Hi, I am using non-blocking sockets in a small server software. Select() seems to indicate that sockets returned by accept() are always ready for write. This causes 99% cpu usage since the socket gets always selected as writeable and after that write returns EAGAIN and again we go to the select().
After looking select.cc fhandler.cc fhandler_socket.cc I think that the reason could be that in fhandler_socket::accept the socket is not set to CONNECTED state and in fhandler_socket::select_write sets write_ready if socket is in unconnected state. (uname -a CYGWIN_NT-5.0 O1400031 1.3.19(0.71/3/2) 2003-01-23 21:31 i686 unknown unknown Cygwin) - Risto - Sunpoint tiedottaa: Puhelintuki numerossa 0600-10 900 1,83E/min+pvm Tutustu myös Etätukeen tästä http://www.sunpoint.net/SunAds/click.htm?mode=footer&id=70&jump=http%3A%2F%2Fwww.pilotvastauspalvelu.fi%2Fdefault.asp%3Fsivu%3D3 -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting: http://cygwin.com/bugs.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/
