From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1755020AbbLDTCY (ORCPT <rfc822;w@1wt.eu>);
	Fri, 4 Dec 2015 14:02:24 -0500
Received: from mx1.redhat.com ([209.132.183.28]:36041 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S1754357AbbLDTCV (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
	Fri, 4 Dec 2015 14:02:21 -0500
Date: Fri, 4 Dec 2015 20:02:50 +0100
From: Oleg Nesterov <oleg@redhat.com>
To: Pavel Machek <pavel@ucw.cz>
Cc: Dmitry Vyukov <dvyukov@google.com>, LKML <linux-kernel@vger.kernel.org>,
        Roland McGrath <roland@hack.frob.com>, syzkaller@googlegroups.com,
        Kostya Serebryany <kcc@google.com>,
        Alexander Potapenko <glider@google.com>,
        Robert Swiecki <swiecki@google.com>, Kees Cook <keescook@google.com>,
        Julien Tinnes <jln@google.com>, Eric Dumazet <edumazet@google.com>
Subject: Re: Unkillable processes due to PTRACE_TRACEME
Message-ID: <20151204190244.GA16199@redhat.com>
References: <CACT4Y+Zu9uvt5z_LRAQaPTo+nwfodwBbjkeTRHoxGKRCPqpx5Q@mail.gmail.com> <20151019194911.GA20063@redhat.com> <CACT4Y+aDeSED9pnERC+gyHAEV=+n+awJ-s-8QGBQWMzuww894A@mail.gmail.com> <CACT4Y+bHNEKdZpgOYc+LaKxDyt1GeD+Vf4jLnmAVWQ3Z8=LKaw@mail.gmail.com> <CACT4Y+YmBeTxXVpoWLybcKCy_EwX2uAA5u2b7KbACJX4k3B4FQ@mail.gmail.com> <20151020105539.GA27706@redhat.com> <20151203205613.GC14427@amd>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20151203205613.GC14427@amd>
User-Agent: Mutt/1.5.18 (2008-05-17)
Sender: linux-kernel-owner@vger.kernel.org
List-ID: <linux-kernel.vger.kernel.org>
X-Mailing-List: linux-kernel@vger.kernel.org

Hi Pavel,

On 12/03, Pavel Machek wrote:
>
> > You can't. This is one of historical oddities. You need to reap the
> > traced sub-thread first. And PTRACE_DETACH doesn't work.
>
> If kill -9 does not take out the process,

Just in case, "kill -9" can't help because the task is already killed and
zombie. The problem is that /sbin/init can't reap it without __WALL unless
we change the kernel.

> surely that sounds like a
> security problem?
>
> I know ptrace is old and tricky and ugly, but ....?

Yes this should be fixed. I'll resend the patches next week, I am a bit busy
now.

And, Dmitry, I didn't forget about another problem you reported ;) I'll try
to redo/resend the fixes for WARN_ON() in in task_participate_group_stop()
as well.

Oleg.