备注

AI Translation Notice

This document was automatically translated by hunyuan-turbos-latest model, for reference only.

Source document: kernel/filesystem/vfs/mountable_fs.md
Translation time: 2025-06-29 09:58:15
Translation model: hunyuan-turbos-latest

Please report issues via Community Channel

备注

Author: Zhuang Kaiye

Design

    graph TD
    subgraph 用户层 / 系统调用
        A[sys_mount] --> B[produce_fs!]
    end

    subgraph 文件系统注册与创建
        B --> C{查找 FSMAKER}
        C -- 找到匹配 --> D[FileSystemMaker::builder && FileSystemMaker::maker]
        D --> G[FileSystem 实例]
        G --> H[挂载成功]
        C -- 未找到 --> I[错误: EINVAL]
    end

    subgraph 文件系统实现者
        J[RamFS / 其他文件系统] --> K[实现 MountableFileSystem trait]
        K --> L[make_mount_data]
        K --> M[make_fs]
        L --> N[register_mountable_fs!宏]
        M --> N
    end

    N --> O[分布式切片 FSMAKER ]
    O --> C

    click J "#" "RamFS - 文件系统示例"
    click B "#" "produce_fs 函数"
    click O "#" "FSMAKER - 文件系统工厂数组"

Process Description:

A specific file system (e.g., RamFS) registers its creation logic with FSMAKER by implementing MountableFileSystem trait and using the register_mountable_fs! macro.
Users request to mount a file system via the sys_mount system call.
sys_mount calls the produce_fs function, passing the file system type, raw mount data, and source path.
produce_fs traverses the global FSMAKER array to find a FileSystemMaker matching the requested file system type name.
If found, it first calls maker.builder (which internally invokes the specific file system’s make_mount_data method) to process the raw data and generate an optional mount_data object.
Next, it calls maker.build (which internally invokes the specific file system’s make_fs method), passing the mount_data generated in the previous step, thereby creating the file system instance.
The successfully created file system instance (Arc<dyn FileSystem>) is returned and used for subsequent mount operations.
If no matching file system type is found, an error is returned.

Additional Notes

Currently, DragonOS supports mounting the following file systems: ramfs, ext4, and vfat. When mounting disk files in DragonOS, note the following:

Since the system cannot directly determine the file system type of a disk, the target partition’s file system type must be explicitly known before mounting.
The mount operation requires specifying the corresponding disk device name (located under /dev).
These disk device files are generated by modifying the tools/run-qemu.sh startup script to load the prepared disk image files into the system. Examples of virtio disk device naming include vda1 and vdb1. The device names of disks in DragonOS are automatically assigned based on the order in which images are loaded in run-qemu.sh (a, b, c, etc.), with numbers indicating partition numbers.

Therefore, to mount a disk currently, you can modify the test-mount-ext4 executable to mount the specified disk file with the corresponding file system format. Below is an example of the mount operation:

use core::ffi::{c_char, c_void};
use libc::{mount, MS_BIND};
use std::fs;
use std::path::Path;

fn main() {
    let ext4_path = Path::new("mnt/ext4");
    let dir = fs::create_dir_all(ext4_path);
    if dir.is_err() {
        panic!("mkdir /mnt/ext4 fail.");
    }

    // 硬盘名称，由传入顺序决定
    let source = b"/dev/vdb1\0".as_ptr() as *const c_char;
    let target = b"/mnt/ext4\0".as_ptr() as *const c_char;
    // 文件系统类型
    let fstype = b"ext4\0".as_ptr() as *const c_char;
    let flags = MS_BIND;
    let data = std::ptr::null() as *const c_void;
    let _ = unsafe { mount(source, target, fstype, flags, data) };

    println!("Mount successfully!");
}

As for disk image creation, you can run sudo bash tools/make_fs_image.sh to simultaneously create disk images for ext4 and fat. After entering the system, two additional disk files (vdb1 and vdc1) will appear in the dev directory. You can then execute the test-mount-ext4 and test-mount-fat test programs to verify the setup.