is there a fix for nvidia in hardy LTS?

Sorry it looks like I had more than just nv driver installed and working in Gusty (OpenGL version string: 1.5.8 NVIDIA 96.39 in glxinfo).
Also I managed to add generic nv to xorg.config file in Hardy (2.6.24.21rt) but nv default is using 800X600 resolution and lower versions of glx than nvidia in Gutsy of course. So still a problem.

Can you give me the output of:

lspci | grep nVidia

This should list the nvidia card... if not then do the output of:

lspci

This list will be much longer, but will include the video card and model.

Can you give me the output of:

lspci | grep nVidia

This should list the nvidia card... if not then do the output of:

lspci

This list will be much longer, but will include the video card and model.

as asked, but thought I already listed it, sorry:

code:

01:00.0 VGA compatible controller: nVidia Corporation NV25GL [Quadro4 900 XGL] (rev a3)

...and by the way, I must mention that with present setup, my screen is displayed off center- 1/2'' almost, to the right, so I have obvious problems using apps. that have a right scroll like firefox for example...
At least in Gutsy, the screen was detected and configured as "custom" by NVIDIA now it's "unknown" and misconfigured by all (vesa, nv default or restricted driver, I believe)!

You did list it before but I wanted to ensure that the xorg.conf was accurate.

Now that I am sure of that here is a thread post that should solve your problem:
http://ubuntuforums.org/showthread.php?t=970237

and another

http://ubuntuforums.org/showthread.php?t=971103

and one from the Nvidia forums

http://www.nvnews.net/vbulletin/showthread.php?t=122140

You may want to regenerate your xorg.conf:

sudo dpkg-reconfigure -phigh xserver-xorg

before trying all of these methods.

Hi!
I am away from home and thus cannot try any of these yet but before I do I would like to confirm the following:

Since the driver I need for my card is: " Version: 96.43.07

Operating System: Linux x86
Release Date: July 16, 2008

Release Highlights

    * Fixed a problem that caused TV output on secondary TVs to be black and white on some GPUs.
    * Improved nvidia-xconfig behavior when updating X configuration files without a "Module" section.
    * Restored compatibility with recent Linux 2.6 kernels."

Are these issues that nvidia report fixed with this driver the same ones I have been having? Sorry if it's obvious but I just want to be sure! By the way I don't think that drivers that work well with Ibex will work in Hardy, (diffrent modules and kernel image) isn't that correct? I want to stay with Hardy LTS.
Also I had read some posts advising to remove (purge) all nvidia related modules and apps. BEFORE installing the restricted driver manually. Is this necessary? And if so what X related modules should I still keep installed? Will regenerating my xorg.conf take care of all X related issues or should I check anything else in Synaptic or apt.?
Finally why did my xorg.config show the resolution I used to use (1024X768) when the included restricted driver was actually installed but I couldn't adjust my screen settings? (see above xorg.file)

"... "Section "Screen"
 Identifier "Default Screen"
 Device "nVidia Corporation NV25GL [Quadro4 900 XGL]"
 Monitor "Generic Monitor"
 Defaultdepth 24
 SubSection "Display"
  Modes "1024x768" "800x600"
 EndSubSection
EndSection... "

Thankyou

I tried the to install the 96.43.07 driver manually as instructed by nvidia readme and related posts; I removed conflicting modules and added "make" and other development modules in synaptic that the driver requires to build the corresponding module. Yet the driver failed to install!

This after I went to trouble trying to stop gdm as my screen will not switch to a console (ctrl alt f1-f6) or boot into one (installed rcconfig). So in the end I had to try nvidia driver install in recovery mode-dropped to shell. I am really frustrated with this.

I also tried to re-install the previous driver via envy but once again, this led me to a blank screen after reboot! I can't believe that I will be incapable of using this LTS version for basic multimedia production use! I cannot find another suitable video card for this computer easily and it is a good card, matching the computer in terms of "age". I don't understand how I can't just get the older nvidia driver to work (via envy) on this computer when the install worked on another cheaper, albeit slightly newer, computer with similar nvidia card-using the same envy nvidia driver successfully!

  And I am using identical monitors as well! The only difference is the rt kernel. NVIDIA did not mention any problems with this particular kernel version though...

I am supposed to send a copy of my log file to nvidia, but I don't know if they will find me a solution-I have never dealt with their bug forum before. Any suggestions? Pls!?

nvidia-installer log file '/var/log/nvidia-installer.log'
creation time: Thu Nov 27 00:55:19 2008
installer version: 1.0.7

option status:
  license pre-accepted : false
  update : false
  force update : false
  expert : false
  uninstall : false
  driver info : false
  precompiled interfaces : true
  no ncurses color : false
  query latest version : false
  OpenGL header files : true
  no questions : false
  silent : false
  no recursion : false
  no backup : false
  kernel module only : false
  sanity : false
  add this kernel : false
  no runlevel check : false
  no network : false
  no ABI note : false
  no RPMs : false
  no kernel module : false
  force SELinux : default
  no X server check : false
  no cc version check : false
  force tls : (not specified)
  X install prefix : (not specified)
  X library install path : (not specified)
  X module install path : (not specified)
  OpenGL install prefix : (not specified)
  OpenGL install libdir : (not specified)
  utility install prefix : (not specified)
  utility install libdir : (not specified)
  doc install prefix : (not specified)
  kernel name : (not specified)
  kernel include path : (not specified)
  kernel source path : (not specified)
  kernel output path : (not specified)
  kernel install path : (not specified)
  proc mount point : /proc
  ui : (not specified)
  tmpdir : /tmp
  ftp mirror : ftp://download.nvidia.com
  RPM file list : (not specified)

Using: nvidia-installer ncurses user interface
-> You appear to be running in runlevel 1; this may cause problems. For exampl
   e: some distributions that use devfs do not run the devfs daemon in runlevel
   1, making it difficult for `nvidia-installer` to correctly setup the kernel
   module configuration files. It is recommended that you quit installation no
   w and switch to runlevel 3 (`telinit 3`) before installing.

   Quit installation now? (select 'No' to continue installation) (Answer: No)
-> License accepted.
-> Installing NVIDIA driver version 96.43.07.
-> No precompiled kernel interface was found to match your kernel; would you li
   ke the installer to attempt to download a kernel interface for your kernel f
   rom the NVIDIA ftp site (ftp://download.nvidia.com)? (Answer: Yes)
-> No matching precompiled kernel interface was found on the NVIDIA ftp site;
   this means that the installer will need to compile a kernel interface for
   your kernel.
-> Performing CC sanity check with CC="cc".
-> Performing CC version check with CC="cc".
-> Kernel source path: '/lib/modules/2.6.24-21-rt/build'
-> Kernel output path: '/lib/modules/2.6.24-21-rt/build'
-> Performing rivafb check.
-> Performing nvidiafb check.
-> Performing Xen check.
-> Cleaning kernel module build directory.
   executing: 'cd ./usr/src/nv; make clean'...
-> Building kernel module:
   executing: 'cd ./usr/src/nv; make module SYSSRC=/lib/modules/2.6.24-21-rt/bu
   ild SYSOUT=/lib/modules/2.6.24-21-rt/build'...
   NVIDIA: calling KBUILD...
   make CC=cc KBUILD_VERBOSE=1 -C /lib/modules/2.6.24-21-rt/build SUBDIRS=/tmp
   /selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv modules
   test -e include/linux/autoconf.h -a -e include/config/auto.conf || ( \
    echo; \
    echo " ERROR: Kernel configuration is invalid."; \
    echo " include/linux/autoconf.h or include/config/auto.conf are mis
   sing."; \
    echo " Run 'make oldconfig && make prepare' on kernel src to fix it
   ."; \
    echo; \
    /bin/false)
   mkdir -p /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/.tmp_vers
   ions ; rm -f /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/.tmp_
   versions/*
   make -f scripts/Makefile.build obj=/tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07
   -pkg1/usr/src/nv
     cc -Wp,-MD,/tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/.nv.o
   .d -nostdinc -isystem /usr/lib/gcc/i486-linux-gnu/4.2.4/include -D__KERNEL_
   _ -Iinclude -include include/linux/autoconf.h -Wall -Wundef -Wstrict-proto
   types -Wno-trigraphs -fno-strict-ali
   asing -fno-common -Werror-implicit-function-declaration -O2 -pipe -msoft-flo
   at -mregparm=3 -freg-struct-return -mpreferred-stack-boundary=2 -march=i586
   -mtune=generic -ffreestanding -maccumulate-outgoing-args -Iinclude/asm-x86
   /mach-default -fomit-frame-pointer -g -fno-stack-protector -Wdeclaration-af
   ter-statement -Wno-pointer-sign -I/tmp/selfgz5338/NVIDIA-Linux-x86-96.43.0
   7-pkg1/usr/src/nv -Wall -Wimplicit -Wreturn-type -Wswitch -Wformat -Wchar-su
   bscripts -Wparentheses -Wpointer-arith -Wno-multichar -Werror -MD -Wsign-c
   ompare -Wno-cast-qual -Wno-error -D_LOOSE_KERNEL_NAMES -D__KERNEL__ -DMODULE
   -DNVRM -DNV_VERSION_STRING=\"96.43.07\" -UDEBUG -U_DEBUG -DNDEBUG -DMODULE
   -D"KBUILD_STR(s)=#s" -D"KBUILD_BASENAME=KBUILD_STR(nv)" -D"KBUILD_MODNAME=K
   BUILD_STR(nvidia)" -c -o /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/
   src/nv/.tmp_nv.o /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/n
   v.c
   In file included from include/linux/list.h:8,
                    from include/linux/preempt.h:11,
                    from include/linux/spinlock.h:85,
                    from include/linux/seqlock.h:29,
                    from include/linux/time.h:8,
                    from include/linux/timex.h:57,
                    from include/linux/sched.h:53,
                    from include/linux/utsname.h:35,
                    from /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src
   /nv/nv-linux.h:19,
                    from /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src
   /nv/nv.c:14:
   include/linux/prefetch.h: In function 'prefetch_range':
   include/linux/prefetch.h:57: warning: pointer of type 'void *' used in arith
   metic
   In file included from include/asm/dma-mapping_32.h:5,
                    from include/asm/dma-mapping.h:2,
                    from include/linux/dma-mapping.h:52,
                    from include/asm-generic/pci-dma-compat.h:7,
                    from include/asm/pci.h:88,
                    from include/linux/pci.h:796,
                    from /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src
   /nv/nv-linux.h:85,
                    from /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src
   /nv/nv.c:14:
   include/linux/scatterlist.h: In function 'sg_virt':
   include/linux/scatterlist.h:293: warning: pointer of type 'void *' used in a
   rithmetic
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c: In function
   '__nv_setup_pat_entries':
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:836: warning:
   comparison between signed and unsigned
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c: In function
   '__nv_restore_pat_entries':
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:862: warning:
   comparison between signed and unsigned
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c: In function
   'nv_kern_cpu_callback':
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:1188: warning
   : comparison between signed and unsigned
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:1195: warning
   : comparison between signed and unsigned
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c: In function
   'nv_lock_init_locks':
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:3517: error:
   implicit declaration of function '__SEMAPHORE_INITIALIZER'
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:3517: error:
   invalid initializer
   /tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.c:3518: error:
   invalid initializer
   make[3]: *** [/tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src/nv/nv.o
   ] Error 1
   make[2]: *** [_module_/tmp/selfgz5338/NVIDIA-Linux-x86-96.43.07-pkg1/usr/src
   /nv] Error 2
   NVIDIA: left KBUILD.
   nvidia.ko failed to build!
   make[1]: *** [module] Error 1
   make: *** [module] Error 2
-> Error.
ERROR: Unable to build the NVIDIA kernel module.
ERROR: Installation has failed. Please see the file
       '/var/log/nvidia-installer.log' for details. You may find suggestions
       on fixing installation problems in the README available on the Linux
       driver download page at www.nvidia.com.

** For the build problems please ensure you have your headers installed by running the following commands:

uname -r
sudo apt-get install linux-headers-'uname -r'
sudo apt-get install linux-restricted-modules-'uname -r'
sudo apt-get install linux-source-2.6.24> build-essential

killall gdm
chmod a+x /location/to/driver
sh /location/to/driver

nano /etc/default/linux-restricted-modules-common

DISABLED_MODULES="nv"

** For the boot problems try looking at these pages:
http://us.download.nvidia.com/XFree86/Linux-x86/173.14.12/README/appendix-h.html

http://us.download.nvidia.com/XFree86/Linux-x86/173.14.12/README/chapter-08.html

Thankyou for your ongoing support with this issue,

I will follow your above instructions and try to manually install the latest stable driver again-for this last round-hopefully! Just to assure success, I have a few questions for building the nvidia kernel module:

Does the apt-get install commands automatically select the right 'make' or 'gcc' version(s) (development tools); as I only had installed the latest version (2.XX) and perhaps NVIDIA requires version 1.XX or earlier (in Synaptic).

Also I had disabled "nv" but this time, should I add:

'DISABLED_MODULES="nv nvidia_new" ' as suggested in NVIDIA requirements?

Does the gdm command above switch me to runlevel 3 (or should I do '# cp /etc/inittab /etc/inittab.original' and ':3:initdefault:' first) and I will be able to work in console mode? (As I have mentioned before, my system setup seems to otherwise not allow me to boot or switch to console mode without X; I get a blank screen with cursor on top left, blinking but unresponsive (exept for Alt F7)).

I was pretty sure I had the right linux-headers installed and had uninstalled via apt command all things NVIDIA, as instructed in one post but...with one important exeption; 'NVDIA source restricted module' or 'NVDIA source common module' remained installed. (unfortunately, I am not at home again and I have no notes on this last detail, just going by memory...)

I am not 100% on the questions you have asked -- since I use the pre-compiled drivers that come with the distribution.

Hello!
This might be my last query!!!

I inadvertently came across someone from Canonical (go figure) in my area (of which I only just found out about) who suggested that my VGA extension could be enough to confound the nvidia drivers, of which I had no idea! I retried envyNG again (does not work) and next, re-installed the restricted driver (as you use) and this time Nvidia detected my screen. Or at least it detected a similar screen with 1024X768 resolution and I can already enable some enhanced effects in 'desktop'!

I still can't believe it and I hope that I can plug back into my needed extension without a problem. I was told to add 'gksudo' just in front of '/usr/bin/nvidia-settings' which is in the command box in 'nvidia x xerver settings' {properties}. Furthermore, I was advised to save to x config file when I found my desired resolution (and metamode) and to not merge this new server setting with the previous one.

The only hitch is that I still rebooted into the lower resolution but I selected the higher resolution in 'Screen Resolution' this time. I don't know if this will allow me to always reboot into my desired resolution. If that doesn't work is there another way to make the highest resolution the default?

When I tried to use 'Monitor Settings' I could not "access" or adjust my monitor. The message is: "The current monitor is in the database but does not support DDC/CI"; it mentions "plugging in the VGA cable on only one of the monitor's 2 inputs and not the one supporting DDC/CI. The fact that this screen happens to have a USB input for touchscreen capability usable only with a specific software (which I don't need or have connected) should be irrelevant since I should be changing my monitor settings via nvidia- settings instead, is this correct?

Well looks like everything is working so I assume I do my other monitor adjustments from Nvidia-Settings... and by the way I will keep a note of the commands you listed further back just in case I ever need to install the drivers manually in the future.

Thanks PrivateVoid, that solved my question.

So I just want to add a thanks to you (PrivateVoid) because you kept refering to the restricted drivers; and to all those who helped with the build/including of those drivers which (pre-compiled) reduce alot of hassle! EnvyNG did work right for me on my other computer (although I had to manually check the right drivers) and both options are really encouraging and simple for any linux newbie. I just wish I had known about not having anything between the monitor plug and the computer, I don't get how an extension is even "seen" by a driver...Well, I'll be more vigilant with any new hardware!