From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4b) Gecko/20030509 Description of problem: After upgrading to kernel-2.4.20-13.9, I loose the ability to enter keystrokes in X after an indefinite amount of time (anything between 2 and 10 minutes). *ALL* keyboard activity is lost (Caps Lock keyboard lights, [Alt-SysReq-S] Emergency Sync, [Ctrl-Alt-Fx], etc.). The mouse still functions. Version-Release number of selected component (if applicable): XFree86-4.3.0-2 How reproducible: Always Steps to Reproduce: 1. [upgrade to 2.4.20-13.9] 2. startx 3. Enter keystrokes in X Actual Results: After some time, the keyboard ceases to function. Expected Results: Ability to continue entering keystrokes. Additional info: After either : - exiting (by means of the mouse) and restarting X (*) ; - suspending and resuming (laptop : IBM ThinkPad A30p) ; keyboard functionality in X seems to be fully and permanently restored, until after the machine has been restarted (**), after which it occurs again. (*) after exiting X, keyboard is functional in the console (however, as mentioned, in X I'm not able to CTRL-ALT-Fx to a virtual console), so X can be restarted again by means of 'startx'. (**) due to crash bug #90993, apparently I'm only able to suspend/resume when not removing USB devices. Note 1: concerning bug #90993 : I've only been able to reproduce that crash when resuming in X ; thought I'd mention that here too. Note 2: this bug is reproducible with both RawHide XFree86-4.3.0-5 and Shrike 4.3.0-2.
If you've upgraded the kernel, and now something doesn't work, it stands to reason that it is a kernel problem, not an XFree86 one, since X has not changed. It sounds like your kernel has crashed. Please attach your /var/log/messages from after this problem occurs. Reassigning to kernel component.
Also attach the output of "lsmod" from while the kernel is running please.
Created attachment 91719 [details] lsmod
Nothing in /var/log/messages. Are you sure this is a kernel issue, and not a kernel upgrade-induced XFree86 issue ? As I reported, I can continue using XFree86 with the keyboard after exiting and restarting X ; there is no indication of a kernel crash/oops in /var/log/messages.
Additional info : Note 1 : I was copying some data between two external IEEE-1394 drives, when /var/log/messages reported FS problems on one of the external drives (whether that is a harddisk hardware problem or bugs within IEEE1394/OHCI1394/SBP2 drivers is not relevant here, I suppose). Since then, I'm experiencing the X keyboard lockup every 5 to 10 minutes. Please note the FS corruption is limited to one of the IEEE1394 drives, so I am not yet concluding my laptop hardware platform is falling apart (would be very coincidental, together with the 2.4.20-13.9 update). Note 2 : I already mentioned suspending/resuming to restore keyboard functionality. I suspended by pressing laptop keys [Fn-F4], not by closing the lid. This implies that at least (and for now, exclusively) the special (hard-wired) keyboard keys are recognized. For now, I'm obliged to rollback to 2.4.20-9, security fix or not. This is pretty painful, as I'm in the midst of setting up several (to be exposed) new servers with RHL 9 (yes, I am aware of RH ES/AS :-( ).
Putting left foot in mouth : I just experienced the lost keyboard behaviour with kernel-2.4.20-9. I'm assessing my latest RPM installations, and apart from the obligatory RHN-updates (kde, xinetd, kernel & tcpdump), I only installed mozilla-1.4b (RawHide) and alsa-0.9.3a (FreshRPMs). I already reverted to OSS, to exclude ALSA, but it strikes me as odd that the keyboard lock usually (but not always) happens when typing in Mozilla. As I've been running mozilla-1.4b_gtk2 (mozilla.org) for a while without problems, I'm very hesitant to cry wolf a second time, so I'll refrain from reassigning to the mozilla component. Any suggestions to debug this ?
As I'm experiencing some strange anomalies with 2.4.20-9 now too, I'm having my machine serviced.