2nd Monitor Shutoff upon grub to Leap on During Boot

[mrmazda]

This works as expected, to set monitor locations. Thanks!

You're welcome! It's always nice to know when an attempt to help produces a working solution!

[linuxvinh]

For EDID...

xrandr --verbose

...which gave me EDID for DVI-D-1 of...

EDID:  
                00ffffffffffff0015c3512517461402 
                28190103803c2278ea1df5ae4f35b325 
                0d5054a10800a94081808140d1c00101 
                010101010101565e00a0a0a029503020 
                350055502100001a283c80a070b02340 
                3020360055502100001a000000fd0017 
                3f1a5919000a202020202020000000fc 
                0043473237370a20202020202020006e

...which seems really short compared to HDMI-A-1...

EDID:
                00ffffffffffff0010acbaa04c445930 
                2e1a010380342078ea0495a9554d9d26 
                105054a54b00714f8180a940d1c0d100 
                010101010101283c80a070b023403020 
                360006442100001e000000ff00434656 
                394e3642383059444c0a000000fc0044 
                454c4c2055323431350a2020000000fd 
                00313d1e5311000a202020202020014a 
                020322f14f9005040302071601141f12 
                132021222309070765030c0010008301 
                0000023a801871382d40582c45000644 
                2100001e011d8018711c1620582c2500 
                06442100009e011d007251d01e206e28 
                550006442100001e8c0ad08a20e02d10 
                103e9600064421000018000000000000 
                00000000000000000000000000000082

I took DVI-D-1 EDID and made a text file, edidDVI-D-1.txt, contents of...

00ffffffffffff0015c351251746140228190103803c2278ea1df5ae4f35b3250d5054a10800a94081808140d1c00101010101010101565e00a0a0a029503020350055502100001a283c80a070b023403020360055502100001a000000fd00173f1a5919000a202020202020000000fc00434732373701a20202020202020006e

...then used...

% xxd -r -p edidDVI-D-1.txt edidDVI-D-1.bin

To check...

% parse-edid < edidDVI-D-1.bin
Checksum Correct 

Section "Monitor" 
        Identifier "CG277" 
        ModelName "CG277" 
        VendorName "ENC" 
        # Monitor Manufactured week 40 of 2015 
        # EDID version 1.3 
        # Digital Display 
        DisplaySize 600 340 
        Gamma 2.20 
        Option "DPMS" "true" 
        Horizsync 26-89 
        VertRefresh 23-63 
        # Maximum pixel clock is 250MHz 
        #Not giving standard mode: 1600x1200, 60Hz 
        #Not giving standard mode: 1280x1024, 60Hz 
        #Not giving standard mode: 1280x960, 60Hz 
        #Not giving standard mode: 1920x1080, 60Hz 
        Modeline        "Mode 0" +hsync -vsync  
        Modeline        "Mode 1" +hsync -vsync  
EndSection

...which falls well short of all the available modes reported by xrandr. HDMI-A-1's EDID looks like this in comparison...

% parse-edid < edidHDMI-A-1.bin 
Checksum Correct 

Section "Monitor" 
        Identifier "DELL U2415" 
        ModelName "DELL U2415" 
        VendorName "DEL" 
        # Monitor Manufactured week 46 of 2016 
        # EDID version 1.3 
        # Digital Display 
        DisplaySize 520 320 
        Gamma 2.20 
        Option "DPMS" "true" 
        Horizsync 30-83 
        VertRefresh 49-61 
        # Maximum pixel clock is 170MHz 
        #Not giving standard mode: 1152x864, 75Hz 
        #Not giving standard mode: 1280x1024, 60Hz 
        #Not giving standard mode: 1600x1200, 60Hz 
        #Not giving standard mode: 1920x1080, 60Hz 
        #Not giving standard mode: 1920x1200, 60Hz 

        #Extension block found. Parsing... 
        Modeline        "Mode 16" +hsync +vsync  
        Modeline        "Mode 0" +hsync +vsync  
        Modeline        "Mode 1" 148.500 1920 2008 2052 2200 1080 1084 1089 112
5 +hsync +vsync 
        Modeline        "Mode 2" 74.250 1920 2008 2052 2200 1080 1082 1087 1125
 +hsync +vsync interlace 
        Modeline        "Mode 3" 74.250 1280 1390 1420 1650 720 725 730 750 +hs
ync +vsync 
        Modeline        "Mode 4" 27.027 720 736 798 858 480 489 495 525 -hsync 
-vsync 
        Modeline        "Mode 5" 27.027 720 736 798 858 480 489 495 525 -hsync 
-vsync 
        Modeline        "Mode 6" 27.027 1440 1478 1602 1716 480 484 487 525 -hs
ync -vsync interlace 
        Modeline        "Mode 7" 27.000 1440 1464 1590 1728 576 578 581 625 -hs
ync -vsync interlace 
        Modeline        "Mode 8" 25.200 640 656 752 800 480 490 492 525 -hsync 
-vsync 
        Modeline        "Mode 9" 74.250 1920 2448 2492 2640 1080 1082 1089 1125
 +hsync +vsync interlace 
        Modeline        "Mode 10" 148.500 1920 2448 2492 2640 1080 1084 1089 11
25 +hsync +vsync 
        Modeline        "Mode 11" 27.000 720 732 796 864 576 581 586 625 -hsync
 -vsync 
        Modeline        "Mode 12" 74.250 1280 1720 1760 1980 720 725 730 750 +h
sync +vsync 
        Modeline        "Mode 13" 74.250 1920 2558 2602 2750 1080 1084 1089 112
5 +hsync +vsync 
        Modeline        "Mode 14" 74.250 1920 2448 2492 2640 1080 1084 1089 112
5 +hsync +vsync 
        Modeline        "Mode 15" 74.250 1920 2008 2052 2200 1080 1084 1089 112
5 +hsync +vsync 
        Modeline        "Mode 17" +hsync +vsync interlace 
        Modeline        "Mode 18" +hsync +vsync  
        Modeline        "Mode 19" -hsync -vsync  
        Option "PreferredMode" "Mode 16" 
EndSection

Basically, the EDID looks incorrect.

So, one problem I'm currently facing is that I can't get the proper EDID, even with monitor 2 directly connected to DVI-D-1.

-----

The other problem I'm curious about is where to put the EDID binary file for grub Kernel command line to find.
/lib/firmware/edid seems like the likely place, as /lib/firmware exists, whereas /usr/lib/firmware, referenced in some docs, doesn't with openSUSE.

/lib/firmware/edid/edidDVI-D-1.bin is ultimately what I settled upon, but dmesg...

[    5.470960] [drm] forcing DVI-D-1 connector on 
[    5.471023] platform DVI-D-1: Direct firmware load for edid/edidDVI-D-1.bin failed with error -2 
[    5.471064] [drm:edid_load [drm]] *ERROR* Requesting EDID firmware "edid/edidDVI-D-1.bin" failed (err=-2)

So the binary isn't in the right place or the binary is bad?

[linuxvinh]

You're welcome! It's always nice to know when an attempt to help produces a working solution!

Sadly, I'm still struggling with DVI-D-0 or DVI-D-1 to monitor 2 working through the KVM...

[linuxvinh]

Current connection DVI-D-1 <-> KVM port 1 <-> monitor 2

Should xrandr be able to turn on outputs regardless of whether a cable is connected or not? i.e. force turn on no matter what?

An odd thing I'm finding is even if DVI-D-1 has a cable connected,

xrandr --output DVI-D-1 --auto

still says...

DVI-D-1 disconnected (normal left inverted right x axis y axis)

...with a follow-up xrandr.

[linuxvinh]

Well, since I only have to do it once in the morning... the hacky workaround...

1. Boot to HDMI-A-0 with HDMI-A-0 and DVI-D-1 connected to KVM, with DVI-D-1 "disconnected" according to the OS
2. Disconnect monitor 2 from KVM and connect to DVI-D-1 (basically DVI-D-1 <-> monitor 2, bypassing KVM altogether)
3. xrandr --output HDMI-A-0 --primary --output DVI-D-1 --auto --left-of HDMI-A-0
4. Connect monitor 2 back to KVM, DVI-D-1 to KVM

monitor 2, at this point, according to xrandr, is still "disconnected" and displays nothing, but it's virtually there...

DVI-D-1 disconnected 2560x1440+0+0 (normal left inverted right x axis y axis) 0
mm x 0mm 
  2560x1440 (0x192) 241.500MHz +HSync -VSync 
        h: width  2560 start 2608 end 2640 total 2720 skew    0 clock  88.79KHz 
        v: height 1440 start 1443 end 1448 total 1481           clock  59.95Hz

I have a use for this, even if I can't see the display.

[linuxvinh]

To summarize, it's possible to use xrandr to initialize the graphics card to use DVI-D-1 (or DVI-D-0), if monitor 2 is directly connected. Once going through the KVM, monitor 2 is no longer recognized as being connected. I tried two different cables with the same result.

Still searching for a solution...

[linuxvinh]

amdgpu.dc = 0 in the grub Kernel command line allows monitor2 to work, but what does this ultimately mean in terms of toggling on/off features?

[linuxvinh]

In post #15, I listed my workaround that required unplugging and plugging in the cables between the KVM and DVI-D-1 directly. It was cumbersome. I've adopted doing this instead...

xrandr --newmode "2560x1440" 241.50 2560 2608 2640 2720 1440 1443 1448 1481 +hsync -vsync
xrandr --addmode DVI-D-1 "2560x1440"
xrandr --output HDMI-A-0 --primary --output DVI-D-1 --mode "2560x1440" --left-of HDMI-A-0

• The first line adds the 2560x1440 mode that my monitor accepts.
• The second line adds the mode to DVI-D-1. (There's a -d xrandr flag that supposed to allow the first line to define the newmode for DVI-D-1, but with the monitor considered "disconnected", it won't allow it.)
• The third line turns on both displays, sets HDMI-A-0 to primary, DVI-D-1 to use 2560x1440 mode and to be left of HDMI-A-0.

xrandr still reports that DVI-D-1 is "disconnected" so the display is blank, as the graphics card isn't outputting.

Is there a way to force output, whether the display is actually "disconnected" or not?

[mrmazda]

Is there a way to force output, whether the display is actually "disconnected" or not?

/etc/X11/xorg.con* might do it. Getting it right with multiple displays is tricky business I never got fully understood. Hwinfo --monitor should produce something similar to the following:

    Vert. Sync Range: 55-75 Hz
    Hor. Sync Range: 30-80 kHz

X will calculate modelines from these thus:

Section "Monitor"
  Identifier "Monitor[2]"
  HorizSync 30-80
  VertRefresh 55-75
EndSection

I am seriously rusty on use of xorg.con* with multiple displays, since I've found xrandr to be much simpler and have not often attempted use of xorg.con* in recent years. These references may be helpful in lieu:

• https://nouveau.freedesktop.org/Mult...orDesktop.html
• https://wiki.archlinux.org/title/Xorg

[linuxvinh]

/etc/X11/xorg.con* might do it. Getting it right with multiple displays is tricky business I never got fully understood. Hwinfo --monitor should produce something similar to the following:

    Vert. Sync Range: 55-75 Hz
    Hor. Sync Range: 30-80 kHz

X will calculate modelines from these thus:

Section "Monitor"
  Identifier "Monitor[2]"
  HorizSync 30-80
  VertRefresh 55-75
EndSection

I am seriously rusty on use of xorg.con* with multiple displays, since I've found xrandr to be much simpler and have not often attempted use of xorg.con* in recent years. These references may be helpful in lieu:

• https://nouveau.freedesktop.org/Mult...orDesktop.html
• https://wiki.archlinux.org/title/Xorg

Thank you again, @mrmazda, for taking the time to respond.

The solution in the previous post was derived from your suggestions.

-----

I think I've finally settled upon a solution that requires the least effort on my part, and that is only adding...

amdgpu.dc=0

to Grub.

DVI-D-1 is seen as connected through the KVM, just like before.

DVI-D-1 connected 2560x1440+0+0 (normal left inverted right x axis y axis) 597m
m x 336mm

This also means no video or edid helpers Grub commands I've tried in previous posts. It's just adding amdgpu.dc.

I am still unclear what the disadvantage this means by turning off AMD GPU DC, but this post only mentions something about sound control over DisplayPort...

https://github.com/rockowitz/ddcutil/issues/99

The official Kernel and AMD page isn't clear...

https://www.kernel.org/doc/html/late...amdgpu-dc.html

And it seems like kernel updates toggle on/off DC by default, so perhaps amdgpu.dc was 0 before and a recent update toggled it to 1.