Bug 9302
Summary: | Kudzu probes locking up keyboard on laptop | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Product: | [Retired] Red Hat Linux | Reporter: | Joe Ceklosky <jceklosk> | ||||||||
Component: | kudzu | Assignee: | Bill Nottingham <notting> | ||||||||
Status: | CLOSED RAWHIDE | QA Contact: | |||||||||
Severity: | high | Docs Contact: | |||||||||
Priority: | medium | ||||||||||
Version: | 6.2 | CC: | elladan, ml, rvokal | ||||||||
Target Milestone: | --- | ||||||||||
Target Release: | --- | ||||||||||
Hardware: | i386 | ||||||||||
OS: | Linux | ||||||||||
Whiteboard: | |||||||||||
Fixed In Version: | Doc Type: | Bug Fix | |||||||||
Doc Text: | Story Points: | --- | |||||||||
Clone Of: | Environment: | ||||||||||
Last Closed: | 2000-04-05 14:13:25 UTC | Type: | --- | ||||||||
Regression: | --- | Mount Type: | --- | ||||||||
Documentation: | --- | CRM: | |||||||||
Verified Versions: | Category: | --- | |||||||||
oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |||||||||
Cloudforms Team: | --- | Target Upstream Version: | |||||||||
Embargoed: | |||||||||||
Attachments: |
|
Description
Joe Ceklosky
2000-02-10 14:38:53 UTC
Are there any messages in /var/log/messages around the time that the keyboard is disabled? It might not be kudzu. What runs just before the language selection dialog comes up in text mode install? I flipped over to another v-terminal and the last message I see is: ... * going to insmod fat.o (path is NULL) * going to insmod vfat.o (path is NULL ) Also there is no /var/log/messages file on the initial install root file system. The very first thing before the keyboard selection is the general newt startup; does it lock up as soon as the screen turns blue? It locks up as soon as the language selection screen comes up. So the screen is blue for a second then language selection is up and the keyboard is locked. Shutdown the notebook down to memory using the hotkey and turn it back on it's unlocked. *** Bug 5813 has been marked as a duplicate of this bug. *** Can we now start looking at how to track this down rather than pointing this back to another bug report. What is it that kudzu is doing that it can mess up the keyboard? Blocking interrupts? I have the same problem when kudzu runs to detect new hardware. The keyboard is dead. What happens if you run something like 'Xconfigurator' or 'setup'? This seems to be happening on my Compaq LTE 5100. It occurs during a network ftp install immediately after the "loading second stage ramdisk" finishes. If I sleep the laptop, then unsleep it, the keyboard works again. Unfortunately, sleeping it turns off the PCMCIA devices, thus shutting down the network, thus making a network install impossible since it doesn't reset the pcmcia bus later. Again, does this happen only in the installer and kudzu, or in other newt-based apps, such as 'Xconfigurator' or 'authconfig'? There are NO lock ups running authconfig or Xconfigurator. OK, I'm going to attach a short program. Does it exhibit the same behavior? Created attachment 129 [details]
short test program.
Yes. This program does lock up my notebook keyboard. (me as of #5813 - sorry for not responding earlier, I was on vacation) Mario (or does it return?)If you strace the program, where is it hanging at? It returns. It detects the Mouse (the built in touchpad) properly and returns to the prompt. However, the keyboard is gone by then.. Doesnt respond to any keys, not even to sleep. The power key works, though :) Created attachment 131 [details]
updated test program
Does the new psaux2.c program fare any better? There's no real reason for it to lock up the keyboard, but you knew that. ;) Nope, keyboard still hangs. No real reason... Well, yeah, no real reason.... Hmm, I have no idea about how PSAUX is supposed to work, but since mouse and keyboard interface are the same controller/port, it kinda makes sense. And since you do only a read on that file, it would look to me like more a problem in the psaux driver than in kudzu.. If you start gpm at boot, does it also hang the keyboard? Also, when the keyboard locks up, are there any messages in the system logs? See #5813 - GPM works just fine. However there is a nice twist: While GPM is running, psaux.c doesnt hang the keyboard either. However as soon as I stop GPM, and try it again -> kaboom PS: Unless you respond today or early tomorrow (saturday) I wont be available for another week - sorry After reinstalling 6.2 the same keyboard locking is there on the language selection screen. Putting the laptop to sleep ad waking it up unlocks the keyboard. Also when kudzu runs (when it's trying to detect new hardware), it counts down from 10 to 1. During this period the keyboard does NOT repsond as well. What is KUDZU doing while it is counting down or what did it do to the keyboard just before starting the count down? The countdown is most likely the countdown to enter the hardware config screen before it times out (I'm assuming the "Welcome to Kudzu" dialog is showing at this point.) (yes, we didn't get a chance to get this fixed before 6.2; whatever fix is most likely in the kernel.) Yes the entire Kudzu screen is up and the counter is counting down starting at 10 9... What is Kudzu doing during the count down. Waiting for keyboard input. We don't want the boot process to hang while we ask the user if they want to configure some new device, so we set it to time out after a while. OK, finally got ahold of a laptop here that exhibits this behavior Does the new attachment (psaux3.c) cause the keyboard to lock up still? Created attachment 187 [details]
yet another test program.
*** Bug 10457 has been marked as a duplicate of this bug. *** The new program does NOT lock the keyboard. I upped the sleep to 10 seconds and switched virtual consoles with no problem. How did the old code work? Because there's no reason to *need* the sleep there; did it work OK before you upped the sleep to 10 seconds? Sure it worked fine before I upped the sleep value. I was refering to the old version of the code where you were opening the device. Fixed in kudzu-0.38; will be fixed in the installer whenever it's rebuilt against that. |