You are not logged in.
[SOLVED FOR OP - Topic kept open for others still experiencing the issue]
After the recent upgrade to linux-3.14 and nvidia-334.21-4, upon boot when startup reaches launching X in my DM (lightdm) the computer instantly powers off (not a clean shutdown, power off as if the power cord has been pulled).
I get zero errors in my journal and the logs for X and lightdm are completely empty .
I have tried isolating the issue, removed all non-default boot options, kernel modules, plymouth, scripts and systemd services and lightdm non default config, same problem.
I have booted to a console and tried startx, same problem.
I have rebuilt xorg-server and xorg-server-common from abs in-case they needed re-building against the new kernel, same problem.
Reverting to a 3.13 kernel and nvidia-334.21-2 resolves the issue.
When using nouveau, everything boots as normal, but I don't want to use nouveau as since kernel 3.12, using nouveau causes a new seat to be created on resume from suspend/hibernate. I am unable to get back to the session I suspended/hibernated.
This issue does not exist with the nvidia binary.
So I am now in a catch 22. I can't use nouveau as I am unable to suspend/hibernate.
I can't use nvidia as I cannot even get an X session.
I cannot upgrade the kernel as my system is rendered unusable.
I am unable to troubleshoot, as I get no errors or logs.
Can anyone suggest anything else to try?
Looking forward to all possible help.
Cheers.
[EDIT] adding rcutree.rcu_idle_gp_delay=1 to my kernel parameters solves the issue with linux-3.14-1 and nvidia-337.12 [/EDIT]
Last edited by Padfoot (2014-04-16 07:20:15)
Offline
I experienced the exact same problem, my laptop simply dies when attempting to start x. I suspect it is the very same issue a lot of laptop users had with the 3.10 kernel as well. There is an old topic about it here: https://bbs.archlinux.org/viewtopic.php?id=167195
As well as a still active topic on Nvidia forums: https://devtalk.nvidia.com/default/topi … r-crash/12
Back with the 3.10 kernel nobody found any solution for it, but it seemed to have been fixed in the 3.11 kernel. It would appear the problem has resurfaced for some with the 3.14 kernel. After reading through the topic on Nvidia forums I found a way to successfully get x working again.
I just have to use the kernel boot option:
rcutree.rcu_idle_gp_delay=1
Some people have posted on Nvidia forums that they must also use the boot option:
CONFIG_HZ=1000
This might be hardware dependent, so you may want to give it a try with both or just the first one and see if it works for you.
Offline
Hi,
it seems i have the same problem.
But when i try startx as root it drops me to xterm. There is no response while i am trying to start gdm.
I noticed that my xorg log file looks like described here : this page.
So, i tried to follow the intructions but there is no effecet.
If someone have any ideas, please let me know!
Offline
I just have to use the kernel boot option:
rcutree.rcu_idle_gp_delay=1
Some people have posted on Nvidia forums that they must also use the boot option:
CONFIG_HZ=1000
This might be hardware dependent, so you may want to give it a try with both or just the first one and see if it works for you.
Adding
rcutree.rcu_idle_gp_delay=1
to kernel boot parameters fixed it for me!
Thanks a million, I was banging my head the past couple of days trying to fix it
% uname -a
Linux archasus 3.14.1-1-ARCH #1 SMP PREEMPT Mon Apr 14 20:40:47 CEST 2014 x86_64 GNU/Linux
% pacman -Qi nvidia | grep Version
Version : 334.21-4
"We don't inherit the land from our ancestors - we borrow it from our children."
Offline
Wow,
Can you post how actually you managed to fix this problem?
Seems I've missed some steps ?... I am not sure...
Thank you
Offline
Can you post how actually you managed to fix this problem?
Seems I've missed some steps ?... I am not sure...
Sure, you'll probably have to look at the Arch Linux Kernel Parameters Wiki, because it really depends on what boot loader you are using.
"We don't inherit the land from our ancestors - we borrow it from our children."
Offline
Saultdon,
thanks again, but I have not resolved this problem.
I am using GRUB and after adding to my /etc/default/grub:
GRUB_CMDLINE_LINUX_DEFAULT="quiet rw recutree.rcu_idle_gp_delay=1"
and running
grub-mkconfig -o /boot/grub/grub.cfg
I have still nothing.
So, I stll have no laptop ((
Last edited by iohanes (2014-04-15 19:07:22)
Offline
iohanes, from what I can understand of your post, you do not seem to have the issue described here. Our laptops are actually powering off in an unclean manner. Some others on nvidia forums are having a slightly different issue, where the screen is dark when starting x.
GRUB_CMDLINE_LINUX_DEFAULT="quiet rw recutree.rcu_idle_gp_delay=1"
You seem to have misspelled the boot option there. Furthermore, if you are indeed having the same problem as the rest of us, then you may also want to add the second option I mentioned, CONFIG_HZ=1000 as some people on nvidia forums said they needed both of these options in order to get x functioning.
Offline
Daerandin,
thank you. I am browsing from phone, so I did not notice that your issue is differnt from mine.
Yes there is a typo in boot option but not in my /etc/default/grub ))
Many thanks!
Best,
Iohanes.
Offline
FYI everyone.
Updated to nvidia(extra), nvidia-utils(testing), and nvidia-libgl(testing) to version 337.12-1 and the problem still persists without the needed kernel parameters.
That coughing noise my laptop made when it died was kind of scary!
Not sure if there will be any changes made to nvidia-libgl and nvidia-utils as they make their way to extra, so I tested it out prematurely anyways. I'll test again when they are officially in extra - holding my breath.
EDIT:
Kernel parameter fix is still needed for NVIDIA 337.12-1
Last edited by saultdon (2014-04-16 14:38:34)
"We don't inherit the land from our ancestors - we borrow it from our children."
Offline
Problems with nvidia and new Arch kernels have been giving me fits for a while. I finally ditched nvidia and installed nouveau.
Offline
Adding rcutree.rcu_idle_gp_delay=1 to my kernel parameters with linux-3.14-1 and nvidia-337.12 solved my problem.
I have marked the OP as solved, but not the thread overall as it seems others are still experiencing this problem.
Cheers.
Offline
Adding rcutree.rcu_idle_gp_delay=1 to my kernel parameters with linux-3.14-1 and nvidia-337.12 solved my problem.
Thanks, all. I can confirm that this fix worked for me. For those using rEFInd to boot via EFI, remember that you add kernel parameters in refind_linux.conf.
Offline
I just have to use the kernel boot option:
rcutree.rcu_idle_gp_delay=1
Finally!! I got solved too!!
My laptop is an Asusg75VX. On kernel-3.14.3.1 and Nvidia 337.12
lspci |grep VGA
01:00.0 VGA compatible controller: NVIDIA Corporation GK104M [GeForce GTX 670MX] (rev a1)
do it good first, it will be faster than do it twice the saint
Offline