Tobias Brunner [Thu, 7 Nov 2013 08:50:12 +0000 (09:50 +0100)]
trap-manager: Prevent deadlock when installing trap policies
Because the write lock was held while calling add_policies() on
child_sa_t, which finishes with a call to child_state_change() on bus_t,
a deadlock would ensue if CHILD_SAs are concurrently being established,
which also causes a call to child_state_change() that will require
the read lock in trap_manager_t.
No locks are now being held while creating the CHILD_SA and installing the
trap policies.
Martin Willi [Wed, 6 Nov 2013 09:16:56 +0000 (10:16 +0100)]
Merge branch 'unit-tests'
Replace the "check" based libstrongswan unit test framework with our own,
giving us more flexibility for our specific needs.
The new framework is more portable and uses complete libstrongswan init/deinit
cycles for each test to properly catch leaks. It fully supports multi-threaded
tests, and brings many of them for all threading primitives, watcher and
streams.
The --enable-unit-tests option is not required anymore for libstrongswan tests,
but still is for the still "check" based charon-tkm tests.
Martin Willi [Tue, 5 Nov 2013 13:40:03 +0000 (14:40 +0100)]
unit-tests: Separate test runner to a library, reusable by other tests
Other users may make use of the noinst libtest.la helper library to implement
unit tests. For libstrongswan, tests.[ch] provide the configuration for test
runner to perform unit tests in a simple manner.
Martin Willi [Thu, 17 Oct 2013 15:05:38 +0000 (17:05 +0200)]
unit-tests: Pass linked_list->invoke* varargs as uintptr_t
Passing integers of unspecified length may result in passing an integer shorter
than uintptr_t. When reading them back, we might get more data than passed,
resulting in a failure.
Martin Willi [Mon, 4 Nov 2013 10:26:33 +0000 (11:26 +0100)]
unit-tests: Enable libstrongswan tests even if --enable-unit-tests not set
As we don't depend on the check framework anymore, we can enable the unit tests
by default. These are built/executed with "make check" only, so it makes no
sense to disable them.
Martin Willi [Fri, 1 Nov 2013 10:28:53 +0000 (11:28 +0100)]
ike: Don't immediately DPD after deferred DELETEs following IKE_SA rekeying
Some peers seem to defer DELETEs a few seconds after rekeying the IKE_SA, which
is perfectly valid. For short(er) DPD delays, this leads to the situation where
we send a DPD request during set_state(), but the IKE_SA has no hosts set yet.
Avoid that DPD by resetting the INBOUND timestamp during set_state().
Ansis Atteka [Wed, 30 Oct 2013 02:48:51 +0000 (19:48 -0700)]
updown: fix segfault when interface name can't be resolved
The child_updown() function sets up environment variables to the updown
script. Sometimes call to hydra->kernel_interface->get_interface() could
fail and iface variable could be left uninitialized. This patch fixes
this issue by passing "unknown" as interface name.
Here is the stacktrace:
0 0x00007fa90791f445 in raise () from /lib/x86_64-linux-gnu/libc.so.6
1 0x00007fa907922bab in abort () from /lib/x86_64-linux-gnu/libc.so.6
2 0x0000000000401ed7 in segv_handler (signal=11) at charon.c:183
3 <signal handler called>
4 0x00007fa90793221f in vfprintf () from /lib/x86_64-linux-gnu/libc.so.6
5 0x00007fa9079f0580 in __vsnprintf_chk () from /lib/x86_64-linux-gnu/libc.so.6
6 0x00007fa9079f04c8 in __snprintf_chk () from /lib/x86_64-linux-gnu/libc.so.6
7 0x00007fa8f9b95b86 in snprintf (
__fmt=0x7fa8f9b961b8 "2>&1 PLUTO_VERSION='1.1' PLUTO_VERB='%s%s%s' PLUTO_CONNECTION='%s' PLUTO_INTERFACE='%s' PLUTO_REQID='%u' PLUTO_ME='%H' PLUTO_MY_ID='%Y' PLUTO_MY_CLIENT='%H/%u' PLUTO_MY_PORT='%u' PLUTO_MY_PROTOCOL='%u"..., __n=1024, __s=0x7fa8f7923440 "2>&1 PLUTO_VERSION='1.1' PLUTO_VERB='up-host' PLUTO_CONNECTION='remote-40.0.0.40' PLUTO_INTERFACE='\367\250\177")
at /usr/include/x86_64-linux-gnu/bits/stdio2.h:65
8 child_updown (this=0x8486b0, ike_sa=0x7fa8e4005f80, child_sa=0x7fa8d4008290, up=true) at updown_listener.c:308
9 0x00007fa907ecc11c in ?? () from /usr/lib/strongswan/libcharon.so.0
10 0x00007fa907ef89bf in ?? () from /usr/lib/strongswan/libcharon.so.0
11 0x00007fa907ef2fc8 in ?? () from /usr/lib/strongswan/libcharon.so.0
12 0x00007fa907ee84ff in ?? () from /usr/lib/strongswan/libcharon.so.0
13 0x00007fa907ee3067 in ?? () from /usr/lib/strongswan/libcharon.so.0
14 0x00007fa90835e8fb in ?? () from /usr/lib/strongswan/libstrongswan.so.0
15 0x00007fa908360d30 in ?? () from /usr/lib/strongswan/libstrongswan.so.0
16 0x00007fa907cade9a in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
17 0x00007fa9079db4bd in clone () from /lib/x86_64-linux-gnu/libc.so.6
18 0x0000000000000000 in ?? ()
Ansis Atteka [Fri, 25 Oct 2013 22:42:10 +0000 (15:42 -0700)]
debian: build debug symbol package
Before this patch all debug symbols were stripped off and simply
discarded. GDB without debug symbols is barely usable, but at
the same time distributing binaries with debug symbols would
drastically increase strongswan/libstrongswan package size.
Instead of discarding debug symbols, it would be better to strip
them off into a dedicated debian package. So that, if needed, one
could still install them and use GDB.
Tobias Brunner [Tue, 29 Oct 2013 14:16:22 +0000 (15:16 +0100)]
Use exact mask when calling umask(2)
Due to the previous negation the high bits of the mask were set, which
at least some versions of the Android build system prevent with a compile-time
check.