![]() |
Slackware current still freezes
I posted this already a cpl of hrs ago but not sure where the post has gone.
Not a newbie, just re-registered as my old account must have been disabled/deleted (as i didnt use it for many years). I have been using slackware for 15+yrs and had never happened to have to hard-restart the system as it kept on freezing once I installed slackware14.2; due to lack of time (and health) i rolled back to 14.1. Recently, i decided to try again and installed slackware current; at the first startx i was welcomed with a total system after just a few seconds. Many hrs of searching around led me to believe that the issue was with noeveau and its video acceleration function. So I permanently disabled noeveau video accelerator by adding a file in /etc/X11/xorg.vesa.d/disable-gpu.conf and edited: Code:
Section "Extensions" Quote:
I gues i could try to install the nVidia proprietary drivers, but i couldnt find an updated how-to tutorial (i tried last night and it went severely wrong). Any suggestions would be welcome. Regards, |
|
I put my vote on installing proprietary drivers, just download it from nvidia website. Make it executable and run it from init 3 as root, reboot.
Also you have a very old card so maybe you need the legacy drivers, please check on the website so that it supports your card Using slackbuild or something alike never worked for me. ____________ Download NVIDIA proprietary drivers Go to nvidia.com and download the Linux 64-bit drivers for your graphic card. Install NVIDIA proprietary drivers chmod +x NVIDIA*.run ./NVIDIA*.run Choose ”yes” when it asks about blacklistin nouveau. (You could write your own but for me its faster to just reboot) Reboot (or do some modprobe/rmmod voodoo) ./NVIDIA*.run When it asks for 32-bit libraries, choose ”yes”. Reboot. ___________________ *for funsies* trite@bifrost:~$ nvidia-smi Wed Jan 16 15:11:28 2019 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 410.78 Driver Version: 410.78 CUDA Version: 10.0 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GTX 108... Off | 00000000:01:00.0 On | N/A | | 29% 60C P0 63W / 250W | 3674MiB / 11175MiB | 0% Default | +-------------------------------+----------------------+----------------------+ | 1 GeForce GTX 980 Off | 00000000:02:00.0 Off | N/A | | 28% 27C P8 11W / 180W | 1MiB / 4043MiB | 0% Default | +-------------------------------+----------------------+----------------------+ | 2 GeForce GTX 970 Off | 00000000:03:00.0 Off | N/A | | 0% 38C P8 12W / 170W | 1MiB / 4043MiB | 0% Default | +-------------------------------+----------------------+----------------------+ | 3 GeForce GTX 960 Off | 00000000:09:00.0 Off | N/A | | 0% 32C P8 5W / 160W | 1MiB / 4043MiB | 0% Default | +-------------------------------+----------------------+----------------------+ | 4 GeForce GTX 960 Off | 00000000:0A:00.0 Off | N/A | | 0% 35C P8 8W / 130W | 1MiB / 2002MiB | 0% Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 1540 G /usr/libexec/Xorg 245MiB | | 0 8556 C+G ...x86)\Origin Games\Battlefield V\bfv.exe 3258MiB | | 0 26427 C+G ...c/Program Files (x86)/Origin/Origin.exe 165MiB | +-----------------------------------------------------------------------------+ |
Thanks for the fast reply.
Quote:
Quote:
Quote:
Anyway, i will try again and feedback shortly. Thanks for your kind help. |
Hello,
Quote:
Quote:
as stated in the documentation: Quote:
-- SeB |
Quote:
If you choose your card model on nvidia site you will get the right one so no worries, I was thinking of pre-package solutions where they usually have two, one for legacy cards and one for newer ones but since you choose model on nvidia site it should give you the right one. Also installing through runfile should be the same on all systems, download the binary. Make it executable and run it from init 3. There are some prerequisits like maybe you need kernel-headers or something like that but those are already there in Slackware system. You also need to have nouveau blacklisted but the runfile does that for you. You go into init 3 by either just writing it in prompt or edit /etc/inittab and change default mode to 3 from 4. |
@however
Sorry, I missed your card model being pretty old, phenixia2003 is right about the driver version. In case you prefer the SlackBuilds: https://slackbuilds.org/repository/1...acy390-kernel/ https://slackbuilds.org/repository/1...acy390-driver/ |
Thanks very much, everyone!
it seems that I have a long evening/night ahead of me (making love with slackware). I'll let you know the details later (or tomorrow). Ciao p.s.: do i also need to install the nVidia kernel? |
Quote:
Some more info from the maintainer of the SlackBuilds, you might find these useful (don't mind the driver version there, it's an old post): https://www.linuxquestions.org/quest...7/#post5599110 |
and here we go again, booting from windows!
what happened: Quote:
Any hints? Thanks, p.s.: I forgot to take a picture of the /var/log/nvidia-installer.log file and try to post it here (although being a newbie member i wouldnt be allowed), or at least re-type the msgs in it. Should this be vital to avoid another full reinstall, I would be glad to try my best and get it & type the whole log here. This time, if i manage to get the system running again, I will try with the binary files installation. Should this be so complex, in the 21st century computer world? (( |
.... can ... you .... be ... a little .... more ... specific.... please?
Quote:
It's difficult to understand your system status now. Quote:
https://www.youtube.com/watch?v=IVpOyKCNZYw (Minute 1:40) I personally don't even bother to install their crap driver on my few systems where I have the Nvidia Optimus (Intel + discrete Nvidia), just using Intel. But that's not possible in your case. Go for "binary files installation", whatever that means for you. |
Quote:
and yes, the install stopped and my system status is, as mentioned: i can start X but with no keyboard and no mouse (they don't work, not recognized, caput, modules corrupted, broken, files damaged, finito, etc; i hope this is clear) Quote:
Quote:
as suggested by Quote:
would the /var/log/nvidia-installer.log file help? |
Somewhat clearer now. I thought you went for the SlackBuilds and failed to build nvidia-legacy390-kernel.
Yes, the /var/log/nvidia-installer.log might help to see what happened and what system files were modified, you can use https://pastebin.com/ to upload it. On your actual status - no mouse - no keyboard - check what you have in /etc/X11/ - is there any xorg.conf file ? If some leftovers from Nvidia, remove (move-backup) them and try to startx The SlackBuilds approach is a safer because while building them you can see if there are errors / failing and you can stop. -current has a new kernel and they might fail, but it's worth trying. Take the time and read the README from: https://slackbuilds.org/repository/1...acy390-driver/ together with: https://www.linuxquestions.org/quest...7/#post5599110 |
Try updating your kernel and then try the runfile again, or try with older version of nvidia driver.
What kernel version do you have? I've had that some versions of nvidia drivers only compile kernel modules with certain kernel versions but if you switch to an older one it works. Easiest is to just use latest kernel with latest drivers since they have the most chance of success. But I've always ran current as rolling release and update regularly so when it sometimes doesnt work I just either update kernel or install the last working older version of driver. |
First off despite that some people don't prefer it, the nvidia proprietary driver is most definitely NOT "crap". I won't belabor that subject and derail this thread but having used and compared it for well over 20 years, as have millions of people, it has been a major benefit for alternative operating systems including Linux and we are lucky to have it. One of the values of installing from the .run script from runlevel 3 is we get useful error messages that offer possible solutions even for those that don't bother to read the requirements, NOT a good idea unless you don't really fit in a manually powerful distro like Slackware and would do better on a "hold your hand" Ubuntu derivative. You want the freedom and power? You gotta earn it.
One of those requirements for legacy drivers is having complete kernel AND SOURCE built with the same GCC as is being used. The Full Recommended Install should provide you with that if you didn't uncheck that option or update GCC. The install will sometimes fail if one hasn't disabled the nouveau driver and nvidia's own attempt wasn't able to manage it, usually due to some user created conflict. The disabling of nouveau should properly be done in "/etc/modprobe.d/blacklist.conf", nowhere else. Before one reboots after having created that blacklist, or letting nvidia do it for you, is to edit two things - "/etc/X11/xorg.conf" and "/etc/lilo.conf", or the appropriate grub file if you use that. The default xorg.conf.example can be renamed or copied to xorg.conf after you know it has the lines Code:
Code:
Code:
Having the kernel with source built by the gcc in use, disabling nouveau with a vesa fallback works 99% of the time. or at least it has for me in 20 years of lots of distro hopping and Slackware as my Main. Distros other than Slackware don't like it much that we don't use their repository rebuilt packages but most can be forced and I force them, and it always works, though on some distros updating can be more involved if one doesn't use DKMS and I don't... ever. It's just vastly easier on Slackware provided you follow nvidia's instructions. There is no need to worry about the VESA line in xorg.conf since nvidia's run script will offer to create a generic and proper one for you. Let it since your existing one will graciously not be deleted but simply renamed as a backup. Like everything else nothing is hard, everything is easy once you learn how. Sorry to have to say this in the 21st Century but RTFM and pay attention to error messages.. |
All times are GMT -5. The time now is 06:53 AM. |