Hello Adrian and Christoph, On 1/23/24 1:43 PM, John Paul Adrian Glaubitz wrote: > Hello Christoph, > > On Tue, 2024-01-23 at 21:21 +0100, Christoph wrote: >> here the log files for Xorg.0.log, .xsession-errors and dmesg command. There >> seem to be a lot of errors in it that indicate a bug with the nouveau driver. > > Looking at the output of "dmesg", this actually indicates problems with the > nouveau kernel module: > > [ 27.578040] nouveau 0000:0a:00.0: gr: intr 00100000 [ERROR] nsource > 00000002 [DATA_ERROR] nstatus 02000000 [BAD_ARGUMENT] ch 0 [000c1030 > (udev-worker)[171]] subc 4 class 0039 mthd 0314 data > 00010000 > [ 27.578078] nouveau 0000:0a:00.0: gr: intr 00100000 [ERROR] nsource > 00000002 [DATA_ERROR] nstatus 02000000 [BAD_ARGUMENT] ch 0 [000c1030 > (udev-worker)[171]] subc 4 class 0039 mthd 0318 data > 00010000 > [ 27.745010] nouveau 0000:0a:00.0: gr: intr 00100000 [ERROR] nsource > 00000002 [DATA_ERROR] nstatus 02000000 [BAD_ARGUMENT] ch 0 [000c1030 > (udev-worker)[171]] subc 4 class 0039 mthd 0314 data > 00010000 > [ 27.745077] nouveau 0000:0a:00.0: gr: intr 00100000 [ERROR] nsource > 00000002 [DATA_ERROR] nstatus 02000000 [BAD_ARGUMENT] ch 0 [000c1030 > (udev-worker)[171]] subc 4 class 0039 mthd 0318 data > 00010000 > > As far as I know, nouveau is not well supported on big-endian systems. > > You can try reporting the bug here: > > https://gitlab.freedesktop.org/drm/nouveau/-/issues > > If you have the possibility, I would recommend replacing the graphics card > with a Radeon-based chip which is much better supported on big-endian systems. > > Also, this indicates that the the nouveau XOrg driver is not installed: > > [ 12.378] (II) LoadModule: "nouveau" > [ 12.379] (WW) Warning, couldn't open module nouveau > [ 12.379] (EE) Failed to load module "nouveau" (module does not exist, 0) > > You are using the modeset XOrg driver instead. > > So, try installing the package xserver-xorg-video-nouveau and see if that > helps: > > # apt install xserver-xorg-video-nouveau > >> Unfortunately the system did not create any files and/or folders in /var/log >> named syslog. > > You can try something like: > > # journalctl > all_logs.txt [1] > > However, that should not be necessary as we already know what's wrong. > > Adrian > >> [1] https://www.putorius.net/export-logs-to-a-file-with-journalctl.html >
On my PowerMac G5, I had to replace a failed Radeon card with a low-end Nvidia card, and I was no longer able to test Debian powerpc64 kernels because the kernel was missing the needed CONFIG options: *** Aug 31, 2023 kernel (doesn't work on my G5) *** $ fgrep -i nvidia config-6.4.0-4-powerpc64 CONFIG_NET_VENDOR_NVIDIA=y # CONFIG_I2C_NVIDIA_GPU is not set # NVidia media platform drivers *** Dec 22, 2023 (also doesn't work) *** $ fgrep -i nvidia config-6.6.8-powerpc64 CONFIG_NET_VENDOR_NVIDIA=y # CONFIG_I2C_NVIDIA_GPU is not set # NVidia media platform drivers # CONFIG_HID_NVIDIA_SHIELD is not set Compiling a custom kernel with the following CONFIG options works: $ fgrep -i nvidia config-6.5.1-ppc64 # CONFIG_NET_VENDOR_NVIDIA is not set # CONFIG_I2C_NVIDIA_GPU is not set CONFIG_FB_NVIDIA=y CONFIG_FB_NVIDIA_I2C=y # CONFIG_FB_NVIDIA_DEBUG is not set CONFIG_FB_NVIDIA_BACKLIGHT=y $ dmesg | grep -i nvidia [ 0.264634] nvidiafb: Device ID: 10de0321 [ 0.264666] nvidiafb 0000:f0:10.0: vgaarb: deactivate vga console [ 0.285134] nvidiafb: CRTC0 analog not found [ 0.309133] nvidiafb: CRTC1 analog found [ 1.153130] nvidiafb: EDID found from BUS2 [ 1.153154] nvidiafb: CRTC 1 appears to have a CRT attached [ 1.153162] nvidiafb: Using CRT on CRTC 1 [ 1.157999] nvidiafb: PCI nVidia NV32 framebuffer (64MB @ 0x98000000) -Stan
