strongswan.git
6 years agoMerge branch 'dnscert'
Tobias Brunner [Fri, 11 Oct 2013 13:46:09 +0000 (15:46 +0200)]
Merge branch 'dnscert'

The new dnscert plugin adds support for authentication via CERT resource
records that are protected with DNSSEC.

6 years agotesting: Add ikev2/net2net-dnscert scenario
Tobias Brunner [Thu, 26 Sep 2013 16:28:48 +0000 (18:28 +0200)]
testing: Add ikev2/net2net-dnscert scenario

6 years agotesting: Provide moon's and sun's certificate as CERT RR
Tobias Brunner [Thu, 26 Sep 2013 16:16:10 +0000 (18:16 +0200)]
testing: Provide moon's and sun's certificate as CERT RR

6 years agotesting: Enable dnscert plugin
Tobias Brunner [Thu, 26 Sep 2013 15:01:11 +0000 (17:01 +0200)]
testing: Enable dnscert plugin

6 years agotesting: Load testing.conf.local from the same directory as testing.conf
Tobias Brunner [Thu, 26 Sep 2013 15:00:21 +0000 (17:00 +0200)]
testing: Load testing.conf.local from the same directory as testing.conf

6 years agodnscert: Add DNS CERT support for pubkey authentication
Ruslan N. Marchenko [Fri, 30 Aug 2013 15:51:12 +0000 (17:51 +0200)]
dnscert: Add DNS CERT support for pubkey authentication

Add DNSSEC protected CERT RR delivered certificate authentication.
The new dnscert plugin is based on the ipseckey plugin and relies on the
existing PEM decoder as well as x509 and PGP parsers.  As such the plugin
expects PEM encoded PKIX(x509) or PGP(GPG) certificate payloads.

The plugin is targeted to improve interoperability with Racoon, which
supports this type of authentication, ignoring in-stream certificates
and using only DNS provided certificates for FQDN IDs.

6 years agoipseckey: Properly handle failure to create a certificate
Tobias Brunner [Thu, 29 Aug 2013 13:58:48 +0000 (15:58 +0200)]
ipseckey: Properly handle failure to create a certificate

Also, try the next key (if available) if parsing an IPSECKEY failed.

6 years agoipseckey: Refactor creation of certificate enumerator
Tobias Brunner [Thu, 29 Aug 2013 13:47:05 +0000 (15:47 +0200)]
ipseckey: Refactor creation of certificate enumerator

Reduces nesting and fixes a memory leak (rrsig_enum).

6 years agoipseckey: Depend on plugin features to create public key and certificate objects
Tobias Brunner [Thu, 29 Aug 2013 13:25:23 +0000 (15:25 +0200)]
ipseckey: Depend on plugin features to create public key and certificate objects

6 years agounbound: Add support for DLV (DNSSEC Lookaside Validation)
Tobias Brunner [Thu, 29 Aug 2013 07:04:36 +0000 (09:04 +0200)]
unbound: Add support for DLV (DNSSEC Lookaside Validation)

Fixes #392.

6 years agoMerge branch 'fwmarks'
Tobias Brunner [Fri, 11 Oct 2013 13:33:06 +0000 (15:33 +0200)]
Merge branch 'fwmarks'

Allows setting a mark on outbound packets and the routing rule
installed by charon.  With those settings it is possible to setup
tunnels with kernel-libipsec where the remote peer is part of the remote
traffic selector.

The following example settings in strongswan.conf show how this can be
configured:

charon {
    plugins {
        kernel-netlink {
            fwmark = !0x42
        }
        socket-default {
            fwmark = 0x42
        }
        kernel-libipsec {
            allow_peer_ts = yes
        }
    }
}

To make it work it is necessary to set

  net.ipv4.conf.all.rp_filter

appropriately, otherwise the kernel drops the packets.

References #380.

6 years agokernel-libipsec: Don't ignore policies of type != POLICY_IPSEC
Tobias Brunner [Thu, 10 Oct 2013 13:41:29 +0000 (15:41 +0200)]
kernel-libipsec: Don't ignore policies of type != POLICY_IPSEC

This actually broke rekeying due to the DROP policies that are
temporarily added, which broke the refcount as the ignored policies
were not ignored in del_policy() (the type is not known there).

6 years agokernel-libipsec: Add an option to allow remote TS to match the IKE peer
Tobias Brunner [Tue, 13 Aug 2013 15:10:00 +0000 (17:10 +0200)]
kernel-libipsec: Add an option to allow remote TS to match the IKE peer

Setting the fwmark options for the kernel-netlink and socket-default
plugins allow this kind of setup.

It is probably required to set net.ipv4.conf.all.rp_filter to 2 to make
it work.

6 years agosocket-default: Allow setting firewall mark on outbound packets
Tobias Brunner [Tue, 13 Aug 2013 14:58:33 +0000 (16:58 +0200)]
socket-default: Allow setting firewall mark on outbound packets

6 years agokernel-netlink: Allow setting firewall marks on routing rule
Tobias Brunner [Tue, 13 Aug 2013 14:53:06 +0000 (16:53 +0200)]
kernel-netlink: Allow setting firewall marks on routing rule

6 years agoipsec_types: Add utility function to parse mark_t from strings
Tobias Brunner [Tue, 13 Aug 2013 13:15:45 +0000 (15:15 +0200)]
ipsec_types: Add utility function to parse mark_t from strings

6 years agoMerge branch 'database-transactions'
Tobias Brunner [Fri, 11 Oct 2013 13:29:30 +0000 (15:29 +0200)]
Merge branch 'database-transactions'

This adds support for transactions to the database_t interface and the two
current implementations.

The pool utility is also moved to its own directory in src/.

6 years agoattr-sql: Use a serializable transaction when inserting identities
Tobias Brunner [Thu, 10 Oct 2013 09:02:16 +0000 (11:02 +0200)]
attr-sql: Use a serializable transaction when inserting identities

6 years agodatabase: Add support for serializable transactions
Tobias Brunner [Thu, 10 Oct 2013 08:58:40 +0000 (10:58 +0200)]
database: Add support for serializable transactions

6 years agosql: Don't use MyISAM engine and set collation/charset for all tables
Tobias Brunner [Fri, 6 Sep 2013 12:09:32 +0000 (14:09 +0200)]
sql: Don't use MyISAM engine and set collation/charset for all tables

The MyISAM engine doesn't support transactions.

6 years agopool: Change transaction handling
Tobias Brunner [Fri, 6 Sep 2013 09:29:17 +0000 (11:29 +0200)]
pool: Change transaction handling

6 years agopool: Move the pool utility to its own directory in src
Tobias Brunner [Thu, 5 Sep 2013 16:00:48 +0000 (18:00 +0200)]
pool: Move the pool utility to its own directory in src

6 years agoattr-sql: Handle concurrent insertion of identities
Tobias Brunner [Fri, 13 Sep 2013 11:25:49 +0000 (13:25 +0200)]
attr-sql: Handle concurrent insertion of identities

If the same identity is added concurrently by two threads (or by the
pool utility) INSERT might fail even though the SELECT was unsuccessful
before.

We are currently not able to lock the identities table in a portable way
(something like SELECT ... FOR UPDATE on MySQL).

6 years agoattr-sql: Don't use database transactions in create_attribute_enumerator
Tobias Brunner [Thu, 5 Sep 2013 15:03:11 +0000 (17:03 +0200)]
attr-sql: Don't use database transactions in create_attribute_enumerator

There could, of course, be race conditions when enumerating the attributes,
but those probably don't matter (e.g. missing an attribute that was
concurrently added).

Transactions are more intended to revert multiple changes if anything
fails in the process.

6 years agosqlite: Implement transaction handling
Tobias Brunner [Thu, 5 Sep 2013 14:50:23 +0000 (16:50 +0200)]
sqlite: Implement transaction handling

6 years agomysql: Implement transaction handling
Tobias Brunner [Thu, 5 Sep 2013 14:46:24 +0000 (16:46 +0200)]
mysql: Implement transaction handling

6 years agodatabase: Add interface to handle transactions
Tobias Brunner [Fri, 6 Sep 2013 06:16:39 +0000 (08:16 +0200)]
database: Add interface to handle transactions

6 years agomysql: Ensure connections are properly released in multi-threaded environments
Tobias Brunner [Thu, 5 Sep 2013 13:33:24 +0000 (15:33 +0200)]
mysql: Ensure connections are properly released in multi-threaded environments

6 years agocrypto-factory: Try next available RNG implementation if constructor fails
Tobias Brunner [Thu, 3 Oct 2013 08:24:59 +0000 (10:24 +0200)]
crypto-factory: Try next available RNG implementation if constructor fails

6 years agocrypto-factory: Order entries by algorithm identifier and (optionally) speed
Tobias Brunner [Thu, 3 Oct 2013 08:23:30 +0000 (10:23 +0200)]
crypto-factory: Order entries by algorithm identifier and (optionally) speed

6 years agoRemove HASH_PREFERRED, usages are replaced with HASH_SHA1, which is required for...
Tobias Brunner [Thu, 3 Oct 2013 08:14:49 +0000 (10:14 +0200)]
Remove HASH_PREFERRED, usages are replaced with HASH_SHA1, which is required for IKEv2 anyway

6 years agovstr: Forward actual field width
Tobias Brunner [Fri, 11 Oct 2013 11:57:05 +0000 (13:57 +0200)]
vstr: Forward actual field width

fmt_field_width is a flag that indicates if a field width
is defined in obj_field_width.

6 years agounit-tests: support testing when leak-detective has not been enabled
Martin Willi [Tue, 25 Jun 2013 15:09:07 +0000 (17:09 +0200)]
unit-tests: support testing when leak-detective has not been enabled

6 years agoNEWS: Updates for the ah, libipsec-usestats and printf-hook merges
Martin Willi [Fri, 11 Oct 2013 09:40:02 +0000 (11:40 +0200)]
NEWS: Updates for the ah, libipsec-usestats and printf-hook merges

6 years agoMerge branch 'printf-hook'
Martin Willi [Fri, 11 Oct 2013 09:12:38 +0000 (11:12 +0200)]
Merge branch 'printf-hook'

Adds a custom printf hook implementation as a fallback if neither the glibc
style hooks nor vstr is available. This can avoid the Vstr dependency on some
systems at the cost of slower and less complete printf functions.

6 years agoprintf-hook-builtin: Print NaN/Infinity floating point values as such
Martin Willi [Fri, 27 Sep 2013 16:16:46 +0000 (18:16 +0200)]
printf-hook-builtin: Print NaN/Infinity floating point values as such

6 years agoprintf-hook-builtin: Correctly round up floating point values
Martin Willi [Fri, 27 Sep 2013 14:13:14 +0000 (16:13 +0200)]
printf-hook-builtin: Correctly round up floating point values

6 years agoprintf-hook-builtin: Add some preliminary floating point support
Martin Willi [Fri, 11 Oct 2013 08:55:05 +0000 (10:55 +0200)]
printf-hook-builtin: Add some preliminary floating point support

This minimalistic implementation has no aspiration for completeness or
accuracy, and just provides what we need.

6 years agoprintf-hook-builtin: Support GNU %m specifier
Martin Willi [Fri, 27 Sep 2013 09:16:11 +0000 (11:16 +0200)]
printf-hook-builtin: Support GNU %m specifier

6 years agoprintf-hook-builtin: Add a new "builtin" backend using its own printf() routines
Martin Willi [Fri, 11 Oct 2013 09:06:02 +0000 (11:06 +0200)]
printf-hook-builtin: Add a new "builtin" backend using its own printf() routines

Overloads printf C library functions by a self-contained implementation,
based on klibc. Does not yet feature all the required default formatters,
including those for floating point values.

6 years agoprintf-hook: Add some basic printf() string/integer test functions
Martin Willi [Fri, 27 Sep 2013 10:19:11 +0000 (12:19 +0200)]
printf-hook: Add some basic printf() string/integer test functions

6 years agoprintf-hook: Move glibc/vstr printf hook backends to separate files
Martin Willi [Fri, 27 Sep 2013 15:30:17 +0000 (17:30 +0200)]
printf-hook: Move glibc/vstr printf hook backends to separate files

6 years agoMerge branch 'libipsec-usestats'
Martin Willi [Fri, 11 Oct 2013 08:24:27 +0000 (10:24 +0200)]
Merge branch 'libipsec-usestats'

Brings SA usage statistics and volume based expiration to libipsec and the
associated kernel-libipsec plugin. Additionally removes any ESPv3 style TFC
padding found in incoming packets.

6 years agolibipsec: Enforce byte/packet lifetimes on SAs
Martin Willi [Mon, 30 Sep 2013 13:47:27 +0000 (15:47 +0200)]
libipsec: Enforce byte/packet lifetimes on SAs

6 years agokernel-libipsec: Support ESPv3 TFC padding
Martin Willi [Mon, 23 Sep 2013 14:26:11 +0000 (16:26 +0200)]
kernel-libipsec: Support ESPv3 TFC padding

6 years agolibipsec: remove extra RFC4303 TFC padding appended to inner payload
Martin Willi [Mon, 23 Sep 2013 14:23:54 +0000 (16:23 +0200)]
libipsec: remove extra RFC4303 TFC padding appended to inner payload

6 years agokernel-libipsec: Support query_sa() to report usage statistics
Martin Willi [Mon, 23 Sep 2013 10:46:43 +0000 (12:46 +0200)]
kernel-libipsec: Support query_sa() to report usage statistics

6 years agolibipsec: Support usage statistics and query_sa() on IPsec SAs
Martin Willi [Mon, 23 Sep 2013 10:10:07 +0000 (12:10 +0200)]
libipsec: Support usage statistics and query_sa() on IPsec SAs

6 years agokernel: Use a time_t to report use time in query_policy()
Martin Willi [Mon, 23 Sep 2013 10:35:33 +0000 (12:35 +0200)]
kernel: Use a time_t to report use time in query_policy()

6 years agokernel: Use a time_t to report use time in query_sa()
Martin Willi [Mon, 23 Sep 2013 10:28:13 +0000 (12:28 +0200)]
kernel: Use a time_t to report use time in query_sa()

6 years agoMerge branch 'ah'
Martin Willi [Fri, 11 Oct 2013 08:15:43 +0000 (10:15 +0200)]
Merge branch 'ah'

Brings support for Security Associations integrity protected by the
Authentication Header protocol, both to IKEv1 and IKEv2. Currently only plain
AH is supported, but no (now deprecated) RFC2401 style AH+ESP bundles.

6 years agoipsec.conf: Add a description for the new 'ah' keyword.
Martin Willi [Thu, 10 Oct 2013 16:09:57 +0000 (18:09 +0200)]
ipsec.conf: Add a description for the new 'ah' keyword.

6 years agotesting: Add an IKEv1 host2host AH transport mode test case
Martin Willi [Wed, 9 Oct 2013 14:10:33 +0000 (16:10 +0200)]
testing: Add an IKEv1 host2host AH transport mode test case

6 years agotesting: Add an IKEv1 net2net AH test case
Martin Willi [Wed, 9 Oct 2013 14:10:08 +0000 (16:10 +0200)]
testing: Add an IKEv1 net2net AH test case

6 years agotesting: Add an IKEv2 host2host AH transport mode test case
Martin Willi [Wed, 9 Oct 2013 13:20:22 +0000 (15:20 +0200)]
testing: Add an IKEv2 host2host AH transport mode test case

6 years agotesting: Add an IKEv2 net2net AH test case
Martin Willi [Wed, 9 Oct 2013 13:10:40 +0000 (15:10 +0200)]
testing: Add an IKEv2 net2net AH test case

6 years agotesting: Allow AH packets in default INPUT/OUTPUT chains
Martin Willi [Wed, 9 Oct 2013 13:05:46 +0000 (15:05 +0200)]
testing: Allow AH packets in default INPUT/OUTPUT chains

6 years agoupdown: Install forwarding rules with the actually used protocol
Martin Willi [Wed, 9 Oct 2013 12:48:50 +0000 (14:48 +0200)]
updown: Install forwarding rules with the actually used protocol

6 years agoupdown: Add a PLUTO_PROTO variable set to 'ah' or 'esp'
Martin Willi [Wed, 9 Oct 2013 12:48:25 +0000 (14:48 +0200)]
updown: Add a PLUTO_PROTO variable set to 'ah' or 'esp'

6 years agostarter: Reject connections having both 'ah' and 'esp' keywords set
Martin Willi [Wed, 9 Oct 2013 12:09:08 +0000 (14:09 +0200)]
starter: Reject connections having both 'ah' and 'esp' keywords set

We currently don't support mixed proposals or bundles, so don't create the
illusion we would.

6 years agoike: Define keylength for aescmac algorithm
Martin Willi [Fri, 21 Jun 2013 14:01:03 +0000 (16:01 +0200)]
ike: Define keylength for aescmac algorithm

6 years agoikev1: Support parsing of AH+IPComp proposals
Martin Willi [Fri, 21 Jun 2013 14:00:22 +0000 (16:00 +0200)]
ikev1: Support parsing of AH+IPComp proposals

6 years agostarter: Remove obsolete 'auth' option
Martin Willi [Thu, 20 Jun 2013 15:10:13 +0000 (17:10 +0200)]
starter: Remove obsolete 'auth' option

6 years agoikev1: Accept more than two certificate payloads
Martin Willi [Thu, 20 Jun 2013 15:07:27 +0000 (17:07 +0200)]
ikev1: Accept more than two certificate payloads

6 years agoikev1: Support en-/decoding of SA payloads with AH algorithms
Martin Willi [Thu, 20 Jun 2013 15:06:46 +0000 (17:06 +0200)]
ikev1: Support en-/decoding of SA payloads with AH algorithms

6 years agokernel-handler: Whitespace cleanups
Martin Willi [Thu, 20 Jun 2013 14:16:39 +0000 (16:16 +0200)]
kernel-handler: Whitespace cleanups

6 years agostroke: List proposals in statusall without leading '/' in AH SAs
Martin Willi [Thu, 20 Jun 2013 14:16:06 +0000 (16:16 +0200)]
stroke: List proposals in statusall without leading '/' in AH SAs

6 years agoikev1: Delete quick modes with the negotiated SA protocol
Martin Willi [Thu, 20 Jun 2013 14:15:31 +0000 (16:15 +0200)]
ikev1: Delete quick modes with the negotiated SA protocol

6 years agotrap-manager: Install trap with SA protocol of the first configured proposal
Martin Willi [Thu, 20 Jun 2013 14:14:52 +0000 (16:14 +0200)]
trap-manager: Install trap with SA protocol of the first configured proposal

6 years agochild-sa: Save protocol during SPI allocation
Martin Willi [Thu, 20 Jun 2013 14:13:35 +0000 (16:13 +0200)]
child-sa: Save protocol during SPI allocation

This allows us to properly delete the incomplete SA with the correct protocol
should negotiation fail.

6 years agoikev1: Negotiate SPI with the first/negotiated proposal protocol
Martin Willi [Thu, 20 Jun 2013 14:12:58 +0000 (16:12 +0200)]
ikev1: Negotiate SPI with the first/negotiated proposal protocol

6 years agoikev2: Allocate SPI with the protocol of the first/negotiated proposal
Martin Willi [Thu, 20 Jun 2013 14:12:14 +0000 (16:12 +0200)]
ikev2: Allocate SPI with the protocol of the first/negotiated proposal

6 years agoproposal: Strip redundant integrity algos for ESP proposals only
Martin Willi [Thu, 20 Jun 2013 14:10:55 +0000 (16:10 +0200)]
proposal: Strip redundant integrity algos for ESP proposals only

6 years agostroke: Configure proposal with AH protocol if 'ah' option set
Martin Willi [Thu, 20 Jun 2013 14:09:51 +0000 (16:09 +0200)]
stroke: Configure proposal with AH protocol if 'ah' option set

6 years agostarter: Add an 'ah' keyword for Authentication Header Security Associations
Martin Willi [Thu, 20 Jun 2013 14:08:23 +0000 (16:08 +0200)]
starter: Add an 'ah' keyword for Authentication Header Security Associations

6 years agoVersion bump to 5.1.1rc1
Andreas Steffen [Fri, 11 Oct 2013 07:53:42 +0000 (09:53 +0200)]
Version bump to 5.1.1rc1

6 years agoKeep a copy of the tnccs instance for PT-TLS handover
Andreas Steffen [Wed, 9 Oct 2013 17:03:07 +0000 (19:03 +0200)]
Keep a copy of the tnccs instance for PT-TLS handover

6 years agoxauth-pam: Make trimming of email addresses optional 5.1.1dr4
Tobias Brunner [Fri, 4 Oct 2013 08:49:54 +0000 (10:49 +0200)]
xauth-pam: Make trimming of email addresses optional

Fixes #430.

6 years agoikev1: Accept reauthentication attempts with a keep unique policy from same host
Martin Willi [Wed, 18 Sep 2013 12:11:40 +0000 (14:11 +0200)]
ikev1: Accept reauthentication attempts with a keep unique policy from same host

When we have a "keep" unique policy in place, we have to be less strict in
rejecting Main/Aggressive Modes to enforce it. If the host/port equals to
that of an existing ISAKMP SA, we assume it is a reauthentication attempt
and accept the new SA (to replace the old).

6 years agoikev1: Don't log a reauthentication detection message if no children adopted
Martin Willi [Wed, 18 Sep 2013 11:59:44 +0000 (13:59 +0200)]
ikev1: Don't log a reauthentication detection message if no children adopted

When a replace unique policy is in place, the children get adopted during
the uniqueness check. In this case the message is just misleading.

6 years agoikev1: Delay a potential delete for a duplicate IKE_SA having a replace policy
Martin Willi [Wed, 18 Sep 2013 11:56:45 +0000 (13:56 +0200)]
ikev1: Delay a potential delete for a duplicate IKE_SA having a replace policy

Sending a DELETE for the replaced SA immediately is problematic during
reauthentication, as the peer might have associated the Quick Modes to the
old SA, and also delete them.

With this change the delete for the old ISAKMP SA is usually omitted, as it
is gets implicitly deleted by the reauth.

6 years agoeap-radius: Increase buffer for attributes sent in RADIUS accounting messages
Tobias Brunner [Tue, 10 Sep 2013 14:51:20 +0000 (16:51 +0200)]
eap-radius: Increase buffer for attributes sent in RADIUS accounting messages

64 bytes might be too short for user names/identities.

6 years agoopenssl: Properly log FIPS mode when enabled via openssl.conf
Tobias Brunner [Fri, 27 Sep 2013 07:11:55 +0000 (09:11 +0200)]
openssl: Properly log FIPS mode when enabled via openssl.conf

Enabling FIPS mode twice will fail, so if it is enabled in openssl.conf
it should be disabled in strongswan.conf (or the other way around).

Either way, we should log whether FIPS mode is enabled or not.

References #412.

6 years agoandroid: New release after fixing remediation instructions regression
Tobias Brunner [Thu, 26 Sep 2013 11:03:38 +0000 (13:03 +0200)]
android: New release after fixing remediation instructions regression

6 years agoandroid: Change progress dialog handling
Tobias Brunner [Thu, 26 Sep 2013 11:50:23 +0000 (13:50 +0200)]
android: Change progress dialog handling

With the previous code the dialog sometimes was hidden for a short while
before it got reopened.

6 years agoandroid: Clear remediation instructions when starting a new connection
Tobias Brunner [Thu, 26 Sep 2013 11:00:45 +0000 (13:00 +0200)]
android: Clear remediation instructions when starting a new connection

6 years agostarter: Don't ignore keyingtries with rekey=no
Tobias Brunner [Thu, 26 Sep 2013 08:15:03 +0000 (10:15 +0200)]
starter: Don't ignore keyingtries with rekey=no

Since keyingtries also affects the number of retries initially or when
reestablishing an SA it should not be affected by the rekey option.

Fixes #418.

6 years agoload-tester: Fix crash if private key was not loaded successfully
Tobias Brunner [Tue, 24 Sep 2013 07:24:59 +0000 (09:24 +0200)]
load-tester: Fix crash if private key was not loaded successfully

Fixes #417.

6 years agoprintf-hook: Write to output stream instead of the FD directly when using Vstr
Tobias Brunner [Mon, 23 Sep 2013 15:01:53 +0000 (17:01 +0200)]
printf-hook: Write to output stream instead of the FD directly when using Vstr

This avoids problems when other stdio functions are used (fputs,
fwrite) as writes via Vstr/FD were always unbuffered.

6 years agoandroid: New release after improving recovery after connectivity changes
Tobias Brunner [Mon, 23 Sep 2013 12:33:29 +0000 (14:33 +0200)]
android: New release after improving recovery after connectivity changes

6 years agoandroid: Change state handling to display errors occurring while the app is hidden
Tobias Brunner [Fri, 20 Sep 2013 13:07:41 +0000 (15:07 +0200)]
android: Change state handling to display errors occurring while the app is hidden

A new connection ID allows listeners to track which errors they have
already shown to the user or were already dismissed by the user.

This was necessary because the state fragment is now unregistered from
state changes when it is not shown.

6 years agoandroid: Don't update state fragments when they are not displayed
Tobias Brunner [Fri, 20 Sep 2013 12:07:40 +0000 (14:07 +0200)]
android: Don't update state fragments when they are not displayed

Besides that updates don't make much sense when the fragments are not
displayed this fixes the following exception:
java.lang.IllegalStateException: Can not perform this action after
onSaveInstanceState

6 years agoikev2: Force an update of the host addresses on the first response
Tobias Brunner [Fri, 20 Sep 2013 12:05:53 +0000 (14:05 +0200)]
ikev2: Force an update of the host addresses on the first response

This is especially useful on Android where we are able to send messages
even if we don't know the correct local address (this is possible
because we don't set source addresses in outbound messages).  This way
we may learn the correct local address if it e.g. changed right before
reestablishing an SA.

Updating the local address later is tricky without MOBIKE as the
responder might not update the associated IPsec SAs properly.

6 years agoike-sa: Resolve hosts before reestablishing an IKE_SA
Tobias Brunner [Fri, 20 Sep 2013 12:03:23 +0000 (14:03 +0200)]
ike-sa: Resolve hosts before reestablishing an IKE_SA

6 years agoandroid: Several plugins were moved from libcharon to libtnccs
Tobias Brunner [Fri, 20 Sep 2013 09:16:21 +0000 (11:16 +0200)]
android: Several plugins were moved from libcharon to libtnccs

These were moved in commits e8f65c5cde and 12b3db5006.

6 years agoandroid: Properly handle failures while initializing charon
Tobias Brunner [Fri, 20 Sep 2013 08:30:02 +0000 (10:30 +0200)]
android: Properly handle failures while initializing charon

6 years agokernel-netlink: Allow to override xfrm_acq_expires value
Ansis Atteka [Mon, 23 Sep 2013 04:21:39 +0000 (21:21 -0700)]
kernel-netlink: Allow to override xfrm_acq_expires value

When using auto=route, current xfrm_acq_expires default value
implies that tunnel can be down for up to 165 seconds, if
other peer rejected first IKE request with an AUTH_FAILED or
NO_PROPOSAL_CHOSEN error message. These error messages are
completely normal in setups where another application
pushes configuration to both strongSwans without waiting
for acknowledgment that they have updated their configurations.

This patch allows strongswan to override xfrm_acq_expires default
value by setting charon.plugins.kernel-netlink.xfrm_acq_expires in
strongswan.conf.

Signed-off-by: Ansis Atteka <aatteka@nicira.com>
6 years agoImplemented TCG/PB-PDP_Referral message
Andreas Steffen [Tue, 17 Sep 2013 19:57:08 +0000 (21:57 +0200)]
Implemented TCG/PB-PDP_Referral message

6 years agoAllow vendor-specific PB-TNC messages
Andreas Steffen [Tue, 17 Sep 2013 09:19:11 +0000 (11:19 +0200)]
Allow vendor-specific PB-TNC messages

6 years agoignore *.1 manpage files
Andreas Steffen [Tue, 17 Sep 2013 08:58:53 +0000 (10:58 +0200)]
ignore *.1 manpage files