From owner-freebsd-scsi@FreeBSD.ORG  Fri Dec 12 19:47:57 2003
Return-Path: <owner-freebsd-scsi@FreeBSD.ORG>
Delivered-To: freebsd-scsi@freebsd.org
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP
	id 16C0616A4CE; Fri, 12 Dec 2003 19:47:57 -0800 (PST)
Received: from mailman.zeta.org.au (mailman.zeta.org.au [203.26.10.16])
	by mx1.FreeBSD.org (Postfix) with ESMTP
	id C2CE343D1F; Fri, 12 Dec 2003 19:47:54 -0800 (PST)
	(envelope-from bde@zeta.org.au)
Received: from gamplex.bde.org (katana.zip.com.au [61.8.7.246])
	by mailman.zeta.org.au (8.9.3p2/8.8.7) with ESMTP id OAA14078;
	Sat, 13 Dec 2003 14:47:52 +1100
Date: Sat, 13 Dec 2003 14:47:54 +1100 (EST)
From: Bruce Evans <bde@zeta.org.au>
X-X-Sender: bde@gamplex.bde.org
To: Scott Long <scottl@freebsd.org>
In-Reply-To: <3FD8FF91.9010702@freebsd.org>
Message-ID: <20031213144044.I3311@gamplex.bde.org>
References: <3FD8868C.4020208@netasq.com> <3FD8FF91.9010702@freebsd.org>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
cc: freebsd-scsi@freebsd.org
cc: Fabien THOMAS <fabien.thomas@netasq.com>
Subject: Re: asr driver with INVARIANT
X-BeenThere: freebsd-scsi@freebsd.org
X-Mailman-Version: 2.1.1
Precedence: list
List-Id: SCSI subsystem <freebsd-scsi.freebsd.org>
List-Unsubscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-scsi>,
	<mailto:freebsd-scsi-request@freebsd.org?subject=unsubscribe>
List-Archive: <http://lists.freebsd.org/pipermail/freebsd-scsi>
List-Post: <mailto:freebsd-scsi@freebsd.org>
List-Help: <mailto:freebsd-scsi-request@freebsd.org?subject=help>
List-Subscribe: <http://lists.freebsd.org/mailman/listinfo/freebsd-scsi>,
	<mailto:freebsd-scsi-request@freebsd.org?subject=subscribe>
X-List-Received-Date: Sat, 13 Dec 2003 03:47:57 -0000

On Thu, 11 Dec 2003, Scott Long wrote:

> Fabien THOMAS wrote:
> > When compiled with INVARIANT the driver failed in debugger because the
> > memory flags is M_WAITOK.
> >
> > Is it really necessary to use M_WAITOK here ?
>
> Allocating any memory from within the strategy() path is generally a
> really bad idea no matter what.  However, changing this to M_NOWAIT
> means that a memory shortage will cause the I/O to fail.  What happens
> if the failed I/O request originated from the vm system trying to push
> pages to the swap partition in order to create more free memory?

Same as in the ata driver: things fail.  I accidentally saw failures
of malloc() in the ata driver when I broke vm.  The MALLOC_MAKE_FAILURES
options would give failures intentionly.  The problems that I saw were
mostly transient, but I mostly don't use swap.

Bruce