| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|/
|
|
|
| |
This initializes conn_info in the ribmgr and fmgr before it is passed
to cacep.
|
| |
|
| |
|
| |
|
|
|
|
| |
This fixes some memleaks as reported by valgrind.
|
|
|
|
|
| |
The QoS spec was not being initialized before being added to the graph
structure, resulting in an error when compiled with clang.
|
|\
| |
| |
| | |
irmd: Fix some missing locks
|
| | |
|
| | |
|
| |
| |
| |
| | |
These solve several bugfixes in the normal.
|
|\ \
| |/
|/| |
|
| |
| |
| |
| |
| |
| | |
When the flow_req_arr returned a negative error, it was not handled
correctly in the local IPCP, and an access was made into a table with
a negative integer. This handles it correctly.
|
|\ \
| | |
| | |
| | | |
ipcpd: normal: Split fmgr init into init and start
|
| | |
| | |
| | |
| | |
| | |
| | | |
This split the initialization of the flow manager into an init part
and a start part. This avoids the usage of data structures that have
not been properly initialized yet.
|
|\ \ \ |
|
| | | |
| | | |
| | | |
| | | |
| | | | |
This solves some race conditions where packets arrive on a flow before
it is added to a flow_set.
|
| |/ /
| | |
| | |
| | |
| | | |
The temporary cube parameter must be initialized if a qosspec is
passed to avoid uninitialized values for the qoscube in the prototype.
|
| |\ \
| | | |
| | | |
| | | | |
ipcpd: Fix cleanup of connmgr
|
| | | |
| | | |
| | | |
| | | |
| | | | |
This lessens the likelyhood that the CDAP SDU thread is not yet active
when a new connection starts sending CDAP messages.
|
| | | |
| | | |
| | | |
| | | | |
Fixed trying to double lock and cleanup order in main.
|
| |\ \ \
| | |/ /
| |/| |
| | | | |
Be graph
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This lets the routing component listen to RIB events. It listens to
/fsdb which is populated with FSOs. The graph that is kept within the
routing component is updated depending on the event that was received.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
This adds a graph structure which will be updated by routing when it
is notified about a new RIB event. The routing can then use this graph
as input for calculating the shortest path to a destination.
|
| | | | |
|
| |/ / |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This makes the routing component create a Flow State Database
(FSDB). An FSDB contains Flow State Objects (FSOs). An FSO is created
when a neighbor is added, it is deleted when a neighbor is removed and
its QoS is updated when a neighbor's QoS changes.
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
You can now add multiple flows to a CDAP instance. This will simplify
sending messages to different peers (e.g. for syncing the RIB). A
request will now return an array of keys terminated by
CDAP_KEY_INVALID. Removes the enum from the CDAP proto file to just
take the opcode as an integer.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
Our mailserver was migrated from intec.ugent.be to the central
ugent.be emailserver. This PR updates the header files to reflect this
change as well. Some header files were also homogenized if the
parameters within the functions were badly aligned.
|
| | |
| | |
| | |
| | | |
The CMakeLists files are now properly indented.
|
| | |
| | |
| | |
| | | |
This removes the CDAP flow class, which is no longer needed.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This refactors the different Application Entities of the normal
IPCP. They all listen to and use the connection manager to establish
new application connections.
This commit also adds a neighbors struct to the normal IPCP. It
contains neighbor structs that contain application
connection. Notifiers can be registered in case a neighbor changes
(added, removed, QoS changed).
The flow manager has an instance of this neighbors struct and listens
to these events to update its flow set. The routing component also
listens to these events so that it can update the FSDB if needed. The
flow manager now also creates the PFF instances and the routing
instances per QoS cube.
The RIB manager also uses this an instance of the neighbors struct and
listens to neighbor events as well.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This adds the connection manager which allows the different AEs of the
normal IPCP to register with it. An AE can then use the connection
manager to allocate a flow to a neighbor, or to wait for a new
connection from a neighbor.
|
| | | |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
By removing authentication as part of CACEP, all policies
disappear. CACEP becomes a policy-free connection establishment
protocol between Application Entities. Authentication can later be
added cleanly as a pure policy function when needed.
|
| | |
| | |
| | |
| | |
| | | |
The AE name should not be passed over the layer boundaries. If an
application has more than one AE it should exchange this in CACEP.
|
| |\ \
| | | |
| | | |
| | | | |
lib: Revise CACEP API
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
The information passed to CACEP is split between the information about
the connection and the information to be used during the
authentication exchange.
|
|\ \ \ \
| |/ / /
|/| | /
| | |/
| |/| |
|
| | |
| | |
| | |
| | |
| | | |
This fixes bad timedwaits for the state of the reg_entry. Also
slightly revised timedwaits throughout the prototype.
|
|\| |
| |/
|/| |
|
| | |
|
|\| |
|
| | |
|
| |
| |
| |
| |
| | |
This prevents assertion failures in the IPCP in some cases. IPCPs can
now safely assert the type.
|
|\| |
|
| |
| |
| |
| |
| |
| | |
When hammering the oping thread, it will have write fails when the
buffer gets full as its flow is non-blocking. It would stop and
deallocate the flow, but should just continue.
|
|\| |
|
| |
| |
| |
| |
| |
| | |
Sometimes the receiver thread got the SDU before the writer thread has
set the sent time when testing over the local. The sent time is now
written before actually sending to avoid this.
|
| |
| |
| |
| |
| |
| |
| | |
When starting two IRMds, the second one exits, but applications can't
contact the first because the second instance took the UNIX socket.
Now the lockfile is checked for running IRMd instances before opening
the UNIX socket.
|