The test runner deinit function often cancels all threads from the pool. This
operation might hang on error conditions, hence we should include that hook in
the test timeout to fail properly.
A worker raises SIGUSR1 to inform the main thread that the test fails. The main
thread then starts cancelling workers, but the offending thread should be
terminated immediately to prevent it from test continuation.
Adds AEAD support to the TLS stack, currently supporting AES-GCM. Brings fixes
for TLS record fragmentation, enforcing TLS versions < 1.2 and proper signature
scheme support indication.
While a hardcoded 1.2 version is fine when we offer that in Client Hello, we
should include the actually offered version if it has been reduced before
starting the exchange.
Fixes fragment reassembling if a buffer contains more than one record, but
the last record contains a partial TLS record header. Thanks to Nick Saunders
and Jamil Nimeh for identifying this issue and providing a fix for it.
To better separate the code path for different TLS versions and modes of
operation, we introduce a TLS AEAD abstraction. We provide three implementations
using traditional transforms, and get prepared for TLS AEAD modes.
The salt, or often called implicit nonce, varies between AEAD algorithms and
their use in protocols. For IKE and ESP, GCM uses 4 bytes, while CCM uses
3 bytes. With TLS, however, AEAD mode uses 4 bytes for both GCM and CCM.
Our GCM backends currently support 4 bytes and CCM 3 bytes only. This is fine
until we go for CCM mode support in TLS, which requires 4 byte nonces.
Limits cached OCSP verification to responses signed by the CA, a directly
delegated signer or a pre-installed OCSP responder certificate. Disables
auth config merge for revocation trust-chain strength checkin, as it breaks
CA constraints in some scenarios.
To avoid considering each cached OCSP response and evaluating its trustchain,
we limit the certificates considered for OCSP signing to:
- The issuing CA of the checked certificate
- A directly delegated signer by the same CA, having the OCSP signer constraint
- Any locally installed (trusted) certificate having the OCSP signer constraint
The first two options cover the requirements from RFC 6960 2.6. For
compatibility with non-conforming CAs, we allow the third option as exception,
but require the installation of such certificates locally.
This behavior was introduced with 6840a6fb to avoid key/signature strength
checking for the revocation trustchain as we do it for end entity certificates.
Unfortunately this breaks CA constraint checking under certain conditions, as
we merge additional intermediate/CA certificates to the auth config.
As key/signature strength checking of the revocation trustchain is a rather
exotic requirement we drop support for that to properly enforce CA constraints.
(Re-)Introduces X.509 Attribute Certificate support in IKE, and cleans up the
x509 AC parser/generator. ACs may be stored locally or exchanged in IKEv2
CERT payloads, Attribute Authorities must be installed locally. pki --acert
issues Attribute Certificates and replaces the removed openac utility.
This validator checks for any attribute certificate it can find for validated
end entity certificates and tries to extract group membership information
used for connection authorization rules.
While it really would be desirable to allow stream destruction during on_read()
callbacks, this does not work anymore since e49b2998. Until we have a proper
solution for this issue, use asynchronous disconnects for the only user doing
so.
Fixes#518.
To properly shut down charon-cmd with leak reports, only the main thread
should catch SIGUSR1 to shut down the application. Work threads should ignore
SIGUSR1 to avoid any hard application termination.
When doing a rekey for a CHILD_SA, the use counters get reset. An inactivity
job is queued for a time unrelated to the rekey time, so it might happen
that the inactivity job gets executed just after rekeying. If this happens,
inactivity is detected even if we had traffic on the rekeyed CHILD_SA just
before rekeying.
This change implies that inactivity checks can't handle inactivity timeouts
for rekeyed CHILD_SAs, and therefore requires that inactivity timeout is shorter
than the rekey time to have any effect.
When acting on files, we can use fstat() to estimate the buffer size. On
non-file FDs, we dynamically increase an allocated buffer.
Additionally we slightly change the function signature to properly handle
zero-length files and add appropriate unit tests.
When changing async callbacks on streams, we have to make sure the watcher
callback is not currently active and has temporarily disabled callbacks. This
could have been the case, as we didn't explicitly removed any pending
watcher registration if both callbacks are NULL.
By enforcing the watcher unregistration, we are sure the watcher callback is
not active and currently is not mangling the callback hooks. This should make
sure we avoid any races for the callback variables.
The hook data counts remaining buffer bytes, not used ones. Counting them
correctly fixes a crash for long hexdumps.
Further, print_in_hook() must return the number of bytes that would have been
written, not the actually written bytes. This is important, as we allocate a
dynamic buffer in bus that relies on the exact byte count. Fixes long hexdumps
that got truncated.
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.
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