| Commit message (Collapse) | Author | Age | Files | Lines |
... | |
| |\
| | |
| | |
| | | |
lib: Stabilise flow allocation
|
| | |
| | |
| | |
| | |
| | | |
Remote requests for a flow that's locally deallocated should not
generate an error. Sending messages should not be under lock.
|
| | |
| | |
| | |
| | | |
This solves some potential data races in the shims.
|
| |/
|/| |
|
|/
|
|
|
|
|
| |
Deallocation was reverted to a synchronoous operation between the AP,
IRMd and IPCP in order to avoid inconsistent states of the port_id.
Fixes some memory leaks, particularly the shm_flow_set is now closed
upon deallocation.
|
|
|
|
|
| |
The default flow options were not set when a flow allocation request
arrives remotely. This went unnoticed due to a previous bug.
|
|
|
|
|
| |
gethostbyname is included in POSIX 2001. Ouroboros requires at least
POSIX 2001 and recommends POSIX 2008 for stable builds.
|
| |
|
|
|
|
|
|
| |
CheckLibraryExists was not included. Apparently the
CHECK_FUNCTION_EXISTS call fails with the strict prototyping flag
turned on.
|
|
|
|
|
| |
This has the code checked with -Wcast-qual and -Wconversion flags.
These flags were removed because SWIG generated code fails.
|
|
|
|
|
| |
This reduces the risk for some bugs, for instance due to
signed/unsigned mismatches and unused variables.
|
|\
| |
| |
| | |
lib: Demultiplex the fast path
|
| |
| |
| |
| |
| |
| |
| |
| | |
The fast path will now use an incoming ring buffer per flow per
process. This necessitated the development of a new method for the
asynchronous io call, which is now based on an event queue system for
scalability (fqueue). The ipcpd's and tools have been updated to this
API.
|
|/
|
|
|
| |
There was no check to see if the normal IPCP was enrolled or not when
a flow allocation request is done.
|
|
|
|
|
|
|
| |
Added the missing implementation of setting read/write options for
flows. This allows applications to block the fast path for
remotes. IPCPs can use this to block the fast path for the N + 1 flow
when receiving remote deallocation requests.
|
| |
|
|
|
|
|
| |
Both the N and N-1 entity must now call flow_dealloc for the port_id
to be released.
|
|
|
|
|
|
|
|
| |
This commit will remove the RMT component from the normal IPCP, as
some of its functionality would else be duplicated in the FMGR. Now
all reading from flows, either N-1 or N+1 is done in the FMGR, then
either passed to the FRCT or a lookup is performed in the PFF (not
there yet) and the PDU is forwarded.
|
|
|
|
|
|
|
| |
This is the first version of the fast path bootstrap in the normal
IPCP. It sets up a connection with the other end, and creates the
appropriate data structures. N+1 and N-1 SDUs are read and written and
passed through the right components.
|
| |
|
|
|
|
|
| |
Flow deallocation from the application will immediately return (void
call). The IRMd will not send a reply message.
|
|
|
|
| |
Now correctly handles and drops non-ouroboros traffic.
|
|
|
|
|
| |
Upon deallocation, the remote's SAP was being used to lookup the fd
instead of our own SAP. This resulted in strange behaviour.
|
|
|
|
|
|
| |
Some locks were taken at an inappropriate time, causing the
shim-eth-llc to block indefinitely upon shutdown and causing flow
allocation to halt since a response could never be given.
|
|
|
|
|
|
|
|
|
|
|
|
| |
IPCPs can now use ap_init() to initialize the memory. All flows are
accessed using flow descriptors, this greatly simplifies IPCP
development. Reverts the fast path to a single ap_rbuff per process.
Splits lib/ipcp into irmd/ipcp and lib/ipcp-dev. Adds a lib/shim-dev
holding tailored functions for shims. Moves the buffer_t to utils.h.
Fixes the shim-eth-llc length field. Removes the flow from shared.h.
Fixes #4
Fixes #5
|
|
|
|
| |
There was some crapped up indentation in the crc32 source file.
|
|\
| |
| |
| | |
ipcpd: normal: Add operations to get and set the PCI
|
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This adds the operations needed in the normal IPCP to get and set the
Protocol Control Information. It allows to allocate or release space
in the current DU. The struct pci can be serialized into newly
allocate space. Vice versa, a struct pci can be deserialized given a
DU. It allows for decreasing the TTL in the DU and for calculating the
CRC32. The TTL and CRC32 can now be selected when creating a new DIF.
|
| | |
|
|\ \
| |/
|/|
| | |
lib: Add northbound ringbuffers
|
| |
| |
| |
| | |
Fast path is split in north and southbound paths.
|
|\ \
| |/
|/|
| | |
lib: ipcp: Decouple flow_dealloc
|
| |
| |
| |
| | |
Sending a dealloc to the irmd moved to its own call.
|
|/
|
|
|
|
| |
The head and tail alloc and release operations were taking an int to
identify the idx instead of a ssize_t. The size was a ssize_t instead
of a size_t.
|
|
|
|
|
|
| |
Disables robust mutexes and clock attributes for condition variables
for compatibility with OSX (SUSv2). Implements clock_gettime and adds
some defines for OSX compatibility in time_utils.
|
|
|
|
|
|
|
|
|
|
| |
The shm_du_map is renamed to shm_rdrbuff to reflect the Random
Deletion Ringbuffer used in the implementation. The close_on_exit call
is removed and SDUs are cleaned up by the application in the ap_fini()
call. This required a non-blocking peek() operation in the shm_ap_rbuff.
Some initial implementation for future support of qos cubes has been
added to the shm_rdrbuff.
|
|
|
|
| |
IPCPs will report their pid on shutdown for faster debugging.
|
|
|
|
| |
Destruction of the object in the reply stage was unsafe.
|
|
|
|
| |
This removes the ret value which was not being set correctly.
|
|\
| |
| |
| | |
Fix locks
|
| | |
|
|\|
| |
| |
| | |
ipcpd: Revised locking
|
| |
| |
| |
| |
| |
| | |
The state lock was reverted to an rwlock to avoid interference of
management functions with the fast path. IPCPs now close without
calling unsafe functions in the signal handler.
|
|\ \
| |/
|/|
| | |
ipcpd: normal: Add initial steps for N+1 flow allocation
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
This adds the initial framework for flow allocation between two N+1
endpoints. The FMGR will receive flow allocation requests and will
create a connection as a result, addressed to the right address, it
will also pass a flow allocation message to this address. Upon receipt
on the other side, the FMGR will be receive a flow allocation message
and a pointer to a new connection. The FMGR can then accept or destroy
the connection. This commit also introduces the RMT function, which is
needed by the FRCT to forward its SDUs on the right file descriptor.
|
| | |
|
|/ |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The complete data model inside the IRMd has been restructured. The
bind operation was revised to allow binding of AP and AP instances and
register those names with different DIFs (see "irm bind" for details).
Server applications require to call ap_init with a server name
argument, client application that do not the be reachable over any DIF
can call ap_init(NULL). Calling ap_init for a client with a specified
name will not have adverse consequences for the application, but will
consume unnecessary resources in the IRMd.
Application servers can now be started at any point after the IRMd has
been started. Starting servers, binding AP names and registering names in
DIFs can be performed in any order that does not defy temporal logic.
Supports naming instances by their pid. In case of IPCP Instances
created with the IRM tool, the name assigned during "irm ipcp create"
can be used.
All the changes required updates in the tools.
|
|
|
|
|
|
|
|
| |
This adds a condition variable with a timeout to the CDAP request so
that we can respond correctly to the answer from the remote. It also
adds a timeout to the condition variable waiting on completion of
enrollment. Furthermore, for every CDAP callback a new thread is now
spawned, to avoid deadlocking in case a callback is stuck.
|
|
|
|
|
|
|
| |
Previously it was up to the admin to choose whether or not to enable
DDNS. Now the build just checks if the executables can be found, and
if they are available DDNS functionality is enabled, else it is
disabled.
|