| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit deprecates ouroboros_init and ouroboros_fini and adds them
as a constructor or destructor, causing these function to be run
automatically when a program that links to the library calls and exits
main(). For this to fully work, the library had to be split so that we
can avoid the irmd calling these functions (the IRMd has to create the
shm structures on which these calls depend).
The library is split in 3 parts: libouroboros-dev, libouroboros-irm
and libouroboros-common. The latter is linked to the other two so that
including libouroboros-dev or libouroboros-irm will also link
libouroboros-common.
|
|
|
|
|
|
|
|
|
|
| |
This adds a cleanup handler to ensure the lock is released in the
event that the thread is cancelled during a write. An alternative
solution could be to make the flow write in a non-blocking way (only
the blocking write contains an execution path that has a cancellation
point).
Fixes #55
|
|
|
|
|
|
|
|
|
| |
This makes the threadpool use pthread_cancel instead of setting an
exit flag that threadpool managed threads check periodically. This
drastically reduces CPU consumption in the irmd when running a lot of
applications. It requires cancellation handlers in the ipcp and irmd
to be implemented to ensure safe cancellation during operation and
shutdown.
|
|
|
|
|
|
| |
This will let the link state policy react to flow up and down events
by notifying the PFFs of the routing instances of this event so they
can take an appropriate action.
|
|
|
|
|
|
| |
This adds the Loop-Free Alternates (LFA) policy. In case a link goes
down a LFA may be selected to route the SDUs on without causing loops
instead of the main hop that just went down.
|
|\
| |
| |
| | |
Be routing
|
| |
| |
| |
| |
| | |
This adds a refcount to the graph edges so that it is only included in
the calculation if both sides announced it.
|
| |
| |
| |
| |
| | |
This keeps the index in the vertex struct so that is more easily
available during Dijkstra.
|
| |
| |
| |
| |
| |
| | |
This simplifies the Dijkstra implementation by immediately setting the
correct next hop during Dijkstra instead of looping through the list
of predecessors afterwards.
|
| |
| |
| |
| |
| | |
This simplifies several internal graph functions by passing an array
of bools instead of an array of vertices.
|
| |
| |
| |
| |
| |
| | |
This returns a list as routing table instead of a pointer to a pointer
to a pointer, which simplifies the looping through the routing table
and makes it more extensible for future additions.
|
|\ \
| | |
| | |
| | | |
ipcpd: normal: Set clock to realtime clock
|
| |/
| |
| |
| |
| |
| | |
A pthread_cond_wait was using the pthread_cond_clock and thereby
timing out constantly, consuming 100% CPU. This changes it to use the
realtime clock.
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The lookup_update was waiting for the LU_INIT state to resolve under
dht->lock which prevented that update.
This PR fixes this bug, but leaves a (very rare) bug when the
lookup_destroy is called while the lookup_update is waiting for the
LU_INIT state to resolve. The solution also is a (harmless) lock
inversion, but this is also not the best.
Fixes #51
Fixes #52
|
|
|
|
|
|
|
|
|
|
|
| |
This adds a threadpool manager to the DHT. This was needed because the
detached thread could cause a data race on shutdown.
The threadpool manager is revised to allow multiple instances in a
single program.
The irmd and ipcp now store commands in a buffer (list) instead of a
single buffer before passing it to handler threads.
|
|\
| |
| |
| | |
Some bugfixes
|
| | |
|
|\|
| |
| |
| | |
Some bugfixes
|
| | |
|
|/ |
|
| |
|
|\
| |
| |
| | |
ipcpd: normal: Add alternate hop PFF
|
| |
| |
| |
| |
| | |
This adds a PFF that returns an alternate hop as next hop in case the
hop that would have been returned is down.
|
|/
|
|
|
|
|
|
|
| |
This revises the sdu_scheduler of the normal IPCP to create a
scheduler thread per QoS cube and let the kernel scheduler schedule
them based on a priority. Priorities can be set at build time in a
range of 0-100, which will be mapped onto a suitable priority for the
kernel scheduler. The current scheduler policy is fixed to SCHED_RR
since it is the most suitable scheduler defined by POSIX.
|
|
|
|
| |
This avoids potential data races on rib.fuse.
|
|\
| |
| |
| | |
ipcpd: Handle DHT SDUs in different thread
|
| |
| |
| |
| |
| |
| |
| | |
The DHT will now spawn a thread when receiving SDUs to avoid
starvation of sdu scheduler threads.
Also fixes some locking issues.
|
|/
|
|
|
| |
This turns the PDU Forwarding Function of the IPCP into a policy. For
now only the simple PFF policy is available.
|
|\
| |
| |
| | |
ipcpd: Refuse query when DHT not in running state
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
| |
The DHT will now enroll or sync when a data transfer connection is
added. This avoids the need to create a temporary data transfer
connection during enrollment (and speeds it up considerably).
The notifier system was modified to take an opaque pointer to the
object that registers as a parameter.
|
|
|
|
|
|
|
|
| |
This adds the flow down event to Ouroboros. In the shim-eth-llc, a
netlink socket is opened which listens to device up/down events. For
each event the flow is then adjusted with fccntl to notify the user
the flow is down or back up again. In the normal IPCP an event is
thrown if a write reports that the flow is down.
|
| |
|
|
|
|
|
|
|
| |
This adds a virtual RIB that is accessible as a filesystem that is
accessed through a fuse mountpoint (configurable , default is
/tmp/ouroboros). Currently, each IPCP will export its link state
database.
|
|
|
|
|
|
|
|
| |
This removes the RIB as a datastructure and CDAP as the protocol
between IPCPs. CDAP, the rib and related sources are deprecated. The
link-state protocol policy is udpated to use its own protocol based on
a simple broadcast strategy along a tree. The neighbors struct is
deprecated and moved to the library as a generic notifier component.
|
|\
| |
| |
| | |
ipcpd: Shutdown DHT work thread first at destroy
|
| |
| |
| |
| | |
This prevents a potential deadlock on the wrlock.
|
|/
|
|
|
| |
When enroll or bootstrap failed, finalize_components could be called
twice, resulting in double frees or use-after-free.
|
|
|
|
|
| |
This avoids a redefinition of a typedef which is a C11 feature, and
thus not supported by clang when the c89 switch is enabled.
|
|
|
|
|
|
|
| |
This will remove the dst parameter from enroll_boot, since it was only
being used in logging. It also turns some other debug logs into info
logs, and makes the component names as they are used in logs all
lowercase.
|
| |
|
| |
|
|\
| |
| |
| | |
ipcpd: Fix state in dht lookup at lookup_wait
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
The correct state was returned but not set internally in the
lookup_struct, causing lookup_wait to return a lookup in LU_PENDING if
the wait timedout.
This also unlocks the IRMd during register calls and fixes some
cleanup if lookup_wait gets interrupted.
Fixes #48
|
|\ \
| |/
|/|
| | |
lib, ipcpd: Fix e-mail addresses in protobuf files
|
| |
| |
| |
| | |
This fixes the wrong e-mail addresses in the protobuf files.
|
|/ |
|
|
|
|
| |
This fixes several bugs as reported by the clang static analyzer.
|