Closed markusressel closed 6 months ago
@markusressel did you manage to find a solution? I have the same hardware and I am facing this issue as well
@sagittaros No :cry:
I too am having this problem. I was able to reproduce this problem both on Arch, EndeavourOS, Manjaro
Hardware: Dell XPS 9510, Integrated Intel & RTX 3050
Xorg log
[ 13.656] (II) systemd-logind: took control of session /org/freedesktop/login1/session/_31
[ 13.657] (II) xfree86: Adding drm device (/dev/dri/card0)
[ 13.657] (II) Platform probe for /sys/devices/pci0000:00/0000:00:02.0/drm/card0
[ 13.657] (II) systemd-logind: got fd for /dev/dri/card0 226:0 fd 11 paused 0
[ 13.659] (--) PCI:*(0@0:2:0) 8086:9a60:1028:0a61 rev 1, Mem @ 0x6164000000/16777216, 0x4000000000/268435456, I/O @ 0x00004000/64, BIOS @ 0x????????/131072
[ 13.659] (--) PCI: (1@0:0:0) 10de:25a2:1028:0a61 rev 161, Mem @ 0xa0000000/16777216, 0x6000000000/4294967296, 0x6100000000/33554432, I/O @ 0x00003000/128
[ 13.659] (WW) Open ACPI failed (/var/run/acpid.socket) (No such file or directory)
[ 13.659] (II) LoadModule: "glx"
[ 13.660] (II) Loading /usr/lib/xorg/modules/extensions/libglx.so
[ 13.664] (II) Module glx: vendor="X.Org Foundation"
[ 13.664] compiled for 1.21.1.3, module version = 1.0.0
[ 13.664] ABI class: X.Org Server Extension, version 10.0
[ 13.664] (II) LoadModule: "modesetting"
[ 13.664] (II) Loading /usr/lib/xorg/modules/drivers/modesetting_drv.so
[ 13.667] (II) Module modesetting: vendor="X.Org Foundation"
[ 13.667] compiled for 1.21.1.3, module version = 1.21.1
[ 13.667] Module class: X.Org Video Driver
[ 13.667] ABI class: X.Org Video Driver, version 25.2
[ 13.667] (II) modesetting: Driver for Modesetting Kernel Drivers: kms
[ 13.667] (WW) Falling back to old probe method for modesetting
[ 13.667] (II) modeset(0): using default device
[ 13.667] (II) modeset(G0): using drv /dev/dri/card0
[ 13.667] (WW) VGA arbiter: cannot open kernel arbiter, no multi-card support
[ 13.667] (EE)
Fatal server error:
[ 13.667] (EE) Cannot run in framebuffer mode. Please specify busIDs for all framebuffer devices
[ 13.667] (EE)
[ 13.667] (EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
[ 13.667] (EE) Please also check the log file at "/home/alphard/.local/share/xorg/Xorg.0.log" for additional information.
[ 13.667] (EE)
[ 13.691] (EE) Server terminated with error (1). Closing log file.
Optimus Manager log
[8] INFO: # Xorg pre-start hook
[8] INFO: Previous state was: {'type': 'pending_pre_xorg_start', 'requested_mode': 'integrated', 'current_mode': None}
[8] INFO: Requested mode is: integrated
[174] INFO: Available modules: ['nouveau', 'nvidia', 'nvidia_drm', 'nvidia_modeset', 'nvidia_uvm']
[174] INFO: Unloading modules ['nvidia_drm', 'nvidia_modeset', 'nvidia_uvm', 'nvidia'] (if loaded)
[178] INFO: switching=none, nothing to do
[1126] INFO: Writing to /etc/X11/xorg.conf.d/10-optimus-manager.conf
[1126] INFO: Writing state {'type': 'pending_post_xorg_start', 'switch_id': '20220209T130719', 'requested_mode': 'integrated'}
[1126] INFO: Xorg pre-start hook completed successfully.
I have tried with gdm, sddm, and startx, following all necessary steps in the optimus-manager wiki.
I'm also having the exact same issue on a ASUS TUF laptop with RTX 3050 mobile GPU and Intel Tiger Lake graphics. Would be nice to know how to fix this.
Having the same issue since last nvidia update, on new hp spectre 16 inch, with rtx3050 and tiger lake on arch.
Anybody found a solution?
chiming in to report the same issue on my new laptop with the following specs:
CPU: intel i7-11370H GPU: nvidia GeForce RTX 3050 Kernel: 5.18.3-arch1-1 Nvidia Driver Version: 515.48.07-9
@markusressel could you check if running startx
after removing the autogenerated file at /etc/X11/xorg.conf.d/10-optimus-manager.conf
solves the issue? it's working on my end.
update: that was an error on my end i hadn't enabled the optimus manager service. sorry for the false alarm.
I have the same issue. I managed to log in by switching to a different tty using alt+ctrl+F2, and then enabling gdm with sudo systemctl enable gdm.service -f
. Then, I rebooted.
Gnome and GDM uses wayland by default and this issue is with X11. So this solution is more like a workaround. Since it is wayland, I still cannot get optimus-manager to work as it only works with X11.
If anyone has any solutions, then I'd finally be able to switch to Arch.
@chebro No. I upgraded to the latest 5.18.6 kernel yesterday, tried to install optimus manager again, and no luck. Just a black screen.
Besides, what would be the point? Optimus manager is supposed to use its config to switch modes. Removing its config will just break it.
Same issue here. I've given up for now, I'll use WSL on Windows until this problem goes away. I couldn't setup Optimus to reduce power draw on my Dell XPS 9710 with the NVIDIA RTX 3050 GPU. I got only 5 hours of usage, compared to 7+ hours on Windows.
Same issue also on a ZenBook Pro Duo UX582ZM, does not run on integrated, nor in any other setting. The only way to get i3 working, for me, is to keep optimus-manager running, but before starting x from terminal I have to manually delete
/etc/X11/xorg.conf.d/10-optimus-manager.conf
which looks exactly like the one posted at the beginning of this report. If I do not delete it, I get this:
[ 11.995] (EE) Fatal server error: [ 11.995] (EE) Cannot run in framebuffer mode. Please specify busIDs for all framebuffer devices
One last thing, if you are running kernel 5.18, try passing the kernel parameter ibt=off.
Due to FS#74886/FS#74891, nvidia does not boot on Linux 5.18 on systems with Intel CPUs. Until this is fixed, a workaround is disabling the Indirect Branch Tracking CPU security feature by setting the ibt=off kernel parameter from the bootloader. This security feature is responsible for mitigating a class of exploit techniques, but is deemed safe as a temporary stopgap solution. You can alternatively try nvidia-open and continue using IBT.
source: https://wiki.archlinux.org/title/NVIDIA#Installation
I have a similar laptop (with a RTX 3080), and I was able to fix this by changing the BusID
s in /etc/X11/xorg.conf.d/10-optimus-manager.conf to
Section "Device"
Identifier "integrated"
Driver "modesetting"
BusID "PCI:0:2:0"
EndSection
Section "Device"
Identifier "nvidia"
Driver "nvidia"
BusID "PCI:1:0:0"
EndSection
After restarting gdm
I see it's the primary vendor
glxinfo | grep vendor
server glx vendor string: NVIDIA Corporation
client glx vendor string: NVIDIA Corporation
OpenGL vendor string: NVIDIA Corporation
I don't think 4 numerical identifiers has ever been valid for the PCI bus in Xorg which is what lead me to try it; I haven't looked at the code, but it probably just needs to trim the identifier down (it would be interesting to see if other laptops work with 4, or if they're getting the expected 3 number identifier)
edit:
Ahh it looks like there is a check for this, but it's disabled in code https://github.com/Askannz/optimus-manager/blob/master/optimus_manager/pci.py#L31
Describe the bug Hi there,
I am in the process of setting up an Asus ROG Zephyrus M16 2021 laptop with an Intel Tiger Lake Graphics and an Nvidia RTX 3070 Mobile.
I followed the advice of the README to remove
/etc/X11/xorg.conf
, however I do not get a picture without letting xorg generate something itself usingsudo Xorg :0 -configure
.Even then, switching only results in a black screen with a blinking cursor. I have tried switching between TTY's, but it doesn't help, the screen stays black.
Let me know if you need further details, thx.
System info
Logs
/usr/share/X11/xorg.conf.d/10-nvidia-drm-outputclass.conf
/etc/X11/xorg.conf.d/10-optimus-manager.conf
/etc/optimus-manager/optimus-manager.conf
/home/markus/.local/share/xorg/Xorg.0.log
optimus-manager.service:
/var/log/optimus-manager/switch/switch-20211115T011615.log