Emmanuel Deloget [Fri, 17 Feb 2017 22:00:48 +0000 (23:00 +0100)]
OpenSSL: don't use direct access to the internal of X509_STORE_CTX
OpenSSL 1.1 does not allow us to directly access the internal of
any data type, including X509_STORE_CTX. We have to use the defined
functions to do so.
Fortunately, these functions have existed since the dawn of time so
we don't have any compatibility issue here.
fix redirect-gateway behaviour when an IPv4 default route does not exist
When no IPv4 default route exists, the "redirect-gateway" routine
aborts even if the sub-option "local" was specified or if we are
connecting to the remote host using IPv6.
This is not expected because in either case OpenVPN should not
bother checking the existence of the default route as it is not
required at all.
Therefore, skip the IPv4 default route check when "local" is
specified or we are connecting to an IPv6 remote host.
Signed-off-by: Antonio Quartulli <a@unstable.cc> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <20170119162518.31752-1-a@unstable.cc>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13905.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
(cherry picked from commit 14670a9d654be48f92b58ac47e6f74d3dcfe1733)
attempt to add IPv6 route even when no IPv6 address was configured
Even if no IPv6 address is configured, OpenVPN still supports
transporting IPv6 segments, therefore adding an IPv6 route
should always be allowed.
However, the route might fail to be installed or may just not work
as expected, therefore, a proper warning should be printed to inform
the user of the possible pitfall.
Always allow adding an IPv6 route and print a WARNING when
no IPv6 address is configured for the interface.
Trac: #832 Signed-off-by: Antonio Quartulli <a@unstable.cc> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <20170131112131.13570-1-a@unstable.cc>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13994.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
(cherry picked from commit 2b7650e7ec9241745e4f66c932d6cffaece927d7)
Ilya Shipitsin [Sun, 29 Jan 2017 06:58:11 +0000 (11:58 +0500)]
Resolve several travis-ci issues
MBEDTLS_VERSION, OPENSSL_VERSION were defined twice - in both
.travis.yml and .travis/build-deps.sh files, the last one
defined OPENSSL_VERSION via nonexistent OPENSSL_VERION
variable, which lead us to use openssl-1.0.1 instead of
openssl-1.0.2, I removed variable definition from build-deps.sh
"cache: [ apt: true ]" is not a travis supported option, it was
introduced by mistake, I removed it
LD_LIBRARY_PATH was defined for the entire test run, it includes
custom openssl build, which was picked by "wget", so "wget"
could not verify SSL cert at https://www.openssl.org sometimes.
We do not want wget to pick our custom LD_LIBRARY_PATH, so I moved
that variable to "script" section
LD_LIBRARY_PATH was defined for both linux and osx environments,
for the second DYLD_LIBRARY_PATH must be defined instead
v2: Upgrade openssl, mbedtls to the most recent versions
v3: DYLD_LIBRARY_PATH was defined via LD_LIBRARY_PATH by mistake
Acked-by: Selva Nair <selva.nair@gmail.com>
Message-Id: <1485673091-7600-1-git-send-email-chipitsine@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13983.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 208c03ea145ed89083c43267733487c99a805069)
Christian Hesse [Tue, 24 Jan 2017 14:39:47 +0000 (15:39 +0100)]
systemd: Do not race on RuntimeDirectory
Different unit instances create and destroy the same RuntimeDirectory.
This leads to running instances where the status file (and possibly
more runtime data) is no longer accessible.
So do not handle this in unit files but provide a tmpfiles.d
configuration and let systemd-tmpfiles do the work.
Nobody will (unintentionally) delete the directories and its content.
As /run is volatile we do not have to care about cleanup.
Signed-off-by: Christian Hesse <mail@eworm.de> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <20170124143947.27385-2-list@eworm.de>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13939.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 3de7be7b17de879a78eea4afe4c918c6104c635d)
Christian Hesse [Tue, 24 Jan 2017 14:39:46 +0000 (15:39 +0100)]
systemd: Use automake tools to install unit files
If systemd is enabled we install unit files to $libdir/systemd/system
(or the path specified by SYSTEMD_UNIT_DIR).
The unit files are generated on the fly with matching $sbindir.
Signed-off-by: Christian Hesse <mail@eworm.de> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <20170124143947.27385-1-list@eworm.de>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13940.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit ca5b4c2aad2370be7862660d274b7485f2d0af71)
David Sommerseth [Tue, 24 Jan 2017 23:23:44 +0000 (00:23 +0100)]
systemd: Move the READY=1 signalling to an earlier point
Currently, OpenVPN will first tell systemd it is ready once the
log will be appended with "Initialization Sequence Completed".
This turns out to cause some issues several places.
First, it adds challenges if --chroot is used in the configuration;
this is already fixed. Secondly, it will cause havoc on static key
p2p mode configurations where the log line above will not happen
before either sides have completed establishing a connection. And
thirdly, if a client configuration fails to establish a connection
within 90 seconds, it will also fail. For the third case this may
not be a critical issue itself, as the host just needs to get
an Internet access established first - which in some scenarios may
take much longer than those 90 seconds systemd grants after the
OpenVPN client configuration is started.
The approach this patch takes is to consider OpenVPN ready when
all the initial preparations and configurations have completed - but
before a connection to a remote side have been attempted. This
also removes the need for specially handling the --chroot scenario.
The final "Initialization Sequence Completed" message update is
kept (though slightly simplified) to indicate we're in a good
state - even though this update will not be visible if --chroot
is used (which was the situation also before this patch).
Trac: #827, #801 Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Gert Doering <gert@greenie.muc.de> Acked-by: Christian Hesse <mail@eworm.de>
Message-Id: <20170124232344.7825-1-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13945.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit e83a8684f0a0d944e9d53cdad2b543cfd1b6fbae)
Steffan Karger [Sun, 22 Jan 2017 16:04:41 +0000 (17:04 +0100)]
Use SHA256 for the internal digest, instead of MD5
Our internal options digest uses MD5 hashes to store the state, instead of
storing the full options string. There's nothing wrong with that, but it
would still be better to use SHA256 because:
* That makes it easier to make OpenVPN "FIPS-compliant" (forbids MD5)
* We don't have to explain anymore that MD5 is fine too
The slightly less bytes for the digest (16 instead of 32) and operations
per connection setup are not worth sticking to MD5.
Note that might SHA256 not be available in de crypto lib, OpenVPN will
refuse to start and shout "Message hash algorithm 'SHA256' not found".
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <1485101081-9784-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13926.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 5b48e8c9f85442936f744c3c550d9d41fe8c7b60)
David Sommerseth [Fri, 20 Jan 2017 21:04:57 +0000 (22:04 +0100)]
git: Merge .gitignore files into a single file
We already track a lot of files over the whole directory structure
in the main .gitignore file. But a few additional ones had been
added into some of the subdirectories.
This unifies all these files into a master file for the whole project,
making it easier to know where to look at and edit if changes needs
to be done.
Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Steffan Karger <steffan@karger.me>
Message-Id: <20170120210457.3383-1-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13916.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit d14b3c60c7796736e07bc3cddb0ab3a58475793e)
Steffan Karger [Sat, 14 Jan 2017 14:10:20 +0000 (15:10 +0100)]
More broadly enforce Allman style and braces-around-conditionals
We want { and } aligned, which means also adding a newline between each
for() and {, while() and {, etc.
Also, we agreed to always use braces with conditionals. The previous
uncrustify config added these for if()s, now also add these for while()
and for().
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1484403020-6857-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13875.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 4cd4899e8e80efae03c584a760fd107251735723)
David Sommerseth [Tue, 10 Jan 2017 20:34:32 +0000 (21:34 +0100)]
management: >REMOTE operation would overwrite ce change indicator
If the management interface on a client received a signal while waiting
for input on the management channel, the "connection entry changed" status
would be overwritten even though nothing was changed. Which could lead
into connecting to the wrong server.
This patch improves this by adding a check if a bool value was changed to
false. This change happens only on signals.
Further, the former 'ret' value have been renamed to 'ce_changed', to
clarify what the expected return value contains. Plus adding some comments
related to this.
And finally do some code style cleanup, breaking up too long lines, adding
some air here and there to improve the readability.
Signed-off-by: David Sommerseth <davids@openvpn.net> Cc: Selva Nair <selva.nair@gmail.com> Acked-by: Selva Nair <selva.nair@gmail.com>
Message-Id: <1484080473-10415-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13851.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit e81f313a71e548638d9e9679226ee84b3b614f13)
Selva Nair [Tue, 3 Jan 2017 20:38:03 +0000 (15:38 -0500)]
Always release dhcp address in close_tun() on Windows.
Also make sure --dhcp-pre-release results in not just dhcp_release()
in open_tun() but a subsequent dhcp_renew() as well. Else dhcp transaction
gets aborted as this call to release() happens after the adapter status
is changed to connected.
Fixes Trac #807 (but can't say the same for Trac #665 without knowing
how to reproduce it)
v2: Mark --dhcp-release as obsolete in manpage and option parser, and
remove the unused dhcp_release variable.
Enforce dhcp-renew with dhcp-pre-release while parsing the option
instead of in open_tun().
Trac: #807
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1483475883-17450-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13814.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
(cherry picked from commit db5b9b45508ea8f66ea80565279af3edd9300499)
Gisle Vanem [Mon, 2 Jan 2017 16:17:51 +0000 (17:17 +0100)]
Crash in options.c
When compiling with --disable-crypto, openvpn would crash on --help as
commit 5d429efd97 introduce and extra %d into the "usage_message" string
but forgot to add it to the #ifndef ENABLE_CRYPTO fprintf() call.
Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <9d41f9dd-a587-5c1e-2e0d-ebb6c921f4ae@yahoo.no>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13808.html
Steffan Karger [Mon, 26 Dec 2016 19:15:43 +0000 (20:15 +0100)]
Textual fixes for Changes.rst
We will likely refer many people to the Changes.rst file once we've
released 2.4. This commits tries to polish the language a bit, and
adds two real changes:
- Remove duplicate mention of the changes --tls-cipher defaults
- Move the 'redirect-gateway' behavioural change from 'features' to
'behavioural changes'.
v2 - On the fly commit changes, based on comments from Selva Nair.
DS also added a few minor corrections on top of that.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <1482779743-9548-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13732.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit f38942d1440575e23d9f8713db435b434381486e)
Steffan Karger [Sun, 25 Dec 2016 22:02:14 +0000 (23:02 +0100)]
man: encourage user to read on about --tls-crypt
As suggested by krzee in trac #790, refer to the --tls-crypt option
form the --tls-auth section of the man page, to encourage users to
check out the --tls-crypt feature.
Trac: #790 Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <1482703334-18949-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13713.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 403dfe1bfdbdf6e5f8abac3401a96852562aec54)
Steffan Karger [Sun, 25 Dec 2016 22:38:25 +0000 (23:38 +0100)]
Document that RSA_SIGN can also request TLS 1.2 signatures
Ever since we support TLS 1.2 (OpenVPN 2.3.3+), the RSA_SIGN might not
only request MD5-SHA1 'TLS signatures', but also other variants.
Document this by updating the implementation hints, and explicitly
stating that we expect a PKCS#1 1.5 signature.
Trac: #764 Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Arne Schwabe <arne@rfc2549.org>
Message-Id: <1482705505-20302-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13714.html Signed-off-by: David Sommerseth <davids@openvpn.net>
(cherry picked from commit 1e36b814073c0f56c77e4922cc105f00b8558e7e)
David Sommerseth [Mon, 19 Dec 2016 18:52:12 +0000 (19:52 +0100)]
dev-tools: Added script for updating copyright years in files
Very simple tool which modifies the Copyright lines in all git checked-in
files with an updated year. Lines only listing a single year (2016) will
be modified to list a range instead.
Only the Copyright lines owners of specific owners will be modified. The
script will need to be slightly updated to cover more owners. See the
UPDATE_COPYRIGHT_LINES line in the script for the currently set owners.
v2 - On-the-fly-commit-update: use vendor/ instead of cmocka and
add @sophos.com to the list of copyright holders to update
Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Steffan Karger <steffan@karger.me>
Message-Id: <1482173532-25132-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13645.html
(cherry picked from commit da8f11f895bb78174d4412d82a6992c398da495a)
Steffan Karger [Thu, 15 Dec 2016 21:46:06 +0000 (22:46 +0100)]
Don't reopen tun if cipher changes
When the pulled options change, OpenVPN will attempt to reopen the tun
device. That might fail if the process has already dropper privileges,
and is not needed unless the tun MTU is changed. This patch therefore
ignores the cipher value for the digest if a fixed tun-mtu is used.
Additionally, this patch changes the md_ctx_update() call to include the
trailing zero byte of each option, to make sure that parsing "foo,bar"
results in a different hash than "foobar". (Sorry for not catching that
during the review...)
The unit tests are a bit lame, but it secretly serves as a way to lower
the bar for adding more buffer.c unit tests.
Trac: #761 Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <1481838366-32335-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13579.html Signed-off-by: David Sommerseth <davids@openvpn.net>
David Sommerseth [Wed, 14 Dec 2016 21:05:00 +0000 (22:05 +0100)]
dev-tools: Add reformat-all.sh for code style unification
This script will run all files related to the currently checked out
git branch through uncrustify using a standardized style configuration.
Due to a bug in uncrustify 0.64, it is needed to add a special treatment
to one of the files at the moment. So this both pre- and post-patched
before/after uncrustify is run. This is to simply to assure that all
file processing will happen consistently each time.
Also added doc/doxygen/doc_key_generation.h to an ignore list, as
it carries some specific Doxygen formatting we should be careful with.
This file is anyhow not so critical and can be managed manually.
The src/compat/compat-lz4.[ch] files are also not touched, as they
are based on upstream formatting. This makes it easier to update
to a newer LZ4 version later on and even see what the differences
are.
v2 - Include updated config from CodeStyle wiki page
Remove line lenght restriction for The Great Reformatting
Update the script with improvements by krzee
v3 - Update with a fixed config from the CodeStyle wiki page
Corrected a typo in the commit message (0.63->0.64)
Minor changes to the reformat script (no pushd/popd,
some new lines moved around, bash->sh)
David Sommerseth [Tue, 13 Dec 2016 12:16:56 +0000 (13:16 +0100)]
Changes.rst: Mainatiner update on C99
Mention for maintainers that we've moved to build with -std=c99 by
default. Also document that 32-bit RHEL5 builds will need -std=gnu99
to be buildable.
Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1481631416-15377-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13518.html Signed-off-by: David Sommerseth <davids@openvpn.net>
systemd: Intermediate --chroot fix with the new sd_notify() implementation
Commit c5931897ae8d663e7e introduced support for talking directly
to the systemd service manager about the situation for the OpenVPN
tunnel. This approach makes a lot of sense and is mostly the proper
way to do it. But it was discovered that it breaks OpenVPN
configurations using --chroot.
The reason sd_notify() calls fails when using chroot() is that
sd_notify() expects to have access to a file as declared in the
$NOTIFY_SOCKET environment variable. It is the main systemd
instance which is responsible to provide both the environment variable
as well as the socket file sd_nodify() should use. When --chroot
comes into play, the $NOTIFY_SOCKET file will not be available
for OpenVPN any more.
As things are getting close to the 2.4_rc2 release we will not dare
to bring a too invasive fix. As well we need some time to discuss
an approrpriate solution. So this intermediate fix will only
provide a "successful start" message to the systemd service manager
right before chroot() happens. This will at least resolve the issue
in a safe and non-intrusive way.
Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Christian Hesse <mail@eworm.de>
Message-Id: <1481079112-22990-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13416.html
Changes: Further improve systemd unit file updates
There were some reports that the directories mentioned should
have trailing /, to make it clearer they are directories and not
files. Also rephrased that sentence slightly to be even clearer
in this aspect.
Signed-off-by: David Sommerseth <davids@openvpn.net>
Magnus Kroken [Fri, 9 Dec 2016 09:07:35 +0000 (10:07 +0100)]
mbedtls: include correct net/net_sockets header according to version
<mbedtls/net.h> is deprecated as of mbedTLS 2.4.0, it is renamed
<mbedtls/net_sockets.h>. OpenVPN will fail to build with
mbedTLS 2.4.0 with MBEDTLS_DEPRECATED_REMOVED defined.
Check MBEDTLS_VERSION_NUMBER, and include net.h for < 2.4.0 and
net_sockets.h for >= 2.4.0.
Signed-off-by: Magnus Kroken <mkroken@gmail.com> Acked-by: Steffan Karger <steffan.karger@fox-it.com>
Message-Id: <1481274455-657-1-git-send-email-mkroken@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13451.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Wed, 7 Dec 2016 19:20:47 +0000 (20:20 +0100)]
Deprecate --no-iv
This fixes the bug of supporting --no-iv (since we're only accepting
bugfixes in the current release phase ;) ).
The --no-iv function decreases security if used (CBC *requires*
unpredictable IVs, other modes don't allow --no-iv at all), and even
marginally decreases other user's security by adding unwanted
complexity to our code.
Let's get rid of this.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de> Acked-by: Arne Schwabe <arne@rfc2549.org>
Message-Id: <1481138447-6292-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13430.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Wed, 7 Dec 2016 18:01:24 +0000 (19:01 +0100)]
Fix (and cleanup) crypto flags in combination with NCP
tls_session_update_crypto_params() did not properly set crypto_flags_or,
but instead set crypto_flags_and twice if a OFB/CFB mode was selected.
Also, the crypto flags in ks->crypto_options.flags were set before
tls_session_update_crypto_params() was called, causing those to not be
adjusted. To fix this, set the crypto flags in
tls_session_generate_data_channel_keys() instead of key_state_init().
While touching that code, remove the to _or and _and variables, which are
not needed at all.
Finally, refuse to accept --no-iv if NCP is enabled (we might otherwise
negotiate invalid combinations and ASSERT out later, and using --no-iv is
a bad idea anyway).
Trac: #784
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1481133684-5325-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13428.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Gert Doering [Tue, 6 Dec 2016 12:26:02 +0000 (13:26 +0100)]
Refactor setting close-on-exec for socket FDs
The existing code can leak socket FDs to the "--up" script, which is
not desired. Brought up by Alberto Gonzalez Iniesta, based on debian
bug 367716.
Since different sockets get create at different times, just moving the
set_cloexec() to link_socket_init_phase1() is not good enough - so move
the call into create_socket_<family>(), so we will catch ALL socket
creations, no matter when or under which conditions they will be
created (SOCKS proxy socket, listening socket, ...).
--inetd gets an extra fd_cloexec() call, as socket FD is inherited.
Christian Hesse [Thu, 1 Dec 2016 21:31:04 +0000 (22:31 +0100)]
Refuse to daemonize when running from systemd
We start with systemd Type=notify, so refuse to daemonize. This does not
affect starting openvpn from script or command line.
v2: Update commit message about script and command line.
Signed-off-by: Christian Hesse <mail@eworm.de> Tested-By: Richard Bonhomme <fragmentux@gmail.com> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <20161201213104.5667-2-list@eworm.de>
URL: http://www.mail-archive.com/search?l=mid&q=20161201213104.5667-2-list@eworm.de Signed-off-by: David Sommerseth <davids@openvpn.net>
Christian Hesse [Thu, 1 Dec 2016 21:31:03 +0000 (22:31 +0100)]
Use systemd service manager notification
Notify systemd service manager when our initialization sequence
completed. This helps ordering services as dependencies can rely on vpn
being available.
v2: Add curly brackets (and indention) to block the else-part, msg()
call was non-conditional before.
v3: Move systemd header include from init.h to init.c.
Signed-off-by: Christian Hesse <mail@eworm.de> Tested-By: Richard Bonhomme <fragmentux@gmail.com> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <20161201213104.5667-1-list@eworm.de>
URL: http://www.mail-archive.com/search?l=mid&q=20161201213104.5667-1-list@eworm.de Signed-off-by: David Sommerseth <davids@openvpn.net>
In order to prevent annoying delays upon client connection,
reload the CRL file only if it was modified since the last
reload operation.
If not, keep on using the already stored CRL.
This change will boost client connection time in instances
where the CRL file is quite large (dropping from several
seconds to few milliseconds).
Cc: Steffan Karger <steffan.karger@fox-it.com> Signed-off-by: Antonio Quartulli <a@unstable.cc> Acked-by: Steffan Karger <steffan.karger@fox-it.com>
Message-Id: <20161201104145.23821-1-a@unstable.cc>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13345.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Selva Nair [Wed, 30 Nov 2016 21:51:36 +0000 (16:51 -0500)]
Do not restart dns client service as a part of --register-dns processing
As reported and discussed on Trac #775, restarting dns service has
unwanted side effects when there are dependent services. And it
appears unnecessary to restart this service to get DNS registered
on Windows.
Resolve by removing two actions from --register-dns:
'net stop dnscache' and 'net start dnscache' run through the service
or directly.
Trac: #775
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480542696-7123-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13331.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Selva Nair [Wed, 30 Nov 2016 00:39:32 +0000 (19:39 -0500)]
Force 'def1' method when --redirect-gateway is done through service
The service deletes all added routes when the client process (openvpn)
exits, causing the re-instated default route to disappear.
Fix by rewriting "--redirect-gateway" to "--redirect-gateway def1" when
routes are set using interactive service.
Only the behaviour on Windows with intereactive service is affected.
Trac: #778
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480466372-2396-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13307.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Selva Nair [Wed, 30 Nov 2016 01:53:14 +0000 (20:53 -0500)]
When parsing '--setenv opt xx ..' make sure a third parameter is present
When no parameters are present, set it to "setenv opt" to trigger a
descriptive error message. And, thus get rid of the pesky NULL pointer
dereferencing.
Trac: #779
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480470794-6349-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13311.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Mon, 28 Nov 2016 22:14:12 +0000 (23:14 +0100)]
Introduce and use secure_memzero() to erase secrets
As described in trac #751, and shortly after reported by Zhaomo Yang, of
the University of California, San Diego, we use memset() (often through
the CLEAR() macro) to erase secrets after use. In some cases however, the
compiler might optimize these calls away.
This patch replaces these memset() calls on secrets by calls to a new
secure_memzero() function, that will not be optimized away.
Since we use CLEAR() a LOT of times, I'm not changing that to use
secure_memzero() to prevent performance impact. I did annotate the macro
to point people at secure_memzero().
This patch also replaces some CLEAR() or memset() calls with a zero-
initialization using "= { 0 }" if that has the same effect, because that
better captures the intend of that code.
Trac: #751
Signed-off-by: Steffan Karger <steffan.karger@fox-it.com> Acked-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480371252-3880-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13278.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Selva Nair [Tue, 29 Nov 2016 02:27:04 +0000 (21:27 -0500)]
Map restart signals from event loop to SIGTERM during exit-notification wait
Commit 63b3e000c9.. fixed SIGTERM getting lost during exit notification
by ignoring any restart signals triggered during this interval. However,
as reported in Trac 777, this could result in repeated triggering of
restart signals when the event loop cannot continue without restart due
to IO errors or timeout.
Avoid by converting soft SIGUSR1 and SIGHUP signals received during
exit-notify wait period to SIGTERM.
Trac #777
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480386424-30876-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13284.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Mon, 28 Nov 2016 14:26:40 +0000 (15:26 +0100)]
Clean up format_hex_ex()
Fix a potential null-pointer dereference, and make the code a bit more
readable while doing so.
The NULL dereference could not be triggered, because the current code
never called format_hex_ex() with maxouput == 0 and separator == NULL.
But it's nicer to not depend on that.
Our use of int vs size_t for lengths needs some attention too, but I'm
not pulling that into this patch. Instead I decided to just make the
(previously existing) assumption that INT_MAX <= SIZE_MAX explicit by
adding a static_assert().
Signed-off-by: Steffan Karger <steffan.karger@fox-it.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480343200-25908-1-git-send-email-steffan.karger@fox-it.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13259.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Mon, 28 Nov 2016 14:53:20 +0000 (15:53 +0100)]
tls_process: don't set variable that's never read
Found by the clang static analyzer: the state_change variable is set,
but never read afterwards. This code has been like this since 2005,
makes sense without setting state_change to true, and has worked fine
for the past 11 years.
Signed-off-by: Steffan Karger <steffan.karger@fox-it.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1480344801-27855-1-git-send-email-steffan.karger@fox-it.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13260.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Wed, 23 Nov 2016 20:02:05 +0000 (21:02 +0100)]
Refactor data channel key generation API
Originally for "poor man's NCP", I introduced a simpler API for generating
data channel keys. That refactoring is no longer needed for that patch,
but I believe still worth a patch on it's own.
This patch should not change any functionality.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479931325-25919-2-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13216.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Wed, 23 Nov 2016 21:21:44 +0000 (22:21 +0100)]
Poor man's NCP for non-NCP peers
Allows non-NCP peers (<= 2.3, or 2.4+ with --ncp-disable) to specify a
--cipher that is different from the one in our config, as long as the new
cipher value is allowed (i.e. in --ncp-ciphers at our side).
This works both client-to-server and server-to-client. I.e. a 2.4 client
with "cipher BF-CBC" and "ncp-ciphers AES-256-GCM:AES-256-CBC" can connect
to both a 2.3 server with "cipher BF-CBC" as well as a server with
"cipher AES-256-CBC" in its config. The other way around, a 2.3 client
with either "cipher BF-CBC" or "cipher AES-256-CBC" can connect to a 2.4
server with e.g. "cipher BF-CBC" and "ncp-ciphers AES-256-GCM:AES-256-CBC"
in its config.
This patch was inspired by Gert's "Poor man's NCP for 2.3 clients" patch,
but takes a different approach to avoid the need for server-side scripts
or client-side 'setenv UV_*' tricks.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479936104-4045-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13218.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
David Sommerseth [Sat, 17 Sep 2016 13:20:15 +0000 (16:20 +0300)]
Document the --auth-token option
This isn't an option to be used directly in any configuration files,
but to be used via --client-connect scripts or --plugin making use of
OPENVPN_PLUGIN_CLIENT_CONNECT or OPENVPN_PLUGIN_CLIENT_CONNECT_V2.
[v2 - Added lacking .B styling of options
- Clarified the token life time ]
Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1474118415-14666-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg12506.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Tue, 22 Nov 2016 20:09:26 +0000 (21:09 +0100)]
generate_key_expansion: make assumption explicit, use C99 features
This function potentially allocates memory, and can therefor not be run
again on an initialized key_ctx_bi. Make this explicit by adding an error
if someone tries do to this anyway.
While touching the function, cleanup it up a bit to make up for the added
lines of code.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479845366-15774-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13202.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Tue, 22 Nov 2016 20:41:26 +0000 (21:41 +0100)]
--tls-crypt fixes
* Check return value of buf_init() (found by coverity)
* Use the TLS frame to determine the buffer size, as is done for the
reliability buffers used for tls-auth. (We previously incorrectly used
the TLS *plaintext* buffer size, which is bigger for typical setups
with tun-mtu <= 1500. Using the frame to calculate the size saves some
bytes for typical setups, and doesn't break setups with big tun-mtu.)
* More carefully handle errors in tls_crypt_wrap() - just drop the packet
instead of ASSERT()ing out (should not happen in the first place, but
this is a bit more friendly if it happens somehow anyway).
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479847286-17518-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13204.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Selva Nair [Tue, 22 Nov 2016 03:12:12 +0000 (22:12 -0500)]
Handle --dhcp-option DNS6 on Windows using netsh
v2: On closing tun delete the ipv6 dns addresses (if any were set).
Also use "validate=no" only in Windows 7 and higher where it is
supported. Its used to skip the time consuming automatic address
validation which is on by default on those platforms.
Tested on Windows Server 2008 (i686), Win 7 (x64) and Win 10 (x64)
TODO: set dns servers using the interactive service
Signed-off-by: Selva Nair <selva.nair@gmail.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479784332-21680-1-git-send-email-selva.nair@gmail.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13193.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Arne Schwabe [Sat, 19 Nov 2016 15:42:44 +0000 (16:42 +0100)]
Fix various compiler warnings
- move p2mp only push_option_fmt to p2mp only section to avoid warning
that struct push_list being defined in the argument list
- incoming_push_message not declared on client without server by putting
it into the right define block
Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479570164-23522-1-git-send-email-arne@rfc2549.org>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13124.html
Arne Schwabe [Sat, 19 Nov 2016 15:35:56 +0000 (16:35 +0100)]
Remove compat-stdbool.h.
Since we use C99, we are guaranteed to have stdbool.h available Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479569756-23302-1-git-send-email-arne@rfc2549.org>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13123.html
Commit c14c4a9e merged the hash_remove() and hash_add() calls in
multi_process_float(), but didn't notice that the hash key (mi->real) was
updated between these calls. So we now try to remove the *new* address
instead of the *old* address from the hash table. This leaks memory and
might break stuff when a different client floats to the old address/port of
this client. Restore that.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479575566-21198-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13128.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
David Sommerseth [Tue, 15 Nov 2016 13:40:56 +0000 (14:40 +0100)]
options: Remove --tls-remote
In OpenVPN 2.3 --tls-remote got deprecated in favour of --verify-x509-name.
The new option solves the same task as --tls-remote but in a more flexible
and improved way. This new option was introduced in commit 9f0fc745664fd0
(release/2.3: f6e12862cefd054eb1). Removing --tls-remote will only require
a minor configuration file change.
The removal of this option has been documented in the man pages since the
release of OpenVPN v2.3, where also the deprecation of --compat-names and
--no-name-remapping was included. However, those two will first be removed
in OpenVPN v2.5.
The reason not to remove --compat-names and --no-name-remapping now is that
such a change will require TLS verification scripts and plug-ins to be
updated to support the new X.509 subject formatting; which
--verify-x509-name already uses.
Signed-off-by: David Sommerseth <davids@openvpn.net> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1479217256-21298-1-git-send-email-davids@openvpn.net>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13070.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Tue, 8 Nov 2016 20:18:22 +0000 (21:18 +0100)]
Add --tls-crypt unit tests
These help verify the tls-crypt functionality - they already caught a
bug during development. We should however probably also add some
t_client tests once this feature is in.
To test --tls-crypt with as few dependencies as possible, this adds a
mock implementation of msg() (or actually x_msg()). For debugging
purposes, the mock implementation can be made to really log by calling
mock_set_debug_level(), but defaults to (almost) no logging.
Signed-off-by: Steffan Karger <steffan.karger@fox-it.com> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1478636302-9678-6-git-send-email-steffan.karger@fox-it.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg12973.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
Steffan Karger [Tue, 15 Nov 2016 13:29:46 +0000 (14:29 +0100)]
Add control channel encryption (--tls-crypt)
This adds a --tls-crypt option, which uses a pre-shared static key (like
the --tls-auth key) to encrypt control channel packets.
Encrypting control channel packets has three main advantages:
* It provides more privacy by hiding the certificate used for the TLS
connection.
* It is harder to identify OpenVPN traffic as such.
* It provides "poor-man's" post-quantum security, against attackers who
will never know the pre-shared key (i.e. no forward secrecy).
Control channel packet encryption
---------------------------------
We propose to use the following encryption method, based on the SIV
construction [0], to achieve nonce misuse-resistant authenticated
encryption:
msg = control channel plaintext
header = opcode (1 byte) || session_id (8 bytes) || packet_id (8
bytes)
Ka = authentication key (256 bits)
Ke = encryption key (256 bits)
(Ka and Ke are pre-shared keys, like with --tls-auth)
auth_tag = HMAC-SHA256(Ka, header || msg)
IV = 128 most-significant bits of auth_tag
ciph = AES256-CTR(Ke, IV, msg)
output = Header || Tag || Ciph
This boils down to the following on-the-wire packet format:
Where
- XXX - means authenticated, and
* XXX * means authenticated and encrypted.
Which is very similar to the current tls-auth packet format, and has the
same overhead as "--tls-auth" with "--auth SHA256".
The use of a nonce misuse-resistant authenticated encryption scheme
allows us to worry less about the risks of nonce collisions. This is
important, because in contrast with the data channel in TLS mode, we
will not be able to rotate tls-crypt keys often or fully guarantee nonce
uniqueness. For non misuse-resistant modes such as GCM [1], [2], the
data channel in TLS mode only has to ensure that the packet counter
never rolls over, while tls-crypt would have to provide nonce uniqueness
over all control channel packets sent by all clients, for the lifetime
of the tls-crypt key.
Unlike with tls-auth, no --key-direction has to be specified for
tls-crypt. TLS servers always use key direction 1, and TLS clients
always use key direction 2, which means that client->server traffic and
server->client traffic always use different keys, without requiring
configuration.
Using fixed, secure, encryption and authentication algorithms makes both
implementation and configuration easier. If we ever want to, we can
extend this to support other crypto primitives. Since tls-crypt should
provide privacy as well as DoS protection, these should not be made
negotiable.
Security considerations:
------------------------
tls-crypt is a best-effort mechanism that aims to provide as much
privacy and security as possible, while staying as simple as possible.
The following are some security considerations for this scheme.
1. The same tls-crypt key is potentially shared by a lot of peers, so it
is quite likely to get compromised. Once an attacker acquires the
tls-crypt key, this mechanism no longer provides any security against
the attacker.
2. Since many peers potentially use the tls-crypt key for a long time, a
lot of data might be encrypted under the tls-crypt key. This leads
to two potential problems:
* The "opcode || session id || packet id" combination might collide.
This might happen in larger setups, because the session id contains
just 64 bits or random. Using the uniqueness requirement from the
GCM spec [3] (a collision probability of less than 2^(-32)),
uniqueness is achieved when using the tls-crypt key for at most
2^16 (65536) connections per process start. (The packet id
includes the daemon start time in the packet ID, which should be
different after stopping and (re)starting OpenPVN.)
And if a collision happens, an attacker can *only* learn whether
colliding packets contain the same plaintext. Attackers will not
be able to learn anything else about the plaintext (unless the
attacker knows the plaintext of one of these packets, of course).
Since the impact is limited, I consider this an acceptable
remaining risk.
* The IVs used in encryption might collide. When two IVs collide, an
attacker can learn the xor of the two plaintexts by xorring the
ciphertexts. This is a serious loss of confidentiality. The IVs
are 128-bit, so when HMAC-SHA256 is a secure PRF (an assumption
that must also hold for TLS), and we use the same uniqueness
requirement from [3], this limits the total amount of control
channel messages for all peers in the setup to 2^48. Assuming a
large setup of 2^16 (65536) clients, and a (conservative) number of
2^16 control channel packets per connection on average, this means
that clients may set up 2^16 connections on average. I think these
numbers are reasonable.
(I have a follow-up proposal to use client-specific tls-auth/tls-crypt
keys to partially mitigate these issues, but let's tackle this patch
first.)
References:
-----------
[0] Rogaway & Shrimpton, A Provable-Security Treatment of the Key-Wrap
Problem, 2006
(https://www.iacr.org/archive/eurocrypt2006/40040377/40040377.pdf)
[1] Ferguson, Authentication weaknesses in GCM, 2005
[3] Dworking, Recommendation for Block Cipher Modes of Operation:
Galois/Counter Mode (GCM) and GMAC, 2007
(http://csrc.nist.gov/publications/nistpubs/800-38D/SP-800-38D.pdf)
Patch history:
--------------
v2 - processed Arne's review comments:
* Error out early with a clear error message when AES-256-CTR or
HMAC-SHA-256 are not supported by the crypto library.
* Clarify that cipher_ctx_reset() sets the IV.
v3 - actually add error messages promised in v2...
Signed-off-by: Steffan Karger <steffan.karger@fox-it.com> Acked-by: Arne Schwabe <arne@rfc2549.org>
Message-Id: <1479216586-20078-1-git-send-email-steffan.karger@fox-it.com>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg13069.html Signed-off-by: Gert Doering <gert@greenie.muc.de>
David Sommerseth [Mon, 14 Nov 2016 22:45:08 +0000 (23:45 +0100)]
file checks: Merge warn_if_group_others_accessible() into check_file_access()
Commit 825e2ec1f358f2e8 cleaned up the usage of
warn_if_group_others_accessible()
and moved it into options.c. At this point there is only one caller of
this
function, check_file_access().
This takes that clean-up one step further and merges everything into
check_file_access(). In addition it removes some no longer needed #ifdefs
and uses platform_stat() to allow a similar check to happen on the Windows
platform as well.
Steffan Karger [Tue, 1 Nov 2016 19:06:47 +0000 (20:06 +0100)]
Restore pre-NCP cipher options on SIGUSR1
As reported by debbie10t on the openvpn-devel list (Message-ID:
<326b8ff7-39a6-1974-c0b0-82fd2abdc7b7@gmail.com>), an NCP client will
attempt to reconnect with the previously pushed cipher, instead of the
cipher from the config file, after a sigusr1 restart. This can be a
problem when the server is reconfigured (as debbie10t explainted), or when
roaming to a differently-configured server. Fix this by restoring the
cipher options from the config file after a sigusr1 restart.
This makes the cipher options behaviour different from other pushable
options, because those are also cached until a sighup restart. We might
want to change this behaviour in general, but for now let's just fix the
issue at hand.
v2: also cache and restore keysize, as that parameter is relevant too.
v3: inherit cached cipher options from parent context.
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: Gert Doering <gert@greenie.muc.de>
Message-Id: <1478027207-28651-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg12869.html Signed-off-by: David Sommerseth <davids@openvpn.net>
Steffan Karger [Tue, 8 Nov 2016 20:07:43 +0000 (21:07 +0100)]
Fix missing return value checks in multi_process_float()
Fix the missing return value checks on hash_remove() and hash_add() by
replacing the calls with an single hash_add() call with the replace
parameters set to true so that is can't fail. Then just ASSERT() that
this is indeed the case.
This also replaces the other add/remove combinations with a single
add-replace, because that should be slightly faster (and this is in the
'hot path').
Signed-off-by: Steffan Karger <steffan@karger.me> Acked-by: David Sommerseth <davids@openvpn.net>
Message-Id: <1478635663-5837-1-git-send-email-steffan@karger.me>
URL: https://www.mail-archive.com/openvpn-devel@lists.sourceforge.net/msg12968.html Signed-off-by: David Sommerseth <davids@openvpn.net>
projects / thirdparty / openvpn.git / log
