|
5b18f130
|
2011-08-24T16:17:05
|
|
Make rate limiting work with common_timeout logic
|
|
|
e7fe9270
|
2011-07-04T12:16:08
|
|
Merge remote-tracking branch 'github/20_global_locks_init' into patches-2.0
|
|
|
e9340964
|
2011-06-22T11:22:35
|
|
Fix AIX build issue with TAILQ_FOREACH definition
Reported by Lawnstein Chan.
|
|
|
b683cae3
|
2011-04-22T12:01:25
|
|
Avoid race-condition when initializing global locks
Previously, we did stuff like
if (!lock)
EVTHREAD_ALLOC_LOCK(lock,0);
for the evsig base global lock, the arc4random lock, and the debug_map
lock. But that's potentially racy! Instead, we move the
responisiblity for global lock initialization to the functions where
we set up the lock callbacks.
(Rationale: We already require that you set up the locking callbacks
before you create any event_base, and that you do so exatly once.)
|
|
|
a459ef70
|
2011-03-03T15:34:22
|
|
Have event_base_gettimeofday_cached() always return wall-clock time
Based on code by Dave Hart
|
|
c4dc3353
|
2010-11-23T00:22:01
|
|
Add some checks since lack of TAILQ_FOREACH doesn't imply lack of FIRST,
END, NEXT, or INSERT_BEFORE. Quiet some warnings in XL C.
|
|
f817bfa4
|
2010-10-27T17:31:52
|
|
Fix some ints to evutil_socket_t; make tests pass on win64.
|
|
|
38d09606
|
2010-09-15T12:50:31
|
|
Remove event_base.evsigbase; nothing used it.
|
|
|
3658b169
|
2010-09-08T14:12:12
|
|
Merge remote branch 'chrisd/iocp-fixes4'
Conflicts:
test/regress_thread.c
|
|
|
4632b78e
|
2010-09-08T13:22:55
|
|
Minimize calls to base_notify implementation functions, thereby avoiding needless syscalls
The trick here is that if we already told the base to wake up, and it
hasn't woken up yet, we don't need to tell it to wake up again. This
should help lots with inherently multithreaded code like IOCP.
|
|
2447fe88
|
2010-08-28T04:07:48
|
|
Add event_config_set_num_cpus_hint for tuning thread pools, etc.
|
|
|
76f7e7ae
|
2010-08-17T05:02:00
|
|
Some IOCP bufferevent tweaks.
- Increment reference count of bufferevents before initiating overlapped
operations to prevent the destructor from being called while operations
are pending. The only portable way of canceling overlapped ops is to
close the socket.
- Translate error codes to WSA* codes.
- Better handling of errors.
- Add an interface to add and del "virtual" events. Because IOCP
bufferevents don't register any events with the base, the event loop
has no way of knowing they exist. This causes the loop to terminate
prematurely. event_base_{add,del}_virtual increment/decrement base's
event count so the loop runs while there are any enabled IOCP
bufferevents.
|
|
|
e0972c21
|
2010-08-17T13:18:18
|
|
Use conditions instead of current_event_lock to fix a deadlock.
Avi Bab correctly noted as bug 3044479 the fact that any thread
blocking on current_event_lock will do so while holding
th_base_lock, making it impossible for the currently running event's
callback to call any other functions that require th_base_lock.
This patch switches the current_event_lock code to instead use a
condition variable that we wait on if we're trying to mess with
a currently-executing event, and that we signal when we're done
executing a callback if anybody is waiting on it.
|
|
60433a0a
|
2010-08-13T17:08:59
|
|
Clean up syntax on TAILQ_ENTRY() usage
Though the C standards allow it, it's apparently possible to get MSVC
upset by saying "struct { int field; } (declarator);" instead of
"struct {int field; } declarator;", so let's just not do that.
Bugfix for 3044492
(commit msg by nickm)
|
|
|
ec347b92
|
2010-07-07T16:45:03
|
|
Move event-config.h to include/event2
This change means that all required include files are in event2, and
all files not in event2/* are optional.
|
|
|
cb670740
|
2010-04-28T11:51:56
|
|
Make debug mode catch mixed ET and non-ET events on an fd
Of the backends that support edge-triggered IO, most (all?) do not
support attempts to mix edge-triggered and level-triggered IO on the
same FD. With debugging mode enabled, we now detect and refuse attempts
to add a level-triggered IO event to an fd that already has an
edge-triggered IO event, and vice versa.
|
|
|
b557b175
|
2010-03-21T13:28:48
|
|
Detect and refuse reentrant event_base_loop() calls
Calling event_base_loop on a base from inside a callback invoked by
that same base, or from two threads at once, has long been a way to
get exceedingly hard-to-diagnose errors. This patch adds code to
detect such reentrant invocatinos, and exit quickly with a warning
that should explain what went wrong.
|
|
|
cdd4c490
|
2010-03-11T00:38:46
|
|
Try to comment some of the event code more
|
|
|
17efc1cd
|
2010-03-04T01:25:51
|
|
Update all our copyright notices to say "2010"
|
|
|
23170a69
|
2010-02-25T16:57:57
|
|
Fix mingw compilation
|
|
|
4faeaea9
|
2010-02-19T03:39:50
|
|
Clean up formatting: function/keyword spacing consistency.
- Keywords always have a space before a paren. Functions never do.
- No more than 3 blank lines in a row.
|
|
|
e5bbd40a
|
2010-02-18T17:41:15
|
|
Clean up formatting: use tabs, not 8-spaces, to indent.
|
|
|
a66e947b
|
2010-01-25T13:44:56
|
|
Use less memory for each entry in a hashtable
Our hash-table implementation stored a copy of the hash code in each
element. But as we were using it, all of our hash codes were
ridiculously easy to calculate: most of them were just a matter of a
load and a shift.
This patch lets ht-internal be built in either of two ways: one caches
the hash-code for each element, and one recalculates it each time it's
needed.
This patch also chooses a slightly better hash code for
event_debug_entry.
|
|
|
cd17c3ac
|
2010-01-22T00:34:37
|
|
Add support for a "debug mode" to try to catch common errors.
Right now it only catches cases where we aren't initializing events,
or where we are re-initializing events without deleting them first.
These are however shockingly common.
|
|
|
27308aae
|
2010-01-14T16:30:40
|
|
Changelist code to defer event changes until just before dispatch
This is necessary or useful for a few reasons:
1) Sometimes applications will add and delete the same event more
than once between calls to dispatch. Processing these changes
immediately is needless, and potentially expensive (especially
if we're on a system that makes one syscall per changed event).
Yes, this actually happens in practice for nonpathological
code, such as in cases where the user's callback conditionally
re-adds a non-persistent event, or where draining a buffer
turns off writing and invokes a user callback which adds more
data which in turn re-enabled writing.
2) Sometimes we can coalesce multiple changes on the same fd into
a single syscall if we know about them in advance. For
example, epoll can do an add and a delete at the same time, but
only if we have found out about both of them before we tell
epoll.
3) Sometimes adding an event that we immediately delete can cause
unintended consequences: in kqueue, this makes pending events
get reported spuriously.
|
|
|
74871cac
|
2009-11-09T19:37:27
|
|
Change event_base.activequeues to "array of eventlist".
Previously, event_base.activequeues was of type "array of pointers to
eventlist." This was pointless: none of the eventlists were allowed
to be NULL. Worse, it was inefficient:
- It made looking up an active event queue take two pointer
deferences instead of one, thus risking extra cache misses.
- It used more RAM than it needed to, because of the extra pointer
and the malloc overhead.
Also, this patch fixes a bug where we were saying
calloc(N,N*sizeof(X)) instead of calloc(N,sizeof(X)) when allocating
activequeues. That part, I'll backport.
Also, we warn and return -1 on failure to allocate activequeues,
rather than calling event_err.
svn:r1525
|
|
|
ab96b5f3
|
2009-11-09T18:30:33
|
|
Add an option to disable the timeval cache.
svn:r1518
|
|
|
693c24ef
|
2009-11-09T17:16:30
|
|
Implement queued timeouts for case where many timeouts are the same.
Libevent's current timeout code is relatively optimized for the
randomly scattered timeout case, where events are added with their
timeouts in no particular order. We add and remove timeouts with
O(lg n) behavior.
Frequently, however, an application will want to have many timeouts
of the same value. For example, we might have 1000 bufferevents,
each with a 2 second timeout on reading or writing. If we knew this
were always the case, we could just put timeouts in a queue and get
O(1) add and remove behavior. Of course, a queue would give O(n)
performance for a scattered timeout pattern, so we don't want to
just switch the implementation.
This patch gives the user the ability to explicitly tag certain
timeout values as being "very common". These timeout values have a
cookie encoded in the high bits of their tv_usec field to indicate
which queue they belong on. The queues themselves are each
triggered by an entry in the minheap.
See the regress_main.c code for an example use.
svn:r1517
|
|
|
a2a7d1d1
|
2009-10-27T05:16:32
|
|
Do not call the locking variant of event_add or event_active in some cases when we know we have the lock.
svn:r1471
|
|
|
b06b2649
|
2009-07-26T01:29:39
|
|
Make "deferred callback queue" independent of event_base.
This way, we can more easily have an IOCP bufferevent implementation
that does not need an event_base at all. Woot.
svn:r1381
|
|
|
5d71b25b
|
2009-07-17T18:38:38
|
|
Remove all trailing whitespace from end-of-line.
svn:r1350
|
|
|
6b4b77a2
|
2009-07-14T16:54:48
|
|
Make event_del(E) block while E is running in another thread.
This gives you the property that once you have called event_del(E),
you know that E is no longer running or pending or active at all, and
so it is safe to delete the resource used by E's callback.
svn:r1341
|
|
|
b4886ec8
|
2009-05-15T18:44:44
|
|
Trim 22 bytes from struct event on 32 bit platforms, more on 64-bit platforms.
svn:r1292
|
|
|
fe47003d
|
2009-05-05T16:52:37
|
|
Make unit tests for bufferevent_async compile and _almost_ work.
Either I need to make the callbacks get deferred in a base with no events (doable), or I need to make it okay to call launch_read from inside the callback for read (tricky).
svn:r1277
|
|
|
b69d03b5
|
2009-05-05T14:18:14
|
|
Add a constructor for bufferevent_async.
svn:r1274
|
|
|
838d0a81
|
2009-04-17T06:55:08
|
|
Document many internal functions and pieces of code.
svn:r1181
|
|
|
4868f4d2
|
2009-04-10T14:22:33
|
|
Initial support for a lightweight 'deferred callbacks'.
A 'deferred callback' is just a function that we've queued in the
event base. This ability is needed for some mt stuff, and for complex
callback chains. For internal use only.
svn:r1150
|
|
|
ec35eb55
|
2009-02-12T22:19:54
|
|
Make threading functions global, like the mm_ functions. Use the libevent_pthread.la library in regress_pthread.
svn:r1121
|
|
|
b85b710c
|
2009-01-27T22:34:36
|
|
Update copyright statements to reflect the facts that:
a) this is 2009
b) niels and nick have been comaintainers for a while
c) saying "all rights reserved" when you then go on to explicitly
disclaim some rights is sheer cargo-cultism.
svn:r1065
|
|
|
8889a770
|
2009-01-27T22:30:46
|
|
Replace all use of config.h with event-config.h.
svn:r1064
|
|
56ea4687
|
2009-01-22T02:33:38
|
|
Change the semantics of timeouts in conjunction with EV_PERSIST; timeouts in that case will now repeat until deleted.
svn:r1032
|
|
|
a5901991
|
2009-01-19T20:37:24
|
|
Use eventfd for main-thread notification where available (i.e., linux).
svn:r1023
|
|
|
554e1493
|
2009-01-14T20:52:32
|
|
Move per-fd info from eventops into evmap. Not done for win32.c yet.
svn:r1008
|
|
|
169321c9
|
2009-01-13T20:26:37
|
|
Rename four internal headers to follow the -internal.h convention.
svn:r1000
|
|
|
55bcd7d2
|
2009-01-09T13:42:21
|
|
On win32, use a hashtable to map sockets to events rather than using an array.
svn:r988
|
|
|
30cba6d0
|
2008-12-25T09:22:13
|
|
we cannot realloc memory used by TAILQ; instead malloc each slot individually
svn:r977
|
|
|
d776f846
|
2008-12-23T22:23:37
|
|
deprecate the usage of signal_{add,del,set} and name it evsignal_{add,del,set} instead; move the old definitions to compat
svn:r973
|
|
|
02b2b4d1
|
2008-12-23T16:37:01
|
|
Restructure the event backends so that they do not need to keep track of events themselves, as a side effect multiple events can use the same fd or signal.
svn:r972
|
|
|
f7e61870
|
2008-07-11T15:49:04
|
|
support multiple events listening on the same signal; make signals regular events that go on the same event queue
svn:r901
|
|
|
05965921
|
2008-05-31T14:37:31
|
|
Add new functions to access backends by their features and to query the features of a backend.
svn:r842
|
|
|
3f56e364
|
2008-05-08T05:56:20
|
|
event_base_new_with_config() and related methods
svn:r789
|
|
|
8c750eaf
|
2008-05-03T21:37:33
|
|
separate signal events from io events
svn:r760
|
|
|
45e6fb0d
|
2008-05-03T18:23:44
|
|
cache clock_gettime/gettimeofday values in base
svn:r758
|
|
|
c182baca
|
2008-03-10T03:17:20
|
|
switch thread support so that locks get allocated as they are needed.
svn:r690
|
|
|
558de9b3
|
2008-03-02T21:18:33
|
|
Provide OpenSSL style support for multiple threads accessing the same event_base
svn:r684
|
|
|
03589ccb
|
2007-12-24T22:49:30
|
|
rollback r594: restructuring to make event activation independent.
changes are going to wait for api design
svn:r612
|
|
|
7aa845b7
|
2007-12-16T04:10:30
|
|
restructure the code to make event activation independent of regular event logic
svn:r594
|
|
|
fbe24f43
|
2007-12-09T05:07:20
|
|
remove obsoleted recalc code
svn:r581
|
|
|
2823cb05
|
2007-11-25T17:15:28
|
|
r14944@tombo: nickm | 2007-11-25 12:12:28 -0500
Make kqueue pass more unit tests.
svn:r544
|
|
|
7eb250e9
|
2007-11-25T17:14:19
|
|
r14939@tombo: nickm | 2007-11-25 11:59:26 -0500
New function event_set_mem_functions to replace internal calls to malloc, free, etc with a user-supplied functions.
svn:r541
|
|
|
88897852
|
2007-11-25T06:57:59
|
|
provide event_reinit() to reinitialized an event_base after fork - necessary for epoll/kqueue
svn:r539
|
|
|
1c23e219
|
2007-11-17T22:21:42
|
|
r14931@tombo: nickm | 2007-11-17 17:21:09 -0500
Patch from Scott Lamb: Implement event_{base_}loopbreak. Includes documentation and tests. From sf.net Feature Request 1826546.
svn:r535
|
|
|
f74e7258
|
2007-11-07T06:01:57
|
|
r16501@catbus: nickm | 2007-11-07 01:00:31 -0500
This is one of those patches which will either make matters far
simpler after the bugs shake out, or will get reverted pretty quick
once we realize that it is a stupid idea.
We now post-process the config.h file into a new event-config.h file,
whose macros are prefixed with _EVENT_ and which is thus safe for
headers to include. Using this, we can define replacement timeval
manipulation functions in evutil.h, and use them uniformly through our
code. We can also detect which headers are needful in event.h, and
include them as required.
This is also the perfect time to remove the long-deprecated acconfig.h
file, so that autoheader no longer warns.
Should resolve the following issues:
[ 1826530 ] Header files should have access to autoconf output.
[ 1826545 ] acconfig.h is deprecated.
[ 1826564 ] On some platforms, event.h can't be included alone.
svn:r492
|
|
|
30ae40cc
|
2007-11-03T18:04:53
|
|
switch timeouts to a min heap; from Maxim Yegorushkin
svn:r467
|
|
|
41b7cbc3
|
2007-03-10T06:37:53
|
|
more the signal base into the event base; this removes global state and makes signals
work better with threading; from Wouter Wijngaards
small fixes for kqueue and style by me
svn:r351
|
|
|
905ee67d
|
2005-02-22T15:47:53
|
|
provide maintainer mode in automake; put event_gotsig back into global
state; return proper error code
svn:r131
|
|
|
8773c4c9
|
2004-11-25T09:50:18
|
|
make libevent thread-safe; first cut
svn:r122
|