On Wed, Jun 28, 2017 at 12:07:28PM -0400, Jean-Francois Dagenais wrote: > Hi Hyun, > > > On Jun 27, 2017, at 20:53, Hyun Kwon <hyun.kwon@xxxxxxxxxx> wrote: > > > > > > As you noted, the Xilinx DRM driver in Xilinx tree is missing many mainline features including the atomic mode setting and needs some restructuring. Please feel free to send patches to Xilinx git-dev, but note that we are also developing / cleaning up the driver internally at Xilinx to upstream in near feature. While doing that, the code base may change due to heavy cleanup. > > I understand. I am right now shifting the _drv part to an > drm_of_component_probe model. I have removed the drm_driver.load and > unload functions in favour of the component bind/unbind. We were seeing > races in the devices appearing at boot time. > > I wonder why you are not sharing the restructuring effort publicly. It > would certainly help us track and align our own restructuring in order > to reduce code clash upon potential merge, as well as gather early > feedback and adoption for the "braver*" ones among us. (*: here "brave" > means not close to production/release;) > > Anyway, I would love to see the state of your restructure, even if it > still doesn't compile or work! To encourage collaboration: For upstreaming the first one to submit a reasonable driver wins, 2nd party needs to get their stuff integrated into the already merged driver (if it's for the same hw). Better if you work together openly from the start :-) Cheers, Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
