xf86-video-intel: man/intel.man src/i810_reg.h src/i830_video.c src/i830_video.h

[Michel Dänzer]

On Fri, 2009-03-06 at 10:46 +0000, Barry Scott wrote:
> Michel Dänzer wrote:
> > On Thu, 2009-03-05 at 21:26 -0800, Haihao Xiang wrote:
> >   
> >> man/intel.man    |   15 ++++++++
> >>  src/i810_reg.h   |    4 ++
> >>  src/i830_video.c |  100 ++++++++++++++++++++++++++++++++++++++++++++-----------
> >>  src/i830_video.h |    2 +
> >>  4 files changed, 102 insertions(+), 19 deletions(-)
> >>     
> >
> > [...]
> >
> >   
> >> +.SS "XV_SYNC_TO_VBLANK"
> >> +XV_SYNC_TO_VBLANK is used to control whether textured adapter synchronizes 
> >> +the screen update to the vblank to eliminate tearing. It has three 
> >> +values 'auto'(-1), 'off'(0) and 'auto'(1). 'off' means never sync, 'on' means 
> >> +always sync, no matter what size, and 'auto' means sync if the Xv image is 
> >> +more than quarter of the pixels on the screen. The default is 'auto'(-1).
> >>     
> >
> > Rather than syncing to vblank, why don't you define an inclusive
> > scanline window using MI_LOAD_SCAN_LINES_INCL and then use an
> > MI_WAIT_FOR_EVENT 'Scan Line Window Wait' bit? The radeon driver uses
> > that principle, it should make the pipeline stall less when the vertical
> > area is smaller and thus remove the need for the 'auto' attribute value.
> >
> >
> >   
> 
> Does the INTEL hardware have the ability to do this?

Yes, the above bits are documented in the publicly available
VOL_1_graphics_core.pdf .


-- 
Earthling Michel Dänzer           |                http://www.vmware.com
Libre software enthusiast         |          Debian, X and DRI developer