[patch] drawing window borders always falls back to software

[Aaron Plattner]

On Thu, Sep 24, 2009 at 02:10:55PM -0700, Michael wrote:
> Hello,
> 
> I finally found out why drawing ops to window borders always fall back  
> to software. It starts with miPaintWindow() - if it has to draw  
> outside the window's drawable in order to draw the border it requests  
> the screen pixmap which, thanks to the insane way in which it is  
> created by miCreateScreenResources(), is a DRAWABLE_PIXMAP without  
> further flags set so when miPaintWindow() calls ValidateGC() the resp.  
> DDX, be it XAA or EXA or whatever, thinks it's drawing into system  
> memory and falls back to software. The following patch fixes this for  
> XAA by  checking the target drawable against pScreen->devPrivate and  

Use pScreen->GetScreenPixmap(pScreen) instead of pScreen->devPrivate.  This
may not be completely sufficient if the driver creates other "special"
pixmaps (e.g. for an overlay), but I think the number of drivers that do
that that also use XAA is zero.

> use accelerated ops if they match.

I'm a little worried that this will break some drivers that expect
accelerated pixmaps to have driver-provided devPrivates attached.  I don't
know offhand of any drivers that do that, but it is a change in behavior.

> Index: xaaGC.c
> ===================================================================
> RCS file: /cvsroot/xsrc/external/mit/xorg-server/dist/hw/xfree86/xaa/ 
> xaaGC.c,v
> retrieving revision 1.1.1.2
> diff -u -w -r1.1.1.2 xaaGC.c
> - --- xaaGC.c     11 Jun 2009 01:52:58 -0000      1.1.1.2
> +++ xaaGC.c     24 Sep 2009 21:06:10 -0000
> @@ -88,7 +88,8 @@
>              pGC->fgPixel = 0x7fffffff;
>       }
> 
> - -    if((pDraw->type == DRAWABLE_PIXMAP) && ! 
> IS_OFFSCREEN_PIXMAP(pDraw)){
> +    if((pDraw->type == DRAWABLE_PIXMAP) &&
> +       !IS_OFFSCREEN_PIXMAP(pDraw) && !PIXMAP_IS_SCREEN(pDraw, pGC)) {
>          pGCPriv->flags = OPS_ARE_PIXMAP;
>           pGCPriv->changes |= changes;
> 
> Index: xaalocal.h
> ===================================================================
> RCS file: /cvsroot/xsrc/external/mit/xorg-server/dist/hw/xfree86/xaa/ 
> xaalocal.h,v
> retrieving revision 1.3
> diff -u -w -r1.3 xaalocal.h
> - --- xaalocal.h  11 Jun 2009 02:13:52 -0000      1.3
> +++ xaalocal.h  24 Sep 2009 21:06:11 -0000
> @@ -1710,6 +1710,9 @@
>   #define IS_OFFSCREEN_PIXMAP(pPix)\
>           ((XAA_GET_PIXMAP_PRIVATE((PixmapPtr)(pPix)))->offscreenArea)
> 
> +#define PIXMAP_IS_SCREEN(pPix, pGC)\
> +        (pPix == pGC->pScreen->devPrivate)
> +
>   #define PIXMAP_IS_SHARED(pPix)\
>           ((XAA_GET_PIXMAP_PRIVATE((PixmapPtr)(pPix)))->flags &  
> SHARED_PIXMAP)
> 
> 
> this isn't really a performance issue but there are graphics chips  
> which don't let us map their framebuffers in a useful way ( think SGI  
> newport and crime for example ) where we really don't want anything to  
> fall back to software if there is any other way.
> 
> A proper way to fix this would be to change the way the screen pixmap  
> is created so the accel lib, be it XAA, EXA or whatever, knows what  
> it's doing and can mark the pixmap as in video memory. As it is now  
> miCreateScreenResources() just calls CreatePixmap() with the right  
> depth but no width or height, then fills in geometry, data pointer and  
> so on.

Drivers can plug whatever else they need into the screen pixmap by wrapping
CreateScreenResources.

> have fun
> Michael