Created attachment 274027 [details] journalclt -b Hello! Since Kernel 4.15 Suspend-To-RAM (S3) is broken. If I suspend via "systemctl suspend" or Fn+Suspend the laptop immediatley resumes. Steps to reproduce: 1.) Boot 2.) Login on TTY or GNOME/Wayland 3.) $ systemctl suspend What happens: Laptop resumes immediatley. What should happend: Laptop should stay in suspend. Thank you
Created attachment 274029 [details] lspci
Created attachment 274031 [details] lsusb
Created attachment 274033 [details] lscpu
Created attachment 274035 [details] installed packages from cmd pacman -Q Archlinux
Created attachment 274037 [details] journalctl -b from old kernel 4.14 with working suspend (some days ago)
Created attachment 274039 [details] dmesg
I'have a workaround "powertop --auto-tune" allows to suspend: 1.) Boot 2.) Login as root 3.) # powertop --auto-tune 4.) logout 5.) Login as any user 6.) $ systemctl suspend # laptops stays in suspend Probably the helps, maybe some hardware is set in wrong state without "powertop --auto-tune"?
An other ThinkPad T420 is not affected, the differences are: X220 | T420 USB 3.0 | - Intel SSD | Samsung SSD UEFI (1.43) | BIOS (1.46) Bluetooth 4.0* | Bluetooth 3.0 * Chip from newer ThinkPad, they are compatible
Argggh :( It not so reliable to reproduce than I thought. If your machine is not immediatley resuming from suspend, suspend and resume the laptop around five or ten times in a row (you can do that quickly). Seem to depend on some internal state of something.
Okay. There is one signifcant difference between the X220 und the T420, the first one has the package intel-ucode installed and it is loaded during bootup. Furthermore, an microcode-update from 2013 is applied, while the BIOS/UEFI is from 2016 and should (likely?) include any existing microcode-updates. I have done this: # pacman -Rns intel-ucode # grub-mkconfig -o /boot/grub/grub.cfg # reboot I've done more than 20 successful suspends to S3 since yesterday evening, none failed. I keep an eye on this, if it resolves the issue I will close the bug.
It's not gone, I was just very lucky.
Created attachment 274093 [details] acpidump -o
Created attachment 274095 [details] cat /proc/interrupts before suspend
Created attachment 274097 [details] cat /proc/interrupts after resume
Related to bluetooth, will move.