Re: Re: [PATCH] Radeon 2D Acceleration for 2.6

[Benjamin Herrenschmidt <benh@kernel.crashing.org>]

On Fri, 2004-01-23 at 12:22, Michel Dänzer wrote:
> On Wed, 2004-01-21 at 00:09, random wrote: 
> > 
> > +void radeonfb_prim_fillrect(struct radeonfb_info *rinfo, 
> > +                       const struct fb_fillrect *region)
> > +{
> > +	radeon_fifo_wait(4);  
> > +  
> > +	OUTREG(DP_GUI_MASTER_CNTL,  
> > +		rinfo->dp_gui_master_cntl  /* contains, like GMC_DST_32BPP */
> > +                | GMC_BRUSH_SOLID_COLOR
> > +                | ROP3_P);
> > +	OUTREG(DP_BRUSH_FRGD_CLR, region->color);
> > +	OUTREG(DP_WRITE_MSK, 0xffffffff);
> 
> You could probably take this out of the acceleration functions, you
> don't seem to need a different write mask?

Yes, we could factor this one out

> > +	OUTREG(DP_CNTL, (DST_X_LEFT_TO_RIGHT | DST_Y_TOP_TO_BOTTOM));
> > +
> > +	radeon_fifo_wait(2);  
> > +	OUTREG(DST_Y_X, (region->dy << 16) | region->dx);
> > +	OUTREG(DST_WIDTH_HEIGHT, (region->width << 16) | region->height);
> > +	radeon_engine_idle ();
> 
> Why idle the engine in every acceleration function? Doesn't the API
> provide a way to only do it when necessary?

I don't do it in the patch I commited to my tree :) There is an fb_sync
callback that you use for that.

> PS: You lost my last reply to your private mails to your temporary
> address scheme.

And you'll lose mines too, please get rid of this shit and get a real
address :)

Ben.




-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn