Skip to content

modesetting: SIGSEGV in modesetting_drv.so via drmHandleEvent on idle Intel Iris Xe + DP-MST dual-head (Atomic+TearFree, XLibre 1.25.2.2) #3709

New issue
New issue
Open
Open
modesetting: SIGSEGV in modesetting_drv.so via drmHandleEvent on idle Intel Iris Xe + DP-MST dual-head (Atomic+TearFree, XLibre 1.25.2.2)#3709

Description

@KarlIsWright
KarlIsWright
opened on Sep 20, 2026

Select the version

25.2.X

Describe your issue

XLibre 1.25.2.2 (Void package xlibre-xserver-25.2.2_1) crashed with SIGSEGV in modesetting_drv.so while handling a DRM event via drmHandleEvent, after the session had been idle overnight.

This was a single-user Plasma/SDDM X11 session on Intel Iris Xe (Alder Lake). The machine itself did not reboot; only the X server died, which tore down the Plasma session and returned SDDM to the greeter.

At crash time the modesetting driver had:

  • Option "Atomic" "true"
  • Option "TearFree" "true"
  • glamor / OpenGL 4.6 / DRI3 on Mesa Intel Iris Xe (ADL GT2)
  • dual external monitors over DP-MST: DP-1-1 and DP-1-2 at 1920x1080@60, laptop eDP-1 connected

Crash signature (from Xorg log):

Segmentation fault at address 0x559266cacffd
Fatal server error:
Caught signal 11 (Segmentation fault). Server aborting
(EE) Server terminated with error (1). Closing log file.

Backtrace (no debug symbols in distro package):

[2026-09-20 01:08:04] (EE) event9  - Keychron Keychron M1 Mouse: client bug: event processing lagging behind by 25ms, your system is too slow
[2026-09-20 02:04:27] (EE) event9  - Keychron Keychron M1 Mouse: client bug: event processing lagging behind by 28ms, your system is too slow
[2026-09-20 02:05:43] (II) event9  - Keychron Keychron M1 Mouse: SYN_DROPPED event - some input events have been lost.
[2026-09-20 02:10:46] (EE) event9  - Keychron Keychron M1 Mouse: client bug: event processing lagging behind by 22ms, your system is too slow
[2026-09-20 02:11:48] (EE) event9  - Keychron Keychron M1 Mouse: client bug: event processing lagging behind by 39ms, your system is too slow
[2026-09-20 02:18:52] (EE) event9  - Keychron Keychron M1 Mouse: client bug: event processing lagging behind by 27ms, your system is too slow
[2026-09-20 05:37:14] (WW) modeset(0): flip queue retry
[2026-09-20 05:37:14] 
[2026-09-20 05:37:14] Backtrace:
[2026-09-20 05:37:14] 0: /usr/libexec/Xorg (xorg_backtrace+0x50) [0x559723f45630]
[2026-09-20 05:37:14] 1: /usr/libexec/Xorg (0x559723e24000+0x121a49) [0x559723f45a49]
[2026-09-20 05:37:14] 2: /usr/lib/libc.so.6 (0x7fc8b1b97000+0x40110) [0x7fc8b1bd7110]
[2026-09-20 05:37:14] 3: /usr/lib64/xorg/modules/xlibre-25/drivers/modesetting_drv.so (0x7fc8b1651000+0x17251) [0x7fc8b1668251]
[2026-09-20 05:37:14] 4: /usr/lib/libdrm.so.2 (drmHandleEvent+0xbc) [0x7fc8b1d94c5c]
[2026-09-20 05:37:14] 5: /usr/libexec/Xorg (0x559723e24000+0x11fd84) [0x559723f43d84]
[2026-09-20 05:37:14] 6: /usr/libexec/Xorg (0x559723e24000+0x63440) [0x559723e87440]
[2026-09-20 05:37:14] 7: /usr/libexec/Xorg (0x559723e24000+0x2bd4a) [0x559723e4fd4a]
[2026-09-20 05:37:14] 8: /usr/lib/libc.so.6 (0x7fc8b1b97000+0x2abfc) [0x7fc8b1bc1bfc]
[2026-09-20 05:37:14] 9: /usr/lib/libc.so.6 (__libc_start_main+0x85) [0x7fc8b1bc1cb5]
[2026-09-20 05:37:14] 10: /usr/libexec/Xorg (_start+0x21) [0x559723e51931]
[2026-09-20 05:37:14] 
[2026-09-20 05:37:14] Segmentation fault at address 0x559266cacffd
[2026-09-20 05:37:14] 
Fatal server error:
[2026-09-20 05:37:14] Caught signal 11 (Segmentation fault). Server aborting
[2026-09-20 05:37:14] 
[2026-09-20 05:37:14] 
Please consult the XLibre support: https://www.xlibre.net/
[2026-09-20 05:37:14] Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[2026-09-20 05:37:14] 
[2026-09-20 05:37:15] (EE) Server terminated with error (1). Closing log file.

Related issues checked (similar stack, different trigger/context — not a duplicate of this idle dual-head case):

I am also seeing ongoing screen tearing / stuttering on this same Intel + DP-MST dual-head setup under KWin X11, with TearFree enabled. I am reporting the crash as the hard bug; tearing may be related or separate.

Steps to reproduce

Not hard-reproducible on demand yet. Observed sequence:

  1. Void Linux, XLibre 1.25.2.2, SDDM, Plasma 6 X11 (kwin_x11).
  2. Intel Alder Lake Iris Xe only (no dGPU).
  3. Docked dual external displays as DP-MST sinks (DP-1-1, DP-1-2) at 1920x1080@60.
  4. /etc/X11/xorg.conf.d/20-modesetting.conf:
Section "Device"
    Identifier "Intel modesetting"
    Driver     "modesetting"
    Option     "TearFree" "true"
    Option     "Atomic"   "true"
EndSection
  1. Log into Plasma X11, leave the session running overnight (no logout, machine stays powered on).
  2. At ~05:37 local time, Xorg aborts with SIGSEGV in modesetting_drv.so / drmHandleEvent.
  3. SDDM recovers X and shows the greeter; previous Plasma session is gone.

I do not yet have a reliable interactive reproduction (e.g. specific hotplug/unplug or app action). Happy to gather a symbolized gdb backtrace if maintainers want that and can suggest the preferred debug build approach for this package.

What did you expect?

XLibre should remain stable while idle on a single-GPU Intel modesetting setup with dual DP-MST outputs and Atomic/TearFree enabled. A DRM event should not crash the server and destroy the session.

Additional Information

Environment

Component Version
Distro Void Linux (x86_64)
XLibre 1.25.2.2 (xlibre-xserver-25.2.2_1)
Kernel 7.2.6_1
Mesa 26.1.8_1
libdrm 2.4.134_1
DE Plasma 6.7.5 / KWin 6.7.5 (kwin_x11)
DM SDDM 0.21.0
GPU Intel Alder Lake-P GT2 Iris Xe [8086:46a6] (driver i915; xe module also loaded)

Display layout at crash

  • eDP-1 connected (laptop panel)
  • DP-1-1 connected primary 1920x1080+0+0
  • DP-1-2 connected 1920x1080+1920+0
  • (DP-MST style connector names)

Xorg log highlights from crashed session

XLibre X Server 1.25.2.2
(II) modeset(0): using drv /dev/dri/card0
(**) modeset(0): Option "Atomic" "true"
(**) modeset(0): Option "TearFree" "true"
(II) modeset(0): Atomic modesetting enabled
(II) modeset(0): TearFree: enabled
(II) glamor: Using OpenGL 4.6 context.
(II) glamor: DRI3 X acceleration enabled on Mesa Intel(R) Iris(R) Xe Graphics (ADL GT2)

Session started: 2026-09-19 14:13:45
Crash: 2026-09-20 05:37:14
Uptime continued; no system reboot.

Notes

  • Full anonymized Xorg crash log attached: Xorg.crash-modesetting-drmHandleEvent.log
  • Distro build has no debug symbols / unwind info is limited.
  • After the crash I switched Atomic to false (TearFree still true) as an A/B test; will report if that changes crash frequency or tearing.
  • No second user session / switch-user was in use (unlike X session crashes when a second graphical login is started with sddm #1065).
  • Hostnames/usernames/disk UUIDs redacted from the attached log.

Extra fields

Activity

  1. KarlIsWright commented on Sep 20, 2026

    @KarlIsWright
    KarlIsWright
    on Sep 20, 2026
    Author

    Full crash log

    Anonymized full Xorg log from the crashed session:

  2. cepelinas9000 commented on Sep 21, 2026

    @cepelinas9000
    cepelinas9000
    on Sep 21, 2026
    Contributor

    @KarlIsWright we need with debug info to figure what problem is and to debug it need second pc.

    To make with debug info:

    1. Compile XServer debug build and libinput driver (you can get avoid if can simply triggered).
    2. Depends what need to be happens, set display manager to start compiled version.
    3. when X started attach gdb (start gdb in tmux or similar screen program), then set handle all nostop to ignore signals (see: https://sourceware.org/gdb/current/onlinedocs/gdb.html/Signals.html).
    4. wait.
    5. get backtrace.., maybe do coredump

    to compile: https://github.com/X11Libre/xserver/wiki/Building-XLibre

    on other hand if you can trigger without DE running, it is enough in tmux from other pc start:

    # gdb --args /tmp/xlibre-path/bin/X :1 -retro`

    and wait for bug. (add -retro parameter, because otherwise it will be black screen and it is easily confused broken state)

    note: that one over the top, but you get most information

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions