summaryrefslogtreecommitdiff
path: root/src/socketengines/socketengine_epoll.cpp
AgeCommit message (Collapse)Author
2017-10-22Deduplicate error handling in the socket engines.Peter Powell
2017-10-22Convert GetMaxFds() to size_t and deduplicate setting code.Peter Powell
2017-07-12Merge pull request #1337 from SaberUK/master+mergePeter Powell
Merge v2.0.23 and v2.0.24 into master.
2017-07-11Improve and centralize socket engine event counters.Robin Burchell
The write counters were close to useless because they were only incremented on a write "event" which is only triggered when writing would block. Read handling was a little more useful in that all reads must happen through the socket engine, so these were happening at the correct time, but we can clean this up by doing it in the SE itself rather than each platform port. This means that both read and write events are now easily and usefully defined as "a syscall of either read or write was attempted". We also count empty read and write events as being an event, because they still were an attempt to poll a socket in some way. This may help to identify "bad" code which is repeatedly trying to read a socket for some reason. Lastly, we check for failed read/write calls, and log them as an error event. A lot of the time, this is how sockets are determined as being disconnected (ie. at read/write time). While we're at it, split Update() in two to make the calls more self-describing. This has no real impact since only one call is made at a time anyway.
2017-07-09Merge v2.0.23 and v2.0.24 into master.Peter Powell
2016-11-22socketengine_epoll: use getrlimit not ulimitAdam
2016-08-11socketengine_epoll Remove unused #define EP_DELAY 5Attila Molnar
2015-04-12Dispatch EventHandler events to dedicated virtual functionsAttila Molnar
Remove enum EventType
2014-06-14Kill needless #includes in source filesAttila Molnar
2014-06-07Change the number reported by SocketEngine::GetMaxFds() to be informalAttila Molnar
Do not exit if we can't determine it
2014-06-07socketengine_epoll Pass a dummy value to epoll_create() that does not depend ↵Attila Molnar
on GetMaxFds() From man epoll: Since Linux 2.6.8, the size argument is unused, but must be greater than zero. (The kernel dynamically sizes the required data structures without needing this initial hint.)
2014-06-07Relax fd bounds checkingAttila Molnar
We can cope with fds greater than SocketEngine::GetMaxFds() since 3752b3f59d5216d7dc6221a361efc76b9ad2273d
2014-02-09Store the EventHandler* in the kevent/epoll_event/portev structAttila Molnar
This removes the need to call GetRef() for every socket in DispatchEvents()
2014-02-08Change all socketengine methods to be staticAttila Molnar
2014-02-08Move socketengine stats into a new classAttila Molnar
2014-02-08Update SocketEngine::CurrentSetSize in AddFdRef()/DelFdRef()Attila Molnar
2014-02-08Rename SocketEngine::AddFd() and DelFd() to AddFdRef() and DelFdRef()Attila Molnar
2014-02-08Move variables in socketenginesAttila Molnar
Move timespec to DispatchEvents() in kqueue, there is no reason for it to be a member variable
2014-02-08Replace SocketEngine::GetName() with INSPIRCD_SOCKETENGINE_NAME defineAttila Molnar
2014-02-05socketengine_epoll Remove needless initialization of epoll_event when ↵Attila Molnar
removing an fd
2014-02-05socketengine_epoll Initialize CurrentSetSize (only used by m_httpd_stats)Attila Molnar
2014-02-05Fix oversight in 3752b3f59d5216d7dc6221a361efc76b9ad2273dAttila Molnar
2014-02-04Coding style changes in socketenginesAdam
2014-02-04New socketengine stuff:Adam
Use vectors that grow as necessary instead of mass allocating everything at once Rework poll engine logic to make sense
2013-08-30Merge insp20attilamolnar
2013-07-07Fix low risk crash when we can't determine maximum open socket count.Peter Powell
2013-05-19Fix spacing in calls to LogManager::Log.Peter Powell
2013-04-12Add LOG_ prefix to the log level enum values.Peter Powell
2012-10-14Replace printf(_c) with iostreamChrisTX
2012-05-20epoll: make sure EPOLLET flag does not get clobberedWilliam Pitcock
EPOLLET is the highest bit, so storing EPOLLET in a signed value leads to undefined behaviour.
2012-04-19Replace copyright headers with headers granting specific authors copyrightRobby-
2010-03-02DelFd should not fail, it will leave a bad dangling pointer in that casedanieldg
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@12584 e03df62e-2008-0410-955e-edbf42e46eb7
2010-03-02Log duplicate UUID events a bit betterdanieldg
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@12580 e03df62e-2008-0410-955e-edbf42e46eb7
2010-02-11When we get events on unknown file descriptors, unregister them rather than ↵danieldg
spinning at 100% CPU git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@12436 e03df62e-2008-0410-955e-edbf42e46eb7
2010-01-18Remove excessive gettimeofday system callsdanieldg
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@12293 e03df62e-2008-0410-955e-edbf42e46eb7
2010-01-11...because every now and again, i have to do a massive commit.brain
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@12248 e03df62e-2008-0410-955e-edbf42e46eb7
2009-09-30Remove include/inspircd_se_config.h and socketengine-specific headersdanieldg
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@11784 e03df62e-2008-0410-955e-edbf42e46eb7
2009-09-26Add FD_WANT_SINGLE_WRITE to efficiently replace FD_WANT_POLL_WRITEdanieldg
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@11762 e03df62e-2008-0410-955e-edbf42e46eb7
2009-09-26Clean up SocketEngine interface to allow edge-triggered I/O and sockets that ↵danieldg
do not force readability. git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@11760 e03df62e-2008-0410-955e-edbf42e46eb7
2009-09-21Create StreamSocket for IO hooking implementationdanieldg
Fixes the SSL SendQ bug Removes duplicate code between User and BufferedSocket Simplify SSL module API Simplify EventHandler API (Readable/Writeable moved to SE) Add hook for culled objects to invoke callbacks prior to destructor Replace SocketCull with GlobalCull now that sockets can close themselves Shorten common case of user read/parse/write path: User::Write is now zero-copy up to syscall/SSL invocation User::Read has only two copy/scan passes from read() to ProcessCommand git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@11752 e03df62e-2008-0410-955e-edbf42e46eb7
2009-03-15Update all wiki links to point to the new wiki. This was done automatically ↵psychon
with the following command, only .Makefile.inc got some indent fixups by hand. for file in $(find -type f -and -not -path '*/.svn/*' -and -not -name '*.so') ; do sed -e 's#http://www.inspircd.org/wiki#http://wiki.inspircd.org#' -e 's#http://wiki.inspircd.org/index.php/#http://wiki.inspircd.org/#' -i $file ; done git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@11223 e03df62e-2008-0410-955e-edbf42e46eb7
2009-01-02Update copyrights for 2009.w00t
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@10932 e03df62e-2008-0410-955e-edbf42e46eb7
2008-09-11These socket engines may now recieve write and read events in the same ↵w00t
cycle, same as select and poll. git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@10514 e03df62e-2008-0410-955e-edbf42e46eb7
2008-08-25Forward-port poll socket engine as it seems reasonably stable, and half-hack ↵w00t
it into configure. If someone could help me (easily) add it via our wonderful configure macros so it's built, that would be swell. git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@10264 e03df62e-2008-0410-955e-edbf42e46eb7
2008-08-23Forward port r10234, thanks psychonw00t
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@10235 e03df62e-2008-0410-955e-edbf42e46eb7
2008-07-20Fixes to some stuff that writes on memory it shouldn't (thanks psychon).w00t
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@10047 e03df62e-2008-0410-955e-edbf42e46eb7
2008-06-01Only prepare epoll for 25% of maxfds initially; it will grow beyond that as ↵special
necessary, but it's pretty unusual to be loading your server that much. Should save some kernel memory, too git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@9832 e03df62e-2008-0410-955e-edbf42e46eb7
2008-06-01Improve debug output on a few key methods of epoll socketengine, a must when ↵peavey
coding for the socket engine. git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@9830 e03df62e-2008-0410-955e-edbf42e46eb7
2008-05-19Debug on out of range FD.w00t
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@9776 e03df62e-2008-0410-955e-edbf42e46eb7
2008-04-06Initialise this (valgrind warning)w00t
git-svn-id: http://svn.inspircd.org/repository/trunk/inspircd@9384 e03df62e-2008-0410-955e-edbf42e46eb7
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-26 20:15:13 +0000