Siduction Forum
Siduction Forum => Upgrade Warnings => Topic started by: vayu on 2020/10/26, 20:11:05
-
5.8.16-towo.1 still boots fine.
Also KDE is not upgradeable for me (lot's of packages to remove) and has been for a while.
-
kernel 5.9 won't boot on my system
And you think, that helps in any direction?
-
I guess since it works fine with the last one and I haven't troubleshooted why, I guess it doesn't help. I get stuck on simple things sometimes. I just thought someone would appreciate a warning.
-
I've been running your kernels for years now and have been infinitely appreciative of them.
-
~
towo:Defiant> inxi -SGMI
System: Host: Defiant Kernel: 5.9.1-towo.1-siduction-amd64 x86_64 bits: 64 Desktop: KDE Plasma 5.19.5
Distro: siduction 18.3.0 Patience - kde - (202010061355)
Machine: Type: Desktop System: CSL- & KG product: A0000001 v: N/A serial: <superuser/root required>
Mobo: ASUSTeK model: ROG STRIX B450-F GAMING v: Rev 1.xx serial: <superuser/root required>
UEFI: American Megatrends v: 3103 date: 06/17/2020
Graphics: Device-1: Advanced Micro Devices [AMD/ATI] Ellesmere [Radeon RX 470/480/570/570X/580/580X/590] driver: amdgpu
v: kernel
Display: x11 server: X.Org 1.20.9 driver: amdgpu FAILED: ati unloaded: fbdev,modesetting,vesa
resolution: 1920x1080~60Hz
OpenGL: renderer: Radeon RX 580 Series (POLARIS10 DRM 3.39.0 5.9.1-towo.1-siduction-amd64 LLVM 11.0.0)
v: 4.6 Mesa 20.3.0-devel
Info: Processes: 421 Uptime: 11h 05m Memory: 31.36 GiB used: 1.82 GiB (5.8%) Shell: Bash inxi: 3.1.07
So you see, that kernel is working in general.
Let me oracle, you have a nvidia acard and you are using the blob, then have fun, a new driver, which supports kernel 5.9 or higher wil not be avaialable in the near future.
-
you have a nvidia acard and you are using the blob, then have fun, a new driver, which supports kernel 5.9 or higher wil not be avaialable in the near future.
Thank you, that must be my problem.
dkms status shows the 5.9 kernel not having an nvidia module
dkms install nvidia-current/450.66 -k 5.9.0-1-amd64
ends with an error:
Building module: cleaning build area... unset ARCH; env NV_VERBOSE=1 make -j16 modules KERNEL_UNAME=5.9.0-1-amd64....(bad exit status: 2) Error! Bad return status for module build on kernel: 5.9.0-1-amd64 (x86_64) Consult /var/lib/dkms/nvidia-current/450.66/build/make.log for more information.
-
OT: I have a new siduction install (siduction 18.3.0 Patience - kde - (202008212153)
) with a ASUSTeK TUF GAMING B550-PLUS with AMD Ryzen 7 PRO 4750G with Radeon Graphics.
Is kernel 5.9.1 already in the repos? I cannot install it.
-
As allways, new kernels are in our experimental repo.
towo:Defiant> apt policy linux-image-siduction-amd64
linux-image-siduction-amd64:
Installiert: 5.9-1~exp1
Installationskandidat: 5.9-1~exp1
Versionstabelle:
*** 5.9-1~exp1 500
500 https://packages.siduction.org/extra experimental/main amd64 Packages
100 /var/lib/dpkg/status
5.8-16 500
500 https://packages.siduction.org/extra unstable/main amd64 Packages
-
(…)
Let me oracle, you have a nvidia acard and you are using the blob, then have fun, a new driver, which supports kernel 5.9 or higher wil not be avaialable in the near future.
Thanks for your information, I will hold the kernel before I install the 5.9.
NVIDIA Doesn't Expect To Have Linux 5.9 Driver Support For Another Month (https://www.phoronix.com/scan.php?page=news_item&px=NVIDIA-Linux-5.9-Delayed)
-
hold the kernel
...or towo patches (http://rglinuxtech.com/?p=2844) his kernel.
-
...or towo patches his kernel.
This will never happen, since such kernel won't be distributable, that patch violates the licence.
-
Those of Nvidia or the GPL of the Linux kernel?
Well, to want to use a distribution with always up-to-date kernel with a GPU that requires the proprietary driver nvidia is rather unfortunate.
Unlike you, I don't recommend buying a Radeon (everyone has to decide that for himself). If nouveau is enough (here intentionally a GF710, no problems with it), you can use it.
Btw., the thread starter could have called also times its concrete Nvidia GPU.
-
Those of Nvidia or the GPL of the Linux kernel
The kernel one.
This patch only removes the change, that non-gpl modules can use gpl'ed parts of the linux kernel.
Reverting that change clearly violates the kernel licence.
What people do on their own, is their part, but i can't do such thing in a distributed kernel.
-
This patch only removes the change, that non-gpl modules can use gpl'ed parts of the linux kernel.
Thnx for the explanation!
-
Thnx for the explanation!
+1
Speaking only for myself, with 2 Nvidia GPUs in the house, it would be good if release of the 5.9er could wait until the updated driver is available. Otherwise, it will be necessary to pin the 5.8x kernel before the du that installs 5.9.
-
it would be good if release of the 5.9er could wait until the updated driver is available.
And here is the problem, no one knows, if there will such driver available in short time.
I'm not willing to hold back kernel 5.9 only because nvidia blob is not working.
Kernel 5.9 has too important functions that I want or can withhold it for all others.
-
Ich verstehe.
Please, can we have a warning when 5.9 is about to be released, so we can do the pinning?
Thank you.
-
I imagine everyone has noticed towo 5.9 kernels have been working with nvidia just shortly after this. I'm up to 5.9.11 now.
-
"Working" -- yes, but with no CUDA, as I learned on this thread:
https://forum.siduction.org/index.php?topic=8125.0
-
Not exactly right, the driver 455.45.01 works indeed fully on kernel 5.9, now even with cuda.
You can find this driver in debian/experimental.
-
Not exactly right, the driver 455.45.01 works indeed fully on kernel 5.9, now even with cuda.
Well ... hmmmmmmm. I installed 455.45.01 this morning, and checked for the nvidia-uvm module. No nvidia-uvm. My hardware is only a few years old -- but maybe it doesn't support cuda. Thanks.
-
You could look, in the dkms.conf if the uvm-module is activated there.
-
I think it is activated, but maybe something is not working correctly. Here are the modules in dkms.conf:
BUILT_MODULE_NAME[0]="nvidia"
DEST_MODULE_NAME[0]="$PACKAGE_NAME"
DEST_MODULE_LOCATION[0]="/updates/dkms"
BUILT_MODULE_NAME[1]="nvidia-modeset"
DEST_MODULE_NAME[1]="$PACKAGE_NAME-modeset"
DEST_MODULE_LOCATION[1]="/updates/dkms"
BUILT_MODULE_NAME[2]="nvidia-drm"
DEST_MODULE_NAME[2]="$PACKAGE_NAME-drm"
DEST_MODULE_LOCATION[2]="/updates/dkms"
BUILT_MODULE_NAME[3]="nvidia-uvm"
DEST_MODULE_NAME[3]="$PACKAGE_NAME-uvm"
DEST_MODULE_LOCATION[3]="/updates/dkms"
And here is the output of the search for nvidia modules:
don@dibl-patience:~$ lsmod | grep nvidia
nvidia_drm 57344 6
drm_kms_helper 258048 1 nvidia_drm
nvidia_modeset 1220608 17 nvidia_drm
nvidia 28286976 1018 nvidia_modeset
drm 618496 9 drm_kms_helper,nvidia_drm
My hardware:
Graphics: Device-1: NVIDIA GP106 [GeForce GTX 1060 6GB] driver: nvidia v: 455.45.01
Display: x11 server: X.Org 1.20.9 driver: nvidia resolution: 1: 1920x1200~60Hz 2: 1920x1080~60Hz
OpenGL: renderer: GeForce GTX 1060 6GB/PCIe/SSE2 v: 4.6.0 NVIDIA 455.45.01
-
Was it build?
modinfo nvidia-uvm
It is not loaded by default, it will be loaded, if you fire up a application, which use cuda.
-
don@dibl-patience:~$ sudo modinfo nvidia-uvm
[sudo] password for don:
modinfo: ERROR: Module nvidia-uvm not found.
it will be loaded, if you fire up a application, which use cuda.
I don't think I have any application that uses cuda -- that's for serious games, correct?
I have not seen any error in any package while working in KDE/Plasma.
-
don@dibl-patience:~$ sudo modinfo nvidia-uvm
[sudo] password for don:
modinfo: ERROR: Module nvidia-uvm not found.
it will be loaded, if you fire up a application, which use cuda.
I don't think I have any application that uses cuda -- that's for serious games, correct?
I have not seen any error in any package while working in KDE/Plasma.
No not serious games, its more used for things like machine learning or video encoding. For example if you want to user blender or davinci resolve you need cuda. Playing games works fine.