DirectFB - Home of the pragmatist Roadmap

[directfb-dev] Re: Adding support for planar blitting
Mailing List archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[directfb-dev] Re: Adding support for planar blitting

Quoting Antonino Daplas (
> Hi,
> Blitting of YV12 and I420 formats should not be too difficult.  It's
> just an 8bpp to 8bpp blit with rect->h = rect->h + rect->h/2, right?
> I think the difficult part is how to add clipping support when you have
> 3 different planes.  
> So, can we just clip/blit each plane separately?  The following patch
> adds planar blitting to dfb_gfxcard_blit.  It actually subdivides the
> region into 5 subregions (1 for Y-plane, 2 for the U and V planes each),
> and each subregion separately undergoes a clip and a blit.  It looks
> pretty ugly right now but I have tested that at least the blitting
> works.  I haven't tested the clipping part.
> Is it worthwhile to pursue this path?  Anything that I might have
> missed? Any comments?

You implemented one of the two possible ways to fix the problem. You
took the higher level generic way easing the development of drivers.
I did the other way on friday that simply delegates the handling of
planar formats to the drivers. I'm not sure yet but I think it's better
because there may be support by the hardware to do blits from/to planar
formats. The software driver now implements blitting from planar to

Maybe both versions should be available to do accelerated software fallbacks
by doing these three 8bit blits if planar blitting is not supported natively.

The triangle filling code uses accelerated rectangle filling (each scanline)
if it's supported but triangles not.

Best regards,
  Denis Oliver Kropp

| DirectFB - Hardware accelerated graphics |
|                 |

                            Convergence GmbH

Info:  To unsubscribe send a mail to with 
"unsubscribe directfb-dev" as subject.

Home | Main Index | Thread Index / Development / Old Archives