summaryrefslogtreecommitdiff
path: root/src/core/socket.c
Commit message (Collapse)AuthorAge
* Add nng_sub0_subscribe and friends.Garrett D'Amore2024-11-11
| | | | | | These are new functions that replace `NNG_OPT_SUBSCRIBE` and `NNG_OPT_UNSUBSCRIBE`. They are provided here as a transition aid before those options are removed in NNG 2.0.
* Add new functions for socket identity.Garrett D'Amore2024-11-11
| | | | | | | | | | These functions were added in the main branch for NNG 2.0, and we add them here to facilitate converting any code that happened to use the options (`NNG_OPT_PROTO`, `NNG_OPT_PROTONAME`, `NNG_OPT_PEER`, `NNG_OPT_PEERNAME`, and `NNG_OPT_RAW`. Its unlikely that many applications need this, but providing these as a transition aid may help applications start adopting new APIs now.
* Integer sign extension bug in socket.c.Garrett D'Amore2024-11-08
| | | | | | | | It turns out that for now this results in early wakeups, due to another bug in the aio framework. But when that bug is fixed, this bug will lead to hangs when redialing. (cherry picked from commit 2dfb99506142f2d59bcc0e0fa7db6b19a3c75d43)
* fixes #1835 setting NNG_OPT_TLS_CONFIG on socket should hold referenceGarrett D'Amore2024-07-21
|
* fixes #1827 Windows a deadlock on nng_close()Garrett D'Amore2024-05-30
|
* Logging improvements (#1816)Garrett D'Amore2024-04-21
| | | | | | | | | | | | * Add nng_str_sockaddr to get string representation of socket address. * Added nng_log_get_level() is meant to allow users to obtain the current level and avoid some possibly expensive operations just to collect debugging information when debugging is not in effect. We use a custom logger for NUTS, and this fits within the NUTS test framework well, so that if -v is supplied we get more content. All tests now get this by default.
* Log rejects by app, bind/listen/connect failures.Garrett D'Amore2024-04-14
|
* fixes #1740 Public ID hash APIGarrett D'Amore2023-12-29
| | | | | | | | | This includes a manual page documenting the entire set of functions in one step. The hash is 64-bit based for now, to be maximally flexible. An internal 32-bit convenience for the common internal use is also provided (not public). The public API includes a test suite.
* fixes #1658 Possible use-after-free in dialerGarrett D'Amore2023-04-23
|
* fixes #1657 Use after free in listener (data race)Garrett D'Amore2023-04-23
|
* fixes #1610 nni_sock_open() in socket.c: use after freeGarrett D'Amore2023-04-19
|
* Socket close may leak messages.Garrett D'Amore2023-02-07
| | | | | | We try to move the msgq close up earler. While here we can stop dropping and reacquiring the lock -- this is likely left over and may lead to races.
* Fixes compiling when NNG_ENABLE_STATS is OFF (#1600)Ruben Valls2022-06-20
| | | The socket member `st_name` is not available when NNG_ENABLE_STATS is OFF.
* fixes #1577 nng_stat_find_socket always returns the same stats nodeGarrett D'Amore2022-04-16
| | | | | Also this fixes problems with uninitialized socket names, and the socket name stat not being adjusted correctly when set via API.
* Introduce nng_device_aio().Garrett D'Amore2021-12-27
| | | | | | This function is like nng_device(), but runs asynchronously. Also, this fixes #1503 nng_device causes nng_close to blocking
* Socket and context initialization never fails.Garrett D'Amore2021-12-25
| | | | | This makes these functions entirely bullet proof, and eliminates yet more error handling cases.
* ID map static initialization support.Garrett D'Amore2021-12-06
| | | | This also arranges to clean up the maps at nng_fini time.
* Use static initialization for lists and mutexes.Garrett D'Amore2021-12-05
| | | | | | This eliminates some run-time initialization, moving it to compile time. Additional follow up work will expand on this to simplify initialization and reduce the need for certain locks.
* Fix incorrect elides of deprecated code.Garrett D'Amore2021-11-29
| | | | | | (Note that this includes some incorrect formatting due to an apparent bug in clang-format 13. Hopefully they'll fix it later, but for now I'm preserving the bad whitespace.)
* Eliminate the pipe mutex and use atomic for pipe closed.Garrett D'Amore2021-09-06
| | | | | | | | | | This eliminates several mutex operations done each time a pipe is created or destroyed. For large scale systems this should reduce overall pressure on the memory subsystem, and scale better as many threads are coming and going. This also reduces the overall size of nni_pipe -- on Linux by 36 bytes typically.
* SP initialization cannot fail.Garrett D'Amore2021-09-06
|
* fixes #1498 Endpoint close/shutdown could be synchronous (#1499)Garrett D'Amore2021-09-04
|
* Several minor cleanups. Fix socket id stat for listener.v1.5.0Garrett D'Amore2021-07-09
|
* Fix leaking pipes after close.Garrett D'Amore2021-07-07
|
* fixes #1456 bad access in OSX thread on nn (#1457)Garrett D'Amore2021-07-01
| | | | * fixes #1456 bad access in OSX thread on nn * Fix broken aio in darwin cloud.
* fixes #1372 nni_reap could be smallerGarrett D'Amore2020-12-19
|
* fixes #1071 tran_chkopt can be cleaned upGarrett D'Amore2020-11-15
| | | | | | | | | | | | | | | | | This is a sweeping cleanup of the transport logic around options, and also harmonizes the names used when setting or getting options. Additionally, legacy methods are now moved into a separate file and can be elided via CMake or a preprocessor define. Fundamentally, the ability to set to transport options via the socket is deprecated; there are numerous problems with this and my earlier approaches to deal with this have been somewhat misguided. Further these approaches will not work with future protocol work that is planned (were some options need to be negotiated with peers at the time of connection establishment.) Documentation has been updated to reflect this. The test suites still make rather broad use of the older APIs, and will be converted later.
* *Fix Add missing #define when disabling NNG_ENABLE_STATS (#1342)JaylinYu2020-11-13
| | | Co-authored-by: Jaylin <oblivionangel@sina.com>
* fixes #1323 stats framework is *way* to heavyGarrett D'Amore2020-11-11
| | | | | | This should reduce the amount of copying, and the overall size used by pipes and other objects quite a bit. (On my system, the sizeof nni_pipe shrank by 400 bytes, for example.)
* fixes #1289 zerotier should have it's own copy of the id hashing codeGarrett D'Amore2020-08-16
| | | | | | | | | | | fixes #1288 id allocation can overallocate fixes #1126 consider removing lock from idhash This substantially refactors the id hash code, giving a cleaner API, and eliminating a extra locking as well as some wasteful allocations. The ZeroTier code has it's own copy, that is 64-bit friendly, as the rest of the consumers need only a simpler 32-bit API.
* fixes #1230 NNG_OPT_RECONNMAXT zero does not prevent exponential back-offGarrett D'Amore2020-05-24
|
* fixes #1104 move allocation of protocol objects to common coreGarrett D'Amore2020-01-03
| | | | fixes #1103 respondent could inline backtrace
* fixes #1064 Potential deadlock in statistics codeGarrett D'Amore2019-12-29
| | | | | | | | fixes #1063 Include sanitizer runs in CI fixes #1068 Wssfile test sometimes fails with wrong error code While here, addressed a number of clang-tidy items, and some light cleanup of code we were already in.
* fixes #954 large message fails with no error messageGarrett D'Amore2019-07-17
| | | | | | This removes the default 1MB limit on maximum receive sizes. Applications intended for deployment in insecure or hostile environments should choose a sensible default for NNG_OPT_RECVMAXSZ.
* fixes #815 Eliminate socket filters on message queuesGarrett D'Amore2019-03-13
| | | | | | | This also eliminates the enforcement of NNG_OPT_RECVMAXSZ for inproc, which never really made much sense. This helps inproc go faster. While here, also clean up the entry point for protocols to support a drain option, since we don't use that anywhere.
* fixes #848 server hang waiting for client handshakeGarrett D'Amore2019-02-23
| | | | | | | | | | | | | | | | | | | | | | | fixes #698 Need TCP stats fixes #699 Need IPC stats fixes #701 Need TLS stats This commit addresses a problem when negotiating using one of the stream based negotiation APIs -- a slow or misbehaving peer can prevent well behaved ones from establishing a connection. The fix is a fairly significant change in how these transports link up, and it does rely on the fact that the socket only has a single accept() or connect() pending at a time (on a given endpoint that is). While here, we have completely revamped the way transport statistics are done, offering a standard API for collecting these statistics. Unfortunately, this completely borks the statistics for inproc. As we are planning to change the way inproc works soon, in order to provide more control and work on performance fixes for the message queue, we feel this is an acceptable trade-off. Furthermore, almost nobody uses inproc for anything, and even fewer people are making use of the statistics at this time.
* IPC option rework (pipe/conn) to reduce code duplication.Garrett D'Amore2018-12-29
|
* fixes #831 Unify option structures, o_type is unusedGarrett D'Amore2018-12-29
|
* fixes #782 stat namespace collision on QNXGarrett D'Amore2018-11-09
|
* fixes #4 Statistics supportGarrett D'Amore2018-09-03
| | | | | | | | | | | | | | | | This introduces new public APIs for obtaining statistics, and adds some generic stats for dialers, listeners, pipes, and sockets. Also added are stats for inproc and pairv1 protocol. The other protocols and transports will have stats added incrementally as time goes on. A simple test program, and man pages are provided for this. Start by looking at nng_stat(5). Statistics does have some impact, and they can be disabled by using the advanced NNG_ENABLE_STATS (setting it to OFF, it's ON by default) if you need to build a minimized configuration.
* fixes #693 Failed dial results in hard spinGarrett D'Amore2018-08-31
|
* fixes #648 REQ protocol can hang on closeGarrett D'Amore2018-08-14
| | | | | | | | | | | Actually the problem was in socket core, in particular in the shutdown code. The socket shutdown is supposed to ensure that no pipes were present on the socket, so that protocols need not concern themselves with this. The code unfortunately was busted, due to an ordering problem compounded by a race condition. This fixes that, and changes the REQ protocol to avoid the blocking condition altogether, and sprinkles a few assertions to validate these rules are being adhered to.
* fixes #208 pipe start should occur before connect / acceptGarrett D'Amore2018-08-14
| | | | | | | | | | fixes #599 nng_dial sync should not return until added to socket This reintroduces the changes for the above fixes, building upon the transport modifications that we have made to eliminate the separate transport pipe start entry point. It also includes slightly reworked code during start to put a hold on the pipe when it is created, which we we drop at the end, hopefully fixing a use-after-free.
* Revert "fixes #599 nng_dial sync should not return until added to socket"Garrett D'Amore2018-08-06
| | | | | This changeset needs work. We are seeing errors described by This reverts commit d7f7c896c0ede24249ef63b1e45b1878bf4bd473.
* fixes #599 nng_dial sync should not return until added to socketGarrett D'Amore2018-08-05
| | | | | | | | | | fixes #208 pipe start should occur before connect / accept fixes #616 Race condition closing between header & body This refactors the transports to handle their own connection handshaking before passing the pipe to the socket. This changes and simplifies the setup. This also fixes a rather challenging race condition described by #616.
* fixes #601 pipe destroy can fail to close pipeGarrett D'Amore2018-07-18
|
* fixes #589 tsan found racesGarrett D'Amore2018-07-16
|
* fixes #568 Want a single reader/write lock on socket child objectsGarrett D'Amore2018-07-06
| | | | | | | | | | | | | | | | | | | | | | | | | | | fixes #170 Make more use of reaper This is a complete restructure/rethink of how child objects interact with the socket. (This also backs out #576 as it turns out not to be needed.) While 568 says reader/writer lock, for now we have settled for a single writer lock. Its likely that this is sufficient. Essentially we use the single socket lock to guard lists of the socket children. We also use deferred deletion in the idhash to facilitate teardown, which means endpoint closes are no longer synchronous. We use the reaper to clean up objects when the reference count drops to zero. We make a special exception for pipes, since they really are not reference counted by their parents, and they are leaf objects anyway. We believe this addresses the main outstanding race conditions in a much more correct and holistic way. Note that endpoint shutdown is a little tricky, as it makes use of atomic flags to guard against double entry, and against recursive lock entry. This is something that would be nice to make a bit more obvious, but what we have is safe, and the complexity is at least confined to one place.
* fixes #572 Several locking errors foundGarrett D'Amore2018-07-03
| | | | | | | | | | fixes #573 atomic flags could help This introduces a new atomic flag, and reduces some of the global locking. The lock refactoring work is not yet complete, but this is a positive step forward, and should help with certain things. While here we also fixed a compile warning due to incorrect types.
* fixes #564 Race regression caused by#522Garrett D'Amore2018-07-02
| | | | | | fixes #565 Option getting should validate sizes more aggressively fixes #563 Reconnect timeouts should be settable on dialers fixes #562 pipe test is fragile
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-01-18 15:58:01 +0000