| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
| |
If an IPCP SEGVs abruptly, the sanitizer in the IRMd would also cause
a SEGV because of looping over a registry linked list inside another
safe loop over the same list. A function is added to the registry to
avoid this.
|
|
|
|
|
|
| |
In extreme cases, the response to a CDAP request can arrive before the
AE doing the request has called cdap_request_wait. This PR handles it
by waiting for the state to move away from INIT before responding.
|
| |
|
| |
|
| |
|
|
|
|
|
|
|
|
| |
The new element must be added before the larger element
(list_add_tail, since it's a circular list).
Also fixes a bug and improves the locking in rib_pack (rib must be
locked until the subtree is fully packed in the buffer).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The rib_pack function allows packing a subtree of the RIB for
dissemination. The options PACK_HASH_ROOT and PACK_HASH_ALL will add
the hashes for the root object of the packed subtree or every object
to the packed message respectively. Checking of the hashes is
currently only performed at the top level object, verifying the
complete operation.
The rib_unpack function unpacks a packed message and inserts its
contents in the RIB. The option UNPACK_CREATE flags that the unpack
operation is allowed to create new objects, else it will only update
existing objects. More advanced options could be added in the future.
The packed message structure uses Google Protocol Buffers, as defined
in ro.proto.
It adds tests for these functions to the rib_test.
|
|
|
|
|
|
| |
When trying to double bootstrap, the entry name was strdup'd before
the bootstrap leaking the previous name. Now it bootstraps first and
sets the name after a successful bootstrap only.
|
|
|
|
|
| |
Returns the names of the nodes (not the full paths). The function
allocates memory to the children pointer that should be freed.
|
| |
|
| |
|
|
|
|
|
| |
The -EAGAIN from a failed non-blocking flow read would trigger the
assertion. This snuck in somehow (we need some unit tests for dev.c).
|
|
|
|
|
|
|
|
|
|
|
|
| |
The new RIB is implemented as a tree with doubly linked nodes (parents
keep a link to each child, each child keeps a link to its parent). An
index is kept in a btree using CRC32 hashes of the path name in the
RIB. Nodes keep an SHA3-256 hash value that is unique for the entire
subtree. This allows quick checks to see if two RIB subtrees are in
sync. The event system for the RIB is based on the event system for
flows (fqueue), but implemented completely in dynamic memory using
linked lists. An initial test is performed for the RIB. This PR does
not modify existing code to use the new RIB.
|
|
|
|
|
| |
Allows moving the elements of one linked list to
another. Re-initializes the source list.
|
| |
|
|
|
|
|
| |
Adds an implementation of B-trees of order k (k children, min fill is
k/2, max fill k - 1). Useful to implement indexes for faster lookups.
|
| |
|
|
|
|
| |
The CRC32 implementation will be used to ID objects in the RIB.
|
|
|
|
|
|
|
|
| |
Sometimes the timerwheel test could fail because of scheduler timings
(an issue more apparent in Virtual Machines). When the first check
fails, the test will wait for another complete round of the
timerwheel. The maximum number of slots in the test is reduced to
reduce the test time.
|
|
|
|
| |
Test the SHA3 function with known hashes.
|
|\
| |
| |
| | |
Be gam
|
| |
| |
| |
| |
| | |
This allows the selection of a policy for the graph adjacency
manager. Currently we only support constructing a complete graph.
|
| |
| |
| |
| |
| | |
After a previous commit the API of the linked list was changed. This
updates the files that were not yet merged upstream.
|
| |\ |
|
| | |
| | |
| | |
| | |
| | | |
This commit adds the graph adjacency manager to the normal IPCP, which
sets up N-1 flows to other members.
|
| | |
| | |
| | |
| | |
| | |
| | | |
This adds CACEP to the library, which can be used for authenticating
the other party after flow allocation and for obtaining other relevant
information.
|
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
This allows IPCPs to bind a name, so that they can announce their name
to neighbors which can then allocate a flow to them. Registering of
the name happens by an administrator. It also moves the irmd_api to
common ground, since it is used by all IPCPs.
|
| | |
| | |
| | |
| | |
| | |
| | | |
Adds an SHA-3 hash implementation that will be used in the rib,
adapted and redistributed from the RHASH project
(https://github.com/rhash/RHash)
|
| |/
|/| |
|
| | |
|
| |
| |
| |
| |
| | |
Copyright is set to 2016 - 2017. License text on includes and sources
in the library are changed to indicate the LGPLv2.1 license.
|
|/
|
|
| |
Adds LGPL license to the ouroboros lists.
|
|\
| |
| |
| | |
irmd: Change return value of register functions
|
| |
| |
| |
| |
| |
| |
| | |
This changes the return value of the register and unregister
functions. If they were able to reg/unreg some entries a positive
value is returned, otherwise an error. Some spelling errors in the
IRMd have also been fixed.
|
| |
| |
| |
| |
| |
| | |
The state must be set to IPCP_INIT before the main_loop threads are
started or else they will exit if they reach the check earlier than
the state is set.
|
| | |
|
|/
|
|
|
| |
The state of the IPCP should revert to INIT when pthread_create fails
in the normal and bootstrapping procedures.
|
|
|
|
|
|
|
| |
These operations separe the starting and joining of the main ipcp
threads into ipcp_boot() and ipcp_shutdown() operations. This allows
the proper cleanup of user data and user threads after the IPCP is
requested to shut down.
|
|
|
|
| |
This removes a bad unlock in the flow manager.
|
|\
| |
| |
| | |
ipcpd: Initialize rnodes in ribmgr_ro_create
|
| |
| |
| |
| | |
Fixes warnings by some compilers about possibly uninitialized access.
|
|\|
| |
| |
| | |
ipcpd: Fix uninitialized variables in ribmgr
|
| | |
|
| | |
|
|\|
| |
| |
| | |
ipcpd: Refactor normal ipcp, initial commit
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Reorganizes the normal IPCP a bit to make sure internal components do
not need to access the state of the IPCP. The IPCP has now a thread
calling accept and delegating it to the correct component based on the
AE name (this used to be in the fmgr).
Internal components are initialized upon enrollment or bootstrap of
the IPCP. If a step fails, the IPCP goes back to the INIT state, if
all components boot correctly, it goes to the operational state.
RIB synchronization is still done by sending a CDAP start/stop and
syncing with a ribmgr state, but needs revision later on.
|
|/
|
|
|
|
|
|
|
|
| |
Now correctly relays the qoscube end-to-end in the stack.
A simple function specifying the cube in the spec is used for initial
testing.
The translation is now done in dev.c, but it could be moved elsewhere
when qos cabability matures and the need arises.
|
| |
|
|
|
|
|
|
| |
The main thread will wait for the IPCP_OPERATIONAL state before
starting the fmgr main thread by calling fmgr_init(), instead of the
fmgr itself waiting for that state.
|
| |
|