Commit Graph
9388 Commits
Author SHA1 Message Date
Tobias Brunner 5abe3c52d3 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).
2013-10-11 15:16:05 +02:00
Tobias Brunner 4b8b1354ce 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.
2013-10-11 15:16:05 +02:00
Tobias Brunner fad11d602d sqlite: Implement transaction handling 2013-10-11 15:16:05 +02:00
Tobias Brunner f3cb889c9b mysql: Implement transaction handling 2013-10-11 15:16:04 +02:00
Tobias Brunner 947b76cda8 database: Add interface to handle transactions 2013-10-11 15:16:04 +02:00
Tobias Brunner 5f6a40827e mysql: Ensure connections are properly released in multi-threaded environments 2013-10-11 15:16:04 +02:00
Tobias Brunner ec91f15e3b crypto-factory: Try next available RNG implementation if constructor fails 2013-10-11 15:13:25 +02:00
Tobias Brunner 2e22333fbc crypto-factory: Order entries by algorithm identifier and (optionally) speed 2013-10-11 15:13:25 +02:00
Tobias Brunner e2c9a03d15 Remove HASH_PREFERRED, usages are replaced with HASH_SHA1, which is required for IKEv2 anyway 2013-10-11 15:13:25 +02:00
Tobias Brunner 3473cbab9c vstr: Forward actual field width
fmt_field_width is a flag that indicates if a field width
is defined in obj_field_width.
2013-10-11 15:12:16 +02:00
Martin Willi fc566632da unit-tests: support testing when leak-detective has not been enabled 2013-10-11 15:12:16 +02:00
Martin Willi 795cbb98c6 printf-hook-builtin: Print NaN/Infinity floating point values as such 2013-10-11 11:06:09 +02:00
Martin Willi 8af9bf70f5 printf-hook-builtin: Correctly round up floating point values 2013-10-11 11:06:09 +02:00
Martin Willi edc7a3d02f 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.
2013-10-11 11:06:09 +02:00
Martin Willi 7e6a4cdc84 printf-hook-builtin: Support GNU %m specifier 2013-10-11 11:06:09 +02:00
Martin Willi cabe5c0ff4 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.
2013-10-11 11:06:02 +02:00
Martin Willi ebca34d782 printf-hook: Add some basic printf() string/integer test functions 2013-10-11 11:05:37 +02:00
Martin Willi 243048248b printf-hook: Move glibc/vstr printf hook backends to separate files 2013-10-11 11:05:30 +02:00
Martin Willi d53002f088 libipsec: Enforce byte/packet lifetimes on SAs 2013-10-11 10:23:18 +02:00
Martin Willi 12fdc2b16b kernel-libipsec: Support ESPv3 TFC padding 2013-10-11 10:23:18 +02:00
Martin Willi 293515f95c libipsec: remove extra RFC4303 TFC padding appended to inner payload 2013-10-11 10:23:17 +02:00
Martin Willi d53f9b9637 kernel-libipsec: Support query_sa() to report usage statistics 2013-10-11 10:23:17 +02:00
Martin Willi b08967d6d8 libipsec: Support usage statistics and query_sa() on IPsec SAs 2013-10-11 10:23:17 +02:00
Martin Willi d7083b6541 kernel: Use a time_t to report use time in query_policy() 2013-10-11 10:23:17 +02:00
Martin Willi c99458e94e kernel: Use a time_t to report use time in query_sa() 2013-10-11 10:23:17 +02:00
Martin Willi 4817595876 updown: Install forwarding rules with the actually used protocol 2013-10-11 10:15:22 +02:00
Martin Willi c5d9b133e0 updown: Add a PLUTO_PROTO variable set to 'ah' or 'esp' 2013-10-11 10:15:21 +02:00
Martin Willi e48e530b44 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.
2013-10-11 10:15:21 +02:00
Martin Willi 757343d90e ike: Define keylength for aescmac algorithm 2013-10-11 10:15:21 +02:00
Martin Willi a1379e3210 ikev1: Support parsing of AH+IPComp proposals 2013-10-11 10:15:21 +02:00
Martin Willi 25f74be8f9 starter: Remove obsolete 'auth' option 2013-10-11 10:15:21 +02:00
Martin Willi d489e75579 ikev1: Accept more than two certificate payloads 2013-10-11 10:15:21 +02:00
Martin Willi 3771b85806 ikev1: Support en-/decoding of SA payloads with AH algorithms 2013-10-11 10:15:21 +02:00
Martin Willi 44e6aa4fb7 kernel-handler: Whitespace cleanups 2013-10-11 10:15:21 +02:00
Martin Willi f6037b5506 stroke: List proposals in statusall without leading '/' in AH SAs 2013-10-11 10:15:21 +02:00
Martin Willi 4bf92306eb ikev1: Delete quick modes with the negotiated SA protocol 2013-10-11 10:15:21 +02:00
Martin Willi 5d569e07fd trap-manager: Install trap with SA protocol of the first configured proposal 2013-10-11 10:15:21 +02:00
Martin Willi 21b096f3b8 child-sa: Save protocol during SPI allocation
This allows us to properly delete the incomplete SA with the correct protocol
should negotiation fail.
2013-10-11 10:15:21 +02:00
Martin Willi 908fe1632d ikev1: Negotiate SPI with the first/negotiated proposal protocol 2013-10-11 10:15:21 +02:00
Martin Willi cdab8630d9 ikev2: Allocate SPI with the protocol of the first/negotiated proposal 2013-10-11 10:15:21 +02:00
Martin Willi f0c59e1cf8 proposal: Strip redundant integrity algos for ESP proposals only 2013-10-11 10:15:21 +02:00
Martin Willi 0576412989 stroke: Configure proposal with AH protocol if 'ah' option set 2013-10-11 10:15:20 +02:00
Martin Willi a07b97e804 starter: Add an 'ah' keyword for Authentication Header Security Associations 2013-10-11 10:15:20 +02:00
Andreas Steffen 3588299fb8 Keep a copy of the tnccs instance for PT-TLS handover 2013-10-09 19:03:07 +02:00
Tobias Brunner 3e3db3743e xauth-pam: Make trimming of email addresses optional
Fixes #430.
2013-10-04 10:49:54 +02:00
Martin Willi d2e4dd75b7 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).
2013-09-30 13:51:12 +02:00
Martin Willi 9c19d7ca31 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.
2013-09-30 13:51:11 +02:00
Martin Willi ee99f37ecc 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.
2013-09-30 13:51:11 +02:00
Tobias Brunner e4b7b48c1e eap-radius: Increase buffer for attributes sent in RADIUS accounting messages
64 bytes might be too short for user names/identities.
2013-09-27 13:37:12 +02:00
Tobias Brunner c8f34ba7b6 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.
2013-09-27 09:24:03 +02:00