[Intel-gfx] [PATCH] gfx: update the mmheight/mmwidth again after parsing EDID

[ykzhao]

On Tue, 2009-10-20 at 22:35 +0800, Adam Jackson wrote:
> On Tue, 2009-10-20 at 14:31 +0800, yakui.zhao at intel.com wrote:
> > From: Zhao Yakui <yakui.zhao at intel.com>
> > 
> > Sometimes the display mmheight/mmwidth defined in EDID detailed timing block
> > is incorrect for some monitors. After calling the function of xf86OutputSetEDID,
> > it will update the height/midth for the corresponding output by using the value
> > defined in detailed timing. In such case it will cause that the dpi is
> > also incorrect, which causes that font is displayed incorrectly.
> >    For example:
> >      The height/width in EDID detailed timing block is :289 x 21 mm
> >      But the height/wdith about monitor size is 330 x 210 mm.
> > 
> > So after calling the function of xf86OutputSetEDID, Update the height/width
> > again by using the output info, which is obtained in course of output detection.
> 
> You seem to be saying "xf86OutputSetEDID is broken, let's work around it
> in the driver".  I'm less than thrilled with that idea.
The issue is caused by the incorrect EDID. The detailed timing block in
EDID gives the incorrect height. And the function of xf86OutputSetEDID
will update the mmwidth/mmheight parsed from EDID, which causes that we
get the incorrect DPI.

And the issue can also be fixed by the following method:
   using the quirk to fix the info defined in EDID(In fact this quirk
for this monitor already  exists. But it is put into the incorrect
case). 

Thanks.
> 
> - ajax