On Tue, Feb 13, 2018 at 11:08:38AM +0000, Daniel Stone wrote: > Hi all, > > On 13 February 2018 at 11:02, Marius-cristian Vlad > <marius-cristian.v...@nxp.com> wrote: > > The sample client I've posted seems to be using the overlay plane. > > Also, Daniel Stone confirmed a while back that they can be used. > > > > However it seems that page flipping on the overlay plane can not > > be used if the connector is used in the same time (and I'm not > > treating that case -- I disable that connector before handling out the > > lease). Or primary plane -> page flipping, overlay -> wait for vblank? > > > > Looking briefly over weston code I see that for overlay planes they use > > wait for vblank mechanism. Maybe Daniel/Pekka can confirm this, or > > maybe this is an issue with legacy API and atomic doesn't suffer from > > this problem. I'm not 100% sure. > > > > I guess it would be interesting to see if we can use it in the same > > time. > > With the atomic API at least, overlay planes cannot be driven > separately from other planes on that CRTC. Atomic only allows one > commit per CRTC per vblank: either you hand out CRTCs in their > entirety, or when you lease an overlay plane to a client, you stop > updating anything else on that CRTC.
We've been talking about exposing async flips more generally, which (under some driver-specific circumstances at least) could bypass the one per-crtc commit queue that atomic has. That would allow us to hand out overlay planes using leases, while the compositor still composites everything else. But that's a quite a bit away, and there isn't really any real code yet for it. -Daniel -- Daniel Vetter Software Engineer, Intel Corporation http://blog.ffwll.ch _______________________________________________ wayland-devel mailing list wayland-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/wayland-devel
