https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=277959
--- Comment #6 from Kyle Evans <[email protected]> --- I have a tentative patch for this, but I'm still thinking about it a little bit. My current approach is to pull the kqueue out of the eventloop and stop recreating it every time; the pipe read event will drop out when we close the read-side and leave just the signals. daemon_sleep() then switches to using a one-shot EVFILT_TIMER instead of nanosleep(2) so that it can still process signals received in the interim following a previously processed SIGCHLD. -- You are receiving this mail because: You are on the CC list for the bug.
