Inspiron 5565 suspend issues after 17.10 upgrade

This change was made by a bot.

Would it be possible for you to test the latest upstream kernel? Refer to https://wiki.ubuntu.com/KernelMainlineBuilds . Please test the latest v4.14 kernel[0].

If this bug is fixed in the mainline kernel, please add the following tag 'kernel-fixed-upstream'.

If the mainline kernel does not fix this bug, please add the tag: 'kernel-bug-exists-upstream'.

Once testing of the upstream kernel is complete, please mark this bug as "Confirmed".

Thanks in advance.

[0] http://kernel.ubuntu.com/~kernel-ppa/mainline/v4.14-rc6

4.14-rc6 is even worse -- I can't wake the machine after suspend at all, except by rebooting with long-press of the power button.

I tested other mainline kernels too. All 4.13-based mainline kernels I tested behave similar to the shipped Ubuntu kernel, from 4.13.4 down back to 4.13rc1. (incomplete suspend, on resume wifi doesn't connect, turn off wifi button does nothing, reboot only hard way)

The mainline 4.12.14 kernel worked fine for me; suspend/resume were OK.

Tested a couple more kernels and it seems that the issue I have with 4.13.4 is not present in mainline 4.13.5, nor in 4.13.9. Resume worked fine.

Tested with newest mainline kernels. 4.3.11 still works fine. 4.14.rc7 still fails to come out of suspend.

sorry, meant to say 4.13.11, not 4.3.11

Does v4.13 (not v4.13.11) have this issue?

To summarize my earlier comments: All the below relates to suspend/resume of 17.10.

Suspend/resume works fine if I boot with the 4.10-based Ubuntu kernel from 17.04, or any 4.13 mainline kernel starting from 4.13.5 .

The mainline kernels from 4.13 through 4.13.4, upon which the Ubuntu kernel is built, all suspend OK but have problems on resume-- the machine comes up but wifi does not connect. Once that happens I cannot reboot or shut down normally, only by holding the power button.

The mainline kernels from 4.13.5 to 4.13.11 resume OK.

The mainline kernels from 4.14.rc1 to 4.14.rc7 have issues on suspend and cannot properly resume -- display does not come on and the machine does not respond to ssh from elsewhere. Again I am forced to hold the power button down for hard shutdown.

It seems to be two separate issues,

> On 7 Nov 2017, at 10:03 PM, Aaron Peromsik <email address hidden> wrote:
>
> To summarize my earlier comments: All the below relates to
> suspend/resume of 17.10.
>
> Suspend/resume works fine if I boot with the 4.10-based Ubuntu kernel
> from 17.04, or any 4.13 mainline kernel starting from 4.13.5 .
>
> The mainline kernels from 4.13 through 4.13.4, upon which the Ubuntu
> kernel is built, all suspend OK but have problems on resume-- the
> machine comes up but wifi does not connect. Once that happens I cannot
> reboot or shut down normally, only by holding the power button.
>
> The mainline kernels from 4.13.5 to 4.13.11 resume OK.

This means some commits from v4.13.4 to v4.13.5 fixes your resume problem.

The good news is that v4.13.5 based Ubuntu kernel is landing soon.

>
> The mainline kernels from 4.14.rc1 to 4.14.rc7 have issues on suspend
> and cannot properly resume -- display does not come on and the machine
> does not respond to ssh from elsewhere. Again I am forced to hold the
> power button down for hard shutdown.

This means some commits from v4.13 to v4.14-rc1 cause the suspend issue.
You probably will hit this issue when you upgrade the system to 18.04, which uses 4.15 based kernel.

Do you know how to do kernel bisect? I can build kernel for you if you don’t know how.

>
> --
> You received this bug notification because you are subscribed to linux
> in Ubuntu.
> https://bugs.launchpad.net/bugs/1726061
>
> Title:
> Inspiron 5565 suspend issues after 17.10 upgrade
>
> Status in linux package in Ubuntu:
> Confirmed
>
> Bug description:
> On my Inspiron 5565, 17.10 seems to work fairly well, unless I try to
> suspend. Suspend shuts off the display but doesn't seem to actually
> shut off the CPU -- fans still run. When I move the mouse the display
> comes back on but upon entering my password sometimes the session
> crashes and I am dumped back to the login screen; sometimes the
> session is still there but wifi is either disabled (all wifi-related
> menu items gone) or visible (with question mark) but non-functional.
> In any case once I have suspended I cannot shut down or reboot any way
> other than the hard way, by holding the power button for several
> seconds. Doesn't seem to matter whether I am using X or Wayland.
>
> Suspend works mostly fine if I choose the 4.10 kernel from GRUB...
> although I am prompted for login more times than necessary on resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 17.10
> Package: linux-image-4.13.0-16-generic 4.13.0-16.19
> ProcVersionSignature: Ubuntu 4.10.0-37.41-generic 4.10.17
> Uname: Linux 4.10.0-37-generic x86_64
> ApportVersion: 2.20.7-0ubuntu3
> Architecture: amd64
> AudioDevicesInUse:
> USER PID ACCESS COMMAND
> /dev/snd/controlC0: aaron 2489 F.... pulseaudio
> /dev/snd/controlC1: aaron 2489 F.... pulseaudio
> CurrentDesktop: ubuntu:GNOME
> Date: Sun Oct 22 12:22:13 2017
> HibernationDevice: RESUME=UUID=3cc55560-1886-446c-8d60-15136ee14d32
> InstallationDate: Installed on 2016-12-25 (301 days ago)
> InstallationMedia: Ubuntu 16.10 "Yakkety Yak" - Release amd64 (20161012....

Still a few rounds of bisect remaining but this is where I am now:

# bad: [126f760ca94dae77425695f9f9238b731de86e32] rds: Fix incorrect statistics counting
git bisect bad 126f760ca94dae77425695f9f9238b731de86e32
# good: [98e93e968e4947cd71c2eb69e323682daa453ee7] mac80211: Complete ampdu work schedule during session tear down
git bisect good 98e93e968e4947cd71c2eb69e323682daa453ee7
# good: [8e0deed92406d93ae0365cb8a6134db5721e7aca] tipc: remove unnecessary call to dev_net()
git bisect good 8e0deed92406d93ae0365cb8a6134db5721e7aca

But that bisection may be leading toward a dead end or at least a moot point. The behavior which I am calling "bad" there is different from what I am currently seeing in later RCs and in 4.14.0.

In, for example, my last "bad" commit 126f760ca94, upon suspend the machine does actually suspend but can't wake.

In 4.14rc6, 4.14rc8, and 4.14, now upon suspend I am seeing the display power off and everything spins down but only for a moment -- then it immediately resumes. wifi and graphics seem to work OK at that point. The main remaining problem is just failure to remain suspended.

That is true at least upon first suspend. In one of those builds I did see a case where a second suspend would spin down, try to spin up, but although the display backlight did come back on the machine remained unresponsive.

I am not sure why the later RCs and in particular rc6 switched to this new failure mode. In my earlier notes it's clear that they used to be failing more like the way 126f760ca94 still behaves.

Kai-Heng

> On 13 Nov 2017, at 10:23 AM, Aaron Peromsik <email address hidden> wrote:
>
> But that bisection may be leading toward a dead end or at least a moot
> point. The behavior which I am calling "bad" there is different from
> what I am currently seeing in later RCs and in 4.14.0.
>
> In, for example, my last "bad" commit 126f760ca94, upon suspend the
> machine does actually suspend but can't wake.
>
> In 4.14rc6, 4.14rc8, and 4.14, now upon suspend I am seeing the display
> power off and everything spins down but only for a moment -- then it
> immediately resumes. wifi and graphics seem to work OK at that point.
> The main remaining problem is just failure to remain suspended.

Can you attach `cat /sys/kernel/debug/wakeup_sources` after the immediate resume?

>
> That is true at least upon first suspend. In one of those builds I did
> see a case where a second suspend would spin down, try to spin up, but
> although the display backlight did come back on the machine remained
> unresponsive.
>
> I am not sure why the later RCs and in particular rc6 switched to this
> new failure mode. In my earlier notes it's clear that they used to be
> failing more like the way 126f760ca94 still behaves.
>
> --
> You received this bug notification because you are subscribed to linux
> in Ubuntu.
> https://bugs.launchpad.net/bugs/1726061
>
> Title:
> Inspiron 5565 suspend issues after 17.10 upgrade
>
> Status in linux package in Ubuntu:
> Confirmed
>
> Bug description:
> On my Inspiron 5565, 17.10 seems to work fairly well, unless I try to
> suspend. Suspend shuts off the display but doesn't seem to actually
> shut off the CPU -- fans still run. When I move the mouse the display
> comes back on but upon entering my password sometimes the session
> crashes and I am dumped back to the login screen; sometimes the
> session is still there but wifi is either disabled (all wifi-related
> menu items gone) or visible (with question mark) but non-functional.
> In any case once I have suspended I cannot shut down or reboot any way
> other than the hard way, by holding the power button for several
> seconds. Doesn't seem to matter whether I am using X or Wayland.
>
> Suspend works mostly fine if I choose the 4.10 kernel from GRUB...
> although I am prompted for login more times than necessary on resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 17.10
> Package: linux-image-4.13.0-16-generic 4.13.0-16.19
> ProcVersionSignature: Ubuntu 4.10.0-37.41-generic 4.10.17
> Uname: Linux 4.10.0-37-generic x86_64
> ApportVersion: 2.20.7-0ubuntu3
> Architecture: amd64
> AudioDevicesInUse:
> USER PID ACCESS COMMAND
> /dev/snd/controlC0: aaron 2489 F.... pulseaudio
> /dev/snd/controlC1: aaron 2489 F.... pulseaudio
> CurrentDesktop: ubuntu:GNOME
> Date: Sun Oct 22 12:22:13 2017
> HibernationDevice: RESUME=UUID=3cc55560-1886-446c-8d60-15136ee14d32
> InstallationDate: Installed on 2016-12-25 (301 days ago)
> InstallationMedia: Ubuntu 16.10 "Yakkety Yak" - Release amd64 (20161012.2)
> MachineType: Dell Inc. Inspiron 5565
> ProcFB: 0 amdgpudrmfb
> ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.10.0-37-generi...

So it may be PNP0C0D:00 (lid switch), LNXPWRBN:00 (power button), or ACAD wakes your system up.

Can you attach output of `cat /proc/acpi/wakeup` here?

Upon initially booting either 4.13.11 or 4.14.0 I see 1 in active_count for ACAD, and zeros elsewhere. When 4.14.0 has immediately resumed from suspend and shows additional events from the lid and the power button, I have not touched either of those in the course of attempting to suspend.

Attached /proc/acpi/wakeup from 4.14.0 ; in 4.13.11 it seems identical.

4.13.11 has been suspending OK but today it is not -- instead it behaves exactly like 4.14.0 .

Perhaps this is no longer a kernel issue or at least no longer a regression. I am getting the same behavior with 4.14 and 4.13.11 and even the kernel from 17.04 . Normally when I am suspending at the end of the day I first close the laptop lid, then choose suspend from the menu on the external monitor... and that seems to work. Not sure how long the suspend menu has not worked with the laptop lid left open but today I still get it if I boot 17.04 from a different partition.

> On 14 Nov 2017, at 10:06 PM, Aaron Peromsik <email address hidden> wrote:
>
> Perhaps this is no longer a kernel issue or at least no longer a
> regression. I am getting the same behavior with 4.14 and 4.13.11 and
> even the kernel from 17.04 . Normally when I am suspending at the end of
> the day I first close the laptop lid, then choose suspend from the menu
> on the external monitor... and that seems to work. Not sure how long the
> suspend menu has not worked with the laptop lid left open but today I
> still get it if I boot 17.04 from a different partition.

Can you provide output of `journalctl -b` when this issue happens?

>
> --
> You received this bug notification because you are subscribed to linux
> in Ubuntu.
> https://bugs.launchpad.net/bugs/1726061
>
> Title:
> Inspiron 5565 suspend issues after 17.10 upgrade
>
> Status in linux package in Ubuntu:
> Confirmed
>
> Bug description:
> On my Inspiron 5565, 17.10 seems to work fairly well, unless I try to
> suspend. Suspend shuts off the display but doesn't seem to actually
> shut off the CPU -- fans still run. When I move the mouse the display
> comes back on but upon entering my password sometimes the session
> crashes and I am dumped back to the login screen; sometimes the
> session is still there but wifi is either disabled (all wifi-related
> menu items gone) or visible (with question mark) but non-functional.
> In any case once I have suspended I cannot shut down or reboot any way
> other than the hard way, by holding the power button for several
> seconds. Doesn't seem to matter whether I am using X or Wayland.
>
> Suspend works mostly fine if I choose the 4.10 kernel from GRUB...
> although I am prompted for login more times than necessary on resume.
>
> ProblemType: Bug
> DistroRelease: Ubuntu 17.10
> Package: linux-image-4.13.0-16-generic 4.13.0-16.19
> ProcVersionSignature: Ubuntu 4.10.0-37.41-generic 4.10.17
> Uname: Linux 4.10.0-37-generic x86_64
> ApportVersion: 2.20.7-0ubuntu3
> Architecture: amd64
> AudioDevicesInUse:
> USER PID ACCESS COMMAND
> /dev/snd/controlC0: aaron 2489 F.... pulseaudio
> /dev/snd/controlC1: aaron 2489 F.... pulseaudio
> CurrentDesktop: ubuntu:GNOME
> Date: Sun Oct 22 12:22:13 2017
> HibernationDevice: RESUME=UUID=3cc55560-1886-446c-8d60-15136ee14d32
> InstallationDate: Installed on 2016-12-25 (301 days ago)
> InstallationMedia: Ubuntu 16.10 "Yakkety Yak" - Release amd64 (20161012.2)
> MachineType: Dell Inc. Inspiron 5565
> ProcFB: 0 amdgpudrmfb
> ProcKernelCmdLine: BOOT_IMAGE=/boot/vmlinuz-4.10.0-37-generic.efi.signed root=UUID=c30eb880-24e4-41af-9203-feff0ab69c8c ro quiet splash vt.handoff=7
> RelatedPackageVersions:
> linux-restricted-modules-4.10.0-37-generic N/A
> linux-backports-modules-4.10.0-37-generic N/A
> linux-firmware 1.169
> SourcePackage: linux
> UpgradeStatus: Upgraded to artful on 2017-10-22 (0 days ago)
> dmi.bios.date: 07/20/2016
> dmi.bios.vendor: Dell Inc.
> dmi.bios.version: 1.0.2
> dmi.board.name: 0RVKF9
> dmi.board.vendor: Dell Inc.
> dmi.board.version: A00
> dmi.chassis.type: 10
> dmi.chassis.vendor: Del...

I first had to resume from suspending this morning. I ran journalctl -b before and after attempting to suspend again with the lid open. The attachment includes the lines added in between.

I think it's not a userspace issue. A solution/workaround should come from kernel or BIOS/EC.

First thing first, let's confirm all sources in "wakeup" are not culprit here.

# echo EHC1 > /proc/acpi/wakeup
# echo XHC0 > /proc/acpi/wakeup
# systemctl suspend
... and see if this issue continues to happen.

Yes, it continues.

I am not sure why power button and lid switch are not listed in wakeup_source.

This seems fishy though:
Nov 14 23:35:57 ernie kernel: ACPI: button: The lid device is not compliant to SW_LID.

This requires Dell's help to investigate.

Meanwhile, since you mentioned BIOS, I checked and found that version 1.1.1 was release in August. I installed it. It didn't help.