| Commit message (Collapse) | Author | Age |
|---|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #360 core should nng_aio_begin before nng_aio_finish_error
fixes #361 nng_send_aio should check for NULL message
fixes #362 nni_msgq does not signal pollable on certain events
This adds support for contexts for both sides of the surveyor pattern.
Prior to this commit, the raw mode was completely broken, and there
were numerous other bugs found and fixed. This integration includes
*much* deeper validation of this pattern.
Some changes to the core and other patterns have been made, where it
was obvioius that we could make such improvements. (The obviousness
stemming from the fact that RESPONDENT in particular is very closely
derived from REP.)
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This closes a fundamental flaw in the way aio structures were
handled. In paticular, aio expiration could race ahead, and
fire before the aio was properly registered by the provider.
This ultimately led to the possibility of duplicate completions
on the same aio.
The solution involved breaking up nni_aio_start into two functions.
nni_aio_begin (which can be run outside of external locks) simply
validates that nni_aio_fini() has not been called, and clears certain
fields in the aio to make it ready for use by the provider.
nni_aio_schedule does the work to register the aio with the expiration
thread, and should only be called when the aio is actually scheduled
for asynchronous completion. nni_aio_schedule_verify does the same thing,
but returns NNG_ETIMEDOUT if the aio has a zero length timeout.
This change has a small negative performance impact. We have plans to
rectify that by converting nni_aio_begin to use a locklesss flag for
the aio->a_fini bit.
While we were here, we fixed some error paths in the POSIX subsystem,
which would have returned incorrect error codes, and we made some
optmizations in the message queues to reduce conditionals while holding
locks in the hot code path.
|
| |
|
|
|
|
|
|
|
| |
While here I've added some code that should help us backtrack on
a crash, by linking back to the pipe from the context when we
are queued on that pipes sendq.
I'm not sure if we've ever seen these or not, but it could explain
certain infrequent crashes we think we've seen.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
This includes the test from legacy libnanomsg and a man page.
We have refactored the message queue notification system so
that it uses nni_pollable, leading we hope to a more consistent
system, and reducing the code size and complexity.
We also fixed the size of the NN_RCVFD and NN_SNDFD so that they
are a SOCKET on Windows systems, rather than an integer. This
addresses 64-bit compilation problems.
|
| | |
|
| |
|
|
|
|
|
| |
* 127.0.0.1.32 is treated as a hostname, returns EAI_NODATA on my fedora 27 box
* since this is not in POSIX, and deprecated from some libc resolvers
protect with an ifdef
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Ultimately, this just removes the support for lingering altogether.
Based on prior experience, lingering has always been unreliable, and
was removed in legacy libnanomsg ages ago.
The problem is that operating system support for lingering is very
inconsistent at best, and for some transports the very concept is somewhat
meaningless.
Making things worse, we were never able to adequately capture an exit()
event from another thread -- so lingering was always a false promise.
Applications that need to be sure that messages are delivered should
either include an ack in their protocol, use req/rep (which has an ack),
or inject a suitable delay of their own.
For things going over local networks, an extra delay of 100 msec should
be sufficient *most of the time*.
|
| | |
|
| | |
|
| |
|
|
|
| |
Applications which may not even be using TLS should not have to
know about the TLS configuration structure.
|
| | |
|
| |
|
|
|
|
|
| |
This technically doesn't need the lock, as we're only trying to catch
a possible problem during development rather than in the field (and
this should never occur), but the false positive in valgrind obscures
other possible errors, so leave it in the lock.
|
| |
|
|
| |
Turns out that shutdown is sufficient for most needs.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This provides context support for REQ and REP sockets.
More discussion around this is in the issue itself.
Optionally we would like to extend this to the surveyor pattern.
Note that we specifically do not support pollable descriptors
for non-default contexts, and the results of using file descriptors
for polling (NNG_OPT_SENDFD and NNG_OPT_RECVFD) is undefined.
In the future, it might be nice to figure out how to factor in
optional use of a message queue for users who want more buffering,
but we think there is little need for this with cooked mode.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
This makes the raw mode something that is immutable, determined
at socket construction. This is an enabling change for the
separate context support coming soon.
As a result, this is an API breaking change for users of the raw
mode option (NNG_OPT_RAW). There aren't many of them out there.
Cooked mode is entirely unaffected.
There are changes to tests and documentation included.
|
| | |
|
| |
|
|
|
|
|
|
| |
fixes #325 synchronous aio completion crash
fixes #327 move nni_clock() operations to outside the nni_aio_lk.
This work was done for the context tree, and is necessary to properly
enable that branch.
|
| | |
|
| | |
|
| |
|
|
|
| |
This uses numeric identifiers and an idhash table to make sure that
the values we get back are still use.
|
| |
|
|
|
| |
This also gives a performance benefit to WebSocket, by making
the completion logic run synchronously.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #22 Consider using synchronous completions sometimes
Transport improvements for IPC, TCP, and TLS.
This change does three things.
First it permits multiple outstanding receives or sends on the transport.
This change is being made to accomodate some other changes in the protocols
where it might be advantageous to post send or receives directly against
the transport pipe without going through another level of indirection.
Second, it changes the normal completions to be performed synchronously.
This translates into a rather major performance improvement, reducing
latency by some 27%, and thereby improving performance altogether. (This
elminates two extra context switches per transaction!)
FInally, we can save some extra checks and conditions because we know
that completions cannot happen if we don't have a pending operation
(we no longer complete out of sequence), and we only call the dosend
operation when we have something to send. This can eliminate some
pipeline stalls.
|
| | |
|
| | |
|
| |
|
|
|
|
| |
This includes platform portability stuff like mutexes and threads,
etc. While here we fixed a problem with nng_opts_parse and we
addressed the type of the platform API.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
fixes #302 nng_dialer/listener/pipe_getopt_sockaddr desired
This adds plumbing to pass and check the type of options
all the way through.
NNG_ZT_OPT_ORBIT is type UINT64, but you can use the untyped form to
pass two of them if needed.
No typed access for retrieving strings yet. I think this should allocate
a pointer and copy that out, but that's for later.
|
| |
|
|
|
|
|
| |
on a python wrapper (cffi).
Mostly this is fixing inconsistencies in our public API and the actual
implementation.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
fixes #275 nng_pipe_getopt_ptr() missing?
fixes #285 nng_setopt_ptr MIS
fixes #297 nng_listener/dialer_close does not validate mode
This change adds some missing APIs, and changes others.
In particular, certain options are now of type bool, with size
of just one. This is a *breaking* change for code that uses those
options -- NNG_OPT_RAW, NNG_OPT_PAIR1_POLY, NNG_OPT_TLS_VERIFIED.
|
| |
|
|
|
| |
While here we documented that certain options are not supported in the
compatibility layer.
|
| |
|
|
| |
fixes #33
|
| |
|
|
| |
fixes #290 sockaddr improvements
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
The first problem was that using nng_sleep_aio
was found to reset the timeout, and this caused subsequent
operations to start failing with timeouts when reusing
the AIO for other operations.
The second thing is that we think it would be nicer if the
presence of real aio timeouts were still honored, so that
if the timeout is shorter than the sleep time, then we get
back an NNG_ETIMEDOUT like every other operation, and we
get back a 0 if the logical sleep operation completes
normally.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This does a few things. First it closes a case where a dropped
message could prevent subsequent connection attempts from getting through.
Second, it changes the rate at which we retry, and the timeout, to be
a lot more aggressive when attempting to establish a connection. We
retry every 500 ms, for up to 2 minutes, before giving up. This gives
a lot more resilience in the face of message loss that is typical of
ZeroTier in some environments when first establishing communication.
Third, makes the values for the connection attempts *tunable*, so
that applications can adjust for different deployment scenarios.
Fourth, it includes the ability to get the UDP socket name. This was
needed during some debug, and may be useful for a real UDP transport
later, so we're keeping it.
Finally, we added documentation for the above items.
|
| | |
|
| |
|
|
| |
fixes #267 zerotier transport should lock ZT_HOME
|
| |
|
|
|
|
|
| |
The dev branch changed, and the timestamps are signed. (They
are also passed around using volatiles -- probably incorrectly
so, but that is not important for us, the compiler understands
that volatile is meaningless when passing values by value.)
|
| |
|
|
|
|
|
|
| |
The primary ZeroTier RFC is in the "rfcs" repository for nanomsg,
and we should remove this one because it is stale and misleading.
Also, the URI format is incorrect in the header so fix that, and
point to the real ZeroTier RFC.
|
