[mythtv-users] OT: Idea for possible workaround to stop bogus EDID data from messing up MythTV

Christopher David Petersen christopher.david.petersen at gmail.com
Thu Oct 11 18:56:59 UTC 2007

On 10/11/07, Michael T. Dean <mtdean at thirdcontact.com> wrote:
> On 10/11/2007 02:07 PM, Christopher David Petersen wrote:
> > On 10/11/07, Michael T. Dean <mtdean at thirdcontact.com> wrote:
> >
> >> On 10/11/2007 01:55 PM, Christopher David Petersen wrote:
> >>
> >>> I have a Olevia LT26HVE LCD TV, that I believe is providing bogus EDID
> >>> information to my video card (xdpyinfo reports the screen size as
> >>> 40x30mm: wrong size, wrong aspect ratio).
> >> Why not just tell your video driver to ignore the EDID information?
> > I haven't found the option to tell my *SiS* video driver to ignore
> > EDID information. I understand that nVidia drivers can use Option
> > "UseEDID" "FALSE".
> >
> > Does anybody know the option for SiS video drivers?
> I have to admit I don't know much about the SiS driver.  However, unless
> the driver is completely broken, it should a) use the information
> provided in the X command-line (or X configuration if none in the
> command line) to specify DPI/DisplaySize even if the monitor reports
> EDID information or b) allow the user to override the (broken) default
> of using EDID information regardless of configuration (as does the
> NVIDIA driver).  NVIDIA's default of using EDID info regardless of
> configuration is really against the X "way of doing things" (command
> line -dpi to the server should take precedence, followed by DisplaySize,
> and finally--if neither of the above are specified--defaulting to
> EDID-reported info).  So, I'd guess if there's no option listed in the
> man page for the driver to disable EDID/DDC (or at least the DPI
> information provided by it), choosing the right X options--either a -dpi
> argument to the server or DisplaySize in the config file--should work
> for you.
> And, if all else fails--and you're using the open-source driver--you can
> just hack it so sisfb_interpret_edid() always returns false or
> something.  That's always safer than hardware modifications :) (though
> it prevents mode validations so it would allow you to send through
> unusable signals with the wrong modelines).
> Mike
> _______________________________________________
> mythtv-users mailing list
> mythtv-users at mythtv.org
> http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users

Does the -dpi option normally force drivers to ignore EDID?

Am I correct that I need to change where X gets started to something like:
startx -- -dpi 57

Where is the startx command called during boot on a MythDora 4.0
install? Wait, don't feel compelled to answer that: I'll RTFM (or at
least google it).

I like the idea of improving the SiS driver to have an Option
"UseEDID" "true/false" feature, but I'm not above a simple "return
false;" hack :) Thanks, for pointing me to where I should hack (that's
always half the battle).

Thanks for the suggestions. I greatly appreciate the *real* support
the linux community provides.

More information about the mythtv-users mailing list