Framework 13 (Intel Ultra 125H) - Random Display Freezes with xe Driver on Fedora 43

Community Support
,
1

Hardware:

  • Framework Laptop 13 with Intel Core Ultra 125H

  • Fedora 43

  • Kernel: 6.11.7-300.fc43.x86_64

  • Desktop: GNOME on Wayland

  • Display: Laptop screen only (no external monitors)

Problem Description:

I’m experiencing random display freezes where the graphics stop responding, but the system continues running in the background. Audio continues playing and I can move the mouse cursor, but the screen content doesn’t update and the fans suddenly begin to run on full speed. In patient moments I have waited up to 15 minutes - nothing happens. This has been happening since I got the laptop.

Frequency: The issue is inconsistent - sometimes it doesn’t happen for weeks, other times it occurs daily or even several times a day. I haven’t been able to identify a specific trigger.

Symptom Details:

  • Display/graphics freeze completely

  • Mouse cursor still moves

  • Audio continues playing normally

  • System appears to be running in background

  • Cannot switch to TTY (Ctrl+Alt+F3 doesn’t work)

  • Requires hard reboot (holding power button) to recover

Error Logs: Looking at journalctl, I see repeated xe driver errors around the freeze times:

xe 0000:00:02.0: [drm] *ERROR* GT0: TLB invalidation fence timeout, seqno=10975 recv=10974
xe 0000:00:02.0: [drm] GT0: Schedule disable failed to respond, guc_id=6
xe 0000:00:02.0: [drm] Xe device coredump has been created
xe 0000:00:02.0: [drm] GT0: trying reset from guc_exec_queue_timedout_job [xe]
xe 0000:00:02.0: [drm] GT0: reset started
xe 0000:00:02.0: [drm] GT0: reset done
xe 0000:00:02.0: [drm] GT0: VM job timed out on non-killed execqueue

Does anyone have a suggestion or have similiar issues?

2

FWIW, i’ve been running ubuntu 24.04 lts, 6.14 kernel, kde Plasma 5, X11 with out any issue (display or otherwise) for several months on the Intel 125H.

Other forum users have reported display freezing issues (mostly with AMD based systems) but the cause is not clear to me. If you can, try switching to X11 (I believe fedora still supports X11 so trying it shouldn’t be too onerous). EDIT, it looks like fedora 43 dropped X11 support with GNOME. Wayland is my best guess as to the cause of this issue. If you can, try fedora 42 with X11 or switch to a distro that supports X11.

But before doing that you could also try a different kernel version with wayland.

If you can/do switch to X11, I’d definitely try a newer kernel with X11 (6.14+).

I’ve been wanting to switch to fedora/wayland for a while now but have put it off due to reading other’s having issues with wayland (across multiple distros and hardware types).

HTH

3

Hmm, I give that a thought and maybe a try.

Which intel driver are you using? Also intel xe?

4

I believe the Xe in Xe device coredump refers to Intel’s Iris Xe gpu that is a component of the 125H cpu.

driver details:

lspci -k | grep -EA3 'VGA|3D|Display'
00:02.0 VGA compatible controller: Intel Corporation Meteor Lake-P [Intel Arc Graphics] (rev 08)
        Subsystem: Framework Computer Inc. Meteor Lake-P [Intel Arc Graphics]
        Kernel driver in use: i915
        Kernel modules: i915, xe

modinfo xe | grep vermagic
vermagic:       6.14.0-35-generic SMP preempt mod_unload modversions
5 
lspci -k | grep -EA3 'VGA|3D|Display'
00:02.0 VGA compatible controller: Intel Corporation Meteor Lake-P [Intel Arc Graphics] (rev 08)
	Subsystem: Framework Computer Inc. Device 0009
	Kernel driver in use: xe
	Kernel modules: i915, xe

modinfo xe | grep vermagic
vermagic:       6.17.7-300.fc43.x86_64 SMP preempt mod_unload

But yeah, I think I had the problem also with i915 driver and then tried the Xe driver.
Anyway, I am not too deep into Linux & Hardware so it’s hard for me to get a feeling where the problem lies without just (slightly) blind trying

6

typo in your original post or did you upgrade your kernel?

I suspect your issue is not specific to framework. Perhaps try a fedora forum as they might be able to better help you trouble shoot.

Good luck.

7

For reference I found this open issue Making sure you're not a bot!. Looks very similar.

Will look more into it and check out other forum and may update it here once I find a solution (and have not forgotten about the post here).

1 Like
8

The issue persists also on Ubuntu 24.04.03 LTS with GNOME on Wayland and kernel 6.14.0-35-generic. Trying GNOME on X11 now (I intended to do it already, but did not check which window system was used as default).

Also, I am already in contact with the Framework support.

At least we identified that my RAM modules (Crucial CT2K16G56C46S5 32GB Kit 16GBx2 DDR5-5600 SODIMM) have not been officially validated by frame work. But I am not sure if this could be cause of the issue.

9

I am using the exact same ram kit, so I don’t think the ram vendor/type is the issue. You could try a memory test to see if your memory checks out.

I am wondering if the amount of ram in use might be a contributing factor for these issues. I typically use 8 Gb of 32 Gb ram (unless I’m running windows 11 in a vm which is infrequent). It’s always possible I just have not run into this issue due to utilizing low ram.

Do you have a feel if this issue occurs more frequently for you when your ram usage is high?

10

I don’t really see a pattern when it occurs.
90% of my use case is having tabs in firefox open, maybe having 1-2 videos running (one paused). So thats also when the most hangs happen. But also occured when playing 4k mp4 file. Or with ubuntu now, starting a download in firefox and opening the downloads window (nothing else heavy really running). But it is not reproducible. Doing the same things again and it works fine, probably for weeks.

Support already requested doing memtest, which I did and it passed - also sent them the logs from the test.

11

FWIW, even tho I’m using the i915 driver, the xe kernel module is in use, (see below, xe is used for drm? e.g. video playback). I watch plenty of videos on firefox. This is part of the reason I suspect something related to wayland as contributing to the cause.

In any event, other users have reported similar issues (possible with X11 - its not clear to me) and then stablity on fedora 42 with wayland. I hope using X11 works for you until this issue get fixed.

lsmod | grep -E 'xe|i915'
xe                   3436544  0
drm_gpuvm              45056  1 xe
gpu_sched              61440  1 xe
drm_ttm_helper         16384  1 xe
drm_exec               12288  2 drm_gpuvm,xe
drm_suballoc_helper    20480  1 xe
i915                 4702208  160
drm_buddy              24576  2 xe,i915
ttm                   118784  3 drm_ttm_helper,xe,i915
drm_display_helper    278528  2 xe,i915
cec                    94208  3 drm_display_helper,xe,i915
i2c_algo_bit           16384  2 xe,i915
intel_vsec             20480  2 intel_pmc_core,xe
video                  77824  2 xe,i915

EDIT: I configure a custom 3:2 display resolution for X11 to make reading the screen easier on old eyes:

cat /etc/X11/xorg.conf.d/00-monitor.conf 
Section "Monitor"
    Identifier "eDP-1"
    Modeline   "1664x1110_60.00"  154.43  1664 1776 1952 2240  1110 1111 1114 1149  -HSync +Vsync
EndSection

I doubt this config is the reason for my not experiencing issues on X11.

12

I’m having this issue too, and it seems like several other framework users seem to be having it here Making sure you're not a bot!

13

Update: Support suggested removing expansion slots and I have removed all but one USB-C (for charging and docking station purporses). I am back on Fedora 43 with Wayland and had no freezes for two weeks now. This is not a final call, since it could just have been lucky. But maybe that already helps others also testing and identifying this issue with this framework model.

Removing of expansion slots won’t be a long term solution. Let’s see what the support suggests now.

1 Like
14

Still got a freeze. Support did not respond now. They suggested trying booting with “nomodeset” but working with this unperformant graphics over several weeks is not a solution for me.

If no new ideas from the support team come I probably try Fedora 42 with X11 for longterm testing. In the few days with Ubuntu on Gnome X11 I had at least no freeze. X11 might also not be long term solution, but the most sustainable solution for now (if it really works there).

1 Like
15

Sorry to hear your still having trouble with this.

If your willing, would you mind sharing which expansion cards you removed, what dock you use and if this issue has happened when the dock has not been used at all since booting?

The dock is something I’d look into before switching to a different OS/X11.

No worries if you don’t - I would not like dealing with this kind of issue (hence the reason I’m sticking with my current state - still no issues like this for me).

My expansion card load out for 99% of my use is 2 usb c, 1 usb a, and the micro sd card reader. I have hdmi and Ethernet expansion cards but rarely use them. On rare occasions, I’ll use an (old) Apple USB-C Digital AV Multiport Adapter to connect an external monitor, power the framework, and provide an extra usb a for a keyboard or mouse. I also don’t use the apple adapter for extended periods.

16

I removed the HDMI, USB-A and 2nd USB-C expansion slot. The issue has happened connecting to dock, on battery or also on AC - but mostly I am working on battery mode.

1 Like
17

Well that changed a little while ago for me.

TLDR: some similar symptoms you describe in your opening post but nothing that indicates that it is the same issue you experience. I thought I’d mention it in the event it helps you or others.

What happened:

  1. I noticed the network cut out first - I’ve seen similar symptoms (not on framework) with a kernel that is having issues. I’ve read networking can be sensitive to kernel issues and usually is one of the first functions to stop working.
  2. Firefox stop responding next. I tried closing it and restating. It would not restart. I was not watching a video at the time. I had watched maybe two youtube videos earlier in the day.
  3. I had the ability to manipulate the display. For example, I switched to a new desktop (using kde plasma); however, right after that the plasma desktop stopped responding and the fan speed increased. The keyboard and trackpad (aka mouse cursor) did still respond.
  4. I was able to get into a console (and back into X11/plasma with a working mouse cursor) with the Ctl-Alt-Fx keys. In the console, I saw some output being generated that indicated the system was unable to write to the disk. I could not get a login prompt. Unfortunately I did not record the console output - I’ll do so if this happens again. Failure to write to disk is also something I’ve read is common when a kernel starts having issues.
  5. After waiting a few minutes, I hard reset the device. After rebooting, I examined every file in /var/log that had an entry around the date/time of this event. I see no (relevant) errors leading up to this. Based on the time stamp between the last log entry and the start of the reboot, it does look like the kernel lost ability to write to disk - none of the console output I saw was recorded. I also looked for kernel/debugging output on other common locations (e.g. /var/lib/systemd/pstore/), but as yet I’ve found nothing.

Beyond some similar symptoms you describe, there is nothing above that says the event I observed is or is not related to what you have observed. There is also nothing that says this was a kernel issue, that is just my speculation.

FWIW In my 30 or so years using linux (mostly on “mature” hardware running “stable” software like the ubuntu lts release I currently use), I’ve probably seen such events less than 10 times. This is why I pay a lot of attention to it when it happens. That said, I don’t consider framework devices as “mature” nor do I think Fedora is “stable” so I’m not too surprised when there are issues with either.

HTH.

EDIT, I just ran a memtest86+ (v7.0.0) memory test with default settings started from the grub console. I only ran 1 complete pass which passed with no errors.

18

While the sequence I describe above is all well and good, for a kernel crash I would have expected events like a loss of network, inability to read & write to storage, loss of ability to output to a display to happen over a fraction of a second - not the many minutes I experienced.

In addition, when trying to figure out why my network went down (item 1 above) I attempted to ping from an already open terminal. The ping command did not even load, i.e. I now think I had already lost ability to read/write to disk at this point.

Short of it is I think I had a hardware issue related to my Samsung 990 pro 1 TB. These devices have firmware that can be updated and I was running on the firmware that came out the box (version: 5B2QJXD7). I’ve just updated to the latest firmware (Version 8B2QJXD7). According to the samsung www site, the latest and prior firmware updates (which I also did not have) fix things like:
*(8B2QJXD7) To improve read-operation stability. (Release : December 2025)
*(7B2QJXD7) To address the intermittent non-recognition and blue screen issue. (Release: September 2025).

I’ll see how this goes. Apologies for the noise as the issue I experienced likely is not related to your issue.

19

After some back and forth with the support, they will send me a replacement motherboard. It seems the error “aligns with a graphic error om the board itself”.

Just for reference, I ran Fedora 43 with Wayland GNOME now also for two weeks without a hang (but yeah sooner or later on should come).

I also asked them about the open gpu hang issue from i915 driver several times. I did not really get an answer, just when I believed a hardware person is gonna check this, I got the RMA info

1 Like
20

Hi @steve_oc does the new hardware work better?

I’m curious why you are using the new xe driver on this CPU?
It can run on Ultra Series 1 but i915 is the Intel supported driver on that CPU series.
xe driver was developed for upcoming CPU generations.

What problems were you having with i915?