Amdgpu Radeon 9070 XT frequent pageflip_timeouts

[3-1415926535 (Necmi Inal)]

Hello!

i just got a new PC with a Radeon 9070 XT GPU and I have been having problems with the amdgpu module very frequently freezing the screen with pageflip timeout errors in the system journal.

It has gotten very annoying to the point that I think of installing Windows and ditching linux altogether. I have a LG 1440p 200Hz monitor and turning it on and off sometimes recovers the system, sometimes it doesn’t work or freezes again just after a minute. Turning adaptive refresh rate off has made the problem better but it still hasn’t solved it. The frame rate usually drops quite a bit a while before the freeze, noticeable by the choppy pointer movement.

I tried some grub command line settings but they didn’t solve the problem. The settings I have tried are turning the runpm off and setting the debug mask to 0x10 (Sorry for not pasting the setting verbatim, I am not currently using it and I don’t fully remember it), another combination of settings I have tried is split_lock_detect=off together with amdgpu.ppfeaturemask=0xfffd3fff but I haven’t got any luck with either of them.

Here are my specs:
Ryzen 5 7500F CPU
Asus Prime Radeon 9070 XT GPU
16GB of Kingston CL30 DDR5 RAM at 6000 MT/s
Asus B650M-R motherboard
Fedora 44 KDE with all of the latest updates as of today
Kernel version 7.2.4-200.fc44.x86_64
Linux-firmware build 20260910-1.fc44
LG G6 27610A-B 1440p 200Hz monitor over DP

I need your help, this problem is very annoying with how often it occurs, roughly every hour or so.

Edit: Split the post into multiple paragraphs for better readability.

Edit: A day has passed and I set the dcdebugmask flag to 0x12 this morning, along with setting runtime power management off. I have not had any freezes since then. The problem seems to be resolved

Edit: It happened again while browsing the web, although it lasted more than it usually would without these settings. Recovered after suspending with Ctrl+Alt+T and typing systemctl suspend then waking the system up again.

Edit (September 12th 2026): I tried out the testing build of the linux-firmware package from September 10th and it seems more stable. However, I get little artifacts of on my desktop wallpaper:

The artifacts only appear on the wallpaper and don’t appear on other windows.

Edit (September 13th 2026): Removing the kernel parameters amdgpu.runpm=0 and amdgpu.dcdebugmask=0x12 fixed the wallpaper artifacts but the freezes started happening again. I guess I can stand the acne on my wallpaper until the bug is finally fixed.

[denalb (Dennis)]

@3-1415926535

The artifacts only appear on the wallpaper and don’t appear on other windows.

Got those artefacts also. It starts with the release of kernel 7.2.4 and is still there with kernel 7.2.5. I have those artefacts also in Arch Linux (EndeavourOS). So it’s not an issue with Fedora only.

Switching to the next wallpaper in my active slideshow fixes the artifacts.

Here you can read more:

[regression][7.2rc1] drm/amdgpu: give ttm entities access to all the sdma...

Switching back to kernel 7.1.13 should fix that for you, too.

[3-1415926535 (Necmi Inal)]

The new linux-firmware 20260916 build seems to have brought back the issue. I am currently on kernel 7.2.7, but I think it doesn’t have an effect on the pageflip timeouts. A wallpaper with artifacts was far more acceptable than a system which freezes every hour. The worst thing is that I am out of kernels which had the 20260910 build of linux-firmware in their initrd. I am genuinely stumped and don’t know what to do. I feel like my computer is an expensive paperweight because of a nerdy OS I chose to use.

[3-1415926535 (Necmi Inal)]

I tried GNOME and the display output froze again just after 5 minutes. I think the most concrete solution to this problem is windows

[3-1415926535 (Necmi Inal)]

After scouring through forums, I tried out amdgpu.recovery=1 and amdgpu.aspm=0. They seemed to have prevented freezing, although I still get some session-wide stutters every so often, like how it happened before a full freeze. The linux-firmware 20260916 build says it has some amdgpu reverts which fixed other peoples problems about DCMUB in mobile onboard graphics, but it also probably reverted the fix in the 20260910 build that fixed my system. Atleast the system is usable now. I hope the bug will be properly patched soon

[3-1415926535 (Necmi Inal)]

I encountered the display freezes again while gaming. The interesting thing was that the audio also froze too, playing the same sound constantly. This is not what usually happens, if I am watching or listening to something it continues. The amdgpu related boot parameters I had are:
amdgpu.dcdebugmask=0x12 amdgpu.runpm=0 amdgpu.gpu_recovery=1 amdgpu.aspm=0
After the first freeze I just turned my monitor off and on, but after the second one I just magic REISUB’ed. Here is the output of journalctl -b -1:
Sep 28 19:35:06 fedora kwin_wayland[2355]: Pageflip timed out! This is a bug in the amdgpu kernel driver
Sep 28 19:35:06 fedora kwin_wayland[2355]: Please report this at Making sure you're not a bot!
Sep 28 19:35:06 fedora kwin_wayland[2355]: With the output of ‘sudo dmesg’ and ‘journalctl --user-unit plasma-kwin_wayland --boot 0’
Sep 28 19:35:07 fedora kwin_wayland[2355]: Pageflip timed out! This is a bug in the amdgpu kernel driver
Sep 28 19:35:07 fedora kwin_wayland[2355]: Please report this at Making sure you're not a bot!
Sep 28 19:35:07 fedora kwin_wayland[2355]: With the output of ‘sudo dmesg’ and ‘journalctl --user-unit plasma-kwin_wayland --boot 0’
Sep 28 19:35:08 fedora kwin_wayland[2355]: Pageflip timed out! This is a bug in the amdgpu kernel driver
Sep 28 19:35:08 fedora kwin_wayland[2355]: Please report this at Making sure you're not a bot!
Sep 28 19:35:08 fedora kwin_wayland[2355]: With the output of ‘sudo dmesg’ and ‘journalctl --user-unit plasma-kwin_wayland --boot 0’
and many more of the same error…

[blazeanator (blaze)]

this is really strange I to am running a 9070XT on the latest 7.2.7 Kernel and I havent encountered any issues what so ever im not sure what makes my system different to be honest but here is my fetch in case it helps. I did see another post earlier today in the forums where someone was saying that if you used a dedicated display port cable directly from the card to your display it would fix this issue as he noticed some bugs like this when using an hdmi cable because of something in the HDMI protocol so you might try that if your not using a dedicated DP cable directly from the card to the display currently.

Edit. This is the link to the post I referenced above..

Workaround for Linux 7.2 AMDGPU Boot/Plasma Freeze on RX 7800XT / RX 9070XT (DCN 4.0.1 / RDNA3 & RDNA4)

Hi everyone, If you are building or testing the Linux 7.2 kernel on Fedora 44 (or other distros) with high-end RDNA 3 (RX 7800 XT) or RDNA 4 (RX 9070 XT) graphics cards (two other, RX9060XT and RX9070 are working fine) and experiencing black screen when KDE Plasma / Wayland attempts to initialize (while TTY/Ctrl+Alt+F2 still works), I found the root cause and a flawless hardware workaround. The Symptom/Log: The kernel shows infinite timing controller loops right before the freeze: “amdgpu 00…

[yurislnx (Yuri)]

@3-1415926535

I tried GNOME and the display output froze again just after 5 minutes. I think the most concrete solution to this problem is windows

Sad to see but I don’t blame ya. Not sure what’s happening lately with amdgpu reliability. It used to be that it would get stable enough some time after launch when the driver got mature enough. But in recent times, there have been a ton of issues on kernels 7.1 and 7.2, and firmware issues. Makes me wonder if something changed with their development/testing processes.

[gungorr (Rauf Gungor)]

I have several computers with discrete AMD GPUs (AI Pro R9700, 9070XT, 9070, 9060XT, 7800XT, 7600). The problem I posted and cited above was only due to HDMI interface changes in the amdgpu code in the kernel and it was isolated to Hisense TVs used as monitors. There were two solutions

1. revert 7e5760f084d06ca57e46e4dbf7ef08f03f857be8 (and rebuild kernel)
2. or use a DP-HDMI adapter to bypass GPUs HDMI port.

But have not seen any of the corruption problems mentioned in this thread (using Fedora 45 and custom kernel 7.2.8). The corruption/freezing could be a totally different problem. Screen corruption starts at the VRAM, it should not happen due to HDMI code. So, we might be discussing different problems.

This news could be a solution for HDMI problems. It probably will not help with screen corruption and freeze issues discussed here.

[gungorr (Rauf Gungor)]

I think the problems are rather isolated in certain specific cases. I have just posted that I have no such issues with a variety of AMD GPUs on Fedora 45 and custom 7.2.x kernels. The separate HDMI issue (not related to the issues discussed here) I posted earlier was already resolved for my case by using DP-HDMI adapter. I also confirmed the software solution by reverting one commit in the kernel amdgpu code. I am not able to confirm any of the problems discussed here.

[3-1415926535 (Necmi Inal)]

Well I use DisplayPort on my LG G6 27610A-B monitor, so HDMI regressions are not the problem here. I first considered investing in an HDMI 2.1 cable as I heard it was more stable compared to DP, but as it turns out, my monitor doesn’t support HDMI 2.1 so I can’t get the full 200hz at 1440p.

[gungorr (Rauf Gungor)]

I only posted since I received a note from Fedora forum. In summary, I cannot confirm any similar problems with my computers running Fedora 45 and 7.2 kernel. Posting kernel logs might be helpful to identify the problems in your cases. You could save kernel logs with “sudo dmesg > dmesg.log” and post the log here. It can also be helpful to upload the log to Google Gemini and ask to identify the amdgpu error messages if there is any.