suspend crashes laptop

[breezer]

Suspend stopped working when upgraded to kernel 5.3.0-41 from -40. Power-button (narrow slit) LED stays fully lit and front indicator power/wifi/drive LEDs also lit as though not suspended. Screen went black but will not respond to keyboard/touchpad input (even Ctrl-Alt-Del). Only way to recover laptop was long-press power button (5 seconds) to force it to turn off then press again as normal to cold boot.

Also note: Builtin stereo speakers also stopped working in the same kernel version change - out of scope for this bug I expect, perhaps related to SOF or Alsa migration/integration changes I saw in changelog: Builtin microphone never worked on 19.10, even after researching issue and trying assigning each pin in turn to it via hdajackretask.

Boot and shutdown and suspend and builtin speakers all working normally in kernels from initial install of 19.10 in November 2019 through 5.3.0-40 (inclusive).

Laptop is an HP Elitebook 830 G6 with ProdId 6XE61EA#...

TRIED:
installed 5.3.0-42 (latest as I write) kernel, same result as -41.
Added "snd_hda_intel.dmic_detect=0" to /etc/modprobe.d/alsa.conf (about to try it in alsa-base.conf instead)

Also tried live boot of daily snapshot of Ubuntu 20.04 from 12 March 2020, fails to suspend and speakers not recognised (same problems). Kern.log and LEDs strongly indicate machine crashes before suspend achieved (having read up on the difference), so appears to be a suspend problem not a resume one; especially as LED goes into dim/brighten cycle on previous kernels when it goes into suspend and resumes fully with no issue.
Cancelling the suspend just after I've pressed the power button to suspend it and the screen has gone black (think I used space key? or did I press power again?) worked the one time I tried it, and dmesg/logs indicate a fair bit of the suspend happens ok.

Appears to match part of this upstream issue (about an HP Spectre x360 laptop):
https://bugzilla.kernel.org/show_bug.cgi?id=205959

May have time to bisect some kernel changes or try other fixes or build kernels/code, given a few days (am a software developer)

ProblemType: Bug
DistroRelease: Ubuntu 19.10
Package: linux-image-5.3.0-42-generic 5.3.0-42.34
ProcVersionSignature: Ubuntu 5.3.0-42.34-generic 5.3.18
Uname: Linux 5.3.0-42-generic x86_64
ApportVersion: 2.20.11-0ubuntu8.5
Architecture: amd64
AudioDevicesInUse: Error: command ['fuser', '-v', '/dev/snd/seq', '/dev/snd/timer'] failed with exit code 1:
CurrentDesktop: ubuntu:GNOME
Date: Fri Mar 13 19:12:23 2020
InstallationDate: Installed on 2019-11-08 (126 days ago)
InstallationMedia: Ubuntu 19.10 "Eoan Ermine" - Release amd64 (20191017)
MachineType: HP HP EliteBook 830 G6
ProcFB: 0 i915drmfb
ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-5.3.0-42-generic root=UUID=4eabc89a-3237-4e0b-91b2-56346c65d607 ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-5.3.0-42-generic N/A
 linux-backports-modules-5.3.0-42-generic N/A
 linux-firmware 1.183.4
SourcePackage: linux
UpgradeStatus: No upgrade log present (probably fresh install)
dmi.bios.date: 06/18/2019
dmi.bios.vendor: HP
dmi.bios.version: R70 Ver. 01.02.00
dmi.board.name: 854A
dmi.board.vendor: HP
dmi.board.version: KBC Version 52.4C.00
dmi.chassis.asset.tag: 5CG9381HTF
dmi.chassis.type: 10
dmi.chassis.vendor: HP
dmi.modalias: dmi:bvnHP:bvrR70Ver.01.02.00:bd06/18/2019:svnHP:pnHPEliteBook830G6:pvr:rvnHP:rn854A:rvrKBCVersion52.4C.00:cvnHP:ct10:cvr:
dmi.product.family: 103C_5336AN HP EliteBook
dmi.product.name: HP EliteBook 830 G6
dmi.product.sku: 6XE61EA#ABU
dmi.sys.vendor: HP

[Ubuntu Kernel Bot (ubuntu-kernel-bot)]

This change was made by a bot.

[Jouni Mettala (jouni-mettala)]

This could be duplicate of bug 1866170 . It is also suspend regression with e1000e driver. Does your laptop work with kernel from that bug?

[breezer (jgbreezer)]

Installed all -43 kernel deb packages from the link in #5 of bug 1866170 and booted with it (after allowing legacy boot, disabling UEFI); cannot notice any change. Still won't suspend, audio still not recognised. Tried a few alsa force-reloads with dmic setting/not and mode specified/not in case it helped with sound device recognition.

I saw link to e1000e driver download from intel there, but as you only mentioned to test the kernel from there I will presume not worth to download and try to build that .tar.gz .

New dmesg attached (any other logs helpful from -43 kernel boot?). Any further logging I can turn on?

[breezer (jgbreezer)]

Thought the kern.log from the (tested-above) kernel might be rather useful too; attached here the last boot (ie. where boot-offset time in the log becomes [0.00000] again).

[roland (roland-micite)]

The sound problem might be related with: https://bugs.launchpad.net/ubuntu/eoan/+source/linux/+bug/1864061

[Jouni Mettala (jouni-mettala)]

This could give more detailed information:
https://wiki.ubuntu.com/DebuggingKernelSuspend#Debugging_information_to_provide_in_your_bug_report

It includes testing latest mainline kernel.
https://kernel.ubuntu.com/~kernel-ppa/mainline/?C=N;O=D

[breezer (jgbreezer)]

Actually, update: suspend works with following added to /etc/modprobe.d files on the new kernel (-43 with this patch, as per #5 above), but audio still does not (didn't test JUST the blacklisted module yet):

line added to blacklist.conf:
blacklist snd_soc_skl

line from alsa-base.conf:
options snd-hda-intel dmic_detect=0 model=auto

[Jouni Mettala (jouni-mettala)]

Does suspend work with latest mainline kernel without extra blacklist lines?

https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.6-rc7/

[breezer (jgbreezer)]

Yes, suspend still works (tried it twice times) without blacklisting the soc-skl (I did a fresh boot having removed the blacklist lines for it). Same situation as with prior -40 mainline (builtin speakers still work, mic still not work). Seems a fix for mic will be targetted at 20.04 from other dupe tickets, I can wait til I upgrade/for a backport.