[PATCH 04/14] drm/exynos: remove struct *_win_data abstraction on planes

[Rob Clark]

On Thu, Feb 5, 2015 at 4:15 AM, Daniel Vetter <daniel at ffwll.ch> wrote:
> On Thu, Feb 05, 2015 at 11:37:13AM +0900, Joonyoung Shim wrote:
>> Hi Daniel,
>>
>> On 02/04/2015 11:28 PM, Daniel Vetter wrote:
>> > On Wed, Feb 04, 2015 at 04:44:12PM +0900, Joonyoung Shim wrote:
>> >> Hi,
>> >>
>> >> On 02/04/2015 04:14 AM, Gustavo Padovan wrote:
>> >>> From: Gustavo Padovan <gustavo.padovan at collabora.co.uk>
>> >>>
>> >>> struct {fimd,mixer,vidi}_win_data was just keeping the same data
>> >>> as struct exynos_drm_plane thus get ride of it and use exynos_drm_plane
>> >>> directly.
>> >>>
>> >>> It changes how planes are created and remove .win_mode_set() callback
>> >>> that was only filling all *_win_data structs.
>> >>>
>> >>
>> >> I commented already on prior patch.
>> >
>> > I think you don't quite understand how this primary/overlay plane stuff
>> > works in drm core. The entire point of the drm core primary plane is to
>> > work _exactly_ like an overlay plane and allow userspace to mangle the
>> > primary plane configuration through the overlay plane. The only reason we
>> > have primary planes is so that old userspace keeps working.
>> >
>>
>> Right, i misunderstood a bit because exynos hw drivers have dependency
>> of zpos(hw overlay position).
>>
>> Current exynos drm driver has each primary plane of hw drivers and five
>> overlay planes. The primary plane is fixed on default hw overlay and all
>> overlay plane can map to all hw overlays using specific zpos property of
>> exynos drm plane.
>>
>> Gustavo approach will include specific hw overlay data in overlay plane
>> and hw driver keeps overlay planes to array by zpos order. But current
>> zpos of overlay plane is 0 always if user doesn't modify it, so hw
>> driver will use only hw overlay data of primary plane always even if
>> user want to use overlay plane.
>>
>> If user is modified zpos of overlay plane, hw driver can get wrong hw
>> overlay data from different overlay plane because hw driver keeps
>> overlay planes by zpos order.
>
> Yeah I noticed the zpos fun when hacking around too. Exynos should
> probably switch defaults so that overlays are visible by default. And we
> need to standardize the zpos property so that other drivers can use it
> too.

jfyi, I have a bit of logic in mdp5_crtc_atomic_check() (and really
mdp4 probably needs the same) to sort attached planes and derive the
actual hw zpos (with each layer having a unique zpos)..

I suspect most hw will end up needing the same (ie. dislike having two
overlays at same zpos), so might not be a horrible idea to make the
atomic helpers do this sorting for us..

BR,
-R

> But that doesn't change anything with the primary plane just being a
> special plane from the sw side (backwards compat), for exynos hw they all
> look the same.
> -Daniel
> --
> Daniel Vetter
> Software Engineer, Intel Corporation
> +41 (0) 79 365 57 48 - http://blog.ffwll.ch
> _______________________________________________
> dri-devel mailing list
> dri-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/dri-devel