[moduli.5]
"Diffie-Hellman" is the usual spelling for the cryptographic protocol
first published by Whitfield Diffie and Martin Hellman in 1976.
ok jmc@
[sandbox-systrace.c]
fail open(2) with EPERM rather than SIGKILLing the whole process. libc
will call open() to do strerror() when NLS is enabled;
feedback and ok markus@
[authfd.c]
bzero the agent address. the kernel was for a while very cranky about
these things. evne though that's fixed, always good to initialize
memory. ok deraadt djm
[sshd.c ssh-sandbox.h sandbox.h sandbox-rlimit.c sandbox-systrace.c]
[sandbox-null.c]
rename sandbox.h => ssh-sandbox.h to make things easier for portable
[channels.c channels.h clientloop.c clientloop.h mux.c ssh.c]
hook up a channel confirm callback to warn the user then requested X11
forwarding was refused by the server; ok markus@
[servconf.c servconf.h sshd.c sshd_config.5 sandbox-rlimit.c]
[sandbox-systrace.c sandbox.h configure.ac Makefile.in]
introduce sandboxing of the pre-auth privsep child using systrace(4).
This introduces a new "UsePrivilegeSeparation=sandbox" option for
sshd_config that applies mandatory restrictions on the syscalls the
privsep child can perform. This prevents a compromised privsep child
from being used to attack other hosts (by opening sockets and proxying)
or probing local kernel attack surface.
The sandbox is implemented using systrace(4) in unsupervised "fast-path"
mode, where a list of permitted syscalls is supplied. Any syscall not
on the list results in SIGKILL being sent to the privsep child. Note
that this requires a kernel with the new SYSTR_POLICY_KILL option.
UsePrivilegeSeparation=sandbox will become the default in the future
so please start testing it now.
feedback dtucker@; ok markus@
[log.c log.h monitor.c monitor.h monitor_wrap.c monitor_wrap.h sshd.c]
make the pre-auth privsep slave log via a socketpair shared with the
monitor rather than /var/empty/dev/log; ok dtucker@ deraadt@ markus@
[authfile.c]
make sure key_parse_public/private_rsa1() no longer consumes its input
buffer. fixes ssh-add for passphrase-protected ssh1-keys;
noted by naddy@; ok djm@
[regress/dynamic-forward.sh]
Retry establishing the port forwarding after a small delay, should make
the tests less flaky when the previous test is slow to shut down and free
up the port.
[ssh-agent.c]
Check current parent process ID against saved one to determine if the parent
has exited, rather than attempting to send a zero signal, since the latter
won't work if the parent has changed privs. bz#1905, patch from Daniel Kahn
Gillmor, ok djm@
[ssh.c]
bz#1883 - setproctitle() to identify mux master; patch from Bert.Wesarg
AT googlemail.com; ok dtucker@
NB. includes additional portability code to enable setproctitle emulation
on platforms that don't support it.
[readconf.c readconf.h ssh.c ssh_config.5 sshconnect.c sshconnect2.c]
Remove undocumented legacy options UserKnownHostsFile2 and
GlobalKnownHostsFile2 by making UserKnownHostsFile/GlobalKnownHostsFile
accept multiple paths per line and making their defaults include
known_hosts2; ok markus
[authfile.c]
read in key comments for v.2 keys (though note that these are not
passed over the agent protocol); bz#439, based on patch from binder
AT arago.de; ok markus@
- djm@cvs.openbsd.org 2011/05/23 03:30:07
[auth-rsa.c auth.c auth.h auth2-pubkey.c monitor.c monitor_wrap.c pathnames.h servconf.c servconf.h sshd.8 sshd_config sshd_config.5]
allow AuthorizedKeysFile to specify multiple files, separated by spaces.
Bring back authorized_keys2 as a default search path (to avoid breaking
existing users of this file), but override this in sshd_config so it will
be no longer used on fresh installs. Maybe in 2015 we can remove it
entierly :)
feedback and ok markus@ dtucker@
[regress/cert-userkey.sh]
fatal() if asked to generate a legacy ECDSA cert (these don't exist)
and fix the regress test that was trying to generate them :)
[monitor.c monitor_wrap.c servconf.c servconf.h]
use a macro to define which string options to copy between configs
for Match. This avoids problems caused by forgetting to keep three
code locations in perfect sync and ordering
"this is at once beautiful and horrible" + ok dtucker@
[servconf.c]
the options TrustedUserCAKeys, RevokedKeysFile, AuthorizedKeysFile
and AuthorizedPrincipalsFile were not being correctly applied in
Match blocks, despite being overridable there; ok dtucker@
options, we should corresponding -W-option when trying to determine
whether it is accepted. Also includes a warning fix on the program
fragment uses (bad main() return type).
bz#1900 and bz#1901 reported by g.esp AT free.fr; ok dtucker@
[auth.c auth.h auth2-pubkey.c pathnames.h servconf.c servconf.h]
remove support for authorized_keys2; it is a relic from the early days
of protocol v.2 support and has been undocumented for many years;
ok markus@
[PROTOCOL.mux clientloop.c clientloop.h mux.c]
improve our behaviour when TTY allocation fails: if we are in
RequestTTY=auto mode (the default), then do not treat at TTY
allocation error as fatal but rather just restore the local TTY
to cooked mode and continue. This is more graceful on devices that
never allocate TTYs.
If RequestTTY is set to "yes" or "force", then failure to allocate
a TTY is fatal.
ok markus@
[clientloop.c mux.c readconf.c readconf.h ssh.c ssh_config.5]
Add a RequestTTY ssh_config option to allow configuration-based
control over tty allocation (like -t/-T); ok markus@
[readconf.c ssh_config.5]
support negated Host matching, e.g.
Host *.example.org !c.example.org
User mekmitasdigoat
Will match "a.example.org", "b.example.org", but not "c.example.org"
ok markus@
[ssh-keygen.c]
certificate options are supposed to be packed in lexical order of
option name (though we don't actually enforce this at present).
Move one up that was out of sequence
[PROTOCOL.mux clientloop.c clientloop.h mux.c ssh.1 ssh.c]
allow graceful shutdown of multiplexing: request that a mux server
removes its listener socket and refuse future multiplexing requests;
ok markus@
[ssh-keygen.1 ssh-keygen.c]
Add -A option. For each of the key types (rsa1, rsa, dsa and ecdsa)
for which host keys do not exist, generate the host keys with the
default key file path, an empty passphrase, default bits for the key
type, and default comment. This will be used by /etc/rc to generate
new host keys. Idea from deraadt.
ok deraadt
[auth.h]
allow GSSAPI authentication to detect when a server-side failure causes
authentication failure and don't count such failures against MaxAuthTries;
bz#1244 from simon AT sxw.org.uk; ok markus@ before lock
- djm@cvs.openbsd.org 2011/03/10 02:52:57
[auth2-gss.c auth2.c]
allow GSSAPI authentication to detect when a server-side failure causes
authentication failure and don't count such failures against MaxAuthTries;
bz#1244 from simon AT sxw.org.uk; ok markus@ before lock
[entropy.c ssh-add.c ssh-agent.c ssh-keygen.c ssh-keyscan.c]
[ssh-keysign.c ssh-pkcs11-helper.c ssh-rand-helper.8 ssh-rand-helper.c]
[ssh.c ssh_prng_cmds.in sshd.c contrib/aix/buildbff.sh]
[regress/README.regress] Remove ssh-rand-helper and all its
tentacles. PRNGd seeding has been rolled into entropy.c directly.
Thanks to tim@ for testing on affected platforms.
Cygwin-specific service installer script ssh-host-config. The actual
functionality is the same, the revisited version is just more
exact when it comes to check for problems which disallow to run
certain aspects of the script. So, part of this script and the also
rearranged service helper script library "csih" is to check if all
the tools required to run the script are available on the system.
The new script also is more thorough to inform the user why the
script failed. Patch from vinschen at redhat com.
- (tim) [configure.ac] Consistent M4 quoting throughout, updated obsolete
AC_TRY_COMPILE with AC_COMPILE_IFELSE, updated obsolete AC_TRY_LINK with
AC_LINK_IFELSE, updated obsolete AC_TRY_RUN with AC_RUN_IFELSE, misc white
space changes for consistency/readability. Makes autoconf 2.68 happy.
"Nice work" djm
openbsd-compat/port-linux.h] Move SELinux-specific code from ssh.c to
port-linux.c to avoid compilation errors. Add -lselinux to ssh when
building with SELinux support to avoid linking failure; report from
amk AT spamfence.net; ok dtucker
RSA_get_default_method() for the benefit of openssl versions that don't
have it (at least openssl-engine-0.9.6b). Found and tested by Kevin Brott,
ok djm@.
[regress/sftp-glob.sh regress/test-exec.sh] Rework how feature tests are
disabled on platforms that do not support them; add a "config_defined()"
shell function that greps for defines in config.h and use them to decide
on feature tests.
Convert a couple of existing grep's over config.h to use the new function
Add a define "FILESYSTEM_NO_BACKSLASH" for filesystem that can't represent
backslash characters in filenames, enable it for Cygwin and use it to turn
of tests for quotes backslashes in sftp-glob.sh.
based on discussion with vinschen AT redhat.com and dtucker@; ok dtucker@
[clientloop.c]
a couple more tweaks to the post-close protocol 1 stderr/stdout flush:
now that we use atomicio(), convert them from while loops to if statements
add test and cast to compile cleanly with -Wsigned
[clientloop.c]
Use atomicio when flushing protocol 1 std{out,err} buffers at
session close. This was a latent bug exposed by setting a SIGCHLD
handler and spotted by kevin.brott AT gmail.com; ok dtucker@
host-key-force target rather than a substitution that is replaced with a
comment so that the Makefile.in is still a syntactically valid Makefile
(useful to run the distprep target)
avoid some integer overflows mostly with GLOB_APPEND and GLOB_DOOFFS
and sanity check arguments (these will be unnecessary when we switch
struct glob members from being type into to size_t in the future);
"looks ok" tedu@ feedback guenther@
[openbsd-compat/glob.c]
Extend GLOB_LIMIT to cover readdir and stat and bump the malloc limit
from ARG_MAX to 64K.
Fixes glob-using programs (notably ftp) able to be triggered to hit
resource limits.
Idea from a similar NetBSD change, original problem reported by jasper@.
ok millert tedu jasper
[clientloop.c ssh-keygen.c sshd.c]
some unsigned long long casts that make things a bit easier for
portable without resorting to dropping PRIu64 formats everywhere
[readpass.c]
fix ControlMaster=ask regression
reset SIGCHLD handler before fork (and restore it after) so we don't miss
the the askpass child's exit status. Correct test for exit status/signal to
account for waitpid() failure; with claudio@ ok claudio@ markus@
[scp.1 scp.c]
add a new -3 option to scp: Copies between two remote hosts are
transferred through the local host. Without this option the data
is copied directly between the two remote hosts. ok djm@ (bugzilla #1837)
formatter if it is present, followed by nroff and groff respectively.
Fixes distprep target on OpenBSD (which has bumped groff/nroff to ports
in favour of mandoc). feedback and ok tim
[sftp-server.c sftp.1 sftp-client.h sftp.c PROTOCOL sftp-client.c]
add a protocol extension to support a hard link operation. It is
available through the "ln" command in the client. The old "ln"
behaviour of creating a symlink is available using its "-s" option
or through the preexisting "symlink" command; based on a patch from
miklos AT szeredi.hu in bz#1555; ok markus@
- djm@cvs.openbsd.org 2010/12/03 23:49:26
[schnorr.c]
check that g^x^q === 1 mod p; recommended by JPAKE author Feng Hao
(this code is still disabled, but apprently people are treating it as
a reference implementation)
[auth.c hostfile.c hostfile.h ssh.c ssh_config.5 sshconnect.c]
[sshconnect.h sshconnect2.c]
automatically order the hostkeys requested by the client based on
which hostkeys are already recorded in known_hosts. This avoids
hostkey warnings when connecting to servers with new ECDSA keys
that are preferred by default; with markus@
[scp.c]
Pass through ssh command-line flags and options when doing remote-remote
transfers, e.g. to enable agent forwarding which is particularly useful
in this case; bz#1837 ok dtucker@
[channels.c]
remove a debug() that pollutes stderr on client connecting to a server
in debug mode (channel_close_fds is called transitively from the session
code post-fork); bz#1719, ok dtucker
[clientloop.c]
avoid NULL deref on receiving a channel request on an unknown or invalid
channel; report bz#1842 from jchadima AT redhat.com; ok dtucker@
[authfile.c]
Refactor internals of private key loading and saving to work on memory
buffers rather than directly on files. This will make a few things
easier to do in the future; ok markus@
