Signed-Off-By:  Solomon Peachy <redacted>

Below is v5b of the ATOM BIOS patch.  Changes from v5:

 * LVDS/DFP with attached DDC port can be queried
 * LVDS/DFP won't be misdetected as CRT if a DDC port is connected.

  http://www.shaftnet.org/users/pizza/radeonfb-atom-2.6.17-v5b.diff
  http://www.shaftnet.org/users/pizza/radeonfb-atom-2.6.18-v5b.diff

Meanwhile, I've rolled in some other work to create v6 of the patch, 
this time against 2.6.18-rc6 only: 

  http://www.shaftnet.org/users/pizza/radeonfb-atom-2.6.18-v6.diff

Changes from v5b:

(Note -- many of these changes were taken in part from an older patch by
 Claudio Ciccan; I kept what was relevant and tossed the rest)

 * Added many new radeon PCI IDs (X300/X600/X700/X800/X850)
 * R4x0 series gets R300 tweaks (from Xorg)
 * Cap VRAM to 128M (Not really sure why.. related to mmap bugs?)
 * Add default_dynclk to builtin argument parsing.
 * Expanded the register mmio size to max (64k vs former 16k)
     (was there a reason why it wasn't already maxed?)
 * Initialization tweaks for R300 and RS300 (from Xorg)

I'd really like to see the v6 patch (or its successors) go in to 2.6.19, 
as the existing radeonfb driver simply won't work with newer radeon 
cards -- to say nothing about its multihead detection quirks.

So, I encourage everyone with a radeon card to give this new driver a 
whirl, especially if you have an X-series radeon board or if you were 
having head detection problems with the existing driver. 

Please run this with CONFIG_FB_RADEON_DEBUG on, so if something goes 
wrong I can get it fixed up.  

There at least one known corner case where things could go wrong
(LVDS/DFP port has an associated DDC port, but no EDID information is
available) but I'm already working on that.

See also:

  http://bugzilla.kernel.org/show_bug.cgi?id=6215

Ciao,

 - Solomon
-- 
Solomon Peachy        		       pizza at shaftnet dot org	 
Melbourne, FL                          ^^ (mail/jabber/gtalk) ^^
Quidquid latine dictum sit, altum viditur.          ICQ: 1318344