[Intel-gfx] [PATCH] sna/video sprite patch: support for multiple xv ports (exposes minimum Xv ports)

Hadley, MichaelX michaelx.hadley at intel.com
Tue Jun 21 09:43:37 UTC 2016


Hi Chris,

Thanks for the feedback and I'm glad to see that you've made the necessary changes and integrated the patches.

Regards

Mike

-----Original Message-----
From: Chris Wilson [mailto:chris at chris-wilson.co.uk] 
Sent: Monday, June 20, 2016 12:10 PM
To: Hadley, MichaelX
Cc: intel-gfx at lists.freedesktop.org
Subject: Re: [PATCH] sna/video sprite patch: support for multiple xv ports (exposes minimum Xv ports)

On Tue, May 10, 2016 at 01:12:30PM +0100, Michael Hadley wrote:
> The current driver exposes a single adaptor port so that only one Xv 
> video can be played using a single sprite at a given time. This change 
> will allow available hardware sprites to be used. Exposes the minimum 
> Xv sprite ports to be compatible with older X11.
> 
> v2: function sna_video_num_sprites() checks and returns the minimum number
>     of hardware sprites available on the adaptor for combatibility reasons.
> 
> v1: updated patch from lornax.mcneill at intel.com. removed/added
>     whitespace, re-formatted code, removed redundent variable allocation.

This still has the concept of a drawable owning a sprite, which is what I believe to be inconsistent with the exclusive ownership of the sprite from the XvAdapter. Abstracting the access, emulating more planes than physically present is intriguing, just requires careful scrutiny of the various runtime errors that the resource conflicts will then create and whether applications can handle that. Versus providing exclusive access to a sprite and the numerous virtual texture ports already provided.
-Chris

--
Chris Wilson, Intel Open Source Technology Centre


More information about the Intel-gfx mailing list