Re: ptrace: selinux_bprm_committed_creds: simplify __wake_up_parent() code and s/parent/real_parent/

From: Roland McGrath <roland@redhat.com>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Stephen Smalley <sds@tycho.nsa.gov>,
	Andrew Morton <akpm@linux-foundation.org>,
	David Howells <dhowells@redhat.com>,
	Eric Paris <eparis@parisplace.org>,
	James Morris <jmorris@namei.org>,
	linux-kernel@vger.kernel.org
Subject: Re: ptrace: selinux_bprm_committed_creds: simplify __wake_up_parent() code and s/parent/real_parent/
Date: Mon,  4 May 2009 10:38:40 -0700 (PDT)	[thread overview]
Message-ID: <20090504173840.D08E9FC32F@magilla.sf.frob.com> (raw)
In-Reply-To: Oleg Nesterov's message of  Sunday, 3 May 2009 22:10:27 +0200 <20090503201027.GB18889@redhat.com>

> > > Afaics, with this patch the only user of ->parent outside of ptrace.c & co
> > > is arch/ia64/kernel/mca.c:format_mca_init_stack(). Hopefully ->parent will
> > > die soon.
> >
> > You could clean it up today with:
> > 	tracehook_finish_clone(p, 0, 0);
> 
> Yes... but we still need to initialize real_parent/group_leader. 

tracehook_finish_clone->ptrace_init_task does this.
That is the sole point of the call.

> This code is so special, perhaps it is better to just remove "p->parent =
> " later and do not add "#include tracehook". We will see.

Whatever you want there is fine by me.

Thanks,
Roland