Disable stretching in ATI driver

Roland Scheidegger rscheidegger_lists at hispeed.ch
Wed Oct 25 04:09:23 PDT 2006


Benjamin Herrenschmidt wrote:
> On Mon, 2006-10-23 at 19:54 +0200, Klaus Reimer wrote:
>> Alex Deucher wrote:
>>> There have been patches floating around to disable RMX but nothing has
>>> ever been applied.  You should just be able to comment out the the
>>> setting of the RADEON_USE_RMX flag in the mode validation.  If I have
>>> time this week I'll commit something to git.
>> Thanks. I tried that but the result is not that nice. The 1024x768
>> screen is now displayed in the  upper left corner of the 1280x800
>> screen, but that's ok.
> 
> The chip has ways to do auto-centering but I haven't quite figured out
> how to properly program that. Also, we don't need/want to completely
> disable RMX. What we want is an option to maintain the aspect ratio or
> not (nvidia binary has that I think) and best if we can toggle that
> option at runtime (using Xv attributes maybe ? We lack a proper way to
> toggle things like that, unless Keith new RandR++ adds arbitrary
> attributes ?). Also we should probably default to maintaining the aspec
> ratio. That is, we calculate the RMX factors based on the smallest of
> the H and V factors and we center the resulting image.
I think if implementing some choice we'd really want to have all 3 
useful options. That is, centered with original size (granted 640x480 on 
a 1920x1200 display is probably not much fun that way, but some people 
prefer that especially if the displayed resolution isn't that much 
smaller than the screen resolution), scaled/centered with keeping aspect 
ratio, and scaled to fill the screen. Though yes, scaled without keeping 
aspect ratio probably shouldn't be the default, personally I don't think 
this mode is useful at all.

Roland



More information about the xorg mailing list