[Lenovo X1 Carbon 9th] [SYNA8009:00 06CB:CE57] touchpad lagging after resume from suspend

[Alexander Karlstad]

Hardware: Thinkpad X1 Carbon, 9th generation
Touchpad: SYNA8009:00 06CB:CE57

Right after a clean reboot, the touchpad works fine. No issues whatsoever. But after resuming from even its first suspend after a clean boot, the touchpad feels less responsive, choppy or even laggy.

I have tried removing and re-inserting some HID kernel modules, "psmouse", and setting i915.psr_enabled=0 in kernelstubs. Haven't found anything that fixes it yet. The trackpoint works flawlessly all the time, so that is the current backup when I don't have an external mouse connected.

My installed OS is Pop!_Os 21.04, but I have tested and confirmed this on a live USB with Ubuntu 21.04 as well.

I have also tried upgrading to the latest 5.13.5 and 5.14.0rc2 kernels from the mainline PPA, and tried bumping the libinput10 to 1.18.0 from the impish repos.

To recreate:
1. Reboot a X1C9 machine and see that the touchpad works fine
2. Suspend it (to RAM, need to change a BiOS setting so it doesn't do a software suspend)
3. Resume from suspend and see that touchpad acts differently and less responsive than right after a clean boot

What is expected:
1. Same responsive touchpad experience after resuming from suspend as right after a clean boot.

---

ProblemType: Bug
ApportVersion: 2.20.11-0ubuntu65.1
Architecture: amd64
BootLog: Error: [Errno 13] Permission denied: '/var/log/boot.log'
CasperMD5CheckResult: unknown
CompizPlugins: No value set for `/apps/compiz-1/general/screen0/options/active_plugins'
CompositorRunning: None
CurrentDesktop: pop:GNOME
DistUpgraded: Fresh install
DistroCodename: hirsute
DistroRelease: Pop!_OS 21.04
DistroVariant: ubuntu
GraphicsCard:
 Intel Corporation TigerLake GT2 [Iris Xe Graphics] [8086:9a49] (rev 01) (prog-if 00 [VGA controller])
   Subsystem: Lenovo Iris Xe Graphics [17aa:22d5]
MachineType: LENOVO 20XW005PMX
Package: libinput10 1.18.0-1 [origin: unknown]
PackageArchitecture: amd64
ProcKernelCmdLine: initrd=\EFI\Pop_OS-e5934199-eb91-4da9-9e38-0ad1de1baeb2\initrd.img root=UUID=e5934199-eb91-4da9-9e38-0ad1de1baeb2 ro quiet systemd.show_status=false loglevel=0 splash
Tags: hirsute third-party-packages ubuntu
Uname: Linux 5.13.5-051305-generic x86_64
UnreportableReason: This does not seem to be an official Pop!_OS package. Please retry after updating the indexes of available packages, if that does not work then remove related third party packages and try again.
UpgradeStatus: No upgrade log present (probably fresh install)
UserGroups: adm lpadmin plugdev sudo
_MarkForUpload: True
dmi.bios.date: 06/15/2021
dmi.bios.release: 1.42
dmi.bios.vendor: LENOVO
dmi.bios.version: N32ET66W (1.42 )
dmi.board.asset.tag: Not Available
dmi.board.name: 20XW005PMX
dmi.board.vendor: LENOVO
dmi.board.version: SDK0J40697 WIN
dmi.chassis.asset.tag: No Asset Information
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: None
dmi.ec.firmware.release: 1.26
dmi.modalias: dmi:bvnLENOVO:bvrN32ET66W(1.42):bd06/15/2021:br1.42:efr1.26:svnLENOVO:pn20XW005PMX:pvrThinkPadX1CarbonGen9:skuLENOVO_MT_20XW_BU_Think_FM_ThinkPadX1CarbonGen9:rvnLENOVO:rn20XW005PMX:rvrSDK0J40697WIN:cvnLENOVO:ct10:cvrNone:
dmi.product.family: ThinkPad X1 Carbon Gen 9
dmi.product.name: 20XW005PMX
dmi.product.sku: LENOVO_MT_20XW_BU_Think_FM_ThinkPad X1 Carbon Gen 9
dmi.product.version: ThinkPad X1 Carbon Gen 9
dmi.sys.vendor: LENOVO
version.compiz: compiz N/A
version.libdrm2: libdrm2 2.4.105-3~21.04.1
version.libgl1-mesa-dri: libgl1-mesa-dri 21.0.1-2pop0~1624392450~21.04~6fbdfd2
version.libgl1-mesa-glx: libgl1-mesa-glx N/A
version.xserver-xorg-core: xserver-xorg-core 2:1.20.11-1ubuntu1.1
version.xserver-xorg-input-evdev: xserver-xorg-input-evdev N/A
version.xserver-xorg-video-ati: xserver-xorg-video-ati 1:19.1.0-2
version.xserver-xorg-video-intel: xserver-xorg-video-intel 2:2.99.917+git20200714-1ubuntu1
version.xserver-xorg-video-nouveau: xserver-xorg-video-nouveau 1:1.0.17-1

[Alexander Karlstad (alexander.karlstad)]

The journal has a lot of these errors after resuming:

juli 27 00:17:41 machine touchegg[1003]: libinput error: event11 - SYNA8009:00 06CB:CE57 Touchpad: kernel bug: Touch jump detected and discarded.

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Alexander Karlstad (alexander.karlstad)]

apport information

[Daniel van Vugt (vanvugt)]

This is probably a firmware or kernel bug but AFAICT you're not using any kernel versions supplied by Ubuntu. So we can't support the kernels you're using here.

If you think this might be a libinput bug then you should report it to Libinput at: https://gitlab.freedesktop.org/libinput/libinput/-/issues

[Alexander Karlstad (alexander.karlstad)]

I am currently using the 5.13.5 kernel from the Ubuntu kernel team here: https://kernel.ubuntu.com/~kernel-ppa/mainline/

I have also tested the 5.14.0rc2 kernel from the same repo. Can give rc3 a try too.

[Alexander Karlstad (alexander.karlstad)]

Tested 5.14rc3 + 5.13.6 from https://kernel.ubuntu.com/~kernel-ppa/mainline now, and the issue occurs in those as well.

[Alexander Karlstad (alexander.karlstad)]

Update: tested new kernels 5.13.7 and 5.14rc4 now and issue still persists in those kernels as well.

[Alexander Karlstad (alexander.karlstad)]

Further investigation reveals that this only occurs when S3 suspend is enabled in BiOS. The default that is shipped is identified as s2idle in systemd (just called "Windows 10 and Linux" in BiOS[1]) and is a sort of software suspend that does not preserve much battery time (which is also why I switched to S3). S3 is suspend to RAM I think.

I have upgraded to a new BiOS release that was made available today (N32ET68W 1.44), and an update to the Embed Controller. This did not help with the issue unfortunately.

Seeing as the touchpad works fine with the software suspend, but not the S3 RAM suspend, is there perhaps some module that is not handled correctly when waking up from a more thorough suspend?

[1] https://i.imgur.com/7ddEvpW.jpg

[Kai-Heng Feng (kaihengfeng)]

Can you please run `watch -n0 cat /proc/interrupts` and check whether "i2c_designware" or "SYNA8009:00" number is increasing without touching the touchpad?

[Mark Pearson (markrhpearson)]

We have an internal bug tracking this and were working with the touchpad vendor but strong recommendation is to use Modern Standby/S0ix suspend support.

Are you sure you're seeing worse battery time with S0ix? Recommendation from Intel is that it is the same as or better than S3 and on these Tigerlake CPU's is what Intel are recommending for use.

Mark

[Alexander Karlstad (alexander.karlstad)]

> Can you please run `watch -n0 cat /proc/interrupts` and check whether "i2c_designware" or "SYNA8009:00" number is increasing without touching the touchpad?

Tried this now, and yes, both of these ("idma64.0, i2c_designware.0" and "SYNA8009:00") are increasing while not using the touchpad at all after resuming from suspend. Right after a reboot these values remain stable and do not increase while not using it.

> We have an internal bug tracking this and were working with the touchpad vendor but strong recommendation is to use Modern Standby/S0ix suspend support.
>
> Are you sure you're seeing worse battery time with S0ix? Recommendation from Intel is that it is the same as or better than S3 and on these Tigerlake CPU's is what Intel are recommending for use.

Awesome!

I only gave the modern standby/sleep a couple of chances in the beginning, and at that point the laptop battery was drained by the next morning if I closed the lid on it after work (around 5-6 pm the evening before). Bear in mind that that was with the stock kernel that came with Pop 21.04, which is 5.11.0. I can give it a try again now that I am running 5.13.9, if a lot has changed regarding that since then.

[Kai-Heng Feng (kaihengfeng)]

Can you please give this kernel a try:
https://people.canonical.com/~khfeng/lp1938096/

[Alexander Karlstad (alexander.karlstad)]

Tested the kernel provided now and it seems to still act the same. Attaching GIF with output of the `watch -n0 cat /proc/interrupts` command.

I also tried switching the sleep mode back to the default (S0ix) yesterday and this seems to work a lot better with 5.13 kernels than the 5.11 one and preserves battery very well now. I had it set back to "Linux S3" again for this kernel test though.

`uname -a` output: Linux lomez 5.13.0-2010-oem #11 SMP Tue Aug 10 13:37:12 CST 2021 x86_64 x86_64 x86_64 GNU/Linux

[Kai-Heng Feng (kaihengfeng)]

Thanks for the testing. The kernel power cycle the ACPI part of the touchpad, but it doesn't work :(

I guess it's better to stick to s2idle for now.

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Koen van Greevenbroek (koenvg)]

I can confirm this exact bug. I have the same machine (Thinkpad X1 Carbon 9th gen). However, I am running an up-to-date Fedora 34 system. For me, changing the sleep mode to what the BIOS calls "Linux and Windows" (apparently software suspend?) also fixes the problem. I have not had a chance yet to compare the impact on battery life during suspend.

Would it be an idea to report this to Fedora as well?

I might be able to help test things out if needed.

[Vasili Burdo (vburdo)]

I also confirm this bug - Lenovo Thinkpad P1 Gen2.
uname: Linux zzzzz 5.11.0-27-generic #29~20.04.1-Ubuntu SMP Wed Aug 11 15:58:17 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux
touchpad: Bus 001 Device 003: ID 06cb:00bd Synaptics, Inc.
Nothing alarming in logs.

The touchpad affected by this bug feels like "lagging".
Sweep and tap actions (without hard touchpad pressing) work after multiple tries.
Also, I noticed "quick" suspend and resume does not cause the bug.
It's necessary to let laptop stay suspended quite of time - in my case it was couple of hours - I didnd't try to narrow it.

UPDATE: Funny thing, hybernating to disk and waking up restores touchpad funtionality.

[joe williams (joetify)]

I am experiencing the same problem. The trackpad seems laggy after suspend. Also, it seems to commonly get "stuck" in two finger scrolling, where after lifting one of the two fingers it stays in scrolling mode. The only way to get it out is to tap on the trackpad randomly to, I assume, get it to detect different finger actions.

Linux xxxxx 5.11.0-27-generic #29~20.04.1-Ubuntu SMP Wed Aug 11 15:58:17 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Ubuntu 20.04.3 LTS \n \l

[Mark Pearson (markrhpearson)]

Just to pick up on some of the above comments:

 - my understanding is this is a touchpad firmware bug so not much point reporting it to other distro's as well - it's not a Canonical/Ubuntu issue either IMO.

 - Using S0ix/modern suspend (or 'Windows and Linux') sleep mode setting in the BIOS is the fix. If you see other problems around S0ix they should be reported so we can fix those (you shouldn't see excessive battery drain). That's the supported/tested/certified sleep mode on these Tigerlake CPUs.

We are working with Synaptics on a S3 fix - but it's been lower priority I'm afraid. I don't have an update or ETA yet.

Mark

[Tim Richardson (tim-richardson)]

I have this laptop. Bug happens on S3 suspend, much better on modern standby (once it took two minutes or so for the touchpad to resume, every other time it has worked).

Modern suspend does not use much battery. Mine slept for 8:20h last night and the battery percentage declined from 96% to 94%.
my kernel is 5.13 and the session is Wayland gnome.

[CLEsteamer (clesteamer)]

Can confirm same issue here when set to S3(Linux only). Same problem also happens in windows - so this is a hardware issue.

I had strange issues when using a non-stock charger as well (lagged but in a different way).

[Andy Barry (abarry-gmail)]

I am experiencing the exact same symptoms on a Razer Blade 15 (early 2022) model running 5.17.1 from Ubuntu mainline:

$ uname -r
5.17.1-051701-generic

The issue only happens after S3 suspend. s2idle does not cause the problem, but I experience huge battery drain in s2idle (10%/hour).

[Vasili Burdo (vburdo)]

Finally I found the fix of this problem (thans archlinux wiki): just reload "psmouse" module.
Like this:
    sudo sh -c 'rmmod psmouse && modprobe psmouse'

Most often touchpad is back to normal after first reload. Sometimes it does not help and I do 2-3 reloads and it helps.

[rdz (reduzierer)]

This same problem affects me.

Ubuntu 22.04
Thinkpad T14s Gen 3
Linux: 6.1.0-1006-oem

When 'Linux S3' is configured in UEFI BIOS as suspend mode, touchpad is laggy after resume and one of the CPU cores shows high load.

When setting 'Windows and Linux' as suspend mode in the UEFI BIOS (which apparently is s2idle), touchpad works normally after resume from suspend. However, the power drain during s2idle is significant. Battery level drops by more than 30% when idling overnight.

The trick with removing/inserting the module psmouse doesn't help here. When I rmmod i2c_hid_acpi, the high CPU load disappears, but comes back when loading i2c_hid_acpi again.