Stop a laptop panel flickering or stuttering from Panel Self Refresh
The Intel i915 row is incomplete. modinfo -p i915 on 7.2.5-3-omarchy lists enable_panel_replay, and in mainline intel_psr.c the enable_psr parameter gates only _psr_compute_config, while panel_replay_global_enabled checks only the debug flag and params.enable_panel_replay. So i915.enable_psr=0 leaves Panel Replay on, the same gap the record already warns about for xe. Meteor Lake class panels on i915 can use Panel Replay. The amdgpu row also needs a note: amd_shared.h defines DC_DISABLE_PSR 0x10 as PSR v1 and PSR-SU only, with Panel Replay on a separate bit, DC_DISABLE_REPLAY 0x400. 0x10 is what the T14 Gen5 thread confirmed and stays the first choice. Everything else held: issue #6853 has the XPS i915_psr_status output, the runtime echo 1 test, the both-knobs reasoning and the enable_psr=2 comment, #11016 confirms debug 1 disables both PSR and Panel Replay, the ASUS B9406 script exists and says xe.enable_psr=0 does not cover Panel Replay, and the BBS thread confirms 0x10 for the T14 Gen5 AMD OLED. Cause and verify are rewritten to match. The debugfs write needs root and was not run.
Turning off PSR or Panel Replay raises idle power draw. The XPS report estimated 0.5 to 1.5 W, which is a visible hit to battery life. On Omarchy 4, a drop-in written with = instead of +=, or with tee instead of tee -a, silently drops Omarchy's own kernel parameters, including initramfs_async=0, which keeps the Plymouth LUKS prompt working. limine-mkinitcpio rewrites the UKI on the ESP, so check df -h /boot first. To recover from a bad parameter, edit it out in the Limine boot menu editor or boot a Limine snapshot entry.
Symptom
The laptop's own screen flickers, or short horizontal strips of artefacts flash across part of the screen when scrolling or playing video. External monitors on the same machine are fine. Variants that users report:
- Intel: flicker, sometimes with
[i915] *ERROR* CPU pipe A FIFO underruninjournalctl -k. - Newer Intel on the
xedriver (Dell XPS 13 DX13260, Wildcat Lake): scrolling judders far below the panel's 120 Hz and the cursor moves in jerks.i915_psr_statusshowsPSR mode: Panel Replay Selective Update enabled (Early Transport). - AMD: ThinkPad T14 Gen5 AMD OLED with intermittent horizontal artefacts during video and scrolling.
If the whole image freezes instead, with flip_done timed out in the log, see the record amdgpu-flip-done-timed-out.
Cause
Panel Self Refresh (PSR) and its successor Panel Replay are eDP power-saving features. During static periods the GPU stops sending frames and the panel refreshes from its own memory. Bugs in the driver or in a particular panel's wake and exit path cause flicker, artefacts, or late frame delivery when the panel has to wake for an update. That is why cursor movement and scrolling show it most. The Arch wiki documents PSR flicker on Intel and recommends turning it off. On Intel, i915 and xe share one display code path in which enable_psr gates PSR only and enable_panel_replay gates Panel Replay separately, so turning off PSR alone leaves Panel Replay running. Omarchy's own installer turns off Panel Replay for the ASUS ExpertBook B9406 for that reason. On amdgpu, dcdebugmask=0x10 turns off PSR v1 and PSR Selective Update, and Panel Replay has its own bit, 0x400. All six Intel and AMD parameters used below were confirmed with modinfo -p on kernel 7.2.5-3-omarchy.
Fix
**Find your display driver:**
lspci -k | grep -A3 -E 'VGA|Display'
**Intel only: test at runtime first** (the path is from the XPS report and may differ by PCI address):
sudo cat /sys/kernel/debug/dri/0000:00:02.0/eDP-1/i915_psr_status
echo 1 | sudo tee /sys/kernel/debug/dri/0000:00:02.0/i915_edp_psr_debug
Writing 1 turns off both PSR and Panel Replay. If that makes the flicker or stutter stop at once, PSR or Panel Replay is the cause. It resets at reboot.
**Pick the parameters for the driver lspci -k showed as in use:**
| Driver in use | Parameters | | --- | --- | | i915 | i915.enable_psr=0 i915.enable_panel_replay=0 | | xe | xe.enable_psr=0 xe.enable_panel_replay=0 | | amdgpu | amdgpu.dcdebugmask=0x10 |
On Intel you need both parameters. enable_psr=0 does not cover Panel Replay, and turning off only Panel Replay falls back to PSR2 selective fetch, which stuttered the same way in the XPS report. One reporter on the same laptop found that xe.enable_psr=2 xe.enable_panel_replay=0 also cured it, at a smaller battery cost, so try that first if battery matters.
On amdgpu, 0x10 turns off PSR v1 and PSR Selective Update, which is what fixed the ThinkPad T14 Gen5 report. It does not turn off Panel Replay. If the artefacts persist, use amdgpu.dcdebugmask=0x410 instead, which adds the kernel's DC_DISABLE_REPLAY bit (0x400).
**Omarchy 4:** write only the row for your driver. The block below is the xe row. For i915 the line is KERNEL_CMDLINE[default]+=" i915.enable_psr=0 i915.enable_panel_replay=0", and for amdgpu it is KERNEL_CMDLINE[default]+=" amdgpu.dcdebugmask=0x10". Always use +=, because a bare = would wipe Omarchy's own defaults.
sudo tee -a /etc/limine-entry-tool.d/zz-local.conf >/dev/null <<'EOF'
KERNEL_CMDLINE[default]+=" xe.enable_psr=0 xe.enable_panel_replay=0"
EOF
sudo limine-mkinitcpio && sudo limine-update && sudo reboot
After the reboot, check that the parameters reached the kernel:
cat /proc/cmdline
**Plain Arch:** add the parameters for your driver to GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub and run sudo grub-mkconfig -o /boot/grub/grub.cfg, or append them to options in /boot/loader/entries/*.conf for systemd-boot. Then reboot.
Verify
After the reboot, cat /proc/cmdline shows the parameters. On Intel, cat /sys/module/i915/parameters/enable_psr and cat /sys/module/i915/parameters/enable_panel_replay (or the same files under /sys/module/xe/parameters/) both print 0, and i915_psr_status no longer reports PSR or Panel Replay as enabled. On AMD, cat /sys/module/amdgpu/parameters/dcdebugmask prints 16 for 0x10, or 1040 for 0x410. Scroll a long page and play a video to check that the flicker and judder are gone.
Sources
- https://wiki.archlinux.org/title/Intel_graphics
- https://wiki.archlinux.org/title/AMDGPU
- https://bbs.archlinux.org/viewtopic.php?id=300966
- https://github.com/omacom/omarchy/issues/6853
- https://github.com/omacom/omarchy/blob/quattro/install/hardware/asus/fix-asus-ptl-b9406-display.sh
- https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/i915/display/intel_display_types.h
- https://github.com/omacom/omarchy/issues/11016
- https://github.com/omacom/omarchy/issues/6768
- https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/i915/display/intel_display_params.c
- https://github.com/omacom/omarchy/issues/5423
- https://github.com/omacom/omarchy/issues/12875
- https://github.com/omacom/omarchy/issues/11176
- https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/i915/display/intel_psr.c
- https://github.com/torvalds/linux/blob/master/drivers/gpu/drm/amd/include/amd_shared.h