Disabling monitors with bad EDIDs?!?!

[Adam Jackson]

On Tue, 2010-09-21 at 14:59 +0100, Richard Hughes wrote:
> On 19 September 2010 11:56, Dave Airlie <airlied at gmail.com> wrote:
> > I'd be interested in checking if your EDID has somehow been corrupted.
> 
> The LP2480zx has a feature where it can update the EDID with the
> correct colorspace co-ordinates depending on the chosen colorspace. It
> would appear if you turn this feature off (to never change the EDID
> even if you change the 3d lut) the EDID checksum is incorrect.

That sounds a lot like a bug in the monitor.

> I'm pretty sure this EDID updating feature is turned on by default,
> although it's kinda hard to turn on if you ever turned it off, as the
> OSD for this feature is only available if you have a valid input,
> which in the case of the new drm, you won't.

The problem is that the checksum is pretty much the only indication we
have that the EDID is trustworthy at all.  If we accept EDID with bad
checksums we start getting into a game of deciding whether a particular
block "looks" believeable, and that requires making the parser _far_
more heuristic.  I mean, there's nothing to physically stop anybody from
making a 2560x32 monitor, and making an EDID block to reflect that, but
if I saw an EDID with that resolution as the physical size I might not
believe it...

That said we're probably being a bit agressive in turning "bad checksum"
into "disconnected".  That's definitely not intended.

- ajax
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
URL: <http://lists.x.org/archives/xorg-devel/attachments/20100922/532d2727/attachment.pgp>