[RFC v2] Revive the work on render-nodes branch

Ilija Hadzic ihadzic at research.bell-labs.com
Thu Apr 12 12:09:13 PDT 2012

On Thu, 12 Apr 2012, Ville [iso-8859-1] Syrjälä wrote:

> Didn't have time for a detailed look yet, but at least one thing
> missing from your patch set is handling of possible_crtcs and
> possible_clones for getplane and getencoder ioctls.

Yes I agree and I know. That is still work in progress. The ioctl for 
constructing the node should do more checks not only for possible_crtcs 
and possible_clones, but also whether a particular encoder can match the 
connector and not allow creation of the node with incompatible resources. 
I have all that on the TODO list, but I wanted to get the patches out 
early so that I can get some feedback (and hopefully motivate others to 

There will definitely be v3 (and beyond) that will include stuff on my 
TODO list along with any feedback that I receive.

> Also the name "render node" is still very confusing. How about just
> callling it a partition or something. Or maybe someone could suggest
> a more self explanatory name for this stuff.

Dave has chosen the name two years ago when he did the initial demo work. 
I am just keeping this name for now (the advantage is that others know of 
this work by this name, so they know what it is about). However, if there 
is a concensus for a better name I'll be glad to adopt it.

-- Ilija

More information about the dri-devel mailing list