From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vladislav Bolkhovitin Subject: Re: [RFC PATCH 0/19] Target_Core_Mod/ConfigFS v3.0 for v2.6.30 Date: Mon, 06 Apr 2009 16:09:20 +0400 Message-ID: <49D9F0F0.9080206@vlnb.net> References: <1238904091.4250.741.camel@haakon2.linux-iscsi.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Cc: LKML , linux-scsi , Linux-fsdevel , James Bottomley , Andrew Morton , Hannes Reinecke , FUJITA Tomonori , Mike Christie , Joel Becker , "Martin K. Petersen" , Ming Zhang , Philipp Reisner , Linus Torvalds , Jens Axboe , Rafiu Fakunle , Douglas Gilbert To: "Nicholas A. Bellinger" Return-path: In-Reply-To: <1238904091.4250.741.camel@haakon2.linux-iscsi.org> Sender: linux-scsi-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Nicholas A. Bellinger, on 04/05/2009 08:01 AM wrote: > Greetings all, > > I am very proud to announce the first RFC and submission for review of > Target_Core_Mod/ConfigFS v3.0 for v2.6.30. > > Target_Core_Mod is a generic target engine and ConfigFS enabled > infrastructure that provides access to the export server side resources > from Linux/SCSI, Linux/BLOCK and Linux/FILEIO (using buffered I/O) > subsytems to target mode storage fabric modules (like the LIO-Target > v3.0 fabric module). It does complete SPC-3 control emulation on top of > Linux/BLOCK and Linux/FILEIO objects, and does a SCSI passthrough (with > intelligent max_sectors handling) for Linux/SCSI objects. > > Target_Core_Mod/ConfigFS is the next generation version of the > LIO-Target v2.9 codebase that is completely SCSI fabric module > independent. The configuration plane has been ported from our legacy > LIO IOCTl control path in v2.9, and is now 100% upstream ConfigFS > infrastructure clean for v3.0. > > Target_Core_Mod/ConfigFS also contains a number of SCSI fabric features > that you will not find in any other open source Linux-SCSI or > Linux-iSCSI target implementation. These now include SPC-3 compliant > persistent reservation support that is passing complete domain > validation from multiple client environments (RHEL SCSI fencing and MSFT > Cluster 2008), Asymmetric Logical Unit Assignment (Linux scsi_dh_alua > and MPxIO on OpenSolaris clients), VPD/WWN information emulation and > SCSI MIBs to name the most interesting ones. Nicholas, What is the ultimate goal of Target_Core_Mod/ConfigFS? Is it to replace currently kernel's target subsystem STGT? If yes, what advantages Target_Core_Mod/ConfigFS has over STGT, which can justify such a move? For convenience of all interested people I summarized comparison between STGT and other existing SCSI targets for Linux, including Target_Core_Mod/ConfigFS (LIO v3.0), in http://scst.sourceforge.net/comparison.html. If you don't like anything in this comparison, we can discuss it. Vlad