Failed to resume from sleep mode below error
Not always but sometime
When resume hanged, Error message was shown and freezed
Jan 19 22:29:24 localhost kernel: pciehp 0000:00:1c.0:pcie004: link training error: status 0x1001
Jan 19 22:29:24 localhost kernel: pciehp 0000:00:1c.0:pcie004: Failed to check link status
Ok. Please run nvidia-bug-report.sh and attach output file to your post so we know what kind of device is at pci:0000:00:1c
Not necessarily the Nvidia GPU.
Come to think of it, it is unlikely that the Nvidia GPU is hung up at pci:0000:00:1c.0, this should be some chipset pcie root port. There’s probably an ethernet port or the like that’s claiming to support hotplug while it really doesn’t (reliably).
Workaround: disable hotplug for it (as root)
Has to be done after every boot, of course.
What kind of device there is you can check with
lspci -t
to get the tree of pci devices and
lspci -nn
to get the port to device mapping.
0000:00:1c.0 is just an empty slot. While this might be triggered by inserting the GPU card it looks more like buggy firmware/bios. Link training fails because there’s nothing to train.
Leaves you with the workaround, did you try it? If it works, you could make this permanent by making a systemd unit file starting it as one-shot service on boot.