android: Retrigger roam events on onLinkPropertiesChanged callbacks
In dual-stack environments the IPv6 connectivity (via autoconfiguration)
might be established before the IPv4 connectivity (via DHCP). It seems
Android triggers the CONNECTIVITY_ACTION broadcast already when the first
family is fully configured. At that time we might not be able to find an
IPv4 source address. And since Android does not trigger the broadcast
again if IPv4 connectivity is established, the connection is broken
afterwards.
Since Android 5.0 ConnectivityManager provides a new callback framework.
There the onLinkPropertiesChanged callback is triggered on the active
interface if e.g. new IP addresses or routes get added. We can use
this to trigger additional roaming events when the IPv4 connectivity is
established later. It might get triggered a bit more often than we
like, and it might get triggered when not needed (e.g. if IPv4
connectivity is established before IPv6).
Tobias Brunner [Mon, 22 Jun 2015 15:15:33 +0000 (17:15 +0200)]
android: Fix kernel-net implementation on Android 4.3 and earlier
Before fwmarks were used protected sockets were bound to the outbound
interface via SO_BINDTODEVICE. This does not always seem to work well
together with our connect()/getsockname() trick if the server is covered
by the traffic selectors. Calling protect() again after disconnecting
the socket seems to help, but if there is no connectivity at all we still
get the virtual IP back (maybe protect() does not bind the socket to any
interface then).
Tobias Brunner [Wed, 17 Jun 2015 15:21:21 +0000 (17:21 +0200)]
android: Add a custom kernel-net implementation to replace kernel-netlink
When roaming from a mobile network to WiFi on Android 5.x the event
received via ConnectivityManager is triggered before the mobile
connection is fully torn down (i.e. before the interface is disabled and
the routes disappear). So for strongSwan the current path still seems
valid and since no roam event is triggered later the daemon never switches
to WiFi and the connection is broken afterwards.
A possible solution to this is enabling roam events in the kernel-netlink
plugin. That would trigger an event when the device is finally disconnected
from the mobile network. However, this could actually take a some time,
during which traffic continues to be sent via mobile network instead of WiFi.
That's because Android now uses multiple routing tables, routing rules and
fwmarks to direct traffic to the appropriate interface/table, but in our
plugin we don't have the information available that would allow us to make
the switch to a different network/routing table earlier (and we actually
prefer the current path if it is still valid). Additionally, the plugin
produces quite a bit more events than ConnectivityManager (which was one
of the reasons to use the latter in the first place).
This custom kernel-net implementation is now specifically tailored for
Android. Roam events are still triggered via ConnectivityManager but
the source address is determined via connect()/getsockname() on a VPN
excluded UDP socket, which does use the correct routing table as intended
by Android. That way the daemon immediately sees a different source IP
when connectivity changes even if the device is connected to multiple
networks concurrently.
Tobias Brunner [Wed, 17 Jun 2015 13:31:24 +0000 (15:31 +0200)]
android: Manually load libraries with dlopen() and RTLD_GLOBAL on Android M
This fixes an issue when using the Android M preview. Bionic's dynamic
linker was changed so that symbols in libraries loaded with RTLD_LOCAL
were not found anymore in dlsym(RTLD_DEFAULT, ...). This is the case
for libraries loaded with System.loadLibrary(), therefore, the plugin
loader in libstrongswan was not able to resolve any symbols defined in
other libraries loaded later. While this seems to have been broken
unintentionally for existing apps (fix at [1]), it will again be a
problem whenever we decide to increase targetSdkVersion beyond 22 (or
until that fix makes it into the system/emulator images).
Unfortunately, the dynamic loader in releases prior to Android 4.3 can't
load libandroidbridge without also loading its dependencies.
Tobias Brunner [Mon, 15 Jun 2015 14:59:12 +0000 (16:59 +0200)]
android: Extend GUI so the MTU can be entered
This also adds a new area for advanced settings that is only displayed
if the user requests it (or if advanced settings already have been set).
The min. MTU for IPv6 is 1280, anything lower lets the TUN device
creation fail if an IPv6 address has been assigned. If lower MTUs are
necessary we might be able to catch that later when setting the MTU and
just use at least 1280 if an IPv6 address was assigned, but let's keep
it simple for now.
Martin Willi [Wed, 3 Jun 2015 08:52:34 +0000 (10:52 +0200)]
ikev2: Enforce remote authentication config before proceeding with own authentication
Previously the constraints in the authentication configuration of an
initiator were enforced only after all authentication rounds were
complete. This posed a problem if an initiator used EAP or PSK
authentication while the responder was authenticated with a certificate
and if a rogue server was able to authenticate itself with a valid
certificate issued by any CA the initiator trusted.
Because any constraints for the responder's identity (rightid) or other
aspects of the authentication (e.g. rightca) the initiator had were not
enforced until the initiator itself finished its authentication such a rogue
responder was able to acquire usernames and password hashes from the client.
And if a client supported EAP-GTC it was even possible to trick it into
sending plaintext passwords.
This patch enforces the configured constraints right after the responder's
authentication successfully finished for each round and before the initiator
starts with its own authentication.
Martin Willi [Mon, 1 Jun 2015 08:10:19 +0000 (10:10 +0200)]
child-sa: Use any fixed reqid configured on the CHILD_SA config
Global reqid allocation (94eb09ac) broke fixed reqid allocation. Resupport them
by bypassing allocation in the kernel if a fixed reqid has been configured.
Martin Willi [Tue, 2 Jun 2015 12:08:42 +0000 (14:08 +0200)]
ivgen: Allow reusing the same message ID twice in sequential IV gen
We use the message ID and fragment number as IV generator. As IKEv2 uses
distinct message ID counters for actively and passively initiated exchanges,
each IV would be used twice. As we explicitly reject such message IDs since d0ed1079, original-responder initiated exchanges fail with counter mode ciphers.
This commit separates IV space in two halves for sequential IVs, and
automatically assigns once reused sequence numbers to the second half.
Tobias Brunner [Fri, 15 May 2015 09:15:57 +0000 (11:15 +0200)]
unknown-payload: Use a new private payload type and make original type available
This fixes a DoS and potential remote code execution vulnerability that was
caused because the original payload type that was returned previously was
used to cast such payload objects to payloads of the indicated type (e.g.
when logging notify payloads with a payload type for the wrong IKE version).
Martin Willi [Fri, 17 Apr 2015 12:19:37 +0000 (14:19 +0200)]
ikev2: Include fragment number into message ID passed to IV gen
When using a cipher with sequential IVs, we never must pass an identical IV
to the IV generator. To avoid it when using IKEv2 fragmentation, use the lower
bits of the 64-bit message ID as the fragment number.
Martin Willi [Wed, 20 May 2015 08:56:23 +0000 (10:56 +0200)]
libtls: As client, reject DH exchanges using primes smaller than 1024 bit
While the server signs the ephemeral DH parameters, it can be tricked to its
lowest supported DH group by a man-in-the-middle:
https://weakdh.org/imperfect-forward-secrecy.pdf
While we at least use 2048-bit DH groups as server, the client accepts any
DH prime the server sends. If it supports export ciphers, only a 512-bit prime
may be used.
As TLS does not define nor negotiate a DH group for cipher suites, the client
actually must accept what the server offers. To avoid downgrades to weak
DH groups, we must reject what we consider insecure. We set this limit to
1024-bit primes. While this breaks compatibility with TLS servers using weaker
primes, this is what we expect servers at least use. Most browser vendors use
the same limit in a similar fix.
Tobias Brunner [Tue, 31 Mar 2015 08:51:22 +0000 (10:51 +0200)]
vici: Explicitly disable --user-install when installing Ruby Gem
Only one of `--user-install` and `--install-dir` may be set and if
`--user-install` is the default on a system installation will fail
unless we disable it explicitly.
Tobias Brunner [Thu, 21 May 2015 09:05:05 +0000 (11:05 +0200)]
vici: Make installation of Ruby Gem and Python Egg optional
Installing them might not work well when building distro packages (e.g.
with DESTDIR installs). It might be easier to install them later with a
script in the distro package.
When building from source on the local system it could still be useful to
install the packages directly, which can be enabled with separate configure
options.
The main problem with DESTDIR installations of the Python Egg is that
easy_install creates or modifies a file called easy-install.pth in the
installation directory. So it's not actually possible to simply copy
the results in DESTDIR over to the actual system as that file would have
to be merged with any existing one.
Tobias Brunner [Tue, 31 Mar 2015 13:40:30 +0000 (15:40 +0200)]
vici: Support out-of-tree build of Python Egg
We also don't require setup.py to exist during cleanup, as e.g. with
make distcheck the source directory is not writable when the build directory
is cleaned, so setup.py can't be created (to just get removed again anyway
if VICI and the Python Eggs haven't been enabled previously).
Tobias Brunner [Tue, 12 May 2015 14:59:20 +0000 (16:59 +0200)]
ikev1: When a reauth is detected explicitly delete the old IKE_SA
Instead of just implicitly destroying the old SA we properly delete it to
notify the other peer (if the other peer keeps the SA up after the
reauthentication and sends DPDs it might consider us dead even though the
new SA is up, that seems to be the case with racoon). We delay the DELETE
a bit to give the other peer time to get the new SA fully established.
Since DELETE messages are not retransmitted it is still possible that
the other peer misses that we deleted the SA.
Tobias Brunner [Wed, 25 Mar 2015 17:05:57 +0000 (18:05 +0100)]
bus: Add new hook called when IKEv1 CHILD_SAs are migrated to a new IKE_SA
The interface is currently not very nice, but if we ever were able to
safely checkout multiple SAs concurrently we could add something similar
to ike_rekey() and call that when we detect a reauthentication.
Tobias Brunner [Tue, 24 Mar 2015 16:38:49 +0000 (17:38 +0100)]
eap-radius: Add cache for usage stats of expired/rekeyed SAs
There are several situations that the previous code didn't handle that
well, for example, interim updates during rekeying (until the rekeyed SA
was deleted the numbers were too high, then suddenly dropped afterwards),
or rekeying for IKEv1 in general because rekeyed IPsec SAs stay installed
until they expire (so if they were still around when the IKE_SA was
terminated, the reported numbers in the Stop message were too high).
If intermediate updates are not used the cache entries for rekeyed
CHILD_SA will accumulate, we can't clean them up as we don't get
child_updown() events for them.
If we sent retransmits for a message and didn't receive a response it might
still arrive later. Such a message will be queued on the socket. The next
read will then return not the expected response but the one for the earlier
request. For this message the verification will fail and the message gets
discarded. But with the earlier code the actual response was never received.
Instead, a subsequent request resulted in the same failure and so on.
Tobias Brunner [Tue, 5 May 2015 16:11:50 +0000 (18:11 +0200)]
starter: Ensure the daemon executable exists when starting up
The only purpose of starter is to control the IKE daemon, so we
terminate it if the daemon executable is not found (e.g. because
DAEMON_NAME is incorrect).
This removes the charonstart setting (it was not actually configurable
anymore).
Tobias Brunner [Tue, 5 May 2015 15:50:27 +0000 (17:50 +0200)]
starter: Remove START_CHARON compile flag
Since the removal of pluto this is quite superfluous. The flag itself
might be useful to avoid starting charon if the executable does not
exist for some reason (e.g. if DAEMON_NAME is incorrect).
Tobias Brunner [Tue, 5 May 2015 07:39:11 +0000 (09:39 +0200)]
ike-init: Fix error handling if nonceg can't be created
Returning FAILED in the constructor is wrong, but returning NULL doesn't work
either as it's currently assumed tasks always can be created.
Therefore, delay this check until we actually try to allocate a nonce.
Martin Willi [Tue, 5 May 2015 08:37:34 +0000 (10:37 +0200)]
swanctl: Fix --uri option
As we now pass the vici connection to the command dispatcher callback, we can't
parse the --uri option to create the connection from the same callback. Instead
pre-process the common command options in a separate loop, and ignore the same
options while processing the actual command.
This allows to control the life-cycle of a nonce in the context of the
ike init task. In the TKM use-case the nonce generator cannot be
destroyed before the ike init task is finalized, otherwise the created
nonce is detected as stale.
child-create: Make nonceg a member of child_create struct
This allows to control the life-cycle of a nonce in the context of the
child create task. In the TKM use-case, it is required to reset the
nonce context if the created nonce is not consumed. This happens if the
child SA negotiation fails and it is detected before the SA is
established via the TKM kernel plugin (i.e. rekey collision).
If the nonce generator detects a stale nonce upon destroy(), it resets
the context in the TKM and releases associated resources in the ID
manager and chunk map.
Also, do not acquire the nonce context ID in tkm_nonceg_create function
but rather when the nonce is actually created by get_nonce().
The nonces created with get_nonce must also be registered in the chunk map.