summaryrefslogtreecommitdiff
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* Merged in dstaesse/ouroboros/be-demux (pull request #267)Sander Vrijders2016-10-2119-1070/+1537
|\ | | | | | | lib: Demultiplex the fast path
| * lib: Move rbuff creation/destruction to IRMddimitri staessens2016-10-216-84/+105
| | | | | | | | | | This stabilises flow allocation now that the rbuffs are created upon flow allocation. Only the IRMd can sync this process sufficiently.
| * lib: Demultiplex the fast pathdimitri staessens2016-10-2117-1051/+1497
| | | | | | | | | | | | | | | | 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.
* | ipcpd: normal: Avoid allocating flow over unenrolled IPCP.Sander Vrijders2016-10-171-0/+10
|/ | | | | There was no check to see if the normal IPCP was enrolled or not when a flow allocation request is done.
* lib, dev: Implement read/write options for flowsdimitri staessens2016-10-162-0/+19
| | | | | | | 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.
* lib: Stabilize fast flow deallocation over local IPCPdimitri staessens2016-10-153-49/+49
|
* lib: Fix initialisation of flow_setdimitri staessens2016-10-151-2/+3
| | | | The safe copy was not correctly initialised.
* lib: Fix flow deallocation logicdimitri staessens2016-10-137-34/+35
| | | | | Both the N and N-1 entity must now call flow_dealloc for the port_id to be released.
* ipcpd: normal: Improve upon the internal designSander Vrijders2016-10-1211-574/+411
| | | | | | | | 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.
* Merge remote-tracking branch 'upstream/be' into be-connectionSander Vrijders2016-10-126-93/+177
|\
| * lib: Track SDUs in the fast pathdimitri staessens2016-10-113-88/+174
| | | | | | | | | | | | This will allow to finalize deallocating flows until all SDUs have been processed. Read and write calls will now block when a flow was deallocated. Replaces NULL checks in the fast path with asserts.
| * lib: Add a type for struct flow_setdimitri staessens2016-10-093-5/+3
| | | | | | | | | | | | The struct flow_set can now be accessed in applications as flow_set_t. Fixes some malformed lines and homogenizes output when receiving bad SDUs in the oping tool.
* | ipcpd: normal: First version of the fast path bootstrapSander Vrijders2016-10-1111-209/+697
|/ | | | | | | 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.
* lib: Add non-copy reading from fd for IPCPsdimitri staessens2016-10-073-10/+45
|
* lib, dev: Add asynchronous deallocationdimitri staessens2016-10-074-55/+72
| | | | | Flow deallocation from the application will immediately return (void call). The IRMd will not send a reply message.
* lib: Fix shutdown with pending SDUsdimitri staessens2016-10-062-2/+4
| | | | | The SDU's were not correctly removed upon shutdown, peek should be pop.
* ipcpd: Filter traffic on shim-eth-llcdimitri staessens2016-10-063-7/+31
| | | | Now correctly handles and drops non-ouroboros traffic.
* ipcpd: shim-eth-llc: Use correct SAP to lookup in the tableSander Vrijders2016-10-051-3/+5
| | | | | Upon deallocation, the remote's SAP was being used to lookup the fd instead of our own SAP. This resulted in strange behaviour.
* lib, ipcpd: Fix bad lockingSander Vrijders2016-10-052-7/+4
| | | | | | 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.
* lib, ipcp: Revise fast path and flow interfacesdimitri staessens2016-10-0432-2763/+1975
| | | | | | | | | | | | 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
* lib: Fix shm for evil devicesdimitri staessens2016-09-083-6/+6
| | | | | | Cupertino's finest don't seem to support a write call with a filedescriptor that points to a shared memory file. ENOTSUP is not mentioned in their manpage either. Classic.
* ipcpd: normal: Fix crapped up indentationSander Vrijders2016-09-071-48/+48
| | | | There was some crapped up indentation in the crc32 source file.
* Merged in sandervrijders/ouroboros/be-shm-pci (pull request #247)dimitri staessens2016-09-0715-239/+372
|\ | | | | | | ipcpd: normal: Add operations to get and set the PCI
| * ipcpd: normal: Add operations to get and set the PCISander Vrijders2016-09-0715-239/+372
| | | | | | | | | | | | | | | | | | 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: Set umask for file creation permissionsdimitri staessens2016-09-074-17/+12
| | | | | | | | | | Permissions are now set correctly upon creation, removing the need to call fchmod.
* | ipcpd: Add missing include for FreeBSDdimitri staessens2016-09-071-0/+1
| |
* | lib: Add lazy copy of flow_setdimitri staessens2016-09-071-8/+16
| | | | | | | | | | | | The flow_set now has a safe copy which is threadsafe for the select call without needing a lock. This greatly speeds up consecutive select calls.
* | irmd: Fix register with unenrolled IPCPs in systemdimitri staessens2016-09-061-0/+3
| | | | | | | | | | If there are unregistered IPCPs (no DIF name set) the lookup should skip over them without segfaulting.
* | lib: dev: Provide a set of fds to flow_selectdimitri staessens2016-09-065-28/+188
| | | | | | | | | | | | | | The flow_select call now takes as a parameter a flow_set_t, which specifies a set of flow descriptors that will unblock the select call when an SDU for one of them arrives. The select call has been moved to its own header.
* | lib: dev: Add fd lookup tabledimitri staessens2016-09-031-23/+25
| |
* | lib: dev: Refactordimitri staessens2016-09-031-182/+188
| | | | | | | | | | Renamed _ap_instance to ai, adds a char * daf_name (to be set by ipcps in ipcp-dev.h, later to be set when we generalise enrolment to DAFs).
* | Merged in dstaesse/ouroboros/be-split-fp (pull request #239)Sander Vrijders2016-09-023-7/+7
|\ \ | | | | | | | | | src: Fix grammar (SDU's -> SDUs)
| * | src: Fix grammar (SDU's -> SDUs)dimitri staessens2016-09-023-7/+7
| | |
* | | Merged in dstaesse/ouroboros/be-split-fp (pull request #238)Sander Vrijders2016-09-029-29/+70
|\| | | |/ |/| | | lib: Add northbound ringbuffers
| * lib: Add northbound ringbuffersdimitri staessens2016-09-029-29/+70
| | | | | | | | Fast path is split in north and southbound paths.
* | Merged in dstaesse/ouroboros/be-ipcp-dev (pull request #237)Sander Vrijders2016-09-024-36/+38
|\ \ | |/ |/| | | lib: ipcp: Decouple flow_dealloc
| * lib: ipcp: Decouple flow_deallocdimitri staessens2016-09-024-36/+38
| | | | | | | | Sending a dealloc to the irmd moved to its own call.
* | lib: Change rdrbuff API to use correct typesSander Vrijders2016-09-012-9/+9
|/ | | | | | 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.
* irmd: Fix registering AP-I before binddimitri staessens2016-09-011-2/+7
|
* lib, rdrbuff: Fix fast path lockupdimitri staessens2016-09-011-2/+5
| | | | Adds missing condition signals.
* lib, shm_rdrbuff: Fix preprocessor definedimitri staessens2016-08-311-15/+15
|
* irmd: Fix binding after registeringdimitri staessens2016-08-311-0/+5
| | | | | Flow will now allocate correctly when the name is registered in a DIF before it is bound to an AP.
* tools, irm: Fix segfault with missing parametersdimitri staessens2016-08-311-2/+2
|
* lib, ipcp: Compile on Apple junkdimitri staessens2016-08-307-35/+111
| | | | | | 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.
* lib: Refactor shm_du_map to shm_rdrbuffdimitri staessens2016-08-2911-865/+950
| | | | | | | | | | 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.
* lib, cdap: Fix message handlerdimitri staessens2016-08-291-1/+1
| | | | When getting a DELETE opcode, create was called instead of delete.
* cdap: Fix destructiondimitri staessens2016-08-271-6/+3
| | | | | Destroying NULL should succeed. Resources need to be freed even if flow_dealloc fails.
* irmd: Add state checks to api_entrydimitri staessens2016-08-272-2/+10
| | | | Prevents double destruction.
* ipcpd: Report pid on shutdowndimitri staessens2016-08-264-8/+8
| | | | IPCPs will report their pid on shutdown for faster debugging.
* irmd: Revise shutdown proceduredimitri staessens2016-08-261-13/+14
| | | | | Flows are deallocated first, then the socket is closed to prevent IPCPs hanging on the connect.