Bugzilla – Bug 1210598
i915 [Kabylake] eglinfo, broken
Last modified: 2023-11-18 10:48:41 UTC
Created attachment 866387 [details] eglinfo crash message when using the kde "data center" when selecting "opengl (EGL)" i get a message "eglinfo tool crashed ..." when using eglinfo i get : :~> eglinfo EGL client extensions string: EGL_EXT_client_extensions, EGL_EXT_device_base, EGL_EXT_device_enumeration, EGL_EXT_device_query, EGL_EXT_explicit_device, EGL_EXT_platform_base, EGL_EXT_platform_device, EGL_EXT_platform_wayland, EGL_EXT_platform_x11, EGL_EXT_platform_xcb, EGL_KHR_client_get_all_proc_addresses, EGL_KHR_debug, EGL_KHR_platform_gbm, EGL_KHR_platform_wayland, EGL_KHR_platform_x11, EGL_MESA_platform_gbm, EGL_MESA_platform_surfaceless ... EGL_NV_stream_reset, EGL_NV_stream_socket, EGL_NV_stream_socket_inet, EGL_NV_stream_socket_unix, EGL_NV_stream_sync, EGL_NV_system_time, EGL_NV_triple_buffer, EGL_WL_bind_wayland_display, EGL_WL_wayland_eglstream corrupted size vs. prev_size Abandon (core dumped)
Corruption in eglinfo, probably caused by the driver. This looks like the NVIDIA proprietary driver. Reassigning.
First let's check which kind of hardware is this. Please add the output of `glxinfo` and `inxi -aG`,
Created attachment 866391 [details] glxinfo
Created attachment 866392 [details] inxi info
Hardware: Intel HD Graphics 630 / NVIDIA GP108 [GeForce GT 1030] glxinfo [...] OpenGL renderer string: Mesa Intel(R) HD Graphics 630 (KBL GT2) [...] So Optimus Laptop. nVidia driver/suse-prime not in use.
It is a desktop PC. The monitor is connected to the Intel iGPU. The Nvidia dGPU is used only for computing. I don't need Prime technology.
Could you verify this with our latest Mesa version of TW, please. This might have been fixed silently by this change ... ------------------------------------------------------------------- Fri Jun 16 11:01:52 UTC 2023 - Joan Torres <joan.torres@suse.com> - Fix of bsc#1212345 * After update to 23.1.1 this change affected our build split: https://gitlab.freedesktop.org/mesa/mesa/-/commit/2630f4b310f534c830250be44d285e485da83869
mesa 23.1.1 problem not solved mesa 23.1.2 problem not solved
Ok. So I guess you see the changelog entry in comment#7 when running 'rpm --changelog -q Mesa-dri', right?
Created attachment 867701 [details] mesa-dri change log
Yes, it's already included.
*** Bug 1211739 has been marked as a duplicate of this bug. ***
Can reproduce on nvidia 3080 Ti: glxinfo: https://paste.opensuse.org/pastes/f07ea7483e94 ``` $ inxi -aG Graphics: Device-1: NVIDIA GA102 [GeForce RTX 3080 Ti] vendor: Gigabyte driver: nvidia v: 535.86.05 alternate: nouveau,nvidia_drm non-free: 530.xx+ status: current (as of 2023-05) arch: Ampere code: GAxxx process: TSMC n7 (7nm) built: 2020-22 pcie: gen: 1 speed: 2.5 GT/s lanes: 16 link-max: gen: 4 speed: 16 GT/s ports: active: none off: DP-2 empty: DP-1, DP-3, HDMI-A-1, HDMI-A-2 bus-ID: 2b:00.0 chip-ID: 10de:2208 class-ID: 0300 Display: wayland server: X.org v: 1.21.1.8 with: Xwayland v: 23.1.2 compositor: kwin_wayland driver: X: loaded: nvidia unloaded: fbdev,modesetting,vesa alternate: nouveau,nv gpu: nvidia display-ID: 0 Monitor-1: DP-2 res: 1920x1080 size: N/A modes: N/A API: OpenGL v: 4.6.0 NVIDIA 535.86.05 renderer: NVIDIA GeForce RTX 3080 Ti/PCIe/SSE2 direct-render: Yes ```
Here is the crash stacktrace captured by gdb: ``` GL_OES_texture_stencil8, GL_OES_texture_storage_multisample_2d_array, GL_OES_texture_view, GL_OES_vertex_array_object, GL_OES_vertex_half_float, GL_OES_viewport_array, GL_OVR_multiview, GL_OVR_multiview2, GL_OVR_multiview_multisampled_render_to_texture corrupted size vs. prev_size Program received signal SIGABRT, Aborted. __pthread_kill_implementation (threadid=<optimized out>, signo=signo@entry=6, no_tid=no_tid@entry=0) at pthread_kill.c:44 Downloading source file /usr/src/debug/glibc-2.38/nptl/pthread_kill.c 44 return INTERNAL_SYSCALL_ERROR_P (ret) ? INTERNAL_SYSCALL_ERRNO (ret) : 0; (gdb) bt #0 __pthread_kill_implementation (threadid=<optimized out>, signo=signo@entry=6, no_tid=no_tid@entry=0) at pthread_kill.c:44 #1 0x00007ffff7c91ed3 in __pthread_kill_internal (signo=6, threadid=<optimized out>) at pthread_kill.c:78 #2 0x00007ffff7c3f126 in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26 #3 0x00007ffff7c26917 in __GI_abort () at abort.c:79 #4 0x00007ffff7c277e3 in __libc_message (fmt=fmt@entry=0x7ffff7db056d "%s\n") at ../sysdeps/posix/libc_fatal.c:150 #5 0x00007ffff7c9c827 in malloc_printerr (str=str@entry=0x7ffff7dadfb8 "corrupted size vs. prev_size") at malloc.c:5879 #6 0x00007ffff7c9d34e in unlink_chunk (p=<optimized out>, av=<optimized out>) at malloc.c:1605 #7 0x00007ffff7c9d4e0 in malloc_consolidate (av=av@entry=0x7ffff7e4bac0 <main_arena>) at malloc.c:4842 #8 0x00007ffff7c9ebd0 in _int_free (av=0x7ffff7e4bac0 <main_arena>, p=0x55555570f790, have_lock=<optimized out>, have_lock@entry=0) at malloc.c:4736 #9 0x00007ffff7ca1873 in __GI___libc_free (mem=<optimized out>) at malloc.c:3386 #10 0x00007ffff5d33166 in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #11 0x00007ffff7204e80 in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #12 0x0000000040000000 in ?? () #13 0x00007ffff7204e80 in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #14 0x00007ffff5ca7aca in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #15 0x0000000040000000 in ?? () #16 0x00007ffff5a3b70d in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #17 0x00007ffff7207a40 in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #18 0x00007ffff7207a40 in ?? () from /lib64/libnvidia-eglcore.so.535.86.05 #19 0x0000000040000000 in ?? () #20 0x0000000000000000 in ?? () ```
Nvidia claimed that they have root caused the issue and fix will be available in the next branch drivers: https://forums.developer.nvidia.com/t/eglinfo-command-crashes-on-opensuse-tumbleweed-nvidia-535-86-05/263913/3
(In reply to Xu Zhao from comment #15) > Nvidia claimed that they have root caused the issue and fix will be > available in the next branch drivers: > https://forums.developer.nvidia.com/t/eglinfo-command-crashes-on-opensuse- > tumbleweed-nvidia-535-86-05/263913/3 Thanks. But comment#14 was not the initital issue. Initial issue was on Intel gfx hardware.
(In reply to Stefan Dirsch from comment #16) > (In reply to Xu Zhao from comment #15) > > Nvidia claimed that they have root caused the issue and fix will be > > available in the next branch drivers: > > https://forums.developer.nvidia.com/t/eglinfo-command-crashes-on-opensuse- > > tumbleweed-nvidia-535-86-05/263913/3 > > Thanks. But comment#14 was not the initital issue. Initial issue was on > Intel gfx hardware. The initial issue was on `Hardware: Intel HD Graphics 630 / NVIDIA GP108 [GeForce GT 1030]`. Only after NVIDIA fixes their driver could we know if there is another issue on Intel gfx hardware or not.
Sorry, I missed that. OTOH suse-prime is not in use. But let's wait for next branch drivers by nVidia ...
I can confirm that this issue has been fixed in NVIDIA 545.29.02.
Thanks for checking again. I think it was the update of libnvidia-egl-wayland.so from 1.1.11 to 1.1.12. Actually in TW, we have this in a separate package called libnvidia-egl-wayland1, which I recently updated even to 1.1.13.
Problem solved.