[Xorg-driver-geode] Glyph rendering
Huang, FrankR
FrankR.Huang at amd.com
Mon Jul 12 20:10:32 PDT 2010
Jonathan,
Thanks. Good explanation for me.
I see the hope of implementing that in our driver:).
See below for my question:
-----Original Message-----
From: xorg-driver-geode-bounces+frankr.huang=amd.com at lists.x.org [mailto:xorg-driver-geode-bounces+frankr.huang=amd.com at lists.x.org] On Behalf Of Jonathan Morton
Sent: 2010?7?12? 23:07
To: Huang, FrankR
Cc: xorg-driver-geode at lists.x.org; michel at daenzer.net
Subject: Re: [Xorg-driver-geode] Glyph rendering
On Mon, 2010-07-12 at 18:06 +0800, Huang, FrankR wrote:
> Our driver is in low performance in glyph rendering.
> Based on the debug and discussion with Michel, There are three operation:
> 1)Op:PictOpAdd, Src:PICT_a8r8g8b8 Mask:0 Dest:PICT_a8
Extract the Alpha component from the source, and add it to the
destination.
[Frank] How to extract the alpha component? Drop the R,G,B of source picture? And what is the formula for PictoOpAdd between (alpha of source) and (dest) in this special condition? Copy ? Or A + (1-a)*B?
> 2)Op:PictOpAdd, Src:PICT_a8 Mask:0 Dest:PICT_a8r8g8b8
Add the source to the Alpha component of the destination.
[Frank] Copy the alpha of source to alpha of Dest? Make is clear? How about the formula?
Both of the above fall obviously out of the general XRender rules, where
missing image components are taken from the ARGB (1,0,0,0) vector during
fetch, and discarded during store.
> 3)Op:PictOpOver, Src:PICT_a8r8g8b8 Mask:PICT_a8r8g8b8 Dst:PICT_r5g6b5
Is Component Alpha turned on for this? If not, extract the Alpha
component from the mask and otherwise treat as for A8 mask.
[Frank] I debug and found it is not CA enabled. That is to say pMask->componentAlpha is zero. At the same time, our driver can not support CAComposite. If that is so, how about the formula for this PictOpOver operation? Only see the pMask as A8 and discard the R,G,B of pMask?
It's not entirely clear to me why the above operations would be commonly
encountered. Possibly some optimisation of higher software layers would
be appropriate.
--
------
From: Jonathan Morton
jonathan.morton at movial.com
_______________________________________________
Xorg-driver-geode mailing list
Xorg-driver-geode at lists.x.org
http://lists.x.org/mailman/listinfo/xorg-driver-geode
More information about the Xorg-driver-geode
mailing list