From mboxrd@z Thu Jan  1 00:00:00 1970
From: arnd@arndb.de (Arnd Bergmann)
Date: Fri, 20 Jun 2014 20:56:49 +0200
Subject: [PATCH] dt/documentation: add specification of dma bus information
In-Reply-To: <53A46CB7.1030109@ti.com>
References: <1401981720-6946-1-git-send-email-santosh.shilimkar@ti.com>
 <CAL_JsqJAtWnHKPVpDsdOOzT_X=npc3hroLNTGqULM2SJr=1OEg@mail.gmail.com>
 <53A46CB7.1030109@ti.com>
Message-ID: <4193075.zZSpRTFEYz@wuerfel>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On Friday 20 June 2014 13:17:43 Santosh Shilimkar wrote:
> >> +                       dma-coherent;
> >> +                       dma-ranges;
> >> +
> >> +                       dwc3 at 2690000 {
> >> +                               compatible = "synopsys,dwc3";
> >> +                               [...]
> >> +                       };
> > 
> > This example is a bit strange. I don't understand the relationship
> > between keystone-dwc3 and synopsys,dwc3, nor do I want to. I'd prefer
> > to see a simple example here.


> > dma-ranges is a property of the parent which you show, but
> > dma-coherent originally was a property of the bus master itself. While
> > we need to support that, are we changing that? We need to be clear on
> > where the property belongs even if the kernel is more lax.
> > 
> I don't think we are changing it fundamentally but may be I missing
> your point. The dma-coherent as is now a per-device property.
> USB is one of the bus master supports coherency and hence showed
> up in above example.

I think it's enough if you just drop the "synopsys,dwc3" node and the
intermediate dma-ranges property from the example, leaving the
dma-coherent property in the "ti,keystone-dwc3" node.

	Arnd