Date: Mon, 28 Nov 2016 12:14:34 +0000
From: Jamie Iles <jamie.iles@oracle.com>
To: Oleg Nesterov <oleg@redhat.com>
Cc: Jamie Iles <jamie.iles@oracle.com>, linux-kernel@vger.kernel.org,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Ingo Molnar <mingo@redhat.com>, Peter Zijlstra <peterz@infradead.org>,
        Andrew Morton <akpm@linux-foundation.org>
Subject: Re: [PATCH] signal: protect SIGNAL_UNKILLABLE from unintentional
 clearing.
Message-ID: <20161128121434.4it7yd7wmpkxv6f4@cedar>
References: <20161116125759.4727-1-jamie.iles@oracle.com>
 <20161117190420.GA23492@redhat.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20161117190420.GA23492@redhat.com>
User-Agent: Mutt/1.6.2-neo (2016-08-21)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 928
Lines: 30

Hi Oleg,

On Thu, Nov 17, 2016 at 08:04:20PM +0100, Oleg Nesterov wrote:
> On 11/16, Jamie Iles wrote:
> >
> > This can result in init becoming stoppable/killable after tracing.  For
> > example, running:
> >
> >   while true; do kill -STOP 1; done &
> >   strace -p 1
> 
> > and then stopping strace and the kill loop will result in init being
> > left in state TASK_STOPPED.  Sending SIGCONT to init will resume it, but
> > init will now respond to future SIGSTOP signals rather than ignoring
> > them.
> 
> Yes, and a lot more... I forgot about these problems again.
> 
> Jamie, sorry for delay, I'll try to read the patch and reply tomorrow.

Did you get chance to look at the patch?  I did have another thought - 
rather than the accessors, we could change signal_struct to have:

  unsigned int unkillable:1;
  unsigned int flags:31;

to separate signal_unkillable from flags, making it a bit safer in the 
future.

Jamie