Re: [patch] thread-exec-2.5.34-B1, BK-curr

[Alan Cox]

On Sun, 2002-09-15 at 19:38, Linus Torvalds wrote:
> 
> On Sun, 15 Sep 2002, Ingo Molnar wrote:
> > 
> > i dont like those semantics either - will verify whether thread-specific
> > exec() works via a helper thread (or vfork) - it really should.
> 
> As long as it works with something sane (and vfork() is sane), I'm happy 
> with the posix behaviour by default. After all, the execve() really _does_ 
> need to "de-thread" anyway, and if we need to make that explicit (with the 
> vfork()) then that's fine.

An execve can be setuid code so it really represents a whole new
security domain. Thats why the thread signal protection refuses to let
strange child exit signals cross it.

There is code that depends on clone()/exec() not killing other threads
in the group - some threaded web servers for example.