...

What type of information should I collect to find the problem? I would really like to fix this.

I'd look at the apm kernel options, most likely apm=off. Try booting the default kernel, but add "apm=off" to the command line (grub let's you add options to the kernel command line at boot time). If this solves the problem (as I suspect that it will), the modify /etc/grub/menu.lst and add "apm=off" to the kernel boot for the default kernel (probably under the Desktop or default sections). Hope this helps. -Nick -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Sorry about the extra message. I was using Kmail and selected send later but it sent it right away. Here is the link to the other post about a intel video driver problem. http://forums.opensuse.org/showthread.php?p=2282171 Bruce Samhaber wrote:

I tried with "apm=off" but it still hung. I will change grub anyways to have "apm=off" since the machine is 2005 vintage and I believe that it used acpi in SUSE101.

I think this is somehow connected to the networking. If I do not configure the Ethernet ports, the machine stays up for hours. Once there is heavy network traffic then the machine hangs.

I tried the online update to see if this would fix teh problem; but the machine hung in the middle of the update which corrupted the system and the update process. I had to re-install fresh from the DVD with a re-format of the disk partition. Then booted in failsafe mode to do the online-update.

The problem still exists but at least this the up-to date version.

I will now try to boot with nomodeset option as I found a forum posting that refers to a VGA driver problem.

- Bruce.

On Wednesday 19 January 2011 01:04:35 Nick LeRoy wrote:

...

...

...

What type of information should I collect to find the problem? I would really like to fix this.

I'd look at the apm kernel options, most likely apm=off. Try booting the default kernel, but add "apm=off" to the command line (grub let's you add options to the kernel command line at boot time). If this solves the problem (as I suspect that it will), the modify /etc/grub/menu.lst and add "apm=off" to the kernel boot for the default kernel (probably under the Desktop or default sections).

Hope this helps.

-Nick

-- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hi Markus,

From lspci

01:00.0 VGA compatible controller: ATI Technologies Inc RV350 [Mobility Radeon 9600 M10]

I get the following messages from /var/log/Xorg.0.log that seem like they might be relevant butnot sure what they mean.

[ 28.443] (EE) Failed to load module "vboxvideo" (module does not exist, 0) [ 29.193] (EE) No input driver/identifier specified (ignoring) [ 29.211] (EE) No input driver/identifier specified (ignoring) [ 29.213] (EE) ioctl EVIOCGNAME failed: Inappropriate ioctl for device [ 29.213] (EE) PreInit returned NULL for "TPPS/2 IBM TrackPoint" [ 29.213] (EE) No input driver/identifier specified (ignoring) [ 281.728] (EE) No input driver/identifier specified (ignoring)

[ 282.896] (EE) No input driver/identifier specified (ignoring) [ 1483.375] (II) FBDEV(0): FBIOBLANK: Invalid argument (Screen blanking not supported by vesafb of Linux Kernel) [ 1483.375] (II) FBDEV(0): FBIOBLANK: Invalid argument (Screen blanking not supported by vesafb of Linux Kernel) [ 1483.379] (II) FBDEV(0): FBIOBLANK: Invalid argument (Screen blanking not supported by vesafb of Linux Kernel) [ 1810.579] (II) FBDEV(0): FBIOBLANK: Invalid argument (Screen blanking not supported by vesafb of Linux Kernel) [ 1810.579] (II) FBDEV(0): FBIOBLANK: Invalid argument (Screen blanking not supported by vesafb of Linux Kernel)

Thank you, Bruce.

On Wednesday 19 January 2011 00:52:22 Markus Koßmann wrote,

...

...

What type of information should I collect to find the problem? I would really like to fix this.

Could you tell us more about the hardware (i.e graphics card) ? Are there any hints to the problem in the log files /var/log/Xorg.0.log and/or /var/log/messages ?

Are you running in a VirtualBox virtual machine? It is strange that otherwise you would be getting an error on loading the VirtualBox video driver. The next error you are getting is a failure to load the fallback vesa framebuffer driver. Possibly for some reason the new automatic driver detection by X is not working for you. In the past I had an RV350 card and it worked fine with both the ATI proprietary driver and radeon open source. Somebody help me out here: I know that when X was changed to do its automatic on-the-fly configuration, it would first use /etc/X11/xorg.conf if that file still existed. But didn't that get subsequently get changed to a rules file (maybe under modprobe.d)??? I don't remember offhand. Bruce, when you do an lsmod, what is the video driver that is loaded? You may need to force X to use to right driver. (Unless of course this is VBox, which will be a different problem.) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

Hi,

I am running this on an IBM T42 laptop. This is not an kind of virtual box. The video driver is just what the install tool picked out.

I don't think this is the right information from lsmod, it does not specify what the driver is. samhaber-nb:/var # lsmod | grep vid video 21141 0 output 2031 1 video thermal_sys 14735 4 video,fan,processor,thermal

Would it help to know what driver is from my SUSE 10.1 installation. Some of those files might still be valid if someone can tell me where to look.

Thanks, Bruce.

There are two open source drivers for ATI cards. The first is radeon, which is included with the kernel. The second is radeonhd, which is in the package xorg-x11-driver-video-radeonhd but I suspect this is for newer chipsets than you have. (It wouldn't hurt to install it anyway.) Do an lsmod without the qualifier, check for fbdev or vesa or vesafb. If you are getting a graphical display, there is some graphics driver running. In your log I would ignore the vboxvideo error. X now does automatic graphics device detection and driver selection. I haven't seen many of the new logs yet, but it is conceivable that X works through a list checking for the presence of possible driver candidates. vboxvideo is the driver used in a virtualbox guest, so in any case this obviously doesn't apply to you. The errors and info remarks in the log suggest to me that X couldn't find a driver and so it fell back to vesafb. While vesa is the actual X vesa driver, vesafb is the framebuffer driver compiled into the kernel; that's what enables the graphical splash. When run level 5 is reached and control of the display passes to X, it selects its own driver. If you don't find an X graphics driver being loaded, my speculation is that X is falling back to the kernel's. That would explain your getting some graphics but having problems, because the framebuffer is very limited. Recommendation: X still provides for the user to manually control the driver it uses. When the automatic change was first introduced, manual control was done through the traditional xorg.conf file, i.e., if that file was still present, it would be used. I finally remembered, the method was changed again: Look in the /etc/X11/xorg.conf.d directory. There you will find config files each of which corresponds to a section of the old xorg.conf file. Look at 50-device-conf. This file corresponds to the old "device" section, and is where the graphics driver can be specified. I noticed that there is even a remark in the file specific to the radeon driver. so here may be your problem and solution. Remove the comment # on the - Driver "radeon" - line, and restart the machine. Hopefully you'll get a gui and lsmod will show radeon loaded. Then take a look at the X log to determine what timings and resolution X chose. The timings are very important on laptops, particularly some older ones. If you need to set the timings manually, that's done in the file 50-monitor-conf. If you need to force the correct resolution too, that can be done in the file 50-screen-conf. Finally, there is also the ATI proprietary driver option. It's been too long since I've used that driver (worked fine with my RV350 card, btw), and there have been a lot of issues on and off with it over the years. If radeon or radeonhd don't work for you (and/or, if you want compositing, I don't think the open source drivers support that) I'd point you to the wiki and maybe the Forums for the most current info. The Forums deal with graphics driver issues all the time, so it's a good resource. Good luck. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On Sunday, January 23, 2011 21:56 dwgallien's reply (fixed the top-posting mess) wrote:

...

Hi,

I have installed OpenSuse 11.3 on two machines. The first has been running for almost a month now with no problems. The second hangs everytime I have Firefox or Chrome running. Sometimes I can browse the web for a few minutes before it hangs other times it just hangs on loading a default set of tabs. Both are using KDE as the window manager. Both machines previously had KDE3.? on them and now have KDE4. The machine that now hangs had some trouble with migrating the KDE3 settings to KDE4. When the machine hangs I even try CTRL-ALT-DEL and CTRL-ALT-2 but there is no response; needs a hard reset.

If I run the Kernel in failsafe mode then the machine does not hang; with both KDE normal or failsafe mode. If I run the Kernel in normal mode and either KDE normal or failsafe mode with Firefox or Chrome running the system will hang. If I run the Kernel in normal mode and KDE in normal mode then browse the web with Konqueror there was no problem.

What type of information should I collect to find the problem? I would really like to fix this.

I was thinking of deleting the .mozilla directory and re build from scratch, except for my bookmark file.

I think this is a mozilla browser problem but why does it hang the Kernel?

Thank you,

Pls take this with a large grain of salt, my memory may be faulty . . .

Are you using the nvidia proprietary driver? IIRC there have been issues reported with the repo driver and the stock kernel that in particular affected performance in Firefox and possibly Plasma (I don't recall anything reported re Chrome, however). The solution that some found to work was upgrading the kernel from :HEAD and then manually installing the newer nvidia driver downloaded from their site (which means you should first uninstall the driver from the repo via YaST or zypper). With some searching through this list you should easily find the thread(s) discussing this issue.

Good luck.

Btw, you should be able to re-install Firefox without any problem just renaming .mozilla, and then migrating back your bookmarks file to the newly created .mozilla. The bookmarks is just an xml file.

The unfortunate thing about downloading a whole new kernel is, it can't be done very easily, if at all, when one is on dial-up, so the 'fix' is pretty much non-existent. This kind of screw-up should not have been allowed to go into production until it was fixed first. Now those of us who *are* stuck on dial-up with no alternative have to fret over *when* the system is going to freeze up on us (or when it will do it again....and again...ad nauseum). It's happened to me already 4 times and there's no telling when it's going to happen, it just does it. BTW, it's doing it with Seamonkey too, not just Firefox. I use Seamonkey almost exclusively on my KDE. Hard rebooting is going to tear up some hard drives eventually, and it's something I can barely afford as is. This 'problem' of freezing up in 11.3, is leaving a very bad taste and I've been using SuSE since 7.3. :( -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

The unfortunate thing about downloading a whole new kernel is, it can't be done very easily, if at all, when one is on dial-up, so the 'fix' is pretty much non-existent.

This kind of screw-up should not have been allowed to go into production until it was fixed first. Now those of us who *are* stuck on dial-up with no alternative have to fret over *when* the system is going to freeze up on us (or when it will do it again....and again...ad nauseum). It's happened to me already 4 times and there's no telling when it's going to happen, it just does it. BTW, it's doing it with Seamonkey too, not just Firefox. I use Seamonkey almost exclusively on my KDE. Hard rebooting is going to tear up some hard drives eventually, and it's something I can barely afford as is. This 'problem' of freezing up in 11.3, is leaving a very bad taste and I've been using SuSE since 7.3. :(

It's not that simple. The thread I had in mind is "11.3 and very slow nvidia performance"; there were others, too. It appears that the problem can be card and user installation specific. Many if not most users have had no problem. Others' problem was resolved with an upgrade to the newer nvidia driver added to the repo, others switching to the nouveau driver (and losing compositing), others needed to download and manually install the driver from the nvidia website, others needed to upgrade to KDE 4.5, others needed a newer kernel version, others needed to add a kernel argument at boot, etc. The extent of aggravation ranged considerably, from zero to a solution relatively simple (once found) to others (such as having to download sources) obviously being prohibitive if one is on dial-up, is inexperienced, etc. IMHO, with the number of variables in the equation, and Linux distros not having the arbitrary control over 3rd-parties that Microsoft does (noting that, even so, this kind of thing periodically happens in the world of Windows, too), I'm afraid that incompatibility glitches are just going to occasionally happen and that it is impossible for distro testing to catch them all. Even when I worked in the world of large proprietary machines where we controlled *everything*, it was still standard customer practice to wait well past release before upgrading or even to wait until the upgrade thereafter was nearing release before doing the current upgrade (unless of course the new release included a needed critical feature or fix unavailable otherwise). And a production upgrade was never attempted before having done a thorough simulation. (Personally, I still follow this practice.) I don't mean to minimize the problems sometimes experienced with an upgrade, and I also realize that with personal systems there can be non-trivial issues that can make taking such precautions also difficult. But for some, waiting and/or long-term support is the best answer. Just my two cents . . . -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

On 2011/01/23 22:56 (GMT-0500) dwgallien composed:

...

...

Btw, you should be able to re-install Firefox without any problem just renaming .mozilla, and then migrating back your bookmarks file to the newly created .mozilla. The bookmarks is just an xml file.

When was the last time you examined your FF profile directory? FF hasn't routinely used xml for bookmarks for many many moons. Now it's in places.sqlite. Migration and sharing require importing/exporting operations. They can be exported to bookmarks.html, automatically even via user.js and/or about:config.

Oops, my bad. I deserve to be spanked for replying past my bed time. Of course, I meant html, not xml. That said, I'm not aware of a relationship with places.sqlite. Not that I'm questioning your accuracy, but in the past at least I have been able to restore a copy of bookmarks.html without any known problem. -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org

...

relationship with places.sqlite. Not that I'm questioning your accuracy, but in the past at least I have been able to restore a copy of bookmarks.html without any known problem.

Ancient past. Last Firefox release that used bookmarks.html routinely was 2.0.0.x. Places.sqlite merged history with bookmarks in FF 3.x. Restoration has since been more complicated than just copying back an old bookmarks.html file.

Thanks Felix, new data for me. You can call me "ancient" for short - I answer to all manner of names :) -- To unsubscribe, e-mail: opensuse+unsubscribe@opensuse.org For additional commands, e-mail: opensuse+help@opensuse.org