aboutsummaryrefslogtreecommitdiff
path: root/src/platform/windows
Commit message (Collapse)AuthorAge
* fixes #863 socket activation: for TCP and IPC (POSIX only)Garrett D'Amore2024-12-30
| | | | | | | | | This introduces a new option "NNG_OPT_LISTEN_FD", understood by TCP, TLS, and (on POSIX systems) IPC. This option is used to pass a file descriptor or handle (Windows) that is already listening (ready for ACCEPT to be called). For TCP and TLS, the socket must be of type AF_INET or AF_INET6, and for IPC it must be of type AF_UNIX.
* resolver: add some additional test coverageGarrett D'Amore2024-12-29
| | | | | | While here, remove some code paths that do not occur by definition. (For example, if the resolver succeeds, we will definitely have a valid set of addresses, but if it fails, we will definitely not.)
* resolver: use explicit resolver item provided by callerGarrett D'Amore2024-12-28
| | | | | | This avoids the need to perform multiple allocations for dialing, eliminating additional potential failures. Cancellation is also made simpler and more perfectly robust.
* windows tcp: fix compilation warningGarrett D'Amore2024-12-28
|
* tcp stream: add test for ESTATE getting port on unbound listenerGarrett D'Amore2024-12-28
|
* windows tcp: Lookup extended TCP function pointers at startupGarrett D'Amore2024-12-28
| | | | | | This avoids the need for a lock during listener or dialer initialization, and it avoids the need to carry these pointers on those objects. It also eliminates a potential failure case "post startup".
* tcp: flatten the listener implementationGarrett D'Amore2024-12-28
| | | | | | | | | | The endpoints both use a nesting level for some common code and some platform dependent code. But the common code isn't that much and we have similar patterns for e.g. IPC. This avoids a layer of indirection in the structs, and extra allocations. The payoff will be even larger for the dialers, but that is next. (Dialers are more complicated because of DNS.)
* udp: use nni_aio_startGarrett D'Amore2024-12-26
| | | | | | This also moves the close of the UDP socket later, to avoid a potential use after free while the aio's are still in-flight. Unfortunately we cannot unbind cleanly without a hard close.
* ipc: use nni_aio_startGarrett D'Amore2024-12-26
|
* resolver: use nni_aio_startGarrett D'Amore2024-12-26
|
* tcp: use nni_aio_startGarrett D'Amore2024-12-26
|
* pollers: use atomic bit masking operations to eliminate lockssGarrett D'Amore2024-12-22
| | | | This is done for kqueue and poll. Others coming soon.
* streams: add explicit stop functionsGarrett D'Amore2024-12-12
| | | | | | | | | | | | This allows us to explicitly stop streams, dialers, and listeners, before we start tearing down things. This hopefully will be useful in resolving use-after-free bugs in http, tls, and websockets. The new functions are not yet documented, but they are nng_stream_stop, nng_stream_dialer_stop, and nng_stream_listener_stop. They should be called after close, and before free. The close functions now close without blocking, but the stop function is allowed to block.
* windows ipc: significant refactorGarrett D'Amore2024-12-09
| | | | | | | | | | | This refactors a lot of the IPC code to hopefully address various hangs on shutdown, etc. The problem is that named pipes are not terrifically reliable when it comes to aborting ConnectNamedPipe. Additionally there were some logic errors in some of our code that left things rather brittle. Ultimately this all needs to be replaced with UNIX domain sockets which are superior in many ways.
* performance: reference counters can use relaxed order when incrementingGarrett D'Amore2024-12-07
|
* fini: add drain mechanism for aio, reap, and task subsystemsGarrett D'Amore2024-12-07
| | | | | Make sure *everything* is drained before proceeding all the way to deallocation.
* windows ipc: check the return value of nni_aio_scheduleGarrett D'Amore2024-12-03
| | | | | This hopefully addresses a long standing bug with Windows IPC sometimes hanging on shut down.
* Remove the NNG_OPT_IPC_SECURITY_DESCRIPTOR option.Garrett D'Amore2024-11-24
| | | | | | | | This is now replaced with nng_listener_set_security_descriptor and nng_stream_listener_set_security_descriptor functions. We may elect to remove these entirely, but for named pipe users they are probably still quite useful. Moving towards UNIX domain sockets would obsolete this functionality.
* ipc: Add NNG_OPT_REMADDR for dialer.Garrett D'Amore2024-11-23
|
* Use 32-bit port numbers.Garrett D'Amore2024-11-21
| | | | | While TCP and UDP port numbers are 16-bits, ZT uses a larger (24-bit) port number.
* URL refactor part 1.Garrett D'Amore2024-11-18
| | | | | | | This eliminates most (but not all) of the dynamic allocations associated with URL objects. A number of convenience fields on the URL are removed, but we are able to use common buffer for most of the details.
* URL u_port should be a number not a string.Garrett D'Amore2024-11-17
| | | | | | | | | | | The idea here is to reduce the dynamic allocations used for URLs, and also the back and forth with parsing begin strings and port numbers. We always resolve to a port number, and this is easier for everyone. The real goal in the long term is to eliminate dynamic allocation of the URL fields altogether, but that requires a little more work. This is a step in the right direction.
* refactor initialization/finalizationGarrett D'Amore2024-11-11
| | | | | | | | | Applications must now call nng_init(), but they can supply a set of parameters optionally. The code is now safe for multiple libraries to do this concurrently, meaning nng_fini no longer can race against another instance starting up. The nni_init checks on all public APIs are removed now.
* remove nni_plat_println - we have nni_plat_printf which is good enoughGarrett D'Amore2024-11-05
|
* Remove untyped nng_pipe_get, support for untyped sockaddr set.Garrett D'Amore2024-11-03
| | | | This is a step on the path to removing unsafe untyped option accesses.
* Remove nng_stream_set function family.Garrett D'Amore2024-10-28
| | | | This was not really used or useful.
* remove legacy support for configuration of transport options on socketGarrett D'Amore2024-10-20
| | | | | | | | | Transport specific options should be configured on the end point. This has the most impact for TLS, as TLS dialers and listeners will need to be allocated apriori, to configure TLS options. Some legacy tests were removed... we're going to remove the legacy libnanomsg compatibility layer anyway.
* fix maybe uninitialized warningshikokuchuo2024-10-05
|
* fix int-to-pointer conversion warningshikokuchuo2024-10-05
|
* fixes #1837 IPC - Use After FreeGarrett D'Amore2024-08-13
| | | | This fixes a problem only found on Windows, that affected both IPC and TCP.
* UDP: Introduce an experimental (undocumented for now) public API for UDP. ↵Garrett D'Amore2024-06-02
| | | | | | | | (#1838) This exposes the UDP methods as nng_ methods, and adds support for Multicast Membership, which is useful in a variety of situations. No documentation is provided, and applications should consider thios API experimental.
* windows: refactor IPC to avoid possible hangs and use after free.Garrett D'Amore2024-05-30
|
* windows: TCP connection use-after-free fixesGarrett D'Amore2024-05-30
|
* windows: ipc listen use after freeGarrett D'Amore2024-05-30
|
* windows: ipc conn conn_aio not usedGarrett D'Amore2024-05-30
|
* windows: fix TCP use-after-free in listenerGarrett D'Amore2024-05-30
| | | | | | | The logic with overlapped structures was fragile as it used overlapped ios for the connections rather than a single common one for the listener. This changes it to be more like POSIX, and robust against this error.
* windows: drop the hEvent initialization for win_io structures.Garrett D'Amore2024-05-30
| | | | We use overlapped I/O, so we don't need a separate hEvent.
* windows: IPC conn->conn_io is unusedGarrett D'Amore2024-05-30
|
* fixes #1825 Compiler warning due to unused variable in win_clock.cGarrett D'Amore2024-05-22
|
* Windows: add a check for broken legacy environments without timespec_get.Garrett D'Amore2024-05-22
| | | | | | | | | | If you want to build on Windows use a toolchain that supports modern APIs. This means, for Microsoft, UCRT (Universal C Runtime), which is supported by default on modern Visual Studio. MinGW users may have to go out of their way to enable it. (New -D_UCRT flag or something.) The supported toolchain for building on Windows is Visual Studio. Use of other tool chains is not officially supported or guaranteed to work. YMMV.
* Added realtime clock support, and better timestamping.Garrett D'Amore2024-04-14
| | | | | The realtime clock is not (yet) exposed for user applications, but it is used for logging timestamps accurately.
* fixes #1496 Provide NNG_ENABLE_IPV6 option (disabled by default)Garrett D'Amore2024-02-25
| | | | | | | | | This also checks if the build system has the definitions for AF_INET6, which might help in some embedded IPv4 only settings. The resolver test is enhanced to include a check for IPv6 enabled in the kernel. IPv6 support is enabled by default, of course.
* fixes #1543 Deadlock in nng_close(socket)Garrett D'Amore2024-02-25
| | | | | | This looks like a possible problem that may be windows specific involving the flow for IO completion ports. This simplifies the logic a little bit, and should ensure that canceled requests on pipes do not restart.
* windows: Add missing header for string.hGarrett D'Amore2024-02-25
|
* Fix warning from inconsistent socket_pair prototype.Garrett D'Amore2024-01-06
|
* fixes #1572 nng creates too many threadsGarrett D'Amore2024-01-01
| | | | | | | | | | | | This further limits some of the thread counts, but principally it offers a new runtime facility, nng_init_set_parameter(), which can be used to set certain runtime parameters on the number of threads, provided it is called before the rest of application start up. This facility is quite intentionally "undocumented", at least for now, as we want to limit our commitment to it. Still this should be helpful for applications that need to reduce the number of threads that are created.
* provide NNG_MAX_POLLER_THREADSGarrett D'Amore2023-12-29
| | | | | | | This (defaults to 8) sets a limit on the number of poller threads that will be used for servicing I/Os. This is the size of the I/O completion port thread pool on Windows. POSIX pollers are generally not concurrent at present.
* fixes #1746 Create a new socket:// transport for socketpair() based connectionsGarrett D'Amore2023-12-29
| | | | | | | | | | | | | This transport only listens, and creates connections when the application calls setopt on the lister with NNG_OPT_SOCKET_FD, to pass a file descriptor. The FD is turned into an nng_stream, and utilized for SP. The protocol over the descriptor is identical to the TCP protocol (not the IPC protocol). The options for peer information are borrowed from the IPC transport, as they may be useful for these purposes. This includes a test suite and full documentation.
* Update win_tcpconn.cNissimBendanan2023-09-14
| | | | | fix compile error in nano using clang version 16.0.5 fix issue: https://github.com/nanomsg/nng/issues/1676
* Fixes the call to setsockopt(SO_UPDATE_ACCEPT_CONTEXT) on Windows (#1604)Ruben Valls2022-07-15
| | | From the Windows docs, this option must receive as parameters the listener socket.
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2026-01-18 18:10:09 +0000