-keygen(1): don't mess up ssh-keygen -l output when the file\n contains CR characters; GHPR236 bz3385.\n\n * sshd(8): add rate limits to logging of connections dropped by\n PerSourcePenalties. Previously these could be noisy in logs.\n\n * ssh(1): fix argument of "Compression" directive in ssh -G config\n dump, which regressed in openssh-9.8.\n\n * sshd(8): fix a corner-case triggered by UpdateHostKeys when sshd\n refuses to accept the signature returned by an agent holding host\n keys during the hostkey rotation sub-protocol. This situation\n could occur in situations where a PKCS#11 smartcard that lacked\n support for particular signature algorithms was used to store\n host keys.\n\n * ssh-keygen(1): when using RSA keys to sign messages with\n "ssh-keygen -Y", select the signature algorithm based on the\n requested hash algorithm ("-Ohashalg=xxx"). This allows using\n something other than the default of rsa-sha2-512, which may not\n be supported on all signing backends, e.g. some smartcards only\n support SHA256.\n\n * ssh(1), sshd(8), ssh-keyscan(1): fix ML-KEM768x25519 KEX on\n big-endian systems.\n\n * Many regression and interop test improvements.\n\nPortability\n-----------\n\n * All: add support for AWS-LC (AWS libcrypto). bz3784\n\n * sshd(8): add wtmpdb support as a Y2038 safe wtmp replacement.\n\n * sshd(8): add support for locking sshd into memory, enabled with\n the --with-linux-memlock-onfault configure flag.\n\n * Add support for building a standalone sk-libfido2 library,\n enabled by --with-security-key-standalone\n\n * ssh(1), sshd(8), ssh-keyscan(1): include __builtin_popcount\n replacement function. for compilers that lack it.\n\n * All: Check for and replace le32toh, le64toh, htole64 separately.\n It appears that at least some versions of endian.h in glibc do\n not have the latter two. bz#3794\n\n * Remove ancient RHL 6.x config in RPM spec.\n\nChecksums:\n==========\n\n - SHA1 (openssh-10.0.tar.gz) = 933f4fded0497ef6a588381257276e156a70f9c3\n - SHA256 (openssh-10.0.tar.gz) = oaJ+cXLCVoCZAz9W5W1vF7ko4GJW6iq7JmblrPUA34Q=\n\n - SHA1 (openssh-10.0p2.tar.gz) = ac4205e827aea383bf316a33a0e2d5b66b85fcf8\n - SHA256 (openssh-10.0p2.tar.gz) = AhoucJoO30JQsSVr1anlAEEakN3avqgw7VnO+Q652Fw=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.9p2 was released on 2025-02-18. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.9p1\n===========================\n\nThis release fixes two security bugs.\n\nSecurity\n========\n\n* Fix CVE-2025-26465 - ssh(1) in OpenSSH versions 6.8p1 to 9.9p1\n (inclusive) contained a logic error that allowed an on-path\n attacker (a.k.a MITM) to impersonate any server when the\n VerifyHostKeyDNS option is enabled. This option is off by default.\n\n* Fix CVE-2025-26466 - sshd(8) in OpenSSH versions 9.5p1 to 9.9p1\n (inclusive) is vulnerable to a memory/CPU denial-of-service related\n to the handling of SSH2_MSG_PING packets. This condition may be\n mitigated using the existing PerSourcePenalties feature.\n\nBoth vulnerabilities were discovered and demonstrated to be exploitable\nby the Qualys Security Advisory team. We thank them for their detailed\nreview of OpenSSH.\n\nFor OpenBSD, fixes to these problems are available as errata; refer\nto https://www.openbsd.org/errata.html\n \nBugfixes\n========\n\n * ssh(1), sshd(8): fix regression in Match directive that caused\n failures when predicates and their arguments were separated by '='\n characters instead of whitespace (bz3739).\n\n * sshd(8): fix the "Match invalid-user" predicate, which was matching\n incorrectly in the initial pass of config evaluation.\n\n * ssh(1), sshd(8), ssh-keyscan(1): fix mlkem768x25519-sha256 key\n exchange on big-endian systems.\n\n * Fix a number of build problems on particular operating systems /\n configurations.\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.9p2.tar.gz) = edefe960645780dee78059c444d4261667ad3056\n - SHA256 (openssh-9.9p2.tar.gz) = karbYD4IzChe3fll4RmdAlhfqU2ZTWyuW0Hhch4hVnM=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.9 was released on 2024-09-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nOpenSSH plans to remove support for the DSA signature algorithm in\nearly 2025. This release disables DSA by default at compile time.\n\nDSA, as specified in the SSHv2 protocol, is inherently weak - being\nlimited to a 160 bit private key and use of the SHA1 digest. Its\nestimated security level is only 80 bits symmetric equivalent.\n\nOpenSSH has disabled DSA keys by default since 2015 but has retained\nrun-time optional support for them. DSA was the only mandatory-to-\nimplement algorithm in the SSHv2 RFCs, mostly because alternative\nalgorithms were encumbered by patents when the SSHv2 protocol was\nspecified.\n\nThis has not been the case for decades at this point and better\nalgorithms are well supported by all actively-maintained SSH\nimplementations. We do not consider the costs of maintaining DSA\nin OpenSSH to be justified and hope that removing it from OpenSSH\ncan accelerate its wider deprecation in supporting cryptography\nlibraries.\n\nCurrently DSA is disabled at compile time. The final step of\nremoving DSA support entirely is planned for the first OpenSSH\nrelease of 2025.\n\nDSA support may be re-enabled on OpenBSD by setting "DSAKEY=yes"\nin Makefile.inc. To enable DSA support in portable OpenSSH, pass\nthe "--enable-dsa-keys" option to configure.\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh(1): remove support for pre-authentication compression.\n OpenSSH has only supported post-authentication compression in\n the server for some years. Compression before authentication\n significantly increases the attack surface of SSH servers and risks\n creating oracles that reveal information about information sent\n during authentication.\n\n * ssh(1), sshd(8): processing of the arguments to the "Match"\n configuration directive now follows more shell-like rules for\n quoted strings, including allowing nested quotes and \\-escaped\n characters. If configurations contained workarounds for the\n previous simplistic quote handling then they may need to be\n adjusted. If this is the case, it's most likely to be in the\n arguments to a "Match exec" confition. In this case, moving the\n command to be evaluated from the Match line to an external shell\n script is easiest way to preserve compatibility with both the old\n and new versions.\n\nChanges since OpenSSH 9.8\n=========================\n\nThis release contains a number of new features and bugfixes.\n\nNew features\n------------\n\n * ssh(1), sshd(8): add support for a new hybrid post-quantum key\n exchange based on the FIPS 203 Module-Lattice Key Enapsulation\n mechanism (ML-KEM) combined with X25519 ECDH as described by\n https://datatracker.ietf.org/doc/html/draft-kampanakis-curdle-ssh-pq-ke-03\n This algorithm "mlkem768x25519-sha256" is available by default.\n\n * ssh(1): the ssh_config "Include" directive can now expand\n environment as well as the same set of %-tokens "Match Exec"\n supports.\n\n * sshd(8): add a sshd_config "RefuseConnection" option that, if set\n will terminate the connection at the first authentication request.\n\n * sshd(8): add a "refuseconnection" penalty class to sshd_config\n PerSourcePenalties that is applied when a connection is dropped by\n the new RefuseConnection keyword.\n\n * sshd(8): add a "Match invalid-user" predicate to sshd_config Match\n options that matches when the target username is not valid on the\n server.\n\n * ssh(1), sshd(8): update the Streamlined NTRUPrime code to a\n substantially faster implementation.\n\n * ssh(1), sshd(8): the hybrid Streamlined NTRUPrime/X25519 key\n exchange algorithm now has an IANA-assigned name in addition to\n the "@openssh.com" vendor extension name. This algorithm is now\n also available under this name "sntrup761x25519-sha512"\n\n * ssh(1), sshd(8), ssh-agent(1): prevent private keys from being\n included in core dump files for most of their lifespans. This is\n in addition to pre-existing controls in ssh-agent(1) and sshd(8)\n that prevented coredumps. This feature is supported on OpenBSD,\n Linux and FreeBSD.\n\n * All: convert key handling to use the libcrypto EVP_PKEY API, with\n the exception of DSA.\n\n * sshd(8): add a random amount of jitter (up to 4 seconds) to the\n grace login time to make its expiry unpredictable.\n\nBugfixes\n--------\n\n * sshd(8): relax absolute path requirement back to what it was prior\n to OpenSSH 9.8, which incorrectly required that sshd was started\n with an absolute path in inetd mode. bz3717\n\n * sshd(8): fix regression introduced in openssh-9.8 that swapped the\n order of source and destination addresses in some sshd log messages.\n\n * sshd(8): do not apply authorized_keys options when signature\n verification fails. Prevents more restrictive key options being\n incorrectly applied to subsequent keys in authorized_keys. bz3733\n\n * ssh-keygen(1): include pathname in some of ssh-keygen's passphrase\n prompts. Helps the user know what's going on when ssh-keygen is\n invoked via other tools. Requested in GHPR503\n\n * ssh(1), ssh-add(1): make parsing user@host consistently look for\n the last '@' in the string rather than the first. This makes it\n possible to more consistently use usernames that contain '@'\n characters.\n\n * ssh(1), sshd(8): be more strict in parsing key type names. Only\n allow short names (e.g "rsa") in user-interface code and require\n full SSH protocol names (e.g. "ssh-rsa") everywhere else. bz3725\n\n * regress: many performance and correctness improvements to the\n re-keying regression test.\n\n * ssh-keygen(1): clarify that ed25519 is the default key type\n generated and clarify that rsa-sha2-512 is the default signature\n scheme when RSA is in use. GHPR505\n\n * sshd(8): fix minor memory leak in Subsystem option parsing; GHPR515\n\n * All: additional hardening and consistency checks for the sshbuf\n code.\n\n * sshd(8): reduce default logingrace penalty to ensure that a single\n forgotten login that times out will be below the penalty threshold.\n\n * ssh(1): fix proxy multiplexing (-O proxy) bug. If a mux started with\n ControlPersist then later has a forwarding added using mux proxy\n connection and the forwarding was used, then when the mux proxy\n session terminated, the mux master process would issue a bad message\n that terminated the connection.\n\nPortability\n-----------\n\n * sync contrib/ssh-copy-id to the latest upstream version.\n\n * regress: improve portablility for some awk(1) usage (e.g. Solaris)\n\n * In the contrib/redhat RPM spec file, without_openssl was previously\n incorrectly enabled unconditionally.\n\n * sshd(8) restore audit call before exit that regressed in openssh-9.8\n Fixes an issue where the SSH_CONNECTION_ABANDON event was not\n recorded.\n\n * sshd(8): add support for class-imposed login restrictions on FreeBSD.\n Allowing auth_hostok(3) and auth_timeok(3) to control logins.\n\n * Build fixes for Musl libc.\n\n * Fix detection of setres*id on GNU/Hurd\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.9.tar.gz) = 080acf6ff0b862e8faa3baa3920a079536d28e85\n - SHA256 (openssh-9.9.tar.gz) = h1xwa7CVcJfN7I9MgxxPBpUELzo+tnmLy+6slYHTUtw=\n\n - SHA1 (openssh-9.9p1.tar.gz) = 5ded7eb0add0b02b5d1a1c4bf5cb2c89d2117b53\n - SHA256 (openssh-9.9p1.tar.gz) = s0P7zb/4fxWxmG5uFdbU/Jp9NgZr5rf7UHCHuo+WbAI=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.8 was released on 2024-07-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nSecurity\n========\n\nThis release contains fixes for two security problems, one critical\nand one minor.\n\n1) Race condition in sshd(8)\n\nA critical vulnerability in sshd(8) was present in Portable OpenSSH\nversions between 8.5p1 and 9.7p1 (inclusive) that may allow arbitrary\ncode execution with root privileges.\n\nSuccessful exploitation has been demonstrated on 32-bit Linux/glibc\nsystems with ASLR. Under lab conditions, the attack requires on\naverage 6-8 hours of continuous connections up to the maximum the\nserver will accept. Exploitation on 64-bit systems is believed to be\npossible but has not been demonstrated at this time. It's likely that\nthese attacks will be improved upon.\n\nExploitation on non-glibc systems is conceivable but has not been\nexamined. Systems that lack ASLR or users of downstream Linux\ndistributions that have modified OpenSSH to disable per-connection\nASLR re-randomisation (yes - this is a thing, no - we don't\nunderstand why) may potentially have an easier path to exploitation.\nOpenBSD is not vulnerable.\n\nWe thank the Qualys Security Advisory Team for discovering, reporting\nand demonstrating exploitability of this problem, and for providing\ndetailed feedback on additional mitigation measures.\n\n2) Logic error in ssh(1) ObscureKeystrokeTiming\n\nIn OpenSSH version 9.5 through 9.7 (inclusive), when connected to an\nOpenSSH server version 9.5 or later, a logic error in the ssh(1)\nObscureKeystrokeTiming feature (on by default) rendered this feature\nineffective - a passive observer could still detect which network\npackets contained real keystrokes when the countermeasure was active\nbecause both fake and real keystroke packets were being sent\nunconditionally.\n\nThis bug was found by Philippos Giavridis and also independently by\nJacky Wei En Kung, Daniel Hugenroth and Alastair Beresford of the\nUniversity of Cambridge Computer Lab.\n\nWorse, the unconditional sending of both fake and real keystroke\npackets broke another long-standing timing attack mitigation. Since\nOpenSSH 2.9.9 sshd(8) has sent fake keystroke echo packets for\ntraffic received on TTYs in echo-off mode, such as when entering a\npassword into su(8) or sudo(8). This bug rendered these fake\nkeystroke echoes ineffective and could allow a passive observer of\na SSH session to once again detect when echo was off and obtain\nfairly limited timing information about keystrokes in this situation\n(20ms granularity by default).\n\nThis additional implication of the bug was identified by Jacky Wei\nEn Kung, Daniel Hugenroth and Alastair Beresford and we thank them\nfor their detailed analysis.\n\nThis bug does not affect connections when ObscureKeystrokeTiming\nwas disabled or sessions where no TTY was requested.\n\nFuture deprecation notice\n=========================\n\nOpenSSH plans to remove support for the DSA signature algorithm in\nearly 2025. This release disables DSA by default at compile time.\n\nDSA, as specified in the SSHv2 protocol, is inherently weak - being\nlimited to a 160 bit private key and use of the SHA1 digest. Its\nestimated security level is only 80 bits symmetric equivalent.\n\nOpenSSH has disabled DSA keys by default since 2015 but has retained\nrun-time optional support for them. DSA was the only mandatory-to-\nimplement algorithm in the SSHv2 RFCs, mostly because alternative\nalgorithms were encumbered by patents when the SSHv2 protocol was\nspecified.\n\nThis has not been the case for decades at this point and better\nalgorithms are well supported by all actively-maintained SSH\nimplementations. We do not consider the costs of maintaining DSA\nin OpenSSH to be justified and hope that removing it from OpenSSH\ncan accelerate its wider deprecation in supporting cryptography\nlibraries.\n\nThis release, and its deactivation of DSA by default at compile-time,\nmarks the second step in our timeline to finally deprecate DSA. The\nfinal step of removing DSA support entirely is planned for the first\nOpenSSH release of 2025.\n\nDSA support may be re-enabled in OpenBSD by setting "DSAKEY=yes"\nin Makefile.inc. To enable DSA support in portable OpenSSH, pass\nthe "--enable-dsa-keys" option to configure.\n\nPotentially-incompatible changes\n--------------------------------\n\n * all: as mentioned above, the DSA signature algorithm is now\n disabled at compile time.\n\n * sshd(8): the server will now block client addresses that\n repeatedly fail authentication, repeatedly connect without ever\n completing authentication or that crash the server. See the\n discussion of PerSourcePenalties below for more information.\n Operators of servers that accept connections from many users, or\n servers that accept connections from addresses behind NAT or\n proxies may need to consider these settings.\n\n * sshd(8): the server has been split into a listener binary, sshd(8),\n and a per-session binary "sshd-session". This allows for a much\n smaller listener binary, as it no longer needs to support the SSH\n protocol. As part of this work, support for disabling privilege\n separation (which previously required code changes to disable) and\n disabling re-execution of sshd(8) has been removed. Further\n separation of sshd-session into additional, minimal binaries is\n planned for the future.\n\n * sshd(8): several log messages have changed. In particular, some\n log messages will be tagged with as originating from a process\n named "sshd-session" rather than "sshd".\n\n * ssh-keyscan(1): this tool previously emitted comment lines\n containing the hostname and SSH protocol banner to standard error.\n This release now emits them to standard output, but adds a new\n "-q" flag to silence them altogether.\n\n * sshd(8): (portable OpenSSH only) sshd will no longer use argv[0]\n as the PAM service name. A new "PAMServiceName" sshd_config(5)\n directive allows selecting the service name at runtime. This\n defaults to "sshd". bz2101\n\n * (portable OpenSSH only) Automatically-generated files, such as\n configure, config.h.in, etc will now be checked in to the portable\n OpenSSH git release branch (e.g. V_9_8). This should ensure that\n the contents of the signed release branch exactly match the\n contents of the signed release tarball.\n\nChanges since OpenSSH 9.7\n=========================\n\nThis release contains mostly bugfixes.\n\nNew features\n------------\n\n * sshd(8): as described above, sshd(8) will now penalise client\n addresses that, for various reasons, do not successfully complete\n authentication. This feature is controlled by a new sshd_config(5)\n PerSourcePenalties option and is on by default.\n\n sshd(8) will now identify situations where the session did not\n authenticate as expected. These conditions include when the client\n repeatedly attempted authentication unsuccessfully (possibly\n indicating an attack against one or more accounts, e.g. password\n guessing), or when client behaviour caused sshd to crash (possibly\n indicating attempts to exploit bugs in sshd).\n\n When such a condition is observed, sshd will record a penalty of\n some duration (e.g. 30 seconds) against the client's address. If\n this time is above a minimum configurable threshold, then all\n connections from the client address will be refused (along with any\n others in the same PerSourceNetBlockSize CIDR range) until the\n penalty expire.\n\n Repeated offenses by the same client address will accrue greater\n penalties, up to a configurable maximum. Address ranges may be\n fully exempted from penalties, e.g. to guarantee access from a set\n of trusted management addresses, using the new sshd_config(5)\n PerSourcePenaltyExemptList option.\n\n We hope these options will make it significantly more difficult for\n attackers to find accounts with weak/guessable passwords or exploit\n bugs in sshd(8) itself. This option is enabled by default.\n\n * ssh(8): allow the HostkeyAlgorithms directive to disable the\n implicit fallback from certificate host key to plain host keys.\n\nBugfixes\n--------\n\n * misc: fix a number of inaccuracies in the PROTOCOL.*\n documentation files. GHPR430 GHPR487\n\n * all: switch to strtonum(3) for more robust integer parsing in most\n places.\n\n * ssh(1), sshd(8): correctly restore sigprocmask around ppoll()\n\n * ssh-keysign(8): stricter validation of messaging socket fd GHPR492\n\n * sftp(1): flush stdout after writing "sftp>" prompt when not using\n editline. GHPR480\n\n * sftp-server(8): fix home-directory extension implementation, it\n previously always returned the current user's home directory\n contrary to the spec. GHPR477\n\n * ssh-keyscan(1): do not close stdin to prevent error messages when\n stdin is read multiple times. E.g.\n echo localhost | ssh-keyscan -f - -f -\n\n * regression tests: fix rekey test that was testing the same KEX\n algorithm repeatedly instead of testing all of them. bz3692\n\n * ssh_config(5), sshd_config(5): clarify the KEXAlgorithms directive\n documentation, especially around what is supported vs available.\n bz3701.\n\nPortability\n-----------\n\n * sshd(8): expose SSH_AUTH_INFO_0 always to PAM auth modules\n unconditionally. The previous behaviour was to expose it only when\n particular authentication methods were in use.\n\n * build: fix OpenSSL ED25519 support detection. An incorrect function\n signature in configure.ac previously prevented enabling the recently\n added support for ED25519 private keys in PEM PKCS8 format.\n\n * ssh(1), ssh-agent(8): allow the presence of the WAYLAND_DISPLAY\n environment variable to enable SSH_ASKPASS, similarly to the X11\n DISPLAY environment variable. GHPR479\n\n * build: improve detection of the -fzero-call-used-regs compiler\n flag. bz3673.\n\n * build: relax OpenSSL version check to accept all OpenSSL 3.x\n versions.\n\n * sshd(8): add support for notifying systemd on server listen and\n reload, using a standalone implementation that doesn't depend on\n libsystemd. bz2641\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.8.tar.gz) = bc45cedae7f70b41e9922ef4c9f56e74b9a659b7\n - SHA256 (openssh-9.8.tar.gz) = Dnc69VLWFBFdiaz8wySPlvHjb7wZfh/kblQ8ISuQr1Y=\n\n - SHA1 (openssh-9.8p1.tar.gz) = a0bb501b11349f5c5c33a269351be091dc2c2727\n - SHA256 (openssh-9.8p1.tar.gz) = 3YvQAqN5tdSZ37BQ3R+pr4Ap6ARh9LtsUjxJlz9aOfM=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.7 was released on 2024-03-11. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nOpenSSH plans to remove support for the DSA signature algorithm in\nearly 2025 and compile-time disable it later this year.\n\nDSA, as specified in the SSHv2 protocol, is inherently weak - being\nlimited to a 160 bit private key and use of the SHA1 digest. Its\nestimated security level is only 80 bits symmetric equivalent.\n\nOpenSSH has disabled DSA keys by default since 2015 but has retained\nrun-time optional support for them. DSA was the only mandatory-to-\nimplement algorithm in the SSHv2 RFCs[3], mostly because alternative\nalgorithms were encumbered by patents when the SSHv2 protocol was\nspecified.\n\nThis has not been the case for decades at this point and better\nalgorithms are well supported by all actively-maintained SSH\nimplementations. We do not consider the costs of maintaining DSA in\nOpenSSH to be justified and hope that removing it from OpenSSH can\naccelerate its wider deprecation in supporting cryptography\nlibraries.\n\nThis release makes DSA support in OpenSSH compile-time optional,\ndefaulting to on. We intend the next release to change the default\nto disable DSA at compile time. The first OpenSSH release of 2025\nwill remove DSA support entirely.\n\nChanges since OpenSSH 9.6\n=========================\n\nThis release contains mostly bugfixes.\n\nNew features\n------------\n\n * ssh(1), sshd(8): add a "global" ChannelTimeout type that watches\n all open channels and will close all open channels if there is no\n traffic on any of them for the specified interval. This is in\n addition to the existing per-channel timeouts added recently.\n\n This supports situations like having both session and x11\n forwarding channels open where one may be idle for an extended\n period but the other is actively used. The global timeout could\n close both channels when both have been idle for too long.\n\n * All: make DSA key support compile-time optional, defaulting to on.\n\nBugfixes\n--------\n\n * sshd(8): don't append an unnecessary space to the end of subsystem\n arguments (bz3667)\n\n * ssh(1): fix the multiplexing "channel proxy" mode, broken when\n keystroke timing obfuscation was added. (GHPR#463)\n\n * ssh(1), sshd(8): fix spurious configuration parsing errors when\n options that accept array arguments are overridden (bz3657).\n\n * ssh-agent(1): fix potential spin in signal handler (bz3670)\n\n * Many fixes to manual pages and other documentation, including\n GHPR#462, GHPR#454, GHPR#442 and GHPR#441.\n\n * Greatly improve interop testing against PuTTY.\n\nPortability\n-----------\n\n * Improve the error message when the autoconf OpenSSL header check\n fails (bz#3668)\n\n * Improve detection of broken toolchain -fzero-call-used-regs support\n (bz3645).\n\n * Fix regress/misc/fuzz-harness fuzzers and make them compile without\n warnings when using clang16\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.7.tar.gz) = 163272058edc20a8fde81661734a6684c9b4db11\n - SHA256 (openssh-9.7.tar.gz) = gXDWrF4wN2UWyPjyjvVhpjjKd7D2qI6LyZiIYhbJQVg=\n\n - SHA1 (openssh-9.7p1.tar.gz) = ce8985ea0ea2f16a5917fd982ade0972848373cc\n - SHA256 (openssh-9.7p1.tar.gz) = SQQm92bYKidj/KzY2D6j1weYdQx70q/y5X3FZg93P/0=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.6 was released on 2023-12-18. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.5\n=========================\n\nThis release contains a number of security fixes, some small features\nand bugfixes.\n\nSecurity\n========\n\nThis release contains fixes for a newly-discovered weakness in the\nSSH transport protocol, a logic error relating to constrained PKCS#11\nkeys in ssh-agent(1) and countermeasures for programs that invoke\nssh(1) with user or hostnames containing invalid characters.\n\n * ssh(1), sshd(8): implement protocol extensions to thwart the\n so-called "Terrapin attack" discovered by Fabian Bäumer, Marcus\n Brinkmann and Jörg Schwenk. This attack allows a MITM to effect a\n limited break of the integrity of the early encrypted SSH transport\n protocol by sending extra messages prior to the commencement of\n encryption, and deleting an equal number of consecutive messages\n immediately after encryption starts. A peer SSH client/server\n would not be able to detect that messages were deleted.\n\n While cryptographically novel, the security impact of this attack\n is fortunately very limited as it only allows deletion of\n consecutive messages, and deleting most messages at this stage of\n the protocol prevents user authentication from proceeding and\n results in a stuck connection.\n\n The most serious identified impact is that it lets a MITM to\n delete the SSH2_MSG_EXT_INFO message sent before authentication\n starts, allowing the attacker to disable a subset of the keystroke\n timing obfuscation features introduced in OpenSSH 9.5. There is no\n other discernible impact to session secrecy or session integrity.\n\n OpenSSH 9.6 addresses this protocol weakness through a new "strict\n KEX" protocol extension that will be automatically enabled when\n both the client and server support it. This extension makes\n two changes to the SSH transport protocol to improve the integrity\n of the initial key exchange.\n\n Firstly, it requires endpoints to terminate the connection if any\n unnecessary or unexpected message is received during key exchange\n (including messages that were previously legal but not strictly\n required like SSH2_MSG_DEBUG). This removes most malleability from\n the early protocol.\n\n Secondly, it resets the Message Authentication Code counter at the\n conclusion of each key exchange, preventing previously inserted\n messages from being able to make persistent changes to the\n sequence number across completion of a key exchange. Either of\n these changes should be sufficient to thwart the Terrapin Attack.\n\n More details of these changes are in the PROTOCOL file in the\n OpenSSH source distribition.\n\n * ssh-agent(1): when adding PKCS#11-hosted private keys while\n specifying destination constraints, if the PKCS#11 token returned\n multiple keys then only the first key had the constraints applied.\n Use of regular private keys, FIDO tokens and unconstrained keys\n are unaffected.\n\n * ssh(1): if an invalid user or hostname that contained shell\n metacharacters was passed to ssh(1), and a ProxyCommand,\n LocalCommand directive or "match exec" predicate referenced the\n user or hostname via %u, %h or similar expansion token, then\n an attacker who could supply arbitrary user/hostnames to ssh(1)\n could potentially perform command injection depending on what\n quoting was present in the user-supplied ssh_config(5) directive.\n\n This situation could arise in the case of git submodules, where\n a repository could contain a submodule with shell characters in\n its user/hostname. Git does not ban shell metacharacters in user\n or host names when checking out repositories from untrusted\n sources.\n\n Although we believe it is the user's responsibility to ensure\n validity of arguments passed to ssh(1), especially across a\n security boundary such as the git example above, OpenSSH 9.6 now\n bans most shell metacharacters from user and hostnames supplied\n via the command-line. This countermeasure is not guaranteed to be\n effective in all situations, as it is infeasible for ssh(1) to\n universally filter shell metacharacters potentially relevant to\n user-supplied commands.\n\n User/hostnames provided via ssh_config(5) are not subject to these\n restrictions, allowing configurations that use strange names to\n continue to be used, under the assumption that the user knows what\n they are doing in their own configuration files.\n\nPotentially incompatible changes\n--------------------------------\n\n * ssh(1), sshd(8): the RFC4254 connection/channels protocol provides\n a TCP-like window mechanism that limits the amount of data that\n can be sent without acceptance from the peer. In cases where this\n limit was exceeded by a non-conforming peer SSH implementation,\n ssh(1)/sshd(8) previously discarded the extra data. From OpenSSH\n 9.6, ssh(1)/sshd(8) will now terminate the connection if a peer\n exceeds the window limit by more than a small grace factor. This\n change should have no effect of SSH implementations that follow\n the specification.\n\nNew features\n------------\n\n * ssh(1): add a %j token that expands to the configured ProxyJump\n hostname (or the empty string if this option is not being used)\n that can be used in a number of ssh_config(5) keywords. bz3610\n\n * ssh(1): add ChannelTimeout support to the client, mirroring the\n same option in the server and allowing ssh(1) to terminate\n quiescent channels.\n\n * ssh(1), sshd(8), ssh-add(1), ssh-keygen(1): add support for\n reading ED25519 private keys in PEM PKCS8 format. Previously\n only the OpenSSH private key format was supported.\n\n * ssh(1), sshd(8): introduce a protocol extension to allow\n renegotiation of acceptable signature algorithms for public key\n authentication after the server has learned the username being\n used for authentication. This allows varying sshd_config(5)\n PubkeyAcceptedAlgorithms in a "Match user" block.\n\n * ssh-add(1), ssh-agent(1): add an agent protocol extension to allow\n specifying certificates when loading PKCS#11 keys. This allows the\n use of certificates backed by PKCS#11 private keys in all OpenSSH\n tools that support ssh-agent(1). Previously only ssh(1) supported\n this use-case.\n\nBugfixes\n--------\n\n * ssh(1): when deciding whether to enable the keystroke timing\n obfuscation, enable it only if a channel with a TTY is active.\n\n * ssh(1): switch mainloop from poll(3) to ppoll(3) and mask signals\n before checking flags set in signal handler. Avoids potential\n race condition between signaling ssh to exit and polling. bz3531\n \n * ssh(1): when connecting to a destination with both the\n AddressFamily and CanonicalizeHostname directives in use,\n the AddressFamily directive could be ignored. bz5326\n\n * sftp(1): correct handling of the limits@openssh.com option when\n the server returned an unexpected message.\n\n * A number of fixes to the PuTTY and Dropbear regress/integration\n tests.\n\n * ssh(1): release GSS OIDs only at end of authentication, avoiding\n unnecessary init/cleanup cycles. bz2982\n\n * ssh_config(5): mention "none" is a valid argument to IdentityFile\n in the manual. bz3080\n\n * scp(1): improved debugging for paths from the server rejected for\n not matching the client's glob(3) pattern in old SCP/RCP protocol\n mode.\n\n * ssh-agent(1): refuse signing operations on destination-constrained\n keys if a previous session-bind operation has failed. This may\n prevent a fail-open situation in future if a user uses a mismatched\n ssh(1) client and ssh-agent(1) where the client supports a key type\n that the agent does not support.\n\nPortability\n-----------\n\n * Better identify unsupported and unstable compiler flags, such as\n -fzero-call-used-regs which has been unstable across a several\n clang releases.\n\n * A number of fixes to regression test reliability and log\n collection.\n\n * Update the OpenSSL dependency in the RPM specification.\n\n * sshd(8): for OpenSolaris systems that support privilege limitation\n via the getpflags() interface, prefer using the newer PRIV_XPOLICY\n to PRIV_LIMIT. bz2833\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.6.tar.gz) = a6d4cb69811e879e2f158c2e597fd9f444b26506\n - SHA256 (openssh-9.6.tar.gz) = nejPUhSnG1R1sOmIBi/t+HMNvsRqfN/DJgjwIU2tvqg=\n\n - SHA1 (openssh-9.6p1.tar.gz) = de300d09ec79fdbf37de4e6672cce4161439f2c3\n - SHA256 (openssh-9.6p1.tar.gz) = kQIRwHJVqMWtZUORtA7lmABxDdgRndU2LeCThap6d3w=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.5 was released on 2023-10-04. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.4\n=========================\n\nThis release fixes a number of bugs and adds some small features.\n\nPotentially incompatible changes\n--------------------------------\n\n * ssh-keygen(1): generate Ed25519 keys by default. Ed25519 public keys\n are very convenient due to their small size. Ed25519 keys are\n specified in RFC 8709 and OpenSSH has supported them since version 6.5\n (January 2014).\n \n * sshd(8): the Subsystem directive now accurately preserves quoting of\n subsystem commands and arguments. This may change behaviour for exotic\n configurations, but the most common subsystem configuration\n (sftp-server) is unlikely to be affected.\n\nNew features\n------------\n\n * ssh(1): add keystroke timing obfuscation to the client. This attempts\n to hide inter-keystroke timings by sending interactive traffic at\n fixed intervals (default: every 20ms) when there is only a small\n amount of data being sent. It also sends fake "chaff" keystrokes for\n a random interval after the last real keystroke. These are\n controlled by a new ssh_config ObscureKeystrokeTiming keyword.\n \n * ssh(1), sshd(8): Introduce a transport-level ping facility. This adds\n a pair of SSH transport protocol messages SSH2_MSG_PING/PONG to\n implement a ping capability. These messages use numbers in the "local\n extensions" number space and are advertised using a "ping@openssh.com"\n ext-info message with a string version number of "0".\n \n * sshd(8): allow override of Subsystem directives in sshd Match blocks.\n\nBugfixes\n--------\n\n * scp(1): fix scp in SFTP mode recursive upload and download of\n directories that contain symlinks to other directories. In scp mode,\n the links would be followed, but in SFTP mode they were not. bz3611\n \n * ssh-keygen(1): handle cr+lf (instead of just cr) line endings in\n sshsig signature files.\n \n * ssh(1): interactive mode for ControlPersist sessions if they\n originally requested a tty.\n \n * sshd(8): make PerSourceMaxStartups first-match-wins\n \n * sshd(8): limit artificial login delay to a reasonable maximum (5s)\n and don't delay at all for the "none" authentication mechanism.cw\n bz3602\n \n * sshd(8): Log errors in kex_exchange_identification() with level\n verbose instead of error to reduce preauth log spam. All of those\n get logged with a more generic error message by sshpkt_fatal().\n \n * sshd(8): correct math for ClientAliveInterval that caused the probes\n to be sent less frequently than configured.\n \n * ssh(1): fix regression in OpenSSH 9.4 (mux.c r1.99) that caused\n multiplexed sessions to ignore SIGINT under some circumstances.\n\nPortability\n-----------\n\n * Avoid clang zero-call-used-regs=all bug on Apple compilers, which\n for some reason have version numbers that do not match the upstream\n clang version numbers. bz#3584\n\n * Fix configure test for zlib 1.3 and later/development versions. bz3604\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.5.tar.gz) = 8a0bd3a91fac338d97d91817af58df731f6509a3\n - SHA256 (openssh-9.5.tar.gz) = sVMxeM3d6g65qBMktJIofxmK4Ipg9dblKif0VnhPeO0=\n\n - SHA1 (openssh-9.5p1.tar.gz) = 35c16dcc6e7d0a9465faa241476ef24f76b196cc\n - SHA256 (openssh-9.5p1.tar.gz) = 8Cbnt5un+1QPdRgq+W3IqPHbOV+SK7yfbKYDZyaGCGs=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.4 was released on 2023-08-10. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.3p2\n===========================\n\nThis release fixes a number of bugs and adds some small features.\n\nPotentially incompatible changes\n--------------------------------\n\n * This release removes support for older versions of libcrypto.\n OpenSSH now requires LibreSSL >= 3.1.0 or OpenSSL >= 1.1.1.\n Note that these versions are already deprecated by their upstream\n vendors.\n\n * ssh-agent(1): PKCS#11 modules must now be specified by their full\n paths. Previously dlopen(3) could search for them in system\n library directories.\n\nNew features\n------------\n\n * ssh(1): allow forwarding Unix Domain sockets via ssh -W.\n\n * ssh(1): add support for configuration tags to ssh(1).\n This adds a ssh_config(5) "Tag" directive and corresponding\n "Match tag" predicate that may be used to select blocks of\n configuration similar to the pf.conf(5) keywords of the same\n name.\n\n * ssh(1): add a "match localnetwork" predicate. This allows matching\n on the addresses of available network interfaces and may be used to\n vary the effective client configuration based on network location.\n\n * ssh(1), sshd(8), ssh-keygen(1): infrastructure support for KRL\n extensions. This defines wire formats for optional KRL extensions\n and implements parsing of the new submessages. No actual extensions\n are supported at this point.\n\n * sshd(8): AuthorizedPrincipalsCommand and AuthorizedKeysCommand now\n accept two additional %-expansion sequences: %D which expands to\n the routing domain of the connected session and %C which expands\n to the addresses and port numbers for the source and destination\n of the connection.\n\n * ssh-keygen(1): increase the default work factor (rounds) for the\n bcrypt KDF used to derive symmetric encryption keys for passphrase\n protected key files by 50%.\n\nBugfixes\n--------\n\n * ssh-agent(1): improve isolation between loaded PKCS#11 modules\n by running separate ssh-pkcs11-helpers for each loaded provider.\n\n * ssh(1): make -f (fork after authentication) work correctly with\n multiplexed connections, including ControlPersist. bz3589 bz3589\n\n * ssh(1): make ConnectTimeout apply to multiplexing sockets and not\n just to network connections.\n\n * ssh-agent(1), ssh(1): improve defences against invalid PKCS#11\n modules being loaded by checking that the requested module\n contains the required symbol before loading it.\n\n * sshd(8): fix AuthorizedPrincipalsCommand when AuthorizedKeysCommand\n appears before it in sshd_config. Since OpenSSH 8.7 the\n AuthorizedPrincipalsCommand directive was incorrectly ignored in\n this situation. bz3574\n\n * sshd(8), ssh(1), ssh-keygen(1): remove vestigal support for KRL\n signatures When the KRL format was originally defined, it included\n support for signing of KRL objects. However, the code to sign KRLs\n and verify KRL signatures was never completed in OpenSSH. This\n release removes the partially-implemented code to verify KRLs.\n All OpenSSH tools now ignore KRL_SECTION_SIGNATURE sections in\n KRL files.\n\n * All: fix a number of memory leaks and unreachable/harmless integer\n overflows.\n\n * ssh-agent(1), ssh(1): don't truncate strings logged from PKCS#11\n modules; GHPR406\n\n * sshd(8), ssh(1): better validate CASignatureAlgorithms in\n ssh_config and sshd_config. Previously this directive would accept\n certificate algorithm names, but these were unusable in practice as\n OpenSSH does not support CA chains. bz3577\n\n * ssh(1): make `ssh -Q CASignatureAlgorithms` only list signature\n algorithms that are valid for CA signing. Previous behaviour was\n to list all signing algorithms, including certificate algorithms.\n\n * ssh-keyscan(1): gracefully handle systems where rlimits or the\n maximum number of open files is larger than INT_MAX; bz3581\n\n * ssh-keygen(1): fix "no comment" not showing on when running\n `ssh-keygen -l` on multiple keys where one has a comment and other\n following keys do not. bz3580\n\n * scp(1), sftp(1): adjust ftruncate() logic to handle servers that\n reorder requests. Previously, if the server reordered requests then\n the resultant file would be erroneously truncated.\n\n * ssh(1): don't incorrectly disable hostname canonicalization when\n CanonicalizeHostname=yes and ProxyJump was explicitly set to\n "none". bz3567\n\n * scp(1): when copying local->remote, check that the source file\n exists before opening an SFTP connection to the server. Based on\n GHPR#370\n\nPortability\n-----------\n\n * All: a number of build fixes for various platforms and\n configuration combinations.\n\n * sshd(8): provide a replacement for the SELinux matchpathcon()\n function, which is deprecated.\n\n * All: relax libcrypto version checks for OpenSSL >=3. Beyond\n OpenSSL 3.0, the ABI compatibility guarantees are wider (only\n the library major must match instead of major and minor in\n earlier versions). bz#3548.\n\n * Tests: fix build problems for the sk-dummy.so FIDO provider module\n used in some tests.\n\nChecksums:\n==========\n\n - SHA1 (openssh-9.4.tar.gz) = d88126d8d7b8e5bf4656587ac4a16055560641cc\n - SHA256 (openssh-9.4.tar.gz) = 7eqFjx2hAunw+1Jy7f1JQXq//3AMr9B3dKtASDtq8go=\n\n - SHA1 (openssh-9.4p1.tar.gz) = 5dea1f3c88f9cfe53a711a3c893ee8b7d3ffecff\n - SHA256 (openssh-9.4p1.tar.gz) = Ngj9kIjbIWPOs+YAyFq3nQ3j0iHlkZLqGSPiMmOGaoU=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.3p2 was released on 2023-07-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.3\n=========================\n\nThis release fixes a security bug.\n\nSecurity\n========\n\nFix CVE-2023-38408 - a condition where specific libraries loaded via\nssh-agent(1)'s PKCS#11 support could be abused to achieve remote\ncode execution via a forwarded agent socket if the following\nconditions are met:\n\n* Exploitation requires the presence of specific libraries on\n the victim system.\n* Remote exploitation requires that the agent was forwarded\n to an attacker-controlled system.\n\nExploitation can also be prevented by starting ssh-agent(1) with an\nempty PKCS#11/FIDO allowlist (ssh-agent -P '') or by configuring\nan allowlist that contains only specific provider libraries.\n\nThis vulnerability was discovered and demonstrated to be exploitable\nby the Qualys Security Advisory team. \n \nIn addition to removing the main precondition for exploitation,\nthis release removes the ability for remote ssh-agent(1) clients\nto load PKCS#11 modules by default (see below).\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh-agent(8): the agent will now refuse requests to load PKCS#11\n modules issued by remote clients by default. A flag has been added\n to restore the previous behaviour "-Oallow-remote-pkcs11".\n\n Note that ssh-agent(8) depends on the SSH client to identify\n requests that are remote. The OpenSSH >=8.9 ssh(1) client does\n this, but forwarding access to an agent socket using other tools\n may circumvent this restriction.\n\nChecksums:\n==========\n\n- SHA1 (openssh-9.3p2.tar.gz) = 219cf700c317f400bb20b001c0406056f7188ea4\n- SHA256 (openssh-9.3p2.tar.gz) = IA6+FH9ss/EB/QzfngJEKvfdyimN/9n0VoeOfMrGdug=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.3 was released on 2023-03-15. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.2\n=========================\n\nThis release fixes a number of security bugs.\n\nSecurity\n========\n\nThis release contains fixes for a security problem and a memory\nsafety problem. The memory safety problem is not believed to be\nexploitable, but we report most network-reachable memory faults as\nsecurity bugs.\n\n * ssh-add(1): when adding smartcard keys to ssh-agent(1) with the\n per-hop destination constraints (ssh-add -h ...) added in OpenSSH\n 8.9, a logic error prevented the constraints from being\n communicated to the agent. This resulted in the keys being added\n without constraints. The common cases of non-smartcard keys and\n keys without destination constraints are unaffected. This problem\n was reported by Luci Stanescu.\n\n * ssh(1): Portable OpenSSH provides an implementation of the\n getrrsetbyname(3) function if the standard library does not\n provide it, for use by the VerifyHostKeyDNS feature. A\n specifically crafted DNS response could cause this function to\n perform an out-of-bounds read of adjacent stack data, but this\n condition does not appear to be exploitable beyond denial-of-\n service to the ssh(1) client.\n\n The getrrsetbyname(3) replacement is only included if the system's\n standard library lacks this function and portable OpenSSH was not\n compiled with the ldns library (--with-ldns). getrrsetbyname(3) is\n only invoked if using VerifyHostKeyDNS to fetch SSHFP records. This\n problem was found by the Coverity static analyzer.\n\nNew features\n------------\n\n * ssh-keygen(1), ssh-keyscan(1): accept -Ohashalg=sha1|sha256 when\n outputting SSHFP fingerprints to allow algorithm selection. bz3493\n \n * sshd(8): add a `sshd -G` option that parses and prints the\n effective configuration without attempting to load private keys\n and perform other checks. This allows usage of the option before\n keys have been generated and for configuration evaluation and\n verification by unprivileged users.\n\nBugfixes\n--------\n\n * scp(1), sftp(1): fix progressmeter corruption on wide displays;\n bz3534\n\n * ssh-add(1), ssh-keygen(1): use RSA/SHA256 when testing usability\n of private keys as some systems are starting to disable RSA/SHA1\n in libcrypto.\n\n * sftp-server(8): fix a memory leak. GHPR363\n\n * ssh(1), sshd(8), ssh-keyscan(1): remove vestigal protocol\n compatibility code and simplify what's left.\n\n * Fix a number of low-impact Coverity static analysis findings.\n These include several reported via bz2687\n\n * ssh_config(5), sshd_config(5): mention that some options are not\n first-match-wins.\n\n * Rework logging for the regression tests. Regression tests will now\n capture separate logs for each ssh and sshd invocation in a test.\n\n * ssh(1): make `ssh -Q CASignatureAlgorithms` work as the manpage\n says it should; bz3532.\n\n * ssh(1): ensure that there is a terminating newline when adding a\n new entry to known_hosts; bz3529\n\nPortability\n-----------\n\n * sshd(8): harden Linux seccomp sandbox. Move to an allowlist of\n mmap(2), madvise(2) and futex(2) flags, removing some concerning\n kernel attack surface.\n\n * sshd(8): improve Linux seccomp-bpf sandbox for older systems;\n bz3537\n\nChecksums:\n==========\n\n- SHA1 (openssh-9.3.tar.gz) = 5f9d2f73ddfe94f3f0a78bdf46704b6ad7b66ec7\n- SHA256 (openssh-9.3.tar.gz) = eRcXkFZByz70DUBUcyIdvU0pVxP2X280FrmV8pyUdrk=\n\n- SHA1 (openssh-9.3p1.tar.gz) = 610959871bf8d6baafc3525811948f85b5dd84ab\n- SHA256 (openssh-9.3p1.tar.gz) = 6bq6dwGnalHz2Fpiw4OjydzZf6kAuFm8fbEUwYaK+Kg=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.2 was released on 2023-02-02. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.1\n=========================\n\nThis release fixes a number of security bugs.\n\nSecurity\n========\n\nThis release contains fixes for two security problems and a memory\nsafety problem. The memory safety problem is not believed to be\nexploitable, but we report most network-reachable memory faults as\nsecurity bugs.\n\n * sshd(8): fix a pre-authentication double-free memory fault\n introduced in OpenSSH 9.1. This is not believed to be exploitable,\n and it occurs in the unprivileged pre-auth process that is\n subject to chroot(2) and is further sandboxed on most major\n platforms.\n\n * ssh(8): in OpenSSH releases after 8.7, the PermitRemoteOpen option\n would ignore its first argument unless it was one of the special\n keywords "any" or "none", causing the permission list to fail open\n if only one permission was specified. bz3515\n\n * ssh(1): if the CanonicalizeHostname and CanonicalizePermittedCNAMEs\n options were enabled, and the system/libc resolver did not check\n that names in DNS responses were valid, then use of these options\n could allow an attacker with control of DNS to include invalid\n characters (possibly including wildcards) in names added to\n known_hosts files when they were updated. These names would still\n have to match the CanonicalizePermittedCNAMEs allow-list, so\n practical exploitation appears unlikely.\n\nPotentially-incompatible changes\n--------------------------------\n\n * ssh(1): add a new EnableEscapeCommandline ssh_config(5) option that\n controls whether the client-side ~C escape sequence that provides a\n command-line is available. Among other things, the ~C command-line\n could be used to add additional port-forwards at runtime.\n\n This option defaults to "no", disabling the ~C command-line that\n was previously enabled by default. Turning off the command-line\n allows platforms that support sandboxing of the ssh(1) client\n (currently only OpenBSD) to use a stricter default sandbox policy.\n\nNew features\n------------\n\n * sshd(8): add support for channel inactivity timeouts via a new\n sshd_config(5) ChannelTimeout directive. This allows channels that\n have not seen traffic in a configurable interval to be\n automatically closed. Different timeouts may be applied to session,\n X11, agent and TCP forwarding channels.\n\n * sshd(8): add a sshd_config UnusedConnectionTimeout option to\n terminate client connections that have no open channels for a\n length of time. This complements the ChannelTimeout option above.\n \n * sshd(8): add a -V (version) option to sshd like the ssh client has.\n\n * ssh(1): add a "Host" line to the output of ssh -G showing the\n original hostname argument. bz3343\n \n * scp(1), sftp(1): add a -X option to both scp(1) and sftp(1) to\n allow control over some SFTP protocol parameters: the copy buffer\n length and the number of in-flight requests, both of which are used\n during upload/download. Previously these could be controlled in\n sftp(1) only. This makes them available in both SFTP protocol\n clients using the same option character sequence.\n \n * ssh-keyscan(1): allow scanning of complete CIDR address ranges,\n e.g. "ssh-keyscan 192.168.0.0/24". If a CIDR range is passed, then\n it will be expanded to all possible addresses in the range\n including the all-0s and all-1s addresses. bz#976\n\n * ssh(1): support dynamic remote port forwarding in escape\n command-line's -R processing. bz#3499\n\nBugfixes\n--------\n\n * ssh(1): when restoring non-blocking mode to stdio fds, restore\n exactly the flags that ssh started with and don't just clobber them\n with zero, as this could also remove the append flag from the set.\n bz3523\n \n * ssh(1): avoid printf("%s", NULL) if using UserKnownHostsFile=none\n and a hostkey in one of the system known hosts file changes.\n \n * scp(1): switch scp from using pipes to a socket-pair for\n communication with its ssh sub-processes, matching how sftp(1)\n operates.\n\n * sshd(8): clear signal mask early in main(); sshd may have been\n started with one or more signals masked (sigprocmask(2) is not\n cleared on fork/exec) and this could interfere with various things,\n e.g. the login grace timer. Execution environments that fail to\n clear the signal mask before running sshd are clearly broken, but\n apparently they do exist.\n \n * ssh(1): warn if no host keys for hostbased auth can be loaded.\n \n * sshd(8): Add server debugging for hostbased auth that is queued and\n sent to the client after successful authentication, but also logged\n to assist in diagnosis of HostbasedAuthentication problems. bz3507\n\n * ssh(1): document use of the IdentityFile option as being usable to\n list public keys as well as private keys. GHPR352\n \n * sshd(8): check for and disallow MaxStartups values less than or\n equal to zero during config parsing, rather than failing later at\n runtime. bz3489\n \n * ssh-keygen(1): fix parsing of hex cert expiry times specified on\n the command-line when acting as a CA.\n \n * scp(1): when scp(1) is using the SFTP protocol for transport (the\n default), better match scp/rcp's handling of globs that don't match\n the globbed characters but do match literally (e.g. trying to\n transfer a file named "foo.[1]"). Previously scp(1) in SFTP mode\n would not match these pathnames but legacy scp/rcp mode would.\n bz3488\n \n * ssh-agent(1): document the "-O no-restrict-websafe" command-line\n option.\n\n * ssh(1): honour user's umask(2) if it is more restrictive then the\n ssh default (022).\n\nPortability\n-----------\n\n * sshd(8): allow writev(2) in the Linux seccomp sandbox. This seems\n to be used by recent glibcs at least in some configurations during\n error conditions. bz3512.\n\n * sshd(8): simply handling of SSH_CONNECTION PAM env var, removing\n global variable and checking the return value from pam_putenv.\n bz3508\n\n * sshd(8): disable SANDBOX_SECCOMP_FILTER_DEBUG that was mistakenly\n enabled during the OpenSSH 9.1 release cycle.\n\n * misc: update autotools and regenerate the config files using the\n latest autotools\n\n * all: use -fzero-call-used-regs=used on clang 15 instead of\n -fzero-call-used-reg=all, as some versions of clang 15 have\n miscompile code when it was enabled. bz3475\n\n * sshd(8): defer PRNG seeding until after the initial closefrom(2)\n call. PRNG seeding will initialize OpenSSL, and some engine\n providers (e.g. Intel's QAT) will open descriptors for their own\n use that closefrom(2) could clobber. bz3483\n\n * misc: in the poll(2)/ppoll(2) compatibility code, avoid assuming\n the layout of fd_set.\n\n * sftp-server(8), ssh-agent(1): fix ptrace(2) disabling on older\n FreeBSD kernels. Some versions do not support using id 0 to refer\n to the current PID for procctl, so try again with getpid()\n explicitly before failing.\n\n * configure.ac: fix -Wstrict-prototypes in configure test code.\n Clang 16 now warns on this and legacy prototypes will be removed\n in C23. GHPR355\n\n * configure.ac: fix setres*id checks to work with clang-16. glibc\n has the prototypes for setresuid behind _GNU_SOURCE, and clang 16\n will error out on implicit function definitions. bz3497\n\nChecksums:\n==========\n\n- SHA1 (openssh-9.2.tar.gz) = e4b806b7c81b87d6c90afe97b3d016ba6cf3ba1c\n- SHA256 (openssh-9.2.tar.gz) = yYe9uaaWSeetXGXOxuaaEiIsLnvITmGW+l5dgMZb9QU=\n\n- SHA1 (openssh-9.2p1.tar.gz) = 3b172b8e971773a7018bbf3231f6589ae539ca4b\n- SHA256 (openssh-9.2p1.tar.gz) = P2bb8WVftF9Q4cVtpiqwEhjCKIB7ITONY068351xz0Y=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.1 was released on 2022-10-04. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 9.0\n=========================\n\nThis release is focused on bug fixing.\n\nSecurity\n========\n\nThis release contains fixes for three minor memory safety problems.\nNone are believed to be exploitable, but we report most memory safety\nproblems as potential security vulnerabilities out of caution.\n\n * ssh-keyscan(1): fix a one-byte overflow in SSH- banner processing.\n Reported by Qualys\n\n * ssh-keygen(1): double free() in error path of file hashing step in\n signing/verify code; GHPR333\n\n * ssh-keysign(8): double-free in error path introduced in openssh-8.9\n\nPotentially-incompatible changes\n--------------------------------\n\n * The portable OpenSSH project now signs commits and release tags\n using git's recent SSH signature support. The list of developer\n signing keys is included in the repository as .git_allowed_signers\n and is cross-signed using the PGP key that is still used to sign\n release artifacts:\n https://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\n * ssh(1), sshd(8): SetEnv directives in ssh_config and sshd_config\n are now first-match-wins to match other directives. Previously\n if an environment variable was multiply specified the last set\n value would have been used. bz3438\n\n * ssh-keygen(8): ssh-keygen -A (generate all default host key types)\n will no longer generate DSA keys, as these are insecure and have\n not been used by default for some years.\n\n\nNew features\n------------\n\n * ssh(1), sshd(8): add a RequiredRSASize directive to set a minimum\n RSA key length. Keys below this length will be ignored for user\n authentication and for host authentication in sshd(8). \n\n ssh(1) will terminate a connection if the server offers an RSA key\n that falls below this limit, as the SSH protocol does not include\n the ability to retry a failed key exchange.\n\n * sftp-server(8): add a "users-groups-by-id@openssh.com" extension\n request that allows the client to obtain user/group names that\n correspond to a set of uids/gids.\n\n * sftp(1): use "users-groups-by-id@openssh.com" sftp-server\n extension (when available) to fill in user/group names for\n directory listings.\n\n * sftp-server(8): support the "home-directory" extension request\n defined in draft-ietf-secsh-filexfer-extensions-00. This overlaps\n a bit with the existing "expand-path@openssh.com", but some other\n clients support it.\n\n * ssh-keygen(1), sshd(8): allow certificate validity intervals,\n sshsig verification times and authorized_keys expiry-time options\n to accept dates in the UTC time zone in addition to the default\n of interpreting them in the system time zone. YYYYMMDD and\n YYMMDDHHMM[SS] dates/times will be interpreted as UTC if suffixed\n with a 'Z' character.\n\n Also allow certificate validity intervals to be specified in raw\n seconds-since-epoch as hex value, e.g. -V 0x1234:0x4567890. This\n is intended for use by regress tests and other tools that call\n ssh-keygen as part of a CA workflow. bz3468\n\n * sftp(1): allow arguments to the sftp -D option, e.g. sftp -D\n "/usr/libexec/sftp-server -el debug3"\n\n * ssh-keygen(1): allow the existing -U (use agent) flag to work\n with "-Y sign" operations, where it will be interpreted to require\n that the private keys is hosted in an agent; bz3429\n\nBugfixes\n--------\n\n * ssh-keygen(1): implement the "verify-required" certificate option.\n This was already documented when support for user-verified FIDO\n keys was added, but the ssh-keygen(1) code was missing.\n\n * ssh-agent(1): hook up the restrict_websafe command-line flag;\n previously the flag was accepted but never actually used.\n\n * sftp(1): improve filename tab completions: never try to complete\n names to non-existent commands, and better match the completion\n type (local or remote filename) against the argument position\n being completed.\n\n * ssh-keygen(1), ssh(1), ssh-agent(1): several fixes to FIDO key\n handling, especially relating to keys that request\n user-verification. These should reduce the number of unnecessary\n PIN prompts for keys that support intrinsic user verification.\n GHPR302, GHPR329\n\n * ssh-keygen(1): when enrolling a FIDO resident key, check if a\n credential with matching application and user ID strings already\n exists and, if so, prompt the user for confirmation before\n overwriting the credential. GHPR329\n\n * sshd(8): improve logging of errors when opening authorized_keys\n files. bz2042\n\n * ssh(1): avoid multiplexing operations that could cause SIGPIPE from\n causing the client to exit early. bz3454\n\n * ssh_config(5), sshd_config(5): clarify that the RekeyLimit\n directive applies to both transmitted and received data. GHPR328\n\n * ssh-keygen(1): avoid double fclose() in error path.\n\n * sshd(8): log an error if pipe() fails while accepting a\n connection. bz3447\n\n * ssh(1), ssh-keygen(1): fix possible NULL deref when built without\n FIDO support. bz3443\n\n * ssh-keyscan(1): add missing *-sk types to ssh-keyscan manpage.\n GHPR294.\n\n * sshd(8): ensure that authentication passwords are cleared from\n memory in error paths. GHPR286\n\n * ssh(1), ssh-agent(1): avoid possibility of notifier code executing\n kill(-1). GHPR286\n\n * ssh_config(5): note that the ProxyJump directive also accepts the\n same tokens as ProxyCommand. GHPR305.\n\n * scp(1): do not ftruncate(3) files early when in sftp mode. The\n previous behaviour of unconditionally truncating the destination\n file would cause "scp ~/foo localhost:foo" and the reverse\n "scp localhost:foo ~/foo" to delete all the contents of their\n destination. bz3431\n\n * ssh-keygen(1): improve error message when 'ssh-keygen -Y sign' is\n unable to load a private key; bz3429\n\n * sftp(1), scp(1): when performing operations that glob(3) a remote\n path, ensure that the implicit working directory used to construct\n that path escapes glob(3) characters. This prevents glob characters\n from being processed in places they shouldn't, e.g. "cd /tmp/a*/",\n "get *.txt" should have the get operation treat the path "/tmp/a*"\n literally and not attempt to expand it.\n\n * ssh(1), sshd(8): be stricter in which characters will be accepted\n in specifying a mask length; allow only 0-9. GHPR278\n\n * ssh-keygen(1): avoid printing hash algorithm twice when dumping a\n KRL\n\n * ssh(1), sshd(8): continue running local I/O for open channels\n during SSH transport rekeying. This should make ~-escapes work in\n the client (e.g. to exit) if the connection happened to have\n stalled during a rekey event.\n\n * ssh(1), sshd(8): avoid potential poll() spin during rekeying\n\n * Further hardening for sshbuf internals: disallow "reparenting" a\n hierarchical sshbuf and zero the entire buffer if reallocation\n fails. GHPR287\n\nPortability\n-----------\n\n * ssh(1), ssh-keygen(1), sshd(8): automatically enable the built-in\n FIDO security key support if libfido2 is found and usable, unless\n --without-security-key-builtin was requested.\n\n * ssh(1), ssh-keygen(1), sshd(8): many fixes to make the WinHello\n FIDO device usable on Cygwin. The windows://hello FIDO device will\n be automatically used by default on this platform unless requested\n otherwise, or when probing resident FIDO credentials (an operation\n not currently supported by WinHello).\n\n * Portable OpenSSH: remove workarounds for obsolete and unsupported\n versions of OpenSSL libcrypto. In particular, this release removes\n fallback support for OpenSSL that lacks AES-CTR or AES-GCM.\n\n Those AES cipher modes were added to OpenSSL prior to the minimum\n version currently supported by OpenSSH, so this is not expected to\n impact any currently supported configurations.\n\n * sshd(8): fix SANDBOX_SECCOMP_FILTER_DEBUG on current Linux/glibc\n\n * All: resync and clean up internal CSPRNG code.\n\n * scp(1), sftp(1), sftp-server(8): avoid linking these programs with\n unnecessary libraries. They are no longer linked against libz and\n libcrypto. This may be of benefit to space constrained systems\n using any of those components in isolation.\n\n * sshd(8): add AUDIT_ARCH_PPC to supported seccomp sandbox\n architectures.\n\n * configure: remove special casing of crypt(). configure will no\n longer search for crypt() in libcrypto, as it was removed from\n there years ago. configure will now only search libc and libcrypt.\n\n * configure: refuse to use OpenSSL 3.0.4 due to potential RCE in its\n RSA implementation (CVE-2022-2274) on x86_64.\n\n * All: request 1.1x API compatibility for OpenSSL >=3.x; GHPR#322\n\n * ssh(1), ssh-keygen(1), sshd(8): fix a number of missing includes\n required by the XMSS code on some platforms.\n\n * sshd(8): cache timezone data in capsicum sandbox.\n\nChecksums:\n==========\n\n- SHA1 (openssh-9.1.tar.gz) = 3ae2d6a3a695d92778c4c4567dcd6ad481092f6c\n- SHA256 (openssh-9.1.tar.gz) = QKfVArlcItV+e8V1Th85TL5//5d/AvOUhYOeHMDEGuE=\n\n- SHA1 (openssh-9.1p1.tar.gz) = 15545440268967511d3194ebf20bcd0c7ff3fcc9\n- SHA256 (openssh-9.1p1.tar.gz) = GfhQCcfj4jeH8CNvuxV4OSq01L+fjsX+a8HNfov90og=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 9.0 was released on 2022-04-08. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nChanges since OpenSSH 8.9\n=========================\n\nThis release is focused on bug fixing.\n\nPotentially-incompatible changes\n--------------------------------\n\nThis release switches scp(1) from using the legacy scp/rcp protocol\nto using the SFTP protocol by default.\n\nLegacy scp/rcp performs wildcard expansion of remote filenames (e.g.\n"scp host:* .") through the remote shell. This has the side effect of\nrequiring double quoting of shell meta-characters in file names\nincluded on scp(1) command-lines, otherwise they could be interpreted\nas shell commands on the remote side.\n\nThis creates one area of potential incompatibility: scp(1) when using\nthe SFTP protocol no longer requires this finicky and brittle quoting,\nand attempts to use it may cause transfers to fail. We consider the\nremoval of the need for double-quoting shell characters in file names\nto be a benefit and do not intend to introduce bug-compatibility for\nlegacy scp/rcp in scp(1) when using the SFTP protocol.\n\nAnother area of potential incompatibility relates to the use of remote\npaths relative to other user's home directories, for example -\n"scp host:~user/file /tmp". The SFTP protocol has no native way to\nexpand a ~user path. However, sftp-server(8) in OpenSSH 8.7 and later\nsupport a protocol extension "expand-path@openssh.com" to support\nthis.\n\nIn case of incompatibility, the scp(1) client may be instructed to use\nthe legacy scp/rcp using the -O flag.\n\nNew features\n------------\n\n * ssh(1), sshd(8): use the hybrid Streamlined NTRU Prime + x25519 key\n exchange method by default ("sntrup761x25519-sha512@openssh.com").\n The NTRU algorithm is believed to resist attacks enabled by future\n quantum computers and is paired with the X25519 ECDH key exchange\n (the previous default) as a backstop against any weaknesses in\n NTRU Prime that may be discovered in the future. The combination\n ensures that the hybrid exchange offers at least as good security\n as the status quo.\n\n We are making this change now (i.e. ahead of cryptographically-\n relevant quantum computers) to prevent "capture now, decrypt\n later" attacks where an adversary who can record and store SSH\n session ciphertext would be able to decrypt it once a sufficiently\n advanced quantum computer is available.\n\n * sftp-server(8): support the "copy-data" extension to allow server-\n side copying of files/data, following the design in\n draft-ietf-secsh-filexfer-extensions-00. bz2948\n\n * sftp(1): add a "cp" command to allow the sftp client to perform\n server-side file copies.\n\nBugfixes\n--------\n\n * ssh(1), sshd(8): upstream: fix poll(2) spin when a channel's output\n fd closes without data in the channel buffer. bz3405 and bz3411\n\n * sshd(8): pack pollfd array in server listen/accept loop. Could\n cause the server to hang/spin when MaxStartups > RLIMIT_NOFILE\n\n * ssh-keygen(1): avoid NULL deref via the find-principals and\n check-novalidate operations. bz3409 and GHPR#307 respectively.\n\n * scp(1): fix a memory leak in argument processing. bz3404\n\n * sshd(8): don't try to resolve ListenAddress directives in the sshd\n re-exec path. They are unused after re-exec and parsing errors\n (possible for example if the host's network configuration changed)\n could prevent connections from being accepted.\n \n * sshd(8): when refusing a public key authentication request from a\n client for using an unapproved or unsupported signature algorithm\n include the algorithm name in the log message to make debugging\n easier.\n \nPortability\n-----------\n\n * sshd(8): refactor platform-specific locked account check, fixing\n an incorrect free() on platforms with both libiaf and shadow\n passwords (probably only Unixware) GHPR#284,\n\n * ssh(1), sshd(8): Fix possible integer underflow in scan_scaled(3)\n parsing of K/M/G/etc quantities. bz#3401.\n\n * sshd(8): provide killpg implementation (mostly for Tandem NonStop)\n GHPR#301.\n\n * Check for missing ftruncate prototype. GHPR#301\n\n * sshd(8): default to not using sandbox when cross compiling. On most\n systems poll(2) does not work when the number of FDs is reduced\n with setrlimit, so assume it doesn't when cross compiling and we\n can't run the test. bz#3398.\n\n * sshd(8): allow ppoll_time64 in seccomp sandbox. Should fix sandbox\n violations on some (at least i386 and armhf) 32bit Linux platforms.\n bz#3396.\n\n * Improve detection of -fzero-call-used-regs=all support in\n configure script.\n\n\nChecksums:\n==========\n\n\n - SHA1 (openssh-9.0.tar.gz) = 05302aa4781e1a69db4261474ed940bd685afc24\n - SHA256 (openssh-9.0.tar.gz) = 9I/FrLf5Gij/4NIPts9A8yWVi0ienyyMqjqn8s0hyLk=\n\n - SHA1 (openssh-9.0p1.tar.gz) = 06dd658874dcd22d66311cf5999bd56c614de509\n - SHA256 (openssh-9.0p1.tar.gz) = A5dDAhYenszjIVPPoQAS8eZcjzdQ9XOnOrG+/Vlyooo=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.9 was released on 2022-02-23. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nA near-future release of OpenSSH will switch scp(1) from using the\nlegacy scp/rcp protocol to using SFTP by default.\n\nLegacy scp/rcp performs wildcard expansion of remote filenames (e.g.\n"scp host:* .") through the remote shell. This has the side effect of\nrequiring double quoting of shell meta-characters in file names\nincluded on scp(1) command-lines, otherwise they could be interpreted\nas shell commands on the remote side.\n\nThis creates one area of potential incompatibility: scp(1) when using\nthe SFTP protocol no longer requires this finicky and brittle quoting,\nand attempts to use it may cause transfers to fail. We consider the\nremoval of the need for double-quoting shell characters in file names\nto be a benefit and do not intend to introduce bug-compatibility for\nlegacy scp/rcp in scp(1) when using the SFTP protocol.\n\nAnother area of potential incompatibility relates to the use of remote\npaths relative to other user's home directories, for example -\n"scp host:~user/file /tmp". The SFTP protocol has no native way to\nexpand a ~user path. However, sftp-server(8) in OpenSSH 8.7 and later\nsupport a protocol extension "expand-path@openssh.com" to support\nthis.\n\nSecurity Near Miss\n==================\n\n * sshd(8): fix an integer overflow in the user authentication path\n that, in conjunction with other logic errors, could have yielded\n unauthenticated access under difficult to exploit conditions.\n\n This situation is not exploitable because of independent checks in\n the privilege separation monitor. Privilege separation has been\n enabled by default in since openssh-3.2.2 (released in 2002) and\n has been mandatory since openssh-7.5 (released in 2017). Moreover,\n portable OpenSSH has used toolchain features available in most\n modern compilers to abort on signed integer overflow since\n openssh-6.5 (released in 2014).\n\n Thanks to Malcolm Stagg for finding and reporting this bug.\n\nPotentially-incompatible changes\n================================\n\n * sshd(8), portable OpenSSH only: this release removes in-built\n support for MD5-hashed passwords. If you require these on your\n system then we recommend linking against libxcrypt or similar.\n\n * This release modifies the FIDO security key middleware interface\n and increments SSH_SK_VERSION_MAJOR.\n\nChanges since OpenSSH 8.8\n=========================\n\nThis release includes a number of new features.\n\nNew features\n------------\n\n * ssh(1), sshd(8), ssh-add(1), ssh-agent(1): add a system for\n restricting forwarding and use of keys added to ssh-agent(1)\n A detailed description of the feature is available at\n https://www.openssh.com/agent-restrict.html and the protocol\n extensions are documented in the PROTOCOL and PROTOCOL.agent\n files in the source release.\n\n * ssh(1), sshd(8): add the sntrup761x25519-sha512@openssh.com hybrid\n ECDH/x25519 + Streamlined NTRU Prime post-quantum KEX to the\n default KEXAlgorithms list (after the ECDH methods but before the\n prime-group DH ones). The next release of OpenSSH is likely to\n make this key exchange the default method.\n \n * ssh-keygen(1): when downloading resident keys from a FIDO token,\n pass back the user ID that was used when the key was created and\n append it to the filename the key is written to (if it is not the\n default). Avoids keys being clobbered if the user created multiple\n resident keys with the same application string but different user\n IDs.\n\n * ssh-keygen(1), ssh(1), ssh-agent(1): better handling for FIDO keys\n on tokens that provide user verification (UV) on the device itself,\n including biometric keys, avoiding unnecessary PIN prompts.\n\n * ssh-keygen(1): add "ssh-keygen -Y match-principals" operation to\n perform matching of principals names against an allowed signers\n file. To be used towards a TOFU model for SSH signatures in git.\n\n * ssh-add(1), ssh-agent(1): allow pin-required FIDO keys to be added\n to ssh-agent(1). $SSH_ASKPASS will be used to request the PIN at\n authentication time.\n \n * ssh-keygen(1): allow selection of hash at sshsig signing time\n (either sha512 (default) or sha256).\n\n * ssh(1), sshd(8): read network data directly to the packet input\n buffer instead of indirectly via a small stack buffer. Provides a\n modest performance improvement.\n\n * ssh(1), sshd(8): read data directly to the channel input buffer,\n providing a similar modest performance improvement.\n\n * ssh(1): extend the PubkeyAuthentication configuration directive to\n accept yes|no|unbound|host-bound to allow control over one of the\n protocol extensions used to implement agent-restricted keys.\n\nBugfixes\n--------\n\n * sshd(8): document that CASignatureAlgorithms, ExposeAuthInfo and\n PubkeyAuthOptions can be used in a Match block. PR#277.\n\n * sshd(8): fix possible string truncation when constructing paths to\n .rhosts/.shosts files with very long user home directory names.\n\n * ssh-keysign(1): unbreak for KEX algorithms that use SHA384/512\n exchange hashes\n \n * ssh(1): don't put the TTY into raw mode when SessionType=none,\n avoids ^C being unable to kill such a session. bz3360\n\n * scp(1): fix some corner-case bugs in SFTP-mode handling of\n ~-prefixed paths.\n\n * ssh(1): unbreak hostbased auth using RSA keys. Allow ssh(1) to\n select RSA keys when only RSA/SHA2 signature algorithms are\n configured (this is the default case). Previously RSA keys were\n not being considered in the default case.\n\n * ssh-keysign(1): make ssh-keysign use the requested signature\n algorithm and not the default for the key type. Part of unbreaking\n hostbased auth for RSA/SHA2 keys.\n\n * ssh(1): stricter UpdateHostkey signature verification logic on\n the client- side. Require RSA/SHA2 signatures for RSA hostkeys\n except when RSA/SHA1 was explicitly negotiated during initial\n KEX; bz3375\n\n * ssh(1), sshd(8): fix signature algorithm selection logic for\n UpdateHostkeys on the server side. The previous code tried to\n prefer RSA/SHA2 for hostkey proofs of RSA keys, but missed some\n cases. This will use RSA/SHA2 signatures for RSA keys if the\n client proposed these algorithms in initial KEX. bz3375\n\n * All: convert all uses of select(2)/pselect(2) to poll(2)/ppoll(2).\n This includes the mainloops in ssh(1), ssh-agent(1), ssh-agent(1)\n and sftp-server(8), as well as the sshd(8) listen loop and all\n other FD read/writability checks. On platforms with missing or\n broken poll(2)/ppoll(2) syscalls a select(2)-based compat shim is\n available.\n \n * ssh-keygen(1): the "-Y find-principals" command was verifying key\n validity when using ca certs but not with simple key lifetimes\n within the allowed signers file.\n \n * ssh-keygen(1): make sshsig verify-time argument parsing optional\n\n * sshd(8): fix truncation in rhosts/shosts path construction.\n\n * ssh(1), ssh-agent(1): avoid xmalloc(0) for PKCS#11 keyid for ECDSA\n keys (we already did this for RSA keys). Avoids fatal errors for\n PKCS#11 libraries that return empty keyid, e.g. Microchip ATECC608B\n "cryptoauthlib"; bz#3364\n\n * ssh(1), ssh-agent(1): improve the testing of credentials against\n inserted FIDO: ask the token whether a particular key belongs to\n it in cases where the token supports on-token user-verification\n (e.g. biometrics) rather than just assuming that it will accept it.\n\n Will reduce spurious "Confirm user presence" notifications for key\n handles that relate to FIDO keys that are not currently inserted in at\n least some cases. bz3366\n \n * ssh(1), sshd(8): correct value for IPTOS_DSCP_LE. It needs to\n allow for the preceding two ECN bits. bz#3373\n\n * ssh-keygen(1): add missing -O option to usage() for the "-Y sign"\n option.\n\n * ssh-keygen(1): fix a NULL deref when using the find-principals\n function, when matching an allowed_signers line that contains a\n namespace restriction, but no restriction specified on the\n command-line\n\n * ssh-agent(1): fix memleak in process_extension(); oss-fuzz\n issue #42719\n\n * ssh(1): suppress "Connection to xxx closed" messages when LogLevel\n is set to "error" or above. bz3378\n\n * ssh(1), sshd(8): use correct zlib flags when inflate(3)-ing\n compressed packet data. bz3372\n \n * scp(1): when recursively transferring files in SFTP mode, create the\n destination directory if it doesn't already exist to match scp(1) in\n legacy RCP mode behaviour.\n \n * scp(1): many improvements in error message consistency between scp(1)\n in SFTP mode vs legacy RCP mode.\n\n * sshd(8): fix potential race in SIGTERM handling PR#289\n\n * ssh(1), ssh(8): since DSA keys are deprecated, move them to the\n end of the default list of public keys so that they will be tried\n last. PR#295\n\n * ssh-keygen(1): allow 'ssh-keygen -Y find-principals' to match\n wildcard principals in allowed_signers files\n \nPortability\n-----------\n\n * ssh(1), sshd(8): don't trust closefrom(2) on Linux. glibc's\n implementation does not work in a chroot when the kernel does not\n have close_range(2). It tries to read from /proc/self/fd and when\n that fails dies with an assertion of sorts. Instead, call\n close_range(2) directly from our compat code and fall back if\n that fails. bz#3349,\n\n * OS X poll(2) is broken; use compat replacement. For character-\n special devices like /dev/null, Darwin's poll(2) returns POLLNVAL\n when polled with POLLIN. Apparently this is Apple bug 3710161 -\n not public but a websearch will find other OSS projects\n rediscovering it periodically since it was first identified in\n 2005.\n\n * Correct handling of exceptfds/POLLPRI in our select(2)-based\n poll(2)/ppoll(2) compat implementation.\n\n * Cygwin: correct checking of mbstowcs() return value.\n\n * Add a basic SECURITY.md that refers people to the openssh.com\n website.\n\n * Enable additional compiler warnings and toolchain hardening flags,\n including -Wbitwise-instead-of-logical, -Wmisleading-indentation,\n -fzero-call-used-regs and -ftrivial-auto-var-init.\n\n * HP/UX. Use compat getline(3) on HP-UX 10.x, where the libc version\n is not reliable.\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.9.tar.gz) = 653310ba1a63959fe2df503fe7ad556445180127\n - SHA256 (openssh-8.9.tar.gz) = mJigktP+Bk0sB7uRPuWgjcCOYZ+mIMdvRlZe66irtQA=\n\n - SHA1 (openssh-8.9p1.tar.gz) = 205cdf0040a238047e2c49f43460e03d76e5d650\n - SHA256 (openssh-8.9p1.tar.gz) = /Ul2VLerFobaxnL7g9+0ukCW6LX/zazNJiOArli+xec=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.8 was released on 2021-09-26. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nA near-future release of OpenSSH will switch scp(1) from using the\nlegacy scp/rcp protocol to using SFTP by default.\n\nLegacy scp/rcp performs wildcard expansion of remote filenames (e.g.\n"scp host:* .") through the remote shell. This has the side effect of\nrequiring double quoting of shell meta-characters in file names\nincluded on scp(1) command-lines, otherwise they could be interpreted\nas shell commands on the remote side.\n\nThis creates one area of potential incompatibility: scp(1) when using\nthe SFTP protocol no longer requires this finicky and brittle quoting,\nand attempts to use it may cause transfers to fail. We consider the\nremoval of the need for double-quoting shell characters in file names\nto be a benefit and do not intend to introduce bug- compatibility for\nlegacy scp/rcp in scp(1) when using the SFTP protocol.\n\nAnother area of potential incompatibility relates to the use of remote\npaths relative to other user's home directories, for example -\n"scp host:~user/file /tmp". The SFTP protocol has no native way to\nexpand a ~user path. However, sftp-server(8) in OpenSSH 8.7 and later\nsupport a protocol extension "expand-path@openssh.com" to support\nthis.\n\nSecurity\n========\n\nsshd(8) from OpenSSH 6.2 through 8.7 failed to correctly initialise\nsupplemental groups when executing an AuthorizedKeysCommand or\nAuthorizedPrincipalsCommand, where a AuthorizedKeysCommandUser or\nAuthorizedPrincipalsCommandUser directive has been set to run the\ncommand as a different user. Instead these commands would inherit\nthe groups that sshd(8) was started with.\n\nDepending on system configuration, inherited groups may allow\nAuthorizedKeysCommand/AuthorizedPrincipalsCommand helper programs to\ngain unintended privilege.\n\nNeither AuthorizedKeysCommand nor AuthorizedPrincipalsCommand are\nenabled by default in sshd_config(5).\n\nPotentially-incompatible changes\n================================\n\nThis release disables RSA signatures using the SHA-1 hash algorithm\nby default. This change has been made as the SHA-1 hash algorithm is\ncryptographically broken, and it is possible to create chosen-prefix\nhash collisions for <USD$50K [1]\n\nFor most users, this change should be invisible and there is\nno need to replace ssh-rsa keys. OpenSSH has supported RFC8332\nRSA/SHA-256/512 signatures since release 7.2 and existing ssh-rsa keys\nwill automatically use the stronger algorithm where possible.\n\nIncompatibility is more likely when connecting to older SSH\nimplementations that have not been upgraded or have not closely tracked\nimprovements in the SSH protocol. For these cases, it may be necessary\nto selectively re-enable RSA/SHA1 to allow connection and/or user\nauthentication via the HostkeyAlgorithms and PubkeyAcceptedAlgorithms\noptions. For example, the following stanza in ~/.ssh/config will enable\nRSA/SHA1 for host and user authentication for a single destination host:\n\n Host old-host\n HostkeyAlgorithms +ssh-rsa\n\tPubkeyAcceptedAlgorithms +ssh-rsa\n\nWe recommend enabling RSA/SHA1 only as a stopgap measure until legacy\nimplementations can be upgraded or reconfigured with another key type\n(such as ECDSA or Ed25519).\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nChanges since OpenSSH 8.7\n=========================\n\nThis release is motivated primarily by the above deprecation and\nsecurity fix.\n\nNew features\n------------\n * ssh(1): allow the ssh_config(5) CanonicalizePermittedCNAMEs\n directive to accept a "none" argument to specify the default\n behaviour.\n\nBugfixes\n--------\n\n * scp(1): when using the SFTP protocol, continue transferring files\n after a transfer error occurs, better matching original scp/rcp\n behaviour.\n \n * ssh(1): fixed a number of memory leaks in multiplexing,\n\n * ssh-keygen(1): avoid crash when using the -Y find-principals\n command.\n\n * A number of documentation and manual improvements, including\n bz#3340, PR#139, PR#215, PR#241, PR#257\n\nPortability\n-----------\n\n * ssh-agent(1): on FreeBSD, use procctl to disable ptrace(2)\n\n * ssh(1)/sshd(8): some fixes to the pselect(2) replacement\n compatibility code. bz#3345\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.8.tar.gz) = 732947082a8998047e839cc0b4c066bf0a7e1a5b\n - SHA256 (openssh-8.8.tar.gz) = AngyrPSQH255hnzU1l7y+LlVAUNcGWtuYQIFEl22nRo=\n\n - SHA1 (openssh-8.8p1.tar.gz) = 1eb964897a4372f6fb96c7effeb509ec71c379c9\n - SHA256 (openssh-8.8p1.tar.gz) = RZCJDqm7ms5Pca4zF4WjpYIyMkNRYZYO1fyGWI8zH+k=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.7 was released on 2021-08-20. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nImminent deprecation notice\n===========================\n\nOpenSSH will disable the ssh-rsa signature scheme by default in the\nnext release.\n\nIn the SSH protocol, the "ssh-rsa" signature scheme uses the SHA-1\nhash algorithm in conjunction with the RSA public key algorithm.\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 algorithm for less than USD$50K.\n\nNote that the deactivation of "ssh-rsa" signatures does not necessarily\nrequire cessation of use for RSA keys. In the SSH protocol, keys may be\ncapable of signing using multiple algorithms. In particular, "ssh-rsa"\nkeys are capable of signing using "rsa-sha2-256" (RSA/SHA256),\n"rsa-sha2-512" (RSA/SHA512) and "ssh-rsa" (RSA/SHA1). Only the last of\nthese is being turned off by default.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs that is still\nenabled by default.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The RFC8709 ssh-ed25519 signature algorithm. It has been supported\n in OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm, for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nOpenSSH recently enabled the UpdateHostKeys option by default to\nassist the client by automatically migrating to better algorithms.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * scp(1): this release changes the behaviour of remote to remote\n copies (e.g. "scp host-a:/path host-b:") to transfer through the\n local host by default. This was previously available via the -3\n flag. This mode avoids the need to expose credentials on the\n origin hop, avoids triplicate interpretation of filenames by the\n shell (by the local system, the copy origin and the destination)\n and, in conjunction with the SFTP support for scp(1) mentioned\n below, allows use of all authentication methods to the remote\n hosts (previously, only non-interactive methods could be used).\n A -R flag has been added to select the old behaviour.\n\n * ssh(1)/sshd(8): both the client and server are now using a\n stricter configuration file parser. The new parser uses more\n shell-like rules for quotes, space and escape characters. It is\n also more strict in rejecting configurations that include options\n lacking arguments. Previously some options (e.g. DenyUsers) could\n appear on a line with no subsequent arguments. This release will\n reject such configurations. The new parser will also reject\n configurations with unterminated quotes and multiple '='\n characters after the option name.\n\n * ssh(1): when using SSHFP DNS records for host key verification,\n ssh(1) will verify all matching records instead of just those\n with the specific signature type requested. This may cause host\n key verification problems if stale SSHFP records of a different\n or legacy signature type exist alongside other records for a\n particular host. bz#3322\n\n * ssh-keygen(1): when generating a FIDO key and specifying an\n explicit attestation challenge (using -Ochallenge), the challenge\n will now be hashed by the builtin security key middleware. This\n removes the (undocumented) requirement that challenges be exactly\n 32 bytes in length and matches the expectations of libfido2.\n\n * sshd(8): environment="..." directives in authorized_keys files are\n now first-match-wins and limited to 1024 discrete environment\n variable names.\n\nChanges since OpenSSH 8.6\n=========================\n\nThis release contains a mix of new features and bug-fixes.\n\nNew features\n------------\n\n - scp(1): experimental support for transfers using the SFTP protocol\n as a replacement for the venerable SCP/RCP protocol that it has\n traditionally used. SFTP offers more predictable filename handling\n and does not require expansion of glob(3) patterns via the shell\n on the remote side.\n\n SFTP support may be enabled via a temporary scp -s flag. It is\n intended for SFTP to become the default transfer mode in the\n near future, at which time the -s flag will be removed. The -O\n flag exists to force use of the original SCP/RCP protocol for\n cases where SFTP may be unavailable or incompatible.\n\n - sftp-server(8): add a protocol extension to support expansion of\n ~/ and ~user/ prefixed paths. This was added to support these\n paths when used by scp(1) while in SFTP mode.\n\n - ssh(1): add a ForkAfterAuthentication ssh_config(5) counterpart to\n the ssh(1) -f flag. GHPR#231\n\n - ssh(1): add a StdinNull directive to ssh_config(5) that allows the\n config file to do the same thing as -n does on the ssh(1) command-\n line. GHPR#231\n\n - ssh(1): add a SessionType directive to ssh_config, allowing the\n configuration file to offer equivalent control to the -N (no\n session) and -s (subsystem) command-line flags. GHPR#231\n\n - ssh-keygen(1): allowed signers files used by ssh-keygen(1)\n signatures now support listing key validity intervals alongside\n they key, and ssh-keygen(1) can optionally check during signature\n verification whether a specified time falls inside this interval.\n This feature is intended for use by git to support signing and\n verifying objects using ssh keys.\n\n - ssh-keygen(8): support printing of the full public key in a sshsig\n signature via a -Oprint-pubkey flag.\n\nBugfixes\n--------\n\n * ssh(1)/sshd(8): start time-based re-keying exactly on schedule in\n the client and server mainloops. Previously the re-key timeout\n could expire but re-keying would not start until a packet was sent\n or received, causing a spin in select() if the connection was\n quiescent.\n\n * ssh-keygen(1): avoid Y2038 problem in printing certificate\n validity lifetimes. Dates past 2^31-1 seconds since epoch were\n displayed incorrectly on some platforms. bz#3329\n\n * scp(1): allow spaces to appear in usernames for local to remote\n and scp -3 remote to remote copies. bz#1164\n\n * ssh(1)/sshd(8): remove references to ChallengeResponseAuthentication\n in favour of KbdInteractiveAuthentication. The former is what was in\n SSHv1, the latter is what is in SSHv2 (RFC4256) and they were\n treated as somewhat but not entirely equivalent. We retain the old\n name as a deprecated alias so configuration files continue to work\n as well as a reference in the man page for people looking for it.\n bz#3303\n\n * ssh(1)/ssh-add(1)/ssh-keygen(1): fix decoding of X.509 subject name\n when extracting a key from a PKCS#11 certificate. bz#3327\n\n * ssh(1): restore blocking status on stdio fds before close. ssh(1)\n needs file descriptors in non-blocking mode to operate but it was\n not restoring the original state on exit. This could cause\n problems with fds shared with other programs via the shell,\n bz#3280 and GHPR#246\n\n * ssh(1)/sshd(8): switch both client and server mainloops from\n select(3) to pselect(3). Avoids race conditions where a signal\n may arrive immediately before select(3) and not be processed until\n an event fires. bz#2158\n\n * ssh(1): sessions started with ControlPersist were incorrectly\n executing a shell when the -N (no shell) option was specified.\n bz#3290\n\n * ssh(1): check if IPQoS or TunnelDevice are already set before\n overriding. Prevents values in config files from overriding values\n supplied on the command line. bz#3319\n\n * ssh(1): fix debug message when finding a private key to match a\n certificate being attempted for user authentication. Previously it\n would print the certificate's path, whereas it was supposed to be\n showing the private key's path. GHPR#247\n\n * sshd(8): match host certificates against host public keys, not\n private keys. Allows use of certificates with private keys held in\n a ssh-agent. bz#3524\n\n * ssh(1): add a workaround for a bug in OpenSSH 7.4 sshd(8), which\n allows RSA/SHA2 signatures for public key authentication but fails\n to advertise this correctly via SSH2_MSG_EXT_INFO. This causes\n clients of these server to incorrectly match\n PubkeyAcceptedAlgorithmse and potentially refuse to offer valid\n keys. bz#3213\n\n * sftp(1)/scp(1): degrade gracefully if a sftp-server offers the\n limits@openssh.com extension but fails when the client tries to\n invoke it. bz#3318\n\n * ssh(1): allow ssh_config SetEnv to override $TERM, which is\n otherwise handled specially by the protocol. Useful in ~/.ssh/config\n to set TERM to something generic (e.g. "xterm" instead of\n "xterm-256color") for destinations that lack terminfo entries.\n\n * sftp-server(8): the limits@openssh.com extension was incorrectly\n marked as an operation that writes to the filesystem, which made it\n unavailable in sftp-server read-only mode. bz#3318\n\n * ssh(1): fix SEGV in UpdateHostkeys debug() message, triggered when\n the update removed more host keys than remain present.\n\n * many manual page fixes.\n\nPortability\n-----------\n\n * ssh(1): move closefrom() to before first malloc. When built against\n tcmalloc, the closefrom() would stomp on file descriptors created\n for tcmalloc's internal use. bz#3321\n\n * sshd(8): handle GIDs > 2^31 in getgrouplist. When compiled in 32bit\n mode, the getgrouplist implementation may fail for GIDs greater than\n LONG_MAX.\n\n * ssh(1): xstrdup environment variable used by ForwardAgent. bz#3328\n\n * sshd(8): don't sigdie() in signal handler in privsep child process;\n this can end up causing sandbox violations per bz3286\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.7.tar.gz) = 61bfa5a55e2b7e851b1d463aa432e4ff508f61cc\n - SHA256 (openssh-8.7.tar.gz) = Q5jPfCaYhTKhkDPYH32jtjFN7qhNhkNHYC9awg2THLs=\n\n - SHA1 (openssh-8.7p1.tar.gz) = 8719032c1e47732c8fdb14adfb24b5e9e71de802\n - SHA256 (openssh-8.7p1.tar.gz) = fKNLi7JK6eUPM3krcJGzhB1+G0QP9XvJ+r3fAeLtHiQ=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.6 was released on 2021-04-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 algorithm for less than USD$50K.\n\nIn the SSH protocol, the "ssh-rsa" signature scheme uses the SHA-1\nhash algorithm in conjunction with the RSA public key algorithm.\nOpenSSH will disable this signature scheme by default in the near\nfuture.\n\nNote that the deactivation of "ssh-rsa" signatures does not necessarily\nrequire cessation of use for RSA keys. In the SSH protocol, keys may be\ncapable of signing using multiple algorithms. In particular, "ssh-rsa"\nkeys are capable of signing using "rsa-sha2-256" (RSA/SHA256),\n"rsa-sha2-512" (RSA/SHA512) and "ssh-rsa" (RSA/SHA1). Only the last of\nthese is being turned off by default.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs that is still\nenabled by default.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The RFC8709 ssh-ed25519 signature algorithm. It has been supported\n in OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm, for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nOpenSSH recently enabled the UpdateHostKeys option by default to assist\nthe client by automatically migrating to better algorithms.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nSecurity\n========\n\n * sshd(8): OpenSSH 8.5 introduced the LogVerbose keyword. When this\n option was enabled with a set of patterns that activated logging\n in code that runs in the low-privilege sandboxed sshd process, the\n log messages were constructed in such a way that printf(3) format\n strings could effectively be specified the low-privilege code.\n\n An attacker who had successfully exploited the low-privilege\n process could use this to escape OpenSSH's sandboxing and attack\n the high-privilege process. Exploitation of this weakness is\n highly unlikely in practice as the LogVerbose option is not\n enabled by default and is typically only used for debugging. No\n vulnerabilities in the low-privilege process are currently known\n to exist.\n\n Thanks to Ilja Van Sprundel for reporting this bug.\n\nChanges since OpenSSH 8.5\n=========================\n\nThis release contains mostly bug fixes.\n\nNew features\n------------\n\n * sftp-server(8): add a new limits@openssh.com protocol extension\n that allows a client to discover various server limits, including\n maximum packet size and maximum read/write length.\n\n * sftp(1): use the new limits@openssh.com extension (when available)\n to select better transfer lengths in the client.\n\n * sshd(8): Add ModuliFile keyword to sshd_config to specify the\n location of the "moduli" file containing the groups for DH-GEX.\n\n * unit tests: Add a TEST_SSH_ELAPSED_TIMES environment variable to\n enable printing of the elapsed time in seconds of each test.\n\nBugfixes\n--------\n\n * ssh_config(5), sshd_config(5): sync CASignatureAlgorithms lists in\n manual pages with the current default. GHPR#174\n\n * ssh(1): ensure that pkcs11_del_provider() is called before exit.\n GHPR#234\n\n * ssh(1), sshd(8): fix problems in string->argv conversion. Multiple\n backslashes were not being dequoted correctly and quoted space in\n the middle of a string was being incorrectly split. GHPR#223\n\n * ssh(1): return non-zero exit status when killed by signal; bz#3281\n\n * sftp-server(8): increase maximum SSH2_FXP_READ to match the maximum\n packet size. Also handle zero-length reads that are not explicitly\n banned by the spec.\n\nPortability\n-----------\n\n * sshd(8): don't mistakenly exit on transient read errors on the\n network socket (e.g. EINTR, EAGAIN); bz3297\n\n * Create a dedicated contrib/gnome-ssk-askpass3.c source instead of\n building it from the same file as used for GNOME2. Use the GNOME3\n gdk_seat_grab() to manage keyboard/mouse/server grabs for better\n compatibility with Wayland.\n\n * Fix portability build errors bz3293 bz3292 bz3291 bz3278\n\n * sshd(8): soft-disallow the fstatat64 syscall in the Linux\n seccomp-bpf sandbox. bz3276\n\n * unit tests: enable autoopt and misc unit tests that were\n previously skipped\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.6.tar.gz) = a3e93347eed6296faaaceb221e8786391530fccb\n - SHA256 (openssh-8.6.tar.gz) = ihmgdEgKfCBRpC0qzdQRwYownrpBf+rsihvk4Rmim8M=\n\n - SHA1 (openssh-8.6p1.tar.gz) = 8f9f0c94317baeb97747d6258f3997b4542762c0\n - SHA256 (openssh-8.6p1.tar.gz) = w+bk2hYhdiyFDQO0fu0eSN/0zJYI3etUcgKiNN+O164=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.5 was released on 2021-03-03. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 algorithm for less than USD$50K.\n\nIn the SSH protocol, the "ssh-rsa" signature scheme uses the SHA-1\nhash algorithm in conjunction with the RSA public key algorithm.\nOpenSSH will disable this signature scheme by default in the near\nfuture.\n\nNote that the deactivation of "ssh-rsa" signatures does not necessarily\nrequire cessation of use for RSA keys. In the SSH protocol, keys may be\ncapable of signing using multiple algorithms. In particular, "ssh-rsa"\nkeys are capable of signing using "rsa-sha2-256" (RSA/SHA256),\n"rsa-sha2-512" (RSA/SHA512) and "ssh-rsa" (RSA/SHA1). Only the last of\nthese is being turned off by default.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs that is still\nenabled by default.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The RFC8709 ssh-ed25519 signature algorithm. It has been supported\n in OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm, for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nThis release enables the UpdateHostKeys option by default to assist\nthe client by automatically migrating to better algorithms.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nSecurity\n========\n\n * ssh-agent(1): fixed a double-free memory corruption that was\n introduced in OpenSSH 8.2 . We treat all such memory faults as\n potentially exploitable. This bug could be reached by an attacker\n with access to the agent socket.\n\n On modern operating systems where the OS can provide information\n about the user identity connected to a socket, OpenSSH ssh-agent\n and sshd limit agent socket access only to the originating user\n and root. Additional mitigation may be afforded by the system's\n malloc(3)/free(3) implementation, if it detects double-free\n conditions.\n\n The most likely scenario for exploitation is a user forwarding an\n agent either to an account shared with a malicious user or to a\n host with an attacker holding root access.\n\n * Portable sshd(8): Prevent excessively long username going to PAM.\n This is a mitigation for a buffer overflow in Solaris' PAM username\n handling (CVE-2020-14871), and is only enabled for Sun-derived PAM\n implementations. This is not a problem in sshd itself, it only\n prevents sshd from being used as a vector to attack Solaris' PAM.\n It does not prevent the bug in PAM from being exploited via some\n other PAM application. GHPR#212\n\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh(1), sshd(8): this release changes the first-preference signature\n algorithm from ECDSA to ED25519.\n\n * ssh(1), sshd(8): set the TOS/DSCP specified in the configuration\n for interactive use prior to TCP connect. The connection phase of\n the SSH session is time-sensitive and often explicitly interactive.\n The ultimate interactive/bulk TOS/DSCP will be set after\n authentication completes.\n\n * ssh(1), sshd(8): remove the pre-standardization cipher\n rijndael-cbc@lysator.liu.se. It is an alias for aes256-cbc before\n it was standardized in RFC4253 (2006), has been deprecated and\n disabled by default since OpenSSH 7.2 (2016) and was only briefly\n documented in ssh.1 in 2001.\n\n * ssh(1), sshd(8): update/replace the experimental post-quantum\n hybrid key exchange method based on Streamlined NTRU Prime coupled\n with X25519.\n\n The previous sntrup4591761x25519-sha512@tinyssh.org method is\n replaced with sntrup761x25519-sha512@openssh.com. Per its\n designers, the sntrup4591761 algorithm was superseded almost two\n years ago by sntrup761.\n\n (note this both the updated method and the one that it replaced are\n disabled by default)\n\n * ssh(1): disable CheckHostIP by default. It provides insignificant\n benefits while making key rotation significantly more difficult,\n especially for hosts behind IP-based load-balancers.\n\nChanges since OpenSSH 8.4\n=========================\n\nNew features\n------------\n\n * ssh(1): this release enables UpdateHostkeys by default subject to\n some conservative preconditions:\n - The key was matched in the UserKnownHostsFile (and not in the\n GlobalKnownHostsFile).\n - The same key does not exist under another name.\n - A certificate host key is not in use.\n - known_hosts contains no matching wildcard hostname pattern.\n - VerifyHostKeyDNS is not enabled.\n - The default UserKnownHostsFile is in use.\n\n We expect some of these conditions will be modified or relaxed in\n future.\n\n * ssh(1), sshd(8): add a new LogVerbose configuration directive for\n that allows forcing maximum debug logging by file/function/line\n pattern-lists.\n\n * ssh(1): when prompting the user to accept a new hostkey, display\n any other host names/addresses already associated with the key.\n\n * ssh(1): allow UserKnownHostsFile=none to indicate that no\n known_hosts file should be used to identify host keys.\n\n * ssh(1): add a ssh_config KnownHostsCommand option that allows the\n client to obtain known_hosts data from a command in addition to\n the usual files.\n\n * ssh(1): add a ssh_config PermitRemoteOpen option that allows the\n client to restrict the destination when RemoteForward is used\n with SOCKS.\n\n * ssh(1): for FIDO keys, if a signature operation fails with a\n "incorrect PIN" reason and no PIN was initially requested from the\n user, then request a PIN and retry the operation. This supports\n some biometric devices that fall back to requiring PIN when reading\n of the biometric failed, and devices that require PINs for all\n hosted credentials.\n\n * sshd(8): implement client address-based rate-limiting via new\n sshd_config(5) PerSourceMaxStartups and PerSourceNetBlockSize\n directives that provide more fine-grained control on a per-origin\n address basis than the global MaxStartups limit.\n\nBugfixes\n--------\n\n * ssh(1): Prefix keyboard interactive prompts with "(user@host)" to\n make it easier to determine which connection they are associated\n with in cases like scp -3, ProxyJump, etc. bz#3224\n\n * sshd(8): fix sshd_config SetEnv directives located inside Match\n blocks. GHPR#201\n\n * ssh(1): when requesting a FIDO token touch on stderr, inform the\n user once the touch has been recorded.\n\n * ssh(1): prevent integer overflow when ridiculously large\n ConnectTimeout values are specified, capping the effective value\n (for most platforms) at 24 days. bz#3229\n\n * ssh(1): consider the ECDSA key subtype when ordering host key\n algorithms in the client.\n\n * ssh(1), sshd(8): rename the PubkeyAcceptedKeyTypes keyword to\n PubkeyAcceptedAlgorithms. The previous name incorrectly suggested\n that it control allowed key algorithms, when this option actually\n specifies the signature algorithms that are accepted. The previous\n name remains available as an alias. bz#3253\n\n * ssh(1), sshd(8): similarly, rename HostbasedKeyTypes (ssh) and\n HostbasedAcceptedKeyTypes (sshd) to HostbasedAcceptedAlgorithms.\n\n * sftp-server(8): add missing lsetstat@openssh.com documentation\n and advertisement in the server's SSH2_FXP_VERSION hello packet.\n\n * ssh(1), sshd(8): more strictly enforce KEX state-machine by\n banning packet types once they are received. Fixes memleak caused\n by duplicate SSH2_MSG_KEX_DH_GEX_REQUEST (oss-fuzz #30078).\n\n * sftp(1): allow the full range of UIDs/GIDs for chown/chgrp on 32bit\n platforms instead of being limited by LONG_MAX. bz#3206\n\n * Minor man page fixes (capitalization, commas, etc.) bz#3223\n\n * sftp(1): when doing an sftp recursive upload or download of a\n read-only directory, ensure that the directory is created with\n write and execute permissions in the interim so that the transfer\n can actually complete, then set the directory permission as the\n final step. bz#3222\n\n * ssh-keygen(1): document the -Z, check the validity of its argument\n earlier and provide a better error message if it's not correct.\n bz#2879\n\n * ssh(1): ignore comments at the end of config lines in ssh_config,\n similar to what we already do for sshd_config. bz#2320\n\n * sshd_config(5): mention that DisableForwarding is valid in a\n sshd_config Match block. bz3239\n\n * sftp(1): fix incorrect sorting of "ls -ltr" under some\n circumstances. bz3248.\n\n * ssh(1), sshd(8): fix potential integer truncation of (unlikely)\n timeout values. bz#3250\n\n * ssh(1): make hostbased authentication send the signature algorithm\n in its SSH2_MSG_USERAUTH_REQUEST packets instead of the key type.\n This make HostbasedAcceptedAlgorithms do what it is supposed to -\n filter on signature algorithm and not key type.\n\nPortability\n-----------\n\n * sshd(8): add a number of platform-specific syscalls to the Linux\n seccomp-bpf sandbox. bz#3232 bz#3260\n\n * sshd(8): remove debug message from sigchld handler that could cause\n deadlock on some platforms. bz#3259\n\n * Sync contrib/ssh-copy-id with upstream.\n\n * unittests: add a hostname function for systems that don't have it.\n Some systems don't have a hostname command (it's not required by\n POSIX). The do have uname -n (which is), but not all of those have\n it report the FQDN.\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.5.tar.gz) = 04cae43c389fb411227c01219e4eb46e3113f34e\n - SHA256 (openssh-8.5.tar.gz) = 5qB2CgzNG4io4DmChTjHgCWqRWvEOvCKJskLdJCz+SU=\n\n - SHA1 (openssh-8.5p1.tar.gz) = 72eadcbe313b07b1dd3b693e41d3cd56d354e24e\n - SHA256 (openssh-8.5p1.tar.gz) = 9S8/QdQpqpkY44zyAK8iXM3Y5m8FLaVyhwyJc3ZG7CU=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available from the mirror sites:\nhttps://cdn.openbsd.org/pub/OpenBSD/OpenSSH/RELEASE_KEY.asc\n\nPlease note that the OpenPGP key used to sign releases has been\nrotated for this release. The new key has been signed by the previous\nkey to provide continuity.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.4 was released on 2020-09-27. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 algorithm for less than USD$50K. For this reason, we will be\ndisabling the "ssh-rsa" public key signature algorithm by default in a\nnear-future release.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The ssh-ed25519 signature algorithm. It has been supported in\n OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm, for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nWe intend to enable UpdateHostKeys by default in the next OpenSSH\nrelease. This will assist the client by automatically migrating to\nbetter algorithms. Users may consider enabling this option manually.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nSecurity\n========\n\n * ssh-agent(1): restrict ssh-agent from signing web challenges for\n FIDO/U2F keys.\n\n When signing messages in ssh-agent using a FIDO key that has an\n application string that does not start with "ssh:", ensure that the\n message being signed is one of the forms expected for the SSH protocol\n (currently public key authentication and sshsig signatures).\n\n This prevents ssh-agent forwarding on a host that has FIDO keys\n attached granting the ability for the remote side to sign challenges\n for web authentication using those keys too.\n\n Note that the converse case of web browsers signing SSH challenges is\n already precluded because no web RP can have the "ssh:" prefix in the\n application string that we require.\n\n * ssh-keygen(1): Enable FIDO 2.1 credProtect extension when generating\n a FIDO resident key.\n\n The recent FIDO 2.1 Client to Authenticator Protocol introduced a\n "credProtect" feature to better protect resident keys. We use this\n option to require a PIN prior to all operations that may retrieve\n a resident key from a FIDO token.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * For FIDO/U2F support, OpenSSH recommends the use of libfido2 1.5.0\n or greater. Older libraries have limited support at the expense of\n disabling particular features. These include resident keys, PIN-\n required keys and multiple attached tokens.\n\n * ssh-keygen(1): the format of the attestation information optionally\n recorded when a FIDO key is generated has changed. It now includes\n the authenticator data needed to validate attestation signatures. \n\n * The API between OpenSSH and the FIDO token middleware has changed\n and the SSH_SK_VERSION_MAJOR version has been incremented as a\n result. Third-party middleware libraries must support the current\n API version (7) to work with OpenSSH 8.4.\n\n * The portable OpenSSH distribution now requires automake to rebuild\n the configure script and supporting files. This is not required when\n simply building portable OpenSSH from a release tar file.\n\nChanges since OpenSSH 8.3\n=========================\n\nNew features\n------------\n\n * ssh(1), ssh-keygen(1): support for FIDO keys that require a PIN for\n each use. These keys may be generated using ssh-keygen using a new\n "verify-required" option. When a PIN-required key is used, the user\n will be prompted for a PIN to complete the signature operation.\n\n * sshd(8): authorized_keys now supports a new "verify-required"\n option to require FIDO signatures assert that the token verified\n that the user was present before making the signature. The FIDO\n protocol supports multiple methods for user-verification, but\n currently OpenSSH only supports PIN verification.\n\n * sshd(8), ssh-keygen(1): add support for verifying FIDO webauthn\n signatures. Webauthn is a standard for using FIDO keys in web\n browsers. These signatures are a slightly different format to plain\n FIDO signatures and thus require explicit support.\n\n * ssh(1): allow some keywords to expand shell-style ${ENV}\n environment variables. The supported keywords are CertificateFile,\n ControlPath, IdentityAgent and IdentityFile, plus LocalForward and\n RemoteForward when used for Unix domain socket paths. bz#3140\n\n * ssh(1), ssh-agent(1): allow some additional control over the use of\n ssh-askpass via a new $SSH_ASKPASS_REQUIRE environment variable,\n including forcibly enabling and disabling its use. bz#69\n\n * ssh(1): allow ssh_config(5)'s AddKeysToAgent keyword accept a time\n limit for keys in addition to its current flag options. Time-\n limited keys will automatically be removed from ssh-agent after\n their expiry time has passed.\n\n * scp(1), sftp(1): allow the -A flag to explicitly enable agent\n forwarding in scp and sftp. The default remains to not forward an\n agent, even when ssh_config enables it.\n\n * ssh(1): add a '%k' TOKEN that expands to the effective HostKey of\n the destination. This allows, e.g., keeping host keys in individual\n files using "UserKnownHostsFile ~/.ssh/known_hosts.d/%k". bz#1654\n\n * ssh(1): add %-TOKEN, environment variable and tilde expansion to\n the UserKnownHostsFile directive, allowing the path to be\n completed by the configuration (e.g. bz#1654)\n\n * ssh-keygen(1): allow "ssh-add -d -" to read keys to be deleted\n from stdin. bz#3180\n\n * sshd(8): improve logging for MaxStartups connection throttling.\n sshd will now log when it starts and stops throttling and periodically\n while in this state. bz#3055\n\nBugfixes\n--------\n\n * ssh(1), ssh-keygen(1): better support for multiple attached FIDO\n tokens. In cases where OpenSSH cannot unambiguously determine which\n token to direct a request to, the user is now required to select a\n token by touching it. In cases of operations that require a PIN to\n be verified, this avoids sending the wrong PIN to the wrong token\n and incrementing the token's PIN failure counter (tokens\n effectively erase their keys after too many PIN failures).\n\n * sshd(8): fix Include before Match in sshd_config; bz#3122\n\n * ssh(1): close stdin/out/error when forking after authentication\n completes ("ssh -f ...") bz#3137\n\n * ssh(1), sshd(8): limit the amount of channel input data buffered,\n avoiding peers that advertise large windows but are slow to read\n from causing high memory consumption.\n\n * ssh-agent(1): handle multiple requests sent in a single write() to\n the agent.\n\n * sshd(8): allow sshd_config longer than 256k\n\n * sshd(8): avoid spurious "Unable to load host key" message when sshd\n load a private key but no public counterpart\n\n * ssh(1): prefer the default hostkey algorithm list whenever we have\n a hostkey that matches its best-preference algorithm.\n\n * sshd(1): when ordering the hostkey algorithms to request from a\n server, prefer certificate types if the known_hosts files contain a key\n marked as a @cert-authority; bz#3157\n\n * ssh(1): perform host key fingerprint comparisons for the "Are you\n sure you want to continue connecting (yes/no/[fingerprint])?"\n prompt with case sensitivity.\n\n * sshd(8): ensure that address/masklen mismatches in sshd_config\n yield fatal errors at daemon start time rather than later when\n they are evaluated.\n\n * ssh-keygen(1): ensure that certificate extensions are lexically\n sorted. Previously if the user specified a custom extension then\n the everything would be in order except the custom ones. bz#3198\n\n * ssh(1): also compare username when checking for JumpHost loops.\n bz#3057\n\n * ssh-keygen(1): preserve group/world read permission on known_hosts\n files across runs of "ssh-keygen -Rf /path". The old behaviour was\n to remove all rights for group/other. bz#3146\n\n * ssh-keygen(1): Mention the [-a rounds] flag in the ssh-keygen\n manual page and usage().\n\n * sshd(8): explicitly construct path to ~/.ssh/rc rather than\n relying on it being relative to the current directory, so that it\n can still be found if the shell startup changes its directory.\n bz#3185\n\n * sshd(8): when redirecting sshd's log output to a file, undo this\n redirection after the session child process is forked(). Fixes\n missing log messages when using this feature under some\n circumstances.\n\n * sshd(8): start ClientAliveInterval bookkeeping before first pass\n through select() loop; fixed theoretical case where busy sshd may\n ignore timeouts from client.\n\n * ssh(1): only reset the ServerAliveInterval check when we receive\n traffic from the server and ignore traffic from a port forwarding\n client, preventing a client from keeping a connection alive when\n it should be terminated. bz#2265\n\n * ssh-keygen(1): avoid spurious error message when ssh-keygen\n creates files outside ~/.ssh\n\n * sftp-client(1): fix off-by-one error that caused sftp downloads to\n make one more concurrent request that desired. This prevented using\n sftp(1) in unpipelined request/response mode, which is useful when\n debugging. bz#3054\n\n * ssh(1), sshd(8): handle EINTR in waitfd() and timeout_connect()\n helpers. bz#3071\n\n * ssh(1), ssh-keygen(1): defer creation of ~/.ssh until we attempt to\n write to it so we don't leave an empty .ssh directory when it's not\n needed. bz#3156\n\n * ssh(1), sshd(8): fix multiplier when parsing time specifications\n when handling seconds after other units. bz#3171\n\nPortability\n-----------\n\n * sshd(8): always send any PAM account messages. If the PAM account\n stack returns any messages, always send them to the user and not\n just if the check succeeds. bz#2049\n\n * Implement some backwards compatibility for libfido2 libraries\n older than 1.5.0. Note that use of an older library will result\n in the loss of certain features including resident key support,\n PIN support and support for multiple attached tokens.\n\n * configure fixes for XCode 12\n\n * gnome-ssh-askpass3: ensure the "close" button is not focused by\n default for SSH_ASKPASS_PROMPT=none prompts. Avoids space/enter\n accidentally dismissing FIDO touch notifications.\n\n * gnome-ssh-askpass3: allow some control over textarea colour via\n $GNOME_SSH_ASKPASS_FG_COLOR and $GNOME_SSH_ASKPASS_BG_COLOR\n environment variables.\n\n * sshd(8): document another PAM spec problem in a frustrated comment\n\n * sshd(8): support NetBSD's utmpx.ut_ss address field. bz#960\n\n * Add the ssh-sk-helper binary and its manpage to the RPM spec file\n\n * Detect the Frankenstein monster of Linux/X32 and allow the sandbox\n to function there. bz#3085\n\nChecksums:\n==========\n\n- SHA1 (openssh-8.4.tar.gz) = 71675139df6807f396e6bd92ff8cb9b0356385d8\n- SHA256 (openssh-8.4.tar.gz) = JhBgLYkyRge/zQK8ylBSRcOYvrV/tHwQcvVXfExGB70=\n\n- SHA1 (openssh-8.4p1.tar.gz) = 69305059e10a60693ebe6f17731f962c9577535c\n- SHA256 (openssh-8.4p1.tar.gz) = WgHSLkB+scBbqKj3xlTTiKE+nyJuTtM704dI2vodKyQ=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.3 was released on 2020-05-27. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 algorithm for less than USD$50K. For this reason, we will be\ndisabling the "ssh-rsa" public key signature algorithm by default in a\nnear-future release.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The ssh-ed25519 signature algorithm. It has been supported in\n OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm, for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nA future release of OpenSSH will enable UpdateHostKeys by default\nto allow the client to automatically migrate to better algorithms.\nUsers may consider enabling this option manually. Vendors of devices\nthat implement the SSH protocol should ensure that they support the\nnew signature algorithms for RSA keys.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nSecurity\n========\n\n * scp(1): when receiving files, scp(1) could be become desynchronised\n if a utimes(2) system call failed. This could allow file contents\n to be interpreted as file metadata and thereby permit an adversary\n to craft a file system that, when copied with scp(1) in a\n configuration that caused utimes(2) to fail (e.g. under a SELinux\n policy or syscall sandbox), transferred different file names and\n contents to the actual file system layout.\n\n Exploitation of this is not likely as utimes(2) does not fail under\n normal circumstances. Successful exploitation is not silent - the\n output of scp(1) would show transfer errors followed by the actual\n file(s) that were received.\n\n Finally, filenames returned from the peer are (since openssh-8.0)\n matched against the user's requested destination, thereby\n disallowing a successful exploit from writing files outside the\n user's selected target glob (or directory, in the case of a\n recursive transfer). This ensures that this attack can achieve no\n more than a hostile peer is already able to achieve within the scp\n protocol.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * sftp(1): reject an argument of "-1" in the same way as ssh(1) and\n scp(1) do instead of accepting and silently ignoring it.\n\nChanges since OpenSSH 8.2\n=========================\n\nThe focus of this release is bug fixing.\n\nNew Features\n------------\n\n * sshd(8): make IgnoreRhosts a tri-state option: "yes" to ignore\n rhosts/shosts, "no" allow rhosts/shosts or (new) "shosts-only"\n to allow .shosts files but not .rhosts.\n\n * sshd(8): allow the IgnoreRhosts directive to appear anywhere in a\n sshd_config, not just before any Match blocks; bz3148\n\n * ssh(1): add %TOKEN percent expansion for the LocalFoward and\n RemoteForward keywords when used for Unix domain socket forwarding.\n bz#3014\n\n * all: allow loading public keys from the unencrypted envelope of a\n private key file if no corresponding public key file is present.\n \n * ssh(1), sshd(8): prefer to use chacha20 from libcrypto where\n possible instead of the (slower) portable C implementation included\n in OpenSSH.\n\n * ssh-keygen(1): add ability to dump the contents of a binary key\n revocation list via "ssh-keygen -lQf /path" bz#3132\n\nBugfixes\n--------\n\n * ssh(1): fix IdentitiesOnly=yes to also apply to keys loaded from\n a PKCS11Provider; bz#3141\n\n * ssh-keygen(1): avoid NULL dereference when trying to convert an\n invalid RFC4716 private key.\n\n * scp(1): when performing remote-to-remote copies using "scp -3",\n start the second ssh(1) channel with BatchMode=yes enabled to\n avoid confusing and non-deterministic ordering of prompts.\n\n * ssh(1), ssh-keygen(1): when signing a challenge using a FIDO token,\n perform hashing of the message to be signed in the middleware layer\n rather than in OpenSSH code. This permits the use of security key\n middlewares that perform the hashing implicitly, such as Windows\n Hello.\n\n * ssh(1): fix incorrect error message for "too many known hosts\n files." bz#3149\n\n * ssh(1): make failures when establishing "Tunnel" forwarding\n terminate the connection when ExitOnForwardFailure is enabled;\n bz#3116\n\n * ssh-keygen(1): fix printing of fingerprints on private keys and add\n a regression test for same.\n\n * sshd(8): document order of checking AuthorizedKeysFile (first) and\n AuthorizedKeysCommand (subsequently, if the file doesn't match);\n bz#3134\n\n * sshd(8): document that /etc/hosts.equiv and /etc/shosts.equiv are\n not considered for HostbasedAuthentication when the target user is\n root; bz#3148\n \n * ssh(1), ssh-keygen(1): fix NULL dereference in private certificate\n key parsing (oss-fuzz #20074).\n\n * ssh(1), sshd(8): more consistency between sets of %TOKENS are\n accepted in various configuration options.\n\n * ssh(1), ssh-keygen(1): improve error messages for some common\n PKCS#11 C_Login failure cases; bz#3130\n\n * ssh(1), sshd(8): make error messages for problems during SSH banner\n exchange consistent with other SSH transport-layer error messages\n and ensure they include the relevant IP addresses bz#3129\n\n * various: fix a number of spelling errors in comments and debug/error\n messages\n\n * ssh-keygen(1), ssh-add(1): when downloading FIDO2 resident keys\n from a token, don't prompt for a PIN until the token has told us\n that it needs one. Avoids double-prompting on devices that\n implement on-device authentication.\n\n * sshd(8), ssh-keygen(1): no-touch-required FIDO certificate option\n should be an extension, not a critical option.\n \n * ssh(1), ssh-keygen(1), ssh-add(1): offer a better error message\n when trying to use a FIDO key function and SecurityKeyProvider is\n empty.\n\n * ssh-add(1), ssh-agent(8): ensure that a key lifetime fits within\n the values allowed by the wire format (u32). Prevents integer\n wraparound of the timeout values. bz#3119\n\n * ssh(1): detect and prevent trivial configuration loops when using\n ProxyJump. bz#3057.\n \nPortability\n-----------\n\n * Detect systems where signals flagged with SA_RESTART will interrupt\n select(2). POSIX permits implementations to choose whether\n select(2) will return when interrupted with a SA_RESTART-flagged\n signal, but OpenSSH requires interrupting behaviour.\n\n * Several compilation fixes for HP/UX and AIX.\n\n * On platforms that do not support setting process-wide routing\n domains (all excepting OpenBSD at present), fail to accept a\n configuration attempts to set one at process start time rather than\n fatally erroring at run time. bz#3126\n\n * Improve detection of egrep (used in regression tests) on platforms\n that offer a poor default one (e.g. Solaris).\n\n * A number of shell portability fixes for the regression tests.\n\n * Fix theoretical infinite loop in the glob(3) replacement\n implementation.\n\n * Fix seccomp sandbox compilation problems for some Linux\n configurations bz#3085\n\n * Improved detection of libfido2 and some compilation fixes for some\n configurations when --with-security-key-builtin is selected.\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.3.tar.gz) = 46c63b7ddbe46a0666222f7988c993866c31fcca\n - SHA256 (openssh-8.3.tar.gz) = M6CnZ+duGs4bzDio8hQNLwyLQChV+3wkUEO8HWLV35c=\n\n - SHA1 (/openssh-8.3p1.tar.gz) = 04c7adb9986f16746588db8988b910530c589819\n - SHA256 (openssh-8.3p1.tar.gz) = 8r774Ecv5+t10jNA6xdTHLazqsJAdeIGa0H4FOEjh7I=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.2 was released on 2020-02-14. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttps://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nIt is now possible[1] to perform chosen-prefix attacks against the\nSHA-1 hash algorithm for less than USD$50K. For this reason, we will\nbe disabling the "ssh-rsa" public key signature algorithm that depends\non SHA-1 by default in a near-future release.\n\nThis algorithm is unfortunately still used widely despite the\nexistence of better alternatives, being the only remaining public key\nsignature algorithm specified by the original SSH RFCs.\n\nThe better alternatives include:\n\n * The RFC8332 RSA SHA-2 signature algorithms rsa-sha2-256/512. These\n algorithms have the advantage of using the same key type as\n "ssh-rsa" but use the safe SHA-2 hash algorithms. These have been\n supported since OpenSSH 7.2 and are already used by default if the\n client and server support them.\n\n * The ssh-ed25519 signature algorithm. It has been supported in\n OpenSSH since release 6.5.\n\n * The RFC5656 ECDSA algorithms: ecdsa-sha2-nistp256/384/521. These\n have been supported by OpenSSH since release 5.7.\n\nTo check whether a server is using the weak ssh-rsa public key\nalgorithm for host authentication, try to connect to it after\nremoving the ssh-rsa algorithm from ssh(1)'s allowed list:\n\n ssh -oHostKeyAlgorithms=-ssh-rsa user@host\n\nIf the host key verification fails and no other supported host key\ntypes are available, the server software on that host should be\nupgraded.\n\nA future release of OpenSSH will enable UpdateHostKeys by default\nto allow the client to automatically migrate to better algorithms.\nUsers may consider enabling this option manually.\n\n "SHA-1 is a Shambles: First Chosen-Prefix Collision on SHA-1 and\n Application to the PGP Web of Trust" Leurent, G and Peyrin, T\n (2020) https://eprint.iacr.org/2020/014.pdf\n\nSecurity\n========\n\n * ssh(1), sshd(8), ssh-keygen(1): this release removes the "ssh-rsa"\n (RSA/SHA1) algorithm from those accepted for certificate signatures\n (i.e. the client and server CASignatureAlgorithms option) and will\n use the rsa-sha2-512 signature algorithm by default when the\n ssh-keygen(1) CA signs new certificates.\n\n Certificates are at special risk to the aforementioned SHA1\n collision vulnerability as an attacker has effectively unlimited\n time in which to craft a collision that yields them a valid\n certificate, far more than the relatively brief LoginGraceTime\n window that they have to forge a host key signature.\n\n The OpenSSH certificate format includes a CA-specified (typically\n random) nonce value near the start of the certificate that should\n make exploitation of chosen-prefix collisions in this context\n challenging, as the attacker does not have full control over the\n prefix that actually gets signed. Nonetheless, SHA1 is now a\n demonstrably broken algorithm and further improvements in attacks\n are highly likely.\n\n OpenSSH releases prior to 7.2 do not support the newer RSA/SHA2\n algorithms and will refuse to accept certificates signed by an\n OpenSSH 8.2+ CA using RSA keys unless the unsafe algorithm is\n explicitly selected during signing ("ssh-keygen -t ssh-rsa").\n Older clients/servers may use another CA key type such as\n ssh-ed25519 (supported since OpenSSH 6.5) or one of the\n ecdsa-sha2-nistp256/384/521 types (supported since OpenSSH 5.7)\n instead if they cannot be upgraded.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh(1), sshd(8): the above removal of "ssh-rsa" from the accepted\n CASignatureAlgorithms list.\n\n * ssh(1), sshd(8): this release removes diffie-hellman-group14-sha1\n from the default key exchange proposal for both the client and\n server.\n\n * ssh-keygen(1): the command-line options related to the generation\n and screening of safe prime numbers used by the\n diffie-hellman-group-exchange-* key exchange algorithms have\n changed. Most options have been folded under the -O flag.\n\n * sshd(8): the sshd listener process title visible to ps(1) has\n changed to include information about the number of connections that\n are currently attempting authentication and the limits configured\n by MaxStartups.\n\n * ssh-sk-helper(8): this is a new binary. It is used by the FIDO/U2F\n support to provide address-space isolation for token middleware\n libraries (including the internal one). It needs to be installed\n in the expected path, typically under /usr/libexec or similar.\n\nChanges since OpenSSH 8.1\n=========================\n\nThis release contains some significant new features.\n\nFIDO/U2F Support\n----------------\n\nThis release adds support for FIDO/U2F hardware authenticators to\nOpenSSH. U2F/FIDO are open standards for inexpensive two-factor\nauthentication hardware that are widely used for website\nauthentication. In OpenSSH FIDO devices are supported by new public\nkey types "ecdsa-sk" and "ed25519-sk", along with corresponding\ncertificate types.\n\nssh-keygen(1) may be used to generate a FIDO token-backed key, after\nwhich they may be used much like any other key type supported by\nOpenSSH, so long as the hardware token is attached when the keys are\nused. FIDO tokens also generally require the user explicitly authorise\noperations by touching or tapping them.\n\nGenerating a FIDO key requires the token be attached, and will usually\nrequire the user tap the token to confirm the operation:\n\n $ ssh-keygen -t ecdsa-sk -f ~/.ssh/id_ecdsa_sk\n Generating public/private ecdsa-sk key pair.\n You may need to touch your security key to authorize key generation.\n Enter file in which to save the key (/home/djm/.ssh/id_ecdsa_sk): \n Enter passphrase (empty for no passphrase): \n Enter same passphrase again: \n Your identification has been saved in /home/djm/.ssh/id_ecdsa_sk\n Your public key has been saved in /home/djm/.ssh/id_ecdsa_sk.pub\n\nThis will yield a public and private key-pair. The private key file\nshould be useless to an attacker who does not have access to the\nphysical token. After generation, this key may be used like any other\nsupported key in OpenSSH and may be listed in authorized_keys, added\nto ssh-agent(1), etc. The only additional stipulation is that the FIDO\ntoken that the key belongs to must be attached when the key is used.\n\nFIDO tokens are most commonly connected via USB but may be attached\nvia other means such as Bluetooth or NFC. In OpenSSH, communication\nwith the token is managed via a middleware library, specified by the\nSecurityKeyProvider directive in ssh/sshd_config(5) or the\n$SSH_SK_PROVIDER environment variable for ssh-keygen(1) and\nssh-add(1). The API for this middleware is documented in the sk-api.h\nand PROTOCOL.u2f files in the source distribution.\n\nOpenSSH includes a middleware ("SecurityKeyProvider=internal") with\nsupport for USB tokens. It is automatically enabled in OpenBSD and may\nbe enabled in portable OpenSSH via the configure flag\n--with-security-key-builtin. If the internal middleware is enabled\nthen it is automatically used by default. This internal middleware\nrequires that libfido2 (https://github.com/Yubico/libfido2)and its\ndependencies be installed. We recommend that packagers of portable\nOpenSSH enable the built-in middleware, as it provides the\nlowest-friction experience for users.\n\nNote: FIDO/U2F tokens are required to implement the ECDSA-P256\n"ecdsa-sk" key type, but hardware support for Ed25519 "ed25519-sk" is\nless common. Similarly, not all hardware tokens support some of the\noptional features such as resident keys.\n\nThe protocol-level changes to support FIDO/U2F keys in SSH are\ndocumented in the PROTOCOL.u2f file in the OpenSSH source\ndistribution.\n\nThere are a number of supporting changes to this feature:\n\n * ssh-keygen(1): add a "no-touch-required" option when generating\n FIDO-hosted keys, that disables their default behaviour of\n requiring a physical touch/tap on the token during authentication.\n Note: not all tokens support disabling the touch requirement.\n\n * sshd(8): add a sshd_config PubkeyAuthOptions directive that\n collects miscellaneous public key authentication-related options\n for sshd(8). At present it supports only a single option\n "no-touch-required". This causes sshd to skip its default check for\n FIDO/U2F keys that the signature was authorised by a touch or press\n event on the token hardware.\n\n * ssh(1), sshd(8), ssh-keygen(1): add a "no-touch-required" option\n for authorized_keys and a similar extension for certificates. This\n option disables the default requirement that FIDO key signatures\n attest that the user touched their key to authorize them, mirroring\n the similar PubkeyAuthOptions sshd_config option.\n\n * ssh-keygen(1): add support for the writing the FIDO attestation\n information that is returned when new keys are generated via the\n "-O write-attestation=/path" option. FIDO attestation certificates\n may be used to verify that a FIDO key is hosted in trusted\n hardware. OpenSSH does not currently make use of this information,\n beyond optionally writing it to disk.\n\nFIDO2 resident keys\n-------------------\n\nFIDO/U2F OpenSSH keys consist of two parts: a "key handle" part stored\nin the private key file on disk, and a per-device private key that is\nunique to each FIDO/U2F token and that cannot be exported from the\ntoken hardware. These are combined by the hardware at authentication\ntime to derive the real key that is used to sign authentication\nchallenges.\n\nFor tokens that are required to move between computers, it can be\ncumbersome to have to move the private key file first. To avoid this\nrequirement, tokens implementing the newer FIDO2 standard support\n"resident keys", where it is possible to effectively retrieve the key\nhandle part of the key from the hardware.\n\nOpenSSH supports this feature, allowing resident keys to be generated\nusing the ssh-keygen(1) "-O resident" flag. This will produce a\npublic/private key pair as usual, but it will be possible to retrieve\nthe private key part from the token later. This may be done using\n"ssh-keygen -K", which will download all available resident keys from\nthe tokens attached to the host and write public/private key files\nfor them. It is also possible to download and add resident keys\ndirectly to ssh-agent(1) without writing files to the file-system\nusing "ssh-add -K".\n\nResident keys are indexed on the token by the application string and\nuser ID. By default, OpenSSH uses an application string of "ssh:" and\nan empty user ID. If multiple resident keys on a single token are\ndesired then it may be necessary to override one or both of these\ndefaults using the ssh-keygen(1) "-O application=" or "-O user="\noptions. Note: OpenSSH will only download and use resident keys whose\napplication string begins with "ssh:"\n\nStoring both parts of a key on a FIDO token increases the likelihood\nof an attacker being able to use a stolen token device. For this\nreason, tokens should enforce PIN authentication before allowing\ndownload of keys, and users should set a PIN on their tokens before\ncreating any resident keys.\n\nOther New Features\n------------------\n\n * sshd(8): add an Include sshd_config keyword that allows including\n additional configuration files via glob(3) patterns. bz2468\n\n * ssh(1)/sshd(8): make the LE (low effort) DSCP code point available\n via the IPQoS directive; bz2986,\n\n * ssh(1): when AddKeysToAgent=yes is set and the key contains no\n comment, add the key to the agent with the key's path as the\n comment. bz2564\n \n * ssh-keygen(1), ssh-agent(1): expose PKCS#11 key labels and X.509\n subjects as key comments, rather than simply listing the PKCS#11\n provider library path. PR138\n\n * ssh-keygen(1): allow PEM export of DSA and ECDSA keys; bz3091\n\n * ssh(1), sshd(8): make zlib compile-time optional, available via the\n Makefile.inc ZLIB flag on OpenBSD or via the --with-zlib configure\n option for OpenSSH portable.\n\n * sshd(8): when clients get denied by MaxStartups, send a\n notification prior to the SSH2 protocol banner according to\n RFC4253 section 4.2.\n\n * ssh(1), ssh-agent(1): when invoking the $SSH_ASKPASS prompt\n program, pass a hint to the program to describe the type of\n desired prompt. The possible values are "confirm" (indicating\n that a yes/no confirmation dialog with no text entry should be\n shown), "none" (to indicate an informational message only), or\n blank for the original ssh-askpass behaviour of requesting a\n password/phrase.\n\n * ssh(1): allow forwarding a different agent socket to the path\n specified by $SSH_AUTH_SOCK, by extending the existing ForwardAgent\n option to accepting an explicit path or the name of an environment\n variable in addition to yes/no.\n \n * ssh-keygen(1): add a new signature operations "find-principals" to\n look up the principal associated with a signature from an allowed-\n signers file.\n \n * sshd(8): expose the number of currently-authenticating connections\n along with the MaxStartups limit in the process title visible to\n "ps".\n\nBugfixes\n--------\n\n * sshd(8): make ClientAliveCountMax=0 have sensible semantics: it\n will now disable connection killing entirely rather than the\n current behaviour of instantly killing the connection after the\n first liveness test regardless of success. bz2627\n \n * sshd(8): clarify order of AllowUsers / DenyUsers vs AllowGroups /\n DenyGroups in the sshd(8) manual page. bz1690\n\n * sshd(8): better describe HashKnownHosts in the manual page. bz2560\n\n * sshd(8): clarify that permitopen=/PermitOpen do no name or\n address translation in the manual page. bz3099\n\n * sshd(8): allow the UpdateHostKeys feature to function when\n multiple known_hosts files are in use. When updating host keys,\n ssh will now search subsequent known_hosts files, but will add\n updated host keys to the first specified file only. bz2738\n \n * All: replace all calls to signal(2) with a wrapper around\n sigaction(2). This wrapper blocks all other signals during the\n handler preventing races between handlers, and sets SA_RESTART\n which should reduce the potential for short read/write operations.\n \n * sftp(1): fix a race condition in the SIGCHILD handler that could\n turn in to a kill(-1); bz3084\n\n * sshd(8): fix a case where valid (but extremely large) SSH channel\n IDs were being incorrectly rejected. bz3098\n\n * ssh(1): when checking host key fingerprints as answers to new\n hostkey prompts, ignore whitespace surrounding the fingerprint\n itself.\n\n * All: wait for file descriptors to be readable or writeable during\n non-blocking connect, not just readable. Prevents a timeout when\n the server doesn't immediately send a banner (e.g. multiplexers\n like sslh)\n \n * sshd_config(5): document the sntrup4591761x25519-sha512@tinyssh.org\n key exchange algorithm. PR#151\n\nPortability\n-----------\n\n * sshd(8): multiple adjustments to the Linux seccomp sandbox:\n - Non-fatally deny IPC syscalls in sandbox\n - Allow clock_gettime64() in sandbox (MIPS / glibc >= 2.31)\n - Allow clock_nanosleep_time64 in sandbox (ARM) bz3100\n - Allow clock_nanosleep() in sandbox (recent glibc) bz3093\n\n * Explicit check for memmem declaration and fix up declaration if the\n system headers lack it. bz3102\n \nChecksums:\n==========\n\n - SHA1 (openssh-8.2.tar.gz) = 0daae2a8c47c489a8784f2c38c4b39e6159ba678\n - SHA256 (openssh-8.2.tar.gz) = +UmInEIoHJqYqWneMb/kgRbLcq8WDCo7+ooYcjzW4jg=\n\n - SHA1 (openssh-8.2p1.tar.gz) = d1ab35a93507321c5db885e02d41ce1414f0507c\n - SHA256 (openssh-8.2p1.tar.gz) = Q5JRUebPbO4UUBkMDpr03Da0HBJzdhnt/4vOvf9k5nE=\n\nNote: the openssh-8.2 tarball for OpenBSD that was initially released\nadvertised an incorrect version for "ssh -V" and the sshd server\nbanner. The above tarball replace the incorrect release, which has\nbeen renamed to openssh-8.2.tar.gz.incorrect. These are the checksums\nfor the original, incorrect tarball:\n\n - SHA1 (openssh-8.2.tar.gz) = 77584c22fbb89269398acdf53c1e554400584ba8\n - SHA256 (openssh-8.2.tar.gz) = UttLaaSYXVK1O65cYvyQzyQ5sCfuJ4Lwrs8zNsPrluQ=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read https://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.1 was released on 2019-10-09. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nSecurity\n========\n\n * ssh(1), sshd(8), ssh-add(1), ssh-keygen(1): an exploitable integer\n overflow bug was found in the private key parsing code for the XMSS\n key type. This key type is still experimental and support for it is\n not compiled by default. No user-facing autoconf option exists in\n portable OpenSSH to enable it. This bug was found by Adam Zabrocki\n and reported via SecuriTeam's SSD program.\n\n * ssh(1), sshd(8), ssh-agent(1): add protection for private keys at\n rest in RAM against speculation and memory side-channel attacks like\n Spectre, Meltdown and Rambleed. This release encrypts private keys\n when they are not in use with a symmetric key that is derived from a\n relatively large "prekey" consisting of random data (currently 16KB).\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh-keygen(1): when acting as a CA and signing certificates with\n an RSA key, default to using the rsa-sha2-512 signature algorithm.\n Certificates signed by RSA keys will therefore be incompatible\n with OpenSSH versions prior to 7.2 unless the default is\n overridden (using "ssh-keygen -t ssh-rsa -s ...").\n\nChanges since OpenSSH 8.0\n=========================\n\nThis release is focused on bug-fixing.\n\nNew Features\n------------\n\n * ssh(1): Allow %n to be expanded in ProxyCommand strings\n\n * ssh(1), sshd(8): Allow prepending a list of algorithms to the\n default set by starting the list with the '^' character, E.g.\n "HostKeyAlgorithms ^ssh-ed25519"\n\n * ssh-keygen(1): add an experimental lightweight signature and\n verification ability. Signatures may be made using regular ssh keys\n held on disk or stored in a ssh-agent and verified against an\n authorized_keys-like list of allowed keys. Signatures embed a\n namespace that prevents confusion and attacks between different\n usage domains (e.g. files vs email).\n\n * ssh-keygen(1): print key comment when extracting public key from a\n private key. bz#3052\n\n * ssh-keygen(1): accept the verbose flag when searching for host keys\n in known hosts (i.e. "ssh-keygen -vF host") to print the matching\n host's random-art signature too. bz#3003\n\n * All: support PKCS8 as an optional format for storage of private\n keys to disk. The OpenSSH native key format remains the default,\n but PKCS8 is a superior format to PEM if interoperability with\n non-OpenSSH software is required, as it may use a less insecure\n key derivation function than PEM's.\n\nBugfixes\n--------\n\n * ssh(1): if a PKCS#11 token returns no keys then try to login and\n refetch them. Based on patch from Jakub Jelen; bz#2430\n\n * ssh(1): produce a useful error message if the user's shell is set\n incorrectly during "match exec" processing. bz#2791\n\n * sftp(1): allow the maximum uint32 value for the argument passed\n to -b which allows better error messages from later validation.\n bz#3050\n\n * ssh(1): avoid pledge sandbox violations in some combinations of\n remote forwarding, connection multiplexing and ControlMaster.\n\n * ssh-keyscan(1): include SHA2-variant RSA key algorithms in KEX\n proposal; allows ssh-keyscan to harvest keys from servers that\n disable old SHA1 ssh-rsa. bz#3029\n\n * sftp(1): print explicit "not modified" message if a file was\n requested for resumed download but was considered already complete.\n bz#2978\n\n * sftp(1): fix a typo and make <esc><right> move right to the\n closest end of a word just like <esc><left> moves left to the\n closest beginning of a word.\n\n * sshd(8): cap the number of permitopen/permitlisten directives\n allowed to appear on a single authorized_keys line.\n\n * All: fix a number of memory leaks (one-off or on exit paths).\n\n * Regression tests: a number of fixes and improvements, including\n fixes to the interop tests, adding the ability to run most tests\n on builds that disable OpenSSL support, better support for running\n tests under Valgrind and a number of bug-fixes.\n\n * ssh(1), sshd(8): check for convtime() refusing to accept times that\n resolve to LONG_MAX Reported by Kirk Wolf bz2977\n\n * ssh(1): slightly more instructive error message when the user\n specifies multiple -J options on the command-line. bz3015\n\n * ssh-agent(1): process agent requests for RSA certificate private\n keys using correct signature algorithm when requested. bz3016\n\n * sftp(1): check for user@host when parsing sftp target. This\n allows user@[1.2.3.4] to work without a path. bz#2999\n\n * sshd(8): enlarge format buffer size for certificate serial\n number so the log message can record any 64-bit integer without\n truncation. bz#3012\n\n * sshd(8): for PermitOpen violations add the remote host and port to\n be able to more easily ascertain the source of the request. Add the\n same logging for PermitListen violations which where not previously\n logged at all.\n\n * scp(1), sftp(1): use the correct POSIX format style for left\n justification for the transfer progress meter. bz#3002\n\n * sshd(8) when examining a configuration using sshd -T, assume any\n attribute not provided by -C does not match, which allows it to work\n when sshd_config contains a Match directive with or without -C.\n bz#2858\n\n * ssh(1), ssh-keygen(1): downgrade PKCS#11 "provider returned no\n slots" warning from log level error to debug. This is common when\n attempting to enumerate keys on smartcard readers with no cards\n plugged in. bz#3058\n\n * ssh(1), ssh-keygen(1): do not unconditionally log in to PKCS#11\n tokens. Avoids spurious PIN prompts for keys not selected for\n authentication in ssh(1) and when listing public keys available in\n a token using ssh-keygen(1). bz#3006\n\nPortability\n-----------\n\n * ssh(1): fix SIGWINCH delivery of Solaris for multiplexed sessions\n bz#3030\n\n * ssh(1), sshd(8): fix typo that prevented detection of Linux VRF\n\n * sshd(8): add no-op implementation of pam_putenv to avoid build\n breakage on platforms where the PAM implementation lacks this\n function (e.g. HP-UX). bz#3008\n\n * sftp-server(8): fix Solaris privilege sandbox from preventing\n the legacy sftp rename operation from working (was refusing to\n allow hard links to files owned by other users). bz#3036\n\n * All: add a proc_pidinfo()-based closefrom() for OS X to avoid\n the need to brute-force close all high-numbered file descriptors.\n bz#3049\n\n * sshd(8): in the Linux seccomp-bpf sandbox, allow mprotect(2) with\n PROT_(READ|WRITE|NONE) only. This syscall is used by some hardened\n heap allocators. Github PR#142\n\n * sshd(8): in the Linux seccomp-bpf sandbox, allow the s390-specific\n ioctl for ECC hardware support.\n\n * All: use "doc" man page format if the mandoc(1) tool is present on\n the system. Previously configure would not select the "doc" man\n page format if mandoc was present but nroff was not.\n\n * sshd(8): don't install duplicate STREAMS modules on Solaris; check\n if STREAMS modules are already installed on a pty before installing\n since when compiling with XPG>=4 they will likely be installed\n already. Prevents hangs and duplicate lines on the terminal.\n bz#2945 and bz#2998,\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.1.tar.gz) = bf7b0c65a7c0afa5ba9c787f345b8a24fa459add\n - SHA256 (openssh-8.1.tar.gz) = vamkKxZTFfgxQXSxGeJ1vbuot0H3Vx9bNBgrvChSrFg=\n\n - SHA1 (openssh-8.1p1.tar.gz) = c44b96094869f177735ae053d92bd5fcab1319de\n - SHA256 (openssh-8.1p1.tar.gz) = AvXb7zg10HU1VvlzzVe0wZtrH2zSTANEXiOsd8obk/8=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 8.0 was released on 2019-04-17. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nSecurity\n========\n\nThis release contains mitigation for a weakness in the scp(1) tool\nand protocol (CVE-2019-6111): when copying files from a remote system\nto a local directory, scp(1) did not verify that the filenames that\nthe server sent matched those requested by the client. This could\nallow a hostile server to create or clobber unexpected local files\nwith attacker-controlled content.\n\nThis release adds client-side checking that the filenames sent from\nthe server match the command-line request,\n\nThe scp protocol is outdated, inflexible and not readily fixed. We\nrecommend the use of more modern protocols like sftp and rsync for\nfile transfer instead.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * scp(1): Relating to the above changes to scp(1); the scp protocol\n relies on the remote shell for wildcard expansion, so there is no\n infallible way for the client's wildcard matching to perfectly\n reflect the server's. If there is a difference between client and\n server wildcard expansion, the client may refuse files from the\n server. For this reason, we have provided a new "-T" flag to scp\n that disables these client-side checks at the risk of\n reintroducing the attack described above.\n\n * sshd(8): Remove support for obsolete "host/port" syntax. Slash-\n separated host/port was added in 2001 as an alternative to\n host:port syntax for the benefit of IPv6 users. These days there\n are established standards for this like [::1]:22 and the slash\n syntax is easily mistaken for CIDR notation, which OpenSSH\n supports for some things. Remove the slash notation from\n ListenAddress and PermitOpen; bz#2335\n\nChanges since OpenSSH 7.9\n=========================\n\nThis release is focused on new features and internal refactoring.\n\nNew Features\n------------\n\n * ssh(1), ssh-agent(1), ssh-add(1): Add support for ECDSA keys in\n PKCS#11 tokens.\n\n * ssh(1), sshd(8): Add experimental quantum-computing resistant\n key exchange method, based on a combination of Streamlined NTRU\n Prime 4591^761 and X25519.\n\n * ssh-keygen(1): Increase the default RSA key size to 3072 bits,\n following NIST Special Publication 800-57's guidance for a\n 128-bit equivalent symmetric security level.\n\n * ssh(1): Allow "PKCS11Provider=none" to override later instances of\n the PKCS11Provider directive in ssh_config; bz#2974\n\n * sshd(8): Add a log message for situations where a connection is\n dropped for attempting to run a command but a sshd_config\n ForceCommand=internal-sftp restriction is in effect; bz#2960\n\n * ssh(1): When prompting whether to record a new host key, accept\n the key fingerprint as a synonym for "yes". This allows the user\n to paste a fingerprint obtained out of band at the prompt and\n have the client do the comparison for you.\n\n * ssh-keygen(1): When signing multiple certificates on a single\n command-line invocation, allow automatically incrementing the\n certificate serial number.\n\n * scp(1), sftp(1): Accept -J option as an alias to ProxyJump on\n the scp and sftp command-lines.\n\n * ssh-agent(1), ssh-pkcs11-helper(8), ssh-add(1): Accept "-v"\n command-line flags to increase the verbosity of output; pass\n verbose flags though to subprocesses, such as ssh-pkcs11-helper\n started from ssh-agent.\n\n * ssh-add(1): Add a "-T" option to allowing testing whether keys in\n an agent are usable by performing a signature and a verification.\n\n * sftp-server(8): Add a "lsetstat@openssh.com" protocol extension\n that replicates the functionality of the existing SSH2_FXP_SETSTAT\n operation but does not follow symlinks. bz#2067\n\n * sftp(1): Add "-h" flag to chown/chgrp/chmod commands to request\n they do not follow symlinks.\n\n * sshd(8): Expose $SSH_CONNECTION in the PAM environment. This makes\n the connection 4-tuple available to PAM modules that wish to use\n it in decision-making. bz#2741\n\n * sshd(8): Add a ssh_config "Match final" predicate Matches in same\n pass as "Match canonical" but doesn't require hostname\n canonicalisation be enabled. bz#2906\n\n * sftp(1): Support a prefix of '@' to suppress echo of sftp batch\n commands; bz#2926\n\n * ssh-keygen(1): When printing certificate contents using\n "ssh-keygen -Lf /path/certificate", include the algorithm that\n the CA used to sign the cert.\n\nBugfixes\n--------\n\n * sshd(8): Fix authentication failures when sshd_config contains\n "AuthenticationMethods any" inside a Match block that overrides\n a more restrictive default.\n\n * sshd(8): Avoid sending duplicate keepalives when ClientAliveCount\n is enabled.\n\n * sshd(8): Fix two race conditions related to SIGHUP daemon restart.\n Remnant file descriptors in recently-forked child processes could\n block the parent sshd's attempt to listen(2) to the configured\n addresses. Also, the restarting parent sshd could exit before any\n child processes that were awaiting their re-execution state had\n completed reading it, leaving them in a fallback path.\n\n * ssh(1): Fix stdout potentially being redirected to /dev/null when\n ProxyCommand=- was in use.\n\n * sshd(8): Avoid sending SIGPIPE to child processes if they attempt\n to write to stderr after their parent processes have exited;\n bz#2071\n\n * ssh(1): Fix bad interaction between the ssh_config ConnectTimeout\n and ConnectionAttempts directives - connection attempts after the\n first were ignoring the requested timeout; bz#2918\n\n * ssh-keyscan(1): Return a non-zero exit status if no keys were\n found; bz#2903\n\n * scp(1): Sanitize scp filenames to allow UTF-8 characters without\n terminal control sequences; bz#2434\n\n * sshd(8): Fix confusion between ClientAliveInterval and time-based\n RekeyLimit that could cause connections to be incorrectly closed.\n bz#2757\n\n * ssh(1), ssh-add(1): Correct some bugs in PKCS#11 token PIN\n handling at initial token login. The attempt to read the PIN\n could be skipped in some cases, particularly on devices with\n integrated PIN readers. This would lead to an inability to\n retrieve keys from these tokens. bz#2652\n\n * ssh(1), ssh-add(1): Support keys on PKCS#11 tokens that set the\n CKA_ALWAYS_AUTHENTICATE flag by requiring a fresh login after the\n C_SignInit operation. bz#2638\n\n * ssh(1): Improve documentation for ProxyJump/-J, clarifying that\n local configuration does not apply to jump hosts.\n\n * ssh-keygen(1): Clarify manual - ssh-keygen -e only writes\n public keys, not private.\n\n * ssh(1), sshd(8): be more strict in processing protocol banners,\n allowing \\r characters only immediately before \\n.\n\n * Various: fix a number of memory leaks, including bz#2942 and\n bz#2938\n\n * scp(1), sftp(1): fix calculation of initial bandwidth limits.\n Account for bytes written before the timer starts and adjust the\n schedule on which recalculations are performed. Avoids an initial\n burst of traffic and yields more accurate bandwidth limits;\n bz#2927\n\n * sshd(8): Only consider the ext-info-c extension during the initial\n key eschange. It shouldn't be sent in subsequent ones, but if it\n is present we should ignore it. This prevents sshd from sending a\n SSH_MSG_EXT_INFO for REKEX for buggy these clients. bz#2929\n\n * ssh-keygen(1): Clarify manual that ssh-keygen -F (find host in \n authorized_keys) and -R (remove host from authorized_keys) options\n may accept either a bare hostname or a [hostname]:port combo.\n bz#2935\n\n * ssh(1): Don't attempt to connect to empty SSH_AUTH_SOCK; bz#2936\n\n * sshd(8): Silence error messages when sshd fails to load some of\n the default host keys. Failure to load an explicitly-configured\n hostkey is still an error, and failure to load any host key is\n still fatal. pr/103\n\n * ssh(1): Redirect stderr of ProxyCommands to /dev/null when ssh is\n started with ControlPersist; prevents random ProxyCommand output\n from interfering with session output.\n\n * ssh(1): The ssh client was keeping a redundant ssh-agent socket\n (leftover from authentication) around for the life of the\n connection; bz#2912\n\n * sshd(8): Fix bug in HostbasedAcceptedKeyTypes and\n PubkeyAcceptedKeyTypes options. If only RSA-SHA2 signature types\n were specified, then authentication would always fail for RSA keys\n as the monitor checks only the base key (not the signature\n algorithm) type against *AcceptedKeyTypes. bz#2746\n\n * ssh(1): Request correct signature types from ssh-agent when\n certificate keys and RSA-SHA2 signatures are in use.\n\nPortability\n-----------\n\n * sshd(8): On Cygwin, run as SYSTEM where possible, using S4U for\n token creation if it supports MsV1_0 S4U Logon.\n\n * sshd(8): On Cygwin, use custom user/group matching code that\n respects the OS' behaviour of case-insensitive matching.\n\n * sshd(8): Don't set $MAIL if UsePAM=yes as PAM typically specifies\n the user environment if it's enabled; bz#2937\n\n * sshd(8) Cygwin: Change service name to cygsshd to avoid collision\n with Microsoft's OpenSSH port.\n\n * Allow building against OpenSSL -dev (3.x)\n\n * Fix a number of build problems against version configurations and\n versions of OpenSSL. Including bz#2931 and bz#2921\n\n * Improve warnings in cygwin service setup. bz#2922\n\n * Remove hardcoded service name in cygwin setup. bz#2922\n\nChecksums:\n==========\n\n - SHA1 (openssh-8.0.tar.gz) = 8aaa99091fc7e5a92a4a320e1e5521046b3f95f0\n - SHA256 (openssh-8.0.tar.gz) = 1xvSJk1KYSnOLPYEUzyCVwTEQ7MHOaCO65DzeNuuLdo=\n\n - SHA1 (openssh-8.0p1.tar.gz) = 756dbb99193f9541c9206a667eaa27b0fa184a4f\n - SHA256 (openssh-8.0p1.tar.gz) = vZQ4eeaUmOgDHra39E0IzcN9WaeraJqgtDcyDDSB/Wg=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 7.9 was released on 2018-10-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh(1), sshd(8): the setting of the new CASignatureAlgorithms\n option (see below) bans the use of DSA keys as certificate\n authorities.\n\n * sshd(8): the authentication success/failure log message has\n changed format slightly. It now includes the certificate\n fingerprint (previously it included only key ID and CA key\n fingerprint).\n\nChanges since OpenSSH 7.8\n=========================\n\nThis is primarily a bugfix release.\n\nNew Features\n------------\n\n * ssh(1), sshd(8): allow most port numbers to be specified using\n service names from getservbyname(3) (typically /etc/services).\n\n * ssh(1): allow the IdentityAgent configuration directive to accept\n environment variable names. This supports the use of multiple\n agent sockets without needing to use fixed paths.\n\n * sshd(8): support signalling sessions via the SSH protocol.\n A limited subset of signals is supported and only for login or\n command sessions (i.e. not subsystems) that were not subject to\n a forced command via authorized_keys or sshd_config. bz#1424\n\n * ssh(1): support "ssh -Q sig" to list supported signature options.\n Also "ssh -Q help" to show the full set of supported queries.\n\n * ssh(1), sshd(8): add a CASignatureAlgorithms option for the\n client and server configs to allow control over which signature\n formats are allowed for CAs to sign certificates. For example,\n this allows banning CAs that sign certificates using the RSA-SHA1\n signature algorithm.\n\n * sshd(8), ssh-keygen(1): allow key revocation lists (KRLs) to\n revoke keys specified by SHA256 hash.\n\n * ssh-keygen(1): allow creation of key revocation lists directly\n from base64-encoded SHA256 fingerprints. This supports revoking\n keys using only the information contained in sshd(8)\n authentication log messages.\n\nBugfixes\n--------\n\n * ssh(1), ssh-keygen(1): avoid spurious "invalid format" errors when\n attempting to load PEM private keys while using an incorrect\n passphrase. bz#2901\n\n * sshd(8): when a channel closed message is received from a client,\n close the stderr file descriptor at the same time stdout is\n closed. This avoids stuck processes if they were waiting for\n stderr to close and were insensitive to stdin/out closing. bz#2863\n\n * ssh(1): allow ForwardX11Timeout=0 to disable the untrusted X11\n forwarding timeout and support X11 forwarding indefinitely.\n Previously the behaviour of ForwardX11Timeout=0 was undefined.\n\n * sshd(8): when compiled with GSSAPI support, cache supported method\n OIDs regardless of whether GSSAPI authentication is enabled in the\n main section of sshd_config. This avoids sandbox violations if\n GSSAPI authentication was later enabled in a Match block. bz#2107\n\n * sshd(8): do not fail closed when configured with a text key\n revocation list that contains a too-short key. bz#2897\n\n * ssh(1): treat connections with ProxyJump specified the same as\n ones with a ProxyCommand set with regards to hostname\n canonicalisation (i.e. don't try to canonicalise the hostname\n unless CanonicalizeHostname is set to 'always'). bz#2896\n\n * ssh(1): fix regression in OpenSSH 7.8 that could prevent public-\n key authentication using certificates hosted in a ssh-agent(1)\n or against sshd(8) from OpenSSH <7.8.\n\nPortability\n-----------\n\n * All: support building against the openssl-1.1 API (releases 1.1.0g\n and later). The openssl-1.0 API will remain supported at least\n until OpenSSL terminates security patch support for that API version.\n\n * sshd(8): allow the futex(2) syscall in the Linux seccomp sandbox;\n apparently required by some glibc/OpenSSL combinations.\n\n * sshd(8): handle getgrouplist(3) returning more than\n _SC_NGROUPS_MAX groups. Some platforms consider this limit more\n as a guideline.\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.9.tar.gz) = 7c50a86b8f591decd172ed7f5527abc533098dec\n - SHA256 (openssh-7.9.tar.gz) = nSVigtHGn3+xKXRqpSnp4YOyEPPAb+pCHdWS9Eh/IPY=\n\n - SHA1 (openssh-7.9p1.tar.gz) = 993aceedea8ecabb1d0dd7293508a361891c4eaa\n - SHA256 (openssh-7.9p1.tar.gz) = a0s7oiU9hO03ccgFByjVl8kc/OiYcTvre2SjBbbxGq0=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 7.8 was released on 2018-08-24. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh-keygen(1): write OpenSSH format private keys by default\n instead of using OpenSSL's PEM format. The OpenSSH format,\n supported in OpenSSH releases since 2014 and described in the\n PROTOCOL.key file in the source distribution, offers substantially\n better protection against offline password guessing and supports\n key comments in private keys. If necessary, it is possible to write\n old PEM-style keys by adding "-m PEM" to ssh-keygen's arguments\n when generating or updating a key.\n\n * sshd(8): remove internal support for S/Key multiple factor\n authentication. S/Key may still be used via PAM or BSD auth.\n\n * ssh(1): remove vestigal support for running ssh(1) as setuid. This\n used to be required for hostbased authentication and the (long\n gone) rhosts-style authentication, but has not been necessary for\n a long time. Attempting to execute ssh as a setuid binary, or with\n uid != effective uid will now yield a fatal error at runtime.\n\n * sshd(8): the semantics of PubkeyAcceptedKeyTypes and the similar\n HostbasedAcceptedKeyTypes options have changed. These now specify\n signature algorithms that are accepted for their respective\n authentication mechanism, where previously they specified accepted\n key types. This distinction matters when using the RSA/SHA2\n signature algorithms "rsa-sha2-256", "rsa-sha2-512" and their\n certificate counterparts. Configurations that override these\n options but omit these algorithm names may cause unexpected\n authentication failures (no action is required for configurations\n that accept the default for these options).\n\n * sshd(8): the precedence of session environment variables has\n changed. ~/.ssh/environment and environment="..." options in\n authorized_keys files can no longer override SSH_* variables set\n implicitly by sshd.\n\n * ssh(1)/sshd(8): the default IPQoS used by ssh/sshd has changed.\n They will now use DSCP AF21 for interactive traffic and CS1 for\n bulk. For a detailed rationale, please see the commit message:\n https://cvsweb.openbsd.org/log/src/usr.bin/ssh/readconf.c,v#rev1.284\n\nChanges since OpenSSH 7.7\n=========================\n\nThis is primarily a bugfix release.\n\nNew Features\n------------\n\n * ssh(1)/sshd(8): add new signature algorithms "rsa-sha2-256-cert-\n v01@openssh.com" and "rsa-sha2-512-cert-v01@openssh.com" to\n explicitly force use of RSA/SHA2 signatures in authentication.\n\n * sshd(8): extend the PermitUserEnvironment option to accept a\n whitelist of environment variable names in addition to global\n "yes" or "no" settings.\n\n * sshd(8): add a PermitListen directive to sshd_config(5) and a \n corresponding permitlisten= authorized_keys option that control\n which listen addresses and port numbers may be used by remote\n forwarding (ssh -R ...).\n\n * sshd(8): add some countermeasures against timing attacks used for\n account validation/enumeration. sshd will enforce a minimum time\n or each failed authentication attempt consisting of a global 5ms\n minimum plus an additional per-user 0-4ms delay derived from a\n host secret.\n\n * sshd(8): add a SetEnv directive to allow an administrator to\n explicitly specify environment variables in sshd_config.\n Variables set by SetEnv override the default and client-specified\n environment.\n\n * ssh(1): add a SetEnv directive to request that the server sets\n an environment variable in the session. Similar to the existing\n SendEnv option, these variables are set subject to server\n configuration.\n\n * ssh(1): allow "SendEnv -PATTERN" to clear environment variables\n previously marked for sending to the server. bz#1285\n\n * ssh(1)/sshd(8): make UID available as a %-expansion everywhere\n that the username is available currently. bz#2870\n\n * ssh(1): allow setting ProxyJump=none to disable ProxyJump\n functionality. bz#2869\n\nBugfixes\n--------\n\n * sshd(8): avoid observable differences in request parsing that could\n be used to determine whether a target user is valid.\n\n * all: substantial internal refactoring\n\n * ssh(1)/sshd(8): fix some memory leaks; bz#2366\n\n * ssh(1): fix a pwent clobber (introduced in openssh-7.7) that could\n occur during key loading, manifesting as crash on some platforms.\n\n * sshd_config(5): clarify documentation for AuthenticationMethods\n option; bz#2663\n\n * ssh(1): ensure that the public key algorithm sent in a\n public key SSH_MSG_USERAUTH_REQUEST matches the content of the\n signature blob. Previously, these could be inconsistent when a\n legacy or non-OpenSSH ssh-agent returned a RSA/SHA1 signature\n when asked to make a RSA/SHA2 signature.\n\n * sshd(8): fix failures to read authorized_keys caused by faulty\n supplemental group caching. bz#2873\n\n * scp(1): apply umask to directories, fixing potential mkdir/chmod\n race when copying directory trees bz#2839\n\n * ssh-keygen(1): return correct exit code when searching for and\n hashing known_hosts entries in a single operation; bz#2772\n\n * ssh(1): prefer the ssh binary pointed to via argv[0] to $PATH when\n re-executing ssh for ProxyJump. bz#2831\n\n * sshd(8): do not ban PTY allocation when a sshd session is\n restricted because the user password is expired as it breaks\n password change dialog. (regression in openssh-7.7).\n\n * ssh(1)/sshd(8): fix error reporting from select() failures. \n\n * ssh(1): improve documentation for -w (tunnel) flag, emphasising\n that -w implicitly sets Tunnel=point-to-point. bz#2365\n\n * ssh-agent(1): implement EMFILE mitigation for ssh-agent. ssh-agent\n will no longer spin when its file descriptor limit is exceeded.\n bz#2576\n\n * ssh(1)/sshd(8): disable SSH2_MSG_DEBUG messages for Twisted Conch\n clients. Twisted Conch versions that lack a version number in\n their identification strings will mishandle these messages when\n running on Python 2.x (https://twistedmatrix.com/trac/ticket/9422)\n * sftp(1): notify user immediately when underlying ssh process dies\n expectedly. bz#2719\n\n * ssh(1)/sshd(8): fix tunnel forwarding; regression in 7.7 release.\n bz#2855\n\n * ssh-agent(1): don't kill ssh-agent's listening socket entirely if\n it fails to accept(2) a connection. bz#2837\n\n * sshd(8): relax checking of authorized_keys environment="..."\n options to allow underscores in variable names (regression\n introduced in 7.7). bz#2851\n\n * ssh(1): add some missing options in the configuration dump output\n (ssh -G). bz#2835\n\nPortability\n-----------\n\n * sshd(8): Expose details of completed authentication to PAM auth\n modules via SSH_AUTH_INFO_0 in the PAM environment. bz#2408\n\n * Fix compilation problems caused by fights between zlib and OpenSSL\n colliding uses of "free_func"\n\n * Improve detection of unsupported compiler options. Recently these\n may have manifested as "unsupported -Wl,-z,retpoline" warnings\n during linking.\n\n * sshd(8): some sandbox support for Linux/s390 bz#2752.\n\n * regress tests: unbreak key-options.sh test on platforms without\n openpty(3). bz#2856\n\n * use getrandom(2) for PRNG seeding when built without OpenSSL.\n\nChecksums:\n==========\n\n- SHA1 (openssh-7.8.tar.gz) = ed5511cd42b543cd15166a9cbc56705f23b847e7\n- SHA256 (openssh-7.8.tar.gz) = TDqIsMEmghsBUNCrSCPyCxChfitntyOLXNC694py1XE\n\n- SHA1 (openssh-7.8p1.tar.gz) = 27e267e370315561de96577fccae563bc2c37a60\n- SHA256 (openssh-7.8p1.tar.gz) = GkhLsVFSwYO7JRThEqow3TQTjDz7Ay7uVJCmbFBxRMo\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 7.7 was released on 2018-04-02. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh(1)/sshd(8): Drop compatibility support for some very old SSH\n implementations, including ssh.com <=2.* and OpenSSH <= 3.*. These\n versions were all released in or before 2001 and predate the final\n SSH RFCs. The support in question isn't necessary for RFC-compliant\n SSH implementations.\n\nChanges since OpenSSH 7.6\n=========================\n\nThis is primarily a bugfix release.\n\nNew Features\n------------\n\n * All: Add experimental support for PQC XMSS keys (Extended Hash-\n Based Signatures) based on the algorithm described in\n https://tools.ietf.org/html/draft-irtf-cfrg-xmss-hash-based-signatures-12\n The XMSS signature code is experimental and not compiled in by\n default.\n\n * sshd(8): Add a "rdomain" criteria for the sshd_config Match keyword\n to allow conditional configuration that depends on which routing\n domain a connection was received on (currently supported on OpenBSD\n and Linux).\n\n * sshd_config(5): Add an optional rdomain qualifier to the\n ListenAddress directive to allow listening on different routing\n domains. This is supported only on OpenBSD and Linux at present.\n\n * sshd_config(5): Add RDomain directive to allow the authenticated\n session to be placed in an explicit routing domain. This is only\n supported on OpenBSD at present.\n\n * sshd(8): Add "expiry-time" option for authorized_keys files to\n allow for expiring keys.\n\n * ssh(1): Add a BindInterface option to allow binding the outgoing\n connection to an interface's address (basically a more usable\n BindAddress)\n\n * ssh(1): Expose device allocated for tun/tap forwarding via a new\n %T expansion for LocalCommand. This allows LocalCommand to be used\n to prepare the interface.\n\n * sshd(8): Expose the device allocated for tun/tap forwarding via a\n new SSH_TUNNEL environment variable. This allows automatic setup of\n the interface and surrounding network configuration automatically on\n the server.\n\n * ssh(1)/scp(1)/sftp(1): Add URI support to ssh, sftp and scp, e.g.\n ssh://user@host or sftp://user@host/path. Additional connection\n parameters described in draft-ietf-secsh-scp-sftp-ssh-uri-04 are not\n implemented since the ssh fingerprint format in the draft uses the\n deprecated MD5 hash with no way to specify the any other algorithm.\n\n * ssh-keygen(1): Allow certificate validity intervals that specify\n only a start or stop time (instead of both or neither).\n\n * sftp(1): Allow "cd" and "lcd" commands with no explicit path\n argument. lcd will change to the local user's home directory as\n usual. cd will change to the starting directory for session (because\n the protocol offers no way to obtain the remote user's home\n directory). bz#2760\n\n * sshd(8): When doing a config test with sshd -T, only require the\n attributes that are actually used in Match criteria rather than (an\n incomplete list of) all criteria.\n\nBugfixes\n--------\n\n * ssh(1)/sshd(8): More strictly check signature types during key\n exchange against what was negotiated. Prevents downgrade of RSA\n signatures made with SHA-256/512 to SHA-1.\n\n * sshd(8): Fix support for client that advertise a protocol version\n of "1.99" (indicating that they are prepared to accept both SSHv1 and\n SSHv2). This was broken in OpenSSH 7.6 during the removal of SSHv1\n support. bz#2810\n\n * ssh(1): Warn when the agent returns a ssh-rsa (SHA1) signature when\n a rsa-sha2-256/512 signature was requested. This condition is possible\n when an old or non-OpenSSH agent is in use. bz#2799\n\n * ssh-agent(1): Fix regression introduced in 7.6 that caused ssh-agent\n to fatally exit if presented an invalid signature request message.\n\n * sshd_config(5): Accept yes/no flag options case-insensitively, as\n has been the case in ssh_config(5) for a long time. bz#2664\n\n * ssh(1): Improve error reporting for failures during connection.\n Under some circumstances misleading errors were being shown. bz#2814\n\n * ssh-keyscan(1): Add -D option to allow printing of results directly\n in SSHFP format. bz#2821\n\n * regress tests: fix PuTTY interop test broken in last release's SSHv1\n removal. bz#2823\n\n * ssh(1): Compatibility fix for some servers that erroneously drop the\n connection when the IUTF8 (RFC8160) option is sent.\n\n * scp(1): Disable RemoteCommand and RequestTTY in the ssh session\n started by scp (sftp was already doing this.)\n\n * ssh-keygen(1): Refuse to create a certificate with an unusable\n number of principals.\n\n * ssh-keygen(1): Fatally exit if ssh-keygen is unable to write all the\n public key during key generation. Previously it would silently\n ignore errors writing the comment and terminating newline.\n\n * ssh(1): Do not modify hostname arguments that are addresses by\n automatically forcing them to lower-case. Instead canonicalise them\n to resolve ambiguities (e.g. ::0001 => ::1) before they are matched\n against known_hosts. bz#2763\n\n * ssh(1): Don't accept junk after "yes" or "no" responses to hostkey\n prompts. bz#2803\n\n * sftp(1): Have sftp print a warning about shell cleanliness when\n decoding the first packet fails, which is usually caused by shells\n polluting stdout of non-interactive startups. bz#2800\n\n * ssh(1)/sshd(8): Switch timers in packet code from using wall-clock\n time to monotonic time, allowing the packet layer to better function\n over a clock step and avoiding possible integer overflows during\n steps.\n\n * Numerous manual page fixes and improvements.\n\nPortability\n-----------\n\n * sshd(8): Correctly detect MIPS ABI in use at configure time. Fixes\n sandbox violations on some environments.\n\n * sshd(8): Remove UNICOS support. The hardware and software are literal\n museum pieces and support in sshd is too intrusive to justify\n maintaining.\n\n * All: Build and link with "retpoline" flags when available to mitigate\n the "branch target injection" style (variant 2) of the Spectre\n branch-prediction vulnerability.\n\n * All: Add auto-generated dependency information to Makefile.\n\n * Numerous fixed to the RPM spec files.\n\nChecksums:\n==========\n\n- SHA1 (openssh-7.7.tar.gz) = 24812e05fa233014c847c7775748316e7f8a836c\n- SHA256 (openssh-7.7.tar.gz) = T4ua1L/vgAYqwB0muRahvnm5ZUr3PLY9nPljaG8egvo=\n\n- SHA1 (openssh-7.7p1.tar.gz) = 446fe9ed171f289f0d62197dffdbfdaaf21c49f2\n- SHA256 (openssh-7.7p1.tar.gz) = 1zvn5oTpnvzQJL4Vowv/y+QbASsvezyQhK7WIXdea48=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\nOpenSSH 7.6 was released on 2017-10-03. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * ssh(1): delete SSH protocol version 1 support, associated\n configuration options and documentation.\n\n * ssh(1)/sshd(8): remove support for the hmac-ripemd160 MAC.\n\n * ssh(1)/sshd(8): remove support for the arcfour, blowfish and CAST\n ciphers.\n\n * Refuse RSA keys <1024 bits in length and improve reporting for keys\n that do not meet this requirement.\n\n * ssh(1): do not offer CBC ciphers by default.\n\nChanges since OpenSSH 7.5\n=========================\n\nThis is primarily a bugfix release. It also contains substantial\ninternal refactoring.\n\nSecurity\n--------\n\n * sftp-server(8): in read-only mode, sftp-server was incorrectly\n permitting creation of zero-length files. Reported by Michal\n Zalewski.\n\nNew Features\n------------\n\n * ssh(1): add RemoteCommand option to specify a command in the ssh\n config file instead of giving it on the client's command line. This\n allows the configuration file to specify the command that will be\n executed on the remote host.\n\n * sshd(8): add ExposeAuthInfo option that enables writing details of\n the authentication methods used (including public keys where\n applicable) to a file that is exposed via a $SSH_USER_AUTH\n environment variable in the subsequent session.\n\n * ssh(1): add support for reverse dynamic forwarding. In this mode,\n ssh will act as a SOCKS4/5 proxy and forward connections\n to destinations requested by the remote SOCKS client. This mode\n is requested using extended syntax for the -R and RemoteForward\n options and, because it is implemented solely at the client,\n does not require the server be updated to be supported.\n\n * sshd(8): allow LogLevel directive in sshd_config Match blocks;\n bz#2717\n\n * ssh-keygen(1): allow inclusion of arbitrary string or flag\n certificate extensions and critical options.\n\n * ssh-keygen(1): allow ssh-keygen to use a key held in ssh-agent as\n a CA when signing certificates. bz#2377\n\n * ssh(1)/sshd(8): allow IPQoS=none in ssh/sshd to not set an explicit\n ToS/DSCP value and just use the operating system default.\n\n * ssh-add(1): added -q option to make ssh-add quiet on success.\n\n * ssh(1): expand the StrictHostKeyChecking option with two new\n settings. The first "accept-new" will automatically accept\n hitherto-unseen keys but will refuse connections for changed or\n invalid hostkeys. This is a safer subset of the current behaviour\n of StrictHostKeyChecking=no. The second setting "off", is a synonym\n for the current behaviour of StrictHostKeyChecking=no: accept new\n host keys, and continue connection for hosts with incorrect\n hostkeys. A future release will change the meaning of\n StrictHostKeyChecking=no to the behaviour of "accept-new". bz#2400\n\n * ssh(1): add SyslogFacility option to ssh(1) matching the equivalent\n option in sshd(8). bz#2705\n\nBugfixes\n--------\n\n * ssh(1): use HostKeyAlias if specified instead of hostname for\n matching host certificate principal names; bz#2728\n\n * sftp(1): implement sorting for globbed ls; bz#2649\n\n * ssh(1): add a user@host prefix to client's "Permission denied"\n messages, useful in particular when using "stacked" connections\n (e.g. ssh -J) where it's not clear which host is denying. bz#2720\n\n * ssh(1): accept unknown EXT_INFO extension values that contain \\0\n characters. These are legal, but would previously cause fatal\n connection errors if received.\n\n * ssh(1)/sshd(8): repair compression statistics printed at\n connection exit\n\n * sftp(1): print '?' instead of incorrect link count (that the\n protocol doesn't provide) for remote listings. bz#2710\n\n * ssh(1): return failure rather than fatal() for more cases during\n session multiplexing negotiations. Causes the session to fall back\n to a non-mux connection if they occur. bz#2707\n\n * ssh(1): mention that the server may send debug messages to explain\n public key authentication problems under some circumstances; bz#2709\n\n * Translate OpenSSL error codes to better report incorrect passphrase\n errors when loading private keys; bz#2699\n\n * sshd(8): adjust compatibility patterns for WinSCP to correctly\n identify versions that implement only the legacy DH group exchange\n scheme. bz#2748\n\n * ssh(1): print the "Killed by signal 1" message only at LogLevel\n verbose so that it is not shown at the default level; prevents it\n from appearing during ssh -J and equivalent ProxyCommand configs.\n bz#1906, bz#2744\n\n * ssh-keygen(1): when generating all hostkeys (ssh-keygen -A), clobber\n existing keys if they exist but are zero length. zero-length keys\n could previously be made if ssh-keygen failed or was interrupted part\n way through generating them. bz#2561\n\n * ssh(1): fix pledge(2) violation in the escape sequence "~&" used to\n place the current session in the background.\n\n * ssh-keyscan(1): avoid double-close() on file descriptors; bz#2734\n\n * sshd(8): avoid reliance on shared use of pointers shared between\n monitor and child sshd processes. bz#2704\n\n * sshd_config(8): document available AuthenticationMethods; bz#2453\n\n * ssh(1): avoid truncation in some login prompts; bz#2768\n\n * sshd(8): Fix various compilations failures, inc bz#2767\n\n * ssh(1): make "--" before the hostname terminate argument processing\n after the hostname too.\n\n * ssh-keygen(1): switch from aes256-cbc to aes256-ctr for encrypting\n new-style private keys. Fixes problems related to private key\n handling for no-OpenSSL builds. bz#2754\n\n * ssh(1): warn and do not attempt to use keys when the public and\n private halves do not match. bz#2737\n\n * sftp(1): don't print verbose error message when ssh disconnects\n from under sftp. bz#2750\n\n * sshd(8): fix keepalive scheduling problem: activity on a forwarded\n port from preventing the keepalive from being sent; bz#2756\n\n * sshd(8): when started without root privileges, don't require the\n privilege separation user or path to exist. Makes running the\n regression tests easier without touching the filesystem.\n\n * Make integrity.sh regression tests more robust against timeouts.\n bz#2658\n\n * ssh(1)/sshd(8): correctness fix for channels implementation: accept\n channel IDs greater than 0x7FFFFFFF.\n\nPortability\n-----------\n\n * sshd(9): drop two more privileges in the Solaris sandbox:\n PRIV_DAX_ACCESS and PRIV_SYS_IB_INFO; bz#2723\n\n * sshd(8): expose list of completed authentication methods to PAM\n via the SSH_AUTH_INFO_0 PAM environment variable. bz#2408\n\n * ssh(1)/sshd(8): fix several problems in the tun/tap forwarding code,\n mostly to do with host/network byte order confusion. bz#2735\n\n * Add --with-cflags-after and --with-ldflags-after configure flags to\n allow setting CFLAGS/LDFLAGS after configure has completed. These\n are useful for setting sanitiser/fuzzing options that may interfere\n with configure's operation.\n\n * sshd(8): avoid Linux seccomp violations on ppc64le over the\n socketcall syscall.\n\n * Fix use of ldns when using ldns-config; bz#2697\n\n * configure: set cache variables when cross-compiling. The cross-\n compiling fallback message was saying it assumed the test passed,\n but it wasn't actually set the cache variables and this would\n cause later tests to fail.\n\n * Add clang libFuzzer harnesses for public key parsing and signature\n verification.\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.6.tar.gz) = 157fe3989a245c58fcdb34d9fe722a3c4e14c008\n - SHA1 (openssh-7.6p1.tar.gz) = a6984bc2c72192bed015c8b879b35dd9f5350b3b\n\n - SHA256 (openssh-7.6.tar.gz) = Xu3bdpCcu65vM2FnW7b6IKLgd4Kvf2P3WBTMw+I7Bao=\n - SHA256 (openssh-7.6p1.tar.gz) = oyPK7t3+FFuqoNsW6Y14Sx+8fdQ2pr8fR539XNHSFyM=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nOpenSSH 7.5 was released on 2017-03-20. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in future releases,\nspecifically:\n\n * In the next major release (expected June-August), removing remaining\n support for the SSH v.1 protocol (currently client-only and compile-\n time disabled).\n\n * In the same release, removing support for Blowfish and RC4 ciphers\n and the RIPE-MD160 HMAC. (These are currently run-time disabled).\n\n * In the same release, removing the remaining CBC ciphers from being\n offered by default in the client (These have not been offered in\n sshd by default for several years).\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\nThis list reflects our current intentions, but please check the final\nrelease notes for future releases.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * This release deprecates the sshd_config UsePrivilegeSeparation\n option, thereby making privilege separation mandatory. Privilege\n separation has been on by default for almost 15 years and\n sandboxing has been on by default for almost the last five.\n\n * The format of several log messages emitted by the packet code has\n changed to include additional information about the user and\n their authentication state. Software that monitors ssh/sshd logs\n may need to account for these changes. For example:\n\n Connection closed by user x 1.1.1.1 port 1234 [preauth]\n Connection closed by authenticating user x 10.1.1.1 port 1234 [preauth]\n Connection closed by invalid user x 1.1.1.1 port 1234 [preauth]\n\n Affected messages include connection closure, timeout, remote\n disconnection, negotiation failure and some other fatal messages\n generated by the packet code.\n\n * This version removes support for building\n against OpenSSL versions prior to 1.0.1. OpenSSL stopped supporting\n versions prior to 1.0.1 over 12 months ago (i.e. they no longer\n receive fixes for security bugs).\n\nChanges since OpenSSH 7.4\n=========================\n\nThis is a bugfix release.\n\nSecurity\n--------\n\n * ssh(1), sshd(8): Fix weakness in CBC padding oracle countermeasures\n that allowed a variant of the attack fixed in OpenSSH 7.3 to proceed.\n Note that the OpenSSH client disables CBC ciphers by default, sshd\n offers them as lowest-preference options and will remove them by\n default entriely in the next release. Reported by Jean Paul\n Degabriele, Kenny Paterson, Martin Albrecht and Torben Hansen of\n Royal Holloway, University of London.\n\n * sftp-client(1): On Cygwin, a client making\n a recursive file transfer could be manipulated by a hostile server to\n perform a path-traversal attack. creating or modifying files outside\n of the intended target directory. Reported by Jann Horn of Google\n Project Zero.\n\nNew Features\n------------\n\n * ssh(1), sshd(8): Support "=-" syntax to easily remove methods from\n algorithm lists, e.g. Ciphers=-*cbc. bz#2671\n\nBugfixes\n--------\n\n * sshd(1): Fix NULL dereference crash when key exchange start\n messages are sent out of sequence.\n\n * ssh(1), sshd(8): Allow form-feed characters to appear in\n configuration files.\n \n * sshd(8): Fix regression in OpenSSH 7.4 support for the\n server-sig-algs extension, where SHA2 RSA signature methods were\n not being correctly advertised. bz#2680\n\n * ssh(1), ssh-keygen(1): Fix a number of case-sensitivity bugs in\n known_hosts processing. bz#2591 bz#2685\n\n * ssh(1): Allow ssh to use certificates accompanied by a private key\n file but no corresponding plain *.pub public key. bz#2617\n\n * ssh(1): When updating hostkeys using the UpdateHostKeys option,\n accept RSA keys if HostkeyAlgorithms contains any RSA keytype.\n Previously, ssh could ignore RSA keys when only the ssh-rsa-sha2-*\n methods were enabled in HostkeyAlgorithms and not the old ssh-rsa\n method. bz#2650\n \n * ssh(1): Detect and report excessively long configuration file\n lines. bz#2651\n\n * Merge a number of fixes found by Coverity and reported via Redhat\n and FreeBSD. Includes fixes for some memory and file descriptor\n leaks in error paths. bz#2687\n \n * ssh-keyscan(1): Correctly hash hosts with a port number. bz#2692\n\n * ssh(1), sshd(8): When logging long messages to stderr, don't truncate\n "\\r\\n" if the length of the message exceeds the buffer. bz#2688\n\n * ssh(1): Fully quote [host]:port in generated ProxyJump/-J command-\n line; avoid confusion over IPv6 addresses and shells that treat\n square bracket characters specially.\n \n * ssh-keygen(1): Fix corruption of known_hosts when running\n "ssh-keygen -H" on a known_hosts containing already-hashed entries.\n\n * Fix various fallout and sharp edges caused by removing SSH protocol\n 1 support from the server, including the server banner string being\n incorrectly terminated with only \\n (instead of \\r\\n), confusing\n error messages from ssh-keyscan bz#2583 and a segfault in sshd\n if protocol v.1 was enabled for the client and sshd_config\n contained references to legacy keys bz#2686.\n\n * ssh(1), sshd(8): Free fd_set on connection timeout. bz#2683\n\n * sshd(8): Fix Unix domain socket forwarding for root (regression in\n OpenSSH 7.4).\n \n * sftp(1): Fix division by zero crash in "df" output when server\n returns zero total filesystem blocks/inodes.\n \n * ssh(1), ssh-add(1), ssh-keygen(1), sshd(8): Translate OpenSSL errors\n encountered during key loading to more meaningful error codes.\n bz#2522 bz#2523\n\n * ssh-keygen(1): Sanitise escape sequences in key comments sent to\n printf but preserve valid UTF-8 when the locale supports it;\n bz#2520\n\n * ssh(1), sshd(8): Return reason for port forwarding failures where\n feasible rather than always "administratively prohibited". bz#2674\n\n * sshd(8): Fix deadlock when AuthorizedKeysCommand or\n AuthorizedPrincipalsCommand produces a lot of output and a key is\n matched early. bz#2655\n\n * Regression tests: several reliability fixes. bz#2654 bz#2658 bz#2659\n \n * ssh(1): Fix typo in ~C error message for bad port forward\n cancellation. bz#2672\n\n * ssh(1): Show a useful error message when included config files\n can't be opened; bz#2653\n\n * sshd(8): Make sshd set GSSAPIStrictAcceptorCheck=yes as the manual page\n (previously incorrectly) advertised. bz#2637\n\n * sshd_config(5): Repair accidentally-deleted mention of %k token\n in AuthorizedKeysCommand; bz#2656\n\n * sshd(8): Remove vestiges of previously removed LOGIN_PROGRAM; bz#2665\n\n * ssh-agent(1): Relax PKCS#11 whitelist to include libexec and\n common 32-bit compatibility library directories.\n\n * sftp-client(1): Fix non-exploitable integer overflow in SSH2_FXP_NAME\n response handling.\n\n * ssh-agent(1): Fix regression in 7.4 of deleting PKCS#11-hosted\n keys. It was not possible to delete them except by specifying\n their full physical path. bz#2682\n\nPortability\n-----------\n\n * sshd(8): Avoid sandbox errors for Linux S390 systems using an ICA\n crypto coprocessor.\n\n * sshd(8): Fix non-exploitable weakness in seccomp-bpf sandbox arg\n inspection.\n\n * ssh(1): Fix X11 forwarding on OSX where X11 was being started by\n launchd. bz#2341\n\n * ssh-keygen(1), ssh(1), sftp(1): Fix output truncation for various that\n contain non-printable characters where the codeset in use is ASCII.\n\n * build: Fix builds that attempt to link a kerberised libldns. bz#2603\n\n * build: Fix compilation problems caused by unconditionally defining\n _XOPEN_SOURCE in wide character detection.\n\n * sshd(8): Fix sandbox violations for clock_gettime VSDO syscall\n fallback on some Linux/X32 kernels. bz#2142\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.5.tar.gz) = 81384df377e38551f7659a4c250383d0bbd25341\n - SHA1 (openssh-7.5p1.tar.gz) = 5e8f185d00afb4f4f89801e9b0f8b9cee9d87ebd\n\n - SHA256 (openssh-7.5.tar.gz) = Gmk8jOdGdKa7NixUN5J+bTMfeum5Vx8Nv+leAdQNq3U=\n - SHA256 (openssh-7.5p1.tar.gz) = mEbjxfq58FR0ALTSwBeZL5FCIrP9H47ubH3GvF5Z+fA=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nOpenSSH 7.4 was released on 2016-12-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in future releases,\nspecifically:\n\n * In approximately August 2017, removing remaining support for the\n SSH v.1 protocol (client-only and currently compile-time disabled).\n\n * In the same release, removing support for Blowfish and RC4 ciphers\n and the RIPE-MD160 HMAC. (These are currently run-time disabled).\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\n * The next release of OpenSSH will remove support for running sshd(8)\n with privilege separation disabled.\n\n * The next release of portable OpenSSH will remove support for\n OpenSSL version prior to 1.0.1.\n\nThis list reflects our current intentions, but please check the final\nrelease notes for future releases.\n\nPotentially-incompatible changes\n================================\n\nThis release includes a number of changes that may affect existing\nconfigurations:\n\n * This release removes server support for the SSH v.1 protocol.\n\n * ssh(1): Remove 3des-cbc from the client's default proposal. 64-bit\n block ciphers are not safe in 2016 and we don't want to wait until\n attacks like SWEET32 are extended to SSH. As 3des-cbc was the\n only mandatory cipher in the SSH RFCs, this may cause problems\n connecting to older devices using the default configuration,\n but it's highly likely that such devices already need explicit\n configuration for key exchange and hostkey algorithms already\n anyway.\n \n * sshd(8): Remove support for pre-authentication compression.\n Doing compression early in the protocol probably seemed reasonable\n in the 1990s, but today it's clearly a bad idea in terms of both\n cryptography (cf. multiple compression oracle attacks in TLS) and\n attack surface. Pre-auth compression support has been disabled by\n default for >10 years. Support remains in the client.\n \n * ssh-agent will refuse to load PKCS#11 modules outside a whitelist\n of trusted paths by default. The path whitelist may be specified\n at run-time.\n\n * sshd(8): When a forced-command appears in both a certificate and\n an authorized keys/principals command= restriction, sshd will now\n refuse to accept the certificate unless they are identical.\n The previous (documented) behaviour of having the certificate\n forced-command override the other could be a bit confusing and\n error-prone.\n \n * sshd(8): Remove the UseLogin configuration directive and support\n for having /bin/login manage login sessions.\n \nChanges since OpenSSH 7.3\n=========================\n\nThis is primarily a bugfix release.\n\nSecurity\n--------\n\n * ssh-agent(1): Will now refuse to load PKCS#11 modules from paths\n outside a trusted whitelist (run-time configurable). Requests to\n load modules could be passed via agent forwarding and an attacker\n could attempt to load a hostile PKCS#11 module across the forwarded\n agent channel: PKCS#11 modules are shared libraries, so this would\n result in code execution on the system running the ssh-agent if the\n attacker has control of the forwarded agent-socket (on the host\n running the sshd server) and the ability to write to the filesystem\n of the host running ssh-agent (usually the host running the ssh\n client). Reported by Jann Horn of Project Zero.\n\n * sshd(8): When privilege separation is disabled, forwarded Unix-\n domain sockets would be created by sshd(8) with the privileges of\n 'root' instead of the authenticated user. This release refuses\n Unix-domain socket forwarding when privilege separation is disabled\n (Privilege separation has been enabled by default for 14 years).\n Reported by Jann Horn of Project Zero.\n\n * sshd(8): Avoid theoretical leak of host private key material to\n privilege-separated child processes via realloc() when reading\n keys. No such leak was observed in practice for normal-sized keys,\n nor does a leak to the child processes directly expose key material\n to unprivileged users. Reported by Jann Horn of Project Zero.\n \n * sshd(8): The shared memory manager used by pre-authentication\n compression support had a bounds checks that could be elided by\n some optimising compilers. Additionally, this memory manager was\n incorrectly accessible when pre-authentication compression was\n disabled. This could potentially allow attacks against the\n privileged monitor process from the sandboxed privilege-separation\n process (a compromise of the latter would be required first).\n This release removes support for pre-authentication compression\n from sshd(8). Reported by Guido Vranken using the Stack unstable\n optimisation identification tool (http://css.csail.mit.edu/stack/)\n * sshd(8): Fix denial-of-service condition where an attacker who\n sends multiple KEXINIT messages may consume up to 128MB per\n connection. Reported by Shi Lei of Gear Team, Qihoo 360.\n\n * sshd(8): Validate address ranges for AllowUser and DenyUsers\n directives at configuration load time and refuse to accept invalid\n ones. It was previously possible to specify invalid CIDR address\n ranges (e.g. user@127.1.2.3/55) and these would always match,\n possibly resulting in granting access where it was not intended.\n Reported by Laurence Parry.\n\nNew Features\n------------\n\n * ssh(1): Add a proxy multiplexing mode to ssh(1) inspired by the\n version in PuTTY by Simon Tatham. This allows a multiplexing\n client to communicate with the master process using a subset of\n the SSH packet and channels protocol over a Unix-domain socket,\n with the main process acting as a proxy that translates channel\n IDs, etc. This allows multiplexing mode to run on systems that\n lack file- descriptor passing (used by current multiplexing\n code) and potentially, in conjunction with Unix-domain socket\n forwarding, with the client and multiplexing master process on\n different machines. Multiplexing proxy mode may be invoked using\n "ssh -O proxy ..."\n\n * sshd(8): Add a sshd_config DisableForwarding option that disables\n X11, agent, TCP, tunnel and Unix domain socket forwarding, as well\n as anything else we might implement in the future. Like the\n 'restrict' authorized_keys flag, this is intended to be a simple\n and future-proof way of restricting an account.\n\n * sshd(8), ssh(1): Support the "curve25519-sha256" key exchange\n method. This is identical to the currently-supported method named\n "curve25519-sha256@libssh.org".\n\n * sshd(8): Improve handling of SIGHUP by checking to see if sshd is\n already daemonised at startup and skipping the call to daemon(3)\n if it is. This ensures that a SIGHUP restart of sshd(8) will\n retain the same process-ID as the initial execution. sshd(8) will\n also now unlink the PidFile prior to SIGHUP restart and re-create\n it after a successful restart, rather than leaving a stale file in\n the case of a configuration error. bz#2641\n\n * sshd(8): Allow ClientAliveInterval and ClientAliveCountMax\n directives to appear in sshd_config Match blocks.\n\n * sshd(8): Add %-escapes to AuthorizedPrincipalsCommand to match\n those supported by AuthorizedKeysCommand (key, key type,\n fingerprint, etc.) and a few more to provide access to the\n contents of the certificate being offered.\n\n * Added regression tests for string matching, address matching and\n string sanitisation functions.\n\n * Improved the key exchange fuzzer harness.\n \nBugfixes\n--------\n\n * ssh(1): Allow IdentityFile to successfully load and use\n certificates that have no corresponding bare public key. bz#2617\n certificate id_rsa-cert.pub (and no id_rsa.pub).\n\n * ssh(1): Fix public key authentication when multiple\n authentication is in use and publickey is not just the first\n method attempted. bz#2642\n\n * regress: Allow the PuTTY interop tests to run unattended. bz#2639\n \n * ssh-agent(1), ssh(1): improve reporting when attempting to load\n keys from PKCS#11 tokens with fewer useless log messages and more\n detail in debug messages. bz#2610\n\n * ssh(1): When tearing down ControlMaster connections, don't\n pollute stderr when LogLevel=quiet.\n\n * sftp(1): On ^Z wait for underlying ssh(1) to suspend before\n suspending sftp(1) to ensure that ssh(1) restores the terminal mode\n correctly if suspended during a password prompt.\n\n * ssh(1): Avoid busy-wait when ssh(1) is suspended during a password\n prompt.\n\n * ssh(1), sshd(8): Correctly report errors during sending of ext-\n info messages.\n\n * sshd(8): fix NULL-deref crash if sshd(8) received an out-of-\n sequence NEWKEYS message.\n\n * sshd(8): Correct list of supported signature algorithms sent in\n the server-sig-algs extension. bz#2547\n\n * sshd(8): Fix sending ext_info message if privsep is disabled.\n\n * sshd(8): more strictly enforce the expected ordering of privilege\n separation monitor calls used for authentication and allow them\n only when their respective authentication methods are enabled\n in the configuration\n\n * sshd(8): Fix uninitialised optlen in getsockopt() call; harmless\n on Unix/BSD but potentially crashy on Cygwin.\n\n * Fix false positive reports caused by explicit_bzero(3) not being\n recognised as a memory initialiser when compiled with\n -fsanitize-memory.\n \n * sshd_config(5): Use 2001:db8::/32, the official IPv6 subnet for\n configuration examples.\n\nPortability\n-----------\n\n * On environments configured with Turkish locales, fall back to the\n C/POSIX locale to avoid errors in configuration parsing caused by\n that locale's unique handling of the letters 'i' and 'I'. bz#2643\n\n * sftp-server(8), ssh-agent(1): Deny ptrace on OS X using\n ptrace(PT_DENY_ATTACH, ..)\n\n * ssh(1), sshd(8): Unbreak AES-CTR ciphers on old (~0.9.8) OpenSSL.\n\n * Fix compilation for libcrypto compiled without RIPEMD160 support.\n\n * contrib: Add a gnome-ssh-askpass3 with GTK+3 support. bz#2640\n \n * sshd(8): Improve PRNG reseeding across privilege separation and\n force libcrypto to obtain a high-quality seed before chroot or\n sandboxing.\n\n * All: Explicitly test for broken strnvis. NetBSD added an strnvis\n and unfortunately made it incompatible with the existing one in\n OpenBSD and Linux's libbsd (the former having existed for over ten\n years). Try to detect this mess, and assume the only safe option\n if we're cross compiling.\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.4.tar.gz) = 1e2073f95d5ead8f2814b4b6c0700bcd533c410f\n - SHA1 (openssh-7.4p1.tar.gz) = 2330bbf82ed08cf3ac70e0acf00186ef3eeb97e0\n\n - SHA256 (openssh-7.4.tar.gz) = +GEXh7Xr2J87cq1uA97hF9e+3lfOQ2LKxXGdmFXREf0\n - SHA256 (openssh-7.4p1.tar.gz) = Gx/EoU4gJCkxgZJO0khy5vLgYpPz6JJqN2uK7EgfGdE=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nOpenSSH 7.3 was released on 2016-08-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in a near-future\nrelease, specifically:\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n * Removing server-side support for the SSH v.1 protocol (currently\n compile-time disabled).\n * In approximately 1 year, removing all support for the SSH v.1\n protocol (currently compile-time disabled).\n\nThis list reflects our current intentions, but please check the final\nrelease notes for future releases.\n\nChanges since OpenSSH 7.2\n=========================\n\nThis is primarily a bugfix release.\n\nSecurity\n--------\n\n * sshd(8): Mitigate a potential denial-of-service attack against\n the system's crypt(3) function via sshd(8). An attacker could\n send very long passwords that would cause excessive CPU use in\n crypt(3). sshd(8) now refuses to accept password authentication\n requests of length greater than 1024 characters. Independently\n reported by Tomas Kuthan (Oracle), Andres Rojas and Javier Nieto.\n\n * sshd(8): Mitigate timing differences in password authentication\n that could be used to discern valid from invalid account names\n when long passwords were sent and particular password hashing\n algorithms are in use on the server. CVE-2016-6210, reported by\n EddieEzra.Harari at verint.com\n\n * ssh(1), sshd(8): Fix observable timing weakness in the CBC padding\n oracle countermeasures. Reported by Jean Paul Degabriele, Kenny\n Paterson, Torben Hansen and Martin Albrecht. Note that CBC ciphers\n are disabled by default and only included for legacy compatibility.\n\n * ssh(1), sshd(8): Improve operation ordering of MAC verification for\n Encrypt-then-MAC (EtM) mode transport MAC algorithms to verify the\n MAC before decrypting any ciphertext. This removes the possibility\n of timing differences leaking facts about the plaintext, though no\n such leakage has been observed. Reported by Jean Paul Degabriele,\n Kenny Paterson, Torben Hansen and Martin Albrecht.\n \n * sshd(8): (portable only) Ignore PAM environment vars when\n UseLogin=yes. If PAM is configured to read user-specified\n environment variables and UseLogin=yes in sshd_config, then a\n hostile local user may attack /bin/login via LD_PRELOAD or\n similar environment variables set via PAM. CVE-2015-8325,\n found by Shayan Sadigh.\n\nNew Features\n------------\n\n * ssh(1): Add a ProxyJump option and corresponding -J command-line\n flag to allow simplified indirection through a one or more SSH\n bastions or "jump hosts".\n\n * ssh(1): Add an IdentityAgent option to allow specifying specific\n agent sockets instead of accepting one from the environment.\n \n * ssh(1): Allow ExitOnForwardFailure and ClearAllForwardings to be\n optionally overridden when using ssh -W. bz#2577\n\n * ssh(1), sshd(8): Implement support for the IUTF8 terminal mode as\n per draft-sgtatham-secsh-iutf8-00.\n \n * ssh(1), sshd(8): Add support for additional fixed Diffie-Hellman\n 2K, 4K and 8K groups from draft-ietf-curdle-ssh-kex-sha2-03.\n\n * ssh-keygen(1), ssh(1), sshd(8): support SHA256 and SHA512 RSA\n signatures in certificates;\n \n * ssh(1): Add an Include directive for ssh_config(5) files.\n\n * ssh(1): Permit UTF-8 characters in pre-authentication banners sent\n from the server. bz#2058\n\nBugfixes\n--------\n\n * ssh(1), sshd(8): Reduce the syslog level of some relatively common\n protocol events from LOG_CRIT. bz#2585\n\n * sshd(8): Refuse AuthenticationMethods="" in configurations and\n accept AuthenticationMethods=any for the default behaviour of not\n requiring multiple authentication. bz#2398\n\n * sshd(8): Remove obsolete and misleading "POSSIBLE BREAK-IN\n ATTEMPT!" message when forward and reverse DNS don't match. bz#2585\n\n * ssh(1): Close ControlPersist background process stderr except\n in debug mode or when logging to syslog. bz#1988\n\n * misc: Make PROTOCOL description for direct-streamlocal@openssh.com\n channel open messages match deployed code. bz#2529\n\n * ssh(1): Deduplicate LocalForward and RemoteForward entries to fix\n failures when both ExitOnForwardFailure and hostname\n canonicalisation are enabled. bz#2562\n\n * sshd(8): Remove fallback from moduli to obsolete "primes" file\n that was deprecated in 2001. bz#2559.\n\n * sshd_config(5): Correct description of UseDNS: it affects ssh\n hostname processing for authorized_keys, not known_hosts; bz#2554\n \n * ssh(1): Fix authentication using lone certificate keys in an agent\n without corresponding private keys on the filesystem. bz#2550\n\n * sshd(8): Send ClientAliveInterval pings when a time-based\n RekeyLimit is set; previously keepalive packets were not being\n sent. bz#2252\n \nPortability\n-----------\n\n * ssh(1), sshd(8): Fix compilation by automatically disabling ciphers\n not supported by OpenSSL. bz#2466\n\n * misc: Fix compilation failures on some versions of AIX's compiler\n related to the definition of the VA_COPY macro. bz#2589\n\n * sshd(8): Whitelist more architectures to enable the seccomp-bpf\n sandbox. bz#2590\n\n * ssh-agent(1), sftp-server(8): Disable process tracing on Solaris\n using setpflags(__PROC_PROTECT, ...). bz#2584\n\n * sshd(8): On Solaris, don't call Solaris setproject() with\n UsePAM=yes it's PAM's responsibility. bz#2425\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.3.tar.gz) = b1641e5265d9ec68a9a19decc3a7edd1203cbd33\n - SHA256 (openssh-7.3.tar.gz) = vS0X35qrX9OOPBkyDMYhOje/DBwHBVEV7nv5rkzw4vM=\n\n - SHA1 (openssh-7.3p1.tar.gz) = bfade84283fcba885e2084343ab19a08c7d123a5\n - SHA256 (openssh-7.3p1.tar.gz) = P/uYmm3KppWUw7VQ1IVaWi4XGMzd5/XjY4e0JCIPvsw=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nPortable OpenSSH 7.2p2 was released on 2016-03-10. It will be available\nfrom the mirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols that\nmay be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for\ntheir continued support of the project, especially those who\ncontributed code or patches, reported bugs, tested snapshots or\ndonated to the project. More information on donations may be found\nat: http://www.openssh.com/donations.html\n\nChanges since OpenSSH 7.2p1\n===========================\n\nThis release fixes a security bug:\n\n * sshd(8): sanitise X11 authentication credentials to avoid xauth\n command injection when X11Forwarding is enabled.\n\n Full details of the vulnerability are available at:\n http://www.openssh.com/txt/x11fwd.adv\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.2p2.tar.gz) = 70e35d7d6386fe08abbd823b3a12a3ca44ac6d38\n - SHA256 (openssh-7.2p2.tar.gz) = pyeB0aBDh2oiT/GwAy2qQJTYdWWmhSh1nBwsq1SCVIw=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nOpenSSH 7.2 was released on 2016-02-29. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in a near-future\nrelease, specifically:\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\nThis list reflects our current intentions, but please check the final\nrelease notes for future releases.\n\nPotentially-incompatible changes\n================================\n\nThis release disables a number of legacy cryptographic algorithms\nby default in ssh:\n\n * Several ciphers blowfish-cbc, cast128-cbc, all arcfour variants\n and the rijndael-cbc aliases for AES.\n\n * MD5-based and truncated HMAC algorithms.\n\nThese algorithms are already disabled by default in sshd.\n\nChanges since OpenSSH 7.1p2\n===========================\n\nThis is primarily a bugfix release.\n\nSecurity\n--------\n\n * ssh(1), sshd(8): remove unfinished and unused roaming code (was\n already forcibly disabled in OpenSSH 7.1p2).\n \n * ssh(1): eliminate fallback from untrusted X11 forwarding to\n trusted forwarding when the X server disables the SECURITY\n extension.\n\n * ssh(1), sshd(8): increase the minimum modulus size supported for\n diffie-hellman-group-exchange to 2048 bits.\n\n * sshd(8): pre-auth sandboxing is now enabled by default (previous\n releases enabled it for new installations via sshd_config).\n\nNew Features\n------------\n\n * all: add support for RSA signatures using SHA-256/512 hash\n algorithms based on draft-rsa-dsa-sha2-256-03.txt and\n draft-ssh-ext-info-04.txt.\n\n * ssh(1): Add an AddKeysToAgent client option which can be set to\n 'yes', 'no', 'ask', or 'confirm', and defaults to 'no'. When\n enabled, a private key that is used during authentication will be\n added to ssh-agent if it is running (with confirmation enabled if\n set to 'confirm').\n \n * sshd(8): add a new authorized_keys option "restrict" that includes\n all current and future key restrictions (no-*-forwarding, etc.).\n Also add permissive versions of the existing restrictions, e.g.\n "no-pty" -> "pty". This simplifies the task of setting up\n restricted keys and ensures they are maximally-restricted,\n regardless of any permissions we might implement in the future.\n \n * ssh(1): add ssh_config CertificateFile option to explicitly list\n certificates. bz#2436\n \n * ssh-keygen(1): allow ssh-keygen to change the key comment for all\n supported formats.\n\n * ssh-keygen(1): allow fingerprinting from standard input, e.g.\n "ssh-keygen -lf -"\n\n * ssh-keygen(1): allow fingerprinting multiple public keys in a\n file, e.g. "ssh-keygen -lf ~/.ssh/authorized_keys" bz#1319\n\n * sshd(8): support "none" as an argument for sshd_config\n Foreground and ChrootDirectory. Useful inside Match blocks to\n override a global default. bz#2486\n\n * ssh-keygen(1): support multiple certificates (one per line) and\n reading from standard input (using "-f -") for "ssh-keygen -L"\n \n * ssh-keyscan(1): add "ssh-keyscan -c ..." flag to allow fetching\n certificates instead of plain keys.\n \n * ssh(1): better handle anchored FQDNs (e.g. 'cvs.openbsd.org.') in\n hostname canonicalisation - treat them as already canonical and\n remove the trailing '.' before matching ssh_config.\n\nBugfixes\n--------\n\n * sftp(1): existing destination directories should not terminate\n recursive uploads (regression in openssh 6.8) bz#2528\n\n * ssh(1), sshd(8): correctly send back SSH2_MSG_UNIMPLEMENTED\n replies to unexpected messages during key exchange. bz#2949\n\n * ssh(1): refuse attempts to set ConnectionAttempts=0, which does\n not make sense and would cause ssh to print an uninitialised stack\n variable. bz#2500\n\n * ssh(1): fix errors when attempting to connect to scoped IPv6\n addresses with hostname canonicalisation enabled.\n\n * sshd_config(5): list a couple more options usable in Match blocks.\n bz#2489\n\n * sshd(8): fix "PubkeyAcceptedKeyTypes +..." inside a Match block.\n \n * ssh(1): expand tilde characters in filenames passed to -i options\n before checking whether or not the identity file exists. Avoids\n confusion for cases where shell doesn't expand (e.g. "-i ~/file"\n vs. "-i~/file"). bz#2481\n\n * ssh(1): do not prepend "exec" to the shell command run by "Match\n exec" in a config file, which could cause some commands to fail\n in certain environments. bz#2471\n\n * ssh-keyscan(1): fix output for multiple hosts/addrs on one line\n when host hashing or a non standard port is in use bz#2479\n \n * sshd(8): skip "Could not chdir to home directory" message when\n ChrootDirectory is active. bz#2485\n\n * ssh(1): include PubkeyAcceptedKeyTypes in ssh -G config dump.\n \n * sshd(8): avoid changing TunnelForwarding device flags if they are\n already what is needed; makes it possible to use tun/tap\n networking as non-root user if device permissions and interface\n flags are pre-established\n\n * ssh(1), sshd(8): RekeyLimits could be exceeded by one packet.\n bz#2521\n\n * ssh(1): fix multiplexing master failure to notice client exit.\n\n * ssh(1), ssh-agent(1): avoid fatal() for PKCS11 tokens that present\n empty key IDs. bz#1773\n\n * sshd(8): avoid printf of NULL argument. bz#2535 \n\n * ssh(1), sshd(8): allow RekeyLimits larger than 4GB. bz#2521\n \n * ssh-keygen(1): sshd(8): fix several bugs in (unused) KRL signature\n support.\n\n * ssh(1), sshd(8): fix connections with peers that use the key\n exchange guess feature of the protocol. bz#2515\n\n * sshd(8): include remote port number in log messages. bz#2503\n\n * ssh(1): don't try to load SSHv1 private key when compiled without\n SSHv1 support. bz#2505\n\n * ssh-agent(1), ssh(1): fix incorrect error messages during key\n loading and signing errors. bz#2507\n\n * ssh-keygen(1): don't leave empty temporary files when performing\n known_hosts file edits when known_hosts doesn't exist.\n\n * sshd(8): correct packet format for tcpip-forward replies for\n requests that don't allocate a port bz#2509\n\n * ssh(1), sshd(8): fix possible hang on closed output. bz#2469\n \n * ssh(1): expand %i in ControlPath to UID. bz#2449\n\n * ssh(1), sshd(8): fix return type of openssh_RSA_verify. bz#2460\n \n * ssh(1), sshd(8): fix some option parsing memory leaks. bz#2182\n\n * ssh(1): add a some debug output before DNS resolution; it's a\n place where ssh could previously silently stall in cases of\n unresponsive DNS servers. bz#2433\n \n * ssh(1): remove spurious newline in visual hostkey. bz#2686\n \n * ssh(1): fix printing (ssh -G ...) of HostKeyAlgorithms=+...\n \n * ssh(1): fix expansion of HostkeyAlgorithms=+...\n\nDocumentation\n-------------\n\n * ssh_config(5), sshd_config(5): update default algorithm lists to\n match current reality. bz#2527\n\n * ssh(1): mention -Q key-plain and -Q key-cert query options.\n bz#2455\n\n * sshd_config(8): more clearly describe what AuthorizedKeysFile=none\n does.\n\n * ssh_config(5): better document ExitOnForwardFailure. bz#2444 \n\n * sshd(5): mention internal DH-GEX fallback groups in manual.\n bz#2302\n\n * sshd_config(5): better description for MaxSessions option.\n bz#2531\n\nPortability\n-----------\n\n * ssh(1), sftp-server(8), ssh-agent(1), sshd(8): Support Illumos/\n Solaris fine-grained privileges. Including a pre-auth privsep\n sandbox and several pledge() emulations. bz#2511\n\n * Renovate redhat/openssh.spec, removing deprecated options and\n syntax.\n\n * configure: allow --without-ssl-engine with --without-openssl\n \n * sshd(8): fix multiple authentication using S/Key. bz#2502\n\n * sshd(8): read back from libcrypto RAND_* before dropping\n privileges. Avoids sandboxing violations with BoringSSL.\n\n * Fix name collision with system-provided glob(3) functions.\n bz#2463\n\n * Adapt Makefile to use ssh-keygen -A when generating host keys.\n bz#2459\n \n * configure: correct default value for --with-ssh1 bz#2457\n\n * configure: better detection of _res symbol bz#2259\n\n * support getrandom() syscall on Linux\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.2.tar.gz) = 9567d00fffe655010c087aeb80c830cecbbecca6\n - SHA256 (openssh-7.2.tar.gz) = 99GsHA8NwSGuEJhMc7hAOQ510y1xfGx27uJqyw73sCI=\n\n - SHA1 (openssh-7.2p1.tar.gz) = d30a6fd472199ab5838a7668c0c5fd885fb8d371\n - SHA256 (openssh-7.2p1.tar.gz) = lzzDey81l+TPWZsJ5gTnnA/l2bb1laJOke0GYoYLSsM=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de\nRaadt, Kevin Steves, Damien Miller, Darren Tucker, Jason McIntyre,\nTim Rice and Ben Lindstrom.\nOpenSSH 7.1p2 was released on 2016-01-14. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 7.1p1\n===========================\n\n * SECURITY: ssh(1): The OpenSSH client code between 5.4 and 7.1\n contains experimental support for resuming SSH-connections (roaming).\n\n The matching server code has never been shipped, but the client\n code was enabled by default and could be tricked by a malicious\n server into leaking client memory to the server, including private\n client user keys.\n\n The authentication of the server host key prevents exploitation\n by a man-in-the-middle, so this information leak is restricted\n to connections to malicious or compromised servers.\n\n MITIGATION: For OpenSSH >= 5.4 the vulnerable code in the client\n can be completely disabled by adding 'UseRoaming no' to the global\n ssh_config(5) file, or to user configuration in ~/.ssh/config,\n or by passing -oUseRoaming=no on the command line.\n\n PATCH: See below for a patch to disable this feature (Disabling\n Roaming in the Source Code).\n\n This problem was reported by the Qualys Security Advisory team.\n\n * SECURITY: Fix an out of-bound read access in the packet handling\n code. Reported by Ben Hawkes.\n\n * PROTOCOL: Correctly interpret the 'first_kex_follows' option during\n the initial key exchange. Reported by Matt Johnston.\n\n * Further use of explicit_bzero has been added in various buffer\n handling code paths to guard against compilers aggressively\n doing dead-store removal.\n\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.1p2.tar.gz) = 9202f5a2a50c8a55ecfb830609df1e1fde97f758\n - SHA256 (openssh-7.1p2.tar.gz) = dd75f024dcf21e06a0d6421d582690bf987a1f6323e32ad6619392f3bfde6bbd\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\n\nDisabling Roaming in the Source Code:\n=====================================\n\n--- readconf.c\t30 Jul 2015 00:01:34 -0000\t1.239\n+++ readconf.c\t13 Jan 2016 23:17:23 -0000\n@@ -1648,7 +1648,7 @@ initialize_options(Options * options)\n \toptions->tun_remote = -1;\n \toptions->local_command = NULL;\n \toptions->permit_local_command = -1;\n-\toptions->use_roaming = -1;\n+\toptions->use_roaming = 0;\n \toptions->visual_host_key = -1;\n \toptions->ip_qos_interactive = -1;\n \toptions->ip_qos_bulk = -1;\n@@ -1819,8 +1819,7 @@ fill_default_options(Options * options)\n \t\toptions->tun_remote = SSH_TUNID_ANY;\n \tif (options->permit_local_command == -1)\n \t\toptions->permit_local_command = 0;\n-\tif (options->use_roaming == -1)\n-\t\toptions->use_roaming = 1;\n+\toptions->use_roaming = 0;\n \tif (options->visual_host_key == -1)\n \t\toptions->visual_host_key = 0;\n \tif (options->ip_qos_interactive == -1)\n--- ssh.c\t30 Jul 2015 00:01:34 -0000\t1.420\n+++ ssh.c\t13 Jan 2016 23:17:23 -0000\n@@ -1882,9 +1882,6 @@ ssh_session2(void)\n \t\t\tfork_postauth();\n \t}\n \n-\tif (options.use_roaming)\n-\t\trequest_roaming();\n-\n \treturn client_loop(tty_flag, tty_flag ?\n \t options.escape_char : SSH_ESCAPECHAR_NONE, id);\n }\nOpenSSH 7.1 was released on 2015-08-21. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in the next release\nincluding:\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\n * Several ciphers will be disabled by default: blowfish-cbc,\n cast128-cbc, all arcfour variants and the rijndael-cbc aliases\n for AES.\n\n * MD5-based HMAC algorithms will be disabled by default.\n\nThis list reflects our current intentions, but please check the final\nrelease notes for OpenSSH 7.2 when it is released.\n\nChanges since OpenSSH 7.0\n=========================\n\nThis is a bugfix release.\n\nSecurity\n--------\n\n * sshd(8): OpenSSH 7.0 contained a logic error in PermitRootLogin=\n prohibit-password/without-password that could, depending on\n compile-time configuration, permit password authentication to\n root while preventing other forms of authentication. This problem\n was reported by Mantas Mikulenas.\n\nBugfixes\n--------\n\n * ssh(1), sshd(8): add compatibility workarounds for FuTTY\n\n * ssh(1), sshd(8): refine compatibility workarounds for WinSCP\n\n * Fix a number of memory faults (double-free, free of uninitialised\n memory, etc) in ssh(1) and ssh-keygen(1). Reported by Mateusz\n Kocielski.\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.1.tar.gz) = 06c1db39f33831fe004726e013b2cf84f1889042\n - SHA256 (openssh-7.1.tar.gz) = H7U1se9EoBmhkKi2i7lqpMX9QHdDTsgpu7kd5VZUGSY=\n\n - SHA1 (openssh-7.1p1.tar.gz) = ed22af19f962262c493fcc6ed8c8826b2761d9b6\n - SHA256 (openssh-7.1p1.tar.gz) = /AptLR0GPVxm3/2VJJPQzaJWytIE9oHeD4TvhbKthCg=\n\nPlease note that the SHA256 signatures are base64 encoded and not\nhexadecimal (which is the default for most checksum tools). The PGP\nkey used to sign the releases is available as RELEASE_KEY.asc from\nthe mirror sites.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 7.0 was released on 2015-08-11. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol 2.0 implementation and\nincludes sftp client and server support. OpenSSH also includes\ntransitional support for the legacy SSH 1.3 and 1.5 protocols\nthat may be enabled at compile-time.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture deprecation notice\n=========================\n\nWe plan on retiring more legacy cryptography in the next release\nincluding:\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\n * Several ciphers will be disabled by default: blowfish-cbc,\n cast128-cbc, all arcfour variants and the rijndael-cbc aliases\n for AES.\n\n * MD5-based HMAC algorithms will be disabled by default.\n\nThis list reflects our current intentions, but please check the final\nrelease notes for OpenSSH 7.1 when it is released.\n\nChanges since OpenSSH 6.9\n=========================\n\nThis focus of this release is primarily to deprecate weak, legacy\nand/or unsafe cryptography.\n\nSecurity\n--------\n\n * sshd(8): OpenSSH 6.8 and 6.9 incorrectly set TTYs to be world-\n writable. Local attackers may be able to write arbitrary messages\n to logged-in users, including terminal escape sequences.\n Reported by Nikolay Edigaryev.\n\n * sshd(8): Portable OpenSSH only: Fixed a privilege separation\n weakness related to PAM support. Attackers who could successfully\n compromise the pre-authentication process for remote code\n execution and who had valid credentials on the host could\n impersonate other users. Reported by Moritz Jodeit.\n\n * sshd(8): Portable OpenSSH only: Fixed a use-after-free bug\n related to PAM support that was reachable by attackers who could\n compromise the pre-authentication process for remote code\n execution. Also reported by Moritz Jodeit.\n\n * sshd(8): fix circumvention of MaxAuthTries using keyboard-\n interactive authentication. By specifying a long, repeating\n keyboard-interactive "devices" string, an attacker could request\n the same authentication method be tried thousands of times in\n a single pass. The LoginGraceTime timeout in sshd(8) and any\n authentication failure delays implemented by the authentication\n mechanism itself were still applied. Found by Kingcope.\n\nPotentially-incompatible Changes\n--------------------------------\n\n * Support for the legacy SSH version 1 protocol is disabled by\n default at compile time.\n\n * Support for the 1024-bit diffie-hellman-group1-sha1 key exchange\n is disabled by default at run-time. It may be re-enabled using\n the instructions at http://www.openssh.com/legacy.html\n\n * Support for ssh-dss, ssh-dss-cert-* host and user keys is disabled\n by default at run-time. These may be re-enabled using the\n instructions at http://www.openssh.com/legacy.html\n\n * Support for the legacy v00 cert format has been removed.\n\n * The default for the sshd_config(5) PermitRootLogin option has\n changed from "yes" to "prohibit-password".\n\n * PermitRootLogin=without-password/prohibit-password now bans all\n interactive authentication methods, allowing only public-key,\n hostbased and GSSAPI authentication (previously it permitted\n keyboard-interactive and password-less authentication if those\n were enabled).\n\nNew Features\n------------\n\n * ssh_config(5): add PubkeyAcceptedKeyTypes option to control which\n public key types are available for user authentication.\n\n * sshd_config(5): add HostKeyAlgorithms option to control which\n public key types are offered for host authentications.\n\n * ssh(1), sshd(8): extend Ciphers, MACs, KexAlgorithms,\n HostKeyAlgorithms, PubkeyAcceptedKeyTypes and HostbasedKeyTypes\n options to allow appending to the default set of algorithms\n instead of replacing it. Options may now be prefixed with a '+'\n to append to the default, e.g. "HostKeyAlgorithms=+ssh-dss".\n\n * sshd_config(5): PermitRootLogin now accepts an argument of\n 'prohibit-password' as a less-ambiguous synonym of 'without-\n password'.\n\nBugfixes\n--------\n\n * ssh(1), sshd(8): add compatibility workarounds for Cisco and more\n PuTTY versions. bz#2424\n\n * Fix some omissions and errors in the PROTOCOL and PROTOCOL.mux\n documentation relating to Unix domain socket forwarding;\n bz#2421 bz#2422\n\n * ssh(1): Improve the ssh(1) manual page to include a better\n description of Unix domain socket forwarding; bz#2423\n\n * ssh(1), ssh-agent(1): skip uninitialised PKCS#11 slots, fixing\n failures to load keys when they are present. bz#2427\n\n * ssh(1), ssh-agent(1): do not ignore PKCS#11 hosted keys that with\n empty CKA_ID; bz#2429\n\n * sshd(8): clarify documentation for UseDNS option; bz#2045\n\nPortable OpenSSH\n----------------\n\n * Check realpath(3) behaviour matches what sftp-server requires and\n use a replacement if necessary.\n\nChecksums:\n==========\n\n - SHA1 (openssh-7.0.tar.gz) = a19ff0bad2a67348b1d01a38a9580236120b7099\n - SHA256 (openssh-7.0.tar.gz) = 4F6HV/ZqT465f3sMB2vIkXO+wrYtL5hnqzAymfbZ1Jk=\n\n - SHA1 (openssh-7.0p1.tar.gz) = d8337c9eab91d360d104f6dd805f8b32089c063c\n - SHA256 (openssh-7.0p1.tar.gz) = /VkySToZ9MgRU9gS7k4EK0m707dZqz2TRKvswrwUheU=\n\nPlease note that the PGP key used to sign releases was recently rotated.\nThe new key has been signed by the old key to provide continuity. It is\navailable from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.9 was released on 2015-07-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nFuture Deprecation Notice\n=========================\n\nThe 7.0 release of OpenSSH, due for release in late July, will\ndeprecate several features, some of which may affect compatibility\nor existing configurations. The intended changes are as follows:\n\n * The default for the sshd_config(5) PermitRootLogin option will\n change from "yes" to "no".\n\n * Support for the legacy version 1.x of the SSH protocol will be\n disabled at compile time by default.\n\n * Support for the 1024-bit diffie-hellman-group1-sha1 key exchange\n will be run-time disabled by default.\n\n * Support for ssh-dss, ssh-dss-cert-* host and user keys will be\n run-time disabled by default.\n\n * Support for the legacy v00 cert format will be removed\n\n * Several ciphers will be disabled by default: blowfish-cbc,\n cast128-cbc, all arcfour variants and the rijndael-cbc aliases\n for AES\n\n * Refusing all RSA keys smaller than 1024 bits (the current minimum\n is 768 bits)\n\nThis list reflects our current intentions, but please check the final\nrelease notes for OpenSSH 7.0 when it is released.\n\nChanges since OpenSSH 6.8\n=========================\n\nThis is primarily a bugfix release.\n\nSecurity\n--------\n\n * ssh(1): when forwarding X11 connections with ForwardX11Trusted=no,\n connections made after ForwardX11Timeout expired could be permitted\n and no longer subject to XSECURITY restrictions because of an\n ineffective timeout check in ssh(1) coupled with "fail open"\n behaviour in the X11 server when clients attempted connections with\n expired credentials. This problem was reported by Jann Horn.\n\n * ssh-agent(1): fix weakness of agent locking (ssh-add -x) to\n password guessing by implementing an increasing failure delay,\n storing a salted hash of the password rather than the password\n itself and using a timing-safe comparison function for verifying\n unlock attempts. This problem was reported by Ryan Castellucci.\n\nNew Features\n------------\n\n * ssh(1), sshd(8): promote chacha20-poly1305@openssh.com to be the\n default cipher\n\n * sshd(8): support admin-specified arguments to AuthorizedKeysCommand;\n bz#2081\n\n * sshd(8): add AuthorizedPrincipalsCommand that allows retrieving\n authorized principals information from a subprocess rather than\n a file.\n\n * ssh(1), ssh-add(1): support PKCS#11 devices with external PIN\n entry devices bz#2240\n\n * sshd(8): allow GSSAPI host credential check to be relaxed for\n multihomed hosts via GSSAPIStrictAcceptorCheck option; bz#928\n\n * ssh-keygen(1): support "ssh-keygen -lF hostname" to search\n known_hosts and print key hashes rather than full keys.\n\n * ssh-agent(1): add -D flag to leave ssh-agent in foreground without\n enabling debug mode; bz#2381\n\nBugfixes\n--------\n\n * ssh(1), sshd(8): deprecate legacy SSH2_MSG_KEX_DH_GEX_REQUEST_OLD\n message and do not try to use it against some 3rd-party SSH\n implementations that use it (older PuTTY, WinSCP).\n\n * Many fixes for problems caused by compile-time deactivation of\n SSH1 support (including bz#2369)\n\n * ssh(1), sshd(8): cap DH-GEX group size at 4Kbits for Cisco\n implementations as some would fail when attempting to use group\n sizes >4K; bz#2209\n\n * ssh(1): fix out-of-bound read in EscapeChar configuration option\n parsing; bz#2396\n\n * sshd(8): fix application of PermitTunnel, LoginGraceTime,\n AuthenticationMethods and StreamLocalBindMask options in Match\n blocks\n\n * ssh(1), sshd(8): improve disconnection message on TCP reset;\n bz#2257\n\n * ssh(1): remove failed remote forwards established by muliplexing\n from the list of active forwards; bz#2363\n\n * sshd(8): make parsing of authorized_keys "environment=" options\n independent of PermitUserEnv being enabled; bz#2329\n\n * sshd(8): fix post-auth crash with permitopen=none; bz#2355\n\n * ssh(1), ssh-add(1), ssh-keygen(1): allow new-format private keys\n to be encrypted with AEAD ciphers; bz#2366\n\n * ssh(1): allow ListenAddress, Port and AddressFamily configuration\n options to appear in any order; bz#86\n\n * sshd(8): check for and reject missing arguments for VersionAddendum\n and ForceCommand; bz#2281\n\n * ssh(1), sshd(8): don't treat unknown certificate extensions as\n fatal; bz#2387\n\n * ssh-keygen(1): make stdout and stderr output consistent; bz#2325\n\n * ssh(1): mention missing DISPLAY environment in debug log when X11\n forwarding requested; bz#1682\n\n * sshd(8): correctly record login when UseLogin is set; bz#378\n\n * sshd(8): Add some missing options to sshd -T output and fix output\n of VersionAddendum and HostCertificate. bz#2346 \n\n * Document and improve consistency of options that accept a "none"\n argument" TrustedUserCAKeys, RevokedKeys (bz#2382),\n AuthorizedPrincipalsFile (bz#2288)\n\n * ssh(1): include remote username in debug output; bz#2368\n\n * sshd(8): avoid compatibility problem with some versions of Tera\n Term, which would crash when they received the hostkeys notification\n message (hostkeys-00@openssh.com)\n\n * sshd(8): mention ssh-keygen -E as useful when comparing legacy MD5\n host key fingerprints; bz#2332\n\n * ssh(1): clarify pseudo-terminal request behaviour and use make\n manual language consistent; bz#1716\n\n * ssh(1): document that the TERM environment variable is not subject\n to SendEnv and AcceptEnv; bz#2386\n\nPortable OpenSSH\n----------------\n\n * sshd(8): Format UsePAM setting when using sshd -T, part of bz#2346\n\n * Look for '${host}-ar' before 'ar', making cross-compilation easier;\n bz#2352.\n\n * Several portable compilation fixes: bz#2402, bz#2337, bz#2370\n\n * moduli(5): update DH-GEX moduli\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.9.tar.gz) = cd5fcb93411025bbc4b4b57753b622769dfb1e0d\n - SHA256 (openssh-6.9.tar.gz) = itCMw0aE/xvrGKWhzRD2UM/9kzIOyFaH2dIWMfX8agQ=\n\n - SHA1 (openssh-6.9p1.tar.gz) = 86ab57f00d0fd9bf302760f2f6deac1b6e9df265\n - SHA256 (openssh-6.9p1.tar.gz) = bgdN9TjzV9RAvmz5PcWBoh8i054jbyF/zY6su2yJbP4=\n\nPlease note that the PGP key used to sign releases was recently rotated.\nThe new key has been signed by the old key to provide continuity. It is\navailable from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.8 was released on 2015-03-18. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 6.7\n=========================\n\nThis is a major release, containing a number of new features as\nwell as a large internal re-factoring.\n\nPotentially-incompatible changes\n--------------------------------\n\n * sshd(8): UseDNS now defaults to 'no'. Configurations that match\n against the client host name (via sshd_config or authorized_keys)\n may need to re-enable it or convert to matching against addresses.\n\nNew Features\n------------\n\n * Much of OpenSSH's internal code has been re-factored to be more\n library-like. These changes are mostly not user-visible, but\n have greatly improved OpenSSH's testability and internal layout.\n\n * Add FingerprintHash option to ssh(1) and sshd(8), and equivalent\n command-line flags to the other tools to control algorithm used\n for key fingerprints. The default changes from MD5 to SHA256 and\n format from hex to base64.\n\n Fingerprints now have the hash algorithm prepended. An example of\n the new format: SHA256:mVPwvezndPv/ARoIadVY98vAC0g+P/5633yTC4d/wXE\n Please note that visual host keys will also be different.\n\n * ssh(1), sshd(8): Experimental host key rotation support. Add a\n protocol extension for a server to inform a client of all its\n available host keys after authentication has completed. The client\n may record the keys in known_hosts, allowing it to upgrade to better\n host key algorithms and a server to gracefully rotate its keys.\n\n The client side of this is controlled by a UpdateHostkeys config\n option (default off).\n\n * ssh(1): Add a ssh_config HostbasedKeyType option to control which\n host public key types are tried during host-based authentication.\n\n * ssh(1), sshd(8): fix connection-killing host key mismatch errors\n when sshd offers multiple ECDSA keys of different lengths.\n\n * ssh(1): when host name canonicalisation is enabled, try to\n parse host names as addresses before looking them up for\n canonicalisation. fixes bz#2074 and avoiding needless DNS\n lookups in some cases.\n\n * ssh-keygen(1), sshd(8): Key Revocation Lists (KRLs) no longer\n require OpenSSH to be compiled with OpenSSL support.\n\n * ssh(1), ssh-keysign(8): Make ed25519 keys work for host based\n authentication.\n\n * sshd(8): SSH protocol v.1 workaround for the Meyer, et al,\n Bleichenbacher Side Channel Attack. Fake up a bignum key before\n RSA decryption.\n\n * sshd(8): Remember which public keys have been used for\n authentication and refuse to accept previously-used keys.\n This allows AuthenticationMethods=publickey,publickey to require\n that users authenticate using two _different_ public keys.\n\n * sshd(8): add sshd_config HostbasedAcceptedKeyTypes and\n PubkeyAcceptedKeyTypes options to allow sshd to control what\n public key types will be accepted. Currently defaults to all.\n\n * sshd(8): Don't count partial authentication success as a failure\n against MaxAuthTries.\n\n * ssh(1): Add RevokedHostKeys option for the client to allow\n text-file or KRL-based revocation of host keys.\n\n * ssh-keygen(1), sshd(8): Permit KRLs that revoke certificates by\n serial number or key ID without scoping to a particular CA.\n\n * ssh(1): Add a "Match canonical" criteria that allows ssh_config\n Match blocks to trigger only in the second config pass.\n\n * ssh(1): Add a -G option to ssh that causes it to parse its\n configuration and dump the result to stdout, similar to "sshd -T".\n\n * ssh(1): Allow Match criteria to be negated. E.g. "Match !host".\n\n * The regression test suite has been extended to cover more OpenSSH\n features. The unit tests have been expanded and now cover key\n exchange.\n\nBugfixes\n\n * ssh-keyscan(1): ssh-keyscan has been made much more robust again\n servers that hang or violate the SSH protocol.\n\n * ssh(1), ssh-keygen(1): Fix regression bz#2306: Key path names were\n being lost as comment fields.\n\n * ssh(1): Allow ssh_config Port options set in the second config\n parse phase to be applied (they were being ignored). bz#2286\n\n * ssh(1): Tweak config re-parsing with host canonicalisation - make\n the second pass through the config files always run when host name\n canonicalisation is enabled (and not whenever the host name\n changes) bz#2267\n\n * ssh(1): Fix passing of wildcard forward bind addresses when\n connection multiplexing is in use; bz#2324;\n\n * ssh-keygen(1): Fix broken private key conversion from non-OpenSSH\n formats; bz#2345.\n\n * ssh-keygen(1): Fix KRL generation bug when multiple CAs are in\n use.\n\n * Various fixes to manual pages: bz#2288, bz#2316, bz#2273\n\nPortable OpenSSH\n\n * Support --without-openssl at configure time\n\n Disables and removes dependency on OpenSSL. Many features,\n including SSH protocol 1 are not supported and the set of crypto\n options is greatly restricted. This will only work on systems\n with native arc4random or /dev/urandom.\n\n Considered highly experimental for now.\n\n * Support --without-ssh1 option at configure time\n\n Allows disabling support for SSH protocol 1.\n\n * sshd(8): Fix compilation on systems with IPv6 support in utmpx; bz#2296\n\n * Allow custom service name for sshd on Cygwin. Permits the use of\n multiple sshd running with different service names.\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.8.tar.gz) = 99903c6ca76e0a2c044711017f81127e12459d37\n - SHA256 (openssh-6.8.tar.gz) = N1uzVarFbrm2CzAwuDu3sRoszmqpK+5phAChP/QNyuw=\n\n - SHA1 (openssh-6.8p1.tar.gz) = cdbc51e46a902b30d263b05fdc71340920e91c92\n - SHA256 (openssh-6.8p1.tar.gz) = P/ZM5z7hJEgLW/dnuYMNfTwDu8tqvnFrePAZLDfOFg4=\n\nPlease note that the PGP key used to sign releases was recently rotated.\nThe new key has been signed by the old key to provide continuity. It is\navailable from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.7 was released on 2014-10-06. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 6.6\n=========================\n\nPotentially-incompatible changes\n\n * sshd(8): The default set of ciphers and MACs has been altered to\n remove unsafe algorithms. In particular, CBC ciphers and arcfour*\n are disabled by default.\n\n The full set of algorithms remains available if configured\n explicitly via the Ciphers and MACs sshd_config options.\n\n * sshd(8): Support for tcpwrappers/libwrap has been removed.\n\n * OpenSSH 6.5 and 6.6 have a bug that causes ~0.2% of connections\n using the curve25519-sha256@libssh.org KEX exchange method to fail\n when connecting with something that implements the specification\n correctly. OpenSSH 6.7 disables this KEX method when speaking to\n one of the affected versions.\n\nNew Features\n\n * Major internal refactoring to begin to make part of OpenSSH usable\n as a library. So far the wire parsing, key handling and KRL code\n has been refactored. Please note that we do not consider the API\n stable yet, nor do we offer the library in separable form.\n\n * ssh(1), sshd(8): Add support for Unix domain socket forwarding.\n A remote TCP port may be forwarded to a local Unix domain socket\n and vice versa or both ends may be a Unix domain socket.\n\n * ssh(1), ssh-keygen(1): Add support for SSHFP DNS records for\n ED25519 key types.\n\n * sftp(1): Allow resumption of interrupted uploads.\n\n * ssh(1): When rekeying, skip file/DNS lookups of the hostkey if it\n is the same as the one sent during initial key exchange; bz#2154\n\n * sshd(8): Allow explicit ::1 and 127.0.0.1 forwarding bind\n addresses when GatewayPorts=no; allows client to choose address\n family; bz#2222\n\n * sshd(8): Add a sshd_config PermitUserRC option to control whether\n ~/.ssh/rc is executed, mirroring the no-user-rc authorized_keys\n option; bz#2160\n\n * ssh(1): Add a %C escape sequence for LocalCommand and ControlPath\n that expands to a unique identifier based on a hash of the tuple of\n (local host, remote user, hostname, port). Helps avoid exceeding\n miserly pathname limits for Unix domain sockets in multiplexing\n control paths; bz#2220\n\n * sshd(8): Make the "Too many authentication failures" message\n include the user, source address, port and protocol in a format\n similar to the authentication success / failure messages; bz#2199\n\n * Added unit and fuzz tests for refactored code. These are run\n automatically in portable OpenSSH via the "make tests" target.\n\nBugfixes\n\n * sshd(8): Fix remote forwarding with the same listen port but\n different listen address.\n\n * ssh(1): Fix inverted test that caused PKCS#11 keys that were\n explicitly listed in ssh_config or on the commandline not to be\n preferred.\n\n * ssh-keygen(1): Fix bug in KRL generation: multiple consecutive\n revoked certificate serial number ranges could be serialised to an\n invalid format. Readers of a broken KRL caused by this bug will\n fail closed, so no should-have-been-revoked key will be accepted.\n\n * ssh(1): Reflect stdio-forward ("ssh -W host:port ...") failures in\n exit status. Previously we were always returning 0; bz#2255\n\n * ssh(1), ssh-keygen(1): Make Ed25519 keys' title fit properly in the\n randomart border; bz#2247\n\n * ssh-agent(1): Only cleanup agent socket in the main agent process\n and not in any subprocesses it may have started (e.g. forked\n askpass). Fixes agent sockets being zapped when askpass processes\n fatal(); bz#2236\n\n * ssh-add(1): Make stdout line-buffered; saves partial output getting\n lost when ssh-add fatal()s part-way through (e.g. when listing keys\n from an agent that supports key types that ssh-add doesn't);\n bz#2234\n\n * ssh-keygen(1): When hashing or removing hosts, don't choke on\n @revoked markers and don't remove @cert-authority markers; bz#2241\n\n * ssh(1): Don't fatal when hostname canonicalisation fails and a\n ProxyCommand is in use; continue and allow the ProxyCommand to\n connect anyway (e.g. to a host with a name outside the DNS behind\n a bastion)\n\n * scp(1): When copying local->remote fails during read, don't send\n uninitialised heap to the remote end.\n\n * sftp(1): Fix fatal "el_insertstr failed" errors when tab-completing\n filenames with a single quote char somewhere in the string;\n bz#2238\n\n * ssh-keyscan(1): Scan for Ed25519 keys by default.\n\n * ssh(1): When using VerifyHostKeyDNS with a DNSSEC resolver, down-\n convert any certificate keys to plain keys and attempt SSHFP\n resolution. Prevents a server from skipping SSHFP lookup and\n forcing a new-hostkey dialog by offering only certificate keys.\n \n * sshd(8): Avoid crash at exit via NULL pointer reference; bz#2225\n\n * Fix some strict-alignment errors.\n\nPortable OpenSSH\n\n * Portable OpenSSH now supports building against libressl-portable.\n\n * Portable OpenSSH now requires openssl 0.9.8f or greater. Older\n versions are no longer supported.\n\n * In the OpenSSL version check, allow fix version upgrades (but not\n downgrades. Debian bug #748150.\n\n * sshd(8): On Cygwin, determine privilege separation user at runtime,\n since it may need to be a domain account.\n\n * sshd(8): Don't attempt to use vhangup on Linux. It doesn't work for\n non-root users, and for them it just messes up the tty settings.\n\n * Use CLOCK_BOOTTIME in preference to CLOCK_MONOTONIC when it is\n available. It considers time spent suspended, thereby ensuring\n timeouts (e.g. for expiring agent keys) fire correctly. bz#2228\n\n * Add support for ed25519 to opensshd.init init script.\n\n * sftp-server(8): On platforms that support it, use prctl() to\n prevent sftp-server from accessing /proc/self/{mem,maps}\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.7.tar.gz) = 315497b27a0186e4aef67987cfc9f3d9ba561cd8\n - SHA256 (openssh-6.7.tar.gz) = /me/hPxDw9Tfd3siNKQubSQph84qiKwftiMsgj6nh5E=\n\n - SHA1 (openssh-6.7p1.tar.gz) = 14e5fbed710ade334d65925e080d1aaeb9c85bf6\n - SHA256 (openssh-6.7p1.tar.gz) = svg5Tq6Fjau9732sELma7ADJVGJ1PoA0LlMLu29yVQc=\n\nPlease note that the PGP key used to sign releases was recently rotated.\nThe new key has been signed by the old key to provide continuity. It is\navailable from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.6 was released on 2014-03-15. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 6.6\n=========================\n\nThis is primarily a bugfix release.\n\nSecurity:\n\n * sshd(8): when using environment passing with a sshd_config(5)\n AcceptEnv pattern with a wildcard. OpenSSH prior to 6.6 could be\n tricked into accepting any environment variable that contains the\n characters before the wildcard character.\n\nNew / changed features:\n\n * ssh(1), sshd(8): this release removes the J-PAKE authentication code.\n This code was experimental, never enabled and had been unmaintained\n for some time.\n\n * ssh(1): when processing Match blocks, skip 'exec' clauses other clauses\n predicates failed to match.\n\n * ssh(1): if hostname canonicalisation is enabled and results in the\n destination hostname being changed, then re-parse ssh_config(5) files\n using the new destination hostname. This gives 'Host' and 'Match'\n directives that use the expanded hostname a chance to be applied.\n\nBugfixes:\n\n * ssh(1): avoid spurious "getsockname failed: Bad file descriptor" in\n ssh -W. bz#2200, debian#738692\n\n * sshd(8): allow the shutdown(2) syscall in seccomp-bpf and systrace\n sandbox modes, as it is reachable if the connection is terminated\n during the pre-auth phase.\n\n * ssh(1), sshd(8): fix unsigned overflow that in SSH protocol 1 bignum\n parsing. Minimum key length checks render this bug unexploitable to\n compromise SSH 1 sessions.\n\n * sshd_config(5): clarify behaviour of a keyword that appears in\n multiple matching Match blocks. bz#2184\n\n * ssh(1): avoid unnecessary hostname lookups when canonicalisation is\n disabled. bz#2205\n\n * sshd(8): avoid sandbox violation crashes in GSSAPI code by caching\n the supported list of GSSAPI mechanism OIDs before entering the\n sandbox. bz#2107\n\n * ssh(1): fix possible crashes in SOCKS4 parsing caused by assumption\n that the SOCKS username is nul-terminated.\n\n * ssh(1): fix regression for UsePrivilegedPort=yes when BindAddress is\n not specified.\n\n * ssh(1), sshd(8): fix memory leak in ECDSA signature verification.\n\n * ssh(1): fix matching of 'Host' directives in ssh_config(5) files\n to be case-insensitive again (regression in 6.5).\n\nPortable OpenSSH:\n\n * sshd(8): don't fatal if the FreeBSD Capsicum is offered by the\n system headers and libc but is not supported by the kernel.\n * Fix build using the HP-UX compiler.\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.6.tar.gz) = bf932d798324ff2502409d3714d0ad8d65c7e1e7\n - SHA256 (openssh-6.6.tar.gz) = jaSJE5aiQRm+91dV6EvVGr/ozo33tbxyjjFSiu+Cy80=\n\n - SHA1 (openssh-6.6p1.tar.gz) = b850fd1af704942d9b3c2eff7ef6b3a59b6a6b6e\n - SHA256 (openssh-6.6p1.tar.gz) = SMHwZktFNIdQOABMxPNVW4MpwqgcHfSNtcUXgA3iA7s=\n\nPlease note that the PGP key used to sign releases was recently rotated.\nThe new key has been signed by the old key to provide continuity. It is\navailable from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nChanges since OpenSSH 6.4\n=========================\n\nThis is a feature-focused release.\n\nNew features:\n\n * ssh(1), sshd(8): Add support for key exchange using elliptic-curve\n Diffie Hellman in Daniel Bernstein's Curve25519. This key exchange\n method is the default when both the client and server support it.\n\n * ssh(1), sshd(8): Add support for Ed25519 as a public key type.\n Ed25519 is a elliptic curve signature scheme that offers\n better security than ECDSA and DSA and good performance. It may be\n used for both user and host keys.\n\n * Add a new private key format that uses a bcrypt KDF to better\n protect keys at rest. This format is used unconditionally for\n Ed25519 keys, but may be requested when generating or saving\n existing keys of other types via the -o ssh-keygen(1) option.\n We intend to make the new format the default in the near future.\n Details of the new format are in the PROTOCOL.key file.\n\n * ssh(1), sshd(8): Add a new transport cipher\n "chacha20-poly1305@openssh.com" that combines Daniel Bernstein's\n ChaCha20 stream cipher and Poly1305 MAC to build an authenticated\n encryption mode. Details are in the PROTOCOL.chacha20poly1305 file.\n\n * ssh(1), sshd(8): Refuse RSA keys from old proprietary clients and\n servers that use the obsolete RSA+MD5 signature scheme. It will\n still be possible to connect with these clients/servers but only\n DSA keys will be accepted, and OpenSSH will refuse connection\n entirely in a future release.\n\n * ssh(1), sshd(8): Refuse old proprietary clients and servers that\n use a weaker key exchange hash calculation.\n\n * ssh(1): Increase the size of the Diffie-Hellman groups requested\n for each symmetric key size. New values from NIST Special\n Publication 800-57 with the upper limit specified by RFC4419.\n\n * ssh(1), ssh-agent(1): Support PKCS#11 tokens that only provide\n X.509 certs instead of raw public keys (requested as bz#1908).\n\n * ssh(1): Add a ssh_config(5) "Match" keyword that allows\n conditional configuration to be applied by matching on hostname,\n user and result of arbitrary commands.\n\n * ssh(1): Add support for client-side hostname canonicalisation\n using a set of DNS suffixes and rules in ssh_config(5). This\n allows unqualified names to be canonicalised to fully-qualified\n domain names to eliminate ambiguity when looking up keys in\n known_hosts or checking host certificate names.\n\n * sftp-server(8): Add the ability to whitelist and/or blacklist sftp\n protocol requests by name.\n\n * sftp-server(8): Add a sftp "fsync@openssh.com" to support calling\n fsync(2) on an open file handle.\n\n * sshd(8): Add a ssh_config(5) PermitTTY to disallow TTY allocation,\n mirroring the longstanding no-pty authorized_keys option.\n\n * ssh(1): Add a ssh_config ProxyUseFDPass option that supports the\n use of ProxyCommands that establish a connection and then pass a\n connected file descriptor back to ssh(1). This allows the\n ProxyCommand to exit rather than staying around to transfer data.\n\nBugfixes:\n\n * ssh(1), sshd(8): Fix potential stack exhaustion caused by nested\n certificates.\n\n * ssh(1): bz#1211: make BindAddress work with UsePrivilegedPort.\n\n * sftp(1): bz#2137: fix the progress meter for resumed transfer.\n\n * ssh-add(1): bz#2187: do not request smartcard PIN when removing\n keys from ssh-agent.\n\n * sshd(8): bz#2139: fix re-exec fallback when original sshd binary\n cannot be executed.\n\n * ssh-keygen(1): Make relative-specified certificate expiry times\n relative to current time and not the validity start time.\n\n * sshd(8): bz#2161: fix AuthorizedKeysCommand inside a Match block.\n\n * sftp(1): bz#2129: symlinking a file would incorrectly canonicalise\n the target path.\n\n * ssh-agent(1): bz#2175: fix a use-after-free in the PKCS#11 agent\n helper executable.\n\n * sshd(8): Improve logging of sessions to include the user name,\n remote host and port, the session type (shell, command, etc.) and\n allocated TTY (if any).\n\n * sshd(8): bz#1297: tell the client (via a debug message) when\n their preferred listen address has been overridden by the\n server's GatewayPorts setting.\n\n * sshd(8): bz#2162: include report port in bad protocol banner\n message.\n\n * sftp(1): bz#2163: fix memory leak in error path in do_readdir().\n\n * sftp(1): bz#2171: don't leak file descriptor on error.\n\n * sshd(8): Include the local address and port in "Connection from\n ..." message (only shown at loglevel>=verbose).\n\nPortable OpenSSH:\n\n * Please note that this is the last version of Portable OpenSSH that\n will support versions of OpenSSL prior to 0.9.6. Support (i.e.\n SSH_OLD_EVP) will be removed following the 6.5p1 release.\n\n * Portable OpenSSH will attempt compile and link as a Position\n Independent Executable on Linux, OS X and OpenBSD on recent gcc-\n like compilers. Other platforms and older/other compilers may\n request this using the --with-pie configure flag.\n\n * A number of other toolchain-related hardening options are used\n automatically if available, including -ftrapv to abort on signed\n integer overflow and options to write-protect dynamic linking\n information. The use of these options may be disabled using the\n --without-hardening configure flag.\n\n * If the toolchain supports it, one of the -fstack-protector-strong,\n -fstack-protector-all or -fstack-protector compilation flag are\n used to add guards to mitigate attacks based on stack overflows.\n The use of these options may be disabled using the\n --without-stackprotect configure option.\n\n * sshd(8): Add support for pre-authentication sandboxing using the\n Capsicum API introduced in FreeBSD 10.\n\n * Switch to a ChaCha20-based arc4random() PRNG for platforms that do\n not provide their own.\n\n * sshd(8): bz#2156: restore Linux oom_adj setting when handling\n SIGHUP to maintain behaviour over restart.\n\n * sshd(8): bz#2032: use local username in krb5_kuserok check rather\n than full client name which may be of form user@REALM.\n\n * ssh(1), sshd(8): Test for both the presence of ECC NID numbers in\n OpenSSL and that they actually work. Fedora (at least) has\n NID_secp521r1 that doesn't work.\n\n * bz#2173: use pkg-config --libs to include correct -L location for\n libedit.\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.5.tar.gz) = 0a375e20d895670489a9241f8faa57670214fbed\n - SHA256 (openssh-6.5.tar.gz) = sK5q2rB0o5JCbEmbeE/6N9DtJkT81dwmeuhogT4i900=\n\n - SHA1 (openssh-6.5p1.tar.gz) = 3363a72b4fee91b29cf2024ff633c17f6cd2f86d\n - SHA256 (openssh-6.5p1.tar.gz) = oRle1V25RSUtWhcw1KKipcHJpqoB7y5a91CpYmI9kCc=\n\nPlease note that the PGP key used to sign releases has been rotated.\nThe new key has been signed by the old key to provide continuity. It\nis available from the mirror sites as RELEASE_KEY.asc.\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nChanges since OpenSSH 6.3\n=========================\n\nThis release fixes a security bug:\n\n * sshd(8): fix a memory corruption problem triggered during rekeying\n when an AES-GCM cipher is selected. Full details of the vulnerability\n are available at: http://www.openssh.com/txt/gcmrekey.adv\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.4.tar.gz) = 4caf1a50eb3a3da821c16298c4aaa576fe24210c\n - SHA1 (openssh-6.4p1.tar.gz) = cf5fe0eb118d7e4f9296fbc5d6884965885fc55d\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nChanges since OpenSSH 6.2\n=========================\n\nThis release is predominantly a bugfix release:\n\nFeatures:\n\n * sshd(8): add ssh-agent(1) support to sshd(8); allows encrypted hostkeys,\n or hostkeys on smartcards.\n\n * ssh(1)/sshd(8): allow optional time-based rekeying via a second argument\n to the existing RekeyLimit option. RekeyLimit is now supported in\n sshd_config as well as on the client.\n\n * sshd(8): standardise logging of information during user authentication.\n\n The presented key/cert and the remote username (if available) is now\n logged in the authentication success/failure message on the same log\n line as the local username, remote host/port and protocol in use.\n Certificates contents and the key fingerprint of the signing CA are\n logged too.\n\n Including all relevant information on a single line simplifies log\n analysis as it is no longer necessary to relate information scattered\n across multiple log entries.\n\n * ssh(1): add the ability to query which ciphers, MAC algorithms, key\n types and key exchange methods are supported in the binary.\n\n * ssh(1): support ProxyCommand=- to allow support cases where stdin and\n stdout already point to the proxy.\n\n * ssh(1): allow IdentityFile=none\n\n * ssh(1)/sshd(8): add -E option to ssh and sshd to append debugging logs\n to a specified file instead of stderr or syslog.\n\n * sftp(1): add support for resuming partial downloads using the "reget"\n command and on the sftp commandline or on the "get" commandline using\n the "-a" (append) option.\n\n * ssh(1): add an "IgnoreUnknown" configuration option to selectively\n suppress errors arising from unknown configuration directives.\n\n * sshd(8): add support for submethods to be appended to required\n authentication methods listed via AuthenticationMethods.\n\nBugfixes:\n\n * sshd(8): fix refusal to accept certificate if a key of a different type\n to the CA key appeared in authorized_keys before the CA key.\n\n * ssh(1)/ssh-agent(1)/sshd(8): Use a monotonic time source for timers so\n that things like keepalives and rekeying will work properly over clock\n steps.\n\n * sftp(1): update progressmeter when data is acknowledged, not when it's\n sent. bz#2108\n\n * ssh(1)/ssh-keygen(1): improve error messages when the current user does\n not exist in /etc/passwd; bz#2125\n\n * ssh(1): reset the order in which public keys are tried after partial\n authentication success.\n\n * ssh-agent(1): clean up socket files after SIGINT when in debug mode;\n bz#2120\n\n * ssh(1) and others: avoid confusing error messages in the case of broken\n system resolver configurations; bz#2122\n\n * ssh(1): set TCP nodelay for connections started with -N; bz#2124\n\n * ssh(1): correct manual for permission requirements on ~/.ssh/config;\n bz#2078\n\n * ssh(1): fix ControlPersist timeout not triggering in cases where TCP\n connections have hung. bz#1917\n\n * ssh(1): properly detach a ControlPersist master from its controlling\n terminal.\n\n * sftp(1): avoid crashes in libedit when it has been compiled with multi-\n byte character support. bz#1990\n\n * sshd(8): when running sshd -D, close stderr unless we have explicitly\n requested logging to stderr. bz#1976,\n\n * ssh(1): fix incomplete bzero; bz#2100\n\n * sshd(8): log and error and exit if ChrootDirectory is specified and\n running without root privileges.\n\n * Many improvements to the regression test suite. In particular log files\n are now saved from ssh and sshd after failures.\n\n * Fix a number of memory leaks. bz#1967 bz#2096 and others\n\n * sshd(8): fix public key authentication when a :style is appended to\n the requested username.\n\n * ssh(1): do not fatally exit when attempting to cleanup multiplexing-\n created channels that are incompletely opened. bz#2079\n\nPortable OpenSSH:\n\n * Major overhaul of contrib/cygwin/README\n\n * Fix unaligned accesses in umac.c for strict-alignment architectures.\n bz#2101\n\n * Enable -Wsizeof-pointer-memaccess if the compiler supports it. bz#2100\n\n * Fix broken incorrect commandline reporting errors. bz#1448\n\n * Only include SHA256 and ECC-based key exchange methods if libcrypto has\n the required support.\n\n * Fix crash in SOCKS5 dynamic forwarding code on strict-alignment\n architectures.\n\n * A number of portability fixes for Android:\n * Don't try to use lastlog on Android; bz#2111\n * Fall back to using openssl's DES_crypt function on platforms that don't\n have a native crypt() function; bz#2112\n * Test for fd_mask, howmany and NFDBITS rather than trying to enumerate\n the platforms that don't have them. bz#2085\n * Replace S_IWRITE, which isn't standardized, with S_IWUSR, which is.\n bz#2085\n * Add a null implementation of endgrent for platforms that don't have\n it (eg Android) bz#2087\n * Support platforms, such as Android, that lack struct passwd.pw_gecos.\n bz#2086\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.3.tar.gz) = 8a6ef99ffc80c19e9afe9fe1e857370f6adcf450\n - SHA1 (openssh-6.3p1.tar.gz) = 70845ca79474258cab29dbefae13d93e41a83ccb\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nChanges since OpenSSH 6.2p1\n===========================\n\nThis is a bugfix release:\n\nBugfixes:\n\n * ssh(1): Only warn for missing identity files that were explicitly\n specified.\n\n * Fix bug in contributed contrib/ssh-copy-id script that could result in\n "rm *" being called on mktemp failure. bz#2105\n\n * sshd(8): Quiet disconnect notifications on the server from error() back\n to logit() from error() for normal, client-initiated disconnections.\n bz#2057\n\n * Avoid conflicting definitions of __int64 on Cygwin\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.2p2.tar.gz) = c2b4909eba6f5ec6f9f75866c202db47f3b501ba\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nChanges since OpenSSH 6.1\n=========================\n\nThis release introduces a number of new features:\n\nFeatures:\n\n * ssh(1)/sshd(8): Added support for AES-GCM authenticated encryption in\n SSH protocol 2. The new cipher is available as aes128-gcm@openssh.com\n and aes256-gcm@openssh.com. It uses an identical packet format to the\n AES-GCM mode specified in RFC 5647, but uses simpler and different\n selection rules during key exchange.\n\n * ssh(1)/sshd(8): Added support for encrypt-then-mac (EtM) MAC modes\n for SSH protocol 2. These modes alter the packet format and compute\n the MAC over the packet length and encrypted packet rather than over\n the plaintext data. These modes are considered more secure and are\n used by default when available.\n\n * ssh(1)/sshd(8): Added support for the UMAC-128 MAC as\n "umac-128@openssh.com" and "umac-128-etm@openssh.com". The latter\n being an encrypt-then-mac mode.\n\n * sshd(8): Added support for multiple required authentication in SSH\n protocol 2 via an AuthenticationMethods option. This option lists\n one or more comma-separated lists of authentication method names.\n Successful completion of all the methods in any list is required for\n authentication to complete. This allows, for example, requiring a\n user having to authenticate via public key or GSSAPI before they\n are offered password authentication.\n\n * sshd(8)/ssh-keygen(1): Added support for Key Revocation Lists\n (KRLs), a compact binary format to represent lists of revoked keys\n and certificates that take as little as one bit per certificate when\n revoking by serial number. KRLs may be generated using ssh-keygen(1)\n and are loaded into sshd(8) via the existing RevokedKeys sshd_config\n option.\n\n * ssh(1): IdentitiesOnly now applies to keys obtained from a\n PKCS11Provider. This allows control of which keys are offered from\n tokens using IdentityFile.\n\n * sshd(8): sshd_config(5)'s AllowTcpForwarding now accepts "local"\n and "remote" in addition to its previous "yes"/"no" keywords to allow\n the server to specify whether just local or remote TCP forwarding is\n enabled.\n\n * sshd(8): Added a sshd_config(5) option AuthorizedKeysCommand to\n support fetching authorized_keys from a command in addition to (or\n instead of) from the filesystem. The command is run under an account\n specified by an AuthorizedKeysCommandUser sshd_config(5) option.\n\n * sftp-server(8): Now supports a -d option to allow the starting\n directory to be something other than the user's home directory.\n\n * ssh-keygen(1): Now allows fingerprinting of keys hosted in PKCS#11\n tokens using "ssh-keygen -lD pkcs11_provider".\n\n * ssh(1): When SSH protocol 2 only is selected (the default), ssh(1)\n now immediately sends its SSH protocol banner to the server without\n waiting to receive the server's banner, saving time when connecting.\n\n * ssh(1): Added ~v and ~V escape sequences to raise and lower the\n logging level respectively.\n\n * ssh(1): Made the escape command help (~?) context sensitive so that\n only commands that will work in the current session are shown.\n\n * ssh-keygen(1): When deleting host lines from known_hosts using\n "ssh-keygen -R host", ssh-keygen(1) now prints details of which lines\n were removed.\n \nBugfixes:\n\n * ssh(1): Force a clean shutdown of ControlMaster client sessions when\n the ~. escape sequence is used. This means that ~. should now work in\n mux clients even if the server is no longer responding.\n\n * ssh(1): Correctly detect errors during local TCP forward setup in\n multiplexed clients. bz#2055\n\n * ssh-add(1): Made deleting explicit keys "ssh-add -d" symmetric with\n adding keys with respect to certificates. It now tries to delete the\n corresponding certificate and respects the -k option to allow deleting\n of the key only.\n\n * sftp(1): Fix a number of parsing and command-editing bugs, including\n bz#1956\n\n * ssh(1): When muxmaster is run with -N, ensured that it shuts down\n gracefully when a client sends it "-O stop" rather than hanging around.\n bz#1985\n\n * ssh-keygen(1): When screening moduli candidates, append to the file\n rather than overwriting to allow resumption. bz#1957\n\n * ssh(1): Record "Received disconnect" messages at ERROR rather than\n INFO priority. bz#2057.\n\n * ssh(1): Loudly warn if explicitly-provided private key is unreadable.\n bz#1981\n\nPortable OpenSSH:\n\n * sshd(8): The Linux seccomp-filter sandbox is now supported on ARM\n platforms where the kernel supports it.\n\n * sshd(8): The seccomp-filter sandbox will not be enabled if the system\n headers support it at compile time, regardless of whether it can be\n enabled then. If the run-time system does not support seccomp-filter,\n sshd will fall back to the rlimit pseudo-sandbox.\n\n * ssh(1): Don't link in the Kerberos libraries. They aren't necessary\n on the client, just on sshd(8). bz#2072\n\n * Fix GSSAPI linking on Solaris, which uses a differently-named GSSAPI\n library. bz#2073\n\n * Fix compilation on systems with openssl-1.0.0-fips.\n\n * Fix a number of errors in the RPM spec files.\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.2.tar.gz) = b3f6cd774d345f22f6d0038cc9464cce131a0676\n - SHA1 (openssh-6.2p1.tar.gz) = 8824708c617cc781b2bb29fa20bd905fd3d2a43d\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.1 was released on 2012-08-29. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 6.0\n=========================\n\nThis is primarily a bugfix release.\n\nFeatures:\n\n * sshd(8): This release turns on pre-auth sandboxing sshd by default for\n new installs, by setting UsePrivilegeSeparation=sandbox in sshd_config.\n * ssh-keygen(1): Add options to specify starting line number and number of\n lines to process when screening moduli candidates, allowing processing\n of different parts of a candidate moduli file in parallel\n * sshd(8): The Match directive now supports matching on the local (listen)\n address and port upon which the incoming connection was received via\n LocalAddress and LocalPort clauses.\n * sshd(8): Extend sshd_config Match directive to allow setting AcceptEnv\n and {Allow,Deny}{Users,Groups}\n * Add support for RFC6594 SSHFP DNS records for ECDSA key types. bz#1978\n * ssh-keygen(1): Allow conversion of RSA1 keys to public PEM and PKCS8\n * sshd(8): Allow the sshd_config PermitOpen directive to accept "none" as\n an argument to refuse all port-forwarding requests.\n * sshd(8): Support "none" as an argument for AuthorizedPrincipalsFile\n * ssh-keyscan(1): Look for ECDSA keys by default. bz#1971\n * sshd(8): Add "VersionAddendum" to sshd_config to allow server operators\n to append some arbitrary text to the server SSH protocol banner.\n\nBugfixes:\n\n * ssh(1)/sshd(8): Don't spin in accept() in situations of file\n descriptor exhaustion. Instead back off for a while.\n * ssh(1)/sshd(8): Remove hmac-sha2-256-96 and hmac-sha2-512-96 MACs as\n they were removed from the specification. bz#2023,\n * sshd(8): Handle long comments in config files better. bz#2025\n * ssh(1): Delay setting tty_flag so RequestTTY options are correctly\n picked up. bz#1995\n * sshd(8): Fix handling of /etc/nologin incorrectly being applied to root\n on platforms that use login_cap.\n\nPortable OpenSSH:\n\n * sshd(8): Allow sshd pre-auth sandboxing to fall-back to the rlimit\n sandbox from the Linux SECCOMP filter sandbox when the latter is\n not available in the kernel.\n * ssh(1): Fix NULL dereference when built with LDNS and using DNSSEC to\n retrieve a CNAME SSHFP record.\n * Fix cross-compilation problems related to pkg-config. bz#1996\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.1.tar.gz) = 7ed5b491cfebcaee2273d1f872314107273c2167\n - SHA1 (openssh-6.1p1.tar.gz) = 751c92c912310c3aa9cadc113e14458f843fc7b3\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 6.0 was released on 2012-04-22. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.9\n=========================\n\nThis is primarily a bugfix release.\n\nFeatures:\n\n * ssh-keygen(1): Add optional checkpoints for moduli screening\n * ssh-add(1): new -k option to load plain keys (skipping certificates)\n * sshd(8): Add wildcard support to PermitOpen, allowing things like\n "PermitOpen localhost:*". bz #1857\n * ssh(1): support for cancelling local and remote port forwards via the\n multiplex socket. Use ssh -O cancel -L xx:xx:xx -R yy:yy:yy user@host"\n to request the cancellation of the specified forwardings\n * support cancellation of local/dynamic forwardings from ~C commandline\n\nBugfixes:\n\n * ssh(1): ensure that $DISPLAY contains only valid characters before\n using it to extract xauth data so that it can't be used to play local\n shell metacharacter games.\n * ssh(1): unbreak remote portforwarding with dynamic allocated listen ports\n * scp(1): uppress adding '--' to remote commandlines when the first\n argument does not start with '-'. saves breakage on some\n difficult-to-upgrade embedded/router platforms\n * ssh(1)/sshd(8): fix typo in IPQoS parsing: there is no "AF14" class,\n but there is an "AF21" class\n * ssh(1)/sshd(8): do not permit SSH2_MSG_SERVICE_REQUEST/ACCEPT during\n rekeying\n * ssh(1): skip attempting to create ~/.ssh when -F is passed\n * sshd(8): unbreak stdio forwarding when ControlPersist is in use; bz#1943\n * sshd(1): send tty break to pty master instead of (probably already\n closed) slave side; bz#1859\n * sftp(1): silence error spam for "ls */foo" in directory with files;\n bz#1683\n * Fixed a number of memory and file descriptor leaks\n\nPortable OpenSSH:\n\n * Add a new privilege separation sandbox implementation for Linux's\n new seccomp sandbox, automatically enabled on platforms that support\n it. (Note: privilege separation sandboxing is still experimental)\n * Fix compilation problems on FreeBSD, where libutil contained openpty()\n but not login().\n * ssh-keygen(1): don't fail in -A on platforms that don't support ECC\n * Add optional support for LDNS, a BSD licensed DNS resolver library\n which supports DNSSEC\n * Relax OpenSSL version check to allow running OpenSSH binaries on\n systems with OpenSSL libraries with a newer "fix" or "patch" level\n than the binaries were originally compiled on (previous check only\n allowed movement within "patch" releases). bz#1991\n * Fix builds using contributed Redhat spec file. bz#1992\n\nChecksums:\n==========\n\n - SHA1 (openssh-6.0.tar.gz) = 5d30aba0423c44e89924bb44c5d2153635506a9f\n - SHA1 (openssh-6.0p1.tar.gz) = f691e53ef83417031a2854b8b1b661c9c08e4422\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.9 was released on 2011-09-06. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.8\n=========================\n\nFeatures:\n\n * Introduce sandboxing of the pre-auth privsep child using an optional\n sshd_config(5) "UsePrivilegeSeparation=sandbox" mode that enables\n mandatory restrictions on the syscalls the privsep child can perform.\n This intention is to prevent a compromised privsep child from being\n used to attack other hosts (by opening sockets and proxying) or\n probing local kernel attack surface.\n \n Three concrete sandbox implementation are provided (selected at\n configure time): systrace, seatbelt and rlimit.\n\n The systrace sandbox uses systrace(4) in unsupervised "fast-path"\n mode, where a list of permitted syscalls is supplied. Any syscall not\n on the list results in SIGKILL being sent to the privsep child. Note\n that this requires a kernel with the new SYSTR_POLICY_KILL option\n (only OpenBSD has this mode at present).\n \n The seatbelt sandbox uses OS X/Darwin sandbox(7) facilities with a\n strict (kSBXProfilePureComputation) policy that disables access to\n filesystem and network resources.\n\n The rlimit sandbox is a fallback choice for platforms that don't\n support a better one; it uses setrlimit() to reset the hard-limit\n of file descriptors and processes to zero, which should prevent\n the privsep child from forking or opening new network connections.\n\n Sandboxing of the privilege separated child process is currently\n experimental but should become the default in a future release.\n Native sandboxes for other platforms are welcome (e.g. Capsicum,\n Linux pid/net namespaces, etc.)\n\n * Add new SHA256-based HMAC transport integrity modes from\n http://www.ietf.org/id/draft-dbider-sha2-mac-for-ssh-02.txt\n These modes are hmac-sha2-256, hmac-sha2-256-96, hmac-sha2-512,\n and hmac-sha2-512-96, and are available by default in ssh(1) and\n sshd(8)\n\n * The pre-authentication sshd(8) privilege separation slave process\n now logs via a socket shared with the master process, avoiding the\n need to maintain /dev/log inside the chroot.\n\n * ssh(1) now warns when a server refuses X11 forwarding\n\n * sshd_config(5)'s AuthorizedKeysFile now accepts multiple paths,\n separated by whitespace. The undocumented AuthorizedKeysFile2\n option is deprecated (though the default for AuthorizedKeysFile\n includes .ssh/authorized_keys2)\n\n * sshd_config(5): similarly deprecate UserKnownHostsFile2 and\n GlobalKnownHostsFile2 by making UserKnownHostsFile and\n GlobalKnownHostsFile accept multiple options and default to\n include known_hosts2\n\n * Retain key comments when loading v.2 keys. These will be visible\n in "ssh-add -l" and other places. bz#439\n\n * ssh(1) and sshd(8): set IPv6 traffic class from IPQoS (as well as\n IPv4 ToS/DSCP). bz#1855\n\n * ssh_config(5)'s ControlPath option now expands %L to the host\n portion of the destination host name.\n\n * ssh_config(5) "Host" options now support negated Host matching, e.g.\n \n Host *.example.org !c.example.org\n User mekmitasdigoat\n \n Will match "a.example.org", "b.example.org", but not "c.example.org"\n\n * ssh_config(5): a new RequestTTY option provides control over when a\n TTY is requested for a connection, similar to the existing -t/-tt/-T\n ssh(1) commandline options.\n\n * sshd(8): allow GSSAPI authentication to detect when a server-side\n failure causes authentication failure and don't count such failures\n against MaxAuthTries; bz#1244\n\n * ssh-keygen(1): Add -A option. For each of the key types (rsa1, rsa,\n dsa and ecdsa) for which host keys do not exist, generate the host\n keys with the default key file path, an empty passphrase, default\n bits for the key type, and default comment. This is useful for\n system initialisation scripts.\n\n * ssh(1): Allow graceful shutdown of multiplexing: request that a mux\n server removes its listener socket and refuse future multiplexing\n requests but don't kill existing connections. This may be requested\n using "ssh -O stop ..."\n\n * ssh-add(1) now accepts keys piped from standard input. E.g.\n "ssh-add - < /path/to/key"\n \n * ssh-keysign(8) now signs hostbased authentication\n challenges correctly using ECDSA keys; bz#1858\n\n * sftp(1): document that sftp accepts square brackets to delimit\n addresses (useful for IPv6); bz#1847a\n\n * ssh(1): when using session multiplexing, the master process will\n change its process title to reflect the control path in use and\n when a ControlPersist-ed master is waiting to close; bz#1883 and\n bz#1911\n\n * Other minor bugs fixed: 1849 1861 1862 1869 1875 1878 1879 1892\n 1900 1905 1913\n\nPortable OpenSSH Bugfixes:\n\n * Fix a compilation error in the SELinux support code. bz#1851\n\n * This release removes support for ssh-rand-helper. OpenSSH now\n obtains its random numbers directly from OpenSSL or from\n a PRNGd/EGD instance specified at configure time.\n\n * sshd(8) now resets the SELinux process execution context before\n executing passwd for password changes; bz#1891\n\n * Since gcc >= 4.x ignores all -Wno-options options, test only the\n corresponding -W-option when trying to determine whether it is\n accepted; bz#1901\n\n * Add ECDSA key generation to the Cygwin ssh-{host,user}-config\n scripts.\n\n * Updated .spec and init files for Linux; bz#1920\n\n * Improved SELinux error messages in context change failures and\n suppress error messages when attempting to change from the\n "unconfined_t" type; bz#1924 bz#1919\n\n * Fix build errors on platforms without dlopen(); bz#1929\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.9.tar.gz) = bc0cb728bbc394769f9a2ce5b8cd99dc41e12632\n - SHA1 (openssh-5.9p1.tar.gz) = ac4e0055421e9543f0af5da607a72cf5922dcc56\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nPortable OpenSSH 5.8p2 was released on 2011-05-03. It will be available\nfrom the mirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.8p1\n===========================\n\nSecurity:\n\n * Fix local private host key compromise on platforms without host-\n level randomness support (e.g. /dev/random) reported by Tomas Mraz\n\n On hosts that did not have a randomness source configured in\n OpenSSL and were not configured to use EGD/PRNGd (using the\n --with-prngd-socket configure option), the ssh-rand-helper command\n was being implicitly executed by ssh-keysign with open file\n descriptors to the host private keys. An attacker could use\n ptrace(2) to attach to ssh-rand-helper and exfiltrate the keys.\n\n Most modern operating systems are not vulnerable. In particular,\n *BSD, Linux, OS X and Cygwin do not use ssh-rand-helper.\n\n A full advisory for this issue is available at:\n http://www.openssh.com/txt/portable-keysign-rand-helper.adv\n\nPortable OpenSSH Bugfixes:\n\n * Fix compilation failure when enabling SELinux support.\n\n * Revised Cygwin ssh-{host,user}-config that include ECDSA key\n support.\n\n * Revised Cygwin ssh-host-config to be more thorough in error checking\n and reporting.\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.8p2.tar.gz) = 64798328d310e4f06c9f01228107520adbc8b3e5\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.8 was released on 2011-02-04. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.7\n=========================\n\nSecurity:\n\n * Fix vulnerability in legacy certificate signing introduced in\n OpenSSH-5.6 and found by Mateusz Kocielski.\n\n Legacy certificates signed by OpenSSH 5.6 or 5.7 included data from\n the stack in place of a random nonce field. The contents of the stack\n do not appear to contain private data at this point, but this cannot\n be stated with certainty for all platform, library and compiler\n combinations. In particular, there exists a risk that some bytes from\n the privileged CA key may be accidentally included.\n\n A full advisory for this issue is available at:\n http://www.openssh.com/txt/legacy-cert.adv\n\nPortable OpenSSH Bugfixes:\n\n * Fix compilation failure when enabling SELinux support.\n\n * Do not attempt to call SELinux functions when SELinux is disabled.\n bz#1851\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.8.tar.gz) = 205dece2c8b41c69b082eb65320d359987aae25b\n - SHA1 (openssh-5.8p1.tar.gz) = adebb2faa9aba2a3a3c8b401b2b19677ab53f0de\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.7 was released on 2011-01-24. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.6\n=========================\n\nFeatures:\n\n * Implement Elliptic Curve Cryptography modes for key exchange (ECDH)\n and host/user keys (ECDSA) as specified by RFC5656. ECDH and ECDSA\n offer better performance than plain DH and DSA at the same equivalent\n symmetric key length, as well as much shorter keys.\n \n Only the mandatory sections of RFC5656 are implemented, specifically\n the three REQUIRED curves nistp256, nistp384 and nistp521 and only\n ECDH and ECDSA. Point compression (optional in RFC5656) is NOT\n implemented.\n \n Certificate host and user keys using the new ECDSA key types are\n supported - an ECDSA key may be certified, and an ECDSA key may act\n as a CA to sign certificates.\n\n ECDH in a 256 bit curve field is the preferred key agreement\n algorithm when both the client and server support it. ECDSA host\n keys are preferred when learning a host's keys for the first time,\n or can be learned using ssh-keyscan(1).\n \n * sftp(1)/sftp-server(8): add a protocol extension to support a hard\n link operation. It is available through the "ln" command in the\n client. The old "ln" behaviour of creating a symlink is available\n using its "-s" option or through the preexisting "symlink" command\n\n * scp(1): Add a new -3 option to scp: Copies between two remote hosts\n are transferred through the local host. Without this option the\n data is copied directly between the two remote hosts. \n\n * ssh(1): automatically order the hostkeys requested by the client\n based on which hostkeys are already recorded in known_hosts. This\n avoids hostkey warnings when connecting to servers with new ECDSA\n keys, since these are now preferred when learning hostkeys for the\n first time.\n\n * ssh(1)/sshd(8): add a new IPQoS option to specify arbitrary\n TOS/DSCP/QoS values instead of hardcoding lowdelay/throughput.\n bz#1733\n\n * sftp(1): the sftp client is now significantly faster at performing\n directory listings, using OpenBSD glob(3) extensions to preserve\n the results of stat(3) operations performed in the course of its\n execution rather than performing expensive round trips to fetch\n them again afterwards.\n\n * ssh(1): "atomically" create the listening mux socket by binding it on\n a temporary name and then linking it into position after listen() has\n succeeded. This allows the mux clients to determine that the server\n socket is either ready or stale without races. stale server sockets\n are now automatically removed. (also fixes bz#1711)\n\n * ssh(1)/sshd(8): add a KexAlgorithms knob to the client and server\n configuration to allow selection of which key exchange methods are\n used by ssh(1) and sshd(8) and their order of preference.\n\n * sftp(1)/scp(1): factor out bandwidth limiting code from scp(1) into\n a generic bandwidth limiter that can be attached using the atomicio\n callback mechanism and use it to add a bandwidth limit option to\n sftp(1). bz#1147\n \nBugFixes:\n\n * ssh(1)/ssh-agent(1): honour $TMPDIR for client xauth and ssh-agent\n temporary directories. bz#1809\n\n * ssh(1): avoid NULL deref on receiving a channel request on an unknown\n or invalid channel; bz#1842\n\n * sshd(8): remove a debug() that pollutes stderr on client connecting\n to a server in debug mode; bz#1719\n\n * scp(1): pass through ssh command-line flags and options when doing\n remote-remote transfers, e.g. to enable agent forwarding which is\n particularly useful in this case; bz#1837\n\n * sftp-server(8): umask should be parsed as octal\n\n * sftp(1): escape '[' in filename tab-completion\n\n * ssh(1): Typo in confirmation message. bz#1827\n\n * sshd(8): prevent free() of string in .rodata when overriding\n AuthorizedKeys in a Match block\n\n * sshd(8): Use default shell /bin/sh if $SHELL is ""\n\n * ssh(1): kill proxy command on fatal() (we already killed it on\n clean exit);\n\n * ssh(1): install a SIGCHLD handler to reap expired child process;\n bz#1812\n\n * Support building against openssl-1.0.0a\n\nPortable OpenSSH Bugfixes:\n\n * Use mandoc as preferred manpage formatter if it is present, followed\n by nroff and groff respectively.\n\n * sshd(8): Relax permission requirement on btmp logs to allow group\n read/write\n\n * bz#1840: fix warning when configuring --with-ssl-engine\n\n * sshd(8): Use correct uid_t/pid_t types instead of int. bz#1817\n\n * sshd(8): bz#1824: Add Solaris Project support.\n\n * sshd(8): Check is_selinux_enabled for exact return code since it can\n apparently return -1 under some conditions.\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.7.tar.gz) = 67cb91772a33fb3a004b39bcdb9148218365494c\n - SHA1 (openssh-5.7p1.tar.gz) = 423e27475f06e1055847dfff7f61e1ac632b5372\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.6 was released on 2010-08-23. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.5\n=========================\n\nFeatures:\n\n * Added a ControlPersist option to ssh_config(5) that automatically\n starts a background ssh(1) multiplex master when connecting. This\n connection can stay alive indefinitely, or can be set to\n automatically close after a user-specified duration of inactivity.\n\n * Hostbased authentication may now use certificate host keys. CA keys\n must be specified in a known_hosts file using the @cert-authority\n marker as described in sshd(8).\n\n * ssh-keygen(1) now supports signing certificate using a CA key that\n has been stored in a PKCS#11 token.\n\n * ssh(1) will now log the hostname and address that we connected to at\n LogLevel=verbose after authentication is successful to mitigate\n "phishing" attacks by servers with trusted keys that accept\n authentication silently and automatically before presenting fake\n password/passphrase prompts.\n\n Note that, for such an attack to be successful, the user must have\n disabled StrictHostKeyChecking (enabled by default) or an attacker\n must have access to a trusted host key for the destination server.\n\n * Expand %h to the hostname in ssh_config Hostname options. While this\n sounds useless, it is actually handy for working with unqualified\n hostnames:\n \n Host *.*\n Hostname %h\n Host *\n Hostname %h.example.org\n \n * Allow ssh-keygen(1) to import (-i) and export (-e) of PEM and PKCS#8\n keys in addition to RFC4716 (SSH.COM) encodings via a new -m option \n (bz#1749)\n\n * sshd(8) will now queue debug messages for bad ownership or\n permissions on the user's keyfiles encountered during authentication\n and will send them after authentication has successfully completed.\n These messages may be viewed in ssh(1) at LogLevel=debug or higher.\n\n * ssh(1) connection multiplexing now supports remote forwarding with\n dynamic port allocation and can report the allocated port back to\n the user:\n\n LPORT=`ssh -S muxsocket -R0:localhost:25 -O forward somehost`\n\n * sshd(8) now supports indirection in matching of principal names\n listed in certificates. By default, if a certificate has an\n embedded principals list then the username on the server must match\n one of the names in the list for it to be accepted for\n authentication.\n\n sshd(8) now has a new AuthorizedPrincipalsFile option to specify a\n file containing a list of names that may be accepted in place of the\n username when authorizing a certificate trusted via the\n sshd_config(5) TrustedCAKeys option. Similarly, authentication\n using a CA trusted in ~/.ssh/authorized_keys now accepts a\n principals="name1[,name2,...]" to specify a list of permitted names.\n \n If either option is absent, the current behaviour of requiring the\n username to appear in principals continues to apply. These options\n are useful for role accounts, disjoint account namespaces and\n "user@realm"-style naming policies in certificates.\n \n * Additional sshd_config(5) options are now valid inside Match blocks:\n\n AuthorizedKeysFile\n AuthorizedPrincipalsFile\n HostbasedUsesNameFromPacketOnly\n PermitTunnel\n\n * Revised the format of certificate keys. The new format, identified as\n ssh-{dss,rsa}-cert-v01@openssh.com includes the following changes:\n \n - Adding a serial number field. This may be specified by the CA at\n the time of certificate signing.\n\n - Moving the nonce field to the beginning of the certificate where\n it can better protect against chosen-prefix attacks on the\n signature hash (currently infeasible against the SHA1 hash used)\n \n - Renaming the "constraints" field to "critical options"\n \n - Addng a new non-critical "extensions" field. The "permit-*"\n options are now extensions, rather than critical options to\n permit non-OpenSSH implementation of this key format to degrade\n gracefully when encountering keys with options they do not\n recognize.\n \n The older format is still supported for authentication and may still\n be used when signing certificates (use "ssh-keygen -t v00 ...").\n The v00 format, introduced in OpenSSH 5.4, will be supported for at\n least one year from this release, after which it will be deprecated\n and removed.\n \nBugFixes:\n\n * The PKCS#11 code now retries a lookup for a private key if there is\n no matching key with CKA_SIGN attribute enabled; this fixes\n MuscleCard support (bz#1736)\n \n * Unbreak strdelim() skipping past quoted strings (bz#1757). For\n example, the following directive was not parsed correctly:\n\n AllowUsers "blah blah" blah\n\n * sftp(1): fix swapped args in upload_dir_internal(), breaking\n recursive upload depth checks and causing verbose printing of\n transfers to always be turned on (bz#1797)\n\n * Fix a longstanding problem where if you suspend scp(1) at the\n password/passphrase prompt the terminal mode is not restored.\n\n * Fix a PKCS#11 crash on some smartcards by validating the length\n returned for C_GetAttributValue (bz#1773)\n\n * sftp(1): fix ls in working directories that contain globbing\n characters in their pathnames (bz#1655)\n\n * Print warning for missing home directory when ChrootDirectory=none\n (bz#1564)\n\n * sftp(1): fix a memory leak in do_realpath() error path (bz#1771)\n\n * ssk-keygen(1): Standardise error messages when attempting to open\n private key files to include "progname: filename: error reason"\n (bz#1783)\n\n * Replace verbose and overflow-prone Linebuf code with\n read_keyfile_line() (bz#1565)\n\n * Include the user name on "subsystem request for ..." log messages\n\n * ssh(1) and sshd(8): remove hardcoded limit of 100 permitopen clauses\n and port forwards per direction (bz#1327)\n\n * sshd(8): ignore stderr output from subsystems to avoid hangs if a\n subsystem or shell initialisation writes to stderr (bz#1750)\n\n * Skip the initial check for access with an empty password when\n PermitEmptyPasswords=no (bz#1638)\n\n * sshd(8): fix logspam when key options (from="..." especially) deny\n non-matching keys (bz#1765)\n\n * ssh-keygen(1): display a more helpful error message when $HOME is\n inaccessible while trying to create .ssh directory (bz#1740)\n\n * ssh(1): fix hang when terminating a mux slave using ~. (bz#1758)\n\n * ssh-keygen(1): refuse to generate keys longer than\n OPENSSL_[RD]SA_MAX_MODULUS_BITS, since we would refuse to use\n them anyway (bz#1516)\n\n * Suppress spurious tty warning when using -O and stdin is not a tty\n (bz#1746)\n\n * Kill channel when pty allocation requests fail. Fixed stuck client\n if the server refuses pty allocation (bz#1698)\n\nPortable OpenSSH Bugfixes:\n\n * sshd(8): increase the maximum username length for login recording\n to 512 characters (bz#1579)\n\n * Initialize the values to be returned from PAM to sane values in\n case the PAM method doesn't write to them. (bz#1795) \n\n * Let configure find OpenSSL libraries in a lib64 subdirectory.\n (bz#1756)\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.6.tar.gz) = fa5ac394b874d6709031306b6ac5c48399697f7f\n - SHA1 (openssh-5.6p1.tar.gz) = 347dd39c91c3529f41dae63714d452fb95efea1e\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.5 was released on 2010-04-16. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed code\nor patches, reported bugs, tested snapshots or donated to the project.\nMore information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nThis is a bugfix release.\n\nChanges since OpenSSH 5.4\n=========================\n\n * Unbreak sshd_config's AuthorizedKeysFile option for $HOME-relative paths\n\n * Fix compilation failures on platforms that lack dlopen()\n\n * Include a language tag when sending a protocol 2 disconnection message.\n\n * Make logging of certificates used for user authentication more clear and\n consistent between CAs specified using TrustedUserCAKeys and\n authorized_keys\n\nPortable OpenSSH:\n\n * Allow contrib/ssh-copy-id to fail gracefully when there are no keys in\n the ssh-agent. bz#1723\n\n * Explicitly link libX11 into contrib/gnome-ssh-askpass2. bz#1725\n\n * Allow ChrootDirectory to work in SELinux platforms. bz#1726\n\n * Add configure.ac stanza for Haiku OS. bz#1741\n\n * Enable utmpx support on FreeBSD where possible. bz#1732\n\n * Use pkg-config to determine libedit linker flags where possible. bz#1744\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.5.tar.gz) = 59864a048b09ad1b6e65a74d5d385d8189ab8c74\n - SHA1 (openssh-5.5p1.tar.gz) = 361c6335e74809b26ea096b34062ba8ff6c97cd6\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.4 was released on 2010-03-08. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed code\nor patches, reported bugs, tested snapshots or donated to the project.\nMore information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nThis is a major feature and bugfix release.\n\nChanges since OpenSSH 5.3\n=========================\n\nFeatures:\n\n * After a transition period of about 10 years, this release disables\n SSH protocol 1 by default. Clients and servers that need to use the\n legacy protocol must explicitly enable it in ssh_config / sshd_config\n or on the command-line.\n\n * Remove the libsectok/OpenSC-based smartcard code and add support for\n PKCS#11 tokens. This support is automatically enabled on all\n platforms that support dlopen(3) and was inspired by patches written\n by Alon Bar-Lev. Details in the ssh(1) and ssh-add(1) manpages.\n\n * Add support for certificate authentication of users and hosts using a\n new, minimal OpenSSH certificate format (not X.509). Certificates\n contain a public key, identity information and some validity\n constraints and are signed with a standard SSH public key using\n ssh-keygen(1). CA keys may be marked as trusted in authorized_keys\n or via a TrustedUserCAKeys option in sshd_config(5) (for user\n authentication), or in known_hosts (for host authentication).\n\n Documentation for certificate support may be found in ssh-keygen(1),\n sshd(8) and ssh(1) and a description of the protocol extensions in\n PROTOCOL.certkeys.\n\n * Added a 'netcat mode' to ssh(1): "ssh -W host:port ..." This connects\n stdio on the client to a single port forward on the server. This\n allows, for example, using ssh as a ProxyCommand to route connections\n via intermediate servers. bz#1618\n\n * Add the ability to revoke keys in sshd(8) and ssh(1). User keys may\n be revoked using a new sshd_config(5) option "RevokedKeys". Host keys\n are revoked through known_hosts (details in the sshd(8) man page).\n Revoked keys cannot be used for user or host authentication and will\n trigger a warning if used.\n\n * Rewrite the ssh(1) multiplexing support to support non-blocking\n operation of the mux master, improve the resilience of the master to\n malformed messages sent to it by the slave and add support for\n requesting port- forwardings via the multiplex protocol. The new\n stdio-to-local forward mode ("ssh -W host:port ...") is also\n supported. The revised multiplexing protocol is documented in the\n file PROTOCOL.mux in the source distribution.\n\n * Add a 'read-only' mode to sftp-server(8) that disables open in write\n mode and all other fs-modifying protocol methods. bz#430\n\n * Allow setting an explicit umask on the sftp-server(8) commandline to\n override whatever default the user has. bz#1229\n\n * Many improvements to the sftp(1) client, many of which were\n implemented by Carlos Silva through the Google Summer of Code\n program:\n - Support the "-h" (human-readable units) flag for ls\n - Implement tab-completion of commands, local and remote filenames\n - Support most of scp(1)'s commandline arguments in sftp(1), as a\n first step towards making sftp(1) a drop-in replacement for scp(1).\n Note that the rarely-used "-P sftp_server_path" option has been\n moved to "-D sftp_server_path" to make way for "-P port" to match\n scp(1).\n - Add recursive transfer support for get/put and on the commandline\n\n * New RSA keys will be generated with a public exponent of RSA_F4 ==\n (2**16)+1 == 65537 instead of the previous value 35.\n\n * Passphrase-protected SSH protocol 2 private keys are now protected\n with AES-128 instead of 3DES. This applied to newly-generated keys\n as well as keys that are reencrypted (e.g. by changing their\n passphrase).\n\nBugfixes:\n\n * Hold authentication debug messages until after successful\n authentication. Fixes a minor information leak of environment\n variables specified in authorized_keys if an attacker happens to\n know the public key in use.\n * When using ChrootDirectory, make sure we test for the existence of\n the user's shell inside the chroot and not outside (bz#1679)\n * Cache user and group name lookups in sftp-server using\n user_from_[ug]id(3) to improve performance on hosts where these\n operations are slow (e.g. NIS or LDAP). bz#1495\n * Fix problem that prevented passphrase reading from being interrupted\n in some circumstances; bz#1590\n * Ignore and log any Protocol 1 keys where the claimed size is not\n equal to the actual size.\n * Make HostBased authentication work with a ProxyCommand. bz#1569\n * Avoid run-time failures when specifying hostkeys via a relative\n path by prepending the current working directory in these cases.\n bz#1290\n * Do not prompt for a passphrase if we fail to open a keyfile, and log\n the reason why the open failed to debug. bz#1693\n * Document that the PubkeyAuthentication directive is allowed in a\n sshd_config(5) Match block. bz#1577\n * When converting keys, truncate key comments at 72 chars as per\n RFC4716. bz#1630\n * Do not allow logins if /etc/nologin exists but is not readable by the\n user logging in.\n * Output a debug log if sshd(8) can't open an existing authorized_keys.\n bz#1694\n * Quell tc[gs]etattr warnings when forcing a tty (ssh -tt), since we\n usually don't actually have a tty to read/set; bz#1686\n * Prevent sftp from crashing when given a "-" without a command.\n Also, allow whitespace to follow a "-". bz#1691\n * After sshd receives a SIGHUP, ignore subsequent HUPs while sshd\n re-execs itself. Prevents two HUPs in quick succession from resulting\n in sshd dying. bz#1692\n * Clarify in sshd_config(5) that StrictModes does not apply to\n ChrootDirectory. Permissions and ownership are always checked when\n chrooting. bz#1532\n * Set close-on-exec on various descriptors so they don't get leaked to\n child processes. bz#1643\n * Fix very rare race condition in x11/agent channel allocation: don't\n read after the end of the select read/write fdset and make sure a\n reused FD is not touched before the pre-handlers are called.\n * Fix incorrect exit status when multiplexing and channel ID 0 is\n recycled. bz#1570\n * Fail with an error when an attempt is made to connect to a server\n with ForceCommand=internal-sftp with a shell session (i.e. not a\n subsystem session). Avoids stuck client when attempting to ssh to\n such a service. bz#1606:\n * Warn but do not fail if stat()ing the subsystem binary fails. This\n helps with chrootdirectory+forcecommand=sftp-server and restricted\n shells. bz #1599\n * Change "Connecting to host..." message to "Connected to host."\n and delay it until after the sftp protocol connection has been\n established. Avoids confusing sequence of messages when the\n underlying ssh connection experiences problems. bz#1588\n * Use the HostKeyAlias rather than the hostname specified on the\n commandline when prompting for passwords. bz#1039\n * Correct off-by-one in percent_expand(): we would fatal() when trying\n to expand EXPAND_MAX_KEYS, allowing only EXPAND_MAX_KEYS-1 to\n actually work. Note that nothing in OpenSSH actually uses close to\n this limit at present. bz#1607\n * Fix passing of empty options from scp(1) and sftp(1) to the\n underlying ssh(1). Also add support for the stop option "--".\n * Fix an incorrect magic number and typo in PROTOCOL; bz#1688\n * Don't escape backslashes when displaying the SSH2 banner. bz#1533\n * Don't unnecessarily dup() the in and out fds for sftp-server. bz#1566\n * Force use of the correct hash function for random-art signature\n display as it was inheriting the wrong one when bubblebabble\n signatures were activated. bz#1611\n * Do not fall back to adding keys without constraints (ssh-add -c /\n -t ...) when the agent refuses the constrained add request. bz#1612\n * Fix a race condition in ssh-agent that could result in a wedged or\n spinning agent. bz#1633\n * Flush stdio before exec() to ensure that everying (motd\n in particular) has made it out before the streams go away. bz#1596\n * Set FD_CLOEXEC on in/out sockets in sshd(8). bz#1706\n\nPortable OpenSSH Bugfixes:\n\n * Use system's kerberos principal name on AIX if it's available.\n bz#1583\n * Disable OOM-killing of the listening sshd on Linux. bz#1470\n * Use pkg-config for opensc config if it's available. bz#1160\n * Unbreak Redhat spec to allow building without askpass. bz#1677\n * If PidFile is set in sshd_config, use it in SMF init file. bz#1628\n * Print error and usage() when ssh-rand-helper is passed command-\n line arguments as none are supported. bz#1568\n * Add missing setsockopt() to set IPV6_V6ONLY for local forwarding\n with GatwayPorts=yes. bz#1648\n * Make GNOME 2 askpass dialog desktop-modal. bz#1645\n * If SELinux is enabled set the security context to "sftpd_t" before\n running the internal sftp server. bz#1637\n * Correctly check libselinux for necessary SELinux functions; bz#1713\n * Unbreak builds on Redhat using the supplied openssh.spec; bz#1731\n * Fix incorrect privilege dropping order on AIX that prevented\n chroot operation; bz#1567\n * Call aix_setauthdb/aix_restoredb at the correct times on AIX to\n prevent authentication failure; bz#1710\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.4.tar.gz) = 1776832d902f7b4c7863afd41a5ec7a14efe95d6\n - SHA1 (openssh-5.4p1.tar.gz) = 2a3042372f08afb1415ceaec8178213276a36302\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.3 was released on 2009-10-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nThis release marks the 10th anniversary of the OpenSSH project.\nWe would like to thank the OpenSSH community for their support,\nespecially those who will continue to contribute code or patches,\nreport bugs, test snapshots or donate to the project during the\nnext 10 years. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nThis is a bugfix release, no new features have been added.\n\nChanges since OpenSSH 5.2\n=========================\n\nGeneral Bugfixes:\n\n * Do not limit home directory paths to 256 characters. bz#1615\n\n * Several minor documentation and correctness fixes.\n\nPortable OpenSSH Bugfixes:\n\n * This release removes for support for very old versions of Cygwin and\n for Windows 95/98/ME\n\n * Move the deletion of PAM credentials on logout to after the session\n close. bz#1534\n\n * Make PrintLastLog work on AIX. bz#1595\n\n * Avoid compile errors on FreeBSD from conflicts in glob.h. bz#1634\n\n * Delay dropping of root privileges on AIX so chroot and pam_open_session\n work correctly. bz#1249 and bz#1567\n\n * Increase client IO buffer on Cygwin to 64K, realising a significant\n performance improvement.\n \n * Roll back bz#1241 (better handling for expired passwords on Tru64).\n The change broke password logins on some configurations.\n\n * Accept ENOSYS as a fallback error when attempting atomic\n rename(). bz#1535\n\n * Fix passing of variables to recursive make(1) invocations on Solaris.\n bz#1505\n\n * Skip the tcgetattr call on the pty master on Solaris, since it never\n succeeds and can hang if large amounts of data is sent to the slave\n (eg a copy-paste). bz#1528 \n\n * Fix detection of krb5-config. bz#1639\n\n * Fix test for server-assigned remote forwarding port for non-root users.\n bz#1578\n\n * Fix detection of libresolv on OSX 10.6.\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.3.tar.gz) = f1b9a280565e916c1f84fd4d944313ec926242a2\n - SHA1 (openssh-5.3p1.tar.gz) = d411fde2584ef6022187f565360b2c63a05602b5\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.2 was released on 2009-02-23. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe have also recently completed another Internet SSH usage scan, the \nresults of which may be found at http://www.openssh.com/usage.html\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nThe focus of this release has been on bugfixes as the previous\nopenssh-5.1 release introduced many new features and made some\ninvasive changes.\n\nChanges since OpenSSH 5.1\n=========================\n\nSecurity:\n\n * This release changes the default cipher order to prefer the AES CTR\n modes and the revised "arcfour256" mode to CBC mode ciphers that are\n susceptible to CPNI-957037 "Plaintext Recovery Attack Against SSH".\n\n * This release also adds countermeasures to mitigate CPNI-957037-style\n attacks against the SSH protocol's use of CBC-mode ciphers. Upon\n detection of an invalid packet length or Message Authentication\n Code, ssh/sshd will continue reading up to the maximum supported\n packet length rather than immediately terminating the connection.\n This eliminates most of the known differences in behaviour that\n leaked information about the plaintext of injected data which formed\n the basis of this attack. We believe that these attacks are rendered\n infeasible by these changes.\n\nNew features:\n\n * Added a -y option to ssh(1) to force logging to syslog rather than\n stderr, which is useful when running daemonised (ssh -f)\n\n * The sshd_config(5) ForceCommand directive now accepts commandline\n arguments for the internal-sftp server.\n\n * The ssh(1) ~C escape commandline now support runtime creation of\n dynamic (-D) port forwards.\n\n * Support the SOCKS4A protocol in ssh(1) dynamic (-D) forwards.\n (bz#1482)\n\n * Support remote port forwarding with a listen port of '0'. This\n informs the server that it should dynamically allocate a listen\n port and report it back to the client. (bz#1003)\n\n * sshd(8) now supports setting PermitEmptyPasswords and\n AllowAgentForwarding in Match blocks\n\nBug and documentation fixes\n\n * Repair a ssh(1) crash introduced in openssh-5.1 when the client is\n sent a zero-length banner (bz#1496)\n\n * Due to interoperability problems with certain\n broken SSH implementations, the eow@openssh.com and\n no-more-sessions@openssh.com protocol extensions are now only sent\n to peers that identify themselves as OpenSSH.\n\n * Make ssh(1) send the correct channel number for\n SSH2_MSG_CHANNEL_SUCCESS and SSH2_MSG_CHANNEL_FAILURE messages to\n avoid triggering 'Non-public channel' error messages on sshd(8) in\n openssh-5.1.\n\n * Avoid printing 'Non-public channel' warnings in sshd(8), since the\n ssh(1) has sent incorrect channel numbers since ~2004 (this reverts\n a behaviour introduced in openssh-5.1).\n\n * Avoid double-free in ssh(1) ~C escape -L handler (bz#1539)\n\n * Correct fail-on-error behaviour in sftp(1) batchmode for remote\n stat operations. (bz#1541)\n\n * Disable nonfunctional ssh(1) ~C escape handler in multiplex slave\n connections. (bz#1543)\n\n * Avoid hang in ssh(1) when attempting to connect to a server that\n has MaxSessions=0 set.\n\n * Multiple fixes to sshd(8) configuration test (-T) mode\n\n * Several core and portable OpenSSH bugs fixed: 1380, 1412, 1418,\n 1419, 1421, 1490, 1491, 1492, 1514, 1515, 1518, 1520, 1538, 1540\n\n * Many manual page improvements.\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.2.tar.gz) = 260074ed466e95f054ac05a4406f613d08575217\n - SHA1 (openssh-5.2p1.tar.gz) = 8273a0237db98179fbdc412207ff8eb14ff3d6de\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.1 was released on 2008-07-22. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe have also recently completed another Internet SSH usage scan, the \nresults of which may be found at http://www.openssh.com/usage.html\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots or donated to the\nproject. More information on donations may be found at:\nhttp://www.openssh.com/donations.html\n\nChanges since OpenSSH 5.0\n=========================\n\nSecurity:\n\n * sshd(8): Avoid X11 man-in-the-middle attack on HP/UX (and possibly\n other platforms) when X11UseLocalhost=no\n\n When attempting to bind(2) to a port that has previously been bound\n with SO_REUSEADDR set, most operating systems check that either the\n effective user-id matches the previous bind (common on BSD-derived\n systems) or that the bind addresses do not overlap (Linux and\n Solaris).\n\n Some operating systems, such as HP/UX, do not perform these checks\n and are vulnerable to an X11 man-in-the-middle attack when the\n sshd_config(5) option X11UseLocalhost has been set to "no" - an\n attacker may establish a more-specific bind, which will be used in\n preference to sshd's wildcard listener.\n\n Modern BSD operating systems, Linux, OS X and Solaris implement the\n above checks and are not vulnerable to this attack, nor are systems\n where the X11UseLocalhost has been left at the default value of\n "yes".\n\n Portable OpenSSH 5.1 avoids this problem for all operating systems\n by not setting SO_REUSEADDR when X11UseLocalhost is set to no.\n\n This vulnerability was reported by sway2004009 AT hotmail.com.\n\nNew features:\n\n * Introduce experimental SSH Fingerprint ASCII Visualisation to ssh(1)\n and ssh-keygen(1). Visual fingerprinnt display is controlled by a new\n ssh_config(5) option "VisualHostKey". The intent is to render\n SSH host keys in a visual form that is amenable to easy recall and\n rejection of changed host keys. This technique inspired by the\n graphical hash visualisation schemes known as "random art[*]", and\n by Dan Kaminsky's musings at 23C3 in Berlin.\n\n Fingerprint visualisation in is currently disabled by default, as the\n algorithm used to generate the random art is still subject to change.\n\n "Hash Visualization: a New Technique to improve Real-World\n Security", Perrig A. and Song D., 1999, International Workshop on\n Cryptographic Techniques and E-Commerce (CrypTEC '99)\n http://sparrow.ece.cmu.edu/~adrian/projects/validation/validation.pdf\n\n * sshd_config(5) now supports CIDR address/masklen matching in "Match\n address" blocks, with a fallback to classic wildcard matching. For \n example:\n Match address 192.0.2.0/24,3ffe:ffff::/32,!10.*\n PasswordAuthentication yes\n\n * sshd(8) now supports CIDR matching in ~/.ssh/authorized_keys\n from="..." restrictions, also with a fallback to classic wildcard\n matching.\n\n * Added an extended test mode (-T) to sshd(8) to request that it write\n its effective configuration to stdout and exit. Extended test mode\n also supports the specification of connection parameters (username,\n source address and hostname) to test the application of\n sshd_config(5) Match rules. \n\n * ssh(1) now prints the number of bytes transferred and the overall\n connection throughput for SSH protocol 2 sessions when in verbose\n mode (previously these statistics were displayed for protocol 1\n connections only).\n\n * sftp-server(8) now supports extension methods statvfs@openssh.com and\n fstatvfs@openssh.com that implement statvfs(2)-like operations.\n (bz#1399)\n\n * sftp(1) now has a "df" command to the sftp client that uses the\n statvfs@openssh.com to produce a df(1)-like display of filesystem\n space and inode utilisation (requires statvfs@openssh.com support on\n the server)\n\n * Added a MaxSessions option to sshd_config(5) to allow control of the\n number of multiplexed sessions supported over a single TCP connection.\n This allows increasing the number of allowed sessions above the\n previous default of 10, disabling connection multiplexing \n (MaxSessions=1) or disallowing login/shell/subsystem sessions\n entirely (MaxSessions=0).\n\n * Added a no-more-sessions@openssh.com global request extension that is\n sent from ssh(1) to sshd(8) when the client knows that it will never\n request another session (i.e. when session multiplexing is disabled). \n This allows a server to disallow further session requests and\n terminate the session in cases where the client has been hijacked.\n\n * ssh-keygen(1) now supports the use of the -l option in combination\n with -F to search for a host in ~/.ssh/known_hosts and display its \n fingerprint.\n\n * ssh-keyscan(1) now defaults to "rsa" (protocol 2) keys, instead of\n "rsa1".\n\n * Added an AllowAgentForwarding option to sshd_config(8) to control\n whether authentication agent forwarding is permitted. Note that this\n is a loose control, as a client may install their own unofficial\n forwarder.\n\n * ssh(1) and sshd(8): avoid unnecessary malloc/copy/free when receiving\n network data, resulting in a ~10% speedup\n\n * ssh(1) and sshd(8) will now try additional addresses when connecting\n to a port forward destination whose DNS name resolves to more than\n one address. The previous behaviour was to try the only first address\n and give up if that failed. (bz#383)\n\n * ssh(1) and sshd(8) now support signalling that channels are\n half-closed for writing, through a channel protocol extension\n notification "eow@openssh.com". This allows propagation of closed\n file descriptors, so that commands such as:\n "ssh -2 localhost od /bin/ls | true"\n do not send unnecessary data over the wire. (bz#85)\n\n * sshd(8): increased the default size of ssh protocol 1 ephemeral keys \n from 768 to 1024 bits.\n\n * When ssh(1) has been requested to fork after authentication\n ("ssh -f") with ExitOnForwardFailure enabled, delay the fork until\n after replies for any -R forwards have been seen. Allows for robust\n detection of -R forward failure when using -f. (bz#92)\n\n * "Match group" blocks in sshd_config(5) now support negation of\n groups. E.g. "Match group staff,!guests" (bz#1315)\n\n * sftp(1) and sftp-server(8) now allow chmod-like operations to set \n set[ug]id/sticky bits. (bz#1310)\n\n * The MaxAuthTries option is now permitted in sshd_config(5) match\n blocks.\n\n * Multiplexed ssh(1) sessions now support a subset of the ~ escapes\n that are available to a primary connection. (bz#1331)\n\n * ssh(1) connection multiplexing will now fall back to creating a new\n connection in most error cases. (bz#1439 bz#1329)\n\n * Added some basic interoperability tests against Twisted Conch.\n\n * Documented OpenSSH's extensions to and deviations from the published\n SSH protocols (the PROTOCOL file in the distribution)\n\n * Documented OpenSSH's ssh-agent protocol (PROTOCOL.agent).\n\nBug and documentation fixes\n\n * Make ssh(1) deal more gracefully with channel requests that fail.\n Previously it would optimistically assume that requests would always \n succeed, which could cause hangs if they did not (e.g. when the\n server runs out of file descriptors). (bz#1384)\n\n * ssh(1) now reports multiplexing errors via the multiplex slave's\n stderr where possible (subject to LogLevel in the mux master).\n\n * ssh(1) and sshd(8) now send terminate protocol banners with CR+LF for\n protocol 2 to comply with RFC 4253. Previously they were terminated\n with CR alone. Protocol 1 banners remain CR terminated. (bz#1443)\n\n * Merged duplicate authentication file checks in sshd(8) and refuse to\n read authorised_keys and .shosts from non-regular files. (bz#1438)\n\n * Ensure that sshd(8)'s umask disallows at least group and world write,\n even if a more permissive one has been inherited. (bz#1433)\n\n * Suppress the warning message from sshd(8) when changing to a\n non-existent user home directory after chrooting. (bz#1461)\n\n * Mention that scp(1) follows symlinks when performing recursive\n copies. (bz#1466)\n\n * Prevent sshd(8) from erroneously applying public key restrictions\n leaned from ~/.ssh/authorized_keys to other authentication methods\n when public key authentication subsequently fails. (bz#1472)\n\n * Fix protocol keepalive timeouts - in some cases, keepalive packets\n were being sent, but the connection was not being closed when the\n limit for missing replies was exceeded. (bz#1465)\n\n * Fix ssh(1) sending invalid TTY modes when a TTY was forced (ssh -tt)\n but stdin was not a TTY. (bz#1199)\n\n * ssh(1) will now exit with a non-zero exit status if\n ExitOnForwardFailure was set and forwardings were disabled due to a\n failed host key check.\n\n * Fix MaxAuthTries tests to disallow a free authentication try to\n clients that skipped the protocol 2 "none" authentication method.\n (part of bz#1432)\n\n * Make keepalive timeouts apply while synchronously waiting\n for a packet, particularly during key renegotiation. (bz#1363)\n\n * sshd(8) has been audited to eliminate fd leaks and calls to fatal()\n in conditions of file descriptor exhaustion.\n\nPortable OpenSSH-specific bugfixes\n\n * Avoid a sshd(8) hang-on-exit on Solaris caused by depending on the \n success of isatty() on a PTY master (undefined behaviour). Probably \n affected other platforms too. (bz#1463)\n\n * Fixed test for locked accounts on HP/UX with shadowed\n passwords disabled. (bz#1083)\n\n * Disable poll() fallback in atomiciov for Tru64. readv\n doesn't seem to be a comparable object there, which lead to\n compilation errors. (bz#1386)\n\n * Fall back to racy rename if link returns EXDEV. (bz#1447)\n\n * Explicitly handle EWOULDBLOCK wherever we handle EAGAIN, on\n some platforms (HP nonstop) it is a distinct errno. (bz#1467)\n\n * Avoid NULL dereferences in ancient sigaction replacement\n code. (bz#1240)\n\n * Avoid linking against libgssapi, which despite its name\n doesn't seem to implement all of GSSAPI. (bz#1276)\n\n * Use explicit noreturn attribute instead of __dead, fixing\n compilation problems on Interix. (bz#1112)\n\n * Added support password expiry on Tru64 SIA systems. (bz#1241)\n\n * Fixed an UMAC alignment problem that manifested on Itanium \n platforms. (bz#1462)\n\n * The sftp-server(8) manual now describes the requirements for\n transfer logging in chroot environments. (bz#1488)\n\n * Strip trailing dot from hostnames when the sshd_config(5)\n HostbasedUsesNameFromPacketOnly option is set. (bz#1200)\n \nChecksums:\n==========\n\n - SHA1 (openssh-5.1.tar.gz) = 1e5b43844ed015e4fbbbe25cfad6f5377c60e759\n - SHA1 (openssh-5.1p1.tar.gz) = 877ea5b283060fe0160e376ea645e8e168047ff5\n\nReporting Bugs:\n===============\n\n- Please read http://www.openssh.com/report.html\n Security bugs should be reported directly to openssh@openssh.org\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 5.0 was released on 2008-04-03. It is available from the\nmirrors listed at https://www.openssh.com/.\nWe apologise for any inconvenience resulting from this release\nbeing made so shortly after 4.9. Unfortunately we only learned of\nthe below security issue from the public CVE report. The Debian\nOpenSSH maintainers responsible for handling the initial report of\nthis bug failed to report it via either the private OpenSSH security\ncontact list (openssh@openssh.org) or the portable OpenSSH Bugzilla\n(http://bugzilla.mindrot.org/).\nWe ask anyone wishing to report security bugs in OpenSSH to please use\nthe openssh@openssh.org contact and to practice responsible disclosure.\n\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n\thttp://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nChanges since OpenSSH 4.9:\n============================\n\nSecurity:\n\n * CVE-2008-1483: Avoid possible hijacking of X11-forwarded connections\n by refusing to listen on a port unless all address families bind\n successfully.\n\nChecksums:\n==========\n\n - SHA1 (openssh-5.0.tar.gz) = 729fb3168edf6a68408223b5ed82e59d13b57c47\n - SHA1 (openssh-5.0p1.tar.gz) = 121cea3a730c0b0353334b6f46f438de30ab4928\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.9 was released on 2008-03-31. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n\thttp://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nNote that OpenSSH 4.8 was an OpenBSD-only release shipped with the\nOpenBSD 4.3 CD.\n\nChanges since OpenSSH 4.7:\n============================\n\nSecurity:\n\n * Disable execution of ~/.ssh/rc for sessions where a command has been\n forced by the sshd_config ForceCommand directive. Users who had\n write access to this file could use it to execute arbitrary commands.\n This behaviour was documented, but was an unsafe default and an extra\n hassle for administrators.\n\nNew features:\n\n * Added chroot(2) support for sshd(8), controlled by a new option\n "ChrootDirectory". Please refer to sshd_config(5) for details, and\n please use this feature carefully. (bz#177 bz#1352)\n * Linked sftp-server(8) into sshd(8). The internal sftp server is\n used when the command "internal-sftp" is specified in a Subsystem\n or ForceCommand declaration. When used with ChrootDirectory, the\n internal sftp server requires no special configuration of files\n inside the chroot environment. Please refer to sshd_config(5) for\n more information.\n * Added a "no-user-rc" option for authorized_keys to disable execution\n of ~/.ssh/rc\n * Added a protocol extension method "posix-rename@openssh.com" for\n sftp-server(8) to perform POSIX atomic rename() operations.\n (bz#1400)\n * Removed the fixed limit of 100 file handles in sftp-server(8). The\n server will now dynamically allocate handles up to the number of\n available file descriptors. (bz#1397)\n * ssh(8) will now skip generation of SSH protocol 1 ephemeral server\n keys when in inetd mode and protocol 2 connections are negotiated.\n This speeds up protocol 2 connections to inetd-mode servers that\n also allow Protocol 1 (bz#440)\n * Accept the PermitRootLogin directive in a sshd_config(5) Match\n block. Allows for, e.g. permitting root only from the local\n network.\n * Reworked sftp(1) argument splitting and escaping to be more\n internally consistent (i.e. between sftp commands) and more\n consistent with sh(1). Please note that this will change the\n interpretation of some quoted strings, especially those with\n embedded backslash escape sequences. (bz#778)\n * Support "Banner=none" in sshd_config(5) to disable sending of a\n pre-login banner (e.g. in a Match block).\n * ssh(1) ProxyCommands are now executed with $SHELL rather than\n /bin/sh.\n * ssh(1)'s ConnectTimeout option is now applied to both the TCP\n connection and the SSH banner exchange (previously it just covered\n the TCP connection). This allows callers of ssh(1) to better detect\n and deal with stuck servers that accept a TCP connection but don't\n progress the protocol, and also makes ConnectTimeout useful for\n connections via a ProxyCommand.\n * Many new regression tests, including interop tests against PuTTY's\n plink.\n * Support BSM auditing on Mac OS X\n\nThe following bugs have been fixed in this release:\n\n - scp(1) incorrectly reported "stalled" on slow copies. (bz#799)\n - scp(1) date underflow for timestamps before epoch. (bz#828)\n - scp(1) and sftp(1) progressmeter type fixes. (bz#842)\n - SSH2_MSG_UNIMPLEMENTED packets did not correctly reset the client\n keepalive logic, causing disconnections on servers that did not\n explicitly implement "keepalive@openssh.com". (bz#1307)\n - ssh(1) used the obsolete SIG DNS RRtype for host keys in DNS,\n instead of the current standard RRSIG. (bz#1317)\n - Extract magic buffer size constants in scp(1) to #defines.\n (bz#1333)\n - Correctly drain ACKs when a sftp(1) upload write fails midway,\n avoids a fatal() exit from what should be a recoverable condition.\n (bz#1354)\n - Avoid pointer arithmetic and strict aliasing warnings. (bz#1355)\n - Fixed packet size advertisements. Previously TCP and agent\n forwarding incorrectly advertised the channel window size as the\n packet size, causing fatal errors under some conditions. (bz#1360)\n - Document KbdInteractiveAuthentication in sshd_config(5). (bz#1367)\n - Fixed sftp(1) file handle leak on download when the local file\n could not be opened. (bz#1375)\n - Fixed ssh-keygen(1) selective host key hashing (i.e.\n "ssh-keygen -HF hostname") to not include any IP address in the\n data to be hashed. (bz#1376)\n - Fix clobbering of struct passwd from multiple getpwid calls,\n resulting in fatal errors inside tilde_expand_filename. (bz#1377)\n - Fix parsing of port-forwarding specifications to correctly\n detect errors in either specified port number. (bz#1378)\n - Fix memory leak in ssh(1) ~ escape commandline handling. (bz#1379)\n - Make ssh(1) skip listening on the IPv6 wildcard address when a\n binding address of 0.0.0.0 is used against an old SSH server that\n does not support the RFC4254 syntax for wildcard bind addresses.\n (bz#1381)\n - Remove extra backslashes in the RB_PROTOTYPE macro definition.\n (bz#1385)\n - Support ssh(1) RekeyLimits up to the maximum allowed by the\n protocol: 2**32-1. (bz#1390)\n - Enable IPV6_V6ONLY socket option on sshd(8) listen socket, as is\n already done for X11/TCP forwarding sockets. (bz#1392)\n - Fix FD leak that could hang a ssh(1) connection multiplexing\n master. (bz#1398)\n - Improve error messages when hostname resolution fails due to a\n system error. (bz#1417)\n - Make ssh(1) -q option documentation consistent with reality.\n (bz#1427 bz#1429)\n\nPortable OpenSSH bugs fixed:\n\n - Fixed sshd(8) PAM support not calling pam_session_close(), or\n failing to call it with root privileges. (bz#926)\n - Made sshd(8) use configuration-specified SyslogFacility for\n hosts_access(3) messages for denied connections. (bz#1042)\n - Implement getgrouplist(3) for AIX, enabling NSS LDAP to work on\n this platform. (bz#1081)\n - Fix compilation errors on AIX due to misdefinition of LLONG_MAX.\n (bz#1347)\n - Force use of local glob(3) implementation on Mac OS X and FreeBSD,\n as the platform versions lack features or have unexpected\n behaviour. (bz#1407)\n - Reduce stdout/stderr noise from ssh-copy-id. (bz#1431)\n - Fix activation of OpenSSL engine support when requested in\n configure. (bz#1437)\n\nChecksums:\n==========\n\n - SHA1 (openssh-4.9.tar.gz) = fa7d1b3dcb093bd0dfc643b33b1a57a26f459373\n - SHA1 (openssh-4.9p1.tar.gz) = 91575878883065bd777f82b47e0d481ac69ee7fe\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.8 was an OpenBSD-only release, included on the OpenBSD 4.3\nCD only.\n\nChanges since OpenSSH 4.8:\n============================\n\nNew features:\n\n * Added chroot(2) support for sshd(8), controlled by a new option\n "ChrootDirectory". Please refer to sshd_config(5) for details, and\n please use this feature carefully. (bz#177 bz#1352)\n * Linked sftp-server(8) into sshd(8). The internal sftp server is\n used when the command "internal-sftp" is specified in a Subsystem\n or ForceCommand declaration. When used with ChrootDirectory, the\n internal sftp server requires no special configuration of files\n inside the chroot environment. Please refer to sshd_config(5) for\n more information.\n * Added a protocol extension method "posix-rename@openssh.com" for\n sftp-server(8) to perform POSIX atomic rename() operations.\n (bz#1400)\n * Removed the fixed limit of 100 file handles in sftp-server(8). The\n server will now dynamically allocate handles up to the number of\n available file descriptors. (bz#1397)\n * ssh(8) will now skip generation of SSH protocol 1 ephemeral server\n keys when in inetd mode and protocol 2 connections are negotiated.\n This speeds up protocol 2 connections to inetd-mode servers that\n also allow Protocol 1 (bz#440)\n * Accept the PermitRootLogin directive in a sshd_config(5) Match\n block. Allows for, e.g. permitting root only from the local\n network.\n * Reworked sftp(1) argument splitting and escaping to be more\n internally consistent (i.e. between sftp commands) and more\n consistent with sh(1). Please note that this will change the\n interpretation of some quoted strings, especially those with\n embedded backslash escape sequences. (bz#778)\n * Support "Banner=none" in sshd_config(5) to disable sending of a\n pre-login banner (e.g. in a Match block).\n * ssh(1) ProxyCommands are now executed with $SHELL rather than\n /bin/sh.\n * ssh(1)'s ConnectTimeout option is now applied to both the TCP\n connection and the SSH banner exchange (previously it just covered\n the TCP connection). This allows callers of ssh(1) to better detect\n and deal with stuck servers that accept a TCP connection but don't\n progress the protocol, and also makes ConnectTimeout useful for\n connections via a ProxyCommand.\n * Many new regression tests, including interop tests against PuTTY's\n plink.\n * Support BSM auditing on Mac OS X\n\nThe following bugs have been fixed in this release:\n\n - scp(1) incorrectly reported "stalled" on slow copies. (bz#799)\n - scp(1) date underflow for timestamps before epoch. (bz#828)\n - scp(1) and sftp(1) progressmeter type fixes. (bz#842)\n - SSH2_MSG_UNIMPLEMENTED packets did not correctly reset the client\n keepalive logic, causing disconnections on servers that did not\n explicitly implement "keepalive@openssh.com". (bz#1307)\n - ssh(1) used the obsolete SIG DNS RRtype for host keys in DNS,\n instead of the current standard RRSIG. (bz#1317)\n - Extract magic buffer size constants in scp(1) to #defines.\n (bz#1333)\n - Correctly drain ACKs when a sftp(1) upload write fails midway,\n avoids a fatal() exit from what should be a recoverable condition.\n (bz#1354)\n - Avoid pointer arithmetic and strict aliasing warnings. (bz#1355)\n - Fixed packet size advertisements. Previously TCP and agent\n forwarding incorrectly advertised the channel window size as the\n packet size, causing fatal errors under some conditions. (bz#1360)\n - Document KbdInteractiveAuthentication in sshd_config(5). (bz#1367)\n - Fixed sftp(1) file handle leak on download when the local file\n could not be opened. (bz#1375)\n - Fixed ssh-keygen(1) selective host key hashing (i.e.\n "ssh-keygen -HF hostname") to not include any IP address in the\n data to be hashed. (bz#1376)\n - Fix clobbering of struct passwd from multiple getpwid calls,\n resulting in fatal errors inside tilde_expand_filename. (bz#1377)\n - Fix parsing of port-forwarding specifications to correctly\n detect errors in either specified port number. (bz#1378)\n - Fix memory leak in ssh(1) ~ escape commandline handling. (bz#1379)\n - Make ssh(1) skip listening on the IPv6 wildcard address when a\n binding address of 0.0.0.0 is used against an old SSH server that\n does not support the RFC4254 syntax for wildcard bind addresses.\n (bz#1381)\n - Remove extra backslashes in the RB_PROTOTYPE macro definition.\n (bz#1385)\n - Support ssh(1) RekeyLimits up to the maximum allowed by the\n protocol: 2**32-1. (bz#1390)\n - Enable IPV6_V6ONLY socket option on sshd(8) listen socket, as is\n already done for X11/TCP forwarding sockets. (bz#1392)\n - Fix FD leak that could hang a ssh(1) connection multiplexing\n master. (bz#1398)\n - Improve error messages when hostname resolution fails due to a\n system error. (bz#1417)\n - Make ssh(1) -q option documentation consistent with reality.\n (bz#1427 bz#1429)\n\nPortable OpenSSH bugs fixed:\n\n - Fixed sshd(8) PAM support not calling pam_session_close(), or\n failing to call it with root privileges. (bz#926)\n - Made sshd(8) use configuration-specified SyslogFacility for\n hosts_access(3) messages for denied connections. (bz#1042)\n - Implement getgrouplist(3) for AIX, enabling NSS LDAP to work on\n this platform. (bz#1081)\n - Fix compilation errors on AIX due to misdefinition of LLONG_MAX.\n (bz#1347)\n - Force use of local glob(3) implementation on Mac OS X and FreeBSD,\n as the platform versions lack features or have unexpected\n behaviour. (bz#1407)\n - Reduce stdout/stderr noise from ssh-copy-id. (bz#1431)\n - Fix activation of OpenSSL engine support when requested in\n configure. (bz#1437)\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.7 was released on 2007-09-05. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n http://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nChanges since OpenSSH 4.6:\n============================\n\nSecurity bugs resolved in this release:\n\n * Prevent ssh(1) from using a trusted X11 cookie if creation of an\n untrusted cookie fails; found and fixed by Jan Pechanec.\n\nOther changes, new functionality and fixes in this release:\n\n * sshd(8) in new installations defaults to SSH Protocol 2 only.\n Existing installations are unchanged.\n\n * The SSH channel window size has been increased, and both ssh(1)\n sshd(8) now send window updates more aggressively. These improves\n performance on high-BDP (Bandwidth Delay Product) networks.\n\n * ssh(1) and sshd(8) now preserve MAC contexts between packets, which\n saves 2 hash calls per packet and results in 12-16% speedup for\n arcfour256/hmac-md5.\n\n * A new MAC algorithm has been added, UMAC-64 (RFC4418) as\n "umac-64@openssh.com". UMAC-64 has been measured to be \n approximately 20% faster than HMAC-MD5.\n\n * A -K flag was added to ssh(1) to set GSSAPIAuthentication=Yes\n\n * Failure to establish a ssh(1) TunnelForward is now treated as a\n fatal error when the ExitOnForwardFailure option is set.\n\n * ssh(1) returns a sensible exit status if the control master goes\n away without passing the full exit status. (bz #1261)\n\n * The following bugs have been fixed in this release:\n\n - When using a ProxyCommand in ssh(1), set the outgoing hostname with\n gethostname(2), allowing hostbased authentication to work (bz #616)\n - Make scp(1) skip FIFOs rather than hanging (bz #856)\n - Encode non-printing characters in scp(1) filenames.\n these could cause copies to be aborted with a "protocol error"\n (bz #891)\n - Handle SIGINT in sshd(8) privilege separation child process to\n ensure that wtmp and lastlog records are correctly updated\n (bz #1196)\n - Report GSSAPI mechanism in errors, for libraries that support\n multiple mechanisms (bz #1220)\n - Improve documentation for ssh-add(1)'s -d option (bz #1224)\n - Rearrange and tidy GSSAPI code, removing server-only code being\n linked into the client. (bz #1225)\n - Delay execution of ssh(1)'s LocalCommand until after all forwadings\n have been established. (bz #1232)\n - In scp(1), do not truncate non-regular files (bz #1236)\n - Improve exit message from ControlMaster clients. (bz #1262)\n - Prevent sftp-server(8) from reading until it runs out of buffer\n space, whereupon it would exit with a fatal error. (bz #1286)\n\n * Portable OpenSSH bugs fixed:\n\n - Fix multiple inclusion of paths.h on AIX 5.1 systems. (bz #1243)\n - Implement getpeereid for Solaris using getpeerucred. Solaris\n systems will now refuse ssh-agent(1) and ssh(1) ControlMaster\n clients from different, non-root users (bz #1287)\n - Fix compilation warnings by including string.h if found. (bz #1294)\n - Remove redefinition of _res in getrrsetbyname.c for platforms that\n already define it. (bz #1299)\n - Fix spurious "chan_read_failed for istate 3" errors from sshd(8),\n a side-effect of the "hang on exit" fix introduced in 4.6p1.\n (bz #1306)\n - pam_end() was not being called if authentication failed (bz #1322)\n - Fix SELinux support when SELinux is in permissive mode. Previously\n sshd(8) was treating SELinux errors as always fatal. (bz #1325)\n - Ensure that pam_setcred(..., PAM_ESTABLISH_CRED) is called before\n pam_setcred(..., PAM_REINITIALIZE_CRED), fixing pam_dhkeys.\n (bz #1339)\n - Fix privilege separation on QNX - pre-auth only, this platform does\n not support file descriptor passing needed for post-auth privilege\n separation. (bz #1343)\n\nThanks to everyone who has contributed patches, reported bugs and\ntested releases.\n\nChecksums:\n==========\n\n- SHA1 (openssh-4.7.tar.gz) = 9ebaab9b31e01bd0d04425dc23536bcc78f8d990\n- SHA1 (openssh-4.7p1.tar.gz) = 58357db9e64ba6382bef3d73d1d386fcdc0508f4\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.6 was released on 2007-03-09. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\n:\n::\n:\n============================\n\n * sshd now allows the enabling and disabling of authentication\n methods on a per user, group, host and network basis via the\n Match directive in sshd_config.\n\n:\n\n - Clear SIGALRM when restarting due to SIGHUP. Prevents stray\n signal from taking down sshd if a connection was pending at\n the time SIGHUP was received\n - sftp returned a zero exit status when upload failed due to write\n errors (bugzilla #1252)\n - fixed an inconsistent check for a terminal when displaying scp\n progress meter (bugzilla #1265)\n - Parsing of time values in Match blocks was incorrectly applied\n to the global configuration (bugzilla #1275)\n - Allow multiple forwarding options to work when specified in a\n PermitOpen directive (bugzilla #1267)\n - Interoperate with ssh.com versions that do not support binding\n remote port forwarding sessions to a hostname (bugzilla #1019)\n\n:\n\n - "hang on exit" when background processes are running at the time\n of exit on a ttyful/login session (bugzilla #52)\n - Fix typos in the ssh-rand-helper(8) man page (bugzilla #1259)\n - Check that some SIG records have been returned in getrrsetbyname\n (bugzilla #1281)\n - Fix contrib/findssl for platforms that lack "which" (bugzilla \n #1237)\n - Work around bug in OpenSSL 0.9.8e that broke aes256-ctr,\n aes192-ctr, arcfour256 (bugzilla #1291)\n\n:\n==========\n\n- SHA1 (openssh-4.6.tar.gz) = c1700845be464a769428f34ef727c1f530728afc\n- SHA1 (openssh-4.6p1.tar.gz) = b2aefeb1861b4688b1777436035239ec32a47da8\n\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.5 was released on 2006-11-07. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\n:\n::\n:\n============================\n\nThis is a bugfix only release. No new features have been added.\n\n:\n\n * Fix a bug in the sshd privilege separation monitor that weakened its\n verification of successful authentication. This bug is not known to\n be exploitable in the absence of additional vulnerabilities.\n \n:\n\n * Several compilation fixes for portable OpenSSH\n\n * Fixes to Solaris SMF/process contract support (bugzilla #1255)\n\nThanks to everyone who has contributed patches, reported bugs and\ntested releases.\n\n:\n==========\n\n- SHA1 (openssh-4.5.tar.gz) = def3de1557181062d788695b9371d02635af39fb\n- SHA1 (openssh-4.5p1.tar.gz) = 2eefcbbeb9e4fa16fa4500dec107d1a09d3d02d7\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.4 was released on 2006-09-27. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased\nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures\n:\n::\n:\n============================\n\n:\n\n * Fix a pre-authentication denial of service found by Tavis Ormandy,\n that would cause sshd(8) to spin until the login grace time\n expired.\n\n * Fix an unsafe signal handler reported by Mark Dowd. The signal\n handler was vulnerable to a race condition that could be exploited\n to perform a pre-authentication denial of service. On portable\n OpenSSH, this vulnerability could theoretically lead to\n pre-authentication remote code execution if GSSAPI authentication\n is enabled, but the likelihood of successful exploitation appears\n remote.\n\n * On portable OpenSSH, fix a GSSAPI authentication abort that could\n be used to determine the validity of usernames on some platforms.\n\n:\n\n * Implemented conditional configuration in sshd_config(5) using the\n "Match" directive. This allows some configuration options to be\n selectively overridden if specific criteria (based on user, group,\n hostname and/or address) are met. So far a useful subset of post-\n authentication options are supported and more are expected to be\n added in future releases.\n\n * Add support for Diffie-Hellman group exchange key agreement with a\n final hash of SHA256.\n\n * Added a "ForceCommand" directive to sshd_config(5). Similar to the\n command="..." option accepted in ~/.ssh/authorized_keys, this forces\n the execution of the specified command regardless of what the user\n requested. This is very useful in conjunction with the new "Match"\n option.\n\n * Add a "PermitOpen" directive to sshd_config(5). This mirrors the\n permitopen="..." authorized_keys option, allowing fine-grained\n control over the port-forwardings that a user is allowed to\n establish.\n\n * Add optional logging of transactions to sftp-server(8).\n\n * ssh(1) will now record port numbers for hosts stored in\n ~/.ssh/authorized_keys when a non-standard port has been requested.\n\n * Add an "ExitOnForwardFailure" option to cause ssh(1) to exit (with\n a non-zero exit code) when requested port forwardings could not be\n established.\n\n * Extend sshd_config(5) "SubSystem" declarations to allow the\n specification of command-line arguments.\n\n * Replacement of all integer overflow susceptible invocations of\n malloc(3) and realloc(3) with overflow-checking equivalents.\n\n * Many manpage fixes and improvements\n\n:\n\n - Add optional support for SELinux, controlled using the\n --with-selinux configure option (experimental)\n\n - Add optional support for Solaris process contracts, enabled\n using the --with-solaris-contracts configure option (experimental)\n This option will also include SMF metadata in Solaris packages\n built using the "make package" target\n\n - Add optional support for OpenSSL hardware accelerators (engines),\n enabled using the --with-ssl-engine configure option.\n\n:Thanks to everyone who has contributed patches, reported bugs and\ntested releases.\n\n:\n==========\n\n- SHA1 (openssh-4.4.tar.gz) = 2294b5e5a591420aa05ff607c1890ab622ace878\n- SHA1 (openssh-4.4p1.tar.gz) = 6a52b1dee1c2c9862923c0008d201d98a7fd9d6c\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nPortable OpenSSH 4.3p2 was released on 2006-02-11. It will be available\nfrom the mirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe have also recently completed another Internet SSH usage scan, the \nresults of which may be found at http://www.openssh.com/usage.html\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased \nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures \nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n\thttp://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nChanges since Portable OpenSSH 4.3p1:\n====================================\n\nThis is a release of Portable OpenSSH only, to resolve some \nportability bugs. There are no new features, only fixes: \n\n * Explicitly test for egrep in ./configure, fixing a problem in 4.3p1\n that caused some platforms to fail to detect the available fields\n in utmp/wtmp/lastlog records. This bug manifested as missing or\n empty login/logout records (as seen by last(1), etc.)\n\n * Fix for logout records not being updated on platforms without \n support for post-authentication privilege separation (e.g. Cygwin)\n\n * Fixed compilation problems on Ultrix, NewsOS and QNX\n\nThanks to everyone who has contributed patches, reported bugs or test\nreleases.\n\nChecksums:\n==========\n\n- SHA1 (openssh-4.3p2.tar.gz) = 2b5b0751fd578283ba7b106025c0ba391fd72f1f\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.3 was released on 2006-02-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe have also recently completed another Internet SSH usage scan, the \nresults of which may be found at http://www.openssh.com/usage.html\n\nOnce again, we would like to thank the OpenSSH community for their\ncontinued support of the project, especially those who contributed\ncode or patches, reported bugs, tested snapshots and purchased \nT-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures \nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n\thttp://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nChanges since OpenSSH 4.2:\n============================ \n\nSecurity bugs resolved in this release:\n\n * CVE-2006-0225: scp (as does rcp, on which it is based) invoked a\n subshell to perform local to local, and remote to remote copy\n operations. This subshell exposed filenames to shell expansion\n twice; allowing a local attacker to create filenames containing\n shell metacharacters that, if matched by a wildcard, could lead\n to execution of attacker-specified commands with the privilege of\n the user running scp (Bugzilla #1094)\n\nThis is primarily a bug-fix release, only one new feature has been\nadded: \n\n * Add support for tunneling arbitrary network packets over a\n connection between an OpenSSH client and server via tun(4) virtual\n network interfaces. This allows the use of OpenSSH (4.3+) to create\n a true VPN between the client and server providing real network\n connectivity at layer 2 or 3. This feature is experimental and is \n currently supported on OpenBSD, Linux, NetBSD (IPv4 only) and \n FreeBSD. Other operating systems with tun/tap interface capability \n may be added in future portable OpenSSH releases. Please refer to \n the README.tun file in the source distribution for further details\n and usage examples.\n\nSome of the other bugs resolved and internal improvements are:\n\n * Reduce default key length for new DSA keys generated by ssh-keygen \n back to 1024 bits. DSA is not specified for longer lengths and does \n not fully benefit from simply making keys longer. As per FIPS 186-2 \n Change Notice 1, ssh-keygen will refuse to generate a new DSA key \n smaller or larger than 1024 bits\n\n * Fixed X forwarding failing to start when a the X11 client is executed\n in background at the time of session exit (Bugzilla #1086)\n\n * Change ssh-keygen to generate a protocol 2 RSA key when invoked\n without arguments (Bugzilla #1064)\n\n * Fix timing variance for valid vs. invalid accounts when attempting\n Kerberos authentication (Bugzilla #975)\n\n * Ensure that ssh always returns code 255 on internal error (Bugzilla \n #1137)\n\n * Cleanup wtmp files on SIGTERM when not using privsep (Bugzilla #1029)\n\n * Set SO_REUSEADDR on X11 listeners to avoid problems caused by\n lingering sockets from previous session (X11 applications can\n sometimes not connect to 127.0.0.1:60xx) (Bugzilla #1076)\n\n * Ensure that fds 0, 1 and 2 are always attached in all programs, by\n duping /dev/null to them if necessary.\n\n * Xauth list invocation had bogus "." argument (Bugzilla #1082)\n\n * Remove internal assumptions on key exchange hash algorithm and output\n length, preparing OpenSSH for KEX methods with alternate hashes.\n\n * Ignore junk sent by a server before it sends the "SSH-" banner\n (Bugzilla #1067)\n\n * The manpages has been significantly improves and rearranged, in \n addition to other specific manpage fixes:\n #1037 - Man page entries for -L and -R should mention -g.\n #1077 - Descriptions for "ssh -D" and DynamicForward should mention\n they can specify "bind_address" optionally.\n #1088 - Incorrect descriptions in ssh_config man page for \n ControlMaster=no.\n #1121 - Several corrections for ssh_agent manpages\n\n * Lots of cleanups, including fixes to memory leaks on error paths\n (Bugzilla #1109, #1110, #1111 and more) and possible crashes (#1092)\n\n * Portable OpenSSH-specific fixes:\n\n - Pass random seed during re-exec for each connection: speeds up\n processing of new connections on platforms using the OpenSSH's\n builtin entropy collector (ssh-rand-helper)\n\n - PAM fixes and improvements:\n #1045 - Missing option for ignoring the /etc/nologin file\n #1087 - Show PAM password expiry message from LDAP on login\n #1028 - Forward final non-query conversations to client\n #1126 - Prevent user from being forced to change an expired\n password repeatedly on AIX in some PAM configurations.\n #1045 - Do not check /etc/nologin when PAM is enabled, instead \n allow PAM to handle it. Note that on platforms using \n PAM, the pam_nologin module should be used in sshd's \n session stack in order to maintain past behaviour\n\n - Portability-related fixes:\n #989 - Fix multiplexing regress test on Solaris\n #1097 - Cross-compile fixes.\n #1096 - ssh-keygen broken on HPUX.\n #1098 - $MAIL being set incorrectly for HPUX server login.\n #1104 - Compile error on Tru64 Unix 4.0f\n #1106 - Updated .spec file and startup for SuSE.\n #1122 - Use _GNU_SOURCE define in favor of __USE_GNU, fixing\n compilation problems on glibc 2.4\n\nThanks to everyone who has contributed patches, reported bugs or test\nreleases.\n\nChecksums:\n==========\n\n- SHA1 (openssh-4.3.tar.gz) = 0cb66e56805d66b51511455423bab88aa58a1455\n- SHA1 (openssh-4.3p1.tar.gz) = b1f379127829e7e820955b2825130edd1601ba59\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Darren Tucker, Jason McIntyre, Tim Rice and\nBen Lindstrom.\nOpenSSH 4.2 was released on 2005-09-01. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport of the project, especially those who contributed source,\nreported bugs, tested snapshots and purchased T-shirts or posters.\n\nT-shirt, poster and CD sales directly support the project. Pictures \nand more information can be found at:\n http://www.openbsd.org/tshirts.html and\n\thttp://www.openbsd.org/orders.html\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nChanges since OpenSSH 4.1:\n============================ \n\n - SECURITY: Fix a bug introduced in OpenSSH 4.0 that caused \n GatewayPorts to be incorrectly activated for dynamic ("-D") port \n forwardings when no listen address was explicitly specified.\n\n - SECURITY: sshd in OpenSSH versions prior to 4.2 allow GSSAPI \n credentials to be delegated to users who log in with methods \n other than GSSAPI authentication (e.g. public key) when the \n client requests it. This behaviour has been changed in OpenSSH \n 4.2 to only delegate credentials to users who authenticate\n using the GSSAPI method. This eliminates the risk of credentials \n being inadvertently exposed to an untrusted user/host (though \n users should not activate GSSAPIDelegateCredentials to begin\n with when the remote user or host is untrusted)\n\n - Added a new compression method that delays the start of zlib\n compression until the user has been authenticated successfully. \n The new method ("Compression delayed") is on by default in the \n server. This eliminates the risk of any zlib vulnerability \n leading to a compromise of the server from unauthenticated users.\n\n NB. Older OpenSSH (<3.5) versions have a bug that will cause them\n to refuse to connect to any server that does not offer compression\n when the client has compression requested. Since the new "delayed"\n server mode isn't supported by these older clients, they will\n refuse to connect to a new server unless compression is disabled\n (on the client end) or the original compression method is enabled\n on the server ("Compression yes" in sshd_config)\n\n - Another round of proactive changes for signed vs unsigned integer\n bugs has been completed, including changing the atomicio() API to\n encourage safer programming. This work is ongoing.\n\n - Added support for the improved arcfour cipher modes from\n draft-harris-ssh-arcfour-fixes-02. The improves the cipher's\n resistance to a number of attacks by discarding early keystream\n output.\n\n - Increase the default size of new RSA/DSA keys generated by\n ssh-keygen from 1024 to 2048 bits.\n\n - Many bugfixes and improvements to connection multiplexing,\n including:\n\n - Added ControlMaster=auto/autoask options to support opportunistic\n multiplexing (see the ssh_config(5) manpage for details).\n\n - The client will now gracefully fallback to starting a new TCP\n connection if it cannot connect to a specified multiplexing\n control socket\n\n - Added %h (target hostname), %p (target port) and %r (remote\n username) expansion sequences to ControlPath. Also allow\n ControlPath=none to disable connection multiplexing.\n\n - Implemented support for X11 and agent forwarding over multiplexed\n connections. Because of protocol limitations, the slave\n connections inherit the master's DISPLAY and SSH_AUTH_SOCK rather\n than distinctly forwarding their own.\n\n - Portable OpenSSH: Added support for long passwords (> 8-char) on\n UnixWare 7.\n\n - The following bugs from http://bugzilla.mindrot.org/ were closed:\n\n #471 - Misleading error message if /dev/tty perms wrong\n #623 - Don't use $HOME in manpages\n #829 - Don't allocate a tty if -n option is set\n #1025 - Correctly handle disabled special character in ttymodes\n #1033 - Fix compile-time warnings\n #1046 - AIX 5.3 Garbage on Login\n #1054 - Don't terminate connection on getpeername() failure\n #1076 - GSSAPIDelegateCredentials issue mentioned above\n\n - Lots of other improvements and fixes. Please refer to the ChangeLog\n for details\n\nThanks to everyone who has contributed patches, problem or test reports.\n\nChecksums:\n==========\n\n- SHA1 (openssh-4.2.tar.gz) = d2bd777986a30e446268ceeb24cddbf2edf51b21\n- SHA1 (openssh-4.2p1.tar.gz) = 5e7231cfa8ec673ea856ce291b78fac8b380eb78\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 4.1 was released on 2005-05-26. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\n\nChanges since OpenSSH 4.0:\n============================ \n\n* This is a bugfix release, no new features have been added. Some notable\n fixes are:\n\n - Fix segfault when using forwardings configured in ssh_config(5) and \n ClearAllForwardings (bugzilla #996)\n\n - Limit input buffer size for channels. A peer could send more data\n than the buffer code was willing to accept. This would cause OpenSSH\n to abort the connection (bugzilla #896)\n\n* Several improvements to the regression tests\n\n* Portable OpenSSH:\n\n - OpenSSH will now always normalise IPv4 in IPv6 mapped addresses back to \n IPv4 addresses. This means that IPv4 addresses in log messages on IPv6\n enabled machines will no longer be prefixed by "::ffff:" and AllowUsers,\n DenyUsers, AllowGroups, DenyGroups will match IPv4-style addresses only \n for 4-in-6 mapped connections. This ensures a consistent representation\n of IPv4 addresses regardless of whether or not the machine is IPv6\n enabled.\n\n* Other bugfixes, including bugzilla #950, #997, #998, #999, #1005, #1006, \n #1024, and #1038\n\nChanges since OpenSSH 3.9:\n============================ \n\n* ssh(1) now allows the optional specification of an address to bind to \n in port forwarding connections (local, remote and dynamic). Please \n refer to the documentation for the -L and -R options in the ssh(1) \n manual page and the LocalForward and RemoteForward options in the \n ssh_config(5) manpage. (Bugzilla #413)\n\n* To control remote bindings while retaining backwards compatibility,\n sshd(8)'s GatewayPorts option has been extended. To allow client\n specified bind addresses for remote (-R) port forwardings, the server\n must be configured with "GatewayPorts clientspecified".\n\n* ssh(1) and ssh-keyscan(1) now support hashing of host names and \n addresses added to known_hosts files, controlled by the ssh(1) \n HashKnownHosts configuration directive. This option improves user \n privacy by hiding which hosts have been visited. At present this \n option is off by default, but may be turned on once it receives \n sufficient testing.\n\n* Added options for managing keys in known_hosts files to ssh-keygen(1),\n including the ability to search for hosts by name, delete hosts by\n name and convert an unhashed known_hosts file into one with hashed\n names. These are particularly useful for managing known_hosts files\n with hashed hostnames.\n\n* Improve account and password expiry support in sshd(8). The server\n will now warn in advance for both account and password expiry.\n\n* sshd(8) will now log the source of connections denied by AllowUsers,\n DenyUsers, AllowGroups and DenyGroups (Bugzilla #909)\n\n* Added AddressFamily option to sshd(8) to allow global control over\n IPv4/IPv6 usage. (Bugzilla #989)\n\n* Improved sftp(1) client, including bugfixes and optimisations for the \n ``ls'' command and command history and editing support using libedit.\n\n* Improved the handling of bad data in authorized_keys files,\n eliminating fatal errors on corrupt or very large keys. (Bugzilla\n #884)\n\n* Improved connection multiplexing support in ssh(1). Several bugs \n have been fixed and a new "command mode" has been added to allow the\n control of a running multiplexing master connection, including \n checking that it is up, determining its PID and asking it to exit.\n\n* Have scp(1) and sftp(1) wait for the spawned ssh to exit before they\n exit themselves. This prevents ssh from being unable to restore \n terminal modes (not normally a problem on OpenBSD but common with \n -Portable on POSIX platforms). (Bugzilla #950)\n\n* Portable OpenSSH:\n\n - Add *EXPERIMENTAL* BSM audit support for Solaris systems \n (Bugzilla #125)\n\n - Enable IPv6 on AIX where possible (see README.platform for\n details), working around a misfeature of AIX's getnameinfo. \n (Bugzilla #835)\n\n - Teach sshd(8) to write failed login records to btmp for\n unsuccessful auth attempts. Currently this is only for password,\n keyboard-interactive and challenge/response authentication methods\n and only on Linux and HP-UX.\n\n - sshd(8) now sends output from failing PAM session modules to the \n user before exiting, similar to the way /etc/nologin is handled\n\n - Store credentials from gssapi-with-mic authentication early enough\n to be available to PAM session modules when privsep=yes.\n\nChecksums:\n==========\n\n- SHA1 (openssh-4.1.tar.gz) = 62fc9596b20244bb559d5fee3ff3ecc0dfd557cb\n- SHA1 (openssh-4.1p1.tar.gz) = e85d389da8ad8290f5031b8f9972e2623c674e46\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 4.0 was released on 2005-03-09. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\n\nChanges since OpenSSH 3.9:\n============================ \n\n* ssh(1) now allows the optional specification of an address to bind to \n in port forwarding connections (local, remote and dynamic). Please \n refer to the documentation for the -L and -R options in the ssh(1) \n manual page and the LocalForward and RemoteForward options in the \n ssh_config(5) manpage. (Bugzilla #413)\n\n* To control remote bindings while retaining backwards compatibility,\n sshd(8)'s GatewayPorts option has been extended. To allow client\n specified bind addresses for remote (-R) port forwardings, the server\n must be configured with "GatewayPorts clientspecified".\n\n* ssh(1) and ssh-keyscan(1) now support hashing of host names and \n addresses added to known_hosts files, controlled by the ssh(1) \n HashKnownHosts configuration directive. This option improves user \n privacy by hiding which hosts have been visited. At present this \n option is off by default, but may be turned on once it receives \n sufficient testing.\n\n* Added options for managing keys in known_hosts files to ssh-keygen(1),\n including the ability to search for hosts by name, delete hosts by\n name and convert an unhashed known_hosts file into one with hashed\n names. These are particularly useful for managing known_hosts files\n with hashed hostnames.\n\n* Improve account and password expiry support in sshd(8). The server\n will now warn in advance for both account and password expiry.\n\n* sshd(8) will now log the source of connections denied by AllowUsers,\n DenyUsers, AllowGroups and DenyGroups (Bugzilla #909)\n\n* Added AddressFamily option to sshd(8) to allow global control over\n IPv4/IPv6 usage. (Bugzilla #989)\n\n* Improved sftp(1) client, including bugfixes and optimisations for the \n ``ls'' command and command history and editing support using libedit.\n\n* Improved the handling of bad data in authorized_keys files,\n eliminating fatal errors on corrupt or very large keys. (Bugzilla\n #884)\n\n* Improved connection multiplexing support in ssh(1). Several bugs \n have been fixed and a new "command mode" has been added to allow the\n control of a running multiplexing master connection, including \n checking that it is up, determining its PID and asking it to exit.\n\n* Have scp(1) and sftp(1) wait for the spawned ssh to exit before they\n exit themselves. This prevents ssh from being unable to restore \n terminal modes (not normally a problem on OpenBSD but common with \n -Portable on POSIX platforms). (Bugzilla #950)\n\n* Portable OpenSSH:\n\n - Add *EXPERIMENTAL* BSM audit support for Solaris systems \n (Bugzilla #125)\n\n - Enable IPv6 on AIX where possible (see README.platform for\n details), working around a misfeature of AIX's getnameinfo. \n (Bugzilla #835)\n\n - Teach sshd(8) to write failed login records to btmp for\n unsuccessful auth attempts. Currently this is only for password,\n keyboard-interactive and challenge/response authentication methods\n and only on Linux and HP-UX.\n\n - sshd(8) now sends output from failing PAM session modules to the \n user before exiting, similar to the way /etc/nologin is handled\n\n - Store credentials from gssapi-with-mic authentication early enough\n to be available to PAM session modules when privsep=yes.\n\nChecksums:\n==========\n\n- MD5 (openssh-4.0.tgz) = 7dbf15fe7c294672e8822127f50107d0\n- MD5 (openssh-4.0p1.tar.gz) = 7b36f28fc16e1b7f4ba3c1dca191ac92\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 3.9 was released on 2004-08-18. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\n\nChanges since OpenSSH 3.8:\n============================ \n\n* Added new "IdentitiesOnly" option to ssh(1), which specifies that it should\n use keys specified in ssh_config, rather than any keys in ssh-agent(1)\n\n* Make sshd(8) re-execute itself on accepting a new connection. This security\n measure ensures that all execute-time randomisations are reapplied for each \n connection rather than once, for the master process' lifetime. This includes\n mmap and malloc mappings, shared library addressing, shared library mapping \n order, ProPolice and StackGhost cookies on systems that support such things\n\n* Add strict permission and ownership checks to programs reading ~/.ssh/config\n NB ssh(1) will now exit instead of trying to process a config with poor \n ownership or permissions\n\n* Implemented the ability to pass selected environment variables between the\n client and the server. See "AcceptEnv" in sshd_config(5) and "SendEnv" in \n ssh_config(5) for details\n \n* Added a "MaxAuthTries" option to sshd(8), allowing control over the maximum\n number of authentication attempts permitted per connection\n\n* Added support for cancellation of active remote port forwarding sessions. \n This may be performed using the ~C escape character, see "Escape Characters"\n in ssh(1) for details\n \n* Many sftp(1) interface improvements, including greatly enhanced "ls" support \n and the ability to cancel active transfers using SIGINT (^C)\n\n* Implement session multiplexing: a single ssh(1) connection can now carry \n multiple login/command/file transfer sessions. Refer to the "ControlMaster" \n and "ControlPath" options in ssh_config(5) for more information\n\n* The sftp-server has improved support for non-POSIX filesystems (e.g. FAT)\n\n* Portable OpenSSH: Re-introduce support for PAM password authentication, in \n addition to the keyboard-interactive driver. PAM password authentication \n is less flexible, and doesn't support pre-authentication password expiry but\n runs in-process so Kerberos tokens, etc are retained\n\n* Improved and more extensive regression tests\n\n* Many bugfixes and small improvements\n\nChecksums:\n==========\n\n- MD5 (openssh-3.9.tgz) = 93f48bfcc1560895ae53de6bfc41689b\n- MD5 (openssh-3.9p1.tar.gz) = 8e1774d0b52aff08f817f3987442a16e\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 3.8.1p1 was released on 2004-04-19. It is available from \nthe mirrors listed at https://www.openssh.com/.\n\nThis release is a bug-fix release for the portable version. There are\nno feature additions and no corresponding OpenBSD-only release.\n\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source, help\nwith testing and have bought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nBugs fixed since OpenSSH 3.8p1:\n=============================== \n\nBug #673 - Fix compilation on NetBSD with S/Key enabled\n\nBug #748 - Detect and workaround broken name resolution on HP-UX \n\nBug #802 - Fix linking on Tru64 when compiled with SIA support\n\nBug #808 - Fix PAM crash on expired password when not authenticated using \n pam/kbdint mechanism\n\nBug #810 - Fix erroneous clearing of TZ environment variable\n\nBug #811 - Improve locked password detection across Linux variants\n\nBug #820 - Fix utmp corruption on Irix\n\nBug #825 - Fix disconnection problem when using IPv4-in-IPv6 mapped\n addresses on Solaris.\n\n- Fix compilation on OS X systems with Kerberos/GSSAPI\n\n- Many more minor fixes, please refer to the ChangeLog file for details\n\nChecksums:\n==========\n\n- MD5 (openssh-3.8.1p1.tar.gz) = 1dbfd40ae683f822ae917eebf171ca42\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\n3.8.1p1 is a Portable-only release, there is no corresponding 3.8.1 release.\nSee http://www.openssh.com/txt/release-3.8.1p1 for the Portable release notes.\nOpenSSH 3.8 was released on 2004-02-24. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\n\nChanges since OpenSSH 3.7.1:\n============================ \n\n* sshd(8) now supports forced changes of expired passwords via\n /usr/bin/passwd or keyboard-interactive authentication.\n\n Note for AIX: sshd will now deny password access to accounts with\n passwords expired longer than their maxexpired attribute. For\n details, see the AIX section in README.platform.\n\n* ssh(1) now uses untrusted cookies for X11-Forwarding.\n Some X11 applications might need full access to the X11 server,\n see ForwardX11Trusted in ssh(1) and xauth(1) for more information.\n\n* ssh(1) now supports sending application layer\n keep-alive messages to the server. See ServerAliveInterval\n in ssh(1) for more information.\n\n* Improved sftp(1) batch file support.\n\n* New KerberosGetAFSToken option for sshd(8).\n\n* Updated /etc/moduli file and improved performance for\n protocol version 2.\n\n* Support for host keys in DNS (draft-ietf-secsh-dns-xx.txt).\n Please see README.dns in the source distribution for details.\n\n* Fix a number of memory leaks.\n\n* The experimental "gssapi" support has been replaced with\n the "gssapi-with-mic" to fix possible MITM attacks.\n The two versions are not compatible.\n\nChecksums:\n==========\n\n- MD5 (openssh-3.8.tgz) = 7d5590a333d8f8aa1fa6f19e24938700\n- MD5 (openssh-3.8p1.tar.gz) = 7861a4c0841ab69a6eec5c747daff6fb\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nPortable OpenSSH 3.7.1p2 was released on 2003-09-23. It will be available \nfrom the mirrors listed at http://www.openssh.com/portable.html shortly.\n\nPlease note that this is a release to address issues in the portable \nversion only. The items mentioned below do not affect the OpenBSD \nversion.\n\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nSecurity Changes:\n=================\n\n Portable OpenSSH version 3.7p1 and 3.7.1p1 contain multiple \n vulnerabilities in the new PAM authentication code. At least one of\n these bugs is remotely exploitable (under a non-standard \n configuration, with privsep disabled).\n\n OpenSSH 3.7.1p2 fixes these bugs. Please note that these bugs do not \n exist in OpenBSD's releases of OpenSSH.\n\nChanges since OpenSSH 3.7.1p1:\n==============================\n\n* This release disables PAM by default. To enable it, set "UsePAM yes" in \n sshd_config. Due to complexity, inconsistencies in the specification and\n differences between vendors' PAM implementations we recommend that PAM \n be left disabled in sshd_config unless there is a need for its use. \n Sites using only public key or simple password authentication usually \n have little need to enable PAM support.\n\n* This release now requires zlib 1.1.4 to build correctly. Previous \n versions have security problems.\n\n* Fix compilation for versions of OpenSSL before 0.9.6. Some cipher modes \n are not supported for older OpenSSL versions.\n\n* Fix compilation problems on systems with a missing or lacking inet_ntoa()\n function.\n\n* Workaround problems related to unimplemented or broken setresuid/setreuid \n functions on several platforms.\n\n* Fix compilation on older OpenBSD systems.\n\n* Fix handling of password-less authentication (PermitEmptyPasswords=yes) \n that has not worked since the 3.7p1 release.\n\nChecksums:\n==========\n\n- MD5 (openssh-3.7.1p2.tar.gz) = 61cf5b059938718308836d00f6764a94\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 3.7.1 was released on 2004-02-25. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use https://https.openbsd.org/cgi-bin/order\nand for European orders, use https://https.openbsd.org/cgi-bin/order.eu\n\nSecurity Changes:\n=================\n\n All versions of OpenSSH's sshd prior to 3.7.1 contain buffer\n management errors. It is uncertain whether these errors are\n potentially exploitable, however, we prefer to see bugs\n fixed proactively.\n\n OpenSSH 3.7 fixed one of these bugs.\n\n OpenSSH 3.7.1 fixes more similar bugs.\n\nChanges since OpenSSH 3.6.1:\n============================ \n\n* The entire OpenSSH code-base has undergone a license review. As\n a result, all non-ssh1.x code is under a BSD-style license with no\n advertising requirement. Please refer to README in the source\n distribution for the exact license terms.\n\n* Rhosts authentication has been removed in ssh(1) and sshd(8).\n\n* Changes in Kerberos support:\n\n - KerberosV password support now uses a file cache instead of\n a memory cache.\n\n - KerberosIV and AFS support has been removed.\n\n - KerberosV support has been removed from SSH protocol 1.\n\n - KerberosV password authentication support remains for SSH\n protocols 1 and 2.\n\n - This release contains some GSSAPI user authentication support\n to replace legacy KerberosV authentication support. At present\n this code is still considered experimental and SHOULD NOT BE\n USED.\n \n* Changed order that keys are tried in public key authentication.\n The ssh(1) client tries the keys in the following order:\n\n 1. ssh-agent(1) keys that are found in the ssh_config(5) file\n 2. remaining ssh-agent(1) keys\n 3. keys that are only listed in the ssh_config(5) file\n\n This helps when an ssh-agent(1) has many keys, where the sshd(8)\n server might close the connection before the correct key is tried.\n\n* SOCKS5 support has been added to the dynamic forwarding mode\n in ssh(1).\n\n* Removed implementation barriers to operation of SSH over SCTP.\n\n* sftp(1) client can now transfer files with quote characters in\n their filenames.\n\n* Replaced sshd(8)'s VerifyReverseMapping with UseDNS option.\n When UseDNS option is on, reverse hostname lookups are always\n performed.\n\n* Fix a number of memory leaks.\n\n* Support for sending tty BREAK over SSH protocol 2.\n\n* Workaround for other vendor bugs in KEX guess handling.\n\n* Support for generating KEX-GEX groups (/etc/moduli) in ssh-keygen(1).\n\n* Automatic re-keying based on amount of data sent over connection.\n\n* New AddressFamily option on client to select protocol to use (IPv4\n or IPv6).\n\n* Experimental support for the "aes128-ctr", "aes192-ctr", and\n "aes256-ctr" ciphers for SSH protocol 2.\n\n* Experimental support for host keys in DNS (draft-ietf-secsh-dns-xx.txt).\n Please see README.dns in the source distribution for details.\n\n* Portable OpenSSH:\n\n - Replace PAM password authentication kludge with a more correct\n PAM challenge-response module from FreeBSD.\n\n - PAM support may now be enabled/disabled at runtime using the\n UsePAM directive.\n\n - Many improvements to the OpenSC smartcard support.\n\n - Regression tests now work with portable OpenSSH.\n Please refer to regress/README.regress in the source distribution.\n\n - On platforms that support it, portable OpenSSH now honors the\n UMASK, PATH and SUPATH attributes set in /etc/default/login.\n\n - Deny access to locked accounts, regardless of authentication\n method in use.\n\nChecksums:\n==========\n\n- MD5 (openssh-3.7.1.tgz) = 3d2f1644d6a3d3267e5e2421f1385129\n- MD5 (openssh-3.7.1p1.tar.gz) = f54e574e606c08ef63ebb1ab2f7689dc\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 3.7 was released on 2003-09-16. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n http://www.openbsd.org/tshirts.html#18\n\nFor international orders use http://https.openbsd.org/cgi-bin/order\nand for European orders, use http://https.openbsd.org/cgi-bin/order.eu\n\nSecurity Changes:\n=================\n\n All versions of OpenSSH's sshd prior to 3.7 contain a buffer\n management error. It is uncertain whether this error is\n potentially exploitable, however, we prefer to see bugs\n fixed proactively.\n\n OpenSSH 3.7 fixes this bug.\n\nChanges since OpenSSH 3.6.1:\n============================ \n\n* The entire OpenSSH code-base has undergone a license review. As\n a result, all non-ssh1.x code is under a BSD-style license with no\n advertising requirement. Please refer to README in the source\n distribution for the exact license terms.\n\n* Rhosts authentication has been removed in ssh(1) and sshd(8).\n\n* Changes in Kerberos support:\n\n - KerberosV password support now uses a file cache instead of\n a memory cache.\n\n - KerberosIV and AFS support has been removed.\n\n - KerberosV support has been removed from SSH protocol 1.\n\n - KerberosV password authentication support remains for SSH\n protocols 1 and 2.\n\n - This release contains some GSSAPI user authentication support\n to replace legacy KerberosV authentication support. At present\n this code is still considered experimental and SHOULD NOT BE\n USED.\n \n* Changed order that keys are tried in public key authentication.\n The ssh(1) client tries the keys in the following order:\n\n 1. ssh-agent(1) keys that are found in the ssh_config(5) file\n 2. remaining ssh-agent(1) keys\n 3. keys that are only listed in the ssh_config(5) file\n\n This helps when an ssh-agent(1) has many keys, where the sshd(8)\n server might close the connection before the correct key is tried.\n\n* SOCKS5 support has been added to the dynamic forwarding mode\n in ssh(1).\n\n* Removed implementation barriers to operation of SSH over SCTP.\n\n* sftp(1) client can now transfer files with quote characters in\n their filenames.\n\n* Replaced sshd(8)'s VerifyReverseMapping with UseDNS option.\n When UseDNS option is on, reverse hostname lookups are always\n performed.\n\n* Fix a number of memory leaks.\n\n* Support for sending tty BREAK over SSH protocol 2.\n\n* Workaround for other vendor bugs in KEX guess handling.\n\n* Support for generating KEX-GEX groups (/etc/moduli) in ssh-keygen(1).\n\n* Automatic re-keying based on amount of data sent over connection.\n\n* New AddressFamily option on client to select protocol to use (IPv4\n or IPv6).\n\n* Experimental support for the "aes128-ctr", "aes192-ctr", and\n "aes256-ctr" ciphers for SSH protocol 2.\n\n* Experimental support for host keys in DNS (draft-ietf-secsh-dns-xx.txt).\n Please see README.dns in the source distribution for details.\n\n* Portable OpenSSH:\n\n - Replace PAM password authentication kludge with a more correct\n PAM challenge-response module from FreeBSD.\n\n - PAM support may now be enabled/disabled at runtime using the\n UsePAM directive.\n\n - Many improvements to the OpenSC smartcard support.\n\n - Regression tests now work with portable OpenSSH.\n Please refer to regress/README.regress in the source distribution.\n\n - On platforms that support it, portable OpenSSH now honors the\n UMASK, PATH and SUPATH attributes set in /etc/default/login.\n\n - Deny access to locked accounts, regardless of authentication\n method in use.\n\nChecksums:\n==========\n\n- MD5 (openssh-3.7.tgz) = 86864ecc276c5f75b06d4872a553fa70\n- MD5 (openssh-3.7p1.tar.gz) = 77662801ba2a9cadc0ac10054bc6cb37\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller, Ben Lindstrom, Darren Tucker and Tim Rice.\nOpenSSH 3.6.1p2 was released on 2003-04-30. It is available from the\nmirrors listed at https://www.openssh.com/.This is a release\nof the Portable version only.\n\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\nChanges since OpenSSH 3.6.1p1:\n============================ \n\n* Security: corrected linking problem on AIX/gcc. AIX users are \n advised to upgrade immediately. For details, please refer to \n separate advisory (aixgcc.adv). \n\n* Corrected build problems on Irix\n\n* Corrected build problem when building with AFS support\n\n* Merged some changes from Openwall Linux\n\n\nChecksums:\n==========\n\n- MD5 (openssh-3.6p1.tar.gz) = f3879270bffe479e1bd057aa36258696\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.6.1 was released on 2003-04-01. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n:#18\n::\n\n:\n========================== \n\n* The 'kex guesses' bugfix from OpenSSH 3.6 triggers a bug\n in a few other SSH v2 implementations and causes connections to\n stall. OpenSSH 3.6.1 disables this bugfix when interoperating\n with these implementations.\n\n\n:\n============================================\n\n* RSA blinding is now used by ssh(1), sshd(8) and ssh-agent(1).\n in order to avoid potential timing attacks against the RSA keys.\n Older versions of OpenSSH have been using RSA blinding in\n ssh-keysign(1) only.\n\n Please note that there is no evidence that the SSH protocol is\n vulnerable to the OpenSSL/TLS timing attack described in\n:\n* ssh-agent(1) optionally requires user confirmation if a key gets\n used, see '-c' in ssh-add(1).\n\n* sshd(8) now handles PermitRootLogin correctly when UsePrivilegeSeparation\n is enabled.\n\n* sshd(8) now removes X11 cookies when a session gets closed.\n\n* ssh-keysign(8) is disabled by default and only enabled if the\n new EnableSSHKeysign option is set in the global ssh_config(5)\n file.\n\n* ssh(1) and sshd(8) now handle 'kex guesses' correctly (key exchange\n guesses).\n\n* ssh(1) no longer overwrites SIG_IGN. This matches behaviour from\n rsh(1) and is used by backup tools.\n\n* setting ProxyCommand to 'none' disables the proxy feature, see\n ssh_config(5).\n\n* scp(1) supports add -1 and -2.\n\n* scp(1) supports bandwidth limiting.\n\n* sftp(1) displays a progressmeter.\n\n* sftp(1) has improved error handling for scripting.\n\n\n:\n==========\n\n- MD5 (openssh-3.6.1p1.tar.gz) = d4c2c88b883f097fe88e327cbb4b2e2a\n- MD5 (openssh-3.6.1.tgz) = aa2acd2be17dc3fd514a1e09336aab51\n\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.6 was released on 2003-03-31. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport to the project, especially those who contributed source and\nbought T-shirts or posters.\n\nWe have a new design of T-shirt available, more info on\n:#18\n::\n\n:\n============================ \n\n\n* RSA blinding is now used by ssh(1), sshd(8) and ssh-agent(1).\n in order to avoid potential timing attacks against the RSA keys.\n Older versions of OpenSSH have been using RSA blinding in\n ssh-keysign(1) only.\n\n Please note that there is no evidence that the SSH protocol is\n vulnerable to the OpenSSL/TLS timing attack described in\n:\n* ssh-agent(1) optionally requires user confirmation if a key gets\n used, see '-c' in ssh-add(1).\n\n* sshd(8) now handles PermitRootLogin correctly when UsePrivilegeSeparation\n is enabled.\n\n* sshd(8) now removes X11 cookies when a session gets closed.\n\n* ssh-keysign(8) is disabled by default and only enabled if the\n new EnableSSHKeysign option is set in the global ssh_config(5)\n file.\n\n* ssh(1) and sshd(8) now handle 'kex guesses' correctly (key exchange\n guesses).\n\n* ssh(1) no longer overwrites SIG_IGN. This matches behaviour from\n rsh(1) and is used by backup tools.\n\n* setting ProxyCommand to 'none' disables the proxy feature, see\n ssh_config(5).\n\n* scp(1) supports add -1 and -2.\n\n* scp(1) supports bandwidth limiting.\n\n* sftp(1) displays a progressmeter.\n\n* sftp(1) has improved error handling for scripting.\n\n\n:\n==========\n\n- MD5 (openssh-3.6p1.tar.gz) = 72ef1134d521cb6926c99256dad17fe0\n- MD5 (openssh-3.6.tgz) = 758822b888c5c3f83a98045aef904254\n\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.5 was released on 2002-10-15. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\nChanges since OpenSSH 3.4:\n============================ \n\n* Improved support for Privilege Separation (Portability, Kerberos,\n PermitRootLogin handling).\n\n* ssh(1) prints out all known host keys for a host if it receives an\n unknown host key of a different type.\n\n* Fixed AES/Rijndael EVP integration for OpenSSL < 0.9.7 (caused\n problems with bounds checking patches for gcc).\n\n* ssh-keysign(8) is disabled by default and only enabled if the\n HostbasedAuthentication option is enabled in the global ssh_config(5)\n file.\n\n* ssh-keysign(8) uses RSA blinding in order to avoid timing attacks\n against the RSA host key.\n\n* A use-after-free bug was fixed in ssh-keysign(8). This bug\n broke hostbased authentication on several platforms.\n\n* ssh-agent(1) is now installed setgid in order to avoid ptrace(2)\n attacks.\n\n* ssh-agent(1) now restricts the access with getpeereid(2) (or\n equivalent, where available).\n\n* sshd(8) no longer uses the ASN.1 parsing code from libcrypto when\n verifying RSA signatures.\n\n* sshd(8) now sets the SSH_CONNECTION environment variable.\n\n* Enhanced "ls" support for the sftp(1) client, including globbing and\n detailed listings.\n\n* ssh(1) now always falls back to uncompressed sessions, if the\n server does not support compression.\n\n* The default behavior of sshd(8) with regard to user settable\n environ variables has changed: the new option PermitUserEnvironment\n is disabled by default, see sshd_config(5).\n\n* The default value for LoginGraceTime has been changed from 600 to 120\n seconds, see sshd_config(5).\n\n* Removed erroneous SO_LINGER handling.\n\n\nChecksums:\n==========\n\n- MD5 (openssh-3.5p1.tar.gz) = 42bd78508d208b55843c84dd54dea848\n- MD5 (openssh-3.5.tgz) = 79fc225dbe0fe71ebb6910f449101d23\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.4 was released on 2002-06-26. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\nChanges since OpenSSH 3.3:\n============================ \n\nSecurity Changes:\n=================\n\n All versions of OpenSSH's sshd between 2.9.9 and 3.3\n contain an input validation error that can result in\n an integer overflow and privilege escalation.\n\n OpenSSH 3.4 fixes this bug.\n\n In addition, OpenSSH 3.4 adds many checks to detect \n invalid input and mitigate resource exhaustion attacks.\n\n OpenSSH 3.2 and later prevent privilege escalation\n if UsePrivilegeSeparation is enabled in sshd_config.\n OpenSSH 3.3 enables UsePrivilegeSeparation by\n default.\n\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.3 was released on 2002-06-21. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\n:\n============================ \n\n:\n=================\n\n:\n\n\tprivilege separation is now enabled by default\n\n See UsePrivilegeSeparation in sshd_config(5)\n:- ssh no longer needs to be installed setuid root for protocol\n version 2 hostbased authentication, see ssh-keysign(8).\n protocol version 1 rhosts-rsa authentication still requires privileges\n and is not recommended.\n\n:\n==============\n\n- documentation for the client and server configuration options have\n been moved to ssh_config(5) and sshd_config(5).\n- the server now supports the Compression option, see sshd_config(5).\n- the client options RhostsRSAAuthentication and RhostsAuthentication now\n default to no, see ssh_config(5).\n- the client options FallBackToRsh and UseRsh are deprecated.\n- ssh-agent now supports locking and timeouts for keys, see ssh-add(1).\n- ssh-agent can now bind to unix-domain sockets given on the command line,\n see ssh-agent(1).\n- fixes problems with valid RSA signatures from putty clients.\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.2.3 was released on 2002-05-23. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\nChanges since OpenSSH 3.2.2:\n============================ \n\n This release fixes several problems in OpenSSH 3.2.2:\n\n - a defect in the BSD_AUTH access control handling for\n OpenBSD and BSD/OS systems:\n\n Under certain conditions, on systems using YP with netgroups\n in the password database, it is possible that sshd does ACL\n checks for the requested user name but uses the password\n database entry of a different user for authentication. This\n means that denied users might authenticate successfully while\n permitted users could be locked out (OpenBSD PR 2659).\n\n - login/tty problems on Solaris (bug #245)\n\n - build problems on Cygwin systems\n\n\nChanges between OpenSSH 3.1 and OpenSSH 3.2.2:\n==============================================\n\n Security Changes:\n =================\n \n - fixed buffer overflow in Kerberos/AFS token passing\n - fixed overflow in Kerberos client code\n - sshd no longer auto-enables Kerberos/AFS\n - experimental support for privilege separation,\n see UsePrivilegeSeparation in sshd(8) and\n \t http://www.citi.umich.edu/u/provos/ssh/privsep.html\n for more information.\n - only accept RSA keys of size SSH_RSA_MINIMUM_MODULUS_SIZE (768) or larger\n \n Other Changes:\n ==============\n \n - improved smartcard support (including support for OpenSC,\n see www.opensc.org)\n - improved Kerberos support (including support for MIT-Kerberos V)\n - fixed stderr handling in protocol v2\n - client reports failure if -R style TCP forwarding fails in protocol v2\n - support configuration of TCP forwarding during interactive sessions (~C)\n - improved support for older sftp servers\n - improved support for importing old DSA keys (from ssh.com software).\n - client side support for PASSWD_CHANGEREQ in protocol v2\n - fixed waitpid race conditions\n - record correct lastlogin time\n\nReporting Bugs:\n===============\n\n- please read http://www.openssh.com/report.html\n and http://bugzilla.mindrot.org/\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.2.2 was released on 2002-05-16. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n:\n=================\n\n- fixed buffer overflow in Kerberos/AFS token passing\n- fixed overflow in Kerberos client code\n- sshd no longer auto-enables Kerberos/AFS\n- experimental support for privilege separation,\n see UsePrivilegeSeparation in sshd(8) and\n:- only accept RSA keys of size SSH_RSA_MINIMUM_MODULUS_SIZE (768) or larger\n\n:\n==============\n\n- improved smartcard support (including support for OpenSC, see www.opensc.org)\n- improved Kerberos support (including support for MIT-Kerberos V)\n- fixed stderr handling in protocol v2\n- client reports failure if -R style TCP forwarding fails in protocol v2\n- support configuration of TCP forwarding during interactive sessions (~C)\n- improved support for older sftp servers\n- improved support for importing old DSA keys (from ssh.com software).\n- client side support for PASSWD_CHANGEREQ in protocol v2\n- fixed waitpid race conditions\n- record correct lastlogin time\n\n:\n===============\n\n:OpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.1 was released on 2004-04-09. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n\nImportant Changes:\n==================\n\n- /etc/ssh/ now default directory for keys and configuration files\n- ssh-keygen no longer defaults to a specific key type (rsa1);\n use ssh-keygen -t {rsa,dsa,rsa1}\n- sshd x11 forwarding listens on localhost by default;\n see sshd X11UseLocalhost option to revert to prior behaviour\n if your older X11 clients do not function with this configuration\n\n\nOther Changes:\n==============\n\n- ssh ~& escape char functions now for both protocol versions\n- sshd ReverseMappingCheck option changed to VerifyReverseMapping\n to clarify its function; ReverseMappingCheck can still be used\n- public key fingerprint is now logged with LogLevel=VERBOSE\n- reason logged for disallowed logins (e.g., no shell, etc.)\n- more robust error handling for x11 forwarding\n- improved packet/window size handling in ssh2\n- use of regex(3) has been removed\n- fix SIGCHLD races in sshd (seen on Solaris)\n- sshd -o option added\n- sftp -B -R -P options added\n- ssh-add now adds all 3 default keys\n- ssh-keyscan bug fixes\n- ssh-askpass for hostkey dialog\n- fix fd leak in sshd on SIGHUP\n- TCP_NODELAY set on X11 and TCP forwarding endpoints\n\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.0.2 was released on 2002-12-04. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\nImportant Changes:\n==================\n\n This release fixes a vulnerability in the UseLogin option\n of OpenSSH. This option is not enabled in the default\n installation of OpenSSH.\n\n However, if UseLogin is enabled by the administrator, all\n versions of OpenSSH prior to 3.0.2 may be vulnerable to\n local attacks.\n\n The vulnerability allows local users to pass environment\n variables (e.g. LD_PRELOAD) to the login process. The login\n process is run with the same privilege as sshd (usually\n with root privilege).\n\n Do not enable UseLogin on your machines or disable UseLogin\n again in /etc/sshd_config:\n\t\tUseLogin no\n\nWe also have received many reports about attacks against the crc32\nbug. This bug has been fixed about 12 months ago in OpenSSH 2.3.0.\nHowever, these attacks cause non-vulnerable daemons to chew a lot\nof cpu since the crc32 attack sends a tremendously large amount of\ndata which must be processed.\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\n\n\n\nThe following patch fixes the UseLogin vulnerability in OpenSSH 3.0.1 and\nearlier releases.\n\n--- session.c\t11 Oct 2001 13:45:21 -0000\t1.108\n+++ session.c\t1 Dec 2001 22:14:39 -0000\n@@ -875,6 +875,7 @@\n \t\tchild_set_env(&env, &envsize, "TZ", getenv("TZ"));\n \n \t/* Set custom environment options from RSA authentication. */\n+\tif (!options.use_login)\n \twhile (custom_environment) {\n \t\tstruct envstring *ce = custom_environment;\n \t\tchar *s = ce->s;\nOpenSSH 3.0.1 was released on 2001-11-19. It is available from the\nmirrors listed at https://www.openssh.com/.\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\nImportant Changes:\n==================\n\n A security hole that may allow an attacker to authenticate\n if -- and only if -- the administrator has enabled KerberosV.\n By default, OpenSSH KerberosV support only becomes active\n after KerberosV has been properly configured.\n\n An excessive memory clearing bug (which we believe to be\n unexploitable) also exists, but since this may cause daemon\n crashes, we are providing a fix as well.\n\n Various other non-critical fixes (~& support and more).\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 3.0 was released on 2001-11-06. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nThis release contains many portability bug-fixes (listed in the\nChangeLog) as well as several new features (listed below).\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n:\n==================\n\n1) SSH protocol v2 is now the default protocol version\n\n\tuse the 'Protocol' option from ssh(1) and sshd(8) if\n\tyou need to change this.\n\n2) The files\n\t/etc/ssh_known_hosts2\n\t~/.ssh/known_hosts2\n\t~/.ssh/authorized_keys2\n are now obsolete, you can use\n\t/etc/ssh_known_hosts\n\t~/.ssh/known_hosts\n\t~/.ssh/authorized_keys\n For backward compatibility ~/.ssh/authorized_keys2 will still used for\n authentication and hostkeys are still read from the known_hosts2.\n However, those deprecated files are considered 'readonly'. Future\n releases are likely not to read these files.\n\n3) The CheckMail option in sshd_config is deprecated, as sshd(8) no longer\n checks for new mail.\n\n4) X11 cookies are now stored in $HOME.\n\n:\n=============\n\n1) Smartcard support in the ssh client and agent based on work by\n University of Michigan CITI (:2) support for Rekeying in protocol version 2\n\n3) improved Kerberos support in protocol v1 (KerbIV and KerbV)\n\n4) backward compatibility with older commercial SSH versions >= 2.0.10\n\n5) getopt(3) is now used by all programs\n\n6) dynamic forwarding (use ssh(1) as your socks server)\n\n7) ClearAllForwardings in ssh(1)\n\n8) ssh(1) now checks the hostkey for localhost (NoHostAuthenticationForLocalhost yes/no).\n\n9) -F option in ssh(1)\n\n10) ssh(1) now has a '-b bindaddress' option\n\n11):\n12) The AuthorizedKeysFile option allows specification of alternative\n files that contain the public keys that can be used for user authentication\n (e.g. /etc/ssh_keys/%u, see sshd(8))\n\n13) extended AllowUsers user@host syntax in sshd(8)\n\n14) improved challenge-response support (especially for systems supporting BSD_AUTH)\n\n15) sshd(8) can specify time args as 1h, 2h30s etc.\n\n16) sshd(8) transmits the correct exit status for remote execution with protocol version 2.\n\n17) ssh-keygen(1) can import private RSA/DSA keys generated with the commercial version\n\n18) ssh-keyscan(1) supports protocol version 2\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nPortable OpenSSH 2.9p2 has just been uploaded and shall be making its\nway to the mirrors listed at http://www.openssh.com/portable.html\nshortly.\n\nThis release fixes the "cookies" file deletion problem reported on\nBUGTRAQ as well as a few other minor (non-security) bugs. No new\nfeatures have been added in this release.\n\nRegards,\nDamien Miller\nOpenSSH 2.9.9 has just been uploaded. It is available from the\n:OpenSSH 2.9.9 fixes a weakness in the key file option handling,\nincluding source IP based access control.\n\nOpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nThis release contains many portability bug-fixes (listed in the\nChangeLog) as well as several new features (listed below).\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n:\n===============\n\nThis release fixes weakness in the source IP based access control\n:\n\n Versions of OpenSSH between 2.5 and 2.9.9 are\n affected if they use the 'from=' key file option in\n combination with both RSA and DSA keys in\n ~/.ssh/authorized_keys2.\n\n Depending on the order of the user keys in\n ~/.ssh/authorized_keys2 sshd might fail to apply the\n source IP based access control restriction (e.g.\n from="10.0.0.1") to the correct key:\n\n If a source IP restricted key (e.g. DSA key) is\n immediately followed by a key of a different type\n (e.g. RSA key), then key options for the second key\n are applied to both keys, which includes 'from='.\n\n\tThis means that users can circumvent the system policy\n\tand login from disallowed source IP addresses.\n\t\n\n:\n==================\n\nOpenSSH 2.9.9 might have upgrade issues introduced by the long time\n:\n\n1) The files\n\t/etc/ssh_known_hosts2\n\t~/.ssh/known_hosts2\n\t~/.ssh/authorized_keys2\n are now obsolete, you can use\n\t/etc/ssh_known_hosts\n\t~/.ssh/known_hosts\n\t~/.ssh/authorized_keys\n For backward compatibility ~/.ssh/authorized_keys2 is still used for\n authentication and hostkeys are still read from the known_hosts2.\n However, old files are considered 'readonly'. Future releases are\n likely to not read these files.\n\n2) The CheckMail option in sshd_config is deprecated, sshd no longer\n checks for new mail.\n\n3) X11 cookies are stored in $HOME\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nOpenSSH 2.9 has just been uploaded. It is available from the\n:OpenSSH is a 100% complete SSH protocol version 1.3, 1.5 and 2.0\nimplementation and includes sftp client and server support.\n\nThis release contains many portability bug-fixes (listed in the\nChangeLog) as well as several new features (listed below).\n\nWe would like to thank the OpenSSH community for their continued\nsupport and encouragement.\n\n:\n==================\n\n: \n\tuse the 'Protocol' option from ssh(1) and sshd(8) if\n\tyou want to change this.\n\n:\n\n HostbasedAuthentication, similar to RhostsRSA in SSH protocol\n v1\n\n Rekeying (negotiate new encryption keys for the current SSH\n session, try ~R in interactive SSH sessions)\n\n:\n:\n\n gid swapping in sshd (fixes access to /home/group/user based\n directory structures)\n\n Dan Kaminsky <dankamin@cisco.com> contributed an experimental\n SOCKS4 proxy to the ssh client (yes, client not the server).\n Use 'ssh -D 1080 server' if you want to try this out.\n\n\tserver option PrintLastLog\n\n\timprovements for scp > 2GB\n\n\timproved ListenAddress option.\n:\n\timproved interoperability (bug detection for older implementations)\n\n\timproved documentation\n\nOpenSSH is brought to you by Markus Friedl, Niels Provos, Theo de Raadt,\nKevin Steves, Damien Miller and Ben Lindstrom.\nPortable OpenSSH 2.5.2p2 is now available from the mirror sites\n:\n:\n:\n:\n:\n:\n:\n:\n:\n-d\n\n:Portable OpenSSH 2.5.1p2 has just been uploaded and will be making its\nway to the mirror sites (http://www.openssh.com/portable.html)in due\ncourse.\n\nThis release contains primarily bug-fixes over 2.5.1p1 but an upgrade is\nrecommended. Specific bug-fixes include:\n\n - Fixed endianness issue causing failures when using Rijndael/AES cipher\n - Fix PAM failures on Solaris and Linux\n - Fix RPM spec file for Redhat systems\n - Fixed several compatibility functions\n - Fix entropy collection code for SCO3 and NeXTStep\n - Many other minor fixes (see Changelog for details)\n\nThis release includes Mark Roth's mdoc2man.pl script which can be used\nto fix up the manpages on systems that lack the full andoc set of\nmacros (e.g. Solaris). A future release of portable OpenSSH will automate\nthis scripts use for systems that require it.\n\n-d\n\n\n\n & \n\n\n\n\n \n\n\n\n\n\n\n\n\n==================\n\n \n\n \n \n \n \n\n \n \n\n \n\n \n\n \n\n \n \n\n \n \n \n\n \n\n \n \n\n \n\n \n\n \n\n \n \n \n\n \n\n \n \n\n \n\n \n\n \n \n\n \n\n \n \n \n \n \n\n\n\nThis is to announce the release of portable openssh-2.3.0p1. This\nrelease includes many new features and bug fixes. This is a\nrecommended upgrade if you are using 2.2.0p1 or an older release.\n\nPortable OpenSSH is available from one of the many mirrors listed at\nhttp://www.openssh.com/portable.html\n\nSome of the more notable features include:\n\n- Rijndael support for SSH2. Use the "Ciphers" configuration directive \n to enable it. (Markus Friedl <markus@cvs.openbsd.org>\n\n- Cygwin support (Corinna Vinschen <vinschen@cygnus.com>)\n\n- sftp-server support (Markus Friedl <markus@cvs.openbsd.org>)\n\n- SSH1 single-des support for interop with Cisco routers. This cipher \n is never enabled automatically, you have to either specify "-c des" \n or enable it using "Cipher des" in a config file. (Markus Friedl\n <markus@cvs.openbsd.org>\n\n- Support expired password change through PAM (Steve VanDevender's\n <stevev@darkwing.uoregon.edu>)\n\n- Better compatibility with buggy SSH implementations (Markus Friedl\n <markus@cvs.openbsd.org>\n\n- S/key support for SSH2, based on kbd-interactive auth (Markus Friedl\n <markus@cvs.openbsd.org> and mkiernan@avantgo.com\n\n- scp now supports "-o" option (Ben Lindstron <mouring@pconline.com>)\n\nPlease refer to the ChangeLog for a full list of features and bugfixes.\n\nRegards,\nDamien Miller\nVersion 2.2.0p1 of portable OpenSSH has just been uploaded to the\nmaster site and should be making its way to the mirrors in due\ncourse. \n\nhttp://www.openssh.com/portable.html\n\nThis release contains several new features and bugfixes relative to\nthe previous 2.1.1p4 release. In particular:\n\n- DSA key support in ssh-agent. Please not that this will not \n interop with ssh.com's ssh-agent (Markus Friedl)\n- sshd now implements Random Early Drop connection rate limiting,\n which can help mitigate DoS attacks against sshd. See the \n `MaxStartups' option in the sshd manpage for details (Markus Friedl)\n- `-u' option to sshd allow logging of hostnames (rather than IP \n addresses) in wtmp when `UseLogin' is set to `yes'. (Markus Friedl)\n- Escape character `~' support in SSH2 (Markus Friedl)\n- Interop with SSH.COM ssh 2.3.0 (Markus Friedl)\n- Fix problems when sshd is run from inetd\n- Better SunOS 4.1.x support (Nate Itkin and Charles Levert)\n- Solaris package support, see contrib/solaris (Rip Loomis)\n- Work around connection freezes on HPUX and SunOS 4 (Lutz Jaenicke,\n Tamito KAJIYAMA)\n- Fix ^C ignored issue on Solaris. (Gert Doering, John Horne and \n Garrick James)\n- Further improved NeXT support. (Ben Lindstrom, Mark Miller)\n- Lots of other minor fixes (see ChangeLog for details)\n\nThis release has been tested on HPUX (10.20, 11.00), Irix (5.3,\n6.5), Linux (Debian, Redhat, Slackware, SuSE), NeXTstep 3 (HPPA,\ni386, m68k), OpenStep (i386, m68k, Sparc), SCO Unixware 7.1.0, SCO\nOpenServer 5.0.5, Solaris 2.7 (Sparc), Solaris 2.8 (i386, Sparc),\nSNI/Reliant Unix, DEC OSF/Tru64 5.0.\n\nMany thanks to those who contributed bug reports, fixes and testing \ntime.\n\nRegards,\nDamien Miller\nI have just uploaded portable OpenSSH 2.1.1p4, it should be making\nits way to the mirrors listed at http://www.openssh.com/portable.html\nsoon.\n\nThis release contains several bugfixes from the OpenBSD team,\nprimarily the config file parsing problem reported by Ralf \nEngelschall <rse@engelschall.com>\n\nRegards,\nDamien Miller\n\n--------------- Changelog:\n\n20000716\n - Release 2.1.1p4\n\n20000715\n - (djm) OpenBSD CVS updates\n - provos@cvs.openbsd.org 2000/07/13 16:53:22\n \n allow multiple whitespace but only one '=' between tokens, bug report from\n Ralf S. Engelschall <rse@engelschall.com> but different fix. okay deraadt@\n - provos@cvs.openbsd.org 2000/07/13 17:14:09\n \n typo; todd@fries.net\n - provos@cvs.openbsd.org 2000/07/13 17:19:31\n \n close can fail on AFS, report error; from Greg Hudson <ghudson@mit.edu>\n - markus@cvs.openbsd.org 2000/07/14 16:59:46\n \n allow leading whitespace. ok niels\n - djm@cvs.openbsd.org 2000/07/14 22:01:38\n \n Always create ~/.ssh with mode 700; ok Markus\n - Fixes for SunOS 4.1.4 from Gordon Atwood <gordon@cs.ualberta.ca>\n - Include floatingpoint.h for entropy.c\n - strerror replacement\n\n---------------\nThe 2.1.1p3 release of portable OpenSSH has been uploaded to the\nOpenBSD ftp master site. In a few hours it will be available from one\nof the many mirrors listed at:\n\nhttp://www.openssh.com/portable.html\n\nThis release fixes several bugs reported since the previous release\nand extends portability to NeXT and Reliant Unix.\n\nAs usual, the OpenBSD team has been hard at work further polishing and\nenhancing OpenSSH. This release brings a new configuration directive\n"MaxStartups" which mitigates connection flooding attacks, further\ndetails are in the sshd man-page.\n\nAnother noteworthy difference from previous releases is that\n'FallBackToRsh' now defaults to 'no'. Users of this feature may need\nto edit their /etc/ssh_config or ~/.ssh/config files to achieve the\nsame behavior.\n\nAgain, thanks to those who reported bugs, tested the snapshot and sent\nfixes.\n\nRegards,\nDamien Miller\n\n------------------ Changelog\n\n20000712\n - (djm) Remove -lresolve for Reliant Unix\n - (djm) OpenBSD CVS Updates:\n - deraadt@cvs.openbsd.org 2000/07/11 02:11:34\n \n make MaxStartups code still work with -d; djm\n - deraadt@cvs.openbsd.org 2000/07/11 13:17:45\n \n disable FallBackToRsh by default\n - (djm) Replace in_addr_t with u_int32_t in bsd-inet_aton.c. Report from\n Ben Lindstrom <mouring@pconline.com>\n - (djm) Make building of X11-Askpass and GNOME-Askpass optional in RPM\n spec file.\n - (djm) Released 2.1.1p3\n\n20000711\n - (djm) Fixup for AIX getuserattr() support from Tom Bertelson\n <tbert@abac.com>\n - (djm) ReliantUNIX support from Udo Schweigert <ust@cert.siemens.de>\n - (djm) NeXT: dirent structures to get scp working from Ben Lindstrom \n <mouring@pconline.com>\n - (djm) Fix broken inet_ntoa check and ut_user/ut_name confusion, report \n from Jim Watt <jimw@peisj.pebio.com>\n - (djm) Replaced bsd-snprintf.c with one from Mutt source tree, it is known\n to compile on more platforms (incl NeXT).\n - (djm) Added bsd-inet_aton and configure support for NeXT\n - (djm) Misc NeXT fixes from Ben Lindstrom <mouring@pconline.com>\n - (djm) OpenBSD CVS updates:\n - markus@cvs.openbsd.org 2000/06/26 03:22:29\n \n cleanup, less cut&paste\n - markus@cvs.openbsd.org 2000/06/26 15:59:19\n \n MaxStartups: limit number of unauthenticated connections, work by \n theo and me\n - deraadt@cvs.openbsd.org 2000/07/05 14:18:07\n \n use no_x11_forwarding_flag correctly; provos ok\n - provos@cvs.openbsd.org 2000/07/05 15:35:57\n \n typo\n - aaron@cvs.openbsd.org 2000/07/05 22:06:58\n \n Insert more missing .El directives. Our troff really should identify \n these and spit out a warning.\n - todd@cvs.openbsd.org 2000/07/06 21:55:04\n \n clean code is good code\n - deraadt@cvs.openbsd.org 2000/07/07 02:14:29\n \n sense of port forwarding flag test was backwards\n - provos@cvs.openbsd.org 2000/07/08 17:17:31\n \n replace strtok with strsep; from David Young <dyoung@onthejob.net>\n - deraadt@cvs.openbsd.org 2000/07/08 19:21:15\n \n KNF\n - ho@cvs.openbsd.org 2000/07/08 19:27:33\n \n Better conditions for strsep() ending.\n - ho@cvs.openbsd.org 2000/07/10 10:27:05\n \n Get the correct message on errors. (niels@ ok)\n - ho@cvs.openbsd.org 2000/07/10 10:30:25\n \n strtok() --> strsep(). (niels@ ok)\n - (djm) Fix problem with debug mode and MaxStartups\n - (djm) Don't generate host keys when $(DESTDIR) is set (e.g. during RPM\n builds)\n - (djm) Add strsep function from OpenBSD libc for systems that lack it\n\n20000709\n - (djm) Only enable PAM_TTY kludge for Linux. Problem report from\n Kevin Steves <stevesk@sweden.hp.com>\n - (djm) Match prototype and function declaration for rresvport_af.\n Problem report from Niklas Edmundsson <nikke@ing.umu.se>\n - (djm) Missing $(DESTDIR) on host-key target causing problems with RPM \n builds. Problem report from Gregory Leblanc <GLeblanc@cu-portland.edu>\n - (djm) Replace ut_name with ut_user. Patch from Jim Watt\n <jimw@peisj.pebio.com>\n - (djm) Fix pam sprintf fix\n - (djm) Cleanup entropy collection code a little more. Split initialisation\n from seeding, perform initialisation immediately at start, be careful with\n uids. Based on problem report from Jim Watt <jimw@peisj.pebio.com>\n - (djm) More NeXT compatibility from Ben Lindstrom <mouring@pconline.com>\n Including sigaction() et al. replacements\n - (djm) AIX getuserattr() session initialisation from Tom Bertelson \n <tbert@abac.com>\n\n20000708\n - (djm) Fix bad fprintf format handling in auth-pam.c. Patch from \n Aaron Hopkins <aaron@die.net>\n - (djm) Fix incorrect configure handling of --with-rsh-path option. Fix from\n Lutz Jaenicke <Lutz.Jaenicke@aet.TU-Cottbus.DE>\n - (djm) Fixed undefined variables for OSF SIA. Report from \n Baars, Henk <Hendrik.Baars@nl.origin-it.com>\n - (djm) Handle EWOULDBLOCK returns from read() and write() in atomicio.c \n Fix from Marquess, Steve Mr JMLFDC <Steve.Marquess@DET.AMEDD.ARMY.MIL>\n - (djm) Don't use inet_addr. \n\n20000702\n - (djm) Fix brace mismatch from Corinna Vinschen <vinschen@cygnus.com>\n - (djm) Stop shadow expiry checking from preventing logins with NIS. Based\n on fix from HARUYAMA Seigo <haruyama@nt.phys.s.u-tokyo.ac.jp>\n - (djm) Use standard OpenSSL functions in auth-skey.c. Patch from\n Chris, the Young One <cky@pobox.com>\n - (djm) Fix scp progress meter on really wide terminals. Based on patch \n from James H. Cloos Jr. <cloos@jhcloos.com>\n\n------------------\nAnnouncing the release of portable OpenSSH 2.1.1p2.\n\nThis release primarily contains fixes to the bugs that have been\nreported over the last month, in particular:\n\n - Invalid time bring written to utmp/wtmp on systems using bash2\n\n - Several lastlog fixes\n\n - AIX, SCO, Irix portability fixes\n\n - Avoid failures on PAM systems when using PAM authentication modules\n which require a tty.\n\n - Entropy collection fixes for Solaris.\n\n - EGD robustness improvements\n\n - Fixes and enhancements from the OpenBSD team:\n - Fixed options processing in authorized_keys2 file\n - Compatibility with commercial SSH 2.0.13 and 2.2.0\n - Numerous minor fixes\n\nThere are also a couple of new features:\n\n - Shadow password expiry support (no password change support yet)\n\n - Irix 6.x array sessions, project IDs and system audit trail IDs\n\n - Beginnings of Tru64 / OSF SIA (Security Integration Architecture) \n support\n\n - Beginnings of NeXT support\n\nVersion 2.1.1p2 will be available from the mirrors listed at\nhttp://www.openssh.com/portable.html (as soon as they update).\n\nMany thanks to all those who tested the snapshots and/or contributed\nbug reports and patches\n\nRegards,\nDamien Miller\nAnnouncing the availability of portable OpenSSH 2.1.1p1.\n\nThis release contains the fix for the "UseLogin yes" vulnerability\nidentified in Markus' release and several other enhancements and\nbugfixes. Including:\n\n - Better login code. Andre Lucas has rewritten the login code to \n be much more modular and extensible. In the process he has fixed \n the problems with Solaris utmp[x].\n\n - Revised the entropy collection code to be faster and more reliable.\n\n - Fix for RSA host restrictions ("from=" in authorized_keys)\n\nIt is recommended that all users upgrade to this version.\n\nPortable OpenSSH 2.1.1p1 is available from one of the many mirrors\nlisted at: http://www.openssh.com/portable.html\n\nRegards,\nDamien Miller\nThis is to announce the release of openssh-2.1.0, the first stable\nrelease of portable OpenSSH to incorporate support for the SSH2\nprotocol.\n\nThe SSH2 protocol offers a number of advantages over the SSH1 protocol\nincluding standards compliance (SSH2 is on the IETF standards\ntrack[1]), improved security and operation without RSA (which is\npatented in some countries).\n\nThe SSH2 support in OpenSSH has been developed by Markus Friedl, with\nsupport from the OpenBSD team.\n\nThis is also the first version of the portable version of OpenSSH\nto offer built-in entropy collection. This removes the requirement\nfor EGD on systems that lack a /dev/random driver. As a result,\nOpenSSH-2.1.0 now requires a recent version of OpenSSL[2] to compile\n(version 0.9.5 or later).\n\nNB. The portable version of OpenSSH is currently in the\nprocess of merging its webpages with the official OpenBSD\nproject. Please use http://www.openssh.com/ from now\non. Distribution files are also available from the mirrors listed at\nhttp://violet.ibs.com.au/openssh/files/MIRRORS.html\n\nPlease read http://www.openssh.com/report.html before reporting bugs.\nPatches, bug reports, developer and user queries are welcome on the\nmailing list (http://www.openssh.com/list.html).\nRegards,\nDamien Miller\n\n[1] http://www.ietf.org/html.charters/secsh-charter.html\n[2] http://www.openssl.org/\nThe Unix/Linux port of OpenSSH 1.2.3 was released yesterday and should\nbe available from a mirror near you. A mirror list is available from:\n\nhttp://violet.ibs.com.au/openssh/files/MIRRORS.html\n\nThis release fixes the bugs reported since 1.2.2p1 and contains many\ncleanups from the OpenBSD tree.\n\nIn particular, the OpenSSL detection problems have been resolved.\n\nThe layout has changed a little bit. The packages/ subdirectory has\nbeen replaced with a contrib/ subdirectory which contains platform\nspecific code and other patches. Submissions are welcome.\n\nEnjoy,\nDamien Miller\n\n20000317\n - Clarified --with-default-path option.\n - Added -blibpath handling for AIX to work around stupid runtime linking.\n Problem elucidated by gshapiro@SENDMAIL.ORG by way of Jim Knoble\n <jmknoble@pobox.com>\n - Checks for 64 bit int types. Problem report from Mats Fredholm\n <matsf@init.se>\n - OpenBSD CVS updates:\n - \n \n \n pedantic: signed vs. unsigned, void*-arithm, etc\n - \n Various cleanups and standardizations.\n - Runtime error fix for HPUX from Otmar Stahl \n <O.Stahl@lsw.uni-heidelberg.de>\n\n20000316\n - Fixed configure not passing LDFLAGS to Solaris. Report from David G. \n Hesprich <dghespri@sprintparanet.com>\n - Propagate LD through to Makefile\n - Doc cleanups\n - Added blurb about "scp: command not found" errors to UPGRADING\n\n20000315\n - Fix broken CFLAGS handling during search for OpenSSL. Fixes va_list\n problems with gcc/Solaris.\n - Don't free argument to putenv() after use (in setenv() replacement). \n Report from Seigo Tanimura <tanimura@r.dl.itc.u-tokyo.ac.jp>\n - Created contrib/ subdirectory. Included helpers from Phil Hands' \n Debian package, README file and chroot patch from Ricardo Cerqueira\n <rmcc@clix.pt>\n - Moved gnome-ssh-askpass.c to contrib directory and removed config \n option.\n - Slight cleanup to doc files\n - Configure fix from Bratislav ILICH <bilic@zepter.ru>\n\n20000314\n - Include macro for IN6_IS_ADDR_V4MAPPED. Report from \n peter@frontierflying.com\n - Include /usr/local/include and /usr/local/lib for systems that don't\n do it themselves\n - -R/usr/local/lib for Solaris\n - Fix RSAref detection\n - Fix IN6_IS_ADDR_V4MAPPED macro\n\n20000311\n - Detect RSAref\n - OpenBSD CVS change\n \n - disallow guessing of root password\n - More configure fixes\n - IPv6 workarounds from Hideaki YOSHIFUJI <yoshfuji@ecei.tohoku.ac.jp>\n\n20000309\n - OpenBSD CVS updates to v1.2.3\n\t\n\t - int atomicio -> ssize_t (for alpha). ok deraadt@\n\t\n\t - delay MD5 computation until client sends response, free() early, cleanup.\n\t\n\t - void* -> unsigned char*, ok niels@\n\t\n\t - remove unused variable 'len'. fix comments.\n\t - remove unused variable\n\t\n\t - rename a cpp symbol, to avoid param.h collision\n\t\n\t - missing xfree()\n\t - getsockname() requires initialized tolen; andy@guildsoftware.com\n\t - use getpeername() in packet_connection_is_on_socket(), fixes sshd -i;\n\tfrom Holger.Trapp@Informatik.TU-Chemnitz.DE\n\t\n\t - register cleanup for pty earlier. move code for pty-owner handling to \n \tpty.c ok provos@, dugsong@\n\t\n\t - turn off x11-fwd for the client, too.\n\t\n\t - PKCS#1 padding\n\t\n\t - allow '.' in usernames; from jedgar@fxp.org\n\t\n\t - typo: ignore_user_known_hosts int->flag; naddy@mips.rhein-neckar.de\n\t - sync with sshd_config\n\t\n\t - enable ssh-keygen -l -f ~/.ssh/known_hosts, ok deraadt@\n\t\n\t - Change invalid 'CHAT' loglevel to 'VERBOSE'\n\t\n\t - suppress AAAA query host when '-4' is used; from shin@nd.net.fujitsu.co.jp\n\t - turn off x11-fwd for the client, too.\n\t\n\t - missing xfree()\n\t - retry rresvport_af(), too. from sumikawa@ebina.hitachi.co.jp.\n\t - read error vs. "Connection closed by remote host"\n\t\n\t - ie. -> i.e.,\n\t - do not link to a commercial page..\n\t - sync with sshd_config\n\t\n\t - no need for poll.h; from bright@wintelcom.net\n\t - log with level log() not fatal() if peer behaves badly.\n\t - don't panic if client behaves strange. ok deraadt@\n\t - make no-port-forwarding for RSA keys deny both -L and -R style fwding\n\t - delay close() of pty until the pty has been chowned back to root\n\t - oops, fix comment, too.\n\t - missing xfree()\n\t - move XAUTHORITY to subdir. ok dugsong@. fixes debian bug #57907, too.\n \t(http://cgi.debian.org/cgi-bin/bugreport.cgi?archive=no&bug=57907)\t - register cleanup for pty earlier. move code for pty-owner handling to \n pty.c ok provos@, dugsong@\n\t - create x11 cookie file\n\t - fix pr 1113, fclose() -> pclose(), todo: remote popen()\n\t - version 1.2.3\n - Cleaned up\n - Removed warning workaround for Linux and devpts filesystems (no longer \n required after OpenBSD updates)\n\n20000308\n - Configure fix from Hiroshi Takekawa <takekawa@sr3.t.u-tokyo.ac.jp>\nIt gives me no little pleasure to announce the first stable release\nof the Unix port of OpenSSH. \n\nIt is available in tar. and RPM format from one of the mirrors\nlisted at:\n\nhttp://violet.ibs.com.au/openssh/files/MIRRORS.html\n\nThis release fixes all known issues and is known to compile and\nrecent releases on Linux, Solaris, HPUX and SCO\nUnixware.\n\nPlease review the ChangeLog details on what has changed since\nthe last release.\n\nI am holding off on a wider announcement until the mirrors have\nupdated.\n\nThanks to everyone who assisted testing, bug reports, success\nstories and most of all, patches :) Special thanks to the OpenBSD\ndevelopers giving us OpenSSH to begin .\n\nRegards,\nDamien Miller\n\n http://violet.ibs.com.au/openssh/files/ChangeLog\n$OpenBSD: releasenotes.html,v 1.97 2026/11/05 15:57:02 naddy Exp $
\n", "date_modified": "2026-11-05T11:40:30+00:00", "date_published": "2026-11-05T11:40:30+00:00", "external_url": "https://www.openssh.org/releasenotes.html#10.6p2", "id": "urn:aggr:item:49714de4a337d9b50e2878468ab1087a6f8624bd", "summary": "OpenSSH 10.6p2 (2026-11-05) Portable OpenSSH 10.6p2 was released on 2026-11-05. It will be available from the mirrors listed at https://www.openssh.com/. OpenSSH is a 100% complete SSH protocol 2.0 implementation and includes sftp client…", "tags": [ "news" ], "title": "OpenSSH 10.6p2 Critical workaround for alpine without PAM", "url": "https://aggr.aymericbeaumet.com/items/hnrss-org/2026-11-05-openssh-10-6p2-critical-workaround-for-alpine-without-pam/" } ], "language": "en", "title": "openssh.org", "version": "https://jsonfeed.org/version/1.1" }