[Intel-gfx] [PATCH] Xv: free tearing on textured video

Zhenyu Wang zhenyu.z.wang at intel.com
Thu Feb 26 02:07:18 CET 2009


On 2009.02.26 00:30:19 +0800, Keith Packard wrote:
> On Wed, 2009-02-25 at 15:56 +0800, Zou, Nanhai wrote:
> 
> > I suggest turn this option default to be true or not use an option at all.
> > We already have too many options to handle.
> 
> Ideally, this code would do the right thing automatically, and be
> configured when it 'automatic' wasn't working right. Nanhai's suggestion
> of turning it on by default is a good one, as is Dave's suggestion of
> using an Xv attribute instead of a config file option. 
> 
> I'd suggest one further change -- only sync when the video is 'large',
> so that videos which are only a small part of the screen don't impact
> other applications running at the same time. 
> 
> How about an Xv attribute which has three values 'on' 'off' and 'auto'?
> 'on' means always sync, no matter what size, 'off' means never sync and
> 'auto' means sync if the Xv image is more than half(?) of the pixels on
> the screen. Then we set the default to 'auto'. Users can adjust the
> behaviour of the driver using 'xvattr'.
> 
> One problem with this plan is that 'xvattr' adjusts the attributes on a
> single port, and as we offer 16 textured video ports (along with 1
> overlay port on hardware with an overlay), the user would need to call
> xvattr up to 17 times to effect the change uniformly. I'd say we should
> just fix 'xvattr' to add a '-p all' option to adjust a parameter for all
> ports?
> 

Looks a good plan. I just checked spec for MI_WAIT_EVENT, it looks for newer
G45s some definition might have been changed. Haihao may try it and possibly
update the patch.

-- 
Open Source Technology Center, Intel ltd.

$gpg --keyserver wwwkeys.pgp.net --recv-keys 4D781827
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: Digital signature
URL: <http://lists.freedesktop.org/archives/intel-gfx/attachments/20090226/929977a0/attachment.sig>


More information about the Intel-gfx mailing list