ext4: add lockdep annotations for i_data_sem

With the internal Quota feature, mke2fs creates empty quota inodes and
quota usage tracking is enabled as soon as the file system is mounted.
Since quotacheck is no longer preallocating all of the blocks in the
quota inode that are likely needed to be written to, we are now seeing
a lockdep false positive caused by needing to allocate a quota block
from inside ext4_map_blocks(), while holding i_data_sem for a data
inode.  This results in this complaint:

  Possible unsafe locking scenario:

        CPU0                    CPU1
        ----                    ----
   lock(&ei->i_data_sem);
                                lock(&s->s_dquot.dqio_mutex);
                                lock(&ei->i_data_sem);
   lock(&s->s_dquot.dqio_mutex);

Google-Bug-Id: 27907753

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@vger.kernel.org

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 7e4e3e9b44942189be5fc4201e5871b8b5b7f9be..6a857af0e6b16108b391b1cf2b0dd4caae4dfbc7 100644 (file)
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -911,6 +911,29 @@ do {                                                                              \
 
 #include "extents_status.h"
 
+/*
+ * Lock subclasses for i_data_sem in the ext4_inode_info structure.
+ *
+ * These are needed to avoid lockdep false positives when we need to
+ * allocate blocks to the quota inode during ext4_map_blocks(), while
+ * holding i_data_sem for a normal (non-quota) inode.  Since we don't
+ * do quota tracking for the quota inode, this avoids deadlock (as
+ * well as infinite recursion, since it isn't turtles all the way
+ * down...)
+ *
+ *  I_DATA_SEM_NORMAL - Used for most inodes
+ *  I_DATA_SEM_OTHER  - Used by move_inode.c for the second normal inode
+ *                       where the second inode has larger inode number
+ *                       than the first
+ *  I_DATA_SEM_QUOTA  - Used for quota inodes only
+ */
+enum {
+       I_DATA_SEM_NORMAL = 0,
+       I_DATA_SEM_OTHER,
+       I_DATA_SEM_QUOTA,
+};
+
+
 /*
  * fourth extended file system inode data in memory
  */

diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 4098acc701c3e66d5e8a229ed49828d69997655f..796ff0eafd3c352beccb2347edcc3df87fc91422 100644 (file)
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -60,10 +60,10 @@ ext4_double_down_write_data_sem(struct inode *first, struct inode *second)
 {
        if (first < second) {
                down_write(&EXT4_I(first)->i_data_sem);
-               down_write_nested(&EXT4_I(second)->i_data_sem, SINGLE_DEPTH_NESTING);
+               down_write_nested(&EXT4_I(second)->i_data_sem, I_DATA_SEM_OTHER);
        } else {
                down_write(&EXT4_I(second)->i_data_sem);
-               down_write_nested(&EXT4_I(first)->i_data_sem, SINGLE_DEPTH_NESTING);
+               down_write_nested(&EXT4_I(first)->i_data_sem, I_DATA_SEM_OTHER);
 
        }
 }
@@ -484,6 +484,13 @@ mext_check_arguments(struct inode *orig_inode,
                return -EBUSY;
        }
 
+       if (IS_NOQUOTA(orig_inode) || IS_NOQUOTA(donor_inode)) {
+               ext4_debug("ext4 move extent: The argument files should "
+                       "not be quota files [ino:orig %lu, donor %lu]\n",
+                       orig_inode->i_ino, donor_inode->i_ino);
+               return -EBUSY;
+       }
+
        /* Ext4 move extent supports only extent based file */
        if (!(ext4_test_inode_flag(orig_inode, EXT4_INODE_EXTENTS))) {
                ext4_debug("ext4 move extent: orig file is not extents "

diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 5392975158963118f7eeba2dcf8d560ec485e53d..7ebd9f040b16c152aa8d62abb16ddc8b685309f0 100644 (file)
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -5028,6 +5028,20 @@ static int ext4_quota_on_mount(struct super_block *sb, int type)
                                        EXT4_SB(sb)->s_jquota_fmt, type);
 }
 
+static void lockdep_set_quota_inode(struct inode *inode, int subclass)
+{
+       struct ext4_inode_info *ei = EXT4_I(inode);
+
+       /* The first argument of lockdep_set_subclass has to be
+        * *exactly* the same as the argument to init_rwsem() --- in
+        * this case, in init_once() --- or lockdep gets unhappy
+        * because the name of the lock is set using the
+        * stringification of the argument to init_rwsem().
+        */
+       (void) ei;      /* shut up clang warning if !CONFIG_LOCKDEP */
+       lockdep_set_subclass(&ei->i_data_sem, subclass);
+}
+
 /*
  * Standard function to be called on quota_on
  */
@@ -5067,8 +5081,12 @@ static int ext4_quota_on(struct super_block *sb, int type, int format_id,
                if (err)
                        return err;
        }
-
-       return dquot_quota_on(sb, type, format_id, path);
+       lockdep_set_quota_inode(path->dentry->d_inode, I_DATA_SEM_QUOTA);
+       err = dquot_quota_on(sb, type, format_id, path);
+       if (err)
+               lockdep_set_quota_inode(path->dentry->d_inode,
+                                            I_DATA_SEM_NORMAL);
+       return err;
 }
 
 static int ext4_quota_enable(struct super_block *sb, int type, int format_id,
@@ -5095,8 +5113,11 @@ static int ext4_quota_enable(struct super_block *sb, int type, int format_id,
 
        /* Don't account quota for quota files to avoid recursion */
        qf_inode->i_flags |= S_NOQUOTA;
+       lockdep_set_quota_inode(qf_inode, I_DATA_SEM_QUOTA);
        err = dquot_enable(qf_inode, type, format_id, flags);
        iput(qf_inode);
+       if (err)
+               lockdep_set_quota_inode(qf_inode, I_DATA_SEM_NORMAL);
 
        return err;
 }