From owner-freebsd-arch@FreeBSD.ORG  Mon Sep  5 11:26:07 2011
Return-Path: <owner-freebsd-arch@FreeBSD.ORG>
Delivered-To: freebsd-arch@FreeBSD.org
Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34])
	by hub.freebsd.org (Postfix) with ESMTP id C17AF106566B
	for <freebsd-arch@FreeBSD.org>; Mon,  5 Sep 2011 11:26:07 +0000 (UTC)
	(envelope-from avg@FreeBSD.org)
Received: from citadel.icyb.net.ua (citadel.icyb.net.ua [212.40.38.140])
	by mx1.freebsd.org (Postfix) with ESMTP id 18E3E8FC1F
	for <freebsd-arch@FreeBSD.org>; Mon,  5 Sep 2011 11:26:06 +0000 (UTC)
Received: from odyssey.starpoint.kiev.ua (alpha-e.starpoint.kiev.ua
	[212.40.38.101])
	by citadel.icyb.net.ua (8.8.8p3/ICyb-2.3exp) with ESMTP id OAA29236
	for <freebsd-arch@FreeBSD.org>; Mon, 05 Sep 2011 14:26:05 +0300 (EEST)
	(envelope-from avg@FreeBSD.org)
Message-ID: <4E64B1CC.9070205@FreeBSD.org>
Date: Mon, 05 Sep 2011 14:26:04 +0300
From: Andriy Gapon <avg@FreeBSD.org>
User-Agent: Mozilla/5.0 (X11; FreeBSD amd64;
	rv:5.0) Gecko/20110705 Thunderbird/5.0
MIME-Version: 1.0
To: freebsd-arch@FreeBSD.org
X-Enigmail-Version: 1.2pre
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Cc: 
Subject: SCHEDULER_STOPPED vs kdb_active
X-BeenThere: freebsd-arch@freebsd.org
X-Mailman-Version: 2.1.5
Precedence: list
List-Id: Discussion related to FreeBSD architecture <freebsd-arch.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-arch>,
	<mailto:freebsd-arch-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-arch>
List-Post: <mailto:freebsd-arch@freebsd.org>
List-Help: <mailto:freebsd-arch-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-arch>,
	<mailto:freebsd-arch-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Mon, 05 Sep 2011 11:26:07 -0000


My simple question is whether kdb_active should be included as another indicator
into SCHEDULER_STOPPED?

I assume that kdb_active == true is visible only to the debugger code and never to
the debugged code.  I understand that there is some code that can be executed in
both roles, but I am specifically asking about the roles, not the physical code.
So, is this true?
Assuming that the above is true, it doesn't make sense to not by-pass locking if
the debugger infrastructure executes some common code.
For example I don't understand why the kdb_active checks in kern_mutex.c are what
they are.  I think that kdb_active means that the currently running code is the
only code that can make forward progress in the system.  If we stumble upon a
locked spinlock, then it doesn't make any sense to spin on it waiting for Godot.

Similarly, I think that mtx_asser() should be a NOP if kdb_active.

What important things/scenarios do I overlook?

-- 
Andriy Gapon