Weird system behavior since 2.6.25 kernel. Help needed.
Linux - KernelThis forum is for all discussion relating to the Linux kernel.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
Weird system behavior since 2.6.25 kernel. Help needed.
I have some strange system behavior on > 2.6.25 kernels=( At the first glance everything works fine (my broadcom wi-fi NIC is working well, finally): suspend, resume, no performance decrease, but at some moment strange things happen. I cannot start any new process. When I click on some icon in KDE the application doesn't start. I cannot move windows around the desktop. Can't reboot system. I can kill the X server with CTRL+ALT+Backspace _sometimes_. Mostly the system freezes with blank screen, but sometimes I can get the console login shell. Anyway, I can only type login there, the password promt never appears...
I have never seen my system act like this before 2.6.25.
The config I used to compile kernels is always the same. Nothing is changed in my system except kernel. All kernel sources are vanilla ones.
It could be that the problem is not the kernel, but XOrg/XWindows.
First, roll-back to your previous kernel version and make sure that the system is again stable.
(I presume that you do keep separate and clearly-identifiable backup copies of each .config file... oui?)
Then, carefully setting-aside your .config file, do a make depclean in the kernel directory, restore the .config file, and recompile/reinstall everything. Reboot and see if you can reproduce the incorrect behavior.
If so, then the probable cause is that the kernel information about your display-device is no longer correct.
Kernel configuration-parameters do change, and the easiest way to find out what may have happened is to diff the old config-file against the new.
First, roll-back to your previous kernel version and make sure that the system is again stable.
I did this. Now I'm on 2.6.24.7 and everything is ok.
Quote:
(I presume that you do keep separate and clearly-identifiable backup copies of each .config file... oui?)
I do
Quote:
Then, carefully setting-aside your .config file, do a make depclean in the kernel directory, restore the .config file, and recompile/reinstall everything. Reboot and see if you can reproduce the incorrect behavior.
It was the first thing I've done, when saw this strange behavior. It didn't help - system behaves badly after clean recompile.
I'm not sure if it is my screen driver's malfunction. As I said before, I can kill X (sometimes) and get to console login promt (no screen corruption or anything like that), type my login, press Enter...but no password promt appears I can type some letters and they appear on screen, though.
I agree with "unSpawn"... time to look at the logs.
Let's face it: an awful lot of things about your system right now work just fine. You can not-only "see 'windows'," but you can type into them! Ergo, the actual disruption {"disruptive" though it may be...} is very localized.[/i]
It almost-definitely isn't "the kernel." Check the XWindows/XOrg log files . . .
Distribution: approximately NixOS (http://nixos.org)
Posts: 1,900
Rep:
By the way, try to switch to text console with Ctrl-Alt-F2 . Does it behave better than after killing X? Try changing default runlevel in inittab to 3 or even 1. Does it work better? Try adding an init script that waits until X is launched and a minute more, then launches xterm (manually specifying DISPLAY). What happens? What output does it generate (surely you should redirect xterm output and stderr to files)?
I think you guys got me wrong. The system is working fine *for some time*...it can be even a day! But in the end I _always_ have a situation I described in the first post
As for text console, I told that after killing X I get to text login promt, but cannot login because the password promt never appears.
Uh...maybe I can't describe the situation clearly in English
Today I'll try to reproduce this bad behavior of my system and look through logs more carefully. But the problem is that I do not know how much time do I need to wait until this sh*t happens.
upd.
kernel 2.6.25.4, uptime 54 min. Everything is ok. I'm not sure how long should I wait until something happens...but it will happen
Okay. After about 8 hours it happended again...
No errors in dmesg output except Hangcheck: hangcheck value past margin!
Nothing odd in ps aux ouput.
The symptomes were: firefox stopped scrolling and opening new tabs (i could close the existing ones, though), pidgin couldn't open new conversations, but the existing ones were usable (but little bit slower) and conky stoped updating. I managed to run dmesg and ps from the xterm window, which I always keep opened. The I tried to switch to text terminal with CTRL+ALT+F# buttons with no effect. And then I tried to kill X. It died, but I got my system stuck with a blank screen
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.