Hi Olaf, On Tue, Nov 4, 2008 at 11:12 AM, Olaf van der Spek <[EMAIL PROTECTED]> wrote: > On Tue, Nov 4, 2008 at 5:05 PM, Michael Kerrisk > <[EMAIL PROTECTED]> wrote: >> I'm curious. How would you ideally improve it? > > With the current epoll wait behaviour and LT, you'd have to unregister > the EPOLLIN event for the listener. > After you close a descriptor, you then reregister it. > > With ET, you'd have to set a flag when you hit EM/ENFILE. After you > close a descriptor and the flag is set, call accept. > > Ideally I'd like to see epoll_wait do this behind the scenes, at least > in the ET case.
Okay -- thanks for the ideas. All of that would be a bit too complicated for this example of course (as I think you probably also agree). Cheers, Michael -- Michael Kerrisk Linux man-pages maintainer; http://www.kernel.org/doc/man-pages/ git://git.kernel.org/pub/scm/docs/man-pages/man-pages.git man-pages online: http://www.kernel.org/doc/man-pages/online_pages.html Found a bug? http://www.kernel.org/doc/man-pages/reporting_bugs.html -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
