http://bugzilla.opensuse.org/show_bug.cgi?id=1207279
Bug ID: 1207279 Summary: When going to suspend, system instantly wakes up again Classification: openSUSE Product: openSUSE Tumbleweed Version: Current Hardware: x86-64 OS: openSUSE Tumbleweed Status: NEW Severity: Normal Priority: P5 - None Component: Kernel Assignee: kernel-bugs@opensuse.org Reporter: ralf.kissmann@uibk.ac.at QA Contact: qa-bugs@suse.de Found By: --- Blocker: ---
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101 Firefox/102.0 Build Identifier:
Since the change from kernel 6.0 to 6.1 suspend does not work any more on my system - neither when using the Gnome desktop menu nor when closing the lid on the laptop. In this case the system seems to briefly go to suspend only to directly wake up again (the power LED briefly turns off, just to turn on again). The system still works fine, but suspend just does not work any more. On kernel 6.0 this worked without any problem. The problem was present for the first version of kernel 6.1 and is still unchanged with kernel 6.1.4 The hardware I use is a Lenovo Thinkpad T14s G2 AMD with AMD with Ryzen 7 Pro 5850U processor.
Reproducible: Always
Steps to Reproduce: 1. Run gnome on recent AMD system 2. Suspend system Actual Results: The computer will instantly wake up again.
Expected Results: System should stay in suspend and only wake up when either opening the lid or pressing the power button
It seems that the bug is related to the kernel according a discussion in the Arch-Linux community, e.g., to be found here: https://bbs.archlinux.org/viewtopic.php?id=282293
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c1
Takashi Iwai tiwai@suse.com changed:
What |Removed |Added ---------------------------------------------------------------------------- CC| |ralf.kissmann@uibk.ac.at, | |tiwai@suse.com Flags| |needinfo?(ralf.kissmann@uib | |k.ac.at)
--- Comment #1 from Takashi Iwai tiwai@suse.com --- Does the boot option pcie_aspm=off work around the problem? If yes, it's a regression in PCIe ASPM. But currently I don't see any relevant fix in the upstream yet.
Please give hwinfo output in anyway.
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c2
--- Comment #2 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Created attachment 864249 --> http://bugzilla.opensuse.org/attachment.cgi?id=864249&action=edit Output of hwinfo
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c3
--- Comment #3 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Unfortunately, the pcie_aspm=off did not help, but it seems that the bug might be resolved anyhow: In the bios there are the options "Linux" or "Windows 10" for suspend. When selecting "Windows 10" suspend is working again. With Kernel 6.0 also the "Linux" setting was working. I will do some more tests, but right now it seems that the "Windows 10" setting has resolved the problem. Thank you for looking into this!
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c4
--- Comment #4 from Takashi Iwai tiwai@suse.com --- Thanks for the update! It's ironic that BIOS setup for Linus doesn't work on Linux...
Could you give the hwinfo outputs from both BIOS modes, as well as the dmesg outputs, for comparison?
Also, I'm building a test kernel with the reverts of two commits that are mentioned in the upstream thread (commit 5e85eba6f50d and 4ff116d0d5fd). It's being built in OBS home:tiwai:bsc1207279 repo. Once after the build finishes (takes an hour or so), it will be available at: http://download.opensuse.org/repositories/home:/tiwai:/bsc1207279/standard/ Could you give it a try later?
If the revert kernel is confirmed to work, it means that it's the same bug as the already reported one.
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c5
--- Comment #5 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Created attachment 864299 --> http://bugzilla.opensuse.org/attachment.cgi?id=864299&action=edit Output of dmesg for Linux setting for suspend in Bios
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c6
--- Comment #6 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Created attachment 864300 --> http://bugzilla.opensuse.org/attachment.cgi?id=864300&action=edit Output of dmesg for Windows 10 setting for suspend in Bios
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c7
Ralf Kissmann ralf.kissmann@uibk.ac.at changed:
What |Removed |Added ---------------------------------------------------------------------------- Attachment #864249|0 |1 is obsolete| |
--- Comment #7 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Created attachment 864301 --> http://bugzilla.opensuse.org/attachment.cgi?id=864301&action=edit Output of hwinfo for Linux setting for suspend in Bios
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c8
--- Comment #8 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Created attachment 864302 --> http://bugzilla.opensuse.org/attachment.cgi?id=864302&action=edit Output of hwinfo for Windows 10 setting for suspend in Bios
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c9
--- Comment #9 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- I just uploaded the output of hwinfo and dmesg for both suspend configurations in the BIOS. I will soon try your modified kernel.
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c10
--- Comment #10 from Takashi Iwai tiwai@suse.com --- The obvious difference is that the suspend is used on Linux mode while the s2idle on Windows mode, and this implies that a regression in PM S3 code path.
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c11
--- Comment #11 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- (In reply to Takashi Iwai from comment #4)
Thanks for the update! It's ironic that BIOS setup for Linus doesn't work on Linux...
Could you give the hwinfo outputs from both BIOS modes, as well as the dmesg outputs, for comparison?
Also, I'm building a test kernel with the reverts of two commits that are mentioned in the upstream thread (commit 5e85eba6f50d and 4ff116d0d5fd). It's being built in OBS home:tiwai:bsc1207279 repo. Once after the build finishes (takes an hour or so), it will be available at: http://download.opensuse.org/repositories/home:/tiwai:/bsc1207279/standard/ Could you give it a try later?
If the revert kernel is confirmed to work, it means that it's the same bug as the already reported one.
I just installed the test kernel. It shows the same behaviour as the other 6.1-kernels: with the "Windows 10" BIOS option suspend works but with the "Linux" option the laptop directly wakes up again.
http://bugzilla.opensuse.org/show_bug.cgi?id=1207279 http://bugzilla.opensuse.org/show_bug.cgi?id=1207279#c14
--- Comment #14 from Ralf Kissmann ralf.kissmann@uibk.ac.at --- Yes, I can confirm that suspend ("Linux mode") is working again.
kernel-bugs@lists.opensuse.org
-
bugzilla_noreply＠suse.com