From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; fr-FR; rv:1.2.1) Gecko/20030225 Description of problem: Sometimes (about once an hour), a key I just pressed keeps pressed after I have released it. For example, it happens that I am typing a text and, for no logical reason, a character repeats itself while I don't have my hands on the keyboard, giving something like thaaaaaaaaaaaaaaaat. If I press this key again, it stops repeating itself. It can be a big problem when the key which keeps pressed is an inviible key (like CTRL) because I don't notice it (as it doesn't _write_ anything) but pressing a key (for example 'c') will act like CTRL+'c' instead of just 'c'. Here again, I have to type 'CTRL' again to go back to the normal behaviour. I gave it 'high' severity because you can lose data by typing the wrong key (for example CTRL-'c' instead of 'c') or keeping the 'SUPPR' key pressed forever while wanting to suppress only once. Version-Release number of selected component (if applicable): 4.3 How reproducible: Sometimes Steps to Reproduce: 1.Type a text 2.Sometimes, a key repeats itself while not pressed anymore 3.Click on this key again 4.It's back to normal Actual Results: Here is an exampleeeeeeeeeeeeeeeee(...I put my hands on my head...)eeeeeeeeeeeeeeeeeeeeeeeeee(...I press 'e'...)e of what I see sometimes. Expected Results: Here is an example of what I expect. Additional info: -I already had this bug on Redhat 8 (XFree 4.2) -My hardware is a Fujitsu/Siemens Amilo-D laptop with an ATI Radeon Mobility M6. -I have DRI enabled on Redhat 9 but I didn't have it on Redhat 8. -I don't know if that bug exists in non-X mode as I only use X.
> -I don't know if that bug exists in non-X mode as I only use X. Please find out, and provide that information here. Also please attach your X server log and config file, and your /var/log/messages as bugzilla file attachments.
Please boot your machine into runlevel 3 and indicate if the problem occurs at the console or not. This type of problem is almost certainly a hardware related flaw, kernel bug or somesuch, and unless I can reproduce it, then I can't debug it at the software level. I've never seen this behaviour on any hardware I've got access to, and nobody else has reported this problem either, so you're going to have to narrow the problem down yourself as to wether it is the kernel, your hardware, X or something else, and then provide those details. Please update ASAP.
Created attachment 91389 [details] It's my XF86Config file
Created attachment 91390 [details] It's my XFree86 logfile
Created attachment 91391 [details] It's /var/log/messages
Created attachment 91392 [details] A tgz archive containing two older /var/log/messages
Sorry for being so long to anwser. I attached the files you asked me for. I also tried to reproduce my problem in runlevel 3 but I couldn't. BUT that doesn't mean that can't happen in runlevel 3 because in fact it happens so rarely that I should use my machine in runlevel 3 for 1 whole day to be perfectly sure that never happens un rl3. If I happen to reproduce it in runlevel 3, I'll tell you here, on bugzilla. Thanks for your time !
I finally happened to reproduce the bug in non-X mode. Here's what happened. I was trying to type spanish characters (just for fun...) so I tried to combine ALT-GR, CTRL, SHIFT, ALT, FN, the WINDOWS KEY etc... with normal letters quite randomly and check what it wrote (I did it in X). After a while, I noticed that when I typed normal letters (without any modifier pressed), strange symbols appeared. I understood that one (or more) of the modifiers I tried was probably blocked. SO, I switched to the first non-X virtual screen (CTRL-ALT-F1). There, I tried to log in. typing my login worked (julo) but trying to type numbers (SHIFT+1 on my keyboard) didn't work. So, I tried to press each of the modifiers one time so stop the strange behaviour. And that worked. I could type numbers again. Then, I went back to X and it worked too. I'll attach my /var/log/messages file.
Created attachment 91393 [details] My lates /var/log/messages file targzipped
As suspected, this problem occurs outside of XFree86, so it can't be an XFree86 related problem. It is either a hardware issue, or a kernel issue. I rather suspect the former, however I'll reassign it to the kernel component for comment.
*** This bug has been marked as a duplicate of 74635 ***
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.