Monitors identity swapped (for taskbar display and idle dimming) with null EDID

[jkopczyn]

Distribution

Mint 22.3

Package version

6.2.9

Graphics hardware in use

Intel Corporation Alder Lake-P Integrated Graphics Controller

Frequency

Always

Bug description

With an external monitor connected to my laptop (Framework 2022) by HDMI and Wayland active, taskbar display doesn't match the Primary/Secondary assignment (I understand this to be a known bug.) and screen dimming from inactivity also seems to be mismatched. (I believe this is not.) No amount of cursor movement on the laptop screen will rouse the screen to full brightness; moving onto the external screen will do so immediately.

Relevant: Frameworks' modular ports have a known issue with their HDMI ports where all EDID information is lost. When I connect with a different method (USB-C hub), I don't see this problem.

Steps to reproduce

1,2. (In either order, doesn't matter.)

• connect the monitor to the laptop without an EDID reported
• log in with Wayland
  (The taskbar should be wrong immediately.)
  (Optional: Open Display and see that primary/secondary are assigned correctly. Reset to Defaults says embedded as Primary, external as Secondary, taskbar shows up on external.)

3. Set the screen dimming timer to a sufficiently low time (I had 90 seconds)
4. Let the cursor sit for that long on the embedded screen.
5. When the embedded screen dims, move the cursor on the embedded screen.
   (This should have no effect.)
6. Move it to the external screen.
   (This should immediately end the dimmed screen.)

Expected behavior

Motion on the embedded screen should wake it from dim, primary monitor assignment should match with taskbar location.

Additional information

No response

[leigh123linux]

Relevant: Frameworks' modular ports have a known issue with their HDMI ports where all EDID information is lost. When I connect with a different method (USB-C hub), I don't see this problem.

I doubt your issue has a solution within cinnamon, muffin relies on the EDID.
Try RMA the broken hardware.

[jkopczyn]

Not possible. This isn't a problem of _failing_ hardware; it does this when new. No third-party module that lacks the issue exists on the market AFAIK, though the specs to create such a thing are deliberately available. I expect this bug report is useful primarily for locating problems that other users see unreliably in circumstances they can't reproduce, rather than for fixing the bug I see myself. Though generally handling missing/malformed EDIDs gracefully would be a property any mature compositor would possess. The X11 version does.

…

On Mon, Mar 23, 2026, 23:57 Leigh Scott ***@***.***> wrote: *leigh123linux* left a comment (linuxmint/wayland#196) <#196 (comment)> Relevant: Frameworks' modular ports have a known issue with their HDMI ports where all EDID information is lost. When I connect with a different method (USB-C hub), I don't see this problem. I doubt your issue has a solution within cinnamon, muffin relies on the EDID. Try RMA the broken hardware. — Reply to this email directly, view it on GitHub <#196?email_source=notifications&email_token=AAJR7ZY3WNLE6R5QZZOVQDT4SIWVHA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJRGU4DMNRWG442M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4115866679>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJR7Z36GLWJBKQASN5NSIL4SIWVHAVCNFSM6AAAAACW465NHCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCMJVHA3DMNRXHE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[leigh123linux]

It is a fundamental issue with how the Wayland compositor (and the underlying libinput and drm drivers) handles input-to-display mapping when a monitor reports a null or invalid EDID. [1, 2]
Because the system cannot uniquely identify the external monitor, it often fails to correctly register which "logical" screen the cursor is on for power management events. Consequently, the idle daemon (like mutter in GNOME or KWin in KDE) may only recognize "activity" if the cursor is on what it considers the primary display path, even if your settings say otherwise. [3, 4, 5]
Why this happens across DEs

• Idle Management: Most Wayland compositors use idle-wayland protocols to track user activity. When a display has no EDID, it is often assigned a generic "Unknown" or "Internal" name, which causes a conflict with the laptop's actual internal display in the system's power management logic.
• Coordinate Mismatch: The compositor might be "seeing" the cursor movement on the embedded screen, but because the display mapping is broken by the missing EDID, it doesn't associate those coordinates with a "wake" event for the dimmed state.
• Taskbar Alignment: The taskbar (or panel) often defaults to the first display index (index 0) in the absence of valid metadata. Without a unique EDID to "pin" settings to, the system resets to this default index on every login or wake. [1, 4, 6, 7, 8]

The most reliable fix: EDID Spoofing
Since you already confirmed it works with a USB-C hub, the most robust way to solve this—regardless of which desktop environment (DE) you use—is to force the system to "see" a valid EDID for that HDMI port at the kernel level. This bypasses the Framework hardware's reporting issue entirely.

1. Extract a working EDID: While connected via the USB-C hub (where it works), use the [NVIDIA Control Panel](https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm) or a tool like edid-decode on Linux to save the EDID as a .bin file.
2. Apply at Boot: Place that file in /lib/firmware/edid/your_monitor.bin.
3. Kernel Parameter: Add drm.edid_firmware=HDMI-A-1:edid/your_monitor.bin to your GRUB configuration (replace HDMI-A-1 with your actual port name found via xrandr or wayland-info). [9, 10, 11]

This "fakes" a perfect hardware connection, which should immediately fix both the taskbar placement and the dimming-wake issue in Cinnamon, KDE, or GNOME.

[1] [https://bbs.archlinux.org](https://bbs.archlinux.org/viewtopic.php?id=163216)
[2] [https://www.extron.com](https://www.extron.com/article/uedid)
[3] [https://www.reddit.com](https://www.reddit.com/r/kde/comments/16npjma/wayland_bug_external_monitor_edge_cursor_wonkiness/)
[4] [https://community.frame.work](https://community.frame.work/t/responded-wayland-multi-monitor/33778#:~:text=I%20was%20wondering%20if%20anyone%20has%20managed,stay%20black%20no%20matter%20what%20I%20do.)
[5] [https://www.reddit.com](https://www.reddit.com/r/linuxsucks/comments/1pnybs9/switched_from_cinnamon_to_kde_signal_desktop/)
[6] [https://forums.linuxmint.com](https://forums.linuxmint.com/viewtopic.php?t=437278)
[7] [https://github.com](swaywm/wlr-protocols#55)
[8] [https://github.com](sddm/sddm#1696)
[9] [https://bbs.archlinux.org](https://bbs.archlinux.org/viewtopic.php?id=270616)
[10] [https://www.nvidia.com](https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm#:~:text=At%20the%20EDID%20Source%20property%2C%20click%20Monitor.,save%20the%20EDID%20information%20in%20a%20file.)
[11] [https://www.reddit.com](https://www.reddit.com/r/linux_gaming/comments/1gl81zq/reported_modes_incomplete_unable_to_set_custom/)

[jkopczyn]

I'm aware of the EDID override methods, and I'm already doing that to resolve the X problems that made me try a switch. There are enough other problems with Wayland (crashes that don't seem to be novel to me and so not worth reporting) that I'm not further pursuing getting Wayland working. It remains the case that mature display software would handle this gracefully. None of those 'fundamental' issues are necessary; they can be caught by error handling. This is a rare bug and not a high priority, but mature, *finished* software would catch such errors and handle them gracefully. I'm not familiar with the full parameters of a minimal valid (if useless) EDID, but one which conveys only the information the software can deduce without it, made and used temporarily, seems like the place I would start if I was trying to fix it. Many of these params look to be arbitrary without a sensible zero value, but it's picking arbitrary values anyway, better to systematize it.

…

On Tue, Mar 24, 2026, 08:57 Leigh Scott ***@***.***> wrote: *leigh123linux* left a comment (linuxmint/wayland#196) <#196 (comment)> It is a fundamental issue with how the Wayland compositor (and the underlying libinput and drm drivers) handles input-to-display mapping when a monitor reports a null or invalid EDID. [1, 2] Because the system cannot uniquely identify the external monitor, it often fails to correctly register which "logical" screen the cursor is on for power management events. Consequently, the idle daemon (like mutter in GNOME or KWin in KDE) may only recognize "activity" if the cursor is on what it considers the primary display path, even if your settings say otherwise. [3, 4, 5] Why this happens across DEs - Idle Management: Most Wayland compositors use idle-wayland protocols to track user activity. When a display has no EDID, it is often assigned a generic "Unknown" or "Internal" name, which causes a conflict with the laptop's actual internal display in the system's power management logic. - Coordinate Mismatch: The compositor might be "seeing" the cursor movement on the embedded screen, but because the display mapping is broken by the missing EDID, it doesn't associate those coordinates with a "wake" event for the dimmed state. - Taskbar Alignment: The taskbar (or panel) often defaults to the first display index (index 0) in the absence of valid metadata. Without a unique EDID to "pin" settings to, the system resets to this default index on every login or wake. [1, 4, 6, 7, 8] The most reliable fix: EDID Spoofing Since you already confirmed it works with a USB-C hub, the most robust way to solve this—regardless of which desktop environment (DE) you use—is to force the system to "see" a valid EDID for that HDMI port at the kernel level. This bypasses the Framework hardware's reporting issue entirely. 1. Extract a working EDID: While connected via the USB-C hub (where it works), use the [NVIDIA Control Panel <https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm> ]( https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm) or a tool like edid-decode on Linux to save the EDID as a .bin file. 2. Apply at Boot: Place that file in /lib/firmware/edid/your_monitor.bin. 3. Kernel Parameter: Add drm.edid_firmware=HDMI-A-1:edid/your_monitor.bin to your GRUB configuration (replace HDMI-A-1 with your actual port name found via xrandr or wayland-info). [9, 10, 11] This "fakes" a perfect hardware connection, which should immediately fix both the taskbar placement and the dimming-wake issue in Cinnamon, KDE, or GNOME. [1] [https://bbs.archlinux.org <https://bbs.archlinux.org/viewtopic.php?id=163216>]( https://bbs.archlinux.org/viewtopic.php?id=163216) [2] [https://www.extron.com <https://www.extron.com/article/uedid>]( https://www.extron.com/article/uedid) [3] [https://www.reddit.com <https://www.reddit.com/r/kde/comments/16npjma/wayland_bug_external_monitor_edge_cursor_wonkiness/> ]( https://www.reddit.com/r/kde/comments/16npjma/wayland_bug_external_monitor_edge_cursor_wonkiness/ ) [4] [https://community.frame.work <https://community.frame.work/t/responded-wayland-multi-monitor/33778#:~:text=I%20was%20wondering%20if%20anyone%20has%20managed,stay%20black%20no%20matter%20what%20I%20do.> ]( https://community.frame.work/t/responded-wayland-multi-monitor/33778#:~:text=I%20was%20wondering%20if%20anyone%20has%20managed,stay%20black%20no%20matter%20what%20I%20do .) [5] [https://www.reddit.com <https://www.reddit.com/r/linuxsucks/comments/1pnybs9/switched_from_cinnamon_to_kde_signal_desktop/> ]( https://www.reddit.com/r/linuxsucks/comments/1pnybs9/switched_from_cinnamon_to_kde_signal_desktop/ ) [6] [https://forums.linuxmint.com <https://forums.linuxmint.com/viewtopic.php?t=437278>]( https://forums.linuxmint.com/viewtopic.php?t=437278) [7] [https://github.com <swaywm/wlr-protocols#55>]( swaywm/wlr-protocols#55 <swaywm/wlr-protocols#55>) [8] [https://github.com <sddm/sddm#1696>]( sddm/sddm#1696 <sddm/sddm#1696>) [9] [https://bbs.archlinux.org <https://bbs.archlinux.org/viewtopic.php?id=270616>]( https://bbs.archlinux.org/viewtopic.php?id=270616) [10] [https://www.nvidia.com <https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm#:~:text=At%20the%20EDID%20Source%20property%2C%20click%20Monitor.,save%20the%20EDID%20information%20in%20a%20file.> ]( https://www.nvidia.com/content/Control-Panel-Help/vLatest/en-us/mergedProjects/nvwks/To_get_an_EDID_from_a_display_and_save_it.htm#:~:text=At%20the%20EDID%20Source%20property%2C%20click%20Monitor.,save%20the%20EDID%20information%20in%20a%20file .) [11] [https://www.reddit.com <https://www.reddit.com/r/linux_gaming/comments/1gl81zq/reported_modes_incomplete_unable_to_set_custom/> ]( https://www.reddit.com/r/linux_gaming/comments/1gl81zq/reported_modes_incomplete_unable_to_set_custom/ ) — Reply to this email directly, view it on GitHub <#196?email_source=notifications&email_token=AAJR7Z2L2XPI4PHCAYZI74L4SKV5PA5CNFSNUABFM5UWIORPF5TWS5BNNB2WEL2JONZXKZKDN5WW2ZLOOQXTIMJRHE2DEOBYGEZ2M4TFMFZW63VGMF2XI2DPOKSWK5TFNZ2KYZTPN52GK4S7MNWGSY3L#issuecomment-4119428813>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAJR7Z6VIT3SOSF2PNNNLOD4SKV5PAVCNFSM6AAAAACW465NHCVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHM2DCMJZGQZDQOBRGM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[mtwebster]

There's definitely an issue with monitor assignment in Wayland sessions - I use three monitors on my desktop and their layout differs between x11 and wayland. I don't know if this is directly related to the issue reported here, but it's something we'll work out eventually, it just isn't a priority for now.

The cinnamon-settings-daemon power plugin handles idle-dim, and did not really work in wayland sessions until linuxmint/cinnamon-settings-daemon@bb4287d (which is recent, and not in a release yet).

I've tested it in wayland sessions on my laptop (new Framework 16), and it definitely works, though I haven't hooked a second monitor up to it yet.