From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755152AbaBSU3P (ORCPT ); Wed, 19 Feb 2014 15:29:15 -0500 Received: from mx1.redhat.com ([209.132.183.28]:29251 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754581AbaBSU3N (ORCPT ); Wed, 19 Feb 2014 15:29:13 -0500 Date: Wed, 19 Feb 2014 15:28:41 -0500 From: Richard Guy Briggs To: Oleg Nesterov Cc: John Johansen , Peter Zijlstra , linux-kernel@vger.kernel.org, linux-audit@redhat.com, "Eric W. Biederman" Subject: Re: [PATCH 4/5] audit: anchor all pid references in the initial pid namespace Message-ID: <20140219202841.GC17952@madcap2.tricolour.ca> References: <20131230180600.GD2457@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20131230180600.GD2457@redhat.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 13/12/30, Oleg Nesterov wrote: > On 12/23, Richard Guy Briggs wrote: > > > > Store and log all PIDs with reference to the initial PID namespace and > > use the access functions task_pid_nr() and task_tgid_nr() for task->pid > > and task->tgid rather than access them directly. > > At first glance this patch looks like a good cleanup, but... > > > @@ -429,6 +429,19 @@ static struct audit_entry *audit_data_to_entry(struct audit_rule_data *data, > > f->val = 0; > > } > > > > + if ((f->type == AUDIT_PID) || (f->type == AUDIT_PPID)) { > > + struct pid *pid; > > + rcu_read_lock(); > > + pid = find_vpid(f->val); > > + if (!pid) { > > + rcu_read_unlock(); > > + err = -ESRCH; > > + goto exit_free; > > + } > > + f->val = pid_nr(pid); > > + rcu_read_unlock(); > > + } > > I do not really understand this change, but this doesn't matter, I do > not understand audit. Is this pid_t handed down from userspace a valid pid_t in its namespace? If not, generate an error. If it is, store its global pid_t for future comparisons. It might be better to store a struct task_struct * or a struct pid *, but then the comparison functions would have to change too, along with the rule reporting functions, which still report pid_t to userspace. > However, I think this deserves a separate patch with the changelog. All *that* would need a seperate patch... > > @@ -278,9 +278,12 @@ static void dump_common_audit_data(struct audit_buffer *ab, > > } > > case LSM_AUDIT_DATA_TASK: > > tsk = a->u.tsk; > > - if (tsk && tsk->pid) { > > - audit_log_format(ab, " pid=%d comm=", tsk->pid); > > - audit_log_untrustedstring(ab, tsk->comm); > > + if (tsk) { > > + pid_t pid = task_pid_nr(tsk); > > + if (pid) { > > + audit_log_format(ab, " pid=%d comm=", pid); > > + audit_log_untrustedstring(ab, tsk->comm); > > Just curious, is it really possible that a->u.tsk is an idle thread? No. It is possible a->u.tsk isn't filled in. > Oleg. - RGB -- Richard Guy Briggs Senior Software Engineer, Kernel Security, AMER ENG Base Operating Systems, Red Hat Remote, Ottawa, Canada Voice: +1.647.777.2635, Internal: (81) 32635, Alt: +1.613.693.0684x3545