[SOLVED] Laptop no longer suspending

amr962 · 2024-02-25 22:58:30

Following a pacman update:

[2024-02-24T14:07:14+0100] [ALPM] transaction started
[2024-02-24T14:07:14+0100] [ALPM] upgraded c-ares (1.26.0-1 -> 1.27.0-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded grpc (1.60.1-1 -> 1.62.0-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded bear (3.1.3-8 -> 3.1.3-9)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libgpg-error (1.47-1 -> 1.48-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded git (2.43.2-1 -> 2.44.0-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded spirv-tools (2023.5-1 -> 2023.6-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded glslang (13.1.1-3 -> 14.0.0-1)
[2024-02-24T14:07:14+0100] [ALPM] installed libblockdev-utils (3.0.4-3)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libblockdev (3.0.4-2 -> 3.0.4-3)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libngtcp2 (1.2.0-1 -> 1.3.0-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libnm (1.44.2-3 -> 1.46.0-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded shaderc (2023.7-2 -> 2023.8-1)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libplacebo (6.338.2-1 -> 6.338.2-2)
[2024-02-24T14:07:14+0100] [ALPM] upgraded libpng (1.6.42-1 -> 1.6.43-1)
[2024-02-24T14:07:15+0100] [ALPM] upgraded xmlsec (1.2.39-1 -> 1.3.3-2)
[2024-02-24T14:07:15+0100] [ALPM] upgraded libsecret (0.21.3-1 -> 0.21.4-1)
[2024-02-24T14:07:16+0100] [ALPM] upgraded libreoffice-fresh (24.2.0-1 -> 24.2.0-2)
[2024-02-24T14:07:17+0100] [ALPM] upgraded linux (6.7.5.arch1-1 -> 6.7.6.arch1-1)
[2024-02-24T14:07:17+0100] [ALPM] upgraded networkmanager (1.44.2-3 -> 1.46.0-1)
[2024-02-24T14:07:17+0100] [ALPM] upgraded openmpi (5.0.2-3 -> 5.0.2-4)
[2024-02-24T14:07:17+0100] [ALPM] upgraded qtractor (0.9.39-2 -> 0.9.39-3)
[2024-02-24T14:07:17+0100] [ALPM] upgraded thunderbird (115.7.0-2 -> 115.8.0-1)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-crypto (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-fs (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-mdraid (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-loop (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-nvme (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-part (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] installed libblockdev-swap (3.0.4-3)
[2024-02-24T14:07:17+0100] [ALPM] upgraded udisks2 (2.10.1-2 -> 2.10.1-3)
[2024-02-24T14:07:17+0100] [ALPM] transaction completed

I've not been able to suspend, the screen goes dark but the fan and power led are still on. Can't do anything beside long pressing the power button to forcefully power off. Journalctl shows this:

Feb 25 23:13:39 lenovo systemd-logind[758]: Lid closed.
Feb 25 23:13:42 lenovo systemd-logind[758]: Suspending...
Feb 25 23:13:42 lenovo NetworkManager[751]: <info>  [1708899222.9876] manager: sleep: sleep requested (sleeping: no  enabled: yes)
Feb 25 23:13:42 lenovo NetworkManager[751]: <info>  [1708899222.9877] device (p2p-dev-wlp1s0): state change: disconnected -> unmanaged (reason 'sleeping', sys-iface-state: 'managed')
Feb 25 23:13:42 lenovo NetworkManager[751]: <info>  [1708899222.9881] manager: NetworkManager state is now ASLEEP
Feb 25 23:13:42 lenovo NetworkManager[751]: <info>  [1708899222.9883] device (wlp1s0): state change: activated -> deactivating (reason 'sleeping', sys-iface-state: 'managed')
Feb 25 23:13:43 lenovo systemd[1]: Starting Network Manager Script Dispatcher Service...
Feb 25 23:13:43 lenovo systemd[1]: Started Network Manager Script Dispatcher Service.
Feb 25 23:13:43 lenovo kernel: wlp1s0: deauthenticating from X by local choice (Reason: 3=DEAUTH_LEAVING)
Feb 25 23:13:43 lenovo wpa_supplicant[863]: wlp1s0: CTRL-EVENT-DISCONNECTED bssid=X reason=3 locally_generated=1
Feb 25 23:13:43 lenovo wpa_supplicant[863]: wlp1s0: CTRL-EVENT-DSCP-POLICY clear_all
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3369] device (wlp1s0): supplicant interface state: completed -> disconnected
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3372] device (wlp1s0): state change: deactivating -> disconnected (reason 'sleeping', sys-iface-state: 'managed')
Feb 25 23:13:43 lenovo wpa_supplicant[863]: wlp1s0: CTRL-EVENT-REGDOM-CHANGE init=CORE type=WORLD
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3690] dhcp4 (wlp1s0): canceled DHCP transaction, DHCP client pid 1189
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3691] dhcp4 (wlp1s0): activation: beginning transaction (timeout in 45 seconds)
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3691] dhcp4 (wlp1s0): state changed no lease
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3695] dhcp6 (wlp1s0): canceled DHCP transaction
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.3696] dhcp6 (wlp1s0): state changed no lease
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.4523] device (wlp1s0): set-hw-addr: set MAC address to X (scanning)
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.6142] device (wlp1s0): supplicant interface state: disconnected -> interface_disabled
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.6143] device (wlp1s0): supplicant interface state: interface_disabled -> disconnected
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.6149] device (wlp1s0): state change: disconnected -> unmanaged (reason 'sleeping', sys-iface-state: 'managed')
Feb 25 23:13:43 lenovo NetworkManager[751]: <info>  [1708899223.6969] device (wlp1s0): set-hw-addr: reset MAC address to X (unmanage)
Feb 25 23:13:43 lenovo systemd[1]: Created slice Slice /system/suspend.
Feb 25 23:13:43 lenovo systemd[1]: Starting Lock the screen...
Feb 25 23:13:43 lenovo wpa_supplicant[863]: wlp1s0: CTRL-EVENT-DSCP-POLICY clear_all
Feb 25 23:13:43 lenovo wpa_supplicant[863]: wlp1s0: CTRL-EVENT-DSCP-POLICY clear_all
Feb 25 23:13:43 lenovo wpa_supplicant[863]: nl80211: deinit ifname=wlp1s0 disabled_11b_rates=0
Feb 25 23:13:43 lenovo lock.bash[1406]: No players found
Feb 25 23:13:43 lenovo systemd[1]: Started Lock the screen.
Feb 25 23:13:43 lenovo systemd[1]: Reached target Sleep.
Feb 25 23:13:43 lenovo systemd[1]: Starting System Suspend...
Feb 25 23:13:43 lenovo systemd-sleep[1414]: Performing sleep operation 'suspend'...
Feb 25 23:13:43 lenovo kernel: PM: suspend entry (s2idle)

Also sometimes without the last line. I suspected the kernel update was the problem, so I installed linux-lts but still had the same issues. Using linux-lts did however solve it, if I try to suspend on battery power, suspending when plugged in gives the same output.
No idea what otherwise could be the problem, any help is appreciated.

[amr@lenovo ~]$ cat /etc/kernel/cmdline 
rw quiet bgrt_disable amd_pstate=passive lsm=landlock,lockdown,yama,integrity,apparmor,bpf audit=1

[amr@lenovo ~]$ uname -a
Linux lenovo 6.6.18-1-lts #1 SMP PREEMPT_DYNAMIC Fri, 23 Feb 2024 14:58:40 +0000 x86_64 GNU/Linux

Last edited by amr962 (2024-04-03 20:32:01)

seth · 2024-02-26 09:50:42

Did you try to downgrade the 6.7 kernel in isolation?

amr962 · 2024-02-26 10:49:48

seth wrote:

Did you try to downgrade the 6.7 kernel in isolation?

I did not, would the be simply installing the older kernel from the pacman cache?

amr962 · 2024-02-26 11:27:32

Well, it just happened on the lts kernel while not being plugged in.

amr962 · 2024-02-26 12:31:13

Simply downgrading the linux package with

pacman -U /var/cache/pacman/pkg/linux-6.7.5.arch1-1-x86_64.pkg.tar.zst

fixed suspending on and off power, so it seems like both kernels are affected.

seth · 2024-02-26 15:47:23

Assuming this isn't a fluke, there 315 commits in the update, before bisecting this:

Can't do anything beside long pressing the power button to forcefully power off.

Does https://wiki.archlinux.org/title/Keyboa … el_(SysRq) work (nb. that you've to explicitly enable it!)?

amr962 · 2024-02-26 16:34:56

After updating to test REISUB reboot:

[2024-02-26T17:11:58+0100] [ALPM] transaction started
[2024-02-26T17:11:58+0100] [ALPM] upgraded xz (5.4.6-1 -> 5.6.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libjxl (0.9.2-1 -> 0.10.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded ffmpeg (2:6.1.1-5 -> 2:6.1.1-6)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libksba (1.6.5-1 -> 1.6.6-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded gimp (2.10.36-4 -> 2.10.36-5)
[2024-02-26T17:11:59+0100] [ALPM] upgraded imagemagick (7.1.1.29-1 -> 7.1.1.29-2)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-utils (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-crypto (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-fs (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-loop (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-mdraid (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-nvme (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-part (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:11:59+0100] [ALPM] upgraded libblockdev-swap (3.0.4-3 -> 3.1.0-1)
[2024-02-26T17:12:00+0100] [ALPM] upgraded linux (6.7.5.arch1-1 -> 6.7.6.arch1-1)
[2024-02-26T17:12:00+0100] [ALPM] upgraded rust-analyzer (20240219-1 -> 20240226-1)
[2024-02-26T17:12:00+0100] [ALPM] upgraded sdl2_image (2.8.2-2 -> 2.8.2-4)
[2024-02-26T17:12:00+0100] [ALPM] upgraded thin-provisioning-tools (1.0.11-1 -> 1.0.12-1)
[2024-02-26T17:12:00+0100] [ALPM] upgraded webkit2gtk (2.42.5-1 -> 2.42.5-2)
[2024-02-26T17:12:01+0100] [ALPM] upgraded webkit2gtk-4.1 (2.42.5-1 -> 2.42.5-2)
[2024-02-26T17:12:01+0100] [ALPM] transaction completed

I tested again whether it was possible to suspend, didn't work (running on battery). I enabled sysrq in the session using sysctl, and tested it to confirm it was working. I then closed the lid and attempted to reboot, didn't work.

andyturfer · 2024-02-26 23:56:21

I have the same issue.

Laptop: Legion 7 gen 7 16ARHA7.
CPU: AMD Ryzen 7 6800H.
iGPU: AMD Radeon RX 680m
dGPU: AMD Radeon RX 6700m

The issue only occurs for me when:

I switch to "dGPU mode" in the BIOS (as opposed to "Switchable Graphics")
Kernel is 6.7 or higher

Suspend still works on kernel 6.6.10:

sudo pacman -U https://archive.archlinux.org/packages/l/linux/linux-6.6.10.arch1-1-x86_64.pkg.tar.zst

If I install a 6.7.x kernel, trying to suspend results in:

Black screen.
Keyboard backlighting remaining on.
CTRL + ALT + F3 (to F12) does not bring up a console.
Rapidly pressing ALT + CTRL + DEL does nothing.
Unable to ssh into the laptop (sshd.service is enabled).

I then have to long-press the power button to power down the laptop.

Suspend works (on the dGPU) with the linux-lts kernel (6.6.18).

Last edited by andyturfer (2024-02-27 10:08:39)

seth · 2024-02-27 13:36:17

Edit: just for cross reference, sorry.
https://bbs.archlinux.org/viewtopic.php?id=292867

Last edited by seth (2024-02-27 13:36:38)

njodoin · 2024-02-28 20:40:15

I'm also having this issue since updating to kernel 6.7.6-arch1-1.

Hibernation works, so it appears to just be suspend that hangs.
Frustratingly, there is no log error in the journal from the previous session.
Everything looks as I would expect from a suspend.

Downgrading from 6.7.6 to 6.7.5 as below works flawlessly after rebooting.

sudo pacman -U https://archive.archlinux.org/packages/l/linux/linux-6.7.5.arch1-1-x86_64.pkg.tar.zst

I could try looking into the kernel changes between 6.7.5 and 6.7.6, but I'm woefully unfamiliar with its structure. Would appreciate if anyone with more experience with the kernel might be able to weigh in.

Any idea what in the kernel update might have caused this?

Model: System76 Pang12
CPU: Ryzen 7 6800U
GFX: Radeon 680M
RAM: 31330MiB

Last edited by njodoin (2024-02-28 21:06:15)

seth · 2024-02-28 21:23:38

For semi-blind guesswork on the OP

commit d855ceb6a5fde668c5431156bc60fae0cc52b764
Author: Mario Limonciello <mario.limonciello@amd.com>
Date:   Wed Feb 7 23:52:54 2024 -0600

    Revert "drm/amd: flush any delayed gfxoff on suspend entry"
    
    commit 916361685319098f696b798ef1560f69ed96e934 upstream.
    
    commit ab4750332dbe ("drm/amdgpu/sdma5.2: add begin/end_use ring
    callbacks") caused GFXOFF control to be used more heavily and the
    codepath that was removed from commit 0dee72639533 ("drm/amd: flush any
    delayed gfxoff on suspend entry") now can be exercised at suspend again.
    
    Users report that by using GNOME to suspend the lockscreen trigger will
    cause SDMA traffic and the system can deadlock.
    
    This reverts commit 0dee726395333fea833eaaf838bc80962df886c8.
    
    Acked-by: Alex Deucher <alexander.deucher@amd.com>
    Fixes: ab4750332dbe ("drm/amdgpu/sdma5.2: add begin/end_use ring callbacks")
    Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
    Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
    Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

looks very suspicious.

But they face this with 6.7.6 and 6.6.18-lts, but not 6.7.5 whereas andyturfer has an issue w/ the 6.7 kernels but not w/ 6.6.18
I assume you rather fall into that category and you're both effectively off topic (here)

However, andyturfer also only faces this with his RX 6700m and NOT his 680m, but you *only* have an 680m?
There's a pending VRAM eviction bug and amdgpu (which would *only* affect APUs) and the S3 seems to simply fail w/ an allocation error.

Can either of you trigger this from the multi-user.target (2nd link below)
Any chance to get out of the state w/ the https://wiki.archlinux.org/title/Keyboa … el_(SysRq)

njodoin · 2024-02-28 22:04:31

seth wrote:

But they face this with 6.7.6 and 6.6.18-lts, but not 6.7.5 whereas andyturfer has an issue w/ the 6.7 kernels but not w/ 6.6.18
I assume you rather fall into that category and you're both effectively off topic (here)
However, andyturfer also only faces this with his RX 6700m and NOT his 680m, but you *only* have an 680m?

It is unclear to me whether the others have tried rolling back incrementally or by major releases/release type.

I would be curious to know OP or andyturfer tried rolling back to 6.7.5 or not, though I do recognize that andyturfer's circumstances differ. I believe my issue may align with the OP's, however, more investigation is needed to be sure.

On that note...
@amr962 & @andyturfer: Did you try rolling back to 6.7.5 instead of the LTS? Could you try 6.7.5 specifically?
I'm hoping to isolate the smallest possible surface area, as that could really help isolate the root cause.

seth wrote:

There's a pending VRAM eviction bug and amdgpu (which would *only* affect APUs) and the S3 seems to simply fail w/ an allocation error.

Thank you for mentioning this. I don't necessarily understand what a VRAM eviction bug is. Would you please elaborate on this, and/or link the bugtracker issue for the bug in question so I can try to better understand?

seth wrote:

Can either of you trigger this from the multi-user.target (2nd link below)
Any chance to get out of the state w/ the https://wiki.archlinux.org/title/Keyboa … el_(SysRq)

Yes, the behavior persists in multi-user.target and was consistent between reboots.

Here's the journal tail from one of the relevant multi-user sessions. This is identical to the journal contents when in graphical.target.

❯ sudo journalctl -b-2 | tail -6
Feb 28 15:46:05 necromorph systemd[1]: Reached target Sleep.
Feb 28 15:46:05 necromorph systemd[1]: Starting System Suspend...
Feb 28 15:46:05 necromorph wpa_supplicant[1053]: wlp2s0: CTRL-EVENT-DSCP-POLICY clear_all
Feb 28 15:46:05 necromorph wpa_supplicant[1053]: nl80211: deinit ifname=wlp2s0 disabled_11b_rates=0
Feb 28 15:46:05 necromorph systemd-sleep[1447]: Performing sleep operation 'suspend'...
Feb 28 15:46:05 necromorph kernel: PM: suspend entry (s2idle)

No, sysrq reboot does not affect the hang state. The only method I've found for recovering is a hard power reset by holding the power button, same as OP. I tested the sysrq shortcut from both multi-user.target and graphical.target environments to be certain I was executing the command sequence correctly.

@seth: Thank you for coming in and helping with this. Maintaining a system with a frozen kernel version is not ideal. Hopefully the next kernel update will resolve this, though I would like to get to the bottom of this if only to better understand how the system works.

Last edited by njodoin (2024-02-28 22:08:51)

njodoin · 2024-02-28 22:10:02

On the note of the differing solution versions, OP's update logs show a kernel version change from 6.7.5 > 6.7.6 in the breaking update, so in my mind trying a single version kernel downgrade to 6.7.5 ought to be the first step rather than switching to the LTS kernel.

seth · 2024-02-28 22:10:44

amr962 wrote:

Simply downgrading the linux package with
pacman -U /var/cache/pacman/pkg/linux-6.7.5.arch1-1-x86_64.pkg.tar.zst
fixed suspending on and off power, so it seems like both kernels are affected.

For context, andyturfer's own thread (I assume it's still the same issue)
https://bbs.archlinux.org/viewtopic.php?id=292633

Thread with the amdgpu VRAM eviction bug, https://bbs.archlinux.org/viewtopic.php … 0#p2152270

njodoin · 2024-02-28 22:18:41

Thanks for pointing out OP's message about the fix, I didn't realize I'd overlooked the message when I initially posted here. Disregard previous thoughts on trying 6.7.5 downgrade then, as it's been done.

I also saw andyturfer's thread as well while doing some digging, but as the behavior was slightly different I didn't dig into it much.

Thank you again for passing along the information on the bug; I'll have to do some reading.

Last edited by njodoin (2024-02-28 22:19:36)

njodoin · 2024-02-28 22:26:04

Alright, I'm caught up on the bug. Highly likely it's related, based on the behavior and affected hardware. Will have to sit tight and see if next release will contain a fix.

Last edited by njodoin (2024-02-28 22:27:31)

andyturfer · 2024-02-29 11:25:23

I've done some further investigation into this, and can replicate the issue others are having here when using X11 instead of Wayland:

X11, Radeon RX 6700m ("discrete graphics" set in BIOS) - sleep works with kernel 6.7.5, does not work with 6.7.6.
Wayland, Radeon RX 6700m ("discrete graphics" set in BIOS) - sleep works with 6.6.x, does not work with 6.7.x.
X11, Radeon RX 680m ("switchable graphics" set in BIOS) - sleep works with all kernels (6.6.x and 6.7.x).
Wayland, Radeon RX 680m ("switchable graphics" set in BIOS) - sleep works with all kernels (6.6.x and 6.7.x).

When testing in "switchable graphics" mode (i.e. so the RX 680m is the primary GPU), I also tested suspend while the RX 6700m was up and running (i.e. DRI_PRIME=1 firefox, and then open a YouTube video).

From the above, "does not work" means:

The screen goes black
Keyboard backlighting stays on
Fans are still spinning
Keyboard is unresponsive
Long pressing the power button to power down is the only option

Last edited by andyturfer (2024-02-29 11:29:08)

seth · 2024-02-29 15:47:05

Are you using xf86-video-amdgpu (just to explain the displayserver-depending behavior)

Also: https://bbs.archlinux.org/viewtopic.php … 4#p2153924

andyturfer · 2024-02-29 16:16:06

I haven't installed xf86-video-amdgpu:

# pacman -Q | grep xf86
libxxf86vm 1.1.5-1
xf86-input-libinput 1.4.0-1
xf86-video-amdgpu 23.0.0-1
xf86-video-ati 1:22.0.0-1
xf86-video-nouveau 1.0.17-2
xf86-video-vmware 13.4.0-1

seth · 2024-02-29 16:43:48

And what is

xf86-video-amdgpu 23.0.0-1

?

andyturfer · 2024-02-29 18:23:59

Oh, sorry - completely missed that! I'll uninstall it...

andyturfer · 2024-02-29 23:34:34

Removed xf86-video-amdgpu, tested again with the discrete Radeon RX 6700m:

+-------------+---------------------+-----------------+----------------------------------------------+
| Kernel      |  Works in Wayland?  |  Works in X11?  |  Notes                                       |
+-------------+---------------------+-----------------+----------------------------------------------+
| 6.6.10      |  Yes                |  Yes            |                                              |
| 6.6.18-lts  |  Yes                |  Yes            |                                              |
| 6.7.0       |  No                 |  Yes            |  Wayland sleeps, but black screen on wake.   |
| 6.7.1       |  No                 |  No             |  Wayland and X11 don't sleep (black screen). |
| 6.7.2       |  No                 |  No             |  Wayland and X11 don't sleep (black screen). |
| 6.7.3       |  Yes                |  Yes            |                                              |
| 6.7.4       |  Yes                |  Yes            |                                              |
| 6.7.5       |  No                 |  Yes            |  Wayland doesn't sleep (black screen).       |
| 6.7.6       |  No                 |  No             |  Wayland and X11 don't sleep (black screen). |
+-------------+---------------------+-----------------+----------------------------------------------+

menghuan · 2024-03-01 02:29:03

I also have the same issue.

Model:Xiaomi RedmiBook Pro 15 2022
CPU: AMD Ryzen 7 6800H.
iGPU: AMD Radeon RX 680m

I can't do suspend either, the screen goes black but the computer fan doesn't stop, the only thing I can do is long press the power button to force a shutdown. Downgrading linux from 6.7.6 to 6.7.5 solved this issue.

amr962 · 2024-03-01 10:51:58

Woah, hello everyone! Didn't realize this was already hijacked multiple times

To provide some more info from my end:

I'm using a Ryzen 6800HS with no dedicated GPU.
```
[amr@lenovo ~]$ cat /sys/power/mem_sleep 
[s2idle]
```
My machine doesn't support S3 sleep, so not sure how VRAM eviction bug would affect me.

seth wrote:

...
looks very suspicious.

I'm clueless on kernel stuff, but would it help if I compiled just the two versions till and including this commit, to see if it's causing the problem?

As to the lts kernel, it wasn't exactly the same behavior; it was firstly only happening when plugged in, but then it randomly also happened on battery power. So it's not as consistent.

seth · 2024-03-01 16:43:15

The better check would be to simply revert that commit
To approximate the situation, loqs has build a bunch of kernels between 6.7.5 and 6.7.6 for what was ultimately a kernel config bug.
https://drive.google.com/drive/folders/ … DW8rfPKbcp

You might also want to try the 6.7.6.arch1-2 rebuild, https://bbs.archlinux.org/viewtopic.php … 7#p2154117 (but this lead to keyboard disfunction over there.)

Arch Linux

#1 2024-02-25 22:58:30

[SOLVED] Laptop no longer suspending

#2 2024-02-26 09:50:42

Re: [SOLVED] Laptop no longer suspending

#3 2024-02-26 10:49:48

Re: [SOLVED] Laptop no longer suspending

#4 2024-02-26 11:27:32

Re: [SOLVED] Laptop no longer suspending

#5 2024-02-26 12:31:13

Re: [SOLVED] Laptop no longer suspending

#6 2024-02-26 15:47:23

Re: [SOLVED] Laptop no longer suspending

#7 2024-02-26 16:34:56

Re: [SOLVED] Laptop no longer suspending

#8 2024-02-26 23:56:21

Re: [SOLVED] Laptop no longer suspending

#9 2024-02-27 13:36:17

Re: [SOLVED] Laptop no longer suspending

#10 2024-02-28 20:40:15

Re: [SOLVED] Laptop no longer suspending

#11 2024-02-28 21:23:38

Re: [SOLVED] Laptop no longer suspending

#12 2024-02-28 22:04:31

Re: [SOLVED] Laptop no longer suspending

#13 2024-02-28 22:10:02

Re: [SOLVED] Laptop no longer suspending

#14 2024-02-28 22:10:44

Re: [SOLVED] Laptop no longer suspending

#15 2024-02-28 22:18:41

Re: [SOLVED] Laptop no longer suspending

#16 2024-02-28 22:26:04

Re: [SOLVED] Laptop no longer suspending

#17 2024-02-29 11:25:23

Re: [SOLVED] Laptop no longer suspending

#18 2024-02-29 15:47:05

Re: [SOLVED] Laptop no longer suspending

#19 2024-02-29 16:16:06

Re: [SOLVED] Laptop no longer suspending

#20 2024-02-29 16:43:48

Re: [SOLVED] Laptop no longer suspending

#21 2024-02-29 18:23:59

Re: [SOLVED] Laptop no longer suspending

#22 2024-02-29 23:34:34

Re: [SOLVED] Laptop no longer suspending

#23 2024-03-01 02:29:03

Re: [SOLVED] Laptop no longer suspending

#24 2024-03-01 10:51:58

Re: [SOLVED] Laptop no longer suspending

#25 2024-03-01 16:43:15

Re: [SOLVED] Laptop no longer suspending

Board footer