[clientloop.h ssh.c clientloop.c]
maintain an ordered queue of outstanding global requests that we
expect replies to, similar to the per-channel confirmation queue.
Use this queue to verify success or failure for remote forward
establishment in a race free way.
ok dtucker@
[clientloop.h mux.c channels.c clientloop.c channels.h]
Enable ~ escapes for multiplex slave sessions; give each channel
its own escape state and hook the escape filters up to muxed
channels. bz #1331
Mux slaves do not currently support the ~^Z and ~& escapes.
NB. this change cranks the mux protocol version, so a new ssh
mux client will not be able to connect to a running old ssh
mux master.
ok dtucker@
[key.c]
#define statements that are not atoms need braces around them, else they
will cause trouble in some cases.
Also do a computation of -1 once, and not in a loop several times.
spotted by otto@
[ssh-keygen.c ssh-keygen.1]
ssh-keygen would write fingerprints to STDOUT, and random art to STDERR,
that is not how it was envisioned.
Also correct manpage saying that -v is needed along with -l for it to work.
spotted by naddy@
[ssh_config.5 key.h readconf.c readconf.h ssh-keygen.1 ssh-keygen.c key.c
sshconnect.c]
Introduce SSH Fingerprint ASCII Visualization, a technique inspired by the
graphical hash visualization schemes known as "random art", and by
Dan Kaminsky's musings on the subject during a BlackOp talk at the
23C3 in Berlin.
Scientific publication (original paper):
"Hash Visualization: a New Technique to improve Real-World Security",
Perrig A. and Song D., 1999, International Workshop on Cryptographic
Techniques and E-Commerce (CrypTEC '99)
http://sparrow.ece.cmu.edu/~adrian/projects/validation/validation.pdf
The algorithm used here is a worm crawling over a discrete plane,
leaving a trace (augmenting the field) everywhere it goes.
Movement is taken from dgst_raw 2bit-wise. Bumping into walls
makes the respective movement vector be ignored for this turn,
thus switching to the other color of the chessboard.
Graphs are not unambiguous for now, because circles in graphs can be
walked in either direction.
discussions with several people,
help, corrections and ok markus@ djm@
[auth-options.c match.c servconf.c addrmatch.c sshd.8]
support CIDR address matching in .ssh/authorized_keys from="..." stanzas
ok and extensive testing dtucker@
[PROTOCOL ssh.c serverloop.c]
Add a no-more-sessions@openssh.com global request extension that the
client sends when it knows that it will never request another session
(i.e. when session multiplexing is disabled). This allows a server to
disallow further session requests and terminate the session.
Why would a non-multiplexing client ever issue additional session
requests? It could have been attacked with something like SSH'jack:
http://www.storm.net.nz/projects/7
feedback & ok markus
[sshd.c channels.h channels.c log.c servconf.c log.h servconf.h sshd.8]
Add extended test mode (-T) and connection parameters for test mode (-C).
-T causes sshd to write its effective configuration to stdout and exit.
-C causes any relevant Match rules to be applied before output. The
combination allows tesing of the parser and config files. ok deraadt djm
- djm@cvs.openbsd.org 2008/06/10 03:57:27
[servconf.c match.h sshd_config.5]
support CIDR address matching in sshd_config "Match address" blocks, with
full support for negation and fall-back to classic wildcard matching.
For example:
Match address 192.0.2.0/24,3ffe:ffff::/32,!10.*
PasswordAuthentication yes
addrmatch.c code mostly lifted from flowd's addr.c
feedback and ok dtucker@
[sftp.c sftp-client.c sftp-client.h]
Have the sftp client store the statvfs replies in wire format,
which prevents problems when the server's native sizes exceed the
client's.
Also extends the sizes of the remaining 32bit wire format to 64bit,
they're specified as unsigned long in the standard.
openbsd-compat/Makefile.in openbsd-compat/openbsd-compat.h
openbsd-compat/bsd-statvfs.{c,h}] Add a null implementation of statvfs and
fstatvfs and remove #defines around statvfs code. ok djm@
[clientloop.c]
unbreak tree by committing this bit that I missed from:
Fix sending tty modes when stdin is not a tty (bz#1199). Previously
we would send the modes corresponding to a zeroed struct termios,
whereas we should have been sending an empty list of modes.
Based on patch from daniel.ritz AT alcatel.ch; ok dtucker@ markus@
[ssh-keygen.c]
support -l (print fingerprint) in combination with -F (find host) to
search for a host in ~/.ssh/known_hosts and display its fingerprint;
ok markus@
[sshtty.c ttymodes.c sshpty.h]
Fix sending tty modes when stdin is not a tty (bz#1199). Previously
we would send the modes corresponding to a zeroed struct termios,
whereas we should have been sending an empty list of modes.
Based on patch from daniel.ritz AT alcatel.ch; ok dtucker@ markus@
[openbsd-compat/fmt_scaled.c openbsd-compat/openbsd-compat.h]
Fix compilation on Linux, including pulling in fmt_scaled(3)
implementation from OpenBSD's libutil.
[PROTOCOL]
document our protocol extensions and deviations; ok markus@
- djm@cvs.openbsd.org 2008/05/17 01:31:56
[PROTOCOL]
grammar and correctness fixes from stevesk@
[channels.h clientloop.c nchan.c serverloop.c]
unbreak
ssh -2 localhost od /bin/ls | true
ignoring SIGPIPE by adding a new channel message (EOW) that signals
the peer that we're not interested in any data it might send.
fixes bz #85; discussion, debugging and ok djm@
[channels.c]
error-fd race: don't enable the error fd in the select bitmask
for channels with both in- and output closed, since the channel
will go away before we call select();
report, lots of debugging help and ok djm@
[session.c]
re-add the USE_PIPES code and enable it.
without pipes shutdown-read from the sshd does not trigger
a SIGPIPE when the forked program does a write.
ok djm@
(Id sync only, USE_PIPES never left portable OpenSSH)
[clientloop.c clientloop.h ssh.c mux.c]
tidy up session multiplexing code, moving it into its own file and
making the function names more consistent - making ssh.c and
clientloop.c a fair bit more readable.
ok markus@
[channels.c channels.h clientloop.c serverloop.c]
Try additional addresses when connecting to a port forward destination
whose DNS name resolves to more than one address. The previous behaviour
was to try the first address and give up.
Reported by stig AT venaas.com in bz#343
great feedback and ok markus@
[clientloop.c clientloop.h ssh.c]
Use new channel status confirmation callback system to properly deal
with "important" channel requests that fail, in particular command exec,
shell and subsystem requests. Previously we would optimistically assume
that the requests would always succeed, which could cause hangs if they
did not (e.g. when the server runs out of fds) or were unimplemented by
the server (bz #1384)
Also, properly report failing multiplex channel requests via the mux
client stderr (subject to LogLevel in the mux master) - better than
silently failing.
most bits ok markus@ (as part of a larger diff)
[monitor.c monitor_wrap.c session.h servconf.c servconf.h session.c]
[sshd_config sshd_config.5]
Make the maximum number of sessions run-time controllable via
a sshd_config MaxSessions knob. This is useful for disabling
login/shell/subsystem access while leaving port-forwarding working
(MaxSessions 0), disabling connection multiplexing (MaxSessions 1) or
simply increasing the number of allows multiplexed sessions.
Because some bozos are sure to configure MaxSessions in excess of the
number of available file descriptors in sshd (which, at peak, might be
as many as 9*MaxSessions), audit sshd to ensure that it doesn't leak fds
on error paths, and make it fail gracefully on out-of-fd conditions -
sending channel errors instead of than exiting with fatal().
bz#1090; MaxSessions config bits and manpage from junyer AT gmail.com
ok markus@
[auth-options.c auth1.c channels.c channels.h clientloop.c gss-serv.c]
[monitor.c monitor_wrap.c nchan.c servconf.c serverloop.c session.c]
[ssh.c sshd.c]
Implement a channel success/failure status confirmation callback
mechanism. Each channel maintains a queue of callbacks, which will
be drained in order (RFC4253 guarantees confirm messages are not
reordered within an channel).
Also includes a abandonment callback to clean up if a channel is
closed without sending confirmation messages. This probably
shouldn't happen in compliant implementations, but it could be
abused to leak memory.
ok markus@ (as part of a larger diff)
[bufaux.c buffer.h channels.c packet.c packet.h]
avoid extra malloc/copy/free when receiving data over the net;
~10% speedup for localhost-scp; ok djm@
[servconf.c servconf.h session.c sshd_config.5]
Enable the AllowAgentForwarding option in sshd_config (global and match
context), to specify if agents should be permitted on the server.
As the man page states:
``Note that disabling Agent forwarding does not improve security
unless users are also denied shell access, as they can always install
their own forwarders.''
ok djm@, ok and a mild frown markus@
[sftp-client.c sftp-client.h sftp-server.c sftp.1 sftp.c sftp.h]
introduce sftp extension methods statvfs@openssh.com and
fstatvfs@openssh.com that implement statvfs(2)-like operations,
based on a patch from miklos AT szeredi.hu (bz#1399)
also add a "df" command to the sftp client that uses the
statvfs@openssh.com to produce a df(1)-like display of filesystem
space and inode utilisation
ok markus@
- djm@cvs.openbsd.org 2008/04/13 00:22:17
[dh.c sshd.c]
Use arc4random_buf() when requesting more than a single word of output
Use arc4random_uniform() when the desired random number upper bound
is not a power of two
ok deraadt@ millert@
- djm@cvs.openbsd.org 2008/04/04 05:14:38
[sshd_config.5]
ChrootDirectory is supported in Match blocks (in fact, it is most useful
there). Spotted by Minstrel AT minstrel.org.uk
[monitor_fdpass.c]
msg_controllen has to be CMSG_SPACE so that the kernel can account for
each cmsg_len (ie. msg_controllen = sum of CMSG_ALIGN(cmsg_len). This
works now that kernel fd passing has been fixed to accept a bit of
sloppiness because of this ABI repair.
lots of discussion with kettenis
[sftp-client.c]
prefer POSIX-style file renaming over filexfer rename behaviour if the
server supports the posix-rename@openssh.com extension.
Note that the old (filexfer) behaviour would refuse to clobber an
existing file. Users who depended on this should adjust their sftp(1)
usage.
ok deraadt@ markus@
[monitor_fdpass.c]
Repair the simple cases for msg_controllen where it should just be
CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because
of alignment; ok kettenis hshoexer
[monitor_fdpass.c]
Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due to
an extensive discussion with otto, kettenis, millert, and hshoexer
chroot. Allows ChrootDirectory to work with selinux support compiled in
but not enabled. Using it with selinux enabled will require some selinux
support inside the chroot. "looks sane" djm@
[regress/Makefile regress/test-exec.sh regress/putty-ciphers.sh]
[regress/putty-kex.sh regress/putty-transfer.sh regress/ssh2putty.sh]
basic (crypto, kex and transfer) interop regression tests against putty
To run these, install putty and run "make interop-tests" from the build
directory - the tests aren't run by default yet.
[regress/sftp-cmds.sh]
unbreak lls command and add a regress test that would have caught the
breakage; spotted by mouring@
NB. sftp code change already committed.
[sftp-server.c]
add an extension method "posix-rename@openssh.com" to perform POSIX atomic
rename() operations. based on patch from miklos AT szeredi.hu in bz#1400;
ok dtucker@ markus@
configure (and there's not much point, as openssh won't work without it)
so HAVE_SELECT is not defined and the poll(2) compat code doesn't get
built in. Remove HAVE_SELECT so we can build on platforms without poll.
getgrouplist via getgrset on AIX, rather than iterating over getgrent.
This allows, eg, Match and AllowGroups directives to work with NIS and
LDAP groups.
openbsd-compat/openssl-compat.{c,h}] Bug #1437 Move the OpenSSL compat
header to after OpenSSL headers, since some versions of OpenSSL have
SSLeay_add_all_algorithms as a macro already.
- (tim) [contrib/cygwin/ssh-host-config]
Grammar changes on SYSCONFDIR LOCALSTATEDIR messages.
Check more thoroughly that it's possible to create the /var/empty directory.
Patch by vinschen AT redhat.com
[servconf.c session.c]
delay ~ expansion for ChrootDirectory so it expands to the logged-in user's
home, rather than the user who starts sshd (probably root)
[clientloop.c packet.c serverloop.c]
Revert the change for bz #1307 as it causes connection aborts if an IGNORE
packet arrives while we're waiting in packet_read_expect (and possibly
elsewhere).
[sftp-client.c]
when a remote write error occurs during an upload, ensure that ACKs for
all issued requests are properly drained. patch from t8m AT centrum.cz
[sftp-server.c]
Remove the fixed 100 handle limit in sftp-server and allocate as many
as we have available file descriptors. Patch from miklos AT szeredi.hu;
ok dtucker@ markus@
[sftp.c]
When uploading, correctly handle the case of an unquoted filename with
glob metacharacters that match a file exactly but not as a glob, e.g. a
file called "[abcd]". report and test cases from duncan2nd AT gmx.de
[readconf.c readconf.h sshconnect2.c]
promote rekeylimit to a int64 so it can hold the maximum useful limit
of 2^32; report and patch from Jan.Pechanec AT Sun.COM, ok dtucker@
[channels.c]
When we added support for specified bind addresses for port forwards, we
added a quirk SSH_OLD_FORWARD_ADDR. There is a bug in our handling of
this for -L port forwards that causes the client to listen on both v4
and v6 addresses when connected to a server with this quirk, despite
having set 0.0.0.0 as a bind_address.
report and patch from Jan.Pechanec AT Sun.COM; ok dtucker@
[ssh-keygen.c]
when hashing individual hosts (ssh-keygen -Hf hostname), make sure we
hash just the specified hostname and not the entire hostspec from the
keyfile. It may be of the form "hostname,ipaddr", which would lead to
a hash that never matches. report and fix from jp AT devnull.cz
[ssh.c]
ignore SIGPIPE in multiplex client mode - we can receive this if the
server runs out of fds on us midway. Report and patch from
gregory_shively AT fanniemae.com
[sftp-client.c sftp-client.h]
disable unused functions
initially from tobias@, but disabled them by placing them in
"#ifdef notyet" which was asked by djm@
ok djm@ tobias@
[sshd_config.5 servconf.c]
Allow PermitRootLogin in a Match block. Allows for, eg, permitting root
only from the local network. ok markus@, man page bit ok jmc@
[sshd.c]
When in inetd mode, have sshd generate a Protocol 1 ephemeral server
key only for connections where the client chooses Protocol 1 as opposed
to when it's enabled in the server's config. Speeds up Protocol 2
connections to inetd-mode servers that also allow Protocol 1. bz #440,
based on a patch from bruno at wolff.to, ok markus@
[clientloop.c]
Use the correct packet maximum sizes for remote port and agent forwarding.
Prevents the server from killing the connection if too much data is queued
and an excessively large packet gets sent. bz #1360, ok djm@.
[clientloop.c serverloop.c packet.c]
Make SSH2_MSG_UNIMPLEMENTED and SSH2_MSG_IGNORE messages reset the
ServerAlive and ClientAlive timers. Prevents dropping a connection
when these are enabled but the peer does not support our keepalives.
bz #1307, ok djm@.
[servconf.c canohost.c misc.c channels.c sshconnect.c misc.h ssh-keyscan.c
sshd.c]
Add a small helper function to consistently handle the EAI_SYSTEM error
code of getaddrinfo. Prompted by vgiffin at apple com via bz #1417.
ok markus@ stevesk@
[monitor_wrap.c monitor.c]
Send config block back to slave for invalid users too so options
set by a Match block (eg Banner) behave the same for non-existent
users. Found by and ok djm@
[openbsd-compat/sys-queue.h]
Introduce debugging aid for queue macros. Disabled by default; but
developers are encouraged to run with this enabled.
ok krw@ fgsch@ deraadt@
[openbsd-compat/sys-queue.h]
Performing certain operations on queue.h data structurs produced
funny results. An example is calling LIST_REMOVE on the same
element twice. This will not fail, but result in a data structure
referencing who knows what. Prevent these accidents by NULLing some
fields on remove and replace. This way, either a panic or segfault
will be produced on the faulty operation.
[openbsd-compat/sys-queue.h]
Remove useless ``elm'' argument from the SIMPLEQ_REMOVE_HEAD macro.
This matches our SLIST behaviour and NetBSD's SIMPLEQ as well.
ok millert krw deraadt
[regress/sftp-glob.sh regress/test-exec.sh]
remove "echo -E" crap that I added in last commit and use printf(1) for
cases where we strictly require echo not to reprocess escape characters.
[scp.c]
factor out network read/write into an atomicio()-like function, and
use it to handle short reads, apply bandwidth limits and update
counters. make network IO non-blocking, so a small trickle of
reads/writes has a chance of updating the progress meter; bz #799
ok dtucker@
[sftp.c]
rework argument splitting and parsing to cope correctly with common
shell escapes and make handling of escaped characters consistent
with sh(1) and between sftp commands (especially between ones that
glob their arguments and ones that don't).
parse command flags using getopt(3) rather than hand-rolled parsers.
ok dtucker@
[ssh-keygen.c]
handles zero-sized strings that fgets can return
properly removes trailing newline
removes an unused variable
correctly counts line number
"looks ok" ray@ markus@
[dh.c]
Don't return -1 on error in dh_pub_is_valid(), since it evaluates
to true.
Also fix a typo.
Initial diff from Matthew Dempsky, input from djm.
OK djm, markus.
[auth-bsdauth.c auth-passwd.c auth.c auth.h auth1.c auth2-chall.c]
[monitor.c monitor_wrap.c]
unifdef -DBSD_AUTH
unifdef -USKEY
These options have been in use for some years;
ok markus@ "no objection" millert@
(NB. RCD ID sync only for portable)
[ssh.c sshconnect.c sshconnect.h]
make ssh(1)'s ConnectTimeout option apply to both the TCP connection and
SSH banner exchange (previously it just covered the TCP connection).
This allows callers of ssh(1) to better detect and deal with stuck servers
that accept a TCP connection but don't progress the protocol, and also
makes ConnectTimeout useful for connections via a ProxyCommand;
feedback and "looks ok" markus@
[clientloop.c monitor.c monitor_fdpass.c monitor_fdpass.h]
[monitor_wrap.c ssh.c]
make file descriptor passing code return an error rather than call fatal()
when it encounters problems, and use this to make session multiplexing
masters survive slaves failing to pass all stdio FDs; ok markus@
Darren Tucker
Damien Miller
Tim Rice