On Monday 10 March 2014 12:18:20 Tomi Valkeinen wrote: > On 08/03/14 13:41, Grant Likely wrote: > >> Ok. If we go for single directional link, the question is then: which > >> way? And is the direction different for display and camera, which are > >> kind of reflections of each other? > > > > In general I would recommend choosing whichever device you would > > sensibly think of as a master. In the camera case I would choose the > > camera controller node instead of the camera itself, and in the display > > case I would choose the display controller instead of the panel. The > > binding author needs to choose what she things makes the most sense, but > > drivers can still use if it it turns out to be 'backwards' > > I would perhaps choose the same approach, but at the same time I think > it's all but clear. The display controller doesn't control the panel any > more than a DMA controller controls, say, the display controller. > > In fact, in earlier versions of OMAP DSS DT support I had a simpler port > description, and in that I had the panel as the master (i.e. link from > panel to dispc) because the panel driver uses the display controller's > features to provide the panel device a data stream. > > And even with the current OMAP DSS DT version, which uses the v4l2 style > ports/endpoints, the driver model is still the same, and only links > towards upstream are used. > > So one reason I'm happy with the dual-linking is that I can easily > follow the links from the downstream entities to upstream entities, and > other people, who have different driver model, can easily do the opposite. > > But I agree that single-linking is enough and this can be handled at > runtime, even if it makes the code more complex. And perhaps requires > extra data in the dts, to give the start points for the graph. In theory unidirectional links in DT are indeed enough. However, let's not forget the following. - There's no such thing as single start points for graphs. Sure, in some simple cases the graph will have a single start point, but that's not a generic rule. For instance the camera graphs http://ideasonboard.org/media/omap3isp.ps and http://ideasonboard.org/media/eyecam.ps have two camera sensors, and thus two starting points from a data flow point of view. And if you want a better understanding of how complex media graphs can become, have a look at http://ideasonboard.org/media/vsp1.0.pdf (that's a real world example, albeit all connections are internal to the SoC in that particular case, and don't need to be described in DT). - There's also no such thing as a master device that can just point to slave devices. Once again simple cases exist where that model could work, but real world examples exist of complex pipelines with dozens of elements all implemented by a separate IP core and handled by separate drivers, forming a graph with long chains and branches. We thus need real graph bindings. - Finally, having no backlinks in DT would make the software implementation very complex. We need to be able to walk the graph in a generic way without having any of the IP core drivers loaded, and without any specific starting point. We would thus need to parse the complete DT tree, looking at all nodes and trying to find out whether they're part of the graph we're trying to walk. The complexity of the operation would be at best quadratic to the number of nodes in the whole DT and to the number of nodes in the graph. -- Regards, Laurent Pinchart
Attachment:
signature.asc
Description: This is a digitally signed message part.
