[Bug] ubuntu下使用mihomo-party时, Nvidia GPU出错

[duchenpaul]

Verify steps

• 我已在标题简短的描述了我所遇到的问题
• 我已在 Issue Tracker 中寻找过我要提出的问题，但未找到相同的问题
• 我已在 常见问题 中寻找过我要提出的问题，并没有找到答案
• 这是 GUI 程序的问题，而不是内核程序的问题
• 我已经关闭所有杀毒软件/代理软件后测试过，问题依旧存在
• 我已经使用最新的测试版本测试过，问题依旧存在

操作系统

Linux

系统版本

Ubuntu 24.04.1 LTS

发生问题 mihomo-party 版本

v1.5.12

描述

在Ubuntu下使用v1.5.12, GPU在持续报错, 具体表现形式是用显示器时没有显著问题, 但是使用rdp连接时, rdp连接中断, 后台log显示GPU出错重置导致rdp连接中断, GPU出错跟mihomo party有关, 如下:
syslog:

2025-01-08T21:08:58.635295+08:00 titan kernel: netlink: 'mihomo': attribute type 22 has an invalid length.
2025-01-08T21:08:58.792278+08:00 titan kernel: mihomo-party[37391]: segfault at 8 ip 00007ed795d5914d sp 00007ffe1ce8c9a0 error 4 in libGLX_nvidia.so.565.57.01[7ed795d27000+5c000] likely on CPU 10 (core 20, socket 0)
2025-01-08T21:08:58.792288+08:00 titan kernel: Code: 00 75 3a 48 8b 83 a8 10 00 00 48 85 c0 74 28 48 8b 4d 08 eb 11 0f 1f 84 00 00 00 00 00 48 8b 40 50 48 85 c0 74 11 48 8b 50 08 <48> 39 4a 08 75 ed 48 89 83 b0 10 00 00 5b 5d 41 5c c3 90 48 89 df
2025-01-08T21:08:58.875795+08:00 titan mihomo-party.desktop[36984]: [36984:0108/210858.875728:ERROR:gpu_process_host.cc(982)] GPU process exited unexpectedly: exit_code=139
...
2025-01-08T21:08:59.196653+08:00 titan mihomo-party.desktop[36984]: [36984:0108/210859.196612:ERROR:gpu_process_host.cc(982)] GPU process exited unexpectedly: exit_code=139
2025-01-08T21:08:59.248819+08:00 titan systemd[21334]: snap.snapd-desktop-integration.snapd-desktop-integration.service: Scheduled restart job, restart counter is at 1.
2025-01-08T21:08:59.271582+08:00 titan systemd[21334]: Started snap.snapd-desktop-integration.snapd-desktop-integration.service - Service for snap application snapd-desktop-integration.snapd-desktop-integration.
2025-01-08T21:08:59.284276+08:00 titan kernel: message repeated 8 times: [ [drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership]
2025-01-08T21:08:59.285280+08:00 titan kernel: audit: type=1400 audit(1736341739.283:281): apparmor="DENIED" operation="open" class="file" profile="snap-update-ns.snapd-desktop-integration" name="/proc/37965/maps" pid=37965 comm="5" requested_mask="r" denied_mask="r" fsuid=1001 ouid=0
2025-01-08T21:08:59.285301+08:00 titan kernel: [drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
2025-01-08T21:08:59.288275+08:00 titan kernel: message repeated 2 times: [ [drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership]
2025-01-08T21:08:59.386304+08:00 titan kernel: mihomo-party[37919]: segfault at 8 ip 00007ed795d5914d sp 00007ffe1ce8c9a0 error 4 in libGLX_nvidia.so.565.57.01[7ed795d27000+5c000] likely on CPU 4 (core 8, socket 0)
2025-01-08T21:08:59.386337+08:00 titan kernel: Code: 00 75 3a 48 8b 83 a8 10 00 00 48 85 c0 74 28 48 8b 4d 08 eb 11 0f 1f 84 00 00 00 00 00 48 8b 40 50 48 85 c0 74 11 48 8b 50 08 <48> 39 4a 08 75 ed 48 89 83 b0 10 00 00 5b 5d 41 5c c3 90 48 89 df

GPU驱动版本:

$ nvidia-smi   

Wed Jan  8 22:15:23 2025       
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 565.57.01              Driver Version: 565.57.01      CUDA Version: 12.7     |
|-----------------------------------------+------------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id          Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |           Memory-Usage | GPU-Util  Compute M. |
|                                         |                        |               MIG M. |
|=========================================+========================+======================|
|   0  NVIDIA GeForce RTX 4070 ...    On  |   00000000:01:00.0  On |                  N/A |
|  0%   38C    P8              4W /  285W |     305MiB /  16376MiB |      0%      Default |
|                                         |                        |                  N/A |
+-----------------------------------------+------------------------+----------------------+
                                                                                         
+-----------------------------------------------------------------------------------------+
| Processes:                                                                              |
|  GPU   GI   CI        PID   Type   Process name                              GPU Memory |
|        ID   ID                                                               Usage      |
|=========================================================================================|
|    0   N/A  N/A      2779      C   python3                                       208MiB |
|    0   N/A  N/A     18816      G   /usr/lib/xorg/Xorg                             57MiB |
|    0   N/A  N/A     18910      G   /usr/bin/gnome-shell                           11MiB |
+-----------------------------------------------------------------------------------------+

尝试将mihomo-party退出, rdp中断症状消失.

重现方式

在Ubuntu 24.04.1 LTS下运行此版本mihomo-party, 观察syslog是否有上述错误

[duchenpaul]

有可能是NVidia GPU驱动问题, 但是无法确定