Nvidia driver istallation woes (or no displayport / hdmi signal)

[chris_alife]

I need help connecting to an external monitor.
The PC cannot 'see' the monitor. I've tried activiating the port, and tried installing the Nvidia driver for the card I have.

PC nr 1 cannot connect to 2nd screen through HDMI port. PC 2 cannot connect through display port (DP to hdmi cable)
HDMI cable works and is tested with work's windows PC.

Main PC I am trying this on has no hdmi port, only display port connecting through a dongle thing to convert to hdmi. All output and errors are from this PC, specs below, but worth mentioning that i also have a second PC with an hdmi port, same version of Salix v14.2, that acts the same way. Both are HP laptops.

PC specs:
HP Elitebook 8740w dual core.
Graphics card: NVIDIA Corporation G92GLM [Quadro FX 2800M] (rev a2)

Driver I am trying to use is 340.107 (correct according to Nvidia and also I had this working on Bodhi Linux so I know it works with this card).

Xrandr output:

Code: Select all

Screen 0: minimum 320 x 200, current 1920 x 1200, maximum 8192 x 8192
LVDS-1 connected 1920x1200+0+0 (normal left inverted right x axis y axis) 367mm x 230mm
   1920x1200     60.04*+
   1920x1080     59.96  
   1600x1200     59.87  
   1680x1050     59.95  
   1400x1050     59.98  
   1280x1024     59.89  
   1280x960      59.94  
   1152x864      59.96  
   1024x768      59.92  
   800x600       59.86  
   640x480       59.38  
   720x400       59.55  
   640x400       59.95  
   640x350       59.77  
DP-1 disconnected (normal left inverted right x axis y axis)
DP-2 disconnected (normal left inverted right x axis y axis)
DP-3 disconnected (normal left inverted right x axis y axis)
eDP-1 disconnected (normal left inverted right x axis y axis)
VGA-1 disconnected (normal left inverted right x axis y axis)

I have tried

Code: Select all

xrandr --addmode DP-1 1920x1080

and a couple other things I forget (was a few days ago).

So i read somewhere that maybe it's nouveau that can't see the ports (this makes no sense as clearly xrandr is listing the port as disconnected while using the nouveau driver). But I tried installing the nvidia driver anyway because why not?

The driver fails to install (I'll add here that i don't care much for nvidia and would just as well have this working with nouveau) but anyway, here is an excerpt of my log file for the error:

Code: Select all

nvidia-installer log file '/var/log/nvidia-installer.log'
creation time: Wed Dec 12 19:10:07 2018
installer version: 340.107

PATH: /usr/local/bin:/usr/bin:/bin:/usr/games:/usr/lib64/java/jre/bin:/usr/local/sbin:/sbin:/usr/sbin

nvidia-installer command line:
    ./nvidia-installer

Unable to load: nvidia-installer ncurses v6 user interface

Using: nvidia-installer ncurses user interface
-> License accepted.
-> Installing NVIDIA driver version 340.107.
-> Performing CC sanity check with CC="/usr/bin/cc".
-> Kernel source path: '/usr/src/linux'
-> Kernel output path: '/usr/src/linux'
-> Performing rivafb check.
-> Performing nvidiafb check.
-> Performing Xen check.
-> Performing PREEMPT_RT check.
-> Cleaning kernel module build directory.
   executing: 'cd ./kernel; /usr/bin/make clean'...
-> Building NVIDIA kernel module:
   executing: 'cd ./kernel; /usr/bin/make module SYSSRC=/usr/src/linux SYSOUT=/usr/src/linux NV_BUILD_MODULE_INSTANCES='...
   NVIDIA: calling KBUILD...
   make[1]: Entering directory '/usr/src/linux-4.4.157'
   test -e include/generated/autoconf.h -a -e include/config/auto.conf || (             \
   echo >&2;                                                    \
   echo >&2 "  ERROR: Kernel configuration is invalid.";                \
   echo >&2 "         include/generated/autoconf.h or include/config/auto.conf are missing.";\
   echo >&2 "         Run 'make oldconfig && make prepare' on kernel src to fix it.";   \
   echo >&2 ;                                                   \
   /bin/false)
   mkdir -p /tmp/selfgz1186/NVIDIA-Linux-x86_64-340.107/kernel/.tmp_versions ; rm -f /tmp/selfgz1186/NVIDIA-Linux-x86_64-340.107/kernel/.tmp_versions/*
   /usr/bin/make -f ./scripts/Makefile.build obj=/tmp/selfgz1186/NVIDIA-Linux-x86_64-340.107/kernel

plus a couple thousand more lines of nothing helpful.

In addition, the installer at the time it ran (yes, I ran it twice) displayed an error about kernel not being built with the correct version of GCC/G++ and another error about not having installed at the time of driver install the the correct version of same. So I downgraded to the version it said (5.3 from 5.5) but that didn't remove the error, and anyway I have no idea how to build the kernel module again using the old version of GCC that updates with system updates (potentially breaking it again?).

Needless to say, given that nvidia wrote a log file hundreds of lines long, refer to it for error reporting when the install fails BUT THEN DON'T INCLUDE THE ERROR IN THE LOG FILE, has me wanting to stick with nouveau

Sooo, how can i get my display port via converter dongle to hdmi to use my HD monitor?
Thanks in advance.

[mimosa]

Do you have the kernel source installed?

[galmei]

My logic tells me, if the problem occurs with and without a DP-to-HDMI adapter, then it's not the adapter. However, two different computers are affected, there is a bit more caution required. It is also of interest, if the adapter is passive or active. Passive may, but does not have to work.

A lack of information in the current situation is whether the BIOS image and the Lilo or Grub image is displayed on both monitors at the same time. If so, then everything is right with the connection. It may then only be about missing configuration of the X server. I do not know how it behaves under Uefi.

An attempt to start from zero is e.g. following: remove the monitor from the computer, switch off and unplug the power supply plug and leave it for a short time. Turn off the computer and boot without the external monitor. Then use xrandr to check the configuration and condition. Then connect the monitor and turn it on. When the monitor is ready wait a few seconds and then use xrand to check the configuration again. Then use xrandr to configure the X server so that it can place the monitor image as desired (left, right, top or bottom). The X server can not know where to place the second image, and then ignores the additional monitor. Placing is the task of the X server and not the driver. Therefore, it is also useless to switch from the Nouveau driver to the Nvidia driver.

Helpful:
man xrandr
man xdpyinfo
man xorg.conf

[chris_alife]

@minosa yes I have the kernel source installed, that is where the error that appears on screen but not in the installer log tells me that it is a problem that the kernel source was built with gcc 5.5 instead of 5.3. No idea how to go about fixing that, except maybe to downgrade, delete kernel source, reinstall etc etc. But that has to do with the nvidia driver, which I don't think will solve my actual problem of getting the display port to activate and work.

@galmei I should clarify: i have tried 3 laptops:
1 as described above, HP 8740w running Salix with a display port (no HDMI) doesn't 'see' the external monitor.
2. older HP Compaq, same Salix installed from same USB installer, has HDMI port, does not see the monitor either.
3. Windows Thinkad from work, has HDMI port, connects no problem.

So, it's not the cable (although the adapter might be faulty but unlikely). It's not the monitor.
It's not likely to be the graphics driver, nouveau supports multiple displays.

I'm reading up on this, and trying to find out what my xorg.conf file should include, but I don't have one, I have /etc/X11/xorg.conf-vesa
and afaik, I'm not using vesa, I AM using nouveau.

This is quite annoying, it's not like I have the wrong resolution or something, I just get "no signal" from the monitor, and xrandr shows all output ports to do with vga as "disconnected", except for the built in screen.

Plus xrandr isn't giving me enough info, I don't know if my display port where I plug in the cable is eDP-1, DP-1, DP-2, or DP-3. All the forums I look at don't have dashes in the names (shrug).

I'll try re-running X -configure from init 3, result:

Code: Select all

$sudo X -configure

X.Org X Server 1.18.3
Release Date: 2016-04-04
X Protocol Version 11, Revision 0
Build Operating System: Slackware 14.2 Slackware Linux Project
Current Operating System: Linux darkstar 4.4.19 #2 SMP Mon Aug 22 14:37:39 CDT 2016 x86_64
Kernel command line: auto BOOT_IMAGE=Salix ro root=801 quiet resume=/dev/sda2 vt.default_utf8=1
Build Date: 18 October 2017  12:21:16PM
 
Current version of pixman: 0.34.0
	Before reporting problems, check http://wiki.x.org
	to make sure that you have the latest version.
Markers: (--) probed, (**) from config file, (==) default setting,
	(++) from command line, (!!) notice, (II) informational,
	(WW) warning, (EE) error, (NI) not implemented, (??) unknown.
(==) Log file: "/var/log/Xorg.0.log", Time: Fri Dec 14 18:58:56 2018
List of video drivers:
	apm
	ark
	ast
	chips
	cirrus
	dummy
	glint
	i128
	i740
	mach64
	mga
	neomagic
	nv
	r128
	rendition
	s3
	s3virge
	savage
	siliconmotion
	sis
	tdfx
	tga
	trident
	v4l
	modesetting
	voodoo
	amdgpu
	ati
	radeon
	intel
	nouveau
	openchrome
	vmware
	vesa
(++) Using config file: "/root/xorg.conf.new"
(==) Using config directory: "/etc/X11/xorg.conf.d"
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
Number of created screens does not match number of detected devices.
  Configuration failed.
(EE) Server terminated with error (2). Closing log file.

[mimosa]

So just to clarify, it seems there are two independent problems - the nvidia driver you need won't build on Salix 14.2, perhaps because of something to do with a minimum gcc version requirement. And quite apart from that, your external monitor is not being recognized, though it is on the same hardware with other distros.

The first problem will likely apply to all slack users and it may be worth searching for solutions (or the lack of them) on that basis, in LQ and elsewhere. But as a quick test, you could try with Slackware Current or Slackel (which is basically Salix Current). I'm not sure what the gcc version is there but it will be recent and probably meets the requirement. Live versions are available of both.

http://www.slackel.gr/forum/about.htm
https://alien.slackbook.org/blog/slackw ... e-edition/

[gapan]

gcc in 14.2 is currently 5.5.0, so there's no problem with that.

@chris_alife
The problem with not being able to install the nvidia driver is quite likely that you have mismatched kernel and kernel-source versions. Run:

Code: Select all

ls /var/log/packages/kernel-*

The latest kernel for 14.2 at his moment is 4.4.157 and all packages should match that (well, except kernel-firmware which has its own versioning, but you still want the latest). If not, you need to upgrade them all to the latest version:

Code: Select all

sudo slapt-get -i kernel-huge kernel-headers kernel-modules kernel-firmware kernel-source

(if in 32bit salix, you probably want the kernel-huge-smp and kernel-modules-smp packages instead)

After upgrading, don't forget to update the boot loader or your system won't boot! Either run lilo on a BIOS system:

Code: Select all

sudo lilo -v

or eliloconfig in an EFI system:

Code: Select all

sudo eliloconfig

Reboot and you'll probably be able to install the nvidia proprietary driver.

[mimosa]

gcc in 14.2 is currently 5.5.0, so there's no problem with that.

I should have checked the version number. I thought as much.

Instructions for both a kernel upgrade and installing the nvidia driver are on the wiki.

[galmei]

I'm sorry that this is a very long text. There is a lot to consider.

From the following, I can see that you are using kernel version 4.4.157 on one of the two computers. That's a good thing.

Code: Select all

by chris_alife » 13. Dec 2018, 19:39
make[1]: Entering directory '/usr/src/linux-4.4.157'

But the following illustration shows that you are working or testing with a completely outdated kernel on the other computer.

Code: Select all

by chris_alife » 14. Dec 2018, 17:48
Current Operating System: Linux darkstar 4.4.19 #2 SMP Mon Aug 22 14:37:39 CDT 2016 x86_64

This old kernel is completely unsuitable for finding the cause of the problem.

If you are using the huge kernel, for the machine running the 4.4.19 kernel, do the following:

Code: Select all

sudo slapt-get --install kernel-huge kernel-modules kernel-headers kernel-firmware

Then update the boot mechanism, then boot. The kernel sources are not needed, so they are not included.
For the machine that already has the 4.4.157 kernel, make sure that all of the above modules are up to date as well. In case of doubt, let the same happen.

The next interesting thing is this statement:

PC specs:
HP Elitebook 8740w dual core.
Graphics card: NVIDIA Corporation G92GLM [Quadro FX 2800M] (rev a2)

It means that the chipset NVIDIA G92GLM is installed on the mainboard. Among other things, it contains the functionality of the add-in graphics card Quadro FX 2800M. Now the Quadro variant of the GeForce FX is not included in the excerpt below from the log file /var/log/Xorg.0.log. Neither in the former part with the NOUVEAU drivers, nor in the other part with the Nouveau drivers. This should mean that the Quadro properties are not supported by this X server. The man page (man nouveau) shows the Quadro FX, but the man page and the actual properties are not so tightly coupled.

For comparison with my graphics adapter:

Code: Select all

$ inxi -G
Graphics:  Card: NVIDIA C51 [GeForce 6150 LE]
           Display Server: X.Org 1.18.3 drivers: nouveau (unloaded: nv,vesa) Resolution: 1280x1024@60.02hz
           GLX Renderer: Gallium 0.4 on NV4E GLX Version: 2.1 Mesa 11.2.2
$

The chipset is NVIDIA C51, which includes the functionality of the GeForce 6150 LE add-in graphics card. Both the family GeForce 6, as well as the name of the graphics functionality NV4E, are included in the excerpt below.

Now two matters are of interest (to me).
1. What information does the inxi call provide on your computers? Preferably, you will only consider one computer to avoid confusion. You can also take this information from the log file with a little effort.

2. In your first message, you indicated that it works on Bodhi Linux. What is the version number and release date of the X server on the Bodhi Linux installation? And is the simulated graphics card and the name of the chipset functionality included in the Bodhi Linux Xorg.0.log?

My idea at the moment is that the Salix distribution X server is either too old or incomplete.

Plus xrandr isn't giving me enough info, I don't know if my display port where I plug in the cable is eDP-1, DP-1, DP-2, or DP-3. All the forums I look at don't have dashes in the names (shrug).

From /var/log/Xorg.0.log more information can be gained. But this information may perhaps only be obtained from the data sheet of the computer. The dashes are included in human-readable descriptions. They are not included in configuration data or program code because they might be misunderstood as subtraction operators.

I'm reading up on this, and trying to find out what my xorg.conf file should include, but I don't have one, I have /etc/X11/xorg.conf-vesa
and afaik, I'm not using vesa, I AM using nouveau.

Today, xorg.conf is almost irrelevant. Configuration is now done via individual files in /etc/X11/xorg.conf.d.
/etc/X11/xorg.conf-vesa has no meaning. You can use this file if nothing else works. Then rename /etc/X11/xorg.conf-vesa to /etc/X11/xorg.conf .


Two excerpts from the Salix log file /var/log/Xorg.0.log:

Code: Select all

X.Org X Server 1.18.3
Release Date: 2016-04-04

Code: Select all

[    22.364] 	compiled for 1.18.0, module version = 2.3.4
[    22.364] 	Module class: X.Org Video Driver
[    22.364] 	ABI class: X.Org Video Driver, version 20.0
[    22.365] (II) NOUVEAU driver 
[    22.365] (II) NOUVEAU driver for NVIDIA chipset families :
[    22.365] 	RIVA TNT        (NV04)
[    22.365] 	RIVA TNT2       (NV05)
[    22.365] 	GeForce 256     (NV10)
[    22.365] 	GeForce 2       (NV11, NV15)
[    22.365] 	GeForce 4MX     (NV17, NV18)
[    22.365] 	GeForce 3       (NV20)
[    22.365] 	GeForce 4Ti     (NV25, NV28)
[    22.365] 	GeForce FX      (NV3x)
[    22.365] 	GeForce 6       (NV4x)
[    22.365] 	GeForce 7       (G7x)
[    22.365] 	GeForce 8       (G8x)
[    22.365] 	GeForce GTX 200 (NVA0)
[    22.365] 	GeForce GTX 400 (NVC0)
[    22.365] (II) NOUVEAU driver 
[    22.365] (II) NOUVEAU driver for NVIDIA chipset families :
[    22.365] 	RIVA TNT        (NV04)
[    22.365] 	RIVA TNT2       (NV05)
[    22.366] 	GeForce 256     (NV10)
[    22.366] 	GeForce 2       (NV11, NV15)
[    22.366] 	GeForce 4MX     (NV17, NV18)
[    22.366] 	GeForce 3       (NV20)
[    22.366] 	GeForce 4Ti     (NV25, NV28)
[    22.366] 	GeForce FX      (NV3x)
[    22.366] 	GeForce 6       (NV4x)
[    22.366] 	GeForce 7       (G7x)
[    22.366] 	GeForce 8       (G8x)
[    22.366] 	GeForce GTX 200 (NVA0)
[    22.366] 	GeForce GTX 400 (NVC0)

[chris_alife]

Thanks for the replies. I really appreciate you providing help.

Here is what i have before updating:

Code: Select all

$ls /var/log/packages/kernel-*
/var/log/packages/kernel-firmware[b]-20160823[/b]git-noarch-1_slack14.2
/var/log/packages/kernel-headers-[b]4.4.19-x86-1[/b]_slack14.2
/var/log/packages/kernel-huge-[b]4.4.19-x86_64-1[/b]_slack14.2
/var/log/packages/kernel-modules-[b]4.4.19-x86_64-1[/b]_slack14.2
/var/log/packages/kernel-source-[b]4.4.157-[/b]noarch-1

and after updating:

Code: Select all

@darkstar [ ~ ] 21:07:28 >
$ls /var/log/packages/kernel-*
/var/log/packages/kernel-firmware[b]-20180913_44d4fca-noarch-1[/b]
/var/log/packages/kernel-headers[b]-4.4.157-x86-1[/b]
/var/log/packages/kernel-huge-[b]4.4.157-x86_64-1[/b]
/var/log/packages/kernel-modules-[b]4.4.157-x86_64-1[/b]
/var/log/packages/kernel-source-[b]4.4.157-noarch-1[/b]

Off-topic question but why don't these updates come in gslapt auto-upgrade-notifier? I thought I was fully up-to-date.

And lilo updated no problem:

Code: Select all

Boot image: /boot/vmlinuz -> vmlinuz-huge-4.4.157
Added Salix  *

Will try all the cabling again and probably the nvidia driver too, but to know if it works I first have to wait until my SO stops using new monitor as a TV

[gapan]

kernel related packages are blacklisted from automatic upgrading. You don't want to unknowingly upgrade your kernel among a handful of other packages, it might render your system unbootable. You have to explicitly state that you want to upgrade the kernel and that can only happen from the command line.