Rev.7を表示中。最新版はこちら。

vfsmntアドレス＋ディレクトリdentryアドレスをハッシュインデックスとするmount_hashtable[]にマウントデバイスをtailリストし、マウントディレクトリdentry->d_flagにCACHE_MOUNTEDが設定され、パス検索でdentryのd_flags & DCACHE_MOUNTEDなら、mount_hashtable[]から、係るdenryのマウントデバイスを取得します。

実装概略サンプル-mount dev dir

static struct list_head mount_hashtable[10];

void	mount(char *dev, struct path *dir)
{
    struct vfsmount   *mnt;
    struct list_head  *head;

    mnt = do_kern_mount(dev);
    mnt->mnt_mountpoint = dir->dentry;
    mnt->mnt_parent     = dir->mnt;

    head = mount_hashtable[(dir->mnt + dir->dentry) % 10];

    mnt->mnt_hash->prev = head->prev;
    mnt->mnt_hash->next = head;
    head->prev->next = mnt->mnt_hash;
    head->prev = mnt->mnt_hash;
    dir.dentry->d_flags |= DCACHE_MOUNTED;
}

struct vfsmount *lookup_mnt(struct path *dir)
{
    struct list_head *head
    struct mount *p, *found=NULL;

    if (dir->dentry->d_flags & DCACHE_MOUNTED) {
        head = mount_hashtable[(dir->mnt + dir->dentry) % 10];
        while () {
            p = list_entry(head, struct mount, mnt_hash);
            if (p->mnt_parent->mnt == dir->mnt && p->mnt_mountpoint == dir->dentry) {
                found = p;
                break;
               }
          }
      }
    return found;		
}

mount_hashtable[]はページキャッシュを領域とする動的メモリで、要素はlist_head->next/prevのアドレスで、サイズはシステムに依存され、従ってmount_hashtable[]要素数は動的で、かかる実装は2進数によるmount_hashtable[]の要素数等の動的配列インデックスの実装となり、mount_hashtable[]の要素インデックスは2進数によるサイズ長のmnt+dentryアドレスのインデックス長に係る下位／上位を加算した値で実装されます。

HASH_SHIFTはmount_hashtable[]の要素数の2進数の対数で、HASH_SIZEは実要素数となります。

#define HASH_SHIFT ilog2(PAGE_SIZE / sizeof(struct list_head))
#define HASH_SIZE (1UL << HASH_SHIFT)

struct mount {
       struct list_head mnt_hash;
       struct mount *mnt_parent;
       struct dentry *mnt_mountpoint;
       struct vfsmount mnt;
#ifdef CONFIG_SMP
       struct mnt_pcp __percpu *mnt_pcp;
#else
       int mnt_count;
       int mnt_writers;
#endif
       struct list_head mnt_mounts;    /* list of children, anchored here */
       struct list_head mnt_child;     /* and going through their mnt_child */
       struct list_head mnt_instance;  /* mount instance on sb->s_mounts */
       const char *mnt_devname;        /* Name of device e.g. /dev/dsk/hda1 */
       struct list_head mnt_list;
       struct list_head mnt_expire;    /* link in fs-specific expiry list */
       struct list_head mnt_share;     /* circular list of shared mounts */
       struct list_head mnt_slave_list;/* list of slave mounts */
       struct list_head mnt_slave;     /* slave list entry */
       struct mount *mnt_master;       /* slave is on master->mnt_slave_list */
       struct mnt_namespace *mnt_ns;   /* containing namespace */
#ifdef CONFIG_FSNOTIFY
       struct hlist_head mnt_fsnotify_marks;
       __u32 mnt_fsnotify_mask;
#endif
       int mnt_id;                     /* mount identifier */
       int mnt_group_id;               /* peer group identifier */
       int mnt_expiry_mark;            /* true if marked for expiry */
       int mnt_pinned;
       int mnt_ghosts;
};
      
void __init mnt_init(void)
{
       unsigned u;
       int err;

       init_rwsem(&namespace_sem);

       mnt_cache = kmem_cache_create("mnt_cache", sizeof(struct mount),
                       0, SLAB_HWCACHE_ALIGN | SLAB_PANIC, NULL);

       mount_hashtable = (struct list_head *)__get_free_page(GFP_ATOMIC);

       if (!mount_hashtable)
               panic("Failed to allocate mount hash table\n");

       printk(KERN_INFO "Mount-cache hash table entries: %lu\n", HASH_SIZE);

       for (u = 0; u < HASH_SIZE; u++)
               INIT_LIST_HEAD(&mount_hashtable[u]);

       br_lock_init(&vfsmount_lock);

       err = sysfs_init();
       if (err)
               printk(KERN_WARNING "%s: sysfs_init error: %d\n",
                       __func__, err);
       fs_kobj = kobject_create_and_add("fs", NULL);
       if (!fs_kobj)
               printk(KERN_WARNING "%s: kobj create error\n", __func__);
       init_rootfs();
       init_mount_tree();
}

static int graft_tree(struct mount *mnt, struct path *path)
{
       if (mnt->mnt.mnt_sb->s_flags & MS_NOUSER)
               return -EINVAL;

       if (S_ISDIR(path->dentry->d_inode->i_mode) !=
             S_ISDIR(mnt->mnt.mnt_root->d_inode->i_mode))
               return -ENOTDIR;

       if (d_unlinked(path->dentry))
               return -ENOENT;

       return attach_recursive_mnt(mnt, path, NULL);
}

static int attach_recursive_mnt(struct mount *source_mnt,
                       struct path *path, struct path *parent_path)
{
       LIST_HEAD(tree_list);
       struct mount *dest_mnt = real_mount(path->mnt);
       struct dentry *dest_dentry = path->dentry;
       struct mount *child, *p;
       int err;

       if (IS_MNT_SHARED(dest_mnt)) {
               err = invent_group_ids(source_mnt, true);
               if (err)
                       goto out;
       }
       err = propagate_mnt(dest_mnt, dest_dentry, source_mnt, &tree_list);
       if (err)
               goto out_cleanup_ids;

       br_write_lock(&vfsmount_lock);

       if (IS_MNT_SHARED(dest_mnt)) {
               for (p = source_mnt; p; p = next_mnt(p, source_mnt))
                       set_mnt_shared(p);
       }
       if (parent_path) {
               detach_mnt(source_mnt, parent_path);
               attach_mnt(source_mnt, path);
               touch_mnt_namespace(source_mnt->mnt_ns);
       } else {
               mnt_set_mountpoint(dest_mnt, dest_dentry, source_mnt);
               commit_tree(source_mnt);
       }

       list_for_each_entry_safe(child, p, &tree_list, mnt_hash) {
               list_del_init(&child->mnt_hash);
               commit_tree(child);
       }
       br_write_unlock(&vfsmount_lock);

       return 0;

out_cleanup_ids:
       if (IS_MNT_SHARED(dest_mnt))
               cleanup_group_ids(source_mnt, NULL);
out:
       return err;
}

static void attach_mnt(struct mount *mnt, struct path *path)
{
       mnt_set_mountpoint(real_mount(path->mnt), path->dentry, mnt);
       list_add_tail(&mnt->mnt_hash, mount_hashtable +
                       hash(path->mnt, path->dentry));
       list_add_tail(&mnt->mnt_child, &real_mount(path->mnt)->mnt_mounts);
}

void mnt_set_mountpoint(struct mount *mnt, struct dentry *dentry,
                       struct mount *child_mnt)
{
       mnt_add_count(mnt, 1);  /* essentially, that's mntget */
       child_mnt->mnt_mountpoint = dget(dentry);
       child_mnt->mnt_parent = mnt;
       spin_lock(&dentry->d_lock);
       dentry->d_flags |= DCACHE_MOUNTED;
       spin_unlock(&dentry->d_lock);
}

static void commit_tree(struct mount *mnt)
{
       struct mount *parent = mnt->mnt_parent;
       struct mount *m;
       LIST_HEAD(head);
       struct mnt_namespace *n = parent->mnt_ns;

       BUG_ON(parent == mnt);

       list_add_tail(&head, &mnt->mnt_list);
       list_for_each_entry(m, &head, mnt_list)
               m->mnt_ns = n;

       list_splice(&head, n->list.prev);

       list_add_tail(&mnt->mnt_hash, mount_hashtable +
                               hash(&parent->mnt, mnt->mnt_mountpoint));
       list_add_tail(&mnt->mnt_child, &parent->mnt_mounts);
       touch_mnt_namespace(n);
}

static inline void list_add_tail(struct list_head *new, struct list_head *head)
{
       __list_add(new, head->prev, head);
}

static inline void __list_add(struct list_head *new,
                             struct list_head *prev,
                             struct list_head *next)
{
       next->prev = new;
       new->next = next;
       new->prev = prev;
       prev->next = new;
}

static inline unsigned long hash(struct vfsmount *mnt, struct dentry *dentry)
{
       unsigned long tmp = ((unsigned long)mnt / L1_CACHE_BYTES);
       tmp += ((unsigned long)dentry / L1_CACHE_BYTES);
       tmp = tmp + (tmp >> hash_SHIFT);
       return tmp & (HASH_SIZE - 1);
}

備考
マウントデバイス数の制約はありません。メモリ取得可能範囲で無制限にmount可能です。

mnt->mnt_childのリストは、unmount/share等の親mnt下の子mntにも処理が必要となるケース故での実装です。

アドレスに係る加算は、係る領域とする配列のインデックスに相当します。
list_add_tail(&mnt->mnt_hash, mount_hashtable + hash(path->mnt, path->dentry));
mount_hashtable + hash(path->mnt, path->dentry) ===> mount_hashtable[hash(path->mnt, path->dentry)]

#include <stdio.h>

struct babakaka {
    char   tmp[100];
    int    val;
};

void	main()
{
    int    i;
    struct    babakaka a[10], *b, *c;

    for (i = 0; i < 10; i++) {
        a[i].val = i;
    }

    b = a;
    for (i = 0; i < 10; i++) {
        c = b + i;
        printf("%d:", c->val);
    }
   printf("\n");
} 

[root@localhost mnt7]# ./a.out 
0:1:2:3:4:5:6:7:8:9: