gcc 5.4 correctly warned that a null pointer was being passed into the
entity-specific "set_qos" function when changing a topic QoS, where that
parameter was tagged as "non-null". As it was never dereferenced in
this case the resulting behaviour was still correct.
It turns out that the entire function was overly complicated and that
simply passing the entity pointer round allows eliminating a few
arguments as well.
(Oddly none of the more modern toolchains used pick this up.)
Signed-off-by: Erik Boasson <eb@ilities.com>
the destination cache of the network stack is in a certain state. The issue
is resolved by binding unicast sockets (incoming unicast and all outgoing
traffic) to the address of the interface instead of inaddr_any (0.0.0.0).
Set the new configuration option internal/BindUnicastToInterfaceAddr to
false to get the old behavior.
Co-authored-by: Erik Boasson <eb@ilities.com>
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
* Fix some typos.
Signed-off-by: ChenYing Kuo <evshary@gmail.com>
* Also update q_config.c, cyclonedds.rnc, cyclonedds.xsd for correct
build.
Signed-off-by: ChenYing Kuo <evshary@gmail.com>
* Remove cdds.md.
Signed-off-by: ChenYing Kuo <evshary@gmail.com>
ddsrt_asprintf did not copy non-English interface names. To fix this memory is
allocated with ddsrt_malloc and UTF-16 encoded interface names are converted to
UTF-8.
Signed-off-by: Jeroen Koekkoek <jeroen@koekkoek.nl>
This adds options to check for "unreasonable" RSS growth, receipt of a
minimum number of samples and having run a minimum number of roundtrips.
Signed-off-by: Erik Boasson <eb@ilities.com>
A keep-last volatile WHC retained data already overwritten by the writer
in the absence of ACKs, introduced by 231cb8c9.
Signed-off-by: Erik Boasson <eb@ilities.com>
The status mask on some readers got reduced to just "data available"
when used in conjunction with a waitset, but the consequence is that the
"subscription matched" listener would be suppressed.
Signed-off-by: Erik Boasson <eb@ilities.com>
This already was leaking out in the interface, so this name change was
needed too. The relationship between plist and xqos being so intimate,
doing the one but not the other made no sense.
Signed-off-by: Erik Boasson <eb@ilities.com>
The name (not its definition) now leaks out in ddsi_sertopic, and the
messy old names really shouldn't pollute the interface any more than
necessary.
Signed-off-by: Erik Boasson <eb@ilities.com>
This also removes the code duplication for the handling delivery from
local vs remote writers. (And it adds a test.)
Signed-off-by: Erik Boasson <eb@ilities.com>
This commit changes the implementation of topics so that multiple topic
entities can exist in a single participant for the same topic.
Different entities may refer to different topic implementations
(sertopics, akin to a type support in the DDS specification). All
entities (for the same participant) always have the same QoS, via the
new "ktopic" table in the participant.
Readers and writers are bound to a topic entity and inherit its
properties. If a topic comes in two definitions, say one for C and one
for C++, one can have a single participant with a reader delivering the
data in C representation and another reader delivering it in C++
representation.
This changes the behaviour of create_topic and find_topic: these now (on
successful return) always return a new entity (and thus with a unique
handle), where previously these would simply return a existing one when
possible.
This also requires some small additions to the sertopic/serdata
interface.
Signed-off-by: Erik Boasson <eb@ilities.com>
* Use the parameter tables to pretty-print QoS and plist, rather than a
hard-coded function supporting only the QoS.
* Support diffing two plists: a single table-driven function can handle
both nn_plist_t and ddsi_qos_t, and it removes the discrepancy between
the two types.
* Log content of discovery samples in trace rather than merely printing
"(plist)"
Signed-off-by: Erik Boasson <eb@ilities.com>
Implement trigger of certificate and permission expiries using the timed callbacks.
Implement directory operations such that trusted CA can be read.
This implements OS abstraction functions such as opendir and stat.
Signed-off-by: Stefan Kimmer <skimmer@s2e-systems.com>
Some of the liveliness qos tests were not using unique topic names
for the tests for local and remote readers. Re-using the participant
for these 2 tests results in unexpected reader-proxywriter matches
in the latter test. Fixed by adding a sequence number in the topic name.
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
This commit adds support for liveliness QoS when using local readers.
The implementation for (liveliness) expiration of writers used here is
similar to that used with proxy writers, and it also supports the three
liveliness kinds (1) automatic, which is trivial when using a local
reader and writer, (2) manual-by-participant and (3) manual-by-topic.
In addition, these changes and fixes are included in this commit:
- Fixed a bug in heartbeat handling in the reader: for manual-by-
participant writers the lease was not updated on reception of a
heartbeat message with liveliness flag set. This is fixed and a
test-case is added.
- Include the liveliness flag in a heartbeat message to the trace
- Trace all lease renewals, including liveliness leases
- Replaced liveliness changed state 'twitch' by 2 subsequent calls
to the status callback
- Added a test for liveliness duration 0 and 1ns (for both local
and remote readers)
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
This commit adds the build-in Access Control plugin that is part of the
DDS Security implementation for Cyclone.
The Access Control Plugin API defines the types and operations necessary
to support an access control mechanism for DDS Domain Participants.
Similar to other builtin plugins, the DDS Security access control plugin
is built as a shared library to allow dynamic library loading on runtime.
This enables DDS participants to use specific plugin implementations with
different configurations.
This commit includes some basic tests for the access control functions.
This initial version of the plugin does not support permissions expiry
(not-valid-after date in permissions configuration).
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Process review comments for access control plugin
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Part 2 of processing review changes for access control
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Add test for topicname dcps, add comment for xml date parser
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Fixed an bug in leap year count for year 2200, changed the rounding for sub-ns fraction and added an additional overflow test in DDS_Security_parse_xml_date
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Signed-off-by: Erik Boasson <eb@ilities.com>
Fix typo in dlopen_unknown test
Signed-off-by: Dennis Potman <dennis.potman@adlinktech.com>
Signed-off-by: Erik Boasson <eb@ilities.com>
This works around a termination issue on Windows caused by the process
sometimes being unable to send a packet to itself to wake up a thread
stuck in a blocking read on a socket.
Signed-off-by: Erik Boasson <eb@ilities.com>
Dispatchers are added to a custom linked list while each dispatcher owns a
fibheap of callbacks that are ordered by expiry timestamp. A seperate
fibheap is use for each dispatcher to allow disabling and enabling of
each dispatcher. A new and free function is added that creates a
timed_cd_data object that is used by every public function. A thread is
initialized at object creation instead of using a counter. Add
a protection for the terminate flag.
Signed-off-by: Joao Rebelo <jrebelo@s2e-systems.com>
A generic FSM has been added to DDS Security Core component to realize authentication handshake process.
The list of the states and the transitions are given in the creation and the FSM is started with a start call.
Passing arguments to transition funstions is possible.
Timeout transitions are possible.
Signed-off-by: Kurtulus Oksuztepe <kurtulus.oksuztepe@adlinktech.com>
Lease handling removes expired leases from the heap, and so a proxy
writer that transitions from not-alive to alive needs to re-schedule
expiry again or it will remain alive forever after having regained
liveliness once. This was missing for a manual-by-topic lease.
Signed-off-by: Erik Boasson <eb@ilities.com>
