2in1 problem thread. (nvidia kernel module vs X module, and strange workbug phenom)
hey all...got two delicious problems for you all today.
First one should be a cinche. 1: on boot, (EE) NVIDIA(0): Failed to initialize the NVIDIA kernel module! Please ensure (EE) NVIDIA(0): that there is a supported NVIDIA GPU in this system, and (EE) NVIDIA(0): that the NVIDIA device files have been created properly. (EE) NVIDIA(0): Please consult the NVIDIA README for details. (EE) NVIDIA(0): *** Aborting *** somewhere else, i found some logging that says the kernel module VERSION is different from the X module version, and that that's bad. i'll post that as soon as i find where it is... anyway, the quick fix is easy enough. all i do is kill gdm, `rmmod nvidia`, and restart gdm and presto, it starts, with the right driver...`nvidia` gets loaded. the kernel module installed is the one from the nvidia binary driver. i'm sure i just have to get the right module loaded at boot, i just don't know how to do it. one more weird thing...although i can run glx apps without X crashing, and beryl works just fine, on glxgears i get like maybe 15 fps, and some glx apps don't do so hot (scorched3d crashes frequently, some screensavers really choke up) is that just a linux-drivers-blow thing, or is my card not being fully used? OKAY, problem 2 is a bit more obtuse, i'll be amazed if anyone can help me with it. 2: sometimes when i'm just happily minding my own beeswax (surfing the web, listening to music...nothing cpu intensive) the CPU starts, out of no where, working on something furiously. so furiously, in fact, that X completely stops responding, i can't get to any other terminals, and the one time i did get to a terminal, the computer was so busy it wouldn't even run my "top" command. eventually i just have to cut the juice. i have a cpu applet in my panel, and it shoots up to 100% and stays there, till it eventually stops showing new data, the computer is so busy. worth noting is that about half of the load is "IOWait" as opposed to actual work. anyway here's a list of apps that i'm usually running when it happens...unfortunately i'm usually running enough apps at one time that i can't really pin it down. gnome-terminal gedit amarok beryl 0.2 firefox 2 bittornado evolution liferea vlc 0.8.6 nautilus also to note is that a. it's not the http cache cleaner b. i'm usually using a lot of my ram, and usually around 50% of my paging file. okay, thanks for any responses to my problems, let me know if there's any more info i can provide, especially with the second one i really don't know what to post. here's some system info: amd64, x2 4200 (dual core) 1gig or so ram ubuntu edgy 6.10, kernel 2.6.17-11-generic nvidia 7600GT using driver 1.0-9755 (for amd64) xorg 7.1 gnome 2.16 thanks again :newbie: |
you need to include more details on your first problem like: which distro are you using? what method did you use to install the drivers?
|
the distro i include at the bottom of the post, and i used the nvidia installer for both the driver and kernel mod.
|
Hi,
Could you include your xorg.conf file? When you used nvidia-installer, were there any error messages? You can check the nvidia log in /var/log location. If there are errors, please post them. Also, what size is your swap partition? Sam |
hey...swap size is 500meg. nvidia installer worked without any errors, but i found that mismatched version message in the installer log. posted the whole end of the log:
Code:
NVIDIA: left KBUILD. hope that helps |
sorry, forgot my xorg. here it is. (edit) i should mention that the x log says everything in here is okay, and i only get errors when it loads the nvidia module.(/edit)
Code:
# nvidia-xconfig: X configuration file generated by nvidia-xconfig |
Do you remember how you installed the binary driver?
Did you tell it to update (nvidia-installer --update) or did you download a new one from nvidia? And do you remember if you installed the 1.0-8776 driver, or if slackware did it for you? (I don't know much about slackware. If they have a different method of installing nvidia proprietary software, it may conflict with nvidia's install method) It might be best to uninstall the nvidia drivers completely before updating them. I read somewhere that nvidia drivers need to be patched in order to install them successfully on a 2.6.17 machine. Let's hope it doesn't come to that. The thread I read it on is below. It's for suse, though, not slackware. http://forums.suselinuxsupport.de/in...d=176917&st=0& When you re-install the nvidia drivers, try it this way first: sh NVIDIA-Linux-x86-1.0-9755-pkg1.run --kernel-source-path=/pathto/linux-source Where "/pathto/linux-source" is the path to your 2.6.17-11-generic kernel source files. |
Quote:
(EE) NVIDIA(0): that the NVIDIA device files have been created properly. They are talking about the entries in /proc that the driver uses to communicate with the hardware. If you load it as the last thing in the startup sequence, this issue should go away. 2. Freezing is a really vague problem. I'd unhook every piece of hardware you own, except what's necessary for the system to run, and see if it still happens. If it does, kill every last service you don't need for the system to run. If it still hangs then, I'd bet on hardware, like a spotty switch, NIC, loose cable, hard drive dying etc. I/O issues are the most common cause of stuff like this. This will happen with a fubar'd CD in the drive, a bad or loose drive cable, intermittent network cable, etc. IOWAIT is "I just sent data to a device and I'm waiting for it to respond with something I understand". This will hang your computer like a former dictator if the driver doesn't get an ACK. -Viz |
haha okay, i'll cut my comp off...nothing but rice crackers and skim milk. hopefully it is just a bunk cd or something.
as for loading the driver...i thought that the driver doesn't load until X does? you mean load the kernel module last? i think the kernel module is getting loaded on runlevel 2... but my bash is a little rusty. this is S20nvidia_kernel (S20 is ubuntu speak for "enabled" i think) /etc/rc2.d/S20nvidia_kernel: Code:
#!/bin/sh Code:
alias char-major-195* nvidia samstar: i'm running ubuntu, not slack...hopefully that's not my prob :D anyway i'll start messing around till i hear back |
For x to start the video driver needs to be loaded. You specify the driver name in your xorg.conf but it needs to be loaded already for x to start.
If the driver's not loaded it will fail. This is why you should set your box to boot up to console and start x manually using startx when using vendor provided video drivers. Less hassle when you need to upgrade the kernel and your video driver dies. If you boot up to shell you just need to re run the nvidia installation script, as opposed to waiting for things to fail so you can drop to a shell. It's just easier on your constitution. There's more ways than one to skin this cat. This is just how I've been doing it since nvidia first put out a driver for linux. It's hard for old habits to die. It's an issue that you deal with so infrequently it really doesn't matter as long as your box works. |
ahh I see...I didn't realize it had to be loaded beforehand.
yea, when i used to run deb I'd start X by hand, but since switching to ubuntu i've just left the default boot...anyway, i've done the changes, i'll let you know what happens. |
okay sorry it's been a while since i posted, been trying stuff out.
new news: booting X last/booting X by hand doesn't make a difference I noticed that the module that gets loaded during boot is actually smaller (nearly half the size) of the one that x loads after I remove the boot one. where can I change the module that loads at boot? for that matter, what do I even change it to? both modules have the same name. are they in fact two differnt modules, or is there something fishier afoot. |
Did you solve this issue? I'm having this same problem now too. I have to manually rmmod nvidia, modprobe nvidia, then restart the GDM, before I can get into X.
|
no, still haven't solved it. let me know if you do.
|
Had similar problem.
It was after one of those automatic updates. Had to re-install the NVIDIA driver to get to to work. Lemme know if that fixes it on your systems too.
|
All times are GMT -5. The time now is 05:40 PM. |