[Trac #281] Geode vsync/hsync polarity strangeness

[Zarro Boogs per Child]

#281: Geode vsync/hsync polarity strangeness
----------------------+-----------------------------------------------------
 Reporter:  dwmw2     |       Owner:  mfoster
     Type:  defect    |      Status:  new    
 Priority:  normal    |   Milestone:  BTest-1
Component:  hardware  |    Keywords:         
----------------------+-----------------------------------------------------
 When doing initial DCON testing in July, we failed to get the Geode to
 operate sanely with Vsync and Hsync active high, which is what _was_
 needed for the DCON to operate correctly.

 It _seemed_ to be the case that it would always start up with active-low,
 regardless of whether we asked for high or low. And then it would change
 every time we asked for a change -- so if we'd started with it active-low
 then it would do what we ask for evermore, and if we'd started with it
 active-high then it would do the _opposite_ of what we ask. This situation
 would continue until the next hard reset.

 I got no coherent answer from AMD about the reasons for this, and
 eventually we just added some inverters in the DCON so it accepts active-
 low sync signals, and it was mostly no longer an issue. We start Linux in
 active-low mode, which is what we get, and all is well.

 Until yesterday, when I _watched_ X switch from active-low to active-high
 vsync+hsync because I happened to have the scope hooked up. And although
 VGA mode works on the panel like that because the panel isn't fussy, DCON
 mode was again broken.

 We _really_ need to understand [vh]sync polarity.

 Oh, and OF seems to be managing to power up with active _high_. Although
 it doesn't matter much for OF since OF will never be using DCON mode.

-- 
Ticket URL: <http://dev.laptop.org/ticket/281>
One Laptop Per Child <http://laptop.org/>