ati xrandr: cannot find crtc for output DVI-0

[Alex Deucher]

On 8/27/07, Alex Deucher <alexdeucher at gmail.com> wrote:
> On 8/27/07, Jan Dittmer <jdi at l4x.org> wrote:
> > Alex Deucher wrote:
> > > On 8/27/07, Jan Dittmer <jdi at l4x.org> wrote:
> > >> Given ati driver 6.7.192 and libxrandr 1.2.1 (Debian experimental)
> > >> I cannot enable the DVI-0 output using:
> > >>
> > >> $ xrandr --output DVI-0 --right-of VGA-0 --mode 1280x1024 --rate 60
> > >> xrandr: cannot find crtc for output DVI-0
> > >>
> >
> > [old xrandr output snipped]
> >
> > > radeons only have two crtcs and you are trying to enable more an two
> > > monitors.  you'll need to turn off S-video in order to turn DVI on.
> > >
> > > xrandr --output S-video --off
> > > xrandr --output DVI-0 --auto
> >
> > Thanks for the quick reply. That indeed fixes the issue.
> >
> > > If you don't have a tv plugged in than I'm guessing you probably have
> > > an analog monitor on DVI-0 which is causing the tv to detect load (TV
> > > and DVI port share the same dac).
> >
> > Exactly. I've a dvi-vga adapter on the DVI port. With a DVI cable the
> > color temperatures of the monitors don't match (under windows).
> >
> > A serious problem remains: The screen attached to the DVI output
> > is a lot brighter. All the aliased stuff around glyphs gets white so that
> > the screen is barely readable. All bright colors look white. Is this a problem
> > with some gamma processing? (I checked the monitor settings more than twice,
> > it's not correctable by lowering it's brightness or contrast)
>
> Hmmm... I haven't seen this myself, but I does remind me of a similar
> problem we had years ago, that may have resurfaced with all the
> refactoring of the driver.  I'll see if I can dig up what the solution
> was.

I think perhaps disp_merge_cntl or disp2_merge_cntl is programmed
wrong, (although I can't see how).  Can you dump the values of those
regs using radeontool from git?

e.g.,
radeontool regmatch DISP_MERGE_CNTL
radeontool regmatch DISP2_MERGE_CNTL

Alex