EDID

Russell Shaw rjshaw at netspace.net.au
Fri Jul 6 19:10:02 PDT 2007


Val wrote:
> I had a very similar problem with my upgrade to Ubuntu Fiesty. Although
> I didn't figure out how to override the erroneous info coming from the
> CRT's EDID, I configured the NVIDIA driver to suppress many of the
> checks it does on a candidate graphical mode before putting it into the
> mode pool. This works if you already know what modes your CRT can do.
> 
> I don't know if you use nvidia, but I assume ATI will have some sort of
> an equivalent:
> 
> Option "ModeValidation" "CRT-1: NoVertRefreshCheck,
> NoDFPNativeResolutionCheck, NoMaxPClkCheck, NoEdidMaxPClkCheck,
> NoHorizSyncCheck, NoVirtualSizeCheck;"

Hi,
I tried various radeon driver settings to stop automatic EDID use, but
didn't work.

These days, i no longer bother bug hunting in X because graphics drivers
are too undocumented and mysterious for me to waste time on. DRI stuff is
barely documented too.

I simply downgraded some Debian packages and it works again now.

If things get too bad, i'll just throw away the ATI card and install
an NVidia one with the proprietory driver.

On a slightly different topic, i found why the NVidia-Albatron FX5200
card in my dads pc had rapid screen ripping and tearing. It only happens
when the kernel is configured to boot in graphical framebuffer mode.
Installing another different NVidia card worked ok.

Normally i'd debug to the lowest levels to find where the fault was
(i suspect some kind of card bios). I wasted too much time looking at
that and it was all too unneccessarily difficult for something that
should be easy. The card is now at the back of a cupboard instead of
having a work-around in the driver like it should.



More information about the xorg mailing list