Hello all,

This is on an up-to-date -current machine with:

And:

up-to-date multilib equivalent packages as well

I am using an integrated Intel GPU on this laptop.

---

So I've run into a weird issue that I'm having trouble debugging myself so I thought I'd come here and ask for some help.

Here's what's going on:
=====
I'll be using X11 for awhile (doing whatever) and suddenly the keyboard and mouse don't seem to respond to anything with the exception of:

1. Mouse movement. I can move the cursor around but clicking on anything does absolutely nothing.

2. Keyboard commands for non-X11 things. i.e., I'm able to switch to a tty but if it's firefox I can't type anything in the address bar, or if it's urxvt I can't type anything.

Still regarding the second point, attempting to do things that are X11 things like killing X11 with Ctrl + Alt + Backspace does not work.

This doesn't seem to be specific to any one program. It just begins to affect the program that's currently in use. Hence I think it's related to X11.
=====
So I was wondering if anyone could help me debug this or, if anyone had run into similar problems?

To assist I've attached the Xorg.0.log to this post with a few notes:

1. I included that event at [82.155] to indicate that for several thousand seconds nothing happened (so basically nothing happened for a long time after I started X11).

2. The events that follow after that time mark was me switching to a tty to kill the program that was being used (I swapped back and forth a few times) since as previously stated, kill(1)ing that program seems to "fix" this issue.

Regarding #2 again, there was nothing in between that time period that showed up in Xorg.0.log so whatever this problem is wasn't logged so I'm not sure if the Xorg.0.log is of much help.

So that also indiciates that Ctrl + Alt + F# was working since it's not an X11-related key combination, but I wasn't able to type or do anything in any programs until the program being used at the moment in X11 was killed.

Attached Files

Xorg.0.log (29.7 KB, 9 views)

I notice that it's an HP machine. I have one of those to and there is an occasional problem with keyboard keys sticking down (logically not physically) that only happens in X11. If one of your modifier keys windows/shift/alt/ctrl is the one that sticks down then it can look like the system is unresponsive. Pressing the stuck key again will sort it out. Next time it happens, try pressing each of alt/shift/windows/ctrl individually and then see if keyboard function is restored.

Only workaround I've found so far for this is turning off keyboard repeat with an 'xset r off', but unfortunately that is annoying.

1 members found this post helpful.

Hi GazL,

Thanks for the info.

Next time this happens I'll definitely try this out and give an update on this post.

===

EDIT: Has this been happening recently on -current for you?

This didn't happen until recently (a few weeks ago) for me.

Sadly, yes, it's been happening ever since I got this laptop. It only happens with the laptops keyboard, not if I plug in an external/usb attached keyboard.

I've found a way to reproduce it at will.

Open An X11 application that takes input, text editors are a good choice. Hold down a letter key and then rapidly feather it (slightly release the pressure just enough to break the contact and then reapply). Keep doing that fast enough and before very long (I can usually trigger it within 10-20 seconds) the key will start repeating on its own and will continue when you remove your finger. Pressing the same key again will generate a key-release event and stop the sequence, but not the key-press event that should normally proceed it.

It happens with both evdev and kbd drivers under X11, but I can't reproduce it either in windows or on the linux console.


I think it's some sort of race condition between a key release and key press. It seems that sometimes the Xserver misses a key-release and starts doing it's key repeat routine. I suspect that the physical construction of many keyboards may not actually allow them to generate the press/release fast enough to trigger the issue, usb latency may likely also prevents the release/press happening fast enough to trigger this on usb keyboards, but that's all guesswork on my part.

'xset r off' does prevent it happening, but means you have to live without key repeat.

P.S. This may not be the same issue as you are having, it's just that what you described sounded exactly like what happens here when X gets confused about the state of alt or the windows key (which isn't as easy to see as with a letter because they aren't printable).

Well, unfortunately hitting the key in question (or any key) did not work.

I ran into the problem once again today and this time the faulty X11 program was zathura (a pdf viewer).

I'm still quite lost on why this occurs. :/