Re: [V5, 2/6] fsl/fman: Add FMan support

[Scott Wood <scottwood@freescale.com>]

On Thu, 2015-10-29 at 10:22 -0500, Liberman Igal-B31950 wrote:
> Regards,
> Igal Liberman
> 
> > -----Original Message-----
> > From: Wood Scott-B07421
> > Sent: Wednesday, October 28, 2015 11:31 PM
> > To: Liberman Igal-B31950 <Igal.Liberman@freescale.com>
> > Cc: netdev@vger.kernel.org; linuxppc-dev@lists.ozlabs.org; linux-
> > kernel@vger.kernel.org; Bucur Madalin-Cristian-B32716
> > <madalin.bucur@freescale.com>
> > Subject: Re: [V5, 2/6] fsl/fman: Add FMan support
> > 
> > On Tue, 2015-10-27 at 11:32 -0500, Liberman Igal-B31950 wrote:
> > 
> > > > > +
> > > > > +struct device *fman_get_device(struct fman *fman) {  return
> > > > > +fman->dev; }
> > > > 
> > > > Is this really necessary?
> > > > 
> > > 
> > > Fman port needs fman->dev, fman structure is opaque, so yes, it's 
> > > needed.
> > 
> > Why is opacity being maintained from one part of the fman driver to
> > another?
> > Isn't this the sort of excessive layering that was complained about?
> > 
> > 
> 
> It's not really layering.
> Fman Port uses Fman resources, it's not completely standalone. 

That's my point -- if it's not standalone, why is "struct fman" opaque to the 
port code?

-Scott