On Tue, Dec 27, 2011 at 09:19:48AM +0000, YAMAMOTO Takashi wrote:
> vfork based implementation has its advantages.  eg. less kernel code
> 
> i'm not sure what kind of "dirty libpthread changes".
> can you explain?

We would need to guarantee thread safeness of posix_spawn() - but vfork
itself is not. So either we would need to change the kernel vfork() to
be thread safe (I guess I could do that, but it would be an intrusive
mess), or deal with it inside libpthread somehow (I have no idea -
enumerate all threads and stop them?).

But worst: I don't see how we could possibly gain the main benefit of
the interface (vfork like performance even for multithreaded apps) any
way if we go this route.

Martin

Reply via email to