ext4: Fix the race between read_inode_bitmap() and ext4_new_inode()

We need to make sure we update the inode bitmap and clear
EXT4_BG_INODE_UNINIT flag with sb_bgl_lock held, since
ext4_read_inode_bitmap() looks at EXT4_BG_INODE_UNINIT to decide
whether to initialize the inode bitmap each time it is called.
(introduced by commit c806e68f.)

ext4_read_inode_bitmap does:

spin_lock(sb_bgl_lock(EXT4_SB(sb), block_group));
if (desc->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
	ext4_init_inode_bitmap(sb, bh, block_group, desc);

and ext4_new_inode does
if (!ext4_set_bit_atomic(sb_bgl_lock(sbi, group),
                   ino, inode_bitmap_bh->b_data))
		   ......
		   ...
spin_lock(sb_bgl_lock(sbi, group));

gdp->bg_flags &= cpu_to_le16(~EXT4_BG_INODE_UNINIT);
i.e., on allocation we update the bitmap then we take the sb_bgl_lock
and clear the EXT4_BG_INODE_UNINIT flag. What can happen is a
parallel ext4_read_inode_bitmap can zero out the bitmap in between
the above ext4_set_bit_atomic and spin_lock(sb_bg_lock..)

The race results in below user visible errors
EXT4-fs error (device sdb1): ext4_free_inode: bit already cleared for inode 168449
EXT4-fs warning (device sdb1): ext4_unlink: Deleting nonexistent file ...
EXT4-fs warning (device sdb1): ext4_rmdir: empty directory has too many links ...
# ls -al /mnt/tmp/f/p369/d3/d6/d39/db2/dee/d10f/d3f/l71
ls: /mnt/tmp/f/p369/d3/d6/d39/db2/dee/d10f/d3f/l71: Stale NFS file handle

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Cc: stable@kernel.org

fs/ext4/ialloc.c

@@ -572,6 +572,79 @@ static int find_group_other(struct super_block *sb, struct inode *parent,
 	return -1;
 }
 
+/*
+ * claim the inode from the inode bitmap. If the group
+ * is uninit we need to take the groups's sb_bgl_lock
+ * and clear the uninit flag. The inode bitmap update
+ * and group desc uninit flag clear should be done
+ * after holding sb_bgl_lock so that ext4_read_inode_bitmap
+ * doesn't race with the ext4_claim_inode
+ */
+static int ext4_claim_inode(struct super_block *sb,
+			struct buffer_head *inode_bitmap_bh,
+			unsigned long ino, ext4_group_t group, int mode)
+{
+	int free = 0, retval = 0, count;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	struct ext4_group_desc *gdp = ext4_get_group_desc(sb, group, NULL);
+
+	spin_lock(sb_bgl_lock(sbi, group));
+	if (ext4_set_bit(ino, inode_bitmap_bh->b_data)) {
+		/* not a free inode */
+		retval = 1;
+		goto err_ret;
+	}
+	ino++;
+	if ((group == 0 && ino < EXT4_FIRST_INO(sb)) ||
+			ino > EXT4_INODES_PER_GROUP(sb)) {
+		spin_unlock(sb_bgl_lock(sbi, group));
+		ext4_error(sb, __func__,
+			   "reserved inode or inode > inodes count - "
+			   "block_group = %u, inode=%lu", group,
+			   ino + group * EXT4_INODES_PER_GROUP(sb));
+		return 1;
+	}
+	/* If we didn't allocate from within the initialized part of the inode
+	 * table then we need to initialize up to this inode. */
+	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) {
+
+		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
+			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_INODE_UNINIT);
+			/* When marking the block group with
+			 * ~EXT4_BG_INODE_UNINIT we don't want to depend
+			 * on the value of bg_itable_unused even though
+			 * mke2fs could have initialized the same for us.
+			 * Instead we calculated the value below
+			 */
+
+			free = 0;
+		} else {
+			free = EXT4_INODES_PER_GROUP(sb) -
+				ext4_itable_unused_count(sb, gdp);
+		}
+
+		/*
+		 * Check the relative inode number against the last used
+		 * relative inode number in this group. if it is greater
+		 * we need to  update the bg_itable_unused count
+		 *
+		 */
+		if (ino > free)
+			ext4_itable_unused_set(sb, gdp,
+					(EXT4_INODES_PER_GROUP(sb) - ino));
+	}
+	count = ext4_free_inodes_count(sb, gdp) - 1;
+	ext4_free_inodes_set(sb, gdp, count);
+	if (S_ISDIR(mode)) {
+		count = ext4_used_dirs_count(sb, gdp) + 1;
+		ext4_used_dirs_set(sb, gdp, count);
+	}
+	gdp->bg_checksum = ext4_group_desc_csum(sbi, group, gdp);
+err_ret:
+	spin_unlock(sb_bgl_lock(sbi, group));
+	return retval;
+}
+
 /*
  * There are two policies for allocating an inode.  If the new inode is
  * a directory, then a forward search is made for a block group with both
@@ -594,7 +667,7 @@ struct inode *ext4_new_inode(handle_t *handle, struct inode *dir, int mode)
 	struct ext4_super_block *es;
 	struct ext4_inode_info *ei;
 	struct ext4_sb_info *sbi;
-	int ret2, err = 0, count;
+	int ret2, err = 0;
 	struct inode *ret;
 	ext4_group_t i;
 	int free = 0;
@@ -658,8 +731,13 @@ repeat_in_this_group:
 			if (err)
 				goto fail;
 
-			if (!ext4_set_bit_atomic(sb_bgl_lock(sbi, group),
-						ino, inode_bitmap_bh->b_data)) {
+			BUFFER_TRACE(group_desc_bh, "get_write_access");
+			err = ext4_journal_get_write_access(handle,
+								group_desc_bh);
+			if (err)
+				goto fail;
+			if (!ext4_claim_inode(sb, inode_bitmap_bh,
+						ino, group, mode)) {
 				/* we won it */
 				BUFFER_TRACE(inode_bitmap_bh,
 					"call ext4_handle_dirty_metadata");
@@ -668,10 +746,13 @@ repeat_in_this_group:
 							inode_bitmap_bh);
 				if (err)
 					goto fail;
+				/* zero bit is inode number 1*/
+				ino++;
 				goto got;
 			}
 			/* we lost it */
 			ext4_handle_release_buffer(handle, inode_bitmap_bh);
+			ext4_handle_release_buffer(handle, group_desc_bh);
 
 			if (++ino < EXT4_INODES_PER_GROUP(sb))
 				goto repeat_in_this_group;
@@ -691,22 +772,6 @@ repeat_in_this_group:
 	goto out;
 
 got:
-	ino++;
-	if ((group == 0 && ino < EXT4_FIRST_INO(sb)) ||
-	    ino > EXT4_INODES_PER_GROUP(sb)) {
-		ext4_error(sb, __func__,
-			   "reserved inode or inode > inodes count - "
-			   "block_group = %u, inode=%lu", group,
-			   ino + group * EXT4_INODES_PER_GROUP(sb));
-		err = -EIO;
-		goto fail;
-	}
-
-	BUFFER_TRACE(group_desc_bh, "get_write_access");
-	err = ext4_journal_get_write_access(handle, group_desc_bh);
-	if (err)
-		goto fail;
-
 	/* We may have to initialize the block bitmap if it isn't already */
 	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM) &&
 	    gdp->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) {
@@ -743,49 +808,10 @@ got:
 		if (err)
 			goto fail;
 	}
-
-	spin_lock(sb_bgl_lock(sbi, group));
-	/* If we didn't allocate from within the initialized part of the inode
-	 * table then we need to initialize up to this inode. */
-	if (EXT4_HAS_RO_COMPAT_FEATURE(sb, EXT4_FEATURE_RO_COMPAT_GDT_CSUM)) {
-		if (gdp->bg_flags & cpu_to_le16(EXT4_BG_INODE_UNINIT)) {
-			gdp->bg_flags &= cpu_to_le16(~EXT4_BG_INODE_UNINIT);
-
-			/* When marking the block group with
-			 * ~EXT4_BG_INODE_UNINIT we don't want to depend
-			 * on the value of bg_itable_unused even though
-			 * mke2fs could have initialized the same for us.
-			 * Instead we calculated the value below
-			 */
-
-			free = 0;
-		} else {
-			free = EXT4_INODES_PER_GROUP(sb) -
-				ext4_itable_unused_count(sb, gdp);
-		}
-
-		/*
-		 * Check the relative inode number against the last used
-		 * relative inode number in this group. if it is greater
-		 * we need to  update the bg_itable_unused count
-		 *
-		 */
-		if (ino > free)
-			ext4_itable_unused_set(sb, gdp,
-					(EXT4_INODES_PER_GROUP(sb) - ino));
-	}
-
-	count = ext4_free_inodes_count(sb, gdp) - 1;
-	ext4_free_inodes_set(sb, gdp, count);
-	if (S_ISDIR(mode)) {
-		count = ext4_used_dirs_count(sb, gdp) + 1;
-		ext4_used_dirs_set(sb, gdp, count);
-	}
-	gdp->bg_checksum = ext4_group_desc_csum(sbi, group, gdp);
-	spin_unlock(sb_bgl_lock(sbi, group));
 	BUFFER_TRACE(group_desc_bh, "call ext4_handle_dirty_metadata");
 	err = ext4_handle_dirty_metadata(handle, NULL, group_desc_bh);
-	if (err) goto fail;
+	if (err)
+		goto fail;
 
 	percpu_counter_dec(&sbi->s_freeinodes_counter);
 	if (S_ISDIR(mode))