From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <archer-return-2373-listarch-archer=sourceware.org@sourceware.org>
Received: (qmail 17503 invoked by alias); 22 Feb 2011 19:39:46 -0000
Mailing-List: contact archer-help@sourceware.org; run by ezmlm
Sender: <archer@sourceware.org>
Precedence: bulk
List-Post: <mailto:archer@sourceware.org>
List-Help: <mailto:archer-help@sourceware.org>
List-Subscribe: <mailto:archer-subscribe@sourceware.org>
List-Id: <archer.sourceware.org>
Received: (qmail 17494 invoked by uid 22791); 22 Feb 2011 19:39:45 -0000
X-SWARE-Spam-Status: No, hits=-6.4 required=5.0
	tests=AWL,BAYES_00,RCVD_IN_DNSWL_HI,SPF_HELO_PASS,T_RP_MATCHES_RCVD
X-Spam-Check-By: sourceware.org
Date: Tue, 22 Feb 2011 19:39:00 -0000
From: Oleg Nesterov <oleg@redhat.com>
To: Roland McGrath <roland@redhat.com>
Cc: Jan Kratochvil <jan.kratochvil@redhat.com>,
        Project Archer <archer@sourceware.org>
Subject: Re: ptrace improvement: PTRACE_O_INHERIT
Message-ID: <20110222193124.GA3257@redhat.com>
References: <20110215003551.BC1EA1802A2@magilla.sf.frob.com> <20110215130805.GA30742@redhat.com> <20110215214333.GA18086@host1.dyn.jankratochvil.net> <20110215215634.2665A1806E0@magilla.sf.frob.com> <20110216193356.GB15576@redhat.com> <20110216194534.2D39018020E@magilla.sf.frob.com> <20110216200110.GA17601@redhat.com> <20110219194758.GA8994@host1.dyn.jankratochvil.net> <20110219202921.GA12280@redhat.com> <20110221195405.D27B11801F8@magilla.sf.frob.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20110221195405.D27B11801F8@magilla.sf.frob.com>
User-Agent: Mutt/1.5.18 (2008-05-17)
X-SW-Source: 2011-q1/txt/msg00091.txt.bz2

On 02/21, Roland McGrath wrote:
>
> > However. With or without PTRACE_O_TRACEEXIT, the thread can't go away
> > until ptracer does do_wait(WEXITED). And until it does do_wait() tkill()
> > succeeds even if the tracee is dead/zombie.
>
> That is true except in the case of MT exec (de_thread).

Ah! Indeed you are right, I forgot about this case again.

> The other threads
> live in the process at the time when another does exec will be immediately
> reaped without the chance for the debugger to see its death report.

Unless I missed something again, other threads are visible to wait().

But, in case de_thread() changes the leader, it does release_task(old_leader).
However, from ptrace pov it looks as if the execing thread goes away, it
changes its pid.

This reminds me. PTRACE_O_TRACEEXIT is not reliable if
arch_ptrace_stop_needed(). This is only sparc/ia64 so far, but we were
going to change ptrace_stop() so that it never stops if the process
was really killed by SIGKILL...

Oleg.