1 files changed, 128 insertions, 95 deletions

diff --git a/fs/namespace.c b/fs/namespace.c
index 3ddfd9046c44..7b0b95371696 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -183,7 +183,7 @@ static inline void mnt_dec_count(struct vfsmount *mnt)
 unsigned int mnt_get_count(struct vfsmount *mnt)
 {
 #ifdef CONFIG_SMP
-        unsigned int count = atomic_read(&mnt->mnt_longrefs);
+        unsigned int count = 0;
         int cpu;
 
         for_each_possible_cpu(cpu) {
@@ -217,7 +217,7 @@ struct vfsmount *alloc_vfsmnt(const char *name)
                 if (!mnt->mnt_pcp)
                         goto out_free_devname;
 
-                atomic_set(&mnt->mnt_longrefs, 1);
+                this_cpu_add(mnt->mnt_pcp->mnt_count, 1);
 #else
                 mnt->mnt_count = 1;
                 mnt->mnt_writers = 0;
@@ -611,6 +611,21 @@ static void attach_mnt(struct vfsmount *mnt, struct path *path)
         list_add_tail(&mnt->mnt_child, &path->mnt->mnt_mounts);
 }
 
+static inline void __mnt_make_longterm(struct vfsmount *mnt)
+{
+#ifdef CONFIG_SMP
+        atomic_inc(&mnt->mnt_longterm);
+#endif
+}
+
+/* needs vfsmount lock for write */
+static inline void __mnt_make_shortterm(struct vfsmount *mnt)
+{
+#ifdef CONFIG_SMP
+        atomic_dec(&mnt->mnt_longterm);
+#endif
+}
+
 /*
  * vfsmount lock must be held for write
  */
@@ -624,8 +639,11 @@ static void commit_tree(struct vfsmount *mnt)
         BUG_ON(parent == mnt);
 
         list_add_tail(&head, &mnt->mnt_list);
-        list_for_each_entry(m, &head, mnt_list)
+        list_for_each_entry(m, &head, mnt_list) {
                 m->mnt_ns = n;
+                __mnt_make_longterm(m);
+        }
+
         list_splice(&head, n->list.prev);
 
         list_add_tail(&mnt->mnt_hash, mount_hashtable +
@@ -734,51 +752,30 @@ static inline void mntfree(struct vfsmount *mnt)
         deactivate_super(sb);
 }
 
-#ifdef CONFIG_SMP
-static inline void __mntput(struct vfsmount *mnt, int longrefs)
+static void mntput_no_expire(struct vfsmount *mnt)
 {
-        if (!longrefs) {
 put_again:
-                br_read_lock(vfsmount_lock);
-                if (likely(atomic_read(&mnt->mnt_longrefs))) {
-                        mnt_dec_count(mnt);
-                        br_read_unlock(vfsmount_lock);
-                        return;
-                }
+#ifdef CONFIG_SMP
+        br_read_lock(vfsmount_lock);
+        if (likely(atomic_read(&mnt->mnt_longterm))) {
+                mnt_dec_count(mnt);
                 br_read_unlock(vfsmount_lock);
-        } else {
-                BUG_ON(!atomic_read(&mnt->mnt_longrefs));
-                if (atomic_add_unless(&mnt->mnt_longrefs, -1, 1))
-                        return;
+                return;
         }
+        br_read_unlock(vfsmount_lock);
 
         br_write_lock(vfsmount_lock);
-        if (!longrefs)
-                mnt_dec_count(mnt);
-        else
-                atomic_dec(&mnt->mnt_longrefs);
+        mnt_dec_count(mnt);
         if (mnt_get_count(mnt)) {
                 br_write_unlock(vfsmount_lock);
                 return;
         }
-        if (unlikely(mnt->mnt_pinned)) {
-                mnt_add_count(mnt, mnt->mnt_pinned + 1);
-                mnt->mnt_pinned = 0;
-                br_write_unlock(vfsmount_lock);
-                acct_auto_close_mnt(mnt);
-                goto put_again;
-        }
-        br_write_unlock(vfsmount_lock);
-        mntfree(mnt);
-}
 #else
-static inline void __mntput(struct vfsmount *mnt, int longrefs)
-{
-put_again:
         mnt_dec_count(mnt);
         if (likely(mnt_get_count(mnt)))
                 return;
         br_write_lock(vfsmount_lock);
+#endif
         if (unlikely(mnt->mnt_pinned)) {
                 mnt_add_count(mnt, mnt->mnt_pinned + 1);
                 mnt->mnt_pinned = 0;
@@ -789,12 +786,6 @@ put_again:
         br_write_unlock(vfsmount_lock);
         mntfree(mnt);
 }
-#endif
-
-static void mntput_no_expire(struct vfsmount *mnt)
-{
-        __mntput(mnt, 0);
-}
 
 void mntput(struct vfsmount *mnt)
 {
@@ -802,7 +793,7 @@ void mntput(struct vfsmount *mnt)
                 /* avoid cacheline pingpong, hope gcc doesn't get "smart" */
                 if (unlikely(mnt->mnt_expiry_mark))
                         mnt->mnt_expiry_mark = 0;
-                __mntput(mnt, 0);
+                mntput_no_expire(mnt);
         }
 }
 EXPORT_SYMBOL(mntput);
@@ -815,33 +806,6 @@ struct vfsmount *mntget(struct vfsmount *mnt)
 }
 EXPORT_SYMBOL(mntget);
 
-void mntput_long(struct vfsmount *mnt)
-{
-#ifdef CONFIG_SMP
-        if (mnt) {
-                /* avoid cacheline pingpong, hope gcc doesn't get "smart" */
-                if (unlikely(mnt->mnt_expiry_mark))
-                        mnt->mnt_expiry_mark = 0;
-                __mntput(mnt, 1);
-        }
-#else
-        mntput(mnt);
-#endif
-}
-EXPORT_SYMBOL(mntput_long);
-
-struct vfsmount *mntget_long(struct vfsmount *mnt)
-{
-#ifdef CONFIG_SMP
-        if (mnt)
-                atomic_inc(&mnt->mnt_longrefs);
-        return mnt;
-#else
-        return mntget(mnt);
-#endif
-}
-EXPORT_SYMBOL(mntget_long);
-
 void mnt_pin(struct vfsmount *mnt)
 {
         br_write_lock(vfsmount_lock);
@@ -1216,7 +1180,7 @@ void release_mounts(struct list_head *head)
                         dput(dentry);
                         mntput(m);
                 }
-                mntput_long(mnt);
+                mntput(mnt);
         }
 }
 
@@ -1226,19 +1190,21 @@ void release_mounts(struct list_head *head)
  */
 void umount_tree(struct vfsmount *mnt, int propagate, struct list_head *kill)
 {
+        LIST_HEAD(tmp_list);
         struct vfsmount *p;
 
         for (p = mnt; p; p = next_mnt(p, mnt))
-                list_move(&p->mnt_hash, kill);
+                list_move(&p->mnt_hash, &tmp_list);
 
         if (propagate)
-                propagate_umount(kill);
+                propagate_umount(&tmp_list);
 
-        list_for_each_entry(p, kill, mnt_hash) {
+        list_for_each_entry(p, &tmp_list, mnt_hash) {
                 list_del_init(&p->mnt_expire);
                 list_del_init(&p->mnt_list);
                 __touch_mnt_namespace(p->mnt_ns);
                 p->mnt_ns = NULL;
+                __mnt_make_shortterm(p);
                 list_del_init(&p->mnt_child);
                 if (p->mnt_parent != p) {
                         p->mnt_parent->mnt_ghosts++;
@@ -1246,6 +1212,7 @@ void umount_tree(struct vfsmount *mnt, int propagate, struct list_head *kill)
                 }
                 change_mnt_propagation(p, MS_PRIVATE);
         }
+        list_splice(&tmp_list, kill);
 }
 
 static void shrink_submounts(struct vfsmount *mnt, struct list_head *umounts);
@@ -1844,9 +1811,10 @@ static int do_move_mount(struct path *path, char *old_name)
                 return err;
 
         down_write(&namespace_sem);
-        while (d_mountpoint(path->dentry) &&
-               follow_down(path))
-                ;
+        err = follow_down(path, true);
+        if (err < 0)
+                goto out;
+
         err = -EINVAL;
         if (!check_mnt(path->mnt) || !check_mnt(old_path.mnt))
                 goto out;
@@ -1904,6 +1872,8 @@ out:
         return err;
 }
 
+static int do_add_mount(struct vfsmount *, struct path *, int);
+
 /*
  * create a new mount for userspace and request it to be added into the
  * namespace's tree
@@ -1912,6 +1882,7 @@ static int do_new_mount(struct path *path, char *type, int flags,
                         int mnt_flags, char *name, void *data)
 {
         struct vfsmount *mnt;
+        int err;
 
         if (!type)
                 return -EINVAL;
@@ -1924,15 +1895,47 @@ static int do_new_mount(struct path *path, char *type, int flags,
         if (IS_ERR(mnt))
                 return PTR_ERR(mnt);
 
-        return do_add_mount(mnt, path, mnt_flags, NULL);
+        err = do_add_mount(mnt, path, mnt_flags);
+        if (err)
+                mntput(mnt);
+        return err;
+}
+
+int finish_automount(struct vfsmount *m, struct path *path)
+{
+        int err;
+        /* The new mount record should have at least 2 refs to prevent it being
+         * expired before we get a chance to add it
+         */
+        BUG_ON(mnt_get_count(m) < 2);
+
+        if (m->mnt_sb == path->mnt->mnt_sb &&
+            m->mnt_root == path->dentry) {
+                err = -ELOOP;
+                goto fail;
+        }
+
+        err = do_add_mount(m, path, path->mnt->mnt_flags | MNT_SHRINKABLE);
+        if (!err)
+                return 0;
+fail:
+        /* remove m from any expiration list it may be on */
+        if (!list_empty(&m->mnt_expire)) {
+                down_write(&namespace_sem);
+                br_write_lock(vfsmount_lock);
+                list_del_init(&m->mnt_expire);
+                br_write_unlock(vfsmount_lock);
+                up_write(&namespace_sem);
+        }
+        mntput(m);
+        mntput(m);
+        return err;
 }
 
 /*
  * add a mount into a namespace's mount tree
- * - provide the option of adding the new mount to an expiration list
  */
-int do_add_mount(struct vfsmount *newmnt, struct path *path,
-                 int mnt_flags, struct list_head *fslist)
+static int do_add_mount(struct vfsmount *newmnt, struct path *path, int mnt_flags)
 {
         int err;
 
@@ -1940,9 +1943,10 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path,
 
         down_write(&namespace_sem);
         /* Something was mounted here while we slept */
-        while (d_mountpoint(path->dentry) &&
-               follow_down(path))
-                ;
+        err = follow_down(path, true);
+        if (err < 0)
+                goto unlock;
+
         err = -EINVAL;
         if (!(mnt_flags & MNT_SHRINKABLE) && !check_mnt(path->mnt))
                 goto unlock;
@@ -1958,22 +1962,29 @@ int do_add_mount(struct vfsmount *newmnt, struct path *path,
                 goto unlock;
 
         newmnt->mnt_flags = mnt_flags;
-        if ((err = graft_tree(newmnt, path)))
-                goto unlock;
-
-        if (fslist) /* add to the specified expiration list */
-                list_add_tail(&newmnt->mnt_expire, fslist);
-
-        up_write(&namespace_sem);
-        return 0;
+        err = graft_tree(newmnt, path);
 
 unlock:
         up_write(&namespace_sem);
-        mntput_long(newmnt);
         return err;
 }
 
-EXPORT_SYMBOL_GPL(do_add_mount);
+/**
+ * mnt_set_expiry - Put a mount on an expiration list
+ * @mnt: The mount to list.
+ * @expiry_list: The list to add the mount to.
+ */
+void mnt_set_expiry(struct vfsmount *mnt, struct list_head *expiry_list)
+{
+        down_write(&namespace_sem);
+        br_write_lock(vfsmount_lock);
+
+        list_add_tail(&mnt->mnt_expire, expiry_list);
+
+        br_write_unlock(vfsmount_lock);
+        up_write(&namespace_sem);
+}
+EXPORT_SYMBOL(mnt_set_expiry);
 
 /*
  * process a list of expirable mountpoints with the intent of discarding any
@@ -2262,6 +2273,22 @@ static struct mnt_namespace *alloc_mnt_ns(void)
         return new_ns;
 }
 
+void mnt_make_longterm(struct vfsmount *mnt)
+{
+        __mnt_make_longterm(mnt);
+}
+
+void mnt_make_shortterm(struct vfsmount *mnt)
+{
+#ifdef CONFIG_SMP
+        if (atomic_add_unless(&mnt->mnt_longterm, -1, 1))
+                return;
+        br_write_lock(vfsmount_lock);
+        atomic_dec(&mnt->mnt_longterm);
+        br_write_unlock(vfsmount_lock);
+#endif
+}
+
 /*
  * Allocate a new namespace structure and populate it with contents
  * copied from the namespace of the passed in task structure.
@@ -2299,14 +2326,19 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns,
         q = new_ns->root;
         while (p) {
                 q->mnt_ns = new_ns;
+                __mnt_make_longterm(q);
                 if (fs) {
                         if (p == fs->root.mnt) {
+                                fs->root.mnt = mntget(q);
+                                __mnt_make_longterm(q);
+                                mnt_make_shortterm(p);
                                 rootmnt = p;
-                                fs->root.mnt = mntget_long(q);
                         }
                         if (p == fs->pwd.mnt) {
+                                fs->pwd.mnt = mntget(q);
+                                __mnt_make_longterm(q);
+                                mnt_make_shortterm(p);
                                 pwdmnt = p;
-                                fs->pwd.mnt = mntget_long(q);
                         }
                 }
                 p = next_mnt(p, mnt_ns->root);
@@ -2315,9 +2347,9 @@ static struct mnt_namespace *dup_mnt_ns(struct mnt_namespace *mnt_ns,
         up_write(&namespace_sem);
 
         if (rootmnt)
-                mntput_long(rootmnt);
+                mntput(rootmnt);
         if (pwdmnt)
-                mntput_long(pwdmnt);
+                mntput(pwdmnt);
 
         return new_ns;
 }
@@ -2350,6 +2382,7 @@ struct mnt_namespace *create_mnt_ns(struct vfsmount *mnt)
         new_ns = alloc_mnt_ns();
         if (!IS_ERR(new_ns)) {
                 mnt->mnt_ns = new_ns;
+                __mnt_make_longterm(mnt);
                 new_ns->root = mnt;
                 list_add(&new_ns->list, &new_ns->root->mnt_list);
         }