You are not logged in.
My display remains off while attempting to resume from suspend. I can click my mouse or hit buttons on my keyboard and my keyboard lights turn back on (so the system isn't entirely hanging), but I get no display output whatsoever, and I have to reboot the computer. I'm seeing this in red text in the systemd journal while resuming from suspend after a reboot:
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: SMU: I'm not done with your previous command: SMN_C2PMSG_66:0x0000003A SMN_C2PMSG_82:0x00000000
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: RunBtc failed!
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Failed to setup smc hw!
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: resume of IP block <smu> failed -62
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: amdgpu_device_ip_resume failed (-62).
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: PM: dpm_run_callback(): pci_pm_resume returns -62
Aug 16 20:12:55 EternityPC kernel: amdgpu 0000:08:00.0: PM: failed to resume async: error -62
I've got a Ryzen 5600X/Radeon RX 5700 XT in this computer, and I'm on the Zen kernel (although the exact same thing happens on the most recent base and lts kernels) and XFCE. This is a very fresh install, so I doubt I have anything installed that would need manually updating (other than the BIOS, which I can't even convince Windows to reflash). Any ideas? If there are more logs or troubleshooting suggestions, let me know.
Last edited by eternalflameiv (2025-08-17 03:24:08)
Offline
SMU: I'm not done with your previous command
Please post your complete system journal for the boot, eg.
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
for the previous (-1) one.
Offline
SMU: I'm not done with your previous command
Please post your complete system journal for the boot, eg.
sudo journalctl -b -1 | curl -F 'file=@-' 0x0.st
for the previous (-1) one.
As requested, a full journal: http://0x0.st/KrR6.txt
All I did this boot was enable SysRq temporarily, tell the computer to sleep and try to wake it up, then sync the logs to disk with a SysRq reboot.
Offline
Aug 17 13:38:22 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Starting sdma0 ring reset
Aug 17 13:38:22 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Ring sdma0 reset succeeded
Aug 17 13:38:22 EternityPC kernel: amdgpu 0000:08:00.0: [drm] device wedged, but recovered through reset
Aug 17 13:38:33 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Starting sdma1 ring reset
Aug 17 13:38:33 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Ring sdma1 reset succeeded
Aug 17 13:38:33 EternityPC kernel: amdgpu 0000:08:00.0: [drm] device wedged, but recovered through reset
Aug 17 13:38:44 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Starting gfx_0.0.0 ring reset
Aug 17 13:38:44 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Ring gfx_0.0.0 reset failed
Aug 17 13:38:44 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: GPU reset begin!
After the resume timeout the GPU goes for multiple resets, but fails on more timeouts
Aug 17 13:38:43 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
Aug 17 13:38:43 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
Aug 17 13:38:44 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=1428, emitted seq=1432
can you get that coredump?
For mitigation efforts try
pci_aspm=off amdgpu.aspm=0
https://wiki.archlinux.org/title/Kernel_parameters
Also, sanity check:
Aug 17 13:36:23 archlinux kernel: nvme0n1: p1 p2 p3 p4 p5 p6
Parallel windows?
=> 3rd link below. Mandatory.
Disable it (it's NOT the BIOS setting!) and reboot windows and linux twice for voodo reasons.
Offline
Coredump before adding kernel cmdline args at https://0x0.st/KsKb.txt.
Fast Startup has been off since the moment I started using this computer: right now,
powercfg -a
on Windows shows S0 as the only available standby mode.
Suggested command line does not resolve the issue. After adding kernel cmdline args, full journal at http://0x0.st/KsKl.txt (notably there are some stack traces from the driver in here) and amdgpu coredump at http://0x0.st/KsKy.txt.
Last edited by eternalflameiv (2025-08-18 16:46:37)
Offline
Aug 18 12:30:13 archlinux kernel: Unknown kernel command line parameters "splash pci_aspm=off", will be passed to user space.
My typobad, sorry
pcie_aspm=off
Offline
Still no dice. Journal at http://0x0.st/KsQZ.txt, coredump at http://0x0.st/KsQN.txt.
Offline
https://bugzilla.kernel.org/show_bug.cgi?id=212655 again?
Try the "iommu=soft" kernel parameter…
Offline
Still nothing. Journal at http://0x0.st/Ks9D.txt, coredump at http://0x0.st/Ks9d.txt.
Offline
The journal doesn't have "iommu=soft" applied?
https://wiki.archlinux.org/title/Kernel_parameters
Offline
Must have forgotten to rebuild the initcpio. Anyways, did that properly but still nothing. Journal at http://0x0.st/Ksfz.txt, coredump at http://0x0.st/Ksfi.txt.
Offline
'key, sanity check: can you suspend and resume the system from the multi-user.target (2nd link below)?
Aug 19 11:32:09 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Process information: process Xorg pid 722 thread Xorg:cs0 pid 740
Aug 19 11:32:44 EternityPC kernel: amdgpu 0000:08:00.0: amdgpu: Process information: process Xorg pid 722 thread Xorg:cs0 pid 740
Do you have xf86-video-amdgpu installed? What if you remove that?
Edit: Please also post your Xorg log, https://wiki.archlinux.org/title/Xorg#General
Last edited by seth (2025-08-20 06:50:01)
Offline
Nothing changed when I logged in to the usual desktop environment (although I don't use a display manager anyways, I've got startxfce4 bound to a tty login). Journal at http://0x0.st/Kzch.txt, Xorg log at http://0x0.st/Kzc7.0.log.
xf86-video-amdgpu is not installed. Same behavior after installing it (with default target still multi-user.target unless I did something wrong); journal at http://0x0.st/KzT-.txt, Xorg log at http://0x0.st/KzTo.0.log.
Last edited by eternalflameiv (2025-08-20 19:43:12)
Offline
Interestingly enough, if I log in to a raw tty (so no DE) and suspend from the command line, I still get no video output when I try to wake, but when I type in the "reboot" command, the display comes back alive for a moment to show me a flash of the command line and then the Plymouth splash. Journal from that boot at http://0x0.st/KzTK.txt.
Offline
You logged in on tty2, what's on tty1?
Have you tried to switch the TTY after the wakeup (forth and back to/from/to X11)?
What if you skip plymouth entirely?
Offline
My .bashrc is set to exec startxfce4 on tty1.
Switching between ttys doesn't give any visible feedback from the system; the display remains off.
No change when removing "splash" from the kernel command line.
Offline
Same bug on NixOS (unresolved, just ftr): https://github.com/NixOS/nixpkgs/issues/287586
Somewhat similar (on the OP): https://bbs.archlinux.org/viewtopic.php?id=306813
Fedora: https://discussion.fedoraproject.org/t/ … gpu/133003
Gentoo: https://forums.gentoo.org/viewtopic-p-8 … bb50d87706
Deps sleep/S3 isn't available?
https://wiki.archlinux.org/title/Power_ … end_method
And before I'm running down a stupid-hole here: what's the output of
cat /sys/module/amdgpu/parameters/ppfeaturemask
(The test would be to remove 0x4 from that)
Offline
$ cat /sys/module/amdgpu/parameters/ppfeaturemask
0xfff7bfff
$ cat /sys/power/mem_sleep
[s2idle]
There's no option in the BIOS related to S3 sleep.
Offline
Does it help to disable SCLK_DEEP_SLEEP?
amdgpu.ppfeaturemask=0xfff7bff7
Offline
Unfortunately not. Journal at https://0x0.st/KiLD.txt, Xorg log at https://0x0.st/KiLd.txt.
Offline
This is a faaaaaaar stretch because of the multi-user.target situation, but https://bbs.archlinux.org/viewtopic.php?id=296954 ?
Since it's your only and active GPU we'll also not be able to apply workarounds like unbinding it and rescanning the bus after the sleep
Offline
No dice, I get the exact same behaviour on attempting to wake after rfkill blocking both WiFi and Bluetooth and putting this into a systemd sleep drop-in:
$ cat /etc/systemd/system/systemd-suspend.service.d/override.conf
[Service]
Environment="SYSTEMD_SLEEP_FREEZE_USER_SESSIONS=false"
Offline
after rfkill blocking both WiFi and Bluetooth
The problem is the GPU anyway
Curveball: since it happens when suspending from the console, does it happen if you keep glxgears running during the suspend?
Ideally in hyperactive mode, "vblank_mode=0 glxgears"
Offline
It still does. Journal (looks slightly different) at https://0x0.st/K-s9.txt, Xorg log at https://0x0.st/K-sp.txt.
I'm wondering if I should report a bug upstream at this point - waking from sleep works fine in Windows, so I suspect it's some sort of arcane bug somewhere in the driver stack. I'm not sure how I'd go about identifying which component exactly the issue is in, though.
Last edited by eternalflameiv (2025-08-24 15:59:02)
Offline
Same problem here. Resume fails since kernel 6.16.2, can be solved going back to 6.16.1.
Offline