Bug 2438442 - Graphics broken unless Panel Self Refresh is disabled (with i915.enable_psr=0)
Summary: Graphics broken unless Panel Self Refresh is disabled (with i915.enable_psr=0)
Keywords:
Status: CLOSED ERRATA
Alias: None
Product: Fedora
Classification: Fedora
Component: kernel
Version: 44
Hardware: x86_64
OS: Linux
unspecified
high
Target Milestone: ---
Assignee: Justin M. Forbes
QA Contact: Fedora Extras Quality Assurance
URL:
Whiteboard: RejectedBlocker https://discussion.fe...
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2026-02-10 11:16 UTC by Tomas Popela
Modified: 2026-03-14 02:22 UTC (History)
24 users (show)

Fixed In Version: kernel-6.19.7-300.fc44 kernel-6.19.7-200.fc43 kernel-6.19.7-100.fc42
Clone Of:
Environment:
Last Closed: 2026-03-14 00:16:38 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
kernel log (111.03 KB, text/plain)
2026-02-10 11:28 UTC, Tomas Popela
no flags Details
flickering example video (10.72 MB, video/mp4)
2026-02-23 14:32 UTC, Kamil Páral
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Red Hat Bugzilla 2441941 0 unspecified NEW Graphics break when trying to type LUKS password 2026-03-16 14:19:11 UTC
freedesktop.org Gitlab drm/i915 kernel issues 15771 0 None opened Screen corruption and stuttering on after upgrading to Kernel 6.19 2026-02-27 08:43:49 UTC

Description Tomas Popela 2026-02-10 11:16:06 UTC
1. Please describe the problem:

When running 6.19.rc8 and also the 6.19 final on F44 the graphics breaks (running GNOME), when moving with mouse cursor and also when i.e. selecting the container in Ptyxis making the machine unusable (I can attach a video if needed). Otherwise if using keyboard only, then things sort of work. Doing i915.enable_psr=0 fixes the problem altogether. What fixes the issue is also using the 6.18 kernel from F43

2. What is the Version-Release number of the kernel:

kernel-6.19.0-0.rc8.54.fc44.x86_64
kernel-6.19.0-59.fc44


3. Did it work previously in Fedora? If so, what kernel version did the issue
   *first* appear?  Old kernels are available for download at
   https://koji.fedoraproject.org/koji/packageinfo?packageID=8 :

Between 6.18 and 6.19

4. Can you reproduce this issue? If so, please provide the steps to reproduce
   the issue below:

On my hardware which is a Lenovo P14s G5 laptop with Intel Ultra 7 165H (Meteor Lake) and 3K display it is enough to boot 6.19 kernel

5. Does this problem occur with the latest Rawhide kernel? To install the
   Rawhide kernel, run ``sudo dnf install fedora-repos-rawhide`` followed by
   ``sudo dnf update --enablerepo=rawhide kernel``:

Yes, it does.

6. Are you running any modules that not shipped with directly Fedora's kernel?:

No.

7. Please attach the kernel logs. You can get the complete kernel log
   for a boot with ``journalctl --no-hostname -k > dmesg.txt``. If the
   issue occurred on a previous boot, use the journalctl ``-b`` flag.

Reproducible: Always

Comment 1 Fedora Blocker Bugs Application 2026-02-10 11:26:37 UTC
Proposed as a Blocker for 44-beta by Fedora user tpopela using the blocker tracking app because:

 The laptop can't be normally used with the default configuration unless Intel's Panel Self Refresh is disabled.

Comment 2 Tomas Popela 2026-02-10 11:28:12 UTC
Created attachment 2128874 [details]
kernel log

Comment 3 plumlis 2026-02-10 13:58:10 UTC
Same here. On the latest Kernel 6.19, graphical issues such as screen tearing and stuttering appear immediately after booting into GNOME. Everything works fine on 6.18. More precisely, this problem started with Kernel 6.19 RC2 and is still present in the stable release.

After testing with i915.enable_psr=0, the issue indeed disappears.

ThinkPad P14s, Intel U285H.

Comment 4 Adam Williamson 2026-02-16 16:57:26 UTC
+5 in https://pagure.io/fedora-qa/blocker-review/issue/2028 , marking accepted.

Comment 5 Justin M. Forbes 2026-02-17 15:15:03 UTC
Has anyone given a rawhide kernel a test to see if the issue was fixed with the drm pull?

Comment 6 Tomas Popela 2026-02-18 07:19:38 UTC
I've tried kernel-6.20.0-0.rc0.260217g9702969978695.10.fc45 and it is still broken.

Comment 7 Kamil Páral 2026-02-23 12:16:44 UTC
Tomas, are you willing to bisect the kernel to figure out the commit that broke it?

Comment 8 Tomas Popela 2026-02-23 12:41:13 UTC
No, I don't have the capacity to do so.

Comment 9 Kamil Páral 2026-02-23 13:53:00 UTC
Tomas, uploading a short video would be nice, so that people have a good idea of how it looks like.

We've tested our QA laptops and unfortunately none of them seem to exhibit this bug.

Comment 10 Kamil Páral 2026-02-23 14:08:18 UTC
Similar but not the same (might be related, though) - bug 2441941

Comment 11 Tomas Popela 2026-02-23 14:26:25 UTC
> Tomas, uploading a short video would be nice, so that people have a good idea of how it looks like.

You internally have the access (at least Petr) to the recordings, feel free to upload it somewhere and share it here.

Comment 12 Kamil Páral 2026-02-23 14:32:34 UTC
Created attachment 2130700 [details]
flickering example video

Comment 13 Adam Williamson 2026-02-23 17:33:29 UTC
Hum, has anyone seen this on anything other than a P14s ? I kinda had the impression it was known to affect a wide range of systems, but if it's just one, it might not be a Beta blocker really...

Comment 14 plumlis 2026-02-24 03:41:24 UTC
Hi, I conducted a test on another Chromebook of mine, also running Fedora 43 with an Intel CPU and integrated GPU. I installed and tested Kernel 6.19.3 and did not encounter any PSR-related issues.

Kernel:https://koji.fedoraproject.org/koji/buildinfo?buildID=2944717
CPU:Intel Celeron 6305
GPU:Intel UHD G4(TGL GT2)


However, it's currently difficult to determine whether this issue is specific to the P14, or to certain ThinkPad models.

Have any users with Intel MTL or ARL performed tests? I personally suspect this might be specific to MTL/ARL GPUs.Intel has had some commits with bugs in both 6.18 and 6.19; 6.18 even caused some Sycl-related issues that were only fixed later.

If this is specific to certain ThinkPad models... I know there's someone named Mark Pearson who should be responsible for Lenovo product Linux support. I wonder if we could seek his help, since ThinkPads also come with Fedora pre-installed and will inevitably encounter Kernel 6.19-related issues.

From what I recall, I installed and tested Kernel 6.19 back when it was still at RC1, and I encountered this issue at that time. I suspect that something in the DRM/next or similar PRs after the 6.19 Merge Window opened may have caused this situation.

Comment 15 Petr Sklenar 2026-02-24 07:38:09 UTC
Hi @plumis,

>> Have any users with Intel MTL or ARL performed tests? 

I have no issues with
Version: ThinkPad P1 Gen 7
+
lspci -k | grep -A 3 -i "VGA\|3D"
00:02.0 VGA compatible controller: Intel Corporation Meteor Lake-P [Intel Arc Graphics] (rev 08)
	Subsystem: Lenovo Device 2235
	Kernel driver in use: i915
	Kernel modules: i915, xe

kernel-6.19.3-300.fc44.x86_64
mesa-dri-drivers-25.3.5-2.fc44.x86_64
---

It would also be great if anyone affected can try past kernel builds from https://koji.fedoraproject.org/koji/packageinfo?packageID=8 and identify the last working one and the first broken one.

Comment 16 plumlis 2026-02-24 08:45:20 UTC
Thank you very much for your testing. It seems the issue is becoming even more complex.
I have tested several devices I have on hand, all running a consistent environment: Fedora 43 + Intel GPU + Kernel 6.19.3. The results are as follows:

- Chromebook (Intel Celeron 6305): Normal
- Surface Go 2 (Intel Core m3-8100Y / 4454Y): Normal
- ThinkPad P14s Gen 6: Faulty (exhibits the same tearing/stuttering)

Combined with the original reporter Tomas’s ThinkPad P14s Gen 5, it appears that this issue specifically affects the base-model ThinkPad P14s series with Intel graphics. Since the Gen 6 and Gen 5 are minor iterations of each other, I suspect this might be related to the Display Panel.

For reference, my screen is a 3K (3072 x 1920) IPS panel. I am not sure if Tomas has the exact same panel configuration.

I also attempted the following troubleshooting steps, but none of them resolved the issue:

- Switching from the i915 driver to the experimental Xe driver.
- Toggling Mutter’s VRR (Variable Refresh Rate) settings (both enabling and disabling).
- Testing the newly released Kernel 7.0 RC1.

I tried to extract useful information from the logs, but unfortunately, they appear silent; there is no mention of PSR or any related errors in the kernel logs (dmesg).

I checked Intel PSR commits log and find nothing useful:
https://github.com/torvalds/linux/commits/master/drivers/gpu/drm/i915/display/intel_psr.c

If this is indeed panel-specific, it is difficult to determine if this is exclusive to the P14s. Lenovo uses this same panel model in many other devices, such as the ThinkBook 14+ and likely in future generations of the P14s as well.

Comment 17 plumlis 2026-02-24 08:50:10 UTC
I accidentally tried connecting an external monitor to test, and I made some new discoveries.

When I adjust the refresh rate to 60Hz, everything works normally. When I set it back to 120Hz, screen tearing and stuttering occur. Both my P14s and Tomas's P14s display panels should support 120Hz refresh rate, so we have the same problem.

Could the reason the P1 Gen 7 has no issues be because its screen refresh rate is 60Hz?

So, is one of the reasons for reproducing this issue the high refresh rate?

After kernel 6.19, will Intel core laptops using high refresh rate screens encounter screen tearing issues, and can disabling PSR solve it?

Comment 18 Tomas Popela 2026-02-24 08:51:55 UTC
> For reference, my screen is a 3K (3072 x 1920) IPS panel. I am not sure if Tomas has the exact same panel configuration.

Yes, I do.

Comment 20 Justin M. Forbes 2026-03-03 19:20:40 UTC
Want to give https://koji.fedoraproject.org/koji/taskinfo?taskID=142945237 a try? It is a scratch build, so not secure boot signed unfortunately.

Comment 21 plumlis 2026-03-04 02:57:03 UTC
(In reply to Justin M. Forbes from comment #20)
> Want to give https://koji.fedoraproject.org/koji/taskinfo?taskID=142945237 a
> try? It is a scratch build, so not secure boot signed unfortunately.

Just tried.
Still have issue.

Comment 22 Adam Williamson 2026-03-04 16:59:36 UTC
I'm calling for a re-vote on blocker status here, since despite sending out a call for testing we've yet to receive any report that this affects any systems other than P14s with 120Hz display, and there is a clear workaround for it. Based on current information I think it'd be OK to release Beta with this as a documented known issue.

Comment 23 Justin M. Forbes 2026-03-04 22:34:52 UTC
I am all for this being a final blocker, but it seems a bit much making it a beta blocker as a bug impacting limited hardware with a known workaround in my opinion. As there is no upstream fix just yet, the timeframe for a fix in Fedora is still unclear.

Comment 24 Kamil Páral 2026-03-06 13:20:04 UTC
During Beta Go/NoGo meeting [1]:

!agreed 2438442 - RejectedBlocker (Beta) - this is rejected as we have no clear indication it affects a very wide range of hardware. It's only positively confirmed to affect certain variants of a single laptop model (Thinkpad P14s with 120Hz screen). Per https://fedoraproject.org/wiki/Blocker_Bug_FAQ#What_about_hardware_and_local_configuration_dependent_issues? our decision is that's not broad enough to merit the bug being a Beta blocker

[1] https://meetbot-raw.fedoraproject.org/meeting_matrix_fedoraproject-org/2026-03-05/

Transferring the proposal to a FinalBlocker instead.

Comment 25 plumlis 2026-03-09 13:00:30 UTC
https://patchwork.freedesktop.org/series/162864/
here is a patch from intel for testing about fix this issue.

Comment 26 Kamil Páral 2026-03-09 17:05:45 UTC
Discussed in a blocker review meeting [1] on 2026-03-09. The outcome was:

!agreed 2438442 - RejectedBlocker (Final) - this is rejected per https://fedoraproject.org/wiki/Blocker_Bug_FAQ#What_about_hardware_and_local_configuration_dependent_issues? , because so far we still have no indication this affects more than a single laptop model, and blocking the release for a workaroundable issue on a single laptop model isn't really in line with past reality. Note "Bugs that affect only a single adapter or small group of closely-related adapters are almost never taken as blockers" in the FAQ.

[1] https://meetbot-raw.fedoraproject.org//blocker-review_matrix_fedoraproject-org/

Comment 27 Justin M. Forbes 2026-03-09 21:13:36 UTC
https://koji.fedoraproject.org/koji/taskinfo?taskID=143171521 is a build you can test that patch with

Comment 28 plumlis 2026-03-10 01:20:05 UTC
(In reply to Justin M. Forbes from comment #27)
> https://koji.fedoraproject.org/koji/taskinfo?taskID=143171521 is a build you
> can test that patch with

Thanks for your building.
It seems still havent fix this issue yet.
It seems we need to continue communicating with the Intel upstream developers, report the logs, and seek their assistance to resolve this.

Comment 29 plumlis 2026-03-10 01:20:26 UTC
In fact, another laptop user with a Lenovo Yoga 7 Pro (Intel) has encountered the same issue, but from openSUSE:
https://bugzilla.opensuse.org/show_bug.cgi?id=1259036
It's possible they are using the same LCD panel, a 14.5" 3K (3072x1920) IPS, or possibly the OLED version, but both are 120Hz with VRR, and also on MTL and newer platforms.

Comment 30 Adam Williamson 2026-03-10 04:55:04 UTC
If anyone has a SUSE bugzilla account, it'd be great if they could comment there and link to this bug and https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15771 .

I have an account for progress.opensuse.org , but apparently it's not the same auth system...

Comment 31 plumlis 2026-03-10 07:47:01 UTC
Currently, the OpenSUSE user who reported the issue has already gotten involved on the i915 side:

https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/15771

Based on the current situation, all affected systems share the same 3K IPS display panel and/or high-refresh-rate screens. Intel appears to have narrowed down the root cause to a potential incompatibility between Panel Self-Refresh (PSR) and Variable Refresh Rate (VRR).

It's unclear whether this issue is related to the display panel manufacturer or to a conflict between VRR and PSR. Given that GNOME 50 enables VRR support by default, it's possible that other users may be affected as well.

Affected users report that the issue disappears after disabling and then re-enabling PSR2:


$ echo 1 > /sys/kernel/debug/dri/1/i915_edp_psr_debug  # disable PSR
$ echo 0 > /sys/kernel/debug/dri/1/i915_edp_psr_debug  # re-enable PSR


Intel has now provided a new patch for testing (the previous one didn't work). 

https://patchwork.freedesktop.org/series/162920/

To be honest, I'm not a kernel developer and I'm not very experienced with applying patches and compiling the kernel—I tried several times yesterday but wasn't successful.

Comment 32 Kamil Páral 2026-03-10 09:04:57 UTC
(In reply to plumlis from comment #31)
> Given that GNOME 50 enables VRR
> support by default, it's possible that other users may be affected as well.

I don't think that's correct. With Fedora 44 Beta and a VRR display, the VRR option in available by default, but *disabled* by default. Also, the display runs at 60 Hz by default (while my display can do 144 Hz). I need to switch both manually. Tested with a Live image.

Also, interestingly, in comment 16 you said that disabling VRR doesn't help (only lowering refresh rate does). So I wonder whether this can really be attributed to VRR and how. But anyway, I just wanted to point out that in a default state, Fedora seems be running with conservative values (60 Hz, no VRR) and changes must be done manually. Unless this is hardware-specific.

Comment 33 Tomas Popela 2026-03-10 10:27:30 UTC
(In reply to plumlis from comment #31)
> Intel has now provided a new patch for testing (the previous one didn't
> work). 
> 
> https://patchwork.freedesktop.org/series/162920/
> 
> To be honest, I'm not a kernel developer and I'm not very experienced with
> applying patches and compiling the kernel—I tried several times yesterday
> but wasn't successful.

I've started the scratch build with that patch in https://koji.fedoraproject.org/koji/taskinfo?taskID=143202465 (it is still building)

Comment 34 Adam Williamson 2026-03-10 15:44:03 UTC
Kamil: I wonder if the reason we're not seeing this more is that on *most* systems we default to 60Hz even if the panel is 120Hz-capable, but on the affected systems we default to 120Hz? It's just a theory, but it might explain things.

Comment 35 Michel Dänzer 2026-03-10 18:02:59 UTC
(In reply to Kamil Páral from comment #32)
> With Fedora 44 Beta and a VRR display, the VRR option in available by default, but *disabled* by default.

That's right.

> Unless this is hardware-specific.

It's not. mutter prefers modes which don't allow VRR over the corresponding modes which allow VRR, so the default configuration never allows VRR.

Moreover, in the attached video, VRR can't really be active (mutter doesn't enable the VRR_ENABLED CRTC property), even if VRR was enabled in mutter's display configuration, because there's no single surface which covers the display (since the top bar is visible).

> So I wonder whether this can really be attributed to VRR and how.

It seems unlikely, though it might still be possible in theory, if the kernel driver enables VRR signalling even while the VRR_ENABLED property is disabled (e.g. to avoid glitches when toggling the property, though I thought this could only be necessary in some cases with HDMI display connections, i.e. not with internal panels). Note that in this case mutter couldn't do anything to avoid the issue, per above VRR is effectively disabled in the KMS API in the attached video anyway.

However, I don't see any mention of VRR in the referenced gitlab issue or patch series, so it's not clear where the "Intel appears to have narrowed down the root cause to a potential incompatibility between Panel Self-Refresh (PSR) and Variable Refresh Rate (VRR)" statement is coming from.

Comment 36 plumlis 2026-03-10 22:02:22 UTC
Hi, English is not my native language, but I will do my best to clearly describe everything I currently know about this issue.

> Is this issue related to VRR and Mutter?

Initially, Intel engineers did speculate about this possibility, but it now seems to have been ruled out. It is also unrelated to Mutter. It appears to be purely a kernel issue where PSR2 fails to be correctly enabled or disabled under certain specific conditions.

> What is the current status/nature of this problem?

Current testing indicates that on some 120Hz panels (possibly those with hardware VRR support, though hardware VRR might not even be required), when the refresh rate is set to 120Hz, the Intel driver fails to correctly apply PSR2. This results in screen corruption and stuttering. However, if you manually disable and then re-enable PSR2, the issue disappears.

> Why is this issue difficult to detect?

If I recall correctly, Fedora (and most GNOME-based distributions) does not enable the 120Hz refresh rate by default after a fresh installation. Therefore, for new installations, even if the hardware has this issue related, users won't encounter the problem until they manually enable 120Hz.

In the case of myself and other users who encountered this, we had already manually enabled 120Hz. Everything worked perfectly on kernel 6.18, and the issue only appeared after we upgraded to kernel 6.19.

Furthermore, this problem might only occur on 120Hz panels that support VRR in hardware, causing an error during the kernel driver loading phase. However, due to a limited number of test machines, we cannot yet determine exactly which users and hardware configurations are affected.

Comment 37 Kamil Páral 2026-03-11 11:34:14 UTC
For the record, a common issue description is here:
https://discussion.fedoraproject.org/t/common-issue/182839

Comment 38 Tomas Popela 2026-03-12 08:51:51 UTC
This issue is fixed with https://patchwork.freedesktop.org/series/163098/ (verified through https://koji.fedoraproject.org/koji/taskinfo?taskID=143249180 scratch build).

Comment 39 Fedora Update System 2026-03-12 21:11:25 UTC
FEDORA-2026-b9bd91408a (kernel-6.19.7-300.fc44) has been submitted as an update to Fedora 44.
https://bodhi.fedoraproject.org/updates/FEDORA-2026-b9bd91408a

Comment 40 Fedora Update System 2026-03-12 21:11:29 UTC
FEDORA-2026-21a28a358f (kernel-6.19.7-100.fc42) has been submitted as an update to Fedora 42.
https://bodhi.fedoraproject.org/updates/FEDORA-2026-21a28a358f

Comment 41 Fedora Update System 2026-03-12 21:11:31 UTC
FEDORA-2026-2f3dbc95bc (kernel-6.19.7-200.fc43) has been submitted as an update to Fedora 43.
https://bodhi.fedoraproject.org/updates/FEDORA-2026-2f3dbc95bc

Comment 42 Fedora Update System 2026-03-13 01:38:43 UTC
FEDORA-2026-b9bd91408a has been pushed to the Fedora 44 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-b9bd91408a`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-b9bd91408a

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 43 Fedora Update System 2026-03-13 02:03:24 UTC
FEDORA-2026-2f3dbc95bc has been pushed to the Fedora 43 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-2f3dbc95bc`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-2f3dbc95bc

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 44 Fedora Update System 2026-03-13 02:25:10 UTC
FEDORA-2026-21a28a358f has been pushed to the Fedora 42 testing repository.
Soon you'll be able to install the update with the following command:
`sudo dnf upgrade --enablerepo=updates-testing --refresh --advisory=FEDORA-2026-21a28a358f`
You can provide feedback for this update here: https://bodhi.fedoraproject.org/updates/FEDORA-2026-21a28a358f

See also https://fedoraproject.org/wiki/QA:Updates_Testing for more information on how to test updates.

Comment 45 Kamil Páral 2026-03-13 08:59:29 UTC
Verified by plumlis here:
https://bodhi.fedoraproject.org/updates/FEDORA-2026-2f3dbc95bc#comment-4579734

Comment 46 Fedora Update System 2026-03-14 00:16:38 UTC
FEDORA-2026-b9bd91408a (kernel-6.19.7-300.fc44) has been pushed to the Fedora 44 stable repository.
If problem still persists, please make note of it in this bug report.

Comment 47 Fedora Update System 2026-03-14 02:19:39 UTC
FEDORA-2026-2f3dbc95bc (kernel-6.19.7-200.fc43) has been pushed to the Fedora 43 stable repository.
If problem still persists, please make note of it in this bug report.

Comment 48 Fedora Update System 2026-03-14 02:22:54 UTC
FEDORA-2026-21a28a358f (kernel-6.19.7-100.fc42) has been pushed to the Fedora 42 stable repository.
If problem still persists, please make note of it in this bug report.


Note You need to log in before you can comment on or make changes to this bug.