From mboxrd@z Thu Jan 1 00:00:00 1970 From: Randy Dunlap Subject: Re: [PATCH] SCSI: ISCSI build failure Date: Wed, 01 Nov 2006 08:03:22 -0800 Message-ID: <4548C54A.2020708@oracle.com> References: <200610291945.16948.toralf.foerster@gmx.de> <20061029190357.98fc9c98.randy.dunlap@oracle.com> <20061030083230.5e889415.randy.dunlap@oracle.com> <1162327892.13859.67.camel@mulgrave.il.steeleye.com> <1162394814.13859.100.camel@mulgrave.il.steeleye.com> <20061101153623.GF11399@parisc-linux.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from rgminet01.oracle.com ([148.87.113.118]:58790 "EHLO rgminet01.oracle.com") by vger.kernel.org with ESMTP id S1423951AbWKAQIA (ORCPT ); Wed, 1 Nov 2006 11:08:00 -0500 In-Reply-To: <20061101153623.GF11399@parisc-linux.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: Matthew Wilcox Cc: James Bottomley , Roman Zippel , Toralf F_rster , scsi , akpm Matthew Wilcox wrote: > On Thu, Nov 02, 2006 at 12:26:54AM +0900, James Bottomley wrote: >> On Wed, 2006-11-01 at 01:00 +0100, Roman Zippel wrote: >>> Please don't select NET, it's way too broad. >> Then you'll make select implied dependencies work in Kbuild so we don't >> need to do anything at all about this? Roman said that CONFIG_NET is too broad to select, and I agree with him. Select is meant to be used to enable one library-like kernel feature/function, not all of "Networking". But yes, there do seem to be some depends/selects interaction whose handling is missing. > This is iSCSI ... how can it possibly be useful without NET? Why > doesn't CONFIG_ISCSI depend on NET, making this whole problem moot (for > now). There is no CONFIG_ISCSI symbol AFAIK. I see: CONFIG_SCSI_ISCSI_ATTRS CONFIG_ISCSI_TCP CONFIG_SCSI_QLA_ISCSI If CONFIG_NET=n, CONFIG_SCSI_QLA_ISCSI can still be enabled by a user and it will "select" CONFIG_SCSI_ISCSI_ATTRS. SCSI_ISCSI_ATTRS depends on SCSI && NET, so there should be some (either) depends or selects action going on there. I think we agree on that part at least. -- ~Randy who still wishes that the SCSI low-level drivers menu were not in some dog-awful random order