Hi Team, I performed a world update today and seem to have lost all OpenGL support for my wayland-based system (Plasma 6.3.0, kver: 6.13.2-gentoo-dist). `dmesg` shows both the i915 and nvidia modules loading successfully, however attempts to use `eglinfo` to debug result in the following: ``` [ 202.738421] eglinfo[4727]: segfault at 8 ip 00007fc477e2780d sp 00007ffdc2ba7040 error 4 in libgallium-24.3.4.so[2780d,7fc477e0f000+d8e000] likely on CPU 1 (core 1, socket 0) ``` dmesg snippet for nvidia: ``` [ 12.410642] NVRM: loading NVIDIA UNIX x86_64 Kernel Module 570.86.16 Fri Jan 24 21:25:51 UTC 2025 [ 12.457430] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms 570.86.16 Fri Jan 24 20:44:10 UTC 2025 [ 12.462587] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver [ 13.522336] [drm] Initialized nvidia-drm 0.0.0 for 0000:01:00.0 on minor 1 ``` same for i915: ``` [ 12.174345] i915 0000:00:02.0: [drm] Found coffeelake (device ID 3e98) display version 9.00 stepping N/A [ 12.175561] i915 0000:00:02.0: vgaarb: deactivate vga console [ 12.197198] i915 0000:00:02.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem [ 12.197616] Loading firmware: i915/kbl_dmc_ver1_04.bin [ 12.198514] i915 0000:00:02.0: [drm] Finished loading DMC firmware i915/kbl_dmc_ver1_04.bin (v1.4) [ 13.040019] i915 0000:00:02.0: [drm] [ENCODER:98:DDI A/PHY A] failed to retrieve link info, disabling eDP [ 13.044637] mei_hdcp 0000:00:16.0-b638ab7e-94e2-4ea2-a552-d1c54b627f04: bound 0000:00:02.0 (ops i9xx_always_on_power_well_ops [i915]) [ 13.135584] [drm] Initialized i915 1.6.0 for 0000:00:02.0 on minor 0 [ 13.196477] fbcon: i915drmfb (fb0) is primary device [ 13.266668] i915 0000:00:02.0: [drm] fb0: i915drmfb frame buffer device [ 13.306131] snd_hda_intel 0000:00:1f.3: bound 0000:00:02.0 (ops i915_fence_ops [i915]) ``` Happy to provide additional troubleshooting info, only have access (directly) to this workstation a few days a week, can SSH most days.
Can you try to get a backtrace via e.g. coredumpctl debug? Also, emerge --info
Downgraded mesa to `media-libs/mesa-24.2.8` - still only llvmpipe but eglinfo stopped segfaulting? Will upload core dump if I still have it, or repro. `emerge --info` to follow.
Created attachment 918903 [details] emerge --info
backtrace (had to rebuild with debug symbols): ``` Program terminated with signal SIGSEGV, Segmentation fault. #0 0x00007efc8ba2780d in drisw_init_screen (screen=0x5623f0292100, driver_name_is_inferred=false) at ../mesa-24.3.4/src/gallium/frontends/dri/drisw.c:623 623 if (loader->base.version >= 4) { (gdb) bt #0 0x00007efc8ba2780d in drisw_init_screen (screen=0x5623f0292100, driver_name_is_inferred=false) at ../mesa-24.3.4/src/gallium/frontends/dri/drisw.c:623 #1 0x00007efc8ba22d75 in driCreateNewScreen3 (scrn=scrn@entry=0, fd=<optimized out>, loader_extensions=0x7efc8d4631c0 <image_loader_extensions>, type=type@entry=DRI_SCREEN_SWRAST, driver_configs=driver_configs@entry=0x5623f0355780, driver_name_is_inferred=driver_name_is_inferred@entry=false, has_multibuffer=true, data=0x5623f1a59e00) at ../mesa-24.3.4/src/gallium/frontends/dri/dri_util.c:144 #2 0x00007efc8d42b79f in dri2_create_screen (disp=disp@entry=0x5623f1a59e00) at ../mesa-24.3.4/src/egl/drivers/dri2/egl_dri2.c:825 #3 0x00007efc8d42ce5b in dri2_initialize_device (disp=disp@entry=0x5623f1a59e00) at ../mesa-24.3.4/src/egl/drivers/dri2/platform_device.c:362 #4 0x00007efc8d42bfd8 in dri2_initialize (disp=0x5623f1a59e00) at ../mesa-24.3.4/src/egl/drivers/dri2/egl_dri2.c:912 #5 0x00007efc8d41abe4 in eglInitialize (dpy=<optimized out>, major=0x7ffde248cc2c, minor=0x7ffde248cc30) at ../mesa-24.3.4/src/egl/main/eglapi.c:719 #6 0x00005623c2be5c5c in ?? () #7 0x00005623c2be4a7d in ?? () #8 0x00007efc93a871ee in ?? () from /usr/lib64/libc.so.6 #9 0x00007efc93a872a9 in __libc_start_main () from /usr/lib64/libc.so.6 #10 0x00005623c2be4b05 in ?? () ```
Have tried kernels as far back as 6.12.8 (which I was running for most of this month) - no change.
At ionen's suggestion, purged the .ko files for nvidia drivers and rebooted. Aside from not being able to initialise screens plugged into my nvidia GPU this appears to have gotten i915 recognised again: ``` $ glxinfo -B . . . Extended renderer info (GLX_MESA_query_renderer): Vendor: Intel (0x8086) Device: Mesa Intel(R) UHD Graphics 630 (CFL GT2) (0x3e98) . . . ```
The eglinfo output should end with: > Device platform: > Device #0: > [...] # normal nvidia output > Device #1: > > Platform Device platform: > nouveau: driver missing > libEGL warning: egl: failed to create dri2 screen > nouveau: driver missing > libEGL warning: egl: failed to create dri2 screen > [1] 25260 segmentation fault (core dumped) eglinfo -B The attempt to use nouveau ( which isn't even build with my mesa ) can be disabled using: > __EGL_VENDOR_LIBRARY_FILENAMES=/usr/share/glvnd/egl_vendor.d/10_nvidia.json eglinfo -B or by enforcing zink (if enabled) via > MESA_LOADER_DRIVER_OVERRIDE=zink eglinfo -B The referenced bug has some additional details. The zink part is missing.
That's exactly what I'm seeing.
It’s related to bug 949411, and in upstream bug 12462 https://gitlab.freedesktop.org/mesa/mesa/-/issues/12462.
Also, based on `strings /usr/lib64/libEGL_mesa.so.0.0.0 | grep -Fie 'failed to create dri2 screen' -`, the message `egl: failed to create dri2 screen` comes from libEGL_mesa instead of libgallium. For notice, I tried commit from upstream in merged MR https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/33123, it did not improve this issue.
