SlackwareThis Forum is for the discussion of Slackware Linux.
Notices
Welcome to LinuxQuestions.org, a friendly and active Linux Community.
You are currently viewing LQ as a guest. By joining our community you will have the ability to post topics, receive our newsletter, use the advanced search, subscribe to threads and access many other special features. Registration is quick, simple and absolutely free. Join our community today!
Note that registered members see fewer ads, and ContentLink is completely disabled once you log in.
If you have any problems with the registration process or your account login, please contact us. If you need to reset your password, click here.
Having a problem logging in? Please visit this page to clear all LQ-related cookies.
Get a virtual cloud desktop with the Linux distro that you want in less than five minutes with Shells! With over 10 pre-installed distros to choose from, the worry-free installation life is here! Whether you are a digital nomad or just looking for flexibility, Shells can put your Linux machine on the device that you want to use.
Exclusive for LQ members, get up to 45% off per month. Click here for more info.
I could easily be wrong but I think aragorn2101 is mainly referring to using nVidia's "nvidia-foo.run" installer which does not employ KMS instead of the SlackBuild which does. The disadvantage is when you install a new kernel you must run the nviidia installer each time but the advantages are you have greater choice in which versions you can use, it's all done at runlevel3 which can have numerous advantages, it asks you what features you wish to include such as 32bit compatibility files, makes checks for kernel and gcc version, checks for conflicting GLX libraries and fixes them, and offers to run nvidia-xconfig for you. It's clear and simple and eschews convenience in favor of power.
I'd already tried the nvidia's installer as described in early of this thread. I cannot start x after the installation, means even worse
I know this must be terribly frustrating for you but hang in there as I'm confident we will get it working or know why not. I only spent maybe 3 minutes going through each of your posts, narke, but I didn't see any obvious reference to getting the driver from nvidia's search and installing via it's ./run file. All I saw was SBO packages. It's not that SBO packages are not good. They are fine especially when no difficulties arise. The reason I use and recommended the nvidia-foo.run method is that it is exceptionally rare that one will install successfully that won't work. I've had the nvidia installer tell me essentially "this version won't work on this card. You need "nvidia-foofoo.run" or warn me complete kernel source was not available or two different gcc compilers were used, etc.
There is no need to direct me to the post if I simply missed it. I'd just like to know what Xorg.0.log had to say and possibly "nvidia-installer.log in the same "/var/log" directory. It may be substantially more revealing when "startx" fails altogether than when it simply falls back to the 1915.
I know this must be terribly frustrating for you but hang in there as I'm confident we will get it working or know why not. I only spent maybe 3 minutes going through each of your posts, narke, but I didn't see any obvious reference to getting the driver from nvidia's search and installing via it's ./run file. All I saw was SBO packages. It's not that SBO packages are not good. They are fine especially when no difficulties arise. The reason I use and recommended the nvidia-foo.run method is that it is exceptionally rare that one will install successfully that won't work. I've had the nvidia installer tell me essentially "this version won't work on this card. You need "nvidia-foofoo.run" or warn me complete kernel source was not available or two different gcc compilers were used, etc.
There is no need to direct me to the post if I simply missed it. I'd just like to know what Xorg.0.log had to say and possibly "nvidia-installer.log in the same "/var/log" directory. It may be substantially more revealing when "startx" fails altogether than when it simply falls back to the 1915.
I tried again and recorded what happened. I firstly removed my nouveau and i915 driver files physically to ensure there would be only nvidia's driver will be running. I also removed my /etc/X11/xorg.conf.d. After that my system should be fully clean. Then I installed the nvidia driver using the .run script. It got succeeded as normally. After that I run nvidia-xorgsetup, which created a new /etc/X11/xorg.conf for me.
Here is the xorg.conf file:
Quote:
# nvidia-xconfig: X configuration file generated by nvidia-xconfig
# nvidia-xconfig: version 390.67 (buildmeister@swio-display-x64-rhel04-18) Fri Jun 1 04:26:22 PDT 2018
After a reboot, I run startx, it quickly end with error. log as attached file Xorg.0.log-0. It seems my nvida device was not detected. I inserted a "BusID "PCI:1:0:0" into the Device section of the config file to give it a hint. This time, the startx run very long with the screen fulled with black and without a return. The log file attached as Xorg.0.log-1. It looks to me the nvidia device was found but it still cannot get X up.
Hello again woody
Considering that Xorg.0-0.log terminates properly in that it stops trying to load and says so, while Xorg.0-1.log does not complete and without a clue, maybe we should view that as a clue itself.
First it is good that the addition of the BusID caused nvidia to be detected but I wonder why that was an issue when the nvidia installer had no problem detecting it and presumably at runlevel 3, before any X libraries are activated. However there are a substantial number of differences in our systems that make it difficult for me to offer anything of solid value. I use Legacy mode BIOS, kernel 4.15.0, and /etc/X11/xorg.conf (not xorg.conf.d) so I have no experience nor clues as to what to expect from EFI, 4.17.1, and xorg.conf.d.
My natural suspicion caused by overly lengthy response to startx but without any noted |W| or |E| failure makes me suspect 4.17.1 or even Xorg's reaction to the kernel. Since Current Slackware is still on 4.14.51 I have to wonder how any customizations in compiling 4.17.1 may be a problem causing the slow load time. I also wonder if left alone if it would ultimately succeed at the very least in generating an error. I can't recall ever in close to 20 years use of Linux and nvidia ever not either succeeding or failing and pretty quickly at that. Maybe see if it will succeed if you just walk away for 15 minutes or so. It also might be worthwhile to try plugging into and activating an external monitor since that is one of the few substitution tests available to laptops.
Also does startx succeed with just the Intel graphics system and driver? If it does is there any data revealed by it's success that might be possible to improve xorg.conf.d in the manner that BusID did?
I suspect from seeing some of your other threads that you chose 4.17.1 for specific hardware support for other components but is it possible to try 4.14.51 from Current? I did a little research on 4.17.1 and there is no mention of problems with nvidia though several with AMD/ATi graphics breakage but it is a variable that possibly falls under "less than certain".
I'm grasping at straws here and I know you are working hard to solve this and how frustrating it must be but I can only hope comments may trigger ideas for you to explore since you are the guy at the keyboard and I'm not.
[QUOTE=millgates;5872059]Hi, what is the output of
In my working system (with it to write the post) I have to disable/uninstall the nvidia, nouveau and only use the i915 intel driver, so the xrandr output at this state is:
Hello again woody
First it is good that the addition of the BusID caused nvidia to be detected but I wonder why that was an issue when the nvidia installer had no problem detecting it and presumably at runlevel 3, before any X libraries are activated. However there are a substantial number of differences in our systems that make it difficult for me to offer anything of solid value. I use Legacy mode BIOS, kernel 4.15.0, and /etc/X11/xorg.conf (not xorg.conf.d) so I have no experience nor clues as to what to expect from EFI, 4.17.1, and xorg.conf.d.
There might be a misunderstanding, when doing the nvida driver test, I was using the /etc/X11/xorg.conf file generated from nvidia-xorgsetup. And, the BusID line also was added to the single file.
Quote:
Also does startx succeed with just the Intel graphics system and driver? If it does is there any data revealed by it's success that might be possible to improve xorg.conf.d in the manner that BusID did?
It works quit well with intel's i915 driver. I actually already did a lot of exploring to the succeeded log file and config files (in this case, that are a lot of files in xorg.conf.d, but I think this is just a different way of organizing files, no different at the end), but I did not find anything that can be used to fix the nvidia driver case.
Quote:
I suspect from seeing some of your other threads that you chose 4.17.1 for specific hardware support for other components but is it possible to try 4.14.51 from Current? I did a little research on 4.17.1 and there is no mention of problems with nvidia though several with AMD/ATi graphics breakage but it is a variable that possibly falls under "less than certain".
I'm grasping at straws here and I know you are working hard to solve this and how frustrating it must be but I can only hope comments may trigger ideas for you to explore since you are the guy at the keyboard and I'm not.
Okay, I will find a time to test with the 4.14.51 and see what it will bring. Thanks you very much for your time and suggestions!
even though apparently nouveau was at that time removed as well as blacklisted and so it seems to me should have generated an error message.
You mean the nouveau.modset=0 parameter? It's not relevant because I had removed the nouveau*.ko when doing the test, so I did not need bother to remove the parameter as well.
LinuxQuestions.org is looking for people interested in writing
Editorials, Articles, Reviews, and more. If you'd like to contribute
content, let us know.