Radeon driver in 1.4 breaks dualhead on Thinkpad with RV250

Mon Feb 15 08:47:47 PST 2010

On Mon, Feb 15, 2010 at 11:10 AM, Martin Cracauer <cracauer at cons.org> wrote:
> I previously reported this to the general Xorg list, but since then I
> checked that things work normally with other X11 servers of the same
> version and no ATI driver, so I thought I re-post here.
>
> Basically, when I went and updated my RV250 equipped Thinkpad to Xorg
> 1.4.2 (from what identifies itself as 7.0.0), the internal LCD refuses
> to work in dual-head with the VGA output.  From the Xorg.0.log output
> it looks like the Radeon/ATI driver incorrectly assumes it's talking
> to DVI, not LVDS.
>
> Cross-checking with the output when starting just the LCD
> single-headed (which works fine) I see what I think is confirmation,
> in that in that case RADEON(0) properly iterates through LVDS,
> S-video, DVI and VGA-0.
>
> In the dual-head config the driver insist on only trying DVI (the
> computer doesn't have a DVI output plug) and subsequently only comes
> up with a single screen, on the VGA, and the LCD stays blank.
>

The DVI port is on the docking station.  This is the problem with
zaphod mode, it assumes one connector per screen but when the system
has 4 connectors which ones do you associate with the each zaphod
instance?  Every user  has a different preference depending on their
setup.  Your best bet is to stick with the old driver if you want to
use zaphod, otherwise, use xrandr.

Alex

> I tried a couple of options, basically all I could find on google
> related to lvds and monitor options but the problem persists.
>
> It would be great if you could have a look and point me in the right
> direction.
>
> Here is the message I composed earlier with the Xorg.0.log ouput:
> The chip is:
> Radeon RV250 [Mobility FireGL 9000]
> I use the ATI/Radeon driver.  Dual-head via screenlayout.
>
> The working Xorg version identified itself as:
> X Window System Version 7.0.0
> Release Date: 21 December 2005
> X Protocol Version 11, Revision 0, Release 7.0
>
> The broken update is:
> X.Org X Server 1.4.2
> Release Date: 11 June 2008
>
> xorg.conf:
> http://www.cons.org/xorg-problem201002/xorg.conf.1.vgaonly-but-at-right-res.txt
> Log:
> http://www.cons.org/xorg-problem201002/Xorg.0.log.1.vgaonly-but-at-right-res.txt
>
> That is after I put in a couple options in an attempt to deal with the
> problem.  The original version working in xorg 7.0.0 is here:
> http://www.cons.org/xorg-problem201002/xorg.conf.previously-working
>
> I think these are the decisive lines:
> (II) RADEON(0): Output LVDS using monitor section LVDS
> (II) RADEON(0): BIOS provided dividers will be used.
> (WW) RADEON(0): LVDS Info:
> XRes: 1400, YRes: 1050, DotClock: 84960
> HBlank: 200, HOverPlus: 72, HSyncWidth: 40
> VBlank: 12, VOverPlus: 2, VSyncWidth: 1
> [correct]
> (II) RADEON(0): Output S-video has no monitor section
> (II) RADEON(0): Default TV standard: NTSC
> (II) RADEON(0): TV standards supported by chip: NTSC PAL NTSC-J
> (II) RADEON(0): Port0:
>  Monitor   -- AUTO
>  Connector -- DVI-D
>  DAC Type  -- None
>  TMDS Type -- Internal
>  DDC Type  -- 0x64
> [...]
> (II) RADEON(0): Output: DVI-0, Detected Monitor Type: 0
> finished output detect: 0
> [...]
> II) RADEON(1): Output VGA-0 using monitor section VGA-0
> (II) RADEON(1): I2C bus "VGA-0" initialized.
> (II) RADEON(1): Output DVI-0 has no monitor section
> (II) RADEON(1): DFP table revision: 3
> (II) RADEON(1): I2C bus "DVI-0" initialized.
> (II) RADEON(1): Output LVDS using monitor section LVDS
> (II) RADEON(1): Panel ID string: SXGA+ Single (85MHz)
> (II) RADEON(0): Panel Size from BIOS: 1400x1050
> (II) RADEON(1): BIOS provided dividers will be used.
> (WW) RADEON(1): LVDS Info:
> XRes: 1400, YRes: 1050, DotClock: 84960
> HBlank: 200, HOverPlus: 72, HSyncWidth: 40
> VBlank: 12, VOverPlus: 2, VSyncWidth: 1
> (II) RADEON(1): Output S-video has no monitor section
> (II) RADEON(1): Default TV standard: NTSC
> (II) RADEON(1): TV standards supported by chip: NTSC PAL NTSC-J
> (II) RADEON(1): Port0:
>  Monitor   -- AUTO
>  Connector -- VGA
>  DAC Type  -- Primary
>  TMDS Type -- None
>  DDC Type  -- 0x60
> [...]
> (II) RADEON(1): Output: VGA-0, Detected Monitor Type: 0
> (II) RADEON(1): Found color CRT connected to primary DAC
> [...]
> (II) RADEON(1): Output VGA-0 connected
> (II) RADEON(1): Output VGA-0 using initial mode 1368x768
>
>
> The result is, as I said, that only VGA has output, at the correct
> resultion.  The internal LCD stays blank.  There is only a
> DISPLAY=:0.0.
>
> To the best of my reading of the logfile my problem is that the driver
> erroneously used DVI instead of LVDS and subsequently deleted that
> screen.  I toyed around with everything I could find in google about
> this but I couldn't make the
> (II) RADEON(0): Output: DVI-0, Detected Monitor Type: 0
> go away nor make the LCD work.
>
> If I just use the LCD section in ServerLayout the LCD works fine and
> the VGA port mirrors the LCD output.  Log:
> http://www.cons.org/xorg-problem201002/Xorg.0.log.justonescreen.txt
> Notice that it now has a line:
> (II) RADEON(0): Output: LVDS, Detected Monitor Type: 0
>
>
> Anyway...
>
> So how do I force the first screen on LVDS in dual-head mode?
>
> Thanks
>        Martin
>
>
> P.S. while I here asking a question: while this worked fine under Xorg
> 7.0.0 I would like to force the server to come up dual-head even if
> corrently nothing is connected to the VGA port.  It didn't do that, if
> VGA wasn't connected it just came up with the LCD.
> --
> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
> Martin Cracauer <cracauer at cons.org>   http://www.cons.org/cracauer/
> FreeBSD - where you want to go, today.      http://www.freebsd.org/
> _______________________________________________
> xorg-driver-ati mailing list
> xorg-driver-ati at lists.x.org
> http://lists.x.org/mailman/listinfo/xorg-driver-ati
>