sysfs: make sysfs_drop_dentry() access inodes using ilookup()
authorTejun Heo <htejun@gmail.com>
Wed, 13 Jun 2007 19:27:21 +0000 (04:27 +0900)
committerGreg Kroah-Hartman <gregkh@suse.de>
Wed, 11 Jul 2007 23:09:08 +0000 (16:09 -0700)
sysfs_drop_dentry() used to go through sd->s_dentry and
sd->s_parent->s_dentry to access the inodes.  This is incorrect
because inode can be cached without dentry.

This patch makes sysfs_drop_dentry() access inodes using ilookup() on
sd->s_ino.  This is both correct and simpler.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
fs/sysfs/inode.c

index ee31bf369a88b1eaf647aae9ea4c30695140dcc2..63daa06c4194bc19df4bd5f222e60b97e5e0645d 100644 (file)
@@ -219,9 +219,9 @@ void sysfs_instantiate(struct dentry *dentry, struct inode *inode)
  */
 void sysfs_drop_dentry(struct sysfs_dirent *sd)
 {
-       struct dentry *dentry = NULL, *parent = NULL;
-       struct inode *dir;
+       struct dentry *dentry = NULL;
        struct timespec curtime;
+       struct inode *inode;
 
        /* We're not holding a reference to ->s_dentry dentry but the
         * field will stay valid as long as sysfs_lock is held.
@@ -229,19 +229,9 @@ void sysfs_drop_dentry(struct sysfs_dirent *sd)
        spin_lock(&sysfs_lock);
        spin_lock(&dcache_lock);
 
+       /* drop dentry if it's there and dput() didn't kill it yet */
        if (sd->s_dentry && sd->s_dentry->d_inode) {
-               /* get dentry if it's there and dput() didn't kill it yet */
                dentry = dget_locked(sd->s_dentry);
-               parent = dentry->d_parent;
-       } else if (sd->s_parent->s_dentry->d_inode) {
-               /* We need to update the parent even if dentry for the
-                * victim itself doesn't exist.
-                */
-               parent = dget_locked(sd->s_parent->s_dentry);
-       }
-
-       /* drop */
-       if (dentry) {
                spin_lock(&dentry->d_lock);
                __d_drop(dentry);
                spin_unlock(&dentry->d_lock);
@@ -250,36 +240,39 @@ void sysfs_drop_dentry(struct sysfs_dirent *sd)
        spin_unlock(&dcache_lock);
        spin_unlock(&sysfs_lock);
 
-       /* nothing to do if the parent isn't in dcache */
-       if (!parent)
-               return;
+       dput(dentry);
+       /* XXX: unpin if directory, this will go away soon */
+       if (sd->s_type & SYSFS_DIR)
+               dput(dentry);
 
        /* adjust nlink and update timestamp */
-       dir = parent->d_inode;
-       mutex_lock(&dir->i_mutex);
-
        curtime = CURRENT_TIME;
 
-       dir->i_ctime = dir->i_mtime = curtime;
+       inode = ilookup(sysfs_sb, sd->s_ino);
+       if (inode) {
+               mutex_lock(&inode->i_mutex);
 
-       if (dentry) {
-               dentry->d_inode->i_ctime = curtime;
-               drop_nlink(dentry->d_inode);
-               if (sd->s_type & SYSFS_DIR) {
-                       drop_nlink(dentry->d_inode);
-                       drop_nlink(dir);
-                       /* XXX: unpin if directory, this will go away soon */
-                       dput(dentry);
-               }
+               inode->i_ctime = curtime;
+               drop_nlink(inode);
+               if (sd->s_type & SYSFS_DIR)
+                       drop_nlink(inode);
+
+               mutex_unlock(&inode->i_mutex);
+               iput(inode);
        }
 
-       mutex_unlock(&dir->i_mutex);
+       /* adjust nlink and udpate timestamp of the parent */
+       inode = ilookup(sysfs_sb, sd->s_parent->s_ino);
+       if (inode) {
+               mutex_lock(&inode->i_mutex);
 
-       /* bye bye */
-       if (dentry)
-               dput(dentry);
-       else
-               dput(parent);
+               inode->i_ctime = inode->i_mtime = curtime;
+               if (sd->s_type & SYSFS_DIR)
+                       drop_nlink(inode);
+
+               mutex_unlock(&inode->i_mutex);
+               iput(inode);
+       }
 }
 
 int sysfs_hash_and_remove(struct dentry * dir, const char * name)