EROFS 源码阅读

发表于 更新于 分类于 阅读次数： 本文字数： 566 阅读时长 ≈ 2 分钟

#erofs in-tree 驱动

in-tree 的代码也是模块的形式，以 module_init 开始

1
2
3
4
5
6
7
8
9
10
11
12
13
14
module_init(erofs_module_init);
module_exit(erofs_module_exit);

// in erofs_module_init()
err = register_filesystem(&erofs_fs_type);

static struct file_system_type erofs_fs_type = {
    .owner          = THIS_MODULE,
    .name           = "erofs",
    .init_fs_context = erofs_init_fs_context, // init ->ops and ->fs_private
    .kill_sb        = erofs_kill_sb, // shutdown a fs
    .fs_flags       = FS_REQUIRES_DEV | FS_ALLOW_IDMAP,
};
MODULE_ALIAS_FS("erofs");

Kconfig 中包含的一些选项：

  • EROFS_FS_DEBUG DEBUG 支持
  • EROFS_FS_XATTR 拓展属性支持
  • EROFS_FS_POSIX_ACL ACL 支持
  • EROFS_FS_SECURITY 安全标签支持
  • EROFS_FS_ZIP 压缩支持，默认只支持读取 LZ4 压缩的数据
    • EROFS_FS_ZIP_LZMA 支持读取 microLZMA 压缩的数据
    • EROFS_FS_ZIP_DEFLATE 支持读取 DEFLATE 压缩的数据
    • EROFS_FS_PCPU_KTHREAD 用 per-cpu 的线程池去作异步数据解压缩
      • EROFS_FS_PCPU_KTHREAD_HIPRI 把这些线程设为高优先级
  • EROFS_FS_ONDEMAND 利用 fscache 实现 on-demand 读

#初始化流程

1
2
3
4
5
6
static const struct fs_context_operations erofs_context_ops = {
    .parse_param	= erofs_fc_parse_param, // add a fs param
    .get_tree       = erofs_fc_get_tree, // ~mount a fs
    .reconfigure    = erofs_fc_reconfigure, // sb reconfiguration
    .free		= erofs_fc_free, // clean up
};

#内核的 fs_context 结构

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
struct fs_context {
    const struct fs_context_operations *ops;
    struct file_system_type *fs_type;
    void			*fs_private;
    struct dentry		*root;  // the root of fs tree
    struct user_namespace	*user_ns;
    struct net		*net_ns;
    const struct cred	*cred;
    char			*source;  // /dev/sda1 or host:/path, etc.
    char			*subtype;
    void			*security;
    void			*s_fs_info;  // user data -> erofs_sb_info
    unsigned int		sb_flags;
    unsigned int		sb_flags_mask;
    unsigned int		s_iflags;
    enum fs_context_purpose	purpose:8;
    ...
};

#EROFS 的元数据 erofs_sb_info

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
struct erofs_sb_info {
    struct erofs_mount_opts opt;	/* options */

#ifdef CONFIG_EROFS_FS_ZIP
    /* list for all registered superblocks, mainly for shrinker */
    struct list_head list;
    struct mutex umount_mutex;

    /* managed XArray arranged in physical block number */
    struct xarray managed_pslots;

    unsigned int shrinker_run_no;
    u16 available_compr_algs;

    /* pseudo inode to manage cached pages */
    struct inode *managed_cache;

    struct erofs_sb_lz4_info lz4;
#endif    /* CONFIG_EROFS_FS_ZIP */
    struct inode *packed_inode;
    struct erofs_dev_context *devs;
    struct dax_device *dax_dev;
    u64 dax_part_off;
    u64 total_blocks;
    u32 primarydevice_blocks;

    u32 meta_blkaddr;
#ifdef CONFIG_EROFS_FS_XATTR
    u32 xattr_blkaddr;
    u32 xattr_prefix_start;
    u8 xattr_prefix_count;
    struct erofs_xattr_prefix_item *xattr_prefixes;
    unsigned int xattr_filter_reserved;
#endif
    u16 device_id_mask;	/* valid bits of device id to be used */

    unsigned char islotbits;	/* inode slot unit size in bit shift */
    unsigned char blkszbits;	/* filesystem block size in bit shift */

    u32 sb_size;			/* total superblock size */
    u32 build_time_nsec;
    u64 build_time;

    /* what we really care is nid, rather than ino.. */
    erofs_nid_t root_nid;
    erofs_nid_t packed_nid;
    /* used for statfs, f_files - f_favail */
    u64 inos;

    u8 uuid[16];                    /* 128-bit uuid for volume */
    u8 volume_name[16];             /* volume name */
    u32 feature_compat;
    u32 feature_incompat;

    /* sysfs support */
    struct kobject s_kobj;		/* /sys/fs/erofs/<devname> */
    struct completion s_kobj_unregister;

    /* fscache support */
    struct fscache_volume *volume;
    struct erofs_fscache *s_fscache;
    struct erofs_domain *domain;
    char *fsid;
    char *domain_id;
};

bdev.c 中负责硬盘块操作

申请 super_block

erofs_fc_fill_super 其实是从硬盘上读取数据，初始化 user 的 sb 信息

#inode 操作

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
const struct inode_operations erofs_generic_iops = {
    .getattr = erofs_getattr,
    .listxattr = erofs_listxattr,
    .get_inode_acl = erofs_get_acl,
    .fiemap = erofs_fiemap,
};

const struct inode_operations erofs_symlink_iops = {
    .get_link = page_get_link,
    .getattr = erofs_getattr,
    .listxattr = erofs_listxattr,
    .get_inode_acl = erofs_get_acl,
};

const struct inode_operations erofs_fast_symlink_iops = {
    .get_link = simple_get_link,
    .getattr = erofs_getattr,
    .listxattr = erofs_listxattr,
    .get_inode_acl = erofs_get_acl,
};

#erofs-utils

  • mkfs.erofs 创建一个 erofs 文件系统；
  • fsck.erofs 检查 erofs 镜像的完整性；
  • dump.erofs 导出 erofs 镜像中的相关细节；
  • erofsfuse 用户态

#mkfs

#fsck

#dump

#erofsfuse