Problems with monitor calibration and MergedFB/TwinView

Graeme Gill graeme2 at
Mon Aug 21 21:56:04 PDT 2006

Alex Deucher wrote:

 > Despite your unpleasant attitude, I'll reply to this...again.

Umm - I received no reply to my question, so I hardly see this
as replying again. As for "unpleasant", well, I've been around
a while, and I call them as I see them. If that grates with
you, then perhaps there is a reason. If one has to goad to get
a response, it's better than being ignored.

 > I replied to this thread a couple days ago, but if it didn't go through
 > then here it is one more time...

I see no follow-ups to my query.

 > At the moment it's not currently
 > possible to adjust the CLUT on the secondary crtc in mergedfb mode.
 > this is due to the fact that mergedfb is a hack to make multiple
 > monitors on a multi-crtc card look like a single screen.  The current
 > driver infrastructure is not really designd to handle multi crtc cards
 > very well (it's generally assumed one screen, monitor per card). As
 > such, there's no standard method for accessing the secondary CLUT.

Well, there is a standard mechanism that works, it's the
XFree86-VidModeExtension. MergedFB seems to intercept it/not
emulate it completely, and stops access to the screens individually.

 > one could add a driver option or an Xv attribute to adjust the gamma
 > on the second crtc, but those would just be hacks as well.  Hopefully
 > Keithp's new multi-head aware update to xrandr or the driver attribute
 > system discussed in the power management thread will provide some
 > semi-standard method of doing this.  If this is of interest to you
 > perhaps you could provide a patch or some ideas on implementing a
 > solution.  The source is available after all.

And if you'd kindly like to take over developing color
calibration and profiling software from me, then I'd
be happy to stop what I'm currently doing and start
contributing code. It might not be a very efficient
division of labour though.

The solution is not all that difficult - both MSWindows
and OS X have got it right without any great effort.
If you're going to provide a "virtual" screen that covers
all the real screens, just make sure there's a way
of accessing the real screens as well, rather than
cutting off all communication channels. Xinerama
seems to work OK, because (probably by accident as much as
design I would guess), it doesn't intercept the
XFree86-VidModeExtension traffic.

Quite a few Linux/Unix/X11 users have expressed considerable
interest and anticipation at the possibility of calibrating their
monitors. I've had to disappoint them by telling them
it doesn't work on their favourite multi-monitor extension.

Graeme Gill.

More information about the xorg mailing list