Can not compile Nvidia driver in BookwormPup64 10.0.6

[bigpup]

A frugal install using a save folder.

The nouveau driver is blacklisted so it boots using the Vesa driver. (required to be able to compile a Nvidia driver)

The devx and kernel sources SFS's are located at /mnt/home/

First getting the devx loaded.
It will not load using the SFS load GUI program.
In the program it is listed with all the other BookwormPup64 SFS's, which are all grayed out and cannot be selected.
To get the devx sfs loaded you have to navigate to the file click on it and select install.

The kernel sources SFS does load by using the SFS load GUI.

This is list of what SFS's are loaded:

▶—— INSTALLED/MOUNTED SFS FILES (losetup) ——◀

adrv_dpupbw64_10.0.6.sfs
bdrv_dpupbw64_10.0.6.sfs
devx_dpupbw64_10.0.6.sfs
fdrv_dpupbw64_10.0.6.sfs
kbuild-6.1.76.sfs
kernel_sources-6.1.76-dpupbw64.sfs
puppy_dpupbw64_10.0.6.sfs
zdrv_dpupbw64_10.0.6.sfs

When you start the Nvidia driver run package to start the process of compiling the driver.
It all seems to work OK until it gets to the end of the process of compiling.

It ends with error message and says the driver modules did not get compiled.

In the Nvidia installer.log these are the errors:

ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_lock'
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_unlock'
make[2]: *** [scripts/Makefile.modpost:126: /initrd/mnt/tmpfs/tmp/selfgz16508/NVIDIA-Linux-x86_64-525.85.05/kernel/Module.symvers] Error 1
make[2]: Target 'modpost' not remade because of errors.
make[1]: *** [Makefile modpost] Error 2
make[1]: Target 'modules' not remade because of errors.
make[1]: Leaving directory '/usr/src/kbuild-6.1.76'
make: *** [Makefile:82: modules] Error 2
-> Checking to see whether the nvidia kernel module was successfully built
executing: 'cd ./kernel; /bin/make -k -j20 NV_EXCLUDE_KERNEL_MODULES="" SYSSRC="/lib/modules/6.1.76/source" SYSOUT="/lib/modules/6.1.76/build" NV_KERNEL_MODULES="nvidia"'...
make[1]: Entering directory '/usr/src/kbuild-6.1.76'
warning: the compiler differs from the one used to build the kernel
The kernel was built by: gcc (Debian 12.2.0-14) 12.2.0
You are using: cc (Debian 12.2.0-14) 12.2.0
MODPOST /initrd/mnt/tmpfs/tmp/selfgz16508/NVIDIA-Linux-x86_64-525.85.05/kernel/Module.symvers
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_lock'
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_unlock'
make[2]: *** [scripts/Makefile.modpost:126: /initrd/mnt/tmpfs/tmp/selfgz16508/NVIDIA-Linux-x86_64-525.85.05/kernel/Module.symvers] Error 1
make[2]: Target 'modpost' not remade because of errors.
make[1]: *** [Makefile modpost] Error 2
make[1]: Target 'modules' not remade because of errors.
make[1]: Leaving directory '/usr/src/kbuild-6.1.76'
make: *** [Makefile:82: modules] Error 2
-> Error.
ERROR: An error occurred while performing the step: "Checking to see whether the nvidia kernel module was successfully built". See /var/log/nvidia-installer.log for details.
-> The command cd ./kernel; /bin/make -k -j20 NV_EXCLUDE_KERNEL_MODULES="" SYSSRC="/lib/modules/6.1.76/source" SYSOUT="/lib/modules/6.1.76/build" NV_KERNEL_MODULES="nvidia" failed with the following output:

make[1]: Entering directory '/usr/src/kbuild-6.1.76'
warning: the compiler differs from the one used to build the kernel
The kernel was built by: gcc (Debian 12.2.0-14) 12.2.0
You are using: cc (Debian 12.2.0-14) 12.2.0
MODPOST /initrd/mnt/tmpfs/tmp/selfgz16508/NVIDIA-Linux-x86_64-525.85.05/kernel/Module.symvers
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_lock'
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_unlock'
make[2]: *** [scripts/Makefile.modpost:126: /initrd/mnt/tmpfs/tmp/selfgz16508/NVIDIA-Linux-x86_64-525.85.05/kernel/Module.symvers] Error 1
make[2]: Target '__modpost' not remade because of errors.
make[1]: *** [Makefile modpost] Error 2
make[1]: Target 'modules' not remade because of errors.
make[1]: Leaving directory '/usr/src/kbuild-6.1.76'
make: *** [Makefile:82: modules] Error 2
ERROR: The nvidia kernel module was not created.
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.

Any idea on what is needed or wrong?

Have been able to compile a Nvidia driver in earlier versions of BookwormPup64 10.0.3 and 10.0.4

Yes I have tried this with several Nvidia driver run packages.
Including the one that did compile in earlier BookwormPup64 versions.
All give these same errors.

[dimkr]

You're far more likely to get a working nvidia kernel driver by installing nvidia-driver-dkms, rather than trying to install the latest from the vendor website. Debian 12 sticks to kernel 6.1.x, which gets older and older, while the latest nvidia drivers target newer kernels.

[ozsouth]

@bigpup - if no joy with dkms, newer 6.6 kernel nouveau driver worth a try:

If you swap to a 6.6 ursmerge kernel, you may find, as mikewalsh did, his nvidia ran fine with nouveau driver for that kernel series.
I made a 6.6.13 usrmerge kernel you may wish to try. See: viewtopic.php?p=110442#p110442
If you do try it, must disable kbuild... .sfs (I add .no to end), & delete ...initmodules.txt if exists.

[bigpup]

Thanks for the ideas!

I will try them.

However the kernel being used really has never had anything to do with compiling a Nvidia driver.
The Nvidia driver run packages seem to only need the software that is required to compile.
What kernel is used has never been an issue.
I have seen the age (version) of the dependency software being an issue.
The Nvidia run package is needing newer versions of the dependency software.

About the nouveau driver.
If it is properly installed it will just work.
The Linux kernel I have never seen affect it working.
If the kernel is configured correctly to be able to use the nouveau driver or allow it to be used.

BookwormPup64 10.0.6 does work using the nouveau driver. That is the default for Nvidia hardware.
However it does not provide full support for all the features the Nvidia hardware may have.

[bigpup]

You're far more likely to get a working nvidia kernel driver by installing nvidia-driver-dkms, rather than trying to install the latest from the vendor website. Debian 12 sticks to kernel 6.1.x, which gets older and older, while the latest nvidia drivers target newer kernels.

If doing this has worked for you.
If anyone has done this and it works.

How did you do it?????

Never has made a difference for me.
But not sure I did what is really needed.

All I have found is the driver is compiled too much for Debian Linux and installs scripts that do nor work correctly for BookwormPup64 and how graphics is controlled in it.

[radky]

@bigpup

RE: failed compile of nvidia drivers with kernel 6.1.76

Code: Select all

ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_lock'
ERROR: modpost: GPL-incompatible module nvidia.ko uses GPL-only symbol 'rcu_read_unlock'

In each release of BW64 the security and kernel updates follow those of upstream Debian point releases (currently Debian 12.5 released 10 Feb 2024). The initial Debian 12.5 and BW64 10.6 point releases are based on kernel 6.1.76.

There are multiple online reports of failure to compile nvidia drivers against the 6.1.76 kernel (with associated 'rcu_read_lock' and 'rcu_read_unlock' errors). In upstream Debian, this error was reported with the nvidia-driver (525.147.05-4_deb12u1) available for the initial release of Bookworm 12.5. Subsequently, the Debian maintainers have upgraded to nvidia-driver (525.147.05-7_deb12u1) (available via bookworm-updates non-free) which fixes the issue for users of Debian Linux.

https://packages.debian.org/bookworm-up ... dia-driver

From your error report, it seems you are compiling (with errors) an older nivida driver (NVIDIA-Linux-x86_64-525.85.05) released 19 Jan 2023. I don't know if the Nivida website provides updated sources (appropriate for your video card) that would compile successfully against the 6.1.76 kernel.

Currently, there are several online patches (provided by nvida users) that one can apply to the nvidia driver sources to fix this issue, but I don't know if the patches are available for the specific nvidia driver required for your video card, or if the patches are even safe.

Lastly, upstream kernel developers are constantly updating code (or fixing regressions). Maybe future kernel releases will support successful builds of your nvidia driver.

[bigpup]

Thanks for finding that information!!

That nvidia-driver (525.147.05-7 is one I have not tried.

But I have tried to use newer ones than that one. 550 series that do support my hardware.
Still got errors compiling.

I need to carefully read what you posted and see what it does to help me.

Maybe they just need to make the universal nouveau driver provide more feature support for the hardware.
Mainly i need the control of the Nvidia card fans that is provided by the Nvidia driver settings controller.
With it you can setup exactly how the fans operate.
Also a lot of other features the Nvidia hardware has.

[dimkr]

If doing this has worked for you.
If anyone has done this and it works.

How did you do it?????

I didn't do anything special, only installed nvidia-driver with apt: it installs everything needed and successful builds a loadable kernel module (unlike your installation procedure - I don't know what it is). I just don't have a supported GPU to load the driver and see if it works, but it should once nouveau is blacklisted.

[Clarity]

... I just don't have a supported GPU to load the driver and see if it works, but ...

Would you post your driver please. Its a solution we can try. We, on the forum, can test and report use instructions for BKWP64.

Further @radky if its works in our tests, could you envision it as a selectable Menu item or something similar to how @666philb had it for FossaPUP64.

Thanks in advance

[dimkr]

Would you post your driver please.

You'll get the same driver (and all dependencies, and the nvidia settings panel) if you install nvidia-driver.

[bigpup]

If doing this has worked for you.
If anyone has done this and it works.

How did you do it?????

I didn't do anything special, only installed nvidia-driver with apt: it installs everything needed and successful builds a loadable kernel module (unlike your installation procedure - I don't know what it is). I just don't have a supported GPU to load the driver and see if it works, but it should once nouveau is blacklisted.

Sorry to say this DOES NOT WORK!
It builds the driver module.

But it does produce and installs a lot of other files that do not work in BookwormPup64.

But they are in locations that activate them to do something during boot up.

Most of them are a huge amount of symbolic links to other symbolic links to other symbolic links and finally to a file that is suppose to do something needed.

The final step in doing the install is to reboot the computer to have it not use the nouveau driver and change to using the Nvidia driver.

All these additional files are suppose to make that happen, BUT ALL IT DOES IS MESS UP STARTING X SERVER.

It boots to needing to run xorgwizard.
Nothing you do in xorgwizard gets it to boot x server.
You cannot even get it to use the Vesa driver.

So that is where boot process stops dead!

Here is what is added for files:
Different locations in /etc/

Note:
This image is from a working bookworm accessing what is in the bad save.
So the icons are going to be wrong, because they do not apply to what is running.
.

Screenshot.jpg (82.16 KiB) Viewed 1244 times

.

Screenshot(1).jpg (15.96 KiB) Viewed 1244 times

.

Screenshot(2).jpg (21.35 KiB) Viewed 1244 times

.

Screenshot(3).jpg (17.1 KiB) Viewed 1244 times

[dimkr]

ALL IT DOES IS MESS UP STARTING X SERVER.

What does /var/log/Xorg.0.log say?

[bigpup]

Not sure but this may have info where I tried to use xorgwizard to select a driver to get x server to start.

Code: Select all

[   362.459] 
X.Org X Server 1.21.1.7
X Protocol Version 11, Revision 0
[   362.459] Current Operating System: Linux puppypc18974 6.1.76 #1 SMP PREEMPT_DYNAMIC Thu Feb 15 01:58:06 UTC 2024 x86_64
[   362.459] Kernel command line: BOOT_IMAGE=/bookwormpup641006/vmlinuz pmedia=usbhd pdrv=b0f092c5-6a1c-4e04-b9d8-ddbeb91d4ab7 psubdir=/bookwormpup641006 pfix=fsck,fsckp,trim TZ=EST5EDT,M3.2.0,M11.1.0
[   362.459] xorg-server 2:21.1.7-3+deb12u5 (https://www.debian.org/support) 
[   362.459] Current version of pixman: 0.42.2
[   362.459] 	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
[   362.459] Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
[   362.459] (==) Log file: "/var/log/Xorg.0.log", Time: Sun Mar 17 00:39:51 2024
[   362.459] (==) Using config file: "/etc/X11/xorg.conf"
[   362.459] (==) Using config directory: "/etc/X11/xorg.conf.d"
[   362.459] (==) Using system config directory "/usr/share/X11/xorg.conf.d"
[   362.460] (==) ServerLayout "X.org config"
[   362.460] (**) |-->Screen "Screen0" (0)
[   362.460] (**) |   |-->Monitor "Monitor0"
[   362.460] (**) |   |-->Device "Card0"
[   362.460] (**) Option "DontZap" "false"
[   362.460] (**) Option "AutoAddDevices" "true"
[   362.460] (**) Automatically adding devices
[   362.460] (==) Automatically enabling devices
[   362.460] (==) Automatically adding GPU devices
[   362.460] (==) Automatically binding GPU devices
[   362.460] (==) Max clients allowed: 256, resource mask: 0x1fffff
[   362.461] (WW) The directory "/usr/share/X11/fonts/Type1/" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) `fonts.dir' not found (or not valid) in "/usr/share/X11/fonts/TTF/".
[   362.461] 	Entry deleted from font path.
[   362.461] 	(Run 'mkfontdir' on "/usr/share/X11/fonts/TTF/").
[   362.461] (WW) The directory "/usr/share/fonts/X11/misc" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) The directory "/usr/share/fonts/X11/cyrillic" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) The directory "/usr/share/fonts/X11/100dpi/" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) The directory "/usr/share/fonts/X11/75dpi/" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) `fonts.dir' not found (or not valid) in "/usr/share/fonts/X11/Type1".
[   362.461] 	Entry deleted from font path.
[   362.461] 	(Run 'mkfontdir' on "/usr/share/fonts/X11/Type1").
[   362.461] (WW) The directory "/usr/share/fonts/X11/100dpi" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (WW) The directory "/usr/share/fonts/X11/75dpi" does not exist.
[   362.461] 	Entry deleted from font path.
[   362.461] (**) FontPath set to:
	/usr/share/X11/fonts/misc/,
	built-ins
[   362.461] (==) ModulePath set to "/usr/lib/xorg/modules"
[   362.461] (II) The server relies on udev to provide the list of input devices.
	If no devices become available, reconfigure udev or disable AutoAddDevices.
[   362.461] (II) Loader magic: 0x561611d8ff00
[   362.461] (II) Module ABI versions:
[   362.461] 	X.Org ANSI C Emulation: 0.4
[   362.461] 	X.Org Video Driver: 25.2
[   362.461] 	X.Org XInput driver : 24.4
[   362.461] 	X.Org Server Extension : 10.0
[   362.462] (--) using VT number 3

[   362.462] (II) systemd-logind: logind integration requires -keeptty and -keeptty was not provided, disabling logind integration
[   362.463] (II) xfree86: Adding drm device (/dev/dri/card0)
[   362.463] (II) Platform probe for /sys/devices/pci0000:00/0000:00:02.0/drm/card0
[   362.466] (--) PCI: (0@0:2:0) 8086:4680:1462:7d25 rev 12, Mem @ 0x71000000/16777216, 0x80000000/268435456, I/O @ 0x00005000/64
[   362.466] (--) PCI:*(1@0:0:0) 10de:1f08:1462:3752 rev 161, Mem @ 0x72000000/16777216, 0x90000000/268435456, 0xa0000000/33554432, I/O @ 0x00004000/128, BIOS @ 0x????????/131072
[   362.466] (II) "glx" will be loaded by default.
[   362.466] (II) LoadModule: "dbe"
[   362.466] (II) Module "dbe" already built-in
[   362.466] (II) LoadModule: "glx"
[   362.466] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[   362.467] (II) Module glx: vendor="X.Org Foundation"
[   362.467] 	compiled for 1.21.1.7, module version = 1.0.0
[   362.467] 	ABI class: X.Org Server Extension, version 10.0
[   362.588] (==) Matched nouveau as autoconfigured driver 0
[   362.588] (==) Matched nv as autoconfigured driver 1
[   362.588] (==) Matched modesetting as autoconfigured driver 2
[   362.588] (==) Matched fbdev as autoconfigured driver 3
[   362.588] (==) Matched vesa as autoconfigured driver 4
[   362.588] (==) Assigned the driver to the xf86ConfigLayout
[   362.588] (II) LoadModule: "nouveau"
[   362.588] (II) Loading /usr/lib/xorg/modules/drivers/nouveau_drv.so
[   362.588] (II) Module nouveau: vendor="X.Org Foundation"
[   362.588] 	compiled for 1.21.1.3, module version = 1.0.17
[   362.588] 	Module class: X.Org Video Driver
[   362.588] 	ABI class: X.Org Video Driver, version 25.2
[   362.588] (II) LoadModule: "nv"
[   362.589] (WW) Warning, couldn't open module nv
[   362.589] (EE) Failed to load module "nv" (module does not exist, 0)
[   362.589] (II) LoadModule: "modesetting"
[   362.589] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[   362.589] (II) Module modesetting: vendor="X.Org Foundation"
[   362.589] 	compiled for 1.21.1.7, module version = 1.21.1
[   362.589] 	Module class: X.Org Video Driver
[   362.589] 	ABI class: X.Org Video Driver, version 25.2
[   362.589] (II) LoadModule: "fbdev"
[   362.589] (II) Loading /usr/lib/xorg/modules/drivers/fbdev_drv.so
[   362.589] (II) Module fbdev: vendor="X.Org Foundation"
[   362.589] 	compiled for 1.21.1.3, module version = 0.5.0
[   362.589] 	Module class: X.Org Video Driver
[   362.589] 	ABI class: X.Org Video Driver, version 25.2
[   362.589] (II) LoadModule: "vesa"
[   362.590] (II) Loading /usr/lib/xorg/modules/drivers/vesa_drv.so
[   362.590] (II) Module vesa: vendor="X.Org Foundation"
[   362.590] 	compiled for 1.21.1.3, module version = 2.5.0
[   362.590] 	Module class: X.Org Video Driver
[   362.590] 	ABI class: X.Org Video Driver, version 25.2
[   362.590] (II) NOUVEAU driver Date:   Sat Jan 23 12:24:42 2021 -0500
[   362.590] (II) NOUVEAU driver for NVIDIA chipset families :
[   362.590] 	RIVA TNT            (NV04)
[   362.590] 	RIVA TNT2           (NV05)
[   362.590] 	GeForce 256         (NV10)
[   362.590] 	GeForce 2           (NV11, NV15)
[   362.590] 	GeForce 4MX         (NV17, NV18)
[   362.590] 	GeForce 3           (NV20)
[   362.590] 	GeForce 4Ti         (NV25, NV28)
[   362.590] 	GeForce FX          (NV3x)
[   362.590] 	GeForce 6           (NV4x)
[   362.590] 	GeForce 7           (G7x)
[   362.590] 	GeForce 8           (G8x)
[   362.590] 	GeForce 9           (G9x)
[   362.590] 	GeForce GTX 2xx/3xx (GT2xx)
[   362.590] 	GeForce GTX 4xx/5xx (GFxxx)
[   362.590] 	GeForce GTX 6xx/7xx (GKxxx)
[   362.590] 	GeForce GTX 9xx     (GMxxx)
[   362.590] 	GeForce GTX 10xx    (GPxxx)
[   362.590] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[   362.590] (II) FBDEV: driver for framebuffer: fbdev
[   362.591] (II) VESA: driver for VESA chipsets: vesa
[   362.724] (EE) [drm] Failed to open DRM device for pci:0000:01:00.0: -19
[   362.724] (WW) Falling back to old probe method for modesetting
[   362.725] (II) modeset(1): using default device
[   362.725] (II) Loading sub module "fbdevhw"
[   362.725] (II) LoadModule: "fbdevhw"
[   362.725] (II) Loading /usr/lib/xorg/modules/libfbdevhw.so
[   362.725] (II) Module fbdevhw: vendor="X.Org Foundation"
[   362.725] 	compiled for 1.21.1.7, module version = 0.0.2
[   362.725] 	ABI class: X.Org Video Driver, version 25.2
[   362.725] (EE) Unable to find a valid framebuffer device
[   362.725] (WW) Falling back to old probe method for fbdev
[   362.725] (II) Loading sub module "fbdevhw"
[   362.725] (II) LoadModule: "fbdevhw"
[   362.726] (II) Loading /usr/lib/xorg/modules/libfbdevhw.so
[   362.726] (II) Module fbdevhw: vendor="X.Org Foundation"
[   362.726] 	compiled for 1.21.1.7, module version = 0.0.2
[   362.726] 	ABI class: X.Org Video Driver, version 25.2
[   362.726] (EE) open /dev/fb0: No such file or directory
[   362.726] vesa: Ignoring device with a bound kernel driver
[   362.726] (II) modeset(G0): using drv /dev/dri/card0
[   362.726] (EE) Screen 0 deleted because of no matching config section.
[   362.726] (II) UnloadModule: "modesetting"
[   362.726] (EE) Screen 1 deleted because of no matching config section.
[   362.726] (II) UnloadModule: "fbdev"
[   362.726] (II) UnloadSubModule: "fbdevhw"
[   362.726] (EE) Screen 1 deleted because of no matching config section.
[   362.726] (II) UnloadModule: "vesa"
[   362.726] (EE) 
Fatal server error:
[   362.726] (EE) Cannot run in framebuffer mode. Please specify busIDs        for all framebuffer devices
[   362.726] (EE) 
[   362.726] (EE) 
Please consult the The X.Org Foundation support 
	 at http://wiki.x.org
 for help. 
[   362.726] (EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
[   362.726] (EE) 
[   362.735] (EE) Server terminated with error (1). Closing log file.

[Clarity]

@dimkr , I think I tried what you suggested before, from you. AND, I think I reported that the install via the PM does not work. I may be wrong on my memory that you had suggested it before, but the path you offer does not work.

That is the reason why I asked for what you had produced in hopes that it would yield a result for testing.

Seems we 'may' need a WoofCE devloper who has a test PC with an nvidia card on this problem to pave the way. ... or @666philb

I haven't tried to download a full Debian and try an install on a live system to determine if that gives the same results we are seeing here in BKWP64.

On a different post I reply:
"@bigpup I concur."

[dimkr]

[ 362.588] (II) LoadModule: "nv"
[ 362.589] (WW) Warning, couldn't open module nv
[ 362.589] (EE) Failed to load module "nv" (module does not exist, 0)

Did you save normally, instead of trying to create a .pet or SFS from the built driver?

[Clarity]

NO offence @dimkr but if you still have either PET/SFS, we'll test and report.

[dimkr]

NO offence @dimkr but if you still have either PET/SFS, we'll test and report.

AFAIK the license agreement and the general software licensing mess (use of GPL symbols from a proprietary driver, etc') forbids redistribution of the drivers. This is exactly why the installer has to build stuff on your computer instead of including a prebuilt binary.

@bigpup From what I see, the right driver to use in xorgwizard is nvidia and not nv (as your log says) - you should have /usr/lib/xorg/modules/nvidia_drv.so once you install nvidia-driver.

And make sure you run xorgwizard only after nouveau is blacklisted and no longer loaded (requires reboot) - from what I see, xorgwizard-automatic goes for nv (isn't this one dead for many years?) instead of nvidia if it doesn't see the nvidia-current.ko loaded. This can explain why your X.Org tries to use nv, which obviously doesn't exist. If you specify the driver manually, it should be nvidia.

[mikewalsh]

Mm.

It does seem a bit fruitless giving advice as to what should happen - from the official 'parent' documentation - IF one is unable to test it oneself.

Theory is one thing. Practical, real-world application of the proffered advice is, unfortunately - as most of us are painfully aware - often something else...!

I can - if necessary, as a last resort - 'fall back' to standard Intel graphics on the main desktop rig. I don't have that choice with the Dell Latitude, 'cos the D630 came as standard with an Nvidia mobile GPU.....so it's got to be either nouveau OR 'official' for me. And fan-speed control is a moot point here; the Lat's NVS-135m employs a shared heat pipe for cooling......whereas the HP's Asus GeForce GT 710 is a passive-cooled, fanless design.

(shrug...)

Mike.

[dimkr]

It does seem a bit fruitless giving advice as to what should happen

Even without a supported GPU, I can install nvidia-driver and see that the correct driver is nvidia and not nv, and I can see the wrong assumption inside xorgwizard-automatic, which makes it choose the wrong, non-existing driver.

[mikewalsh]

@dimkr :-

Well, I've got to confess; it never would have occurred to me to install packages for which I had no use.....simply to look at them. Personally, I would have downloaded the individual .debs, extracted them, then looked at them that way.....

(shrug...)

To each their own!

Mike.

[dimkr]

I just boot a VM, install the package and look at the files. 2 minutes max, no harm done to my system

[bigpup]

[ 362.588] (II) LoadModule: "nv"
[ 362.589] (WW) Warning, couldn't open module nv
[ 362.589] (EE) Failed to load module "nv" (module does not exist, 0)

Did you save normally, instead of trying to create a .pet or SFS from the built driver?

This frugal install is running in pupmode 12 so the save is written to exactly at the time something needs to be written to the save.

I did exactly what you stated about installing the driver by using apt install nvidia-driver.

This is a direct download from the Debian repository.

Apt in Bookworm is setup to only look in Debian repositories.

Again I think the driver provided by apt from the Debian repositories are to much specifically compiled for Debian and how it operates and it's file system is setup.
I can tell you for sure that a driver installed by using a Nvidia run package or a pet that is made by using Getnvidia Puppy program does not install a bunch of other files under /etc/

Really if you want to keep talking about trying to get the driver that installs by using apt.

Here is my old topic that I started trying to do it in the first version release of BookwormPup64 10.0.3:
viewtopic.php?p=89250#p89250

You can see that ETP was finally able to get a Nvidia run package to just compile a driver instead of using apt to download one from a Debian repository.

This topic provides a driver for BookwormPup64 10.0.3 and 10.0.4
viewtopic.php?p=102730#p102730

I have had PM's with ETP to try and get a Nvidia run package to compile a driver for BookwormPup64 10.0.6 and we both cannot get it to compile.
That is what this topic is trying to figure out what to do!

[bigpup]

I just boot a VM, install the package and look at the files. 2 minutes max, no harm done to my system

You do not have Nvidia hardware.

Have you rebooted?

So what driver is getting used after reboot?

If you are using the Intel driver or Radeon driver.

The reboot is not going to do anything to stop using them.
It is suppose to blacklist the nouveau driver and change to using the Nvidia driver.
That is the big issue This is not correctly getting done by the script files that apt install nvidia-driver installs.

[dimkr]

The Debian package doesn't run xorgwizard, it's a Puppy thing. And it can't blacklist nouveau the Debian way. If you do the blacklisting yourself and select the nvidia driver in xorgwizard - from what I see, it should work, because it's built and installed correctly.

[Clarity]

@dimkr , again this post is not to upset, but, as I understand developers-users "CAN" distribute propietary drivers for test ... which is what we, here, are attempting. Would you consider sharing, where you would indicate to "test this ..." (or wording that makes clear it is provided for testing) as we try to get to the bottom of what's wrong?

If we have a model that works, we can backtrack to figure out where WE might be getting it wrong and why, for this WoofCE PUP.

Consider it.

[bigpup]

The Debian package doesn't run xorgwizard, it's a Puppy thing. And it can't blacklist nouveau the Debian way. If you do the blacklisting yourself and select the nvidia driver in xorgwizard - from what I see, it should work, because it's built and installed correctly.

This is not the first time I tried using apt install nvidia-driver.

I have already tried to blacklist by doing it with the Boot Manager ->Modules ->Blacklist modules in the first time I tried when BookwormPup64 10.0.6 was released.
Use apt install nvidia-driver.
Reboot.
The nouveau driver did not load on reboot, but it still would not use the Nvidia driver.

But I will try it again with a fresh new save folder and clean boot before doing apt install nvidia-driver.

Be back soon with my results.

[bigpup]

Results of new test.

Blacklisting nouveau driver, booting using the Vesa driver, before trying to install the Nvidia driver.

Still does not work using apt install nvidia-driver.

Already posted why it does not work.

Here is something else it is doing that is not correct for Puppy Linux.

Installing in wrong location and using wrong names for the files.

nvidia-current.ko:
Running module version sanity check.

• Original module
  • No original module exists within this kernel
• Installation
  • Installing to /lib/modules/6.1.76/updates/dkms/

nvidia-current-modeset.ko:
Running module version sanity check.

• Original module
  • No original module exists within this kernel
• Installation
  • Installing to /lib/modules/6.1.76/updates/dkms/

nvidia-current-drm.ko:
Running module version sanity check.

• Original module
  • No original module exists within this kernel
• Installation
  • Installing to /lib/modules/6.1.76/updates/dkms/

nvidia-current-uvm.ko:
Running module version sanity check.

• Original module
  • No original module exists within this kernel
• Installation
  • Installing to /lib/modules/6.1.76/updates/dkms/

nvidia-current-peermem.ko:
Running module version sanity check.

• Original module
  • No original module exists within this kernel
• Installation
  • Installing to /lib/modules/6.1.76/updates/dkms/

None of these files should have current in the name.

Normally in Puppy Linux the files are installed in this location (/usr/lib/modules/kernel number/kernel/drivers/video/ ) with these names.
.
examjple working install from BookwormPup64 10.0.4 installed by Nvidia pet package compiled for it using Nvidia run package.

examjple working install from BookwormPup64 10.0.4 installed by Nvidia pet package compiled for it using Nvidia run package

Screenshot(4).jpg (34.23 KiB) Viewed 839 times

.
.
Yes I even tried to hack the install and move these files to the correct location and give them the correct names.
No help.
.
Boots only to need to run xorgwizard and nothing you do gets it past that point.
.
.

[dimkr]

@bigpup Maybe try

Code: Select all

depmod
modprobe nvidia-current
xorgwizard

On my system, modprobe finds the driver (it only refuses to load because there's no supported GPU), so the problem is not wrong location or name:

If modprobe nvidia-current fails on your machine, with a supported GPU, check what dmesg says because the driver explains why it refuses to load.

[bigpup]

Are you saying to try this when it boots to the point of needing to run xorgwizard?

That is as far as it will boot.

It never goes past that point in the boot process.

If yes.

Enter modprobe nvidia-current there before running xorgwizard?

[dimkr]

Enter modprobe nvidia-current there before running xorgwizard?

Yes - modprobe nvidia-current, if it fails check dmesg, otherwise use xorgwizard-cli and manually specify the nvidia driver, then xwin.