0

I had this problem when i installed Ubuntu 20.04, i manage to solve it by installing Ubuntu again , and installing old drivers. Now a couple months later, i suspend my computer and get a black screen after i come back.

uname -a

Linux noblepc 5.8.0-55-generic #62~20.04.1-Ubuntu SMP Wed Jun 2 08:55:04 UTC 2021 x86_64 x86_64 x86_64 GNU/Linux

Nvidia Driver - 460.80 Cant install old drivers , without them getting updated automatically!

GPU - GTX 1050

What is strange is that i had this problem solved , and for no reason it came back, and now i cant manage to solve the problem , even after a bunch of solutions.

Error

Nvidia-Modeset: Failed to allocate display engine core DMA push buffer!
Nvidia-Modeset: GPU 0: Display engine push buffer channel allocation failed 0x65

Edit:

I Managed to solve this problem installing Driver-server

  • I don't know if you tried it (you may have as it would have been the first thing I would have tried given your description), but did you try just using the GA kernel? ie. remain on 5.4 for the life of Ubuntu 20.04 rather than the HWE stack. – guiverc Jun 11 '21 at 11:58
  • @guiverc, im really new to linux , i dont really know what GA kernel is or even HWE stack. – Filipe Santos Jun 11 '21 at 12:10
  • GA~=general kernel, the initial kernel that is supplied with a LTS release, ie. 5.4 for Ubuntu 20.04 LTS. The HWE (hardware enablement) kernel stack changes allowing later kernel modules (drivers) for newer hardware), ie. 5.4 gets updated to 20.10's 5.8 kernel at 20.04.2, then to 21.04's 5.11 kernel at 20.04.3 etc.. ie. the HWE stack changes kernel until it reaches and settles on the GA kernel of the next LTS which is 22.04 (at 20.04.5). See https://wiki.ubuntu.com/Kernel/LTSEnablementStack (more detail can also be learnt at https://wiki.ubuntu.com/Kernel/RollingLTSEnablementStack) – guiverc Jun 11 '21 at 12:17
  • Many times it happens that the computer tries to hibernate instead of suspend. And if you don't have enough swap memory available that causes a crash. Make sure you have enough swap at least enough free swap memory as memory you are using. You can use command free -h – Carles Mateo Jun 11 '21 at 14:15
  • @CarlesMateo i have about 2GB of Swap memory – Filipe Santos Jun 11 '21 at 14:55
  • I answer you in an answer in order to add images. – Carles Mateo Jun 12 '21 at 11:43

3 Answers3

0

as told previously

If you have only 2 GB of SWAP and the computer attempts to hibernate you'll get it frozen for sure. Your computer has 4GB of RAM, 8GB of RAM or more for sure.

In this example I have create a VM with Virtual Box with 3.3GB of RAM and 2GB of swap and forced hibernation and sleep. Try to locate in your logs if you have something like this:

enter image description here

cat /var/log/syslog | grep -i "hibern\|suspen"

If you find messages like hibernation is likely your computer is attempting to hibernate, and with you swap memory this will almost ever cause a crash explaining your black screen after trying to resume.

This is cause by insufficient swap memory to map the RAM memory.

Sleep keeps the computer on, spending a very little amount of energy to keep what is in RAM memory.

Hibernate swaps all the RAM memory contents to the swap disk and completely shutdowns the computer, spending no energy until it is turned on.

Cheers

Carles Mateo
  • 1,577
  • 6
  • 12
0

I had the same issue and resolved it by installing nvidia-450-server driver as suggested in this post

sudo apt purge nvidia-*
sudo apt autoremove
sudo apt install nvidia-driver-450-server
Richard
  • 102
0

I have an XMG neo 15 m20 with an RTX 2070 and I struggled with this problem for months after a kernel update in november 2021.
The only thing that worked for me -and I tried a lot of stuff- was switching to discrete-graphics-only mode in the BIOS. It was set to "msHybrid" and I switched it to "dGPU". I think it's an ACPI power mangement issue. I still haven't measured the inpact of using dGPU only on battery life.