Stop Microsoft
Operating Systems => Linux and UNIX => Topic started by: Siplus on 9 December 2002, 04:31
-
after using linux for about 6 weeks now, and i'll never go back to windows!! just a few days ago, i went in to play a game, and it locked up after doing nothing, so i was going to open up a console and type 'killall -9 [whatever the game was]', untill i realized that i didn't have the console (i guess that means i'm converted...)
anyway, i only have a few problems w/ linux, but most are b/c i don't have the knowledge i need. the one problem i can't stand is that while i'm in GNOME, sometimes the dialog boxes and/or windows will just be grey when i switch between virtual desktops, and they stay that way for sometimes a few minutes. this is not cool. does anyone know why these windows are not displaying the information for (what seems like) long periods of time? i don't think i've seen this happen in windows *cringe...evil thought, AHH*
-
use KDE. (i dont know the answer, but i havent had this in KDE or Enlightenment)
-
lag in the gui toolkit. its a design issue/*threads are you friends*/
-
Hmmm, I haven't seen that behavior. Maybe it's because I am not running the stock kernel and am using the nVidia accelerated video drivers and have lotsa RAM and a fast processor. I know with the stock kernel I did have strange problems when using VMware but upgrading the kernel and not using the Red Hat patches fixed that. And I didn't know you could switch between virtual desktops on Windows (without adding 3rd party software that is).
[ December 08, 2002: Message edited by: void main ]
-
well, it also screws up when i goto another app in linux like cliking on something else in the taskbar or minimizing something...
but anyway, i am not use the stock kernal, have a AMD 1.1, and 384mb ram, i just didn't dl the nvidia drivers cause i don't know how to compile them. i looked at the instructions b4 i had linux, but none of it made sence and haven't looked at it since
-
If you have an nVidia card I would *highly* recommend installing those drivers. It makes a huge difference. The nVidia code included with stock XFree sucks because nVidea is fairly tight lipped when it comes to giving out details on their cards.
You basically download two RPMs and install them and make one simple little change to your /etc/X11/XF86Config. Once you do it you'll be pissed at yourself for not having done it sooner. What model is your card, I can check and make sure it's supported by the driver. Also is your AMD chip an Athlon? And you say you are not running the stock kernel. What kernel are you running then?
[ December 08, 2002: Message edited by: void main ]
-
ok, i'm not at my house right now, i'm in my programming class, so i'll check and edit this post when i get back:
Athlon 1.1ghz i think
Nvidia Geforce 4 ti 4100, 128 mb ram, AGP
384 mb of system ram
kernal=2.4.18-18.8.0 (checked and this is the one i'm running)
edit:
thanks for the help void main. if you need any more info, just ask. also: when i was at the nvidia site last time, i didn't see any rpms, where would i find the driver rpms?
[ December 09, 2002: Message edited by: Siplus ]
-
If you use Red Hat 8.0 you won't find any pre-made rpm's, you'll have to download src.rpm's and build them yourself.
It's not difficult at all, even I could do it.
Go here http://www.osnews.com/story.php?news_id=1890&page=2 (http://www.osnews.com/story.php?news_id=1890&page=2)
and follow the instructions, you'll be set in about 5 minutes.
Edit: typos
[ December 09, 2002: Message edited by: Ice9 ]
-
It's even easier than that article shows. Instead of doing the "rpm -i xxx.rpm", changing directories and doing the "rpmbuild -bb xxx.spec" you can do it with one caommand "rpmbuild --rebuild xxx.rpm" which will unpack and build the binary RPM automatically. Then just install the resulting NV* binaries (from /usr/src/redhat/RPMS/i386 directory). But of course you still have to make the minor change to the XF86Config file.
-
i went to the site ice9 gave a link to, dl the 2 files, and tried it uptill this happened:
quote:
NVIDIA_GLX-1.0-3123.src.rpm (at top of page)
NVIDIA_kernel-1.0-3123.src.rpm (at bottom of page)
Save these files to your machine. Now, open up a terminal (konsole, gnome-terminal) and become root (also called superuser). Change to the directory containing the packages and install them with:
cd /path/to/packages
rpm -i NVIDIA* (or rpm -ivh NVIDIA* for verbose output and to print hash marks - indicating installation progress)
but when i did it, i got this
any thoughts?
-
Looks like your src.rpm file did not transfer correctly. Try downloading it again. And instead of using the wildcard (*) in your rpm command try each file individually by typing:
# rpmbuild --rebuild NVIDIA_GLX-1.0-3123.src.rpm
# rpmbuild --rebuild NVIDIA_kernel-1.0-3123.src.rpm
which should create the binary RPMs in the "/usr/src/redhat/RPMS/i386" directory. Once you get the binary RPMs install them by:
# rpm -Uvh /usr/src/redhat/RPMS/i386/NV*
Then just change "nv" to "nvidia" in your /etc/X11/XF86Config and when you restart X it should load with the nVidia drivers.
-
ya, i guess if a something downloads at 666 bytes/sec, there *might* be something wrong w/ it...
-
quote:
Originally posted by Siplus:
ya, i guess if a something downloads at 666 bytes/sec, there *might* be something wrong w/ it...
Is this over dialup or an Ethernet interface? If this is over your Ethernet interface (via Cable/DSL, etc) does it also exhibit this slow/erratic behavior to any server, even on your local network? If so, this may also be the root cause of your erratic X operation. X is highly network dependent. Especially if you don't have it configured optimally. If this is the case you should try adding your local host name at the end of the localhost line in your /etc/hosts and see if that helps solve your X problems.
For instance, if your computer name is "siplus" then your localhost line in /etc/hosts would look like this:
127.0.0.1 localhost.localdomain localhost siplus
Once you make this change you might have to restart X (although I don't believe it is necessary).
And of course if that seems to fix or help your problem you might want to figure out what is wrong with your network. Whether it's driver or hardware related...
[ December 09, 2002: Message edited by: void main ]
-
[ December 09, 2002: Message edited by: Siplus ]
-
quote:
Originally posted by Siplus:
nah, i have dial-up through AT&T and i once in awhile my connection goes real slow (although 666 bytes/sec is byfar the slowest i have ever seen it short of not dling at all)
but back to the driver install, you said:
# rpmbuild --rebuild NVIDIA_GLX-1.0-3123.src.rpm
# rpmbuild --rebuild NVIDIA_kernel-1.0-3123.src.rpm
which should create the binary RPMs in the "/usr/src/redhat/RPMS/i386" directory. Once you get the binary RPMs install them by:
# rpm -Uvh /usr/src/redhat/RPMS/i386/NV*
[ December 09, 2002: Message edited by: Siplus ][/b]
-
[ December 09, 2002: Message edited by: Siplus ]
-
Did you successfully rebuild the NVIDIA kernel RPM? If it wasn't successful and you don't have both the kernel and GLX binary RPMs you don't want to install either. Figure out what happened on the NVIDIA kernel driver build.
-
hmm, now that i look back in the terminal i used to build what i dled, i'm seeing a lot of what looks to me like errors, so i'm jsut going to start completely over and redl everything
-
If they are compile errors then redownloading will not help. You already have what is necessary. Can you post a copy of the error messages (ignore any warnings, just errors).
-
just the last few lines in my terminal:
-
Gonna need more of the log. I need to see the first place things go wrong.
-
ok, umm. i'll go up as far as my terminal will go, i just hope it isn't too long in the forum:
edit: i'm taking out the huge terminal output i posted cause it isn't good enough, i'm going to make a log file and put it on my site if anyone else would for some reason want to look at it
[ December 09, 2002: Message edited by: Siplus ]
-
Nope not far enough back. Why don't you run this command:
# rpmbuild --rebuild NVIDIA_kernel-1.0-3123.src.rpm > /tmp/nv.log 2>&1
And then email me the resulting /tmp/nv.log file at [email protected]. I'll look it over.
Also do you have your kernel headers and/or the kernel source installed? What do you get for output from the command:
$ rpm -qa | grep kern
[ December 09, 2002: Message edited by: void main ]
-
ok, i removed that huge post of mine, i e-mailed the log to you void, and posted it on my webpage:
http://www.geocities.com/siplus01/linux/nvidia_driver_log.html (http://www.geocities.com/siplus01/linux/nvidia_driver_log.html)
(i would suggest looking at the source, cause i don't feel like formatting the entire log file (http://smile.gif) )
as for $ rpm -qa | grep kern,
-
According to your log it can not find <linux/vmalloc.h>. I do have that file but I have the kernel source. I also notice you have more than one kernel RPM installed which I am sure is screwing it up. You have a mish mash of kernels with non matching headers. You are going to hate me for this but I would suggest downloading and installing the source for the current kernel version you are running and then the NVIDIA kernel RPM should compile and install. Either that or reinstall the original kernel for RedHat 8.0 (kernel-2.4.18-14.i686.rpm from your CD) and then click on the "Kernel Development" in the "Packages" utility.
[ December 09, 2002: Message edited by: void main ]
-
erm....plz excuse my noobness once again...
i installed the 'kernal development' package, and found the original kernal rpm, but:
1) how to i force the rpm to reinstall??
2) how do i use the kernal development package?
thanks...once again...
-
It appears that your 2.4.18-14 kernel is already installed on your system. You might have it on your boot menu. Just set that one to be the default in your /etc/grub.conf. When you boot that kernel (2.4.18-14), and if you have the Kernel Development checked (which matches that kernel version) in your Software Package tool you should be all set. Then all you have to do is run those rpmbuild commands on your NVIDIA RPMs.
-
so i should always bootup in the stock kernal? if i do boot into the stock k, and make the driver updates, will it still be updated in the newer kernals (i can't see why it wouldn't be....)
also, what would these newer kernals that i dled from the rhn do differently then the stock? are they more stable? (i can't imagine that much improvment, i've never had the stock kernal lock up on me, but my newest one has when i was trying to get winex to work (which i still don't have games working))
-
quote:
It's not difficult at all, even I could do it.
Go here http://www.osnews.com/story.php?news_id=1890&page=2 (http://www.osnews.com/story.php?news_id=1890&page=2)
and follow the instructions, you'll be set in about 5 minutes.
Ice9 thanks for posting that link... been trying to figure this out for ages! :D
Also thanks very much to everyone else that posted thus helping me fix my graphics support... www.mandrakeexpert.com (http://www.mandrakeexpert.com) is no help at all
-
quote:
Originally posted by Siplus:
so i should always bootup in the stock kernal? if i do boot into the stock k, and make the driver updates, will it still be updated in the newer kernals (i can't see why it wouldn't be....)
I personally like the 2.4.20 kernel compiled from source, without all the Red Hat packages. It's up to you which one you use, but you need the source for whichever one you use in order to compile any modules for that kernel.
quote:
also, what would these newer kernals that i dled from the rhn do differently then the stock? are they more stable? (i can't imagine that much improvment, i've never had the stock kernal lock up on me, but my newest one has when i was trying to get winex to work (which i still don't have games working))
No, kernel modules must match the running kernel exactly. If you will notice you have a directory called "/lib/modules". Under that directory there will be a subdirectory or subdirectories. You should have a subdirectory for each kernel version you have. For instance if you are running the stock kernel, it will use modules from under the "/usr/lib/2.4.18-14" directory structure. You will not be able to use modules compiled for the 2.4.18-14 kernel with the 2.4.19 kernel.
So if you want to be able to boot all three of those kernels you have installed, and use the NVIDIA kernel modules on all of them you will need a copy of those modules for every kernel version you have. Now, NVIDIA *does* include these modules in binary form which would make this compiling process mute, but only for kernels included with Red Hat 7.3 and below (which all use different kernels).
If you want a nice running kernel that should speed up your system somewhat I would suggest getting the source for 2.4.20 and build a nice new kernel without the Red Hat patches according to my HOWTO on this:
http://voidmain.kicks-ass.net/redhat/redhat_8_kernel_2.4.20_from_source.html (http://voidmain.kicks-ass.net/redhat/redhat_8_kernel_2.4.20_from_source.html)
After following the directions for installing this kernel boot it, make sure it works, set it as the default kernel, then follow the instructions for installing the NVIDIA RPMS which will work because you have the kernel source code installed that matches the running kernel.
But if you don't want to be bothered with building a new kernel or downloading the kernel source then you will have to revert back to your stock kernel, install the kernel development packages from your CD (using the package tool on the menu). Then you can compile the modules.
The only other option is to find someone who has the same kernel installed that you have and get them to create the binary modules and give you instructions on how to install those binary modules manually.
Hope that helps...
-
ok, i'm starting to understand this more. thanks for your time
-
Wow man, sorry you went through such a pain :(
-
hmm...it's been a little while since i've done anything w/ my video drivers, i guess b/c while the boards were down i didn't have this link:
http://voidmain.kicks-ass.net/redhat/redhat_8_kernel_2.4.20_from_source.html (http://voidmain.kicks-ass.net/redhat/redhat_8_kernel_2.4.20_from_source.html)
but after i dled it (for about 2 hours), i followed the easy instructions, but i got an error when i did "make xconfig"
[ December 18, 2002: Message edited by: Siplus ]
-
You must not have followed the instructions fully. The first paragraph (starting with "Let us begin:") states you need to have the "Kernel Development" checked in your "Packages" utility on your menu. This will install "Tk" which contains "wish" which is what the error is complaining that is missing.