Is IgnoreEdid broken?

James C. Georgas jgeorgas at
Sun Oct 14 05:42:21 PDT 2007

On Sat, 2007-13-10 at 10:46 -0400, Alex Deucher wrote:
> whoops, yeah, that option looks like currently broken.  The problem
> is, in a randr 1.2 world, when you enable that option, do you want to
> ignore the edid on all outputs or just a specific one?  I suppose we
> should probably just remove it as there's a better method now.  see
> this page for more info:
> Basically you specify monitor sections and assign them to outputs.
> from there you can add modelines and specify what you want as the
> preferred mode in each monitor section.
> Alex
Tried it, and it still ignores my override, and uses the EDID.

My problem is that the reported dotclock range for my panel is wrong.
Either the EDID on the panel is wrong, or the driver is reading it
wrong. I'm not sure which, but the result is that the reported maxclock
is 170, when it should be 155 or thereabouts. All the other EDID data
seems to be reasonable, so I just want to adjust the clock range a bit,
and use EDID values for everything else.

I tried using the maxclock option, and the server reads it in, but
doesn't apply it to my modelines.

With 6.6.*, I still couldn't selectively modify the clock range, but at
least I could blow away all the EDID data and manually enter the values
in the monitor section. I can't do that any more.

Any thoughts on how to do this?


More information about the xorg mailing list