From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Nicholas A. Bellinger" Subject: Re: [PATCH] target: Call proc_mkdir + remove_proc_entry with NULL parameter Date: Mon, 24 Jan 2011 12:54:22 -0800 Message-ID: <1295902462.24778.40.camel@haakon2.linux-iscsi.org> References: <1295738425-4814-1-git-send-email-nab@linux-iscsi.org> <1295791971.3007.0.camel@mulgrave.site> Mime-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Return-path: Received: from mail.linux-iscsi.org ([67.23.28.174]:56636 "EHLO linux-iscsi.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752671Ab1AXUy0 (ORCPT ); Mon, 24 Jan 2011 15:54:26 -0500 In-Reply-To: <1295791971.3007.0.camel@mulgrave.site> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-scsi@vger.kernel.org To: James Bottomley Cc: linux-scsi , Fubo Chen , Christoph Hellwig , "J.H." , Joel Becker On Sun, 2011-01-23 at 08:12 -0600, James Bottomley wrote: > On Sat, 2011-01-22 at 15:20 -0800, Nicholas A. Bellinger wrote: > > From: Nicholas Bellinger > > > > This patch makes proc_mkdir() and remove_proc_entry() use a NULL > > parameter to fix the following sparse warning: > > > > CHECK drivers/target/target_core_configfs.c > > drivers/target/target_core_configfs.c:3131:54: warning: Using plain integer as NULL pointer > > drivers/target/target_core_configfs.c:3145:50: warning: Using plain integer as NULL pointer > > drivers/target/target_core_configfs.c:3212:42: warning: Using plain integer as NULL pointer > > Um, so I thought this interface was being replaced ... there's not a lot > of point in sparse fixing it. > Hi James, So my plan here is to move these statistics in /proc/scsi_target/ into configfs specific contexts for TCM v4.1. This also includes moving iscsi_target_mod -> /proc/iscsi_target/ to use /sys/kernel/config/target/iscsi/ context attributes using our existing CONFIGFS_EATTR() based macros in target_core_fabric_configfs.h to provide the fabric dependent attributes. So for target core this should be easy enough to split up short term in 'for-39', but the generic fabric statistics piece will require some new logic that allows CONFIGFS_EATTR() macros to be seq_list() aware in order to cat more than PAGE_SIZE for a individual configfs statistic attribute some (one..?) special case. (jlbec CC'ed) The reason we need this is to handle the N possible dynamic demo mode struct se_node_acl (Initiator NodeACL) per TargetName+TargetPortalGroup endpoint that we need for BKO ops. (J.H CC'ed) --nab