Subject: Re: Conversion of LIO-Target to use include/scsi/iscsi_proto.h defs
From: "Nicholas A. Bellinger" <nab@linux-iscsi.org>
To: Mike Christie <michaelc@cs.wisc.edu>
Cc: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>,
        linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org, hch@lst.de,
        James.Bottomley@suse.de, hare@suse.de
In-Reply-To: <4C6EF67E.2060908@cs.wisc.edu>
References: <1282257947.30453.296.camel@haakon2.linux-iscsi.org>
	 <20100820163058P.fujita.tomonori@lab.ntt.co.jp>
	 <4C6EF67E.2060908@cs.wisc.edu>
Content-Type: text/plain
Date: Fri, 20 Aug 2010 16:55:28 -0700
Message-Id: <1282348528.30453.450.camel@haakon2.linux-iscsi.org>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2870
Lines: 70

On Fri, 2010-08-20 at 16:41 -0500, Mike Christie wrote:
> On 08/20/2010 02:32 AM, FUJITA Tomonori wrote:
> > On Thu, 19 Aug 2010 15:45:47 -0700
> > "Nicholas A. Bellinger"<nab@linux-iscsi.org>  wrote:
> >
> >> So, assuming that the conversion of all struct iscsi_init_* and struct
> >> iscsi_target* to use include/scsi/iscsi_proto.h defs just involves
> >> checking structure size and renaming the member use in
> >> drivers/target/lio-target, would it be acceptable to do something like:
> >>
> >> /* iSCSI PDU Header */
> >> struct iscsi_cmd {
> >>          uint8_t opcode;
> >>          uint8_t flags;
> >>          __be16 rsvd2;
> >>          uint8_t hlength;
> >>          uint8_t dlength[3];
> >>          uint8_t lun[8];
> >>          itt_t    itt;   /* Initiator Task Tag */
> >>          __be32 data_length;
> >>          __be32 cmdsn;
> >>          __be32 exp_statsn;
> >>          uint8_t cdb[ISCSI_CDB_SIZE];    /* SCSI Command Block */
> >>          /* Additional Data (Command Dependent) */
> >> #ifdef ISCSI_TARGET_MODE
> >>          __be32 header_digest;
> >> #endif
> >> };
> >
> > It's up to Mike but it looks hacky to me. struct iscsi_pdu is defined
> > in the exact way as the spec does.
> >
> > I don't think that iscsi_proto.h conversion is a must for the mainline
> > inclusion. I prefer to let it alone for now.
> 
> For the iscsi target, I think it is. For iscsi initiator drivers like 
> bnx2i and be2iscsi we have made them convert from their headers to 
> common linux ones including switching from driver specific defs to 
> iscsi_proto.h.

This is kinda what I figured myself..  So in that case, I will go ahead
and plan on using the initial target mode stubs above to get up and
running, and start doing this sometime after the next (seperate) posting
of TCM_Core + TCM_Loop for Tomo-san's review. 

> 
> For the merging of any target core stuff though, I do not think it is 
> critical. The core target stuff does not need to be merged with a 
> software iscsi target. They can do in at different times.

I think in order to merge the software iSCSI target fabric bits for
v2.6.37, it would sense to do it initially with something similar to
these stubs.  From there I can have a more time to make necessary
modifications to LIO-Target and plan removing the header_digest target
stubs from iscsi_proto.h in for the .38 timeframe.

But really, since you so kindly volunteered to review the
drivers/target/lio-target code to start with, this decision is most
certainly yours to me made as the LIO-Target conversion to iscsi_proto.h
starts rolling..  8-)

Thanks for your comments Mike!

--nab

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/