| Commit message (Collapse) | Author | Age |
|---|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This allows greatly increased scalability for kqueue based systems
with lots of cores (more likely FreeBSD than Darwin, as most macs
only have a smattering of cores), but even for macs we can engage
a few cores for system calls giving improvements.
The implementation here is pretty simple -- each file descriptor
gets assigned to its own kqueue by taking the numeric value of the
file descriptor modulo the number of kqueues we have opened.
The same approach will be adopted for epoll and Solaris/illumos port events.
|
| |
|
|
| |
This is done for kqueue and poll. Others coming soon.
|
| |
|
|
|
|
|
| |
This change moves the posix pollers to inline the PFD and makes
the callbacks constant, so that we can dispense with tests, failures,
and locks. It is anticipated that this will reduce lock based
pressure on the bus and increase performance modestly.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
The poller selection in the previous poller changes for select were
not quite functional. Also, while testing poll() based poller, there
were problems where it simply did not work correctly, so this addresses
those, and it seems to work now.
The pfd structures are exposed as we intend to allow inlining them
to eliminate the separate allocation and potential for failure during
initialization. We also want to have plans afoot to eliminate a
lot of the extra locking done done on each I/O iteration, and this
is setting the foundation for that.
|
| |
|
|
|
| |
This should reduce lock pressure during I/O for FreeBSD and macOS,
and should provide a small performance benefit.
|
| |
|
|
|
| |
Some platforms or configurations may not have more modern options
like kqueue or epoll, or may be constrained by policy.
|
| |
|
|
|
|
|
|
|
| |
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.
|
| |
|
|
|
|
| |
This also exposes an nng_thread_set_name() function for
applications to use. All NNG thread names start with "nng:".
Note that support is highly dependent on the operating system.
|
| |
|
|
|
|
|
|
|
|
| |
This was responsible for hangs in close on FreeBSD. Apparently
our use of EVFILT_USER was incorrect, and rather than fix it, we
have switched to using a notification pipe for synchronizing
closing pipes. In addition to fixing this problem, it should
significantly improve things for NetBSD and OpenBSD, which will
now be able tbenefit from kqueue(), since we no longer depend on
EVFILT_USER.
|
| |
|
|
|
|
|
|
|
|
|
| |
We reap the connections when closing, to ensure that the clean up is
done outside the pollq thread. This also reduces pressure on the
pollq, we think. But more importantly it eliminates some complex
code that was meant to avoid deadlocks, but ultimately created other
use-after-free challenges. This work is an enabler for further
simplifications in the aio/task logic. While here we converted some
potentially racy locking of the dialers and reference counts to simpler
lock-free reference counting.
|
| | |
|
| |
|
|
|
|
|
|
|
| |
Basically, we can ignore EV_EOF, as we wind up still alerting
the corresponding events. EV_ERROR we still treat as HUP.
(The EV_EOF was responsible for prematurely closing the socket
and aborting transactions while there was still data in the socket
buffers.)
|
| |
|
|
|
| |
This sets the kqueue events to autoclear, reducing CPU usage to normal
sane levels, and eliminating the hard spin.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #490 posix_epdesc use-after-free bug
fixes #489 Sanitizer based testing would help
fixes #492 Numerous memory leaks found with sanitizer
This introduces support for compiler-based sanitizers when using
clang or gcc (and not on Windows). See NNG_SANITIZER for possible
settings such as "thread" or "address".
Furthermore, we have fixed the issues we found with both the
thread and address sanitizers. We believe that the thread issues
pointed to a low frequency use-after-free responsible for rare
crashes in some of the tests.
The tests generally have their timeouts doubled when running under
a sanitizer, to account for the extra long times that the sanitizer
can cause these to take.
While here, we also changed the compat_ws test to avoid a particularly
painful and time consuming DNS lookup, and we made the nngcat_unlimited
test a bit more robust by waiting before sending traffic.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #326 consider nni_taskq_exec_synch()
fixes #410 kqueue implementation could be smarter
fixes #411 epoll_implementation could be smarter
fixes #426 synchronous completion can lead to panic
fixes #421 pipe close race condition/duplicate destroy
This is a major refactoring of two significant parts of the code base,
which are closely interrelated.
First the aio and taskq framework have undergone a number of simplifications,
and improvements. We have ditched a few parts of the internal API (for
example tasks no longer support cancellation) that weren't terribly useful
but added a lot of complexity, and we've made aio_schedule something that
now checks for cancellation or other "premature" completions. The
aio framework now uses the tasks more tightly, so that aio wait can
devolve into just nni_task_wait(). We did have to add a "task_prep()"
step to prevent race conditions.
Second, the entire POSIX poller framework has been simplified, and made
more robust, and more scalable. There were some fairly inherent race
conditions around the shutdown/close code, where we *thought* we were
synchronizing against the other thread, but weren't doing so adequately.
With a cleaner design, we've been able to tighten up the implementation
to remove these race conditions, while substantially reducing the chance
for lock contention, thereby improving scalability. The illumos poller
also got a performance boost by polling for multiple events.
In highly "busy" systems, we expect to see vast reductions in lock
contention, and therefore greater scalability, in addition to overall
improved reliability.
One area where we currently can do better is that there is still only
a single poller thread run. Scaling this out is a task that has to be done
differently for each poller, and carefuly to ensure that close conditions
are safe on all pollers, and that no chance for deadlock/livelock waiting
for pfd finalizers can occur.
|
| |
|
|
|
|
|
|
| |
This replaces the epoll support with proper illumos/SunOS port
events. The port event support is structured so that it actually
is superior to epoll and kqueue, because it avoids a single master
lock on the poller. In the future we will explore this for macOS
and Linux pollers.
|
| | |
|
| | |
|
| |
|
