FireMV 2200 vs CustomEDID

Csillag Kristof csillag.kristof at
Tue Oct 6 07:57:38 PDT 2009

Adam Jackson wrote:
> On Tue, 2009-10-06 at 03:38 +0200, Csillag Kristof wrote:
>> Therefore, I have extended the syntax of the CustomEDID option to accept
>> commands after the EDID file name, and implemented one command ("digital")
>> to override this information in the EDID.
> I'm going to bet that what you implemented, won't work.  The radeon
> driver in UMS mode, to check whether the monitor is digital, does:
> static Bool
> monitor_is_digital(xf86MonPtr MonInfo)
> {
>     return (MonInfo->rawData[0x14] & 0x80) != 0;
> }
> So you need to mangle the actual EDID data itself, not just what the
> interpreter decides it says.
Thank you for the information; I will look into that.
> But even doing that is unlikely to help too much, because it looks like
> you're getting utter garbage for the EDID data:
> [...]
> That's not even close to being valid EDID.  Among other things, it
> doesn't start with 0x00ffffffffffff00.
I don't know why the displayed data is garbage,
but this EDID does work
    a) then I connect up the monitor via Analog,
    b) using CustomEDID with a RS690M radeon chipset (DVI connection),
    c) using CustomEDID with a FireMV 2400 PCI card (DVI connection)

so I guess it can not be completely worthless.

   * * *

Now that I looked, the X log files displays different EDID data
almost each time. However, the syntax check immediately after
loading the data (which checks for 0xff) succeeds.

I will check this later.


More information about the xorg mailing list