Random keyboard keys repeating indefinitely / keyboard dies completely

burebista · 2016-02-13 21:19:26

Hello,

I get randomly on my new laptop into the situation where a random keyboard key starts repeating indefinitely, like if it was physically stuck, but it is not. This happens especially under work load. When it happens, I cannot stop it even if I press the repeating key. The only way to stop it is to restart the laptop.

This does not happen in windows (I have dual windows 10 / arch booting on my laptop), which suggests that there is something wrong with some linux drivers and not with the BIOS.

My laptop configuration: a CLEVO P670RE6 chassis having an AMI bios (BIOS version 1.05.11TPCS1, KBC/EC Firmware Revision 1.05.06, ME FW Version 11.00.1183), Intel Core i7 Quad Core Processor 6820HK (2.7GHz) and a hybrid NVIDIA GeForce GTX 970M card. I am using the latest nvidia proprietary driver from the “nvidia” pacman package. I also tried the nouveau driver and still got the repeating problem. Also I got this problem both in cinnamon and xfce.

I am opening another thread hereby because I saw another SOLVED thread where a BIOS update solved the problem, but like I said above, this does not look like a BIOS problem, because I do not get this problem when I boot in windows 10.

I also want to mention another problem, just in case that it is related to the above one, or perhaps even caused by the above one or having the same cause as the above one: in addition to dual booting, I also run sometimes windows 10 in a virtualbox machine inside arch and each time when I run that machine, after a while my keyboard dies completely both inside virtualbox and outside it and I can only fix it by rebooting the laptop. Without starting the windows 10 virtualbox, this second problem never happens.

All the packages in my arch system are up-to-date.

Just a small update:
I think I saw at least several times when this happens the following in the journal right at tye moment when it happens:

Feb 13 22:02:59 vali-defiance2 cinnamon-session[1091]: Gtk-CRITICAL: t+67.15229s: gtk_image_set_from_icon_name: assertion 'GTK_IS_IMAGE (image)' failed

Please help.

Thank you.

Last edited by burebista (2016-02-19 06:06:30)

ugjka · 2016-02-13 21:24:27

Nothing in Journal or dmesg?

burebista · 2016-02-13 22:55:50

For both situations (either key starting repeating indefinitely OR no key repeat but keyboard stopped responding) I see this in journal:

Gtk-CRITICAL: t+455.21963s: gtk_image_set_from_icon_name: assertion 'GTK_IS_IMAGE (image)' failed

And one more update: I run into the second situation (keyboard dying without a key repeating indefinitely) also without starting virtualbox.

mich41 · 2016-02-13 23:19:05

The above is unlikely to be relevant.

Does the key continue repeating if you switch to some VT?

Does this help?

rmmod atkbd ; modprobe atkbd
-- or --
rmmod atkbd ; modprobe atkbd reset=1

edit:
Type the command in advance before it happens and add # at the end of the line so that spam from the repeating key is ignored by shell

Last edited by mich41 (2016-02-13 23:22:39)

burebista · 2016-02-16 10:11:59

The keyboard died again without virtualbox windows machine running, which suggests that it is not related to virtualbox. In the moment when the keyboard died the processor fans speeded up briefly, like there was a short higher processing load. The only thing I was doing at that moment was quitting yaourt-gui. The only thing I could see in the journal at that moment (after reboot of course, because when any of the 2 keyboard problems happen, I cannot recover from them and I have to reboot) using the "journalctl -b -1" command was:

Feb 15 19:01:41 computer-name sudo[6684]: pam_unix(sudo:auth): auth could not identify password for [my-user-name]
Feb 15 19:01:45 computer-name cinnamon-session[6541]: Gtk-CRITICAL: t+37.58483s: gtk_image_set_from_icon_name: assertion 'GTK_IS_IMAGE (image)' failed

mich41 · 2016-02-16 11:46:28

If you have another machine on the LAN you could ssh into this laptop.

burebista · 2016-02-16 15:57:40

@mich41: yes, I could perhaps ssh from another laptop and perhaps reload atkbd or so to see if the keyboard starts working again, but that is not a solution. A solution means to do something so that my laptop does not get into these 2 situations (problems) any more.

Last edited by burebista (2016-02-16 16:45:49)

burebista · 2016-02-16 16:47:34

Do you guys know what packages might need to be changed in order to fix these 2 issues?

Last edited by burebista (2016-02-16 16:47:53)

alphaniner · 2016-02-16 17:59:10

I'm pretty sure mich41 proposed ssh'ing as part of a troubleshooting process rather than as a solution. As you alluded, valuable information may be lost when you reboot, information which would still be there if you ssh...

burebista · 2016-02-16 19:22:08

@alphaniner, @mich41: sorry guys, I was lazy to say more before: I thought of what alphaniner just said as well before my previous reply to mich41, but I thought that the only source of info during ssh-ing would be dmesg and that I could see that as well after reboot by using "journal -b -1" instead of dmesg. If that is not correct, please let me know. Also please let me know if there are any other log files I should look at.

However I might ssh into the troubled laptop when one of the 2 above exposed problems happen in order to see if reloading atkbd fixes the keyboard, because if it does, then it narrows down a bit where to look for a solution perhaps...

Could these problems be bound to my Skylake processor or to the Nvidia maxwell graphical card or a combinations of the 2?

Thank you.

alphaniner · 2016-02-16 19:58:23

AFAIK the journal doesn't necessarily get everything from dmesg, but I may be mistaken.

Also, am I under-medicated or did your name change? O_o

Or maybe you're a pod person, because your demeanor seems to have changed a bit too.

burebista · 2016-02-16 20:29:56

@alphaniner: I did change my username. When I registered I did not think much about what gets displayed and I just entered my email address as my user name...

Last edited by burebista (2016-02-17 09:06:21)

jasonwryan · 2016-02-16 20:34:02

To be clear: our policy about username changes and account deletions stands, but as this was clearly an error upon registration and everyone is given the opportunity of hiding their email address if they so wish, an exception was made.

burebista · 2016-02-17 20:14:41

So, here is an update:

The keyboard died again (meaning not the infinite repetition, but the second problem: no effect for any key pressed) and I ssh-ed into laptop, became root and have run:

# rmmod atkbd ; modprobe atkbd reset=1

That brought the keyboard back from death and I saw in journalctl for the above command:

kernel: input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input20

Then the keyboard died again after a few minutes, then I repeated the above procedure (ssh, …) and saw in journalctl:

kernel: input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input21

But keyboard WAS NOT RESURRECTED ANY MORE. I tried again and no keyboard resurrection.

Looking in dmesg for other similar messages, I see after boot:

input: AT Translated Set 2 keyboard as /devices/platform/i8042/serio0/input/input0

Also, running automatically after booting the command

# rmmod atkbd ; modprobe atkbd reset=1

does not fix my problem.

However, another thing also crossed my mind: each time I boot, I see in dmesg the following:

[Wed Feb 17 19:11:06 2016] ACPI Warning: SystemIO range 0x000000000000F000-0x000000000000F01F conflicts with OpRegion 0x000000000000F000-0x000000000000F00F (\_SB_.PCI0.SBUS.SMBI) (20150930/utaddress-254)
[Wed Feb 17 19:11:06 2016] ACPI: If an ACPI driver is available for this device, you should use it instead of the native driver

Could this memory conflict have anything to do with my problem?

I am waiting for ideas.

Thank you.

burebista · 2016-02-18 21:09:09

Hi,

I got again today both problems. Once the repetition and twice the dead keyboard. For the repetition I could not log in from another machine through ssh to see if there was something in the log because the ssh server deamon was not running, but for the dead keyboard I could and there was no message in neither the dmesg nor journalctl. Any suggestions please?

Also I would like to report these problems as a bug somewhere, because they are really annoying for me. Could you please tell me where is the best place to report these problems and against what components (or what categories?)?

Also a question for admins:
How can I change the name of this thread to "Random keyboard keys repeating indefinitely OR keyboard dies completely" because I strongly feel that both problems may have the same cause.

Thank you.

Slithery · 2016-02-18 21:25:50

burebista wrote:

Also a question for admins:
How can I change the name of this thread to "Random keyboard keys repeating indefinitely OR keyboard dies completely" because I strongly feel that both problems may have the same cause.

I'm no admin but you can change the thread title by editing your first post (although the title you want may be too long).

Last edited by Slithery (2016-02-18 21:26:31)

mich41 · 2016-02-19 22:05:45

The dying could theoretically be caused by many factors, but repetition really suggests f^&*ed Embedded Controller. Too many people use AT keyboards without issues. And if furthermore it turns out that reloading atkbd without reset=1 doesn't help for this then well...

I'd try BIOS updates, especially if there are known cases of this helping. In particular, mind KBC/EC Firmware Revision. W10 may just be lucky that it doesn't touch this chip in some particular ways it doesn't appreciate.

This warning about resource conflict isn't a problem, you are seeing it exactly because a driver which could interfere with ACPI has been disabled to avoid interference. Don't boot with acpi_enforce_resources=lax and you should be fine.

burebista · 2016-02-26 07:31:36

Just some updates:

I downloaded from laptop seller's website (pcspecialist.co.uk) a bios "update" and flashed that... and that actually downgraded(!) my bios (including the KBC/EC Firmware) and did not solve my problems. Pcspecialist is waiting to receive from the manufacturer of the chassis (a company named CLEVO) a proper bios update (not a downgrade) hopefully in a few weeks. When that happens, I will flash it and post the results.

Other than that I observed a few things about these 2 problems:
* Some days they do not happen at all, other days they happen a lot (sometimes even every few minutes!)
* Issuing "rmmod atkbd ; modprobe atkbd reset=1" (via ssh) fixes the 2 problems for a while and the "a while" ranges from 0 (basically no fix, so I retry the command again until it works) to several minutes or sometimes even hours. Afterwards it happens again and I have to issue again the command.
* Each time when these problems happen, a fan inside the laptop starts and spins quickly (so that I can hear it) for a little while (several seconds) and when it stops spinning, I already entered one of the 2 problems. Sometimes this happen even when I am not doing anything at laptop (no work load). The following thought may sound stupid, but I want to put it here to see what others think of it: can this speeding up burst of the fan when the problems happen suggest perhaps some multi-threading issues and/or an infinite recursive call loop somewhere which may blow up a stack and after the blowup the keyboard is one of the 2 exposed problems?

mich41 · 2016-02-26 07:47:44

I suspect it's the EC going crazy. EC is responsible for keyboard AND power management.

There are DE panel plugins and programs which display CPU load graphs in real time, you can install one to see if it correlates with actual CPU load - my guess is it doesn't.

If something really bad (like stack overflow) happened in the kernel, you'd get info in dmesg.

burebista · 2016-02-26 08:03:10

I have already installed in the task bar such a panel that displays a CPU graph and each time this happened I looked at it and indeed I have seen no CPU spike. Could it be that the fan speeding up briefly is not the CPU's fan? I mean could it be that the EC has its own fan that speeds up when this happens?

Last edited by burebista (2016-02-26 08:36:12)

mich41 · 2016-02-26 08:37:25

EC controls the main fan but it doesn't need active cooling itself because it's a simple chip which has no reason to run faster than single- or double-digit MHz.

WorMzy · 2016-02-26 10:15:37

Please post the complete journal from a session where this problem occurred.

burebista · 2016-02-29 20:30:04

Hi guys, I think that partially discovered what is causing the 2 problems:

I am using a wireless Logitech mouse M525 which works with a small Logitech Unifying Receiver USB dongle. These receiver can be configured using a small unifying software app from Logitech (which runs only in Windows) to talk to multiple Logitech wireless devices. I had my receiver configured to talk to, in addition to the above cited mouse, a keyboard that I do not have anymore and another mouse that I did not use for a long time. Once I unbound the receiver (using the windows mentioned software) from anything but the M525 mouse, the 2 problems happen much less frequent than before. How did I discover this: I bought a new Logitech wireless K360 keyboard with the intention that when any of the above 2 problems happens, then I would try to just connect the new wireless keyboard and hopefully I could enter with it the atkbd reload command instead of using ssh. So I booted in Windows and using the unifying software I added the new keyboard to the unifying USB dongle. That operation and the keyboard worked without problems in Windows. Then I booted in Arch and all hell broke loose: random system hangs with kernel panics either during system booting (before I would get a graphical login), or immediately after the graphical login and ALWAYS(!) during copying huge files. Yes, I was always getting kernel panic while copying a huge file (of more than 10GB for a Windows virtual machine that was not running). So I was very puzzled about that kernel panic during copying a huge file, because I never had that before configuring the new wireless keyboard. So I tried copying the file from one disk partition to another, or from one disk partition to an USB stick or from one disk partition to an external HDD and I was ALWAYS getting the kernel panic. Then I thought that it must be because of the unifying stick that was used for multiple devices. Then, once I unbound all devices excepting the mouse that I use, all kernel panics went away and the above 2 problems happened much less frequent. Another bad thing about these kernel panics was that I could not see anything after a reboot after a kernel panic in journal (as root with journalctl -b -1/2/...) about the kernel panic and sometimes it was telling me after reboot that journal file for previous boot was corrupted!

Hmm(!), in fact it just happened again (keyboard died) after 2 days during which it only happened once (I "cleaned" the unifying stick associations on Friday evening). The journal is to huge to post it here, BUT going through it, I could not reach the point where this keyboard dying just happened because of a "Failed to get journal fields: Bad message"! The output from dmesg is also to huge to list here but when this happened, I could see nothing about it in dmesg.

Last edited by burebista (2016-02-29 20:32:31)

mich41 · 2016-02-29 23:04:54

$ cat >xxx.c
#include <stdlib.h>
#include <unistd.h>

int main() {
  setuid(0);
  system("rmmod atkbd ; modprobe atkbd reset=1");
}
CTRL+D
$ su
# make xxx
# chmod 4755 xxx

Now delete xxx.c, put xxx on desktop and start it with mouse when you need it

BTW, it doesn't work without reset=1? Or does it?

burebista · 2016-03-01 19:22:25

Hi guys,

I do update my system every day using yaourt-gui. Today I noticed that libinput was updated but nothing else that was part of the system (I mean not an application). After that update I restarted the laptop and the fan speed bursts came back without me doing anything particularly and without any messages in journal or dmesg. Could these 2 things by related (libinput and fan speedings)? I am asking because from what I have read, libinput is used by Wayland, but I am using X and I am wondering if libinput is also used somehow for X as well and if it could play a role in the keyboard (and fan) problems that I am having?

@mich41: I created that small tool, when the problem happens, I will try it. Thank you.

Also I am waiting in the next several days (hopefully!) to see a BIOS update for my laptop released (conform to the laptop's seller) and flash it.

Thank you.

Last edited by burebista (2016-03-01 20:25:28)

Arch Linux

#1 2016-02-13 21:19:26

Random keyboard keys repeating indefinitely / keyboard dies completely

#2 2016-02-13 21:24:27

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#3 2016-02-13 22:55:50

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#4 2016-02-13 23:19:05

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#5 2016-02-16 10:11:59

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#6 2016-02-16 11:46:28

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#7 2016-02-16 15:57:40

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#8 2016-02-16 16:47:34

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#9 2016-02-16 17:59:10

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#10 2016-02-16 19:22:08

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#11 2016-02-16 19:58:23

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#12 2016-02-16 20:29:56

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#13 2016-02-16 20:34:02

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#14 2016-02-17 20:14:41

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#15 2016-02-18 21:09:09

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#16 2016-02-18 21:25:50

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#17 2016-02-19 22:05:45

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#18 2016-02-26 07:31:36

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#19 2016-02-26 07:47:44

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#20 2016-02-26 08:03:10

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#21 2016-02-26 08:37:25

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#22 2016-02-26 10:15:37

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#23 2016-02-29 20:30:04

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#24 2016-02-29 23:04:54

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

#25 2016-03-01 19:22:25

Re: Random keyboard keys repeating indefinitely / keyboard dies completely

Board footer