- dtucker@cvs.openbsd.org 2006/07/12 11:34:58
[sshd.c servconf.h servconf.c sshd_config.5 auth.c]
Add support for conditional directives to sshd_config via a "Match"
keyword, which works similarly to the "Host" directive in ssh_config.
Lines after a Match line override the default set in the main section
if the condition on the Match line is true, eg
AllowTcpForwarding yes
Match User anoncvs
AllowTcpForwarding no
will allow port forwarding by all users except "anoncvs".
Currently only a very small subset of directives are supported.
ok djm@
- markus@cvs.openbsd.org 2006/07/11 18:50:48
[clientloop.c ssh.1 ssh.c channels.c ssh_config.5 readconf.h session.c
channels.h readconf.c]
add ExitOnForwardFailure: terminate the connection if ssh(1)
cannot set up all requested dynamic, local, and remote port
forwardings. ok djm, dtucker, stevesk, jmc
- dtucker@cvs.openbsd.org 2006/07/11 10:12:07
[ssh.c]
Only copy the part of environment variable that we actually use. Prevents
ssh bailing when SendEnv is used and an environment variable with a really
long value exists. ok djm@
- stevesk@cvs.openbsd.org 2006/07/10 16:01:57
[sftp-glob.c sftp-common.h sftp.c]
buffer.h only needed in sftp-common.h and remove some unneeded
user includes; ok djm@
- (dtucker) [configure.ac defines.h] Only define SHUT_RD (and friends) and O_NONBLOCK
if they're really needed. Fixes build errors on HP-UX, old Linuxes and probably
more.
- (dtucker) [configure.ac ssh-keygen.c openbsd-compat/bsd-openpty.c
openbsd-compat/daemon.c] Add includes needed by open(2). Conditionally
include paths.h. Fixes build error on Solaris.
- (dtucker) [configure.ac ssh-keygen.c openbsd-compat/bsd-openpty.c
openbsd-compat/daemon.c] Add includes needed by open(2). Conditionally
include paths.h. Fixes build error on Solaris.
- dtucker@cvs.openbsd.org 2006/07/10 12:46:51
[misc.c misc.h sshd.8 sshconnect.c]
Add port identifier to known_hosts for non-default ports, based originally
on a patch from Devin Nate in bz#910.
For any connection using the default port or using a HostKeyAlias the
format is unchanged, otherwise the host name or address is enclosed
within square brackets in the same format as sshd's ListenAddress.
Tested by many, ok markus@.
- djm@cvs.openbsd.org 2006/07/10 12:03:20
[scp.c]
duplicate argv at the start of main() because it gets modified later;
pointed out by deraadt@ ok markus@
- stevesk@cvs.openbsd.org 2006/07/03 17:59:32
[channels.c includes.h]
move #include <arpa/inet.h> out of includes.h; old ok djm@
(portable needed session.c too)
- stevesk@cvs.openbsd.org 2006/07/02 18:36:47
[gss-serv-krb5.c gss-serv.c]
no "servconf.h" needed here
(gss-serv-krb5.c change not applied, portable needs the server options)
- stevesk@cvs.openbsd.org 2006/07/02 17:12:58
[ssh.1 ssh.c ssh_config.5 sshd_config.5]
more details and clarity for tun(4) device forwarding; ok and help
jmc@
- djm@cvs.openbsd.org 2006/06/26 10:36:15
[clientloop.c]
mention optional bind_address in runtime port forwarding setup
command-line help. patch from santhi.amirta AT gmail.com
- (dtucker) [configure.ac] Try AIX blibpath test in different order when
compiling with gcc. gcc 4.1.x will accept (but ignore) -b flags so
configure would not select the correct libpath linker flags.
dtucker [Sat, 24 Jun 2006 02:10:07 +0000 (02:10 +0000)]
- (dtucker) [configure.ac] Bug #1193: Define PASSWD_NEEDS_USERNAME on Solaris.
Works around limitation in Solaris' passwd program for changing passwords
where the username is longer than 8 characters. ok djm@
dtucker [Fri, 23 Jun 2006 11:24:12 +0000 (11:24 +0000)]
- (dtucker) [channels.c configure.ac serverloop.c] Bug #1102: Around AIX
4.3.3 ML3 or so, the AIX pty layer starting passing zero-length writes
on the pty slave as zero-length reads on the pty master, which sshd
interprets as the descriptor closing. Since most things don't do zero
length writes this rarely matters, but occasionally it happens, and when
it does the SSH pty session appears to hang, so we add a special case for
this condition. ok djm@
dtucker [Fri, 23 Jun 2006 11:05:12 +0000 (11:05 +0000)]
- (dtucker) [README.platform configure.ac openbsd-compat/port-tun.c] Add
tunnel support for Mac OS X/Darwin via a third-party tun driver. Patch
from reyk@, tested by anil@
djm [Tue, 13 Jun 2006 03:10:18 +0000 (03:10 +0000)]
- djm@cvs.openbsd.org 2006/06/13 01:18:36
[ssh-agent.c]
always use a format string, even when printing a constant
- djm@cvs.openbsd.org 2006/06/13 02:17:07
[ssh-agent.c]
revert; i am on drugs. spotted by alexander AT beard.se
djm [Tue, 13 Jun 2006 03:01:09 +0000 (03:01 +0000)]
- dtucker@cvs.openbsd.org 2006/05/29 12:56:33
[ssh_config]
Add GSSAPIAuthentication and GSSAPIDelegateCredentials to examples in sample
ssh_config. ok markus@
djm [Tue, 13 Jun 2006 02:59:53 +0000 (02:59 +0000)]
- djm@cvs.openbsd.org 2006/05/08 10:49:48
[sshconnect2.c]
uint32_t -> u_int32_t (which we use everywhere else)
(Id sync only - portable already had this)
dtucker [Sun, 21 May 2006 08:26:40 +0000 (08:26 +0000)]
- (dtucker) [auth.c monitor.c] Now that we don't log from both the monitor
and slave, we can remove the special-case handling in the audit hook in
auth_log.
dtucker [Mon, 15 May 2006 07:15:56 +0000 (07:15 +0000)]
- (dtucker) [openbsd-compat/getrrsetbyname.c] Use _compat_res instead of
_res, prevents problems on some platforms that have _res as a global but
don't have getrrsetbyname(), eg IRIX 5.3. Found and tested by
georg.schwarz at freenet.de, ok djm@.
dtucker [Sat, 6 May 2006 07:41:51 +0000 (07:41 +0000)]
- dtucker@cvs.openbsd.org 2006/04/25 08:02:27
[authfile.c authfile.h sshconnect2.c ssh.c sshconnect1.c]
Prevent ssh from trying to open private keys with bad permissions more than
once or prompting for their passphrases (which it subsequently ignores
anyway), similar to a previous change in ssh-add. bz #1186, ok djm@
dtucker [Wed, 3 May 2006 09:01:09 +0000 (09:01 +0000)]
- (dtucker) [packet.c] Remove in_systm.h since it's also in includes.h
and double including it on IRIX 5.3 causes problems. From Georg Schwarz,
"no objections" tim@
- dtucker@cvs.openbsd.org 2006/04/18 10:44:28
[bufaux.c bufbn.c Makefile.in]
Move Buffer bignum functions into their own file, bufbn.c. This means
that sftp and sftp-server (which use the Buffer functions in bufaux.c
but not the bignum ones) no longer need to be linked with libcrypto.
ok markus@
- djm@cvs.openbsd.org 2006/04/22 04:06:51
[uidswap.c]
use setres[ug]id() to permanently revoke privileges; ok deraadt@
(ID Sync only - portable already uses setres[ug]id() whenever possible)
- djm@cvs.openbsd.org 2006/04/20 21:53:44
[includes.h session.c sftp.c]
Switch from using pipes to socketpairs for communication between
sftp/scp and ssh, and between sshd and its subprocesses. This saves
a file descriptor per session and apparently makes userland ppp over
ssh work; ok markus@ deraadt@ (ID Sync only - portable makes this
decision on a per-platform basis)
- djm@cvs.openbsd.org 2006/04/20 09:27:09
[auth.h clientloop.c dispatch.c dispatch.h kex.h]
replace the last non-sig_atomic_t flag used in a signal handler with a
sig_atomic_t, unfortunately with some knock-on effects in other (non-
signal) contexts in which it is used; ok markus@
- dtucker@cvs.openbsd.org 2006/04/18 10:44:28
[bufaux.c bufbn.c]
Move Buffer bignum functions into their own file, bufbn.c. This means
that sftp and sftp-server (which use the Buffer functions in bufaux.c
but not the bignum ones) no longer need to be linked with libcrypto.
ok markus@
- djm@cvs.openbsd.org 2006/04/16 07:59:00
[atomicio.c]
reorder sanity test so that it cannot dereference past the end of the
iov array; well spotted canacar@!
- djm@cvs.openbsd.org 2006/04/16 00:54:10
[sftp-client.c]
avoid making a tiny 4-byte write to send the packet length of sftp
commands, which would result in a separate tiny packet on the wire by
using atomiciov(writev, ...) to write the length and the command in one
pass; ok deraadt@
- djm@cvs.openbsd.org 2006/04/16 00:52:55
[atomicio.c atomicio.h]
introduce atomiciov() function that wraps readv/writev to retry
interrupted transfers like atomicio() does for read/write;
feedback deraadt@ dtucker@ stevesk@ ok deraadt@
- djm@cvs.openbsd.org 2006/04/16 00:48:52
[buffer.c buffer.h channels.c]
Fix condition where we could exit with a fatal error when an input
buffer became too large and the remote end had advertised a big window.
The problem was a mismatch in the backoff math between the channels code
and the buffer code, so make a buffer_check_alloc() function that the
channels code can use to propsectivly check whether an incremental
allocation will succeed. bz #1131, debugged with the assistance of
cove AT wildpackets.com; ok dtucker@ deraadt@