Re: [Cooker] Re: call for testing: radeonhd shadowfb acceleration

[Adam Williamson]

On Sat, 2008-03-22 at 12:44 +0100, eatdirt wrote:
> Adam Williamson wrote:
> 
> > Option "AccelMethod" "ShadowFB"
> > 
> > and let me know -
> > 
> > a) does it break anything?
> > b) is it noticeably faster?
> 
> 
> Hi, tested on
> 
> Card:ATI Radeon X1300 and later: ATI Technologies Inc|R580 [Radeon X1900 
> XT] (Primary) [DISPLAY_VGA]
> 
> It is more than noticeably faster, the compositing under "xcompmgr" is 
> smooth with this option while it is not without, and there is apparently 
> no more bug than without (I am using fvwm2+xcompmgr).

Thanks to you and Fred for the testing. I also got feedback from Michael
Larabel (Phoronix editor). Based on this, I've patched radeonhd to
default to ShadowFB acceleration on all chipsets (it was already
defaulting to it on r6xx, now I look at the code). Can you (and anyone
else with an r5xx chip - Radeon X1xxx) please test with the updated
driver, remove the shadowfb option from xorg.conf if you added it, and
check that shadowfb is still used (and everything works fine)? When
ShadowFB is used, this is logged to /var/log/Xorg.0.log , so grep -i
shadow /var/log/Xorg.0.log should tell you. Thanks.
-- 
adamw