Backspace key above numberpad spams infinite backspace on a single press

[Joe]

I'm guessing the bug is in gsd-keyboard, but I can't imagine it being from anywhere else.

I have also submitted a bug report with the Mint team, where I found the bug, and so I will copy all the extra information from that report in a follow-up comment.

1) Ubuntu 24.04 LTS
2) I can't find a package version. I get variations on "N: Unable to locate package gsd-keyboard"
3) Should behave as an ordinary key, where a single press means a single action.
4) Behaved as if the key were held down, until pressing any other key ended the non-stop backspacing.

Note that the ordinary backspace behaves as expected. More info will be added with additional comment.

ProblemType: Bug
DistroRelease: Ubuntu 24.04
Package: gnome-settings-daemon 46.0-1ubuntu1
ProcVersionSignature: Ubuntu 6.8.0-31.31-generic 6.8.1
Uname: Linux 6.8.0-31-generic x86_64
NonfreeKernelModules: zfs
ApportVersion: 2.28.1-0ubuntu2
Architecture: amd64
CasperMD5CheckResult: pass
CasperVersion: 1.498
CloudArchitecture: x86_64
CloudID: nocloud
CloudName: unknown
CloudPlatform: nocloud
CloudSubPlatform: seed-dir (/var/lib/cloud/seed/nocloud)
CurrentDesktop: ubuntu:GNOME
Date: Sat May 18 02:09:04 2024
ExecutablePath: /usr/libexec/gsd-keyboard
LiveMediaBuild: Ubuntu 24.04 LTS "Noble Numbat" - Release amd64 (20240424)
ProcEnviron:
 LANG=C.UTF-8
 PATH=(custom, no user)
 SHELL=/bin/bash
 XDG_RUNTIME_DIR=<set>
SourcePackage: gnome-settings-daemon
UpgradeStatus: No upgrade log present (probably fresh install)

[Joe (ohcomeon)]

Copied from other report

Distribution

Linux Mint 21.3 Virginia
Package version

6.0.4
Graphics hardware in use

NVIDIA GM107M [GeForce GTX 950M]
Frequency

Always
Bug description

A single press on the numberpad backspace key causes the computer to think the key is held down. The continuous backspace only stops when another key is pressed. From playing around with a CLI program called evtest, it seems the computer sets the key value to 1 on press but doesn't set it to 0 on release. For other keys, a return to 0 is the normal behavior. Evtest will repeatedly report a value of 2 if a key is pressed and held, and it does this correctly with the numberpad backspace. However, a single press on the numberpad backspace and then a single press with the regular backspace shows a value of 2 for the second key even though the first was not held.

The standard backspace key behaves as normal.
Steps to reproduce

Open a text file and populate it with text

Press the numberpad backspace

watch the cursor eat the text without your help

press any other key to stop the madness
Expected behavior

Should behave the same as the ordinary backspace
Additional information

Output from evtest on the following button combination:

Down
Backspace
Numberpad Backspace
Numberpad Backspace
Backspace
Backspace
Numberpad Backspace
Down
Numberpad Backspace
Down
Backspace
Down
Backspace

[code]

^[[BEvent: time 1715798385.192908, type 4 (EV_MSC), code 4 (MSC_SCAN), value d0
Event: time 1715798385.192908, type 1 (EV_KEY), code 108 (KEY_DOWN), value 0
Event: time 1715798385.192908, -------------- SYN_REPORT ------------

Event: time 1715798390.545673, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798390.545673, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 1
Event: time 1715798390.545673, -------------- SYN_REPORT ------------
Event: time 1715798390.752025, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798390.752025, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 0
Event: time 1715798390.752025, -------------- SYN_REPORT ------------

Event: time 1715798392.173154, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798392.173154, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 1
Event: time 1715798392.173154, -------------- SYN_REPORT ------------

Event: time 1715798393.289422, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798393.289422, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 2
Event: time 1715798393.289422, -------------- SYN_REPORT ------------

Event: time 1715798394.592006, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798394.592006, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 2
Event: time 1715798394.592006, -------------- SYN_REPORT ------------
Event: time 1715798394.722603, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798394.722603, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 0

Event: time 1715798394.722603, -------------- SYN_REPORT ------------
Event: time 1715798396.069202, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715798396.069202, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 1
Event: time 1715798396.069202, -------------- SYN_REPORT ---...

[Joe (ohcomeon)]

Copied from other report

I just tried it with Arch but I couldn't figure out how to get past the initial setup screen. The bug doesn't exist on that initial terminal screen in set-up. That prompted me to try typing on that temporary CLI login that shows up when starting Mint from a full power off and the bug does not exist there.

So recap so far:

Mint - Yes
Mint Temp CLI Login - No
Ubuntu - Yes
Arch - Maybe not?
Windows - No

[Joe (ohcomeon)]

Copied from other report

Kernel: 5.15.0-107-generic

Keyboard: 01AX120 (the integrated keyboard on the laptop)

I don't believe this is a hardware issue because the button behaves as expected in windows and in the Arch install command line interface. Plus the evtest output doesn't spit out the following over and over like it does if you hold it down:

[code]
Event: time 1715974930.271262, type 4 (EV_MSC), code 4 (MSC_SCAN), value 0e
Event: time 1715974930.271262, type 1 (EV_KEY), code 14 (KEY_BACKSPACE), value 0
Event: time 1715974930.271262, -------------- SYN_REPORT ------------
[/code]

I checked Driver Manager but for some reason I'm only seeing the GPU driver options and nothing else. However, I figured out how to get the terminal to spit out a list of all the drivers. The problem being I don't know which one is the one I'm supposed to be interested in, so I don't know which one to ask the terminal to give me details about.

Here's the list, which ones should I get more info on?

[code]

Module Size Used by
rfcomm 81920 16
ccm 20480 6
cmac 16384 3
algif_hash 16384 1
algif_skcipher 16384 1
af_alg 32768 6 algif_hash,algif_skcipher
bnep 28672 2
ip6t_REJECT 16384 1
nf_reject_ipv6 20480 1 ip6t_REJECT
xt_hl 16384 22
ip6_tables 32768 52
ip6t_rt 20480 3
ipt_REJECT 16384 1
nf_reject_ipv4 16384 1 ipt_REJECT
xt_LOG 20480 10
nf_log_syslog 20480 10
nft_limit 16384 13
xt_limit 16384 0
xt_addrtype 16384 4
xt_tcpudp 20480 36
xt_conntrack 16384 16
nf_conntrack 172032 1 xt_conntrack
nf_defrag_ipv6 24576 1 nf_conntrack
nf_defrag_ipv4 16384 1 nf_conntrack
nft_compat 20480 153
nft_counter 16384 188
nf_tables 266240 443 nft_compat,nft_counter,nft_limit
nfnetlink 20480 2 nft_compat,nf_tables
binfmt_misc 24576 1
zfs 4112384 0
zunicode 348160 1 zfs
zzstd 491520 1 zfs
zlua 163840 1 zfs
zavl 16384 1 zfs
icp 311296 1 zfs
zcommon 106496 2 zfs,icp
znvpair 98304 2 zfs,zcommon
spl 118784 6 zfs,icp,zzstd,znvpair,zcommon,zavl
snd_soc_skl 172032 0
snd_soc_hdac_hda 24576 1 snd_soc_skl
snd_hda_ext_core 32768 2 snd_soc_hdac_hda,snd_soc_skl
snd_soc_sst_ipc 20480 1 snd_soc_skl
snd_soc_sst_dsp 36864 1 snd_soc_skl
snd_soc_acpi_intel_match 61440 1 snd_soc_skl
snd_soc_acpi 16384 2 snd_soc_acpi_intel_match,snd_soc_skl
snd_soc_core 348160 2 snd_soc_hdac_hda,snd_soc_skl
snd_hda_codec_hdmi 77824 1
snd_compress 24576 1 snd_soc_core
ac97_bus 16384 1 snd_soc_core
snd_pcm_dmaengine 16384 1 snd_soc_core
snd_ctl_led 24576 0
snd_hda_codec_conexant 28672 1
intel_rapl_msr 20480 0
intel_rapl_common 40960 1 in...

[Joe (ohcomeon)]

copied from other report

Okay, I'm now on kernel: 6.5.0-35-generic

The bug persists with the same behavior and the same output from evtest when I test the keyboard.

[Joe (ohcomeon)]

From what I've read, the evtest output is telling us this is a kernel issue? Who is in charge of the kernel? I'm perfectly happy to work with them to get to the bottom of this.

[Joe (ohcomeon)]

Given the following Terminal Command
```
sudo dpkg-reconfigure keyboard-configuration
```
----------------------------------------------
The Default keyboard setting was
```
Generic 105-key PC
```
----------------------------------------------
and the new settings were
```
Do not configure keyboard; keep kernel keymap
English (US)
English (US) - English (Dvorak)
No AltGr key
No compose key
"No" to ctrl + alt + backspace shortcut
```

Following a system reboot after letting the computer finish making the changes, the bug remains. Also, there was not a specific keyboard layout for the this laptop in the presets (Lenovo Thinkpad E570)

[Joe (ohcomeon)]

Also, there was not a specific keyboard layout for the this laptop in the presets (Lenovo Thinkpad E570)

[Joe (ohcomeon)]

Okay it looks like Lenovo uses some kind of special layer for the number pad when working with windows.

https://unix.stackexchange.com/questions/494698/thinkpad-second-sticky-backspace-key

Although that still doesn't fully explain why it works properly before the kernel loads. I don't really know when the kernel loads or even what it takes over from, I'm just guessing that something else is in control for those CLI parts where there's no bug.