On 19.01.2025 05:28, Felix Miata wrote:
It seems to me you probably need to go backport. I tried the "experimental"
equivalent
with i5-11400's Rocket Lake S GT1 (Gen12.1) in openSUSE Tumbleweed, resulting
in loss
of about 2/3 in graphics performance testing with glmark2.
Thanks for reminding me about new xe driver, but unfortunately I wasn't
able to get it working.
I've tried 6.11 kernel from backports, but it look like "xe" module is
absent and only "i915" available.
$ uname -a
Linux hostname 6.11.10+bpo-amd64 #1 SMP PREEMPT_DYNAMIC Debian
6.11.10-1~bpo12+1 (2024-12-19) x86_64 GNU/Linux
$ sudo modinfo i915 --filename
/lib/modules/6.11.10+bpo-amd64/kernel/drivers/gpu/drm/i915/i915.ko.xz
$ sudo modinfo xe --filename
modinfo: ERROR: Module xe not found.
I've still tried "force_probe" parameters you suggested, but with them
Xorg server fails to start.
Relevant lines from /var/log/Xorg.0.log :
...
[ 4.777] Current Operating System: Linux hostname 6.11.10+bpo-amd64
#1 SMP PREEMPT_DYNAMIC Debian 6.11.10-1~bpo12+1 (2024-12-19) x86_64
[ 4.777] Kernel command line: BOOT_IMAGE=/vmlinuz-6.11.10+bpo-amd64
root=/dev/mapper/vg0--nvme-lv0--debian--root ro intel_iommu=on
mitigations=off i915.force_probe=!46a3 xe.force_probe=46a3 quiet splash
[ 4.777] xorg-server 2:21.1.7-3+deb12u8 (https://www.debian.org/support)
...
[ 4.778] (--) PCI:*(0@0:2:0) 8086:46a3:1b50:1018 rev 12, Mem @
0x6001000000/16777216, 0x4000000000/268435456, I/O @ 0x00003000/64, BIOS
@ 0x????????/131072
[ 4.778] (II) LoadModule: "modesetting"
[ 4.778] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[ 4.778] (II) Module modesetting: vendor="X.Org Foundation"
[ 4.778] compiled for 1.21.1.7, module version = 1.21.1
[ 4.778] Module class: X.Org Video Driver
[ 4.778] ABI class: X.Org Video Driver, version 25.2
[ 4.778] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[ 4.778] (EE) open /dev/dri/card0: No such file or directory
[ 4.778] (WW) Falling back to old probe method for modesetting
[ 4.778] (EE) open /dev/dri/card0: No such file or directory
[ 4.778] (EE) Screen 0 deleted because of no matching config section.
[ 4.778] (II) UnloadModule: "modesetting"
[ 4.778] (EE) Device(s) detected, but none match those in the config
file.
[ 4.778] (EE)
Fatal server error:
[ 4.778] (EE) no screens found(EE)
...
Maybe backported kernel was built without new xe driver support?
Once again, I need to know the answer to the question from Subj.
According to Phoronix, new xe driver should be available from kernel
version 6.8 and up.
https://www.phoronix.com/news/Linux-6.8-DRM
For now I'll try to fiddle with i915 module options and see if they will
fix the issue.
As last resort, there is an option to install out-of-tree i915 driver
from Intel, which builds DKMS module. I don't know how different it is
from native kernel sources, though.
I can't comprehend how these bugs are not fixed after so many years.
First patch acknowledging this problem was added to kernel in 2016 and
patches in 2021.
https://patchwork.kernel.org/project/intel-gfx/patch/1481120559-17413-1-git-send-email-joonas.lahti...@linux.intel.com/
https://lists.freedesktop.org/archives/intel-gfx/2021-August/274340.html
Why there are "Uninitialized forcewake domain(s)"? What piece of
software should initialize them? What they even are?
If these "domains" were described by someone in code of a driver, did
they forgot to add initialization procedures for them?
Judging by "gpu hang" recovery procedure, Intel graphics drivers seems
to be similar garbage quality as AMD drivers. :(
--
With kindest regards, Alexander.
Debian - The universal operating system
https://www.debian.org
