Can't get two monitors driven by two separate video cards running in 11.3

English Hardware
#1

I have a computer with a video on the motherboard (radeon)
and a separate video (radeon) card. It will not boot unless
I do “nomodeset” on startup. When it does, I get video from
the motherboard video card, but nothing from the separate
video card. The monitor indicates that it is getting no signal
at all.

I think both the onboard and separate cards are fairly basic
cards – nothing fancy.

I have edited /etc/X11/xorg.conf.d/50-device.conf to
include a line:

  Driver  "radeonhd"

And I have an /etc/X11/xorg.conf file:

Section “ServerLayout”
Identifier “X.org Configured”
Screen 0 “Screen0” 0 0
Screen 1 “Screen1” RightOf “Screen0”
InputDevice “Mouse0” “CorePointer”
InputDevice “Keyboard0” “CoreKeyboard”
EndSection

Section “Files”
ModulePath “/usr/lib/xorg/modules/updates”
ModulePath “/usr/lib/xorg/modules”
FontPath “/usr/share/fonts/misc:unscaled”
FontPath “/usr/share/fonts/TTF/”
FontPath “/usr/share/fonts/OTF/”
FontPath “/usr/share/fonts/Type1/”
FontPath “/usr/share/fonts/100dpi:unscaled”
FontPath “/usr/share/fonts/75dpi:unscaled”
FontPath “/usr/share/fonts/URW/”
FontPath “/usr/share/fonts/cyrillic:unscaled”
FontPath “/usr/share/fonts/misc/sgi:unscaled”
FontPath “/usr/share/fonts/truetype/”
EndSection

Section “Module”
Load “dri”
Load “extmod”
Load “dri2”
Load “dbe”
Load “glx”
Load “record”
Load “vnc”
EndSection

Section “InputDevice”
Identifier “Keyboard0”
Driver “kbd”
EndSection

Section “InputDevice”
Identifier “Mouse0”
Driver “mouse”
Option “Protocol” “auto”
Option “Device” “/dev/input/mice”
Option “ZAxisMapping” “4 5 6 7”
EndSection

Section “Monitor”
#DisplaySize 380 300 # mm
Identifier “Monitor0”
VendorName “SAM”
ModelName “SyncMaster”
HorizSync 30.0 - 81.0
VertRefresh 56.0 - 75.0
Option “DPMS”
EndSection

Section “Monitor”
#DisplaySize 380 300 # mm
Identifier “Monitor1”
VendorName “SAM”
ModelName “SyncMaster”
HorizSync 30.0 - 81.0
VertRefresh 56.0 - 75.0
Option “DPMS”
EndSection

Section “Device”
Identifier “Card0”
Driver “radeonhd”
VendorName “ATI Technologies Inc”
BoardName “RS880 [Radeon HD 4200]”
BusID “PCI:1:5:0”
EndSection

Section “Device”
Identifier “Card1”
Driver “radeonhd”
VendorName “ATI Technologies Inc”
BoardName “RV280 [Radeon 9200 PRO]”
BusID “PCI:3:6:0”
EndSection

Section “Screen”
Identifier “Screen0”
Device “Card0”
Monitor “Monitor0”
SubSection “Display”
Viewport 0 0
Depth 1
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 4
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 8
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 15
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 16
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 24
EndSubSection
EndSection

Section “Screen”
Identifier “Screen1”
Device “Card1”
Monitor “Monitor1”
SubSection “Display”
Viewport 0 0
Depth 1
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 4
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 8
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 15
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 16
EndSubSection
SubSection “Display”
Viewport 0 0
Depth 24
EndSubSection
EndSection

Nothing seems to make any difference.

Larry

#2

Your BIOS setting may have something to do with which card is active. In any case (I might be wrong here), I think only one card with be used with a single x-session. You’d need to activate the other in a separate x-session (manually).

#3

I can’t give you a solution, (and I’m not sure what you’re trying to achieve anyway), but these may be useful to you:

https://wiki.archlinux.org/index.php/Xorg_multiseat

Tri-head Xinerama

#4

I’d like to get a two monitor xinerama display. This worked just fine in 11.1 but
went away when I upgraded to 11.3 (a year or so ago) so I’ve been living with
limited real estate since then. I was never able to figure out why and recently
decided to give it another try. I’ll take a look at the locations you suggested.

Larry

#5

Have a look at /var/log/Xorg.0.log for any clues as to what may be failing. If you need to post the lengthy output here, use

SUSE Paste

and just post the link to it here.

#6

Funny you should mention that. I did take a look at that file this afternoon. I discovered that there was
all kinds of information about the first (onboard) card loading with RADEONHD[0] tags, but NOTHING
about the second standalone board. So, I dug deeper into the stuff that is posted about ATI cards. Turns
out the stand alone card is a 9800PRO which needs the radeon driver, not the radeonhd which I had set
both of them to. So I changed it in xorg.conf, and the bootup went catatonic. BUT, the second monitor
now showed a real video signal from the 9800PRO card, so the change had one positive effect.
But, I had to remotely log into my computer to put things back the way they were.

Digging further, I now find that the onboard video on the motherboard is HD4200 which
actually requires the fglrx driver and not radeonhd.

I think I had the wrong driver for BOTH boards – the HD4200 needs fglrx and the 9200PRO
needs radeon. So, Monday when I’m back in my office, I will download the fglrx driver, set
the (I hope) appropriate drivers for the two video sources in xorg.conf and try try again. Stay tuned
and thanks for the comments.

Larry

#7

I think I had the wrong driver for BOTH boards – the HD4200 needs fglrx and the 9200PRO
needs radeon. So, Monday when I’m back in my office, I will download the fglrx driver, set
the (I hope) appropriate drivers for the two video sources in xorg.conf and try try again. Stay tuned
and thanks for the comments.

I wish you the best success with this. I must admit I didn’t check your hardware listed in your xorg.conf, but yes, the 9200 (RV280) is consigned to legacy support via the open source radeon driver. I’m sure this will be informative to others trying to configure multiple cards with Xorg.

#8

There really needs to be an easier way to do this…

I did the following

  1. Downloaded the fglrx driver

  2. Changed my xorg.conf file so that
    a) the 4200 board used the fglrx driver
    b) the 9200 board used the radeonhd driver (this is the WRONG driver for reasons explained below)

  3. Reboot
    a) The monitor conneced to the 4200 (fglrx) came up fine (although I still needed to a “nomodeset” on boot
    b) The monitor connected to the 9200 was dead (as expected)
    c) The Xorg.0.log file showed the fglrx as [0] as expected. No mention of a second driver as expected.

  4. Now, I changed the xorg.conf file so that the 9200 was the “radeon” driver. This should be the correct
    driver and rebooted. This time it started to boot in (the progress bar got about 1/3 way across) and then
    I got a completely blank screen. Completely hung up.
    a) This time the Xorg.0.log again showed the fglrx[0] messages as before AND it showed radeon[1]
    but it just ended abruptly. The last few lines were:

    15.923] (–) RADEON(1): Chipset: “ATI Radeon 9250 5960 (AGP)” (ChipID = 0x5960)
    15.923] (–) RADEON(1): Linear framebuffer at 0x00000000d8000000
    15.923] (II) RADEON(1): PCI card detected
    15.923] (II) Loading sub module “int10”
    15.923] (II) LoadModule: “int10”
    15.924] (II) Loading /usr/lib/xorg/modules/libint10.so
    15.924] (II) Module int10: vendor=“X.Org Foundation”
    15.924] compiled for 1.8.0, module version = 1.0.0
    15.924] ABI class: X.Org Video Driver, version 7.0
    15.924] (II) RADEON(1): initializing int10

#9

This may well have to do with the proprietary driver overwriting the Xorg GL libraries ie I’m not sure whether both can exist concurrently. (Others may know more specifically about this). You may have to upgrade your 9200 graphics card to something more recent and supported by the fglrx driver. That would be the more sensible choice.

#10

Only one word – BLECHH…

I use the monitor for text, simple graphics, web browsing and an occasional video. No gaming,
no 3D. So all I want is a CHEAP basic video card. A lot of the newer cards have expensive
bells and whistles I don’t want or need. OK, it’s back to newegg.com.

So much for the rant. Thanks for the help.

Larry

#11

I don’t have my notes on a recent presentation given at my local Linux User Group recently,

During that evening complete display probe and configuration using the CLI only was shown, and it was pretty amazing.

I can only point you in that direction because I remember one of the main apps used (and AFAIK it’s installed by default in every openSUSE so it’s probably one of those “fundamental” apps others build on), use it to probe your display info. With any luck I’d guess that a Google search for multi-monitor configuration using this app as one of the keywords should also return the other app or two you would use to configure multiple monitors from the CLI

The CLI app:


$ xrandr

Good Luck,
TS

#12

I can only point you in that direction because I remember one of the main apps used (and AFAIK it’s installed by default in every openSUSE so it’s probably one of those “fundamental” apps others build on), use it to probe your display info. With any luck I’d guess that a Google search for multi-monitor configuration using this app as one of the keywords should also return the other app or two you would use to configure multiple monitors from the CLI

I don’t think it is as simple as that unfortunately. The OP is trying to use the proprietary fglrx driver and open source driver (for the other graphics card) simultaneously. However, for hardware that is auto-detected and configured correctly by the X-server, xrandr is a great utility for dynamic reporting and display configuration. Sometimes, there is no escaping manual configuration (using the X.Org config files).

#13

And for only $80.00 plus shipping I have now ordered a Radeon HD 5450 board
which is fglrx compatible to replace an otherwise perfectly good board. The
new one has HD1080 imaging plus all sorts of other stuff I have no use for.

Oh well, I hope this finally cures the problem.

P.S. What does fglrx mean anyway?

#14

OK, so I’ve now ordered a Radeon HD 5450 board which is fglrx compatible, so I
should then have two boards which use the same driver. And, hopefully (for $80
plus shipping) also fix the problem. And, of course, this new board does all sorts
of stuff I have no interest in.

BTW – fglrx is an awfully odd term. I assume it’s an acronym for something?

#15

BTW – fglrx is an awfully odd term. I assume it’s an acronym for something?

Here you go:
AMD Catalyst - Wikipedia, the free encyclopedia

#16

OK – so I got the new board (ATI CEDAR Radeon HD 5450) and installed it.

lspci sees it, so I changed my xorg.conf file to make both video boards
fglrx. I also changed the PCI address for the new board to match the
address reported by lspci. And when I booted, both screens went dead.

The Xorg.0.log file reports errors. Here’s the tail end:

16.016] (==) fglrx(1): NoDRI = NO
16.016] (==) fglrx(1): Capabilities: 0x00000000
16.016] (==) fglrx(1): CapabilitiesEx: 0x00000000
16.016] (==) fglrx(1): OpenGL ClientDriverName: "fglrx_dri.so"
16.016] (==) fglrx(1): UseFastTLS=0
16.016] (==) fglrx(1): BlockSignalsOnLock=1
16.016] (EE) fglrx(1): Multiview is only supported with acceleration; this screen will now shutdown.
16.016] (EE) fglrx(1): PreInit failed
16.016] (II) fglrx(1): === [xdl_x750_atiddxPreInit] === end
16.016] (EE) fglrx(0): Failed shutdown interrupts. Error -9
16.016] (II) UnloadModule: "fglrx"
16.016] (II) UnloadModule: "fb"
16.016] (II) UnloadModule: "fglrxdrm"
16.016] (II) UnloadModule: "fglrxdrm"
16.016] (--) Depth 24 pixmap format is 32 bpp
16.018] (EE) fglrx(0): atiddxDriScreenInit failed, GPS not been initialized. 
16.018] (WW) fglrx(0): ***********************************************************
16.018] (WW) fglrx(0): * DRI initialization failed                               *
16.018] (WW) fglrx(0): * kernel module (fglrx.ko) may be missing or incompatible *
16.018] (WW) fglrx(0): * 2D and 3D acceleration disabled                         *
16.018] (WW) fglrx(0): ***********************************************************
16.018] (II) fglrx(0): FBADPhys: 0xc4900000 FBMappedSize: 0x10000000
16.018] (II) fglrx(0): Reserved 0x04900000 bytes of sideport memory for power saving
16.018] (EE) fglrx(0): Failed to map FB memory
16.018] (II) fglrx(0): === [xdl_x750_atiddxScreenInit] === end
16.018]

Fatal server error:
16.018] AddScreen/ScreenInit failed for driver 0
16.018]
16.018]

So what’s wrong now?

Larry

#17

I’m not sure. Did you reinstall the fglrx driver? What method did you use?

First, I want to point you at this thread:

http://forums.opensuse.org/english/other-forums/development/programming-scripting/449058-upgrading-ati-driver-atiupgrade-10.html#post2413818

User please-try-again is very familiar with ATI hardware and has created a script which simplifies installing and upgrading the proprietary ATI driver. He may even have a few clues as to what specifically might be the issue here. I don’t have your hardware to test unfortunately. Those errors concerning DRI don’t look promising.

Next, I suggest you do some research on the errors. For example, I do know a quick search using some of those error messages shows others with similar problems where the particular card does not play nicely with a particular revision driver.

[linux] [12.1] drivers installation - doesn’t w… | AMD Developer Forums](http://devgurus.amd.com/message/1279098)

Just as an aside: Are your BusIDs for both cards still correct? (lspci will give you that info)

#18

I installed fglrx using yast2 and lspci gives an address of 01:05.0 for the motherboard
(HD 4200) and 04:00.0 for the (new) HD5450 board. I will check the locations you suggested
but this may have to wait a few days. I’m on a night schedule for the next two nights
and things are a bit disorganized right now. I note that I had to change the bus address
in xorg.conf for the 5450 since it shows up at a different PCI address than the old board
(which was in the same physical slot).

#19

I know you went an got a newer card, but HD4200 works fine with the radeon driver, so it does not require the proprietary fglrx. Was surprised to see that the radeonhd driver doesn’t support that adapter (I checked “man radeonhd” and, indeed, its not listed), but radeonhd was effectively abandoned a good awhile back, so maybe support for that board never made it in before the sun set on it. Anyway, the radeon driver should be the defacto driver, and it does indeed support both those devices … as well as your new adapter too (see “man radeon”). … That said, I have no experience attempting to get the radeon driver to function with multiple device adapters, which is, of course, the central goal of your endeavours.

I believe fglrx dropped support for the older card(s) that you were formerly trying to get to work (9x00PRO) a while back. fglrx will indeed support the HD4200 and your newer adapter too.

#20

Sorry, I’ll add some further clarification to that post, just in case anyone is unaware:

There is a distinct difference between both open source “radeon” and “radeonhd” drivers … in terms of development history, device support, and functionality. Anyway, as previously implied, “radeonhd” is all but dead, whereas the “radeon” is heavily developed and maintained.

fglrx, on the other hand, is the closed source, proprietary AMD/ATI driver offering