From mboxrd@z Thu Jan 1 00:00:00 1970 From: Pete Zaitcev Subject: Re: GFP_ATOMIC allocations... Date: Tue, 3 Sep 2002 12:07:15 -0400 Sender: linux-scsi-owner@vger.kernel.org Message-ID: <20020903120715.A11328@devserv.devel.redhat.com> References: <20020828202534.G30927@redhat.com> <20020828174737.A27554@eng2.beaverton.ibm.com> <20020828214215.A31167@redhat.com> <20020830092257.A19730@eng2.beaverton.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20020830092257.A19730@eng2.beaverton.ibm.com>; from patmans@us.ibm.com on Fri, Aug 30, 2002 at 09:22:57AM -0700 List-Id: linux-scsi@vger.kernel.org To: Patrick Mansfield Cc: linux-scsi@vger.kernel.org, zaitcev@redhat.com > Date: Fri, 30 Aug 2002 09:22:57 -0700 > From: Patrick Mansfield >[...] > Do we really call scan_scsis and scsi_build_commandblocks while not in user > context? Who is the caller (to scsi_register_host)? > > Here is Pete's response saying he found such a case, but not where: I did a call graph, but I discarded it since. IIRC it showed some paths, but perhaps I made a mistake. Please ignore it. The "fix" did not go anywhere, actually. Arjan explained that if a box runs out of atomics, it's bad anyway. We reduced the queue size in the driver that the guy used. -- Pete