Red Hat Bugzilla – Bug 481154
Sporadic change of keyboard to a "bad"/slow state
Last modified: 2009-02-25 18:23:31 EST
Description of problem:
Every few days, I've had Xorg suddenly transition to a bad keyboard state, where all keys need to be held down for about half a second in order to register (making typing very slow and difficult). Neither dmesg nor /var/log/Xorg.0.log show any unusual events. A bit of experimenting with xev shows that X doesn't even get the keystroke until after that half-a-second (these are all keystrokes, from control to the letter keys). Changing to a text console lets me type without any difficulty - switching back into the X session restores the unusable configuration. Starting a second X session from the console gives me a usable setup, as does logging out of the first session and letting it restart.
Version-Release number of selected component (if applicable):
Limited. Problem occurs only every few days, and always when the system is active. It happens during normal use - if I notice any common threads that may lead to the problem I'll add them here.
Steps to Reproduce:
System is a laptop. I have yet to try plugging in an additional USB keyboard to the system when this happens to see if it's affected.
There is some problem with spontaneous switching on of the "Slow keys" accessibility. Run
and switch it off. Is that correct?
I'll try that the next time it happens, and let you know if it works. Thanks.
*** Bug 479204 has been marked as a duplicate of this bug. ***
It happened again, and bringing up that panel, switching on slow keys, and then switching it off fixed the keyboard.
However, two things make me curious:
That panel did not think that slow keys was on, and only by toggling it on and then off was it fixed.
I don't normally run GNOME on my system (meaning no gnome-settings-daemon was running until I started gnome-keyboard-properties), so I'm a bit doubtful it was anything GNOME that set this.
It seems likely that the underlying mechanism gnome-keyboard-properties/GNOME is using is what was somehow triggered. I wonder if there's a way to disable that mechanism globally in X.
*** This bug has been marked as a duplicate of bug 445898 ***
Has this got anything to do with the mouse cursor being corrupted and appearing as a big pile of pixels? Why was my bug closed and marked as a DUP?