You are not logged in.
Using systemd hook and nvidia-open
I had a working hibernate some time ago.
Outputs shown below comes from OCR. May contain errors.
Screen output after entering systemctl hibernate, before the screen turns off:
4263.7484351 NVRM: _kgspIsHeartbeatTimedOut: Heartbeat timed out, currentTimeMs 2353322420 heartbeat 0 heartbeatWithOffsetsMs 0 diff 2353322452 timeout 5200
4263.7485021 NVRM: _kgspRpcRecvPoll: GSP RM heartbeat timed out
4265.3682191 spd5118 19-0050: Failed to write b = 0: -6
4265.3682251 spd5118 19-0050: PM: dpa_run_callback(): spd5118_resume [spd5118] returns -6
4265.3682321 spd5118 19-0050: PM: failed to thaw async: error -6
4265.3682771 spd5118 19-0052: Failed to write b = 0: -6
4265.3682801 spd5118 19-0052: PM: dpa_run_callback(): spd5118_resume [spd5118] returns -6
4265.3682831 spd5118 19-0052: PM: failed to thaw async: error -6
after reboot:
4.1281061 NVRM: GPU0 nvAssertOkFailedNoLog: Insertion failed: Invalid data passed (NV_ERR_INVALID_DATA) (0x00000225) returned from PlatformRequestHandler failed to get target temp from SB10S @ platform_request_handler_ctrl.c:2117
[ 4.1282081 NVRM: GPU0 nvAssertOkFailedNoLog: Insertion failed: Invalid data passed (NV_ERR_INVALID_DATA) (0x00000225) returned from PlatformRequestHandler failed to get platform power mode from SB10S @ platform_request_handler_ctrl.c:2114
[ 4.2867171] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device DP-4
[ 4.2993691] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device DP-4
4.1281065) WRN: GPU nvRegister0xFailedLog: Insertion failed: Invalid data passed (NV_ERR_INVALID_DATA) (0x00000225) returned from PlatformRequestHandler failed to get target temp from SB105 @ platform_request_handler_ctrl.c:2117
[ 4.1282088] WRN: GPU nvRegister0xFailedLog: Insertion failed: Invalid data passed (NV_ERR_INVALID_DATA) (0x00000225) returned from PlatformRequestHandler failed to get platform power mode from SB108 @ platform_request_handler_ctrl.c:2114
[ 4.2867177] nvidia-modeset: WARNING: GPU:0: Unable to read EDID for display device DP-4
[ 9.4358722] WRN: GPU 0001:01:00.0: PreservedHardwareAllocation module parameter is set. System Power Management attempted without driver procfs suspend interface. Please refer to the 'Configuring Power Management Support' section in the driver README.
[ 9.4358811] nvidia-modeset 0001:01:00.0: PM: pci_pm_freeze(): nv_pm_pause_freeze (nvidia-modeset) returns -5
[ 9.4362481] nvidia-modeset 0001:01:00.0: PM: dpm_run_callback(): pci_pm_freeze returns -5
[ 9.4362481] nvidia-modeset 0001:01:00.0: PM: failed to quiesce async: error -5
[10.485856] PM: hibernate: Failed to load image, recovering.
Last edited by TBSword (2026-04-15 10:07:03)
Offline
I might solved the problem.
Running dmesg -T, I found this line almost right after 'Image loading progress: 100%' :
NVRM: GPU 0000:01:00.0: PreserveVideoMemoryAllocations module parameter is set. System Power Management attempted without driver procfs suspend interface. Please refer to the 'Configuring Power Management Support' section in the driver README.
After reading the wiki , I ran sort /proc/driver/nvidia/params. The outputs seemed OK:
UseKernelSuspendNotifiers: 1
TemporaryFilePath: "/var/tmp"
except there's this line
PreserveVideoMemoryAllocations: 2
I guess it is supposed be zero since I'm using nvidia-open 595+.
Here are the last set of commands I had run before the machine finally stopped telling me 'resume failed'. I might have been fooled by Gemini.
echo "options nvidia NVreg_PreserveVideoMemoryAllocations=0" | sudo tee /etc/modprobe.d/nvidia-pm-override.conf
sudo mv /usr/lib/systemd/system-sleep/nvidia /usr/lib/systemd/system-sleep/nvidia.bak
sudo mkinitcpio -P
I hope I didn't mess up anything.
Last edited by TBSword (2026-04-15 10:04:09)
Offline