ext4: make ext4_init_dot_dotdot for inline dir usage
authorTao Ma <boyu.mt@taobao.com>
Mon, 10 Dec 2012 19:05:57 +0000 (14:05 -0500)
committerTheodore Ts'o <tytso@mit.edu>
Mon, 10 Dec 2012 19:05:57 +0000 (14:05 -0500)
Currently, the initialization of dot and dotdot are encapsulated in
ext4_mkdir and also bond with dir_block. So create a new function
named ext4_init_new_dir and the initialization is moved to
ext4_init_dot_dotdot. Now it will called either in the normal non-inline
case(rec_len of ".." will cover the whole block) or when we converting an
inline dir to a block(rec len of ".." will be the real length). The start
of the next entry is also returned for inline dir usage.

Signed-off-by: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
fs/ext4/ext4.h
fs/ext4/namei.c

index 268636af7f5ca92a55e607c60bd8df04a6144f62..cf840146ce814749537d611fc9c272b83667053f 100644 (file)
@@ -2415,6 +2415,10 @@ extern void ext4_unwritten_wait(struct inode *inode);
 extern const struct inode_operations ext4_dir_inode_operations;
 extern const struct inode_operations ext4_special_inode_operations;
 extern struct dentry *ext4_get_parent(struct dentry *child);
+extern struct ext4_dir_entry_2 *ext4_init_dot_dotdot(struct inode *inode,
+                                struct ext4_dir_entry_2 *de,
+                                int blocksize, int csum_size,
+                                unsigned int parent_ino, int dotdot_real_len);
 
 /* symlink.c */
 extern const struct inode_operations ext4_symlink_inode_operations;
index 88e9a2c7e328e74d3518c2e9b698353f46080f3b..edb9f10c14559cbc667c42ef3b1460d5991a3888 100644 (file)
@@ -2230,21 +2230,87 @@ retry:
        return err;
 }
 
-static int ext4_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
+struct ext4_dir_entry_2 *ext4_init_dot_dotdot(struct inode *inode,
+                         struct ext4_dir_entry_2 *de,
+                         int blocksize, int csum_size,
+                         unsigned int parent_ino, int dotdot_real_len)
+{
+       de->inode = cpu_to_le32(inode->i_ino);
+       de->name_len = 1;
+       de->rec_len = ext4_rec_len_to_disk(EXT4_DIR_REC_LEN(de->name_len),
+                                          blocksize);
+       strcpy(de->name, ".");
+       ext4_set_de_type(inode->i_sb, de, S_IFDIR);
+
+       de = ext4_next_entry(de, blocksize);
+       de->inode = cpu_to_le32(parent_ino);
+       de->name_len = 2;
+       if (!dotdot_real_len)
+               de->rec_len = ext4_rec_len_to_disk(blocksize -
+                                       (csum_size + EXT4_DIR_REC_LEN(1)),
+                                       blocksize);
+       else
+               de->rec_len = ext4_rec_len_to_disk(
+                               EXT4_DIR_REC_LEN(de->name_len), blocksize);
+       strcpy(de->name, "..");
+       ext4_set_de_type(inode->i_sb, de, S_IFDIR);
+
+       return ext4_next_entry(de, blocksize);
+}
+
+static int ext4_init_new_dir(handle_t *handle, struct inode *dir,
+                            struct inode *inode)
 {
-       handle_t *handle;
-       struct inode *inode;
        struct buffer_head *dir_block = NULL;
        struct ext4_dir_entry_2 *de;
        struct ext4_dir_entry_tail *t;
        unsigned int blocksize = dir->i_sb->s_blocksize;
        int csum_size = 0;
-       int err, retries = 0;
+       int err;
 
        if (EXT4_HAS_RO_COMPAT_FEATURE(dir->i_sb,
                                       EXT4_FEATURE_RO_COMPAT_METADATA_CSUM))
                csum_size = sizeof(struct ext4_dir_entry_tail);
 
+       inode->i_size = EXT4_I(inode)->i_disksize = blocksize;
+       dir_block = ext4_bread(handle, inode, 0, 1, &err);
+       if (!(dir_block = ext4_bread(handle, inode, 0, 1, &err))) {
+               if (!err) {
+                       err = -EIO;
+                       ext4_error(inode->i_sb,
+                                  "Directory hole detected on inode %lu\n",
+                                  inode->i_ino);
+               }
+               goto out;
+       }
+       BUFFER_TRACE(dir_block, "get_write_access");
+       err = ext4_journal_get_write_access(handle, dir_block);
+       if (err)
+               goto out;
+       de = (struct ext4_dir_entry_2 *)dir_block->b_data;
+       ext4_init_dot_dotdot(inode, de, blocksize, csum_size, dir->i_ino, 0);
+       set_nlink(inode, 2);
+       if (csum_size) {
+               t = EXT4_DIRENT_TAIL(dir_block->b_data, blocksize);
+               initialize_dirent_tail(t, blocksize);
+       }
+
+       BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
+       err = ext4_handle_dirty_dirent_node(handle, inode, dir_block);
+       if (err)
+               goto out;
+       set_buffer_verified(dir_block);
+out:
+       brelse(dir_block);
+       return err;
+}
+
+static int ext4_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode)
+{
+       handle_t *handle;
+       struct inode *inode;
+       int err, retries = 0;
+
        if (EXT4_DIR_LINK_MAX(dir))
                return -EMLINK;
 
@@ -2268,47 +2334,9 @@ retry:
 
        inode->i_op = &ext4_dir_inode_operations;
        inode->i_fop = &ext4_dir_operations;
-       inode->i_size = EXT4_I(inode)->i_disksize = inode->i_sb->s_blocksize;
-       if (!(dir_block = ext4_bread(handle, inode, 0, 1, &err))) {
-               if (!err) {
-                       err = -EIO;
-                       ext4_error(inode->i_sb,
-                                  "Directory hole detected on inode %lu\n",
-                                  inode->i_ino);
-               }
-               goto out_clear_inode;
-       }
-       BUFFER_TRACE(dir_block, "get_write_access");
-       err = ext4_journal_get_write_access(handle, dir_block);
+       err = ext4_init_new_dir(handle, dir, inode);
        if (err)
                goto out_clear_inode;
-       de = (struct ext4_dir_entry_2 *) dir_block->b_data;
-       de->inode = cpu_to_le32(inode->i_ino);
-       de->name_len = 1;
-       de->rec_len = ext4_rec_len_to_disk(EXT4_DIR_REC_LEN(de->name_len),
-                                          blocksize);
-       strcpy(de->name, ".");
-       ext4_set_de_type(dir->i_sb, de, S_IFDIR);
-       de = ext4_next_entry(de, blocksize);
-       de->inode = cpu_to_le32(dir->i_ino);
-       de->rec_len = ext4_rec_len_to_disk(blocksize -
-                                          (csum_size + EXT4_DIR_REC_LEN(1)),
-                                          blocksize);
-       de->name_len = 2;
-       strcpy(de->name, "..");
-       ext4_set_de_type(dir->i_sb, de, S_IFDIR);
-       set_nlink(inode, 2);
-
-       if (csum_size) {
-               t = EXT4_DIRENT_TAIL(dir_block->b_data, blocksize);
-               initialize_dirent_tail(t, blocksize);
-       }
-
-       BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
-       err = ext4_handle_dirty_dirent_node(handle, inode, dir_block);
-       if (err)
-               goto out_clear_inode;
-       set_buffer_verified(dir_block);
        err = ext4_mark_inode_dirty(handle, inode);
        if (!err)
                err = ext4_add_entry(handle, dentry, inode);
@@ -2328,7 +2356,6 @@ out_clear_inode:
        unlock_new_inode(inode);
        d_instantiate(dentry, inode);
 out_stop:
-       brelse(dir_block);
        ext4_journal_stop(handle);
        if (err == -ENOSPC && ext4_should_retry_alloc(dir->i_sb, &retries))
                goto retry;