Laptop keyboard lag when not charging

SpaceDog · 2024-12-11 01:07:53

seth wrote:

Dec 08 11:25:19 SpaceDog kernel: i8042: Warning: Keylock active
Try to add " i8042.reset i8042.unlock=1" to the https://wiki.archlinux.org/title/Kernel_parameters

Damn I hit "report" instead of "quote" again! Sorry!

That change got rid of the keylock message on boot but didn't fix the problem with the keyboard.

I will do more testing to see if booting without a power connection is the only trigger for the problem.

Thanks a lot for all your help on this! I will be a bit slow to work on it for the next week or two as I'm moving house by car from the West coast of Australia to the East coast.

ewaller · 2024-12-11 03:39:48

SpaceDog wrote:

Damn I hit "report" instead of "quote" again! Sorry!

Yeah, I saw that. It happens more than you might guess. No worries.

Last edited by ewaller (2024-12-12 02:09:40)

seth · 2024-12-11 08:02:41

https://bbs.archlinux.org/viewtopic.php?id=279719
See the .postreport link css - might prevent that

Next shot:

i8042.noaux=1

or toggle the touchpad (there's likely an Fn key for that)

SpaceDog · 2024-12-22 14:21:31

seth wrote:

https://bbs.archlinux.org/viewtopic.php?id=279719
See the .postreport link css - might prevent that
Next shot:
i8042.noaux=1
or toggle the touchpad (there's likely an Fn key for that)

I made it across the country. This kernel parameter didn't work.
I also tried updating the BIOS with the latest firmware from Dell, and that didn't work.

But I think I found the solution.
I just did it and my keyboard is working properly now, even though I booted the computer unplugged and kept it unplugged.

Kernel setting:

acpi_osi="Windows 2013"

I think this did it, because usually the keyboard would be laggy if I booted the computer while it was unplugged. I'll report back if the laggy keyboard problem returns.

Last edited by SpaceDog (2024-12-22 14:31:41)

seth · 2024-12-22 14:25:08

That's not a valid parameter (typo?),

acpi_osi=! acpi_osi="Windows 2013"

might actually do something.
https://learn.microsoft.com/en-us/windo … inacpi-osi

SpaceDog · 2024-12-22 14:32:02

Sorry, that's right, it was just a typo in my post. I've edited it now.

Marking this thread as [SOLVED] as long as the problem doesn't come back.
Thanks for all your help!

Last edited by SpaceDog (2024-12-22 14:40:56)

SpaceDog · 2024-12-23 00:18:26

I spoke to soon, the problem came back.
I'll look more into ACPI as a potential part of the problem.

seth · 2024-12-23 09:14:03

For clarification: when the keyboard is slow, plugging the charger *immediately* remedies that?
Compare

time echo "scale=5000; a(1)*4" | bc -l

before and after plugging the charger to "fix" this (not the result - it's gonna be the same and should be familiar - but the time the calculation takes)

Sanity check: is there a parallel windows installation?

SpaceDog · 2024-12-24 02:54:00

seth wrote:

For clarification: when the keyboard is slow, plugging the charger *immediately* remedies that?
Compare
time echo "scale=5000; a(1)*4" | bc -l
before and after plugging the charger to "fix" this (not the result - it's gonna be the same and should be familiar - but the time the calculation takes)
Sanity check: is there a parallel windows installation?

There's no parallel windows installation but the computer came bundled with windows, which has been deleted.
I noticed in journalctl that the UEFI loads some Windows certification keys.

The laggy keyboard is fixed immediately when the laptop is plugged in.

Here's the output of that command, minus the wall of numbers:

Unplugged, laggy keyboard:

echo "scale=5000; a(1)*4"  0.00s user 0.00s system 48% cpu 0.002 total
bc -ql -l  23.13s user 0.01s system 99% cpu 23.157 total

Plugged in a few seconds later:

echo "scale=5000; a(1)*4"  0.00s user 0.00s system 44% cpu 0.001 total
bc -ql -l  14.94s user 0.00s system 99% cpu 14.956 total

I don't know if it's related, but in my journalctl this error appears each time I plug in the AC adapter:

Dec 24 13:49:03 SpaceDog root[16177]: ACPI action undefined: ACPI0003:00
Dec 24 13:49:04 SpaceDog kernel: pcieport 0000:00:1c.0: AER: Corrected error message received from 0000:00:1c.0
Dec 24 13:49:04 SpaceDog kernel: pcieport 0000:00:1c.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)
Dec 24 13:49:04 SpaceDog kernel: pcieport 0000:00:1c.0:   device [8086:9d10] error status/mask=00000001/00002000
Dec 24 13:49:04 SpaceDog kernel: pcieport 0000:00:1c.0:    [ 0] RxErr                  (First)
Dec 24 13:49:07 SpaceDog root[16360]: ACPI action undefined: PNP0C0A:00

The first and last lines are the ACPI action of plugging in the AC, and then the battery state changing.
There's no error when I unplug it, just a bunch of undefined ACPI actions regarding the AC power, battery, and processors:

Dec 24 13:51:34 SpaceDog root[17500]: ACPI action undefined: ACPI0003:00
Dec 24 13:51:34 SpaceDog root[17509]: ACPI action undefined: PNP0C0A:00
Dec 24 13:51:34 SpaceDog root[17511]: ACPI group/action undefined: processor / LNXCPU:00
Dec 24 13:51:34 SpaceDog root[17514]: ACPI group/action undefined: processor / LNXCPU:01
Dec 24 13:51:34 SpaceDog root[17517]: ACPI group/action undefined: processor / LNXCPU:02
Dec 24 13:51:34 SpaceDog root[17519]: ACPI group/action undefined: processor / LNXCPU:03
Dec 24 13:51:34 SpaceDog root[17525]: ACPI group/action undefined: processor / LNXCPU:04
Dec 24 13:51:34 SpaceDog root[17533]: ACPI group/action undefined: processor / LNXCPU:05
Dec 24 13:51:34 SpaceDog root[17539]: ACPI group/action undefined: processor / LNXCPU:06
Dec 24 13:51:34 SpaceDog root[17543]: ACPI group/action undefined: processor / LNXCPU:07

But as I said, the problem doesn't return immediately when I unplug the laptop. The only reliable way to trigger it is to shut down and then start up the laptop unplugged.

seth · 2024-12-24 10:01:28

bc -ql -l  23.13s user 0.01s system 99% cpu 23.157 total
bc -ql -l  14.94s user 0.00s system 99% cpu 14.956 total

So the cpu clearly steps down (when booted?) on battery.

watch cat /sys/devices/system/cpu/cpu[0-9]*/cpufreq/scaling_cur_freq

Also check "top" on whether the keyboard lags because the CPU is… hold on, the external keyboard wasn't affected, right?

You AC adapter shows up as iput device, does it generate noise in evtest when this happens?

Dec 08 11:25:19 SpaceDog kernel: ACPI: AC: AC Adapter [AC] (off-line)
Dec 08 11:25:19 SpaceDog kernel: input: Lid Switch as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0

SpaceDog · 2025-01-02 05:26:29

seth wrote:

bc -ql -l  23.13s user 0.01s system 99% cpu 23.157 total
bc -ql -l  14.94s user 0.00s system 99% cpu 14.956 total
So the cpu clearly steps down (when booted?) on battery.
watch cat /sys/devices/system/cpu/cpu[0-9]*/cpufreq/scaling_cur_freq
Also check "top" on whether the keyboard lags because the CPU is… hold on, the external keyboard wasn't affected, right?
You AC adapter shows up as iput device, does it generate noise in evtest when this happens?
Dec 08 11:25:19 SpaceDog kernel: ACPI: AC: AC Adapter [AC] (off-line)
Dec 08 11:25:19 SpaceDog kernel: input: Lid Switch as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0

I can confirm that the same amplitude of CPU speed step-down happens when I go from plugged-in to unplugged, without the keyboard being effected, so I don't think it's a CPU issue.

The AC adapter doesn't make any noise in evtest, and I think it doesn't really show up as an input device - evtest has the lid switch as input event 0 and the full log makes it look like it's the lid switch that shows up as an input device, not the AC adapter:

Dec 08 11:25:19 SpaceDog kernel: ACPI: AC: AC Adapter [AC] (off-line)
Dec 08 11:25:19 SpaceDog kernel: input: Lid Switch as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0D:00/input/input0
Dec 08 11:25:19 SpaceDog kernel: ACPI: button: Lid Switch [LID0]
Dec 08 11:25:19 SpaceDog kernel: input: Power Button as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0C0C:00/input/input1
Dec 08 11:25:19 SpaceDog kernel: ACPI: button: Power Button [PBTN]

I have found that if I start with the kernel parameter

acpi=off

, then the keyboard doesn't work at all.
This seems strange to me, shouldn't the keyboard work independently of whether acpi is enabled?

I haven't been able to test if my external keyboard is effected the same way as I'm away from home until the 4th of January.

seth · 2025-01-02 14:43:16

This seems strange to me, shouldn't the keyboard work independently of whether acpi is enabled?

Nope.

Dec 08 11:25:19 SpaceDog kernel: ACPI: LAPIC_NMI (acpi_id[0x03] high edge l8042int[0x1])
Dec 08 11:25:19 SpaceDog kernel: i8042: PNP: PS/2 Controller [PNP0303:PS2K,PNP0f13:PS2M] at 0x60,0x64 irq 1,12

Dec 08 11:25:19 SpaceDog kernel: ACPI: [Firmware Bug]: BIOS _OSI(Linux) query ignored

Let's just lie.

acpi_osi=! acpi_osi="Windows 2015"

https://learn.microsoft.com/en-us/windo … inacpi-osi
https://wiki.archlinux.org/title/Kernel_parameters

Arch Linux

#26 2024-12-11 01:07:53

Re: Laptop keyboard lag when not charging

#27 2024-12-11 03:39:48

Re: Laptop keyboard lag when not charging

#28 2024-12-11 08:02:41

Re: Laptop keyboard lag when not charging

#29 2024-12-22 14:21:31

Re: Laptop keyboard lag when not charging

#30 2024-12-22 14:25:08

Re: Laptop keyboard lag when not charging

#31 2024-12-22 14:32:02

Re: Laptop keyboard lag when not charging

#32 2024-12-23 00:18:26

Re: Laptop keyboard lag when not charging

#33 2024-12-23 09:14:03

Re: Laptop keyboard lag when not charging

#34 2024-12-24 02:54:00

Re: Laptop keyboard lag when not charging

#35 2024-12-24 10:01:28

Re: Laptop keyboard lag when not charging

#36 2025-01-02 05:26:29

Re: Laptop keyboard lag when not charging

#37 2025-01-02 14:43:16

Re: Laptop keyboard lag when not charging

Board footer