> Because the shutdown/restart code has traditionally been > trouble-prone, I'm posting this for comments before > committing. This patch changes the worker MPM to use the > ap_mpm_pod_* routines for pipe-of-death management in the > same manner as prefork. My main goal was to eliminate the > poll before the accept in the single-listener case; as a > side-effect, this allows for single-listener unserialized > accept support. Once worker and prefork are both using the > ap_mpm_pod_* API, it should be straightforward to switch to > using the POD instead of signals for all the stop/restart > cases (based on Ryan's recent suggestion).
It actually isn't possible for prefork to use the POD for all messages, because it requires a thread that can just catch messages, and prefork doesn't use threads. Other than that, +1. Ryan
