Dual head, display driven by mga driver has "X" cursor and is unusable

[Chris Fisichella]

  Quoting Chris Fisichella <chris at communityrenewables.com>:

> Hi,
>
> Does anyone have any experience with removing the X cursor from a second
> monitor to make it workable?
>
> Thanks,
> Chris
>
> I have been trying to add a second monitor to my machine. I can' quite
> get it. I am running Debian 7.6.0 and can get the desktop to show either:
> 1. On the original monitor (HP 2025, Nouveau driver), second monitor
> does not load X
> 2. On the second monitor (Haier 760, mga driver), first monitor is black
> with just an underscore cursor showing
> 3. The first monitor showing a usable window manager. I am typing in
> that window manager right now. Off to the right is the second monitor. I
> can move my cursor onto it and it turns to an X. I can't drag any
> windows onto that display. I arrived at this configuration by using the
> xorg.conf file that was generated by running
>
> #Xorg - configure
>
> In the Xorg.0.log file, I see:
> [    18.165] (EE) [drm] drmOpen failed.
> [    18.165] (EE) MGA(1): [drm] DRIScreenInit failed.  Disabling DRI.
> .
> .
> .
> [    18.166] (WW) MGA(1): Direct rendering disabled
> [    18.166] (--) RandR disabled
>
> The entire Xorg.0.log file is below
> Indeed, when I run
> #xrandr
> I don't see the second display.
>
>  

I swapped out the Matrox card and installed an ATI card. It's interesting
because I had the same failure mode. Can I have two cards on separate PCI
busses? I noticed for both cards it first does this:

[    16.987] (--) PCI: (0:0:13:0) 10de:03d1:105b:0d15 rev 162, Mem @
0xfb000000/16777216, 0xe0000000/268435456, 0xfa000000/16777216, BIOS @
0x????????/131072
[    16.987] (--) PCI:*(0:1:7:0) 102b:0520:102b:ff03 rev 1, Mem @
0xf9000000/16777216, 0xfd7fc000/16384, 0xfc800000/8388608, BIOS @
0x????????/65536

It finds both cards. One card is on bus 0 at location 13 (0xD) and the
other is on bus 1 at location 7.

For the card on bus 0, It finds and successfully calls
[    17.295] drmOpenDevice: node name is /dev/dri/card0
[    17.295] drmOpenDevice: open result is 9, (OK)
[    17.295] drmOpenByBusid: Searching for BusID pci:0000:00:0d.0
[    17.295] drmOpenDevice: node name is /dev/dri/card0
[    17.295] drmOpenDevice: open result is 9, (OK)
[    17.295] drmOpenByBusid: drmOpenMinor returns 9
[    17.295] drmOpenByBusid: drmGetBusid reports pci:0000:00:0d.0
[    17.295] (II) [drm] DRM interface version 1.4
[    17.295] (II) [drm] DRM open master succeeded.

The call to drmOpenByBusid returns 0:0:0d.0. I believe that is the onboard
video

When it comes time to use the video card in the PCI slot, something
different happens:
[    18.043] drmOpenDevice: node name is /dev/dri/card0
[    18.044] drmOpenDevice: open result is 11, (OK)
[    18.044] drmOpenDevice: node name is /dev/dri/card0
[    18.044] drmOpenDevice: open result is 11, (OK)
[    18.044] drmOpenByBusid: Searching for BusID pci:0000:01:07.0
[    18.044] drmOpenDevice: node name is /dev/dri/card0
[    18.044] drmOpenDevice: open result is 11, (OK)
[    18.044] drmOpenByBusid: drmOpenMinor returns 11
[    18.044] drmOpenByBusid: Interface 1.4 failed, trying 1.1
[    18.044] drmOpenByBusid: drmGetBusid reports pci:0000:00:0d.0
[    18.044] drmOpenDevice: node name is /dev/dri/card1

I see the call to drmOpenByBusid, and it seems to be trying to find the
card on 1:07:0.  I don't know why it is still using /dev/dri/card0. The
second call to drmGetByBusid then goes back to 0:d:0. /dev/dri/card1 should
be the card that is on the PCI conector. Anyway, as you can see from the
last email everything fails from there:

[    18.165] (EE) [drm] drmOpen failed.
[    18.165] (EE) MGA(1): [drm] DRIScreenInit failed.  Disabling DRI.

The same thing happened to the ATI card.

Any pointers on how to troubleshoot this? Is it a software bug? User
inexperience? I tried following the many posted "dual head" how-to's but
they just say to use xrandr. I can't get that far with my setup.

Thanks,
Chris
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.x.org/archives/xorg/attachments/20150306/a68d619b/attachment.html>