Sleep is broken (amd gpu)

Hello nice people from the forum, Some days ago I updated my cachy os and now when it sleeps, it does not power off as normal and when try to revive the computer the screen goes black. Previously, my pc would shut down shoving whats in the memory into swap (from what I know this is hibernate, or suspend to disk). But now It simply does not shut down, the screen gets turned off and the only way to use the computer again is by rebooting it (by pressing the reset button). I notice it tries to shut down usb, as my mouse led is disabled, and when I click it returns. Here are my settings (from fastfetch): OS: CachyOS x86_64 Kernel: Linux 6.17.1-2-cachyos Display (VA2715-QHD): 2560x1440 in 27", 75 Hz [External] DE: KDE Plasma 6.4.5 WM: KWin (Wayland) WM Theme: Breeze Theme: Breeze (EmeraldDark) [Qt], Breeze-Dark [GTK2], Breeze [GTK3] CPU: AMD Ryzen 5 5500 (12) @ 4.27 GHz GPU: AMD Radeon RX 6700 XT [Discrete] Memory: 4.76 GiB / 31.23 GiB (15%) Swap: 0 B / 31.23 GiB (0%) This is not a back-breaking bug and I want to use this as an opportunity to learn about my system. I am still a bit of a noob, so sorry if I get some terms incorrect. But I would like to know:

  • How to make the log of the system entering and exiting sleep to see what might be wrong,

  • How to see what the sleep button does (i know that there is a configuration file somewhere to show what terminal code it runs),

  • What is the sequence of programs it calls when it goes to sleep to know when to transfer the memory to swap and turn off the power,

My guess is that with the update, it deconfigured sleep into probably a hybrid mode? And if that is the case I would like to return how it was before, suspending to disk. Thanks a lot for any brave soul that wants to help this poor guy trying to learn more about linux and his computer.

You did not specify the motherboard, but I will rely an issue I solved in hopes it provides some ideas:

You might be victim of this Gigabyte Ryzen 7 bug. Since kernel 6.8, one of my sensors reports an over heating state upon wake:

https://forum.manjaro.org/t/acpi-thermal-limit-triggers-thermal-shutdown-on-sleep/154502/15

Use this command to review your logs after a wake/crash/reboot:
journalctl -b0 > journal.txt && journalctl -b-1 > journalpre.txt && kate journ* &

On my system, thermal_zone1/mode is the problem sensor. I create the following file:
/etc/systemd/system/gigabyte-x570-fix.service

[Unit]
Description=Gigabyte X570 Platform Sensor Fix

[Service]
ExecStart=/bin/sh -c ā€˜echo disabled > /sys/class/thermal/thermal_zone1/mode’

[Install]
WantedBy=multi-user.target

Then run this command to eanble the above changes:

sudo systemctl enable --now gigabyte-x570-fix.service

Note, ensure to check if your exact sensor(s) that exceed the limit on wake.

Hello,

My motherboard is MSI and I dont see temperature warning in the journalctl

I have the same issue and my motherboard is MSI B450 Tomahawk MAX.

I’ve also posted this issue on the discord group (its posted with the same name in case you want to take a look there). And from what I’m observing, there are MANY different causes for this problem. One of them reported that their computer would sleep and then give up, entering a zombie state. But my case is different.

Here are some logs:
journalctl -b -1 | grep -i "suspend\|sleep\|PM:"
https://paste.cachyos.org/p/c08e16a.txt

the log journalctl -b0 > journal.txt && journalctl -b-1 > journalpre.txt && kate journ* & was too long and I did not catch anything useful.

Thanks a lot, I will be trying to study more this problem, for now I will not put my pc to sleep and try to not go too deep on this situation.

There are many troubles with sleep on amd platforms.

Last kernel fix it for my laptop, after i opened a bug on bugzilla, Here, one of the AMD engineer helped to understand then fix the trouble. You should do the same, bugzilla power management forum is here

Nice! thanks I will create an account there and publish this issue there: Making sure you're not a bot!

Same issue here. Even the new 6.18RC1 kernel release does not correct the issue.

I’m also using a MSI AM5 mother board with a 9070xt.

Oct 16 22:10:13 cachyos-7950x systemd-logind[1044]: The system will suspend now!
Oct 16 22:10:13 cachyos-7950x NetworkManager[1039]: <info>  [1760613013.8118] manager: sleep: sleep requested (sleeping: no  enabled: yes)
Oct 16 22:10:13 cachyos-7950x lact[1124]: 2025-10-16T11:10:13.811894Z  INFO lact_daemon::suspend: suspend/resume event detected, reloading config
Oct 16 22:10:13 cachyos-7950x NetworkManager[1039]: <info>  [1760613013.8119] device (wlan0): state change: disconnected -> unmanaged (reason 'unmanaged-sleeping', managed-type: 'full')
Oct 16 22:10:13 cachyos-7950x kwin_wayland[1513]: Failed to delay sleep: Sender is not authorized to send message
Oct 16 22:10:13 cachyos-7950x NetworkManager[1039]: <info>  [1760613013.9183] device (p2p-dev-wlan0): state change: disconnected -> unmanaged (reason 'unmanaged-sleeping', managed-type: 'full')
Oct 16 22:10:13 cachyos-7950x NetworkManager[1039]: <info>  [1760613013.9184] manager: NetworkManager state is now ASLEEP
Oct 16 22:10:13 cachyos-7950x systemd[1]: Reached target Sleep.
Oct 16 22:10:13 cachyos-7950x systemd[1]: Starting System Suspend...
Oct 16 22:10:13 cachyos-7950x systemd-sleep[45848]: Successfully froze unit 'user.slice'.
Oct 16 22:10:13 cachyos-7950x systemd-sleep[45848]: Performing sleep operation 'suspend'...
Oct 16 22:10:13 cachyos-7950x kernel: PM: suspend entry (deep)
Oct 16 22:10:14 cachyos-7950x kernel: printk: Suspending console(s) (use no_console_suspend to debug)
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: Root hub is not suspended
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: pci_pm_suspend(): hcd_pci_suspend returns -16
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: dpm_run_callback(): pci_pm_suspend returns -16
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: failed to suspend async: error -16
Oct 16 22:10:14 cachyos-7950x kernel: queueing ieee80211 work while going to suspend
Oct 16 22:10:14 cachyos-7950x kernel: PM: Some devices failed to suspend, or early wake event detected
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: Root hub is not suspended
Oct 16 22:10:14 cachyos-7950x kernel: PM: suspend exit
Oct 16 22:10:14 cachyos-7950x kernel: PM: suspend entry (s2idle)
Oct 16 22:10:14 cachyos-7950x kernel: printk: Suspending console(s) (use no_console_suspend to debug)
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: Root hub is not suspended
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: pci_pm_suspend(): hcd_pci_suspend returns -16
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: dpm_run_callback(): pci_pm_suspend returns -16
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: PM: failed to suspend async: error -16
Oct 16 22:10:14 cachyos-7950x kernel: PM: Some devices failed to suspend, or early wake event detected
Oct 16 22:10:14 cachyos-7950x kernel: xhci_hcd 0000:19:00.0: Root hub is not suspended
Oct 16 22:10:14 cachyos-7950x systemd-sleep[45848]: Failed to put system to sleep. System resumed again: Device or resource busy
Oct 16 22:10:14 cachyos-7950x kernel: PM: suspend exit
Oct 16 22:10:14 cachyos-7950x lact[1124]: 2025-10-16T11:10:14.838917Z  INFO lact_daemon::suspend: suspend/resume event detected, reloading config
Oct 16 22:10:14 cachyos-7950x systemd-sleep[45848]: Successfully thawed unit 'user.slice'.
Oct 16 22:10:14 cachyos-7950x systemd[1]: systemd-suspend.service: Main process exited, code=exited, status=1/FAILURE
Oct 16 22:10:14 cachyos-7950x systemd[1]: systemd-suspend.service: Failed with result 'exit-code'.
Oct 16 22:10:14 cachyos-7950x systemd[1]: Failed to start System Suspend.
Oct 16 22:10:14 cachyos-7950x systemd[1]: Dependency failed for Suspend.
Oct 16 22:10:14 cachyos-7950x systemd[1]: suspend.target: Job suspend.target/start failed with result 'dependency'.
Oct 16 22:10:14 cachyos-7950x systemd-logind[1044]: Operation 'suspend' finished.
Oct 16 22:10:14 cachyos-7950x NetworkManager[1039]: <info>  [1760613014.8388] manager: sleep: wake requested (sleeping: yes  enabled: yes)
Oct 16 22:10:14 cachyos-7950x NetworkManager[1039]: <info>  [1760613014.8389] device (enp9s0): state change: activated -> unmanaged (reason 'unmanaged-sleeping', managed-type: 'full')
Oct 16 22:10:14 cachyos-7950x systemd[1]: Stopped target Sleep.


I searched my cachyOS sleep problem and ended up on this thread.

I think I have the same problem. I have a MSI MAG X870 TOMAHAWK motherboard with Ryzen 7 CPU. When I didn’t have an external GPU, the sleep feature worked perfectly. The fans from the PC were stopped, the lights were off, but when I pressed the power, everything was back again.

Now I have a RX 9070 XT GPU and my PC doesn’t go to sleep anymore. When ā€œsleepingā€, the fans of the PC are still spinning and the lights are on. When I move the mouse, click or press any keys, or press the power button, it’s random… sometimes all the fans jump to max speed, something nothing happens and I can’t wake the PC anymore.

I disabled the sleep feature in KDE, I selected ā€œPower saveā€ after 15 minutes. Not sure if it does anything, but I need to wake up my PC.

I can reproduce this and post some journalctl logs if anyone wants to take a look.

Thank you!

I am facing the same problem… At first it didn’t work, after an update it did work, but since the very last update, it doesn’t work again. The screen turns off, but the computer stays in a stalled state, I have to reset to continue. Mobo is a B75M-ITX. I did a debug a while ago, I don’t remember the details but it was certainly the AMD GPU driver that stalled the process.

I have same issue, ryzen 5 5600G, b450 aorus elite v2, rx 6600

When I added amdgpu.runpm=0’ to the GRUB_CMDLINE_LINUX_DEFAULT in /etc/default/grub my system got back from standby and sleep-modes.

Hello, I was wondering if you had solved this issue, as I am running into the same exact issue with a 9070xt as well, using a MSI mobo and a AM5 CPU, or if you hadn’t found any concrete solutions

Hello everybody.

Thanks a lot for all the help.

I’m here to say that I did absolutely anything (only kept my system updated using cachy-update), and tried my best to avoid putting it to sleep.

Recently, I accidentally pressed the sleep button in my keyboard but then, I could revive it back! The problem went away in an update!

My guess is that some developer looked at some of those issues somewhere and decided to fix it.

If anybody gets this problem, I guess the best to do is to post the logs everywhere in the hopes somebody fixes your issue.

Thanks all you guys for the support!

I have the same problem on MSI Motherboard and Radeon RX 9060 XT GPU. I’ll try do add amdgpu.aspm=0, amdgpu.runpm=0 in the boot parameters and I will see if this helps

lip 03 23:50:12 CachyDesktop kernel: Modules linked in: rfcomm snd_seq_dummy snd_hrtimer snd_seq algif_hash algif_skcipher af_alg bnep vfat fat amd_atl intel_rapl_msr intel_rapl_common snd_hda_codec_alc662 kvm_amd snd_hda_codec_atihdmi>
lip 03 23:50:12 CachyDesktop kernel: nf_tables nfnetlink dm_mod i2c_dev ntsync crypto_user pkcs8_key_parser zram 842_decompress 842_compress lz4hc_compress lz4_compress amdgpu drm_panel_backlight_quirks drm_buddy drm_suballoc_helper i>
lip 03 23:50:12 CachyDesktop kernel: CPU: 1 UID: 0 PID: 16412 Comm: kworker/u64:51 Tainted: G W 7.1.2-3-cachyos#1PREEMPT(full) 9d32af03482519c3acf7808a26f1cfbefc6f6ed5
lip 03 23:50:12 CachyDesktop kernel: Tainted: [W]=WARN
lip 03 23:50:12 CachyDesktop kernel: Hardware name: Micro-Star International Co., Ltd. MS-7E56/B850 GAMING PLUS WIFI (MS-7E56), BIOS 1.AA2 06/25/2026
lip 03 23:50:12 CachyDesktop kernel: Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched]
lip 03 23:50:12 CachyDesktop kernel: RIP: 0010:amdgpu_irq_put+0x120/0x130 [amdgpu]
lip 03 23:50:12 CachyDesktop kernel: Code: 89 da 4c 8b 18 2e 2e 2e 41 ff d3 89 c3 4c 89 e7 4c 89 ee e8 12 53 ee d9 89 d8 5b 41 5c 41 5d 41 5e 41 5f 5d e9 9c ce 9d d8 cc <0f> 0b eb ec cc cc cc cc cc cc cc cc cc cc cc cc 90 90 90 90 90 90
lip 03 23:50:12 CachyDesktop kernel: RSP: 0018:ffffd04e45dcfbe8 EFLAGS: 00010246
lip 03 23:50:12 CachyDesktop kernel: RAX: 00000000ffffffea RBX: 0000000000000000 RCX: ffff8ce8cd06a148
lip 03 23:50:12 CachyDesktop kernel: RDX: ffffffffc1994360 RSI: ffff8ce8cdd2f008 RDI: ffff8ce8dcf00000
lip 03 23:50:12 CachyDesktop kernel: RBP: 0000000000000000 R08: 0000000000000002 R09: 3fffffffffffefff
lip 03 23:50:12 CachyDesktop kernel: R10: ffffffff9be74ab0 R11: ffffffffc0d76210 R12: ffff8ce8dcf00000
lip 03 23:50:12 CachyDesktop kernel: R13: ffff8ce8dcf571f0 R14: ffff8ce8cdd2f008 R15: ffff8ce8dcf00000
lip 03 23:50:12 CachyDesktop kernel: FS: 0000000000000000(0000) GS:ffff8cf0612c3000(0000) knlGS:0000000000000000
lip 03 23:50:12 CachyDesktop kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
lip 03 23:50:12 CachyDesktop kernel: CR2: 000055c6971ec138 CR3: 0000000560c12000 CR4: 0000000000f50ef0
lip 03 23:50:12 CachyDesktop kernel: PKRU: 55555554
lip 03 23:50:12 CachyDesktop kernel: Call Trace:
lip 03 23:50:12 CachyDesktop kernel:
lip 03 23:50:12 CachyDesktop kernel: smu_smc_hw_cleanup+0x6d/0x3b0 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: smu_suspend.llvm.8808524015206972483+0x61/0xf0 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_ip_block_suspend+0x27/0x70 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_device_ip_suspend_phase2+0x13a/0x2a0 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_device_pre_asic_reset+0x520/0x5e0 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_device_asic_reset+0x49/0x540 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_device_gpu_recover+0x2f7/0x440 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: amdgpu_job_timedout.llvm.12578067240936237985+0x378/0x490 [amdgpu bfc8741eb51204d2993fab05f087dc442e636643]
lip 03 23:50:12 CachyDesktop kernel: drm_sched_job_timedout+0x89/0x160 [gpu_sched dd9f7df1cd15a1f18e21d06403d5d36d9c2bd8dd]
lip 03 23:50:12 CachyDesktop kernel: process_scheduled_works+0x23d/0x5c0
lip 03 23:50:12 CachyDesktop kernel: worker_thread+0x198/0x380
lip 03 23:50:12 CachyDesktop kernel: ? __pfx_worker_thread+0x10/0x10
lip 03 23:50:12 CachyDesktop kernel: kthread+0xfb/0x120
lip 03 23:50:12 CachyDesktop kernel: ? __pfx_kthread+0x10/0x10
lip 03 23:50:12 CachyDesktop kernel: ret_from_fork+0xee/0x2b0
lip 03 23:50:12 CachyDesktop kernel: ? __pfx_kthread+0x10/0x10
lip 03 23:50:12 CachyDesktop kernel: ret_from_fork_asm+0x1a/0x30
lip 03 23:50:12 CachyDesktop kernel:
lip 03 23:50:12 CachyDesktop kernel: ---[ end trace 0000000000000000 ]---
lip 03 23:50:12 CachyDesktop kernel: amdgpu 0000:03:00.0: Fail to disable thermal alert!
lip 03 23:50:12 CachyDesktop kernel: amdgpu 0000:03:00.0: suspend of IP block failed -22
lip 03 23:50:12 CachyDesktop kernel: amdgpu 0000:03:00.0: SMU is resuming...
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: SMU: No response msg_reg: e resp_reg: 0
lip 03 23:50:15 CachyDesktop kernel: in params:00000083
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: Failed to SetDriverDramAddr!
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: Failed to setup smc hw!
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: resume of IP block failed -62
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: amdgpu_device_ip_resume_phase2 failed during unwind: -62
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU pre asic reset failed with err, -22 for drm dev, 0000:03:00.0
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: MODE1 reset
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU mode1 reset
lip 03 23:50:15 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU smu mode1 reset
lip 03 23:50:16 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU reset succeeded, trying to resume
lip 03 23:50:16 CachyDesktop kernel: amdgpu 0000:03:00.0: VRAM is lost due to GPU reset!
lip 03 23:50:16 CachyDesktop kernel: amdgpu 0000:03:00.0: SMU is resuming...
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: SMU: No response msg_reg: e resp_reg: 0
lip 03 23:50:19 CachyDesktop kernel: in params:00000083
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: Failed to SetDriverDramAddr!
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: Failed to setup smc hw!
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: resume of IP block failed -62
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU reset end with ret = -62
lip 03 23:50:19 CachyDesktop kernel: amdgpu 0000:03:00.0: GPU Recovery Failed: -62

Your problem when it comes to hibernation is most likely this:

zram can easily break the sleep cycle if your memory usage is too high. I had the same problem until I disabled it.

This blog post shows a reference to this happening and explains some of the cons of using it on certain systems.

It’s also mentioned in the arch wiki that it doesn’t support hibernating to disk aka swap hibernation.

It is easy to change over from zram or disable it, and a guide is included here:

You most likely will want to use zswap, not just disable zram as zswap provides a lot of the same benefits for your drives but still allows you to suspend the system.

This thread helped me fix this problem on my own machine. Thanks y’all!

Because of that, and because I had spare time and scratch for the model requests, I used opencode and a rogue’s gallery of recent models to grind out a tool for automating the zram-zswap change with backups and automated rollback: GitHub - boundring/cachyos-zswap-migrate: Interactive, logged, reversible CachyOS zram->zswap + hibernation swapfile migration tool Ā· GitHub

hope that helps anybody at all, haven’t put in the time and effort for getting dracut handled in that yet