mirror of
https://github.com/comfyanonymous/ComfyUI.git
synced 2025-12-22 20:40:49 +08:00
ce445263e6
Implements Linux Landlock sandboxing to restrict filesystem access when ComfyUI is running. This provides defense-in-depth against malicious custom nodes or workflows that attempt to access sensitive files. How it works: - Uses Linux Landlock LSM (kernel 5.13+) via direct syscalls - Restricts write access to specific directories (output, input, temp, user) - Restricts read access to only what's needed (codebase, models, system libs) - Handles ABI versions 1-5, including IOCTL_DEV for GPU access on v5+ - Exits with error if --enable-landlock is set but Landlock unavailable Write access granted to: - ComfyUI output, input, temp, and user directories - System temp directory (for torch/backends) - SQLite database directory (if configured) - Paths specified via --landlock-allow-writable Read access granted to: - ComfyUI codebase directory - All configured model directories (including extra_model_paths.yaml) - Python installation and site-packages - System libraries (/usr, /lib, /lib64, /opt, /etc, /proc, /sys) - /nix (on NixOS systems) - /dev (with ioctl for GPU access) - Paths specified via --landlock-allow-readable Usage: python main.py --enable-landlock python main.py --enable-landlock --landlock-allow-writable /extra/dir python main.py --enable-landlock --landlock-allow-readable ~/.cache/huggingface Requirements: - Linux with kernel 5.13+ (fails with error on unsupported systems) - Once enabled, restrictions cannot be lifted for the process lifetime - Network access is not restricted (Landlock FS only) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com> |
||
|---|---|---|
| .. | ||
| audio_encoders | ||
| cldm | ||
| comfy_types | ||
| extra_samplers | ||
| image_encoders | ||
| k_diffusion | ||
| ldm | ||
| sd1_tokenizer | ||
| t2i_adapter | ||
| taesd | ||
| text_encoders | ||
| weight_adapter | ||
| checkpoint_pickle.py | ||
| cli_args.py | ||
| clip_config_bigg.json | ||
| clip_model.py | ||
| clip_vision_config_g.json | ||
| clip_vision_config_h.json | ||
| clip_vision_config_vitl_336_llava.json | ||
| clip_vision_config_vitl_336.json | ||
| clip_vision_config_vitl.json | ||
| clip_vision_siglip_384.json | ||
| clip_vision_siglip_512.json | ||
| clip_vision.py | ||
| conds.py | ||
| context_windows.py | ||
| controlnet.py | ||
| diffusers_convert.py | ||
| diffusers_load.py | ||
| float.py | ||
| gligen.py | ||
| hooks.py | ||
| latent_formats.py | ||
| lora_convert.py | ||
| lora.py | ||
| model_base.py | ||
| model_detection.py | ||
| model_management.py | ||
| model_patcher.py | ||
| model_sampling.py | ||
| nested_tensor.py | ||
| ops.py | ||
| options.py | ||
| patcher_extension.py | ||
| pixel_space_convert.py | ||
| quant_ops.py | ||
| rmsnorm.py | ||
| sample.py | ||
| sampler_helpers.py | ||
| samplers.py | ||
| sd1_clip_config.json | ||
| sd1_clip.py | ||
| sd.py | ||
| sdxl_clip.py | ||
| supported_models_base.py | ||
| supported_models.py | ||
| utils.py | ||