forks
/
arsd
mirror of https://github.com/adamdruppe/arsd.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

important commentary on edge-triggeredness

This commit is contained in:
Adam D. Ruppe 2014-12-24 12:11:45 -05:00
parent 4d73f858fd
commit f9a105b914
1 changed files with 30 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

34
eventloop.d
View File

@ -323,16 +323,23 @@ public struct FileEventDispatcher {
WrappedListener[3] handlerSet; WrappedListener[3] handlerSet;
if(readEventHandler !is null) int events;
if(readEventHandler !is null) {
handlerSet[0] = wrap(readEventHandler); handlerSet[0] = wrap(readEventHandler);
if(writeEventHandler !is null) events |= FileEvents.read;
}
if(writeEventHandler !is null) {
handlerSet[0] = wrap(writeEventHandler); handlerSet[0] = wrap(writeEventHandler);
events |= FileEvents.write;
}
if(errorEventHandler !is null) if(errorEventHandler !is null)
handlerSet[0] = wrap(errorEventHandler); handlerSet[0] = wrap(errorEventHandler);
listeners[handle] = handlerSet; listeners[handle] = handlerSet;
addFileToLoop(handle, FileEvents.read | FileEvents.write);
addFileToLoop(handle, events);
} }
public void removeFile(OsFileHandle handle) { public void removeFile(OsFileHandle handle) {
@ -446,6 +453,24 @@ version(linux) {
// ev.events = EPOLL_EVENTS.EPOLLET; // edge triggered // ev.events = EPOLL_EVENTS.EPOLLET; // edge triggered
// Oh I think I know why I did this: if it is level triggered
// and the data is not actually handled, it infinite loops
// on it. So either way, the application needs to do its thing:
// either consume all available data every single time it is
// triggered - read until you get EAGAIN, OR make sure that
// data is never ignored; that every trigger leads to at LEAST
// ONE read.
//
// With writes, it is important to be extremely careful with
// level triggered - a file is often ready to write, especially
// if you aren't actually using it! I like to do blocking
// writes with non-blocking reads, so any level-triggered epoll
// on write is probably not what I want.
//
// Bottom line is this is a kinda leaky abstraction either way
// and we all need to understand what is going on to make the
// best of it. Also watch your CPU usage for infinite loops!
if(events & FileEvents.read) if(events & FileEvents.read)
ev.events |= EPOLL_EVENTS.EPOLLIN; ev.events |= EPOLL_EVENTS.EPOLLIN;
if(events & FileEvents.write) if(events & FileEvents.write)
@ -515,8 +540,9 @@ version(linux) {
auto flags = events[n].events; auto flags = events[n].events;
if(flags & EPOLL_EVENTS.EPOLLIN) if(flags & EPOLL_EVENTS.EPOLLIN)
sendSync(FileReadyToRead(fd)); sendSync(FileReadyToRead(fd));
if(flags & EPOLL_EVENTS.EPOLLOUT) if(flags & EPOLL_EVENTS.EPOLLOUT) {
sendSync(FileReadyToWrite(fd)); sendSync(FileReadyToWrite(fd));
}
if((flags & EPOLL_EVENTS.EPOLLERR) || (flags & EPOLL_EVENTS.EPOLLHUP)) if((flags & EPOLL_EVENTS.EPOLLERR) || (flags & EPOLL_EVENTS.EPOLLHUP))
sendSync(FileError(fd)); sendSync(FileError(fd));
} }