Re: [PATCH 21/21] tcm_loop: Add multi-fabric Linux/SCSI LLD fabricmodule

From: Nicholas A. Bellinger
Date: Wed Mar 16 2011 - 20:57:52 EST


On Tue, 2011-03-15 at 11:18 +0100, Christoph Hellwig wrote:
> On Mon, Mar 14, 2011 at 03:02:55PM -0700, Nicholas A. Bellinger wrote:
> > > Please descide on a naming scheme for the frontend modules. You have
> > > tcm_loop for thise one, and just iscsi for the iscsi frontend.
> >
> > Ok, curently everything that appears under drivers/target/ beside
> > iscsi_target_mod.ko is using a matching drivers/target/$FABRIC_MOD/ and
> > $FABRIC_MOD.ko naming scheme.
> >
> > I would prefer to keep iscsi_target_mod.ko for the software iSCSI case
> > in order to prevent chaos with existing userspace tools and prevent
> > breakage with existing other iSCSI target modules (IET is using
> > iscsi_trgt.ko IIRC).
> >
> > How about dropping the 'tcm_' prefix from the drivers/target/tcm_subdir/
> > to follow how the fabric module actually appears once loaded under
> > /sys/kernel/config/target/$FABRIC/..? We will need to keep the common
> > 'tcm_' prefix for the actual target fabric modules (or choose a
> > different one) to handle those that interface with existing SCSI LLD
> > code.
>
> Dropping the tcm sounds fine to me.
>

<nod>

> > This makes it a bit easier to grok as a reference for newcomers by
> > having seperate sections for configfs control plane, struct
> > target_core_fabric_ops API callers, and Linux/SCSI LLD specific glue.
>
> I don't think it's a very good reference, though. E.g. currently the
> fabric ops methods are implemented in tcm_loop_fabric.c, and all need
> non-static scope because they are assigned in tcm_loop_configfs.c, which
> creates a lot of global scope and pointless prototypes. Either one
> could clean all that up or just merge the files into one which would
> be a lot easier. Especiall as the way the target_fabric_configfs structure
> and the fabric ops are structured will get a big overhaul from me soon.
>

Ok, fair enough.. I will go ahead and look at getting this cleaned up.

In the mean time, James please go ahead and apply patches #1 -> #20 from
this target core series to scsi-misc as I have no futher changes to
report at this time. I will send out an updated PATCH 21/21 addressing
Christoph's comments for merging tcm_loop.ko shortly, and another RFC-v3
series addressing his comments for the initial iscsi_target_mod merge in
the next 24 hours.

--nab

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