Squashfs: reverse order of filesystem table reading
Phillip Lougher [Tue, 24 May 2011 01:57:05 +0000 (02:57 +0100)]
Reverse order of table reading from mostly first to last in placement
order, to last to first.  This is to enable extra superblock sanity
checks to be added in later patches.

Signed-off-by: Phillip Lougher <phillip@lougher.demon.co.uk>

fs/squashfs/super.c

index d16c392..401cc8c 100644 (file)
@@ -214,6 +214,24 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
                goto failed_mount;
        }
 
+       /* Handle xattrs */
+       sb->s_xattr = squashfs_xattr_handlers;
+       xattr_id_table_start = le64_to_cpu(sblk->xattr_id_table_start);
+       if (xattr_id_table_start == SQUASHFS_INVALID_BLK)
+               goto allocate_id_index_table;
+
+       /* Allocate and read xattr id lookup table */
+       msblk->xattr_id_table = squashfs_read_xattr_id_table(sb,
+               xattr_id_table_start, &msblk->xattr_table, &msblk->xattr_ids);
+       if (IS_ERR(msblk->xattr_id_table)) {
+               ERROR("unable to read xattr id index table\n");
+               err = PTR_ERR(msblk->xattr_id_table);
+               msblk->xattr_id_table = NULL;
+               if (err != -ENOTSUPP)
+                       goto failed_mount;
+       }
+
+allocate_id_index_table:
        /* Allocate and read id index table */
        msblk->id_table = squashfs_read_id_index_table(sb,
                le64_to_cpu(sblk->id_table_start), le16_to_cpu(sblk->no_ids));
@@ -224,9 +242,27 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
                goto failed_mount;
        }
 
+       /* Handle inode lookup table */
+       lookup_table_start = le64_to_cpu(sblk->lookup_table_start);
+       if (lookup_table_start == SQUASHFS_INVALID_BLK)
+               goto handle_fragments;
+
+       /* Allocate and read inode lookup table */
+       msblk->inode_lookup_table = squashfs_read_inode_lookup_table(sb,
+               lookup_table_start, msblk->inodes);
+       if (IS_ERR(msblk->inode_lookup_table)) {
+               ERROR("unable to read inode lookup table\n");
+               err = PTR_ERR(msblk->inode_lookup_table);
+               msblk->inode_lookup_table = NULL;
+               goto failed_mount;
+       }
+
+       sb->s_export_op = &squashfs_export_ops;
+
+handle_fragments:
        fragments = le32_to_cpu(sblk->fragments);
        if (fragments == 0)
-               goto allocate_lookup_table;
+               goto allocate_root;
 
        msblk->fragment_cache = squashfs_cache_init("fragment",
                SQUASHFS_CACHED_FRAGMENTS, msblk->block_size);
@@ -245,39 +281,6 @@ static int squashfs_fill_super(struct super_block *sb, void *data, int silent)
                goto failed_mount;
        }
 
-allocate_lookup_table:
-       lookup_table_start = le64_to_cpu(sblk->lookup_table_start);
-       if (lookup_table_start == SQUASHFS_INVALID_BLK)
-               goto allocate_xattr_table;
-
-       /* Allocate and read inode lookup table */
-       msblk->inode_lookup_table = squashfs_read_inode_lookup_table(sb,
-               lookup_table_start, msblk->inodes);
-       if (IS_ERR(msblk->inode_lookup_table)) {
-               ERROR("unable to read inode lookup table\n");
-               err = PTR_ERR(msblk->inode_lookup_table);
-               msblk->inode_lookup_table = NULL;
-               goto failed_mount;
-       }
-
-       sb->s_export_op = &squashfs_export_ops;
-
-allocate_xattr_table:
-       sb->s_xattr = squashfs_xattr_handlers;
-       xattr_id_table_start = le64_to_cpu(sblk->xattr_id_table_start);
-       if (xattr_id_table_start == SQUASHFS_INVALID_BLK)
-               goto allocate_root;
-
-       /* Allocate and read xattr id lookup table */
-       msblk->xattr_id_table = squashfs_read_xattr_id_table(sb,
-               xattr_id_table_start, &msblk->xattr_table, &msblk->xattr_ids);
-       if (IS_ERR(msblk->xattr_id_table)) {
-               ERROR("unable to read xattr id index table\n");
-               err = PTR_ERR(msblk->xattr_id_table);
-               msblk->xattr_id_table = NULL;
-               if (err != -ENOTSUPP)
-                       goto failed_mount;
-       }
 allocate_root:
        root = new_inode(sb);
        if (!root) {