From: Hongli Lai <hongli@...99.net>
Date: 2009-07-27T19:51:38+09:00
Subject: [ruby-core:24572] Re: [Bug #1525] Deadlock in Ruby 1.9's VM caused by ConditionVariable.wait and fork?

Tanaka Akira wrote:
> I'm not sure we can help forking-without-exec situation.

Are you sure nothing can be done about this? I understand that you can't
do anything about native threads that you don't control, but wouldn't it
be at least possible to make threads that Ruby does control synchronize
at fork-time, so that when a fork happens we know that at least Ruby's
own threads won't interfere?
Phusion Passenger relies pretty heavily on a working fork implementation
for all kinds of optimizations. Not being able to fork will result in a
lot of inefficiencies.