Re: [RFC 2/9] Revert "lsm: Remove the socket_post_accept() hook"

["Serge E. Hallyn" <serue@us.ibm.com>]

Quoting Tetsuo Handa (penguin-kernel@I-love.SAKURA.ne.jp):
> Serge E. Hallyn wrote:
> > Quoting Samir Bellabes (sam@synack.fr):
> > > This reverts commit 8651d5c0b1f874c5b8307ae2b858bc40f9f02482.
> > > 
> > > snet needs to reintroduce this hook, as it was designed to be: a hook for
> > > updating security informations on objects.
> > > 
> > > Signed-off-by: Samir Bellabes <sam@synack.fr>
> > 
> > Acked-by: Serge Hallyn <serue@us.ibm.com>
> > 
> > (contingent of course on the proposed user actually going in :)
> > 
> > > diff --git a/net/socket.c b/net/socket.c
> > > index 8984973..fcd4f2b 100644
> > > --- a/net/socket.c
> > > +++ b/net/socket.c
> > > @@ -1557,6 +1557,8 @@ SYSCALL_DEFINE4(accept4, int, fd, struct sockaddr __user *, upeer_sockaddr,
> > >  	fd_install(newfd, newfile);
> > >  	err = newfd;
> > > 
> > > +	security_socket_post_accept(sock, newsock);
> > > +
> > >  out_put:
> > >  	fput_light(sock->file, fput_needed);
> > >  out:
> 
> I think we should call security_socket_post_accept() before fd_install().
> Otherwise, other threads which share fd tables can use
> security-informations-not-yet-updated accept()ed sockets.

That makes sense.