Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  [PATCH] get rid of __exit_files(), __exit_fs() and __put_fs_struct()
  [PATCH] proc_readfd_common() race fix
  [PATCH] double-free of inode on alloc_file() failure exit in create_write_pipe()
  [PATCH] teach seq_file to discard entries
  [PATCH] umount_tree() will unhash everything itself
  [PATCH] get rid of more nameidata passing in namespace.c
  [PATCH] switch a bunch of LSM hooks from nameidata to path
  [PATCH] lock exclusively in collect_mounts() and drop_collected_mounts()
  [PATCH] move a bunch of declarations to fs/internal.h

8 files changed, 63 insertions, 43 deletions

diff --git a/fs/internal.h b/fs/internal.h
index 392e8cc..80aa9a0 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -43,3 +43,14 @@ extern void __init chrdev_init(void);
  * namespace.c
  */
 extern int copy_mount_options(const void __user *, unsigned long *);
+
+extern void free_vfsmnt(struct vfsmount *);
+extern struct vfsmount *alloc_vfsmnt(const char *);
+extern struct vfsmount *__lookup_mnt(struct vfsmount *, struct dentry *, int);
+extern void mnt_set_mountpoint(struct vfsmount *, struct dentry *,
+				struct vfsmount *);
+extern void release_mounts(struct list_head *);
+extern void umount_tree(struct vfsmount *, int, struct list_head *);
+extern struct vfsmount *copy_tree(struct vfsmount *, struct dentry *, int);
+
+extern void __init mnt_init(void);
diff --git a/fs/namespace.c b/fs/namespace.c
index 678f7ce..1bf302d 100644
--- a/fs/namespace.c
+++ b/fs/namespace.c
@@ -1091,20 +1091,20 @@ Enomem:
 struct vfsmount *collect_mounts(struct vfsmount *mnt, struct dentry *dentry)
 {
 	struct vfsmount *tree;
-	down_read(&namespace_sem);
+	down_write(&namespace_sem);
 	tree = copy_tree(mnt, dentry, CL_COPY_ALL | CL_PRIVATE);
-	up_read(&namespace_sem);
+	up_write(&namespace_sem);
 	return tree;
 }
 
 void drop_collected_mounts(struct vfsmount *mnt)
 {
 	LIST_HEAD(umount_list);
-	down_read(&namespace_sem);
+	down_write(&namespace_sem);
 	spin_lock(&vfsmount_lock);
 	umount_tree(mnt, 0, &umount_list);
 	spin_unlock(&vfsmount_lock);
-	up_read(&namespace_sem);
+	up_write(&namespace_sem);
 	release_mounts(&umount_list);
 }
 
@@ -1205,32 +1205,32 @@ static int attach_recursive_mnt(struct vfsmount *source_mnt,
 	return 0;
 }
 
-static int graft_tree(struct vfsmount *mnt, struct nameidata *nd)
+static int graft_tree(struct vfsmount *mnt, struct path *path)
 {
 	int err;
 	if (mnt->mnt_sb->s_flags & MS_NOUSER)
 		return -EINVAL;
 
-	if (S_ISDIR(nd->path.dentry->d_inode->i_mode) !=
+	if (S_ISDIR(path->dentry->d_inode->i_mode) !=
 	      S_ISDIR(mnt->mnt_root->d_inode->i_mode))
 		return -ENOTDIR;
 
 	err = -ENOENT;
-	mutex_lock(&nd->path.dentry->d_inode->i_mutex);
-	if (IS_DEADDIR(nd->path.dentry->d_inode))
+	mutex_lock(&path->dentry->d_inode->i_mutex);
+	if (IS_DEADDIR(path->dentry->d_inode))
 		goto out_unlock;
 
-	err = security_sb_check_sb(mnt, nd);
+	err = security_sb_check_sb(mnt, path);
 	if (err)
 		goto out_unlock;
 
 	err = -ENOENT;
-	if (IS_ROOT(nd->path.dentry) || !d_unhashed(nd->path.dentry))
-		err = attach_recursive_mnt(mnt, &nd->path, NULL);
+	if (IS_ROOT(path->dentry) || !d_unhashed(path->dentry))
+		err = attach_recursive_mnt(mnt, path, NULL);
 out_unlock:
-	mutex_unlock(&nd->path.dentry->d_inode->i_mutex);
+	mutex_unlock(&path->dentry->d_inode->i_mutex);
 	if (!err)
-		security_sb_post_addmount(mnt, nd);
+		security_sb_post_addmount(mnt, path);
 	return err;
 }
 
@@ -1294,7 +1294,7 @@ static noinline int do_loopback(struct nameidata *nd, char *old_name,
 	if (!mnt)
 		goto out;
 
-	err = graft_tree(mnt, nd);
+	err = graft_tree(mnt, &nd->path);
 	if (err) {
 		LIST_HEAD(umount_list);
 		spin_lock(&vfsmount_lock);
@@ -1501,7 +1501,7 @@ int do_add_mount(struct vfsmount *newmnt, struct nameidata *nd,
 		goto unlock;
 
 	newmnt->mnt_flags = mnt_flags;
-	if ((err = graft_tree(newmnt, nd)))
+	if ((err = graft_tree(newmnt, &nd->path)))
 		goto unlock;
 
 	if (fslist) /* add to the specified expiration list */
@@ -1746,7 +1746,8 @@ long do_mount(char *dev_name, char *dir_name, char *type_page,
 	if (retval)
 		return retval;
 
-	retval = security_sb_mount(dev_name, &nd, type_page, flags, data_page);
+	retval = security_sb_mount(dev_name, &nd.path,
+				   type_page, flags, data_page);
 	if (retval)
 		goto dput_out;
 
@@ -1986,15 +1987,13 @@ asmlinkage long sys_pivot_root(const char __user * new_root,
 			       const char __user * put_old)
 {
 	struct vfsmount *tmp;
-	struct nameidata new_nd, old_nd, user_nd;
-	struct path parent_path, root_parent;
+	struct nameidata new_nd, old_nd;
+	struct path parent_path, root_parent, root;
 	int error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
 
-	lock_kernel();
-
 	error = __user_walk(new_root, LOOKUP_FOLLOW | LOOKUP_DIRECTORY,
 			    &new_nd);
 	if (error)
@@ -2007,14 +2006,14 @@ asmlinkage long sys_pivot_root(const char __user * new_root,
 	if (error)
 		goto out1;
 
-	error = security_sb_pivotroot(&old_nd, &new_nd);
+	error = security_sb_pivotroot(&old_nd.path, &new_nd.path);
 	if (error) {
 		path_put(&old_nd.path);
 		goto out1;
 	}
 
 	read_lock(&current->fs->lock);
-	user_nd.path = current->fs->root;
+	root = current->fs->root;
 	path_get(&current->fs->root);
 	read_unlock(&current->fs->lock);
 	down_write(&namespace_sem);
@@ -2022,9 +2021,9 @@ asmlinkage long sys_pivot_root(const char __user * new_root,
 	error = -EINVAL;
 	if (IS_MNT_SHARED(old_nd.path.mnt) ||
 		IS_MNT_SHARED(new_nd.path.mnt->mnt_parent) ||
-		IS_MNT_SHARED(user_nd.path.mnt->mnt_parent))
+		IS_MNT_SHARED(root.mnt->mnt_parent))
 		goto out2;
-	if (!check_mnt(user_nd.path.mnt))
+	if (!check_mnt(root.mnt))
 		goto out2;
 	error = -ENOENT;
 	if (IS_DEADDIR(new_nd.path.dentry->d_inode))
@@ -2034,13 +2033,13 @@ asmlinkage long sys_pivot_root(const char __user * new_root,
 	if (d_unhashed(old_nd.path.dentry) && !IS_ROOT(old_nd.path.dentry))
 		goto out2;
 	error = -EBUSY;
-	if (new_nd.path.mnt == user_nd.path.mnt ||
-	    old_nd.path.mnt == user_nd.path.mnt)
+	if (new_nd.path.mnt == root.mnt ||
+	    old_nd.path.mnt == root.mnt)
 		goto out2; /* loop, on the same file system  */
 	error = -EINVAL;
-	if (user_nd.path.mnt->mnt_root != user_nd.path.dentry)
+	if (root.mnt->mnt_root != root.dentry)
 		goto out2; /* not a mountpoint */
-	if (user_nd.path.mnt->mnt_parent == user_nd.path.mnt)
+	if (root.mnt->mnt_parent == root.mnt)
 		goto out2; /* not attached */
 	if (new_nd.path.mnt->mnt_root != new_nd.path.dentry)
 		goto out2; /* not a mountpoint */
@@ -2062,27 +2061,26 @@ asmlinkage long sys_pivot_root(const char __user * new_root,
 	} else if (!is_subdir(old_nd.path.dentry, new_nd.path.dentry))
 		goto out3;
 	detach_mnt(new_nd.path.mnt, &parent_path);
-	detach_mnt(user_nd.path.mnt, &root_parent);
+	detach_mnt(root.mnt, &root_parent);
 	/* mount old root on put_old */
-	attach_mnt(user_nd.path.mnt, &old_nd.path);
+	attach_mnt(root.mnt, &old_nd.path);
 	/* mount new_root on / */
 	attach_mnt(new_nd.path.mnt, &root_parent);
 	touch_mnt_namespace(current->nsproxy->mnt_ns);
 	spin_unlock(&vfsmount_lock);
-	chroot_fs_refs(&user_nd.path, &new_nd.path);
-	security_sb_post_pivotroot(&user_nd, &new_nd);
+	chroot_fs_refs(&root, &new_nd.path);
+	security_sb_post_pivotroot(&root, &new_nd.path);
 	error = 0;
 	path_put(&root_parent);
 	path_put(&parent_path);
 out2:
 	mutex_unlock(&old_nd.path.dentry->d_inode->i_mutex);
 	up_write(&namespace_sem);
-	path_put(&user_nd.path);
+	path_put(&root);
 	path_put(&old_nd.path);
 out1:
 	path_put(&new_nd.path);
 out0:
-	unlock_kernel();
 	return error;
 out3:
 	spin_unlock(&vfsmount_lock);
diff --git a/fs/pipe.c b/fs/pipe.c
index 8be381b..f73492b 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -988,7 +988,10 @@ struct file *create_write_pipe(void)
 	return f;
 
  err_dentry:
+	free_pipe_info(inode);
 	dput(dentry);
+	return ERR_PTR(err);
+
  err_inode:
 	free_pipe_info(inode);
 	iput(inode);
diff --git a/fs/pnode.c b/fs/pnode.c
index 1d8f544..f968e35 100644
--- a/fs/pnode.c
+++ b/fs/pnode.c
@@ -9,6 +9,7 @@
 #include <linux/mnt_namespace.h>
 #include <linux/mount.h>
 #include <linux/fs.h>
+#include "internal.h"
 #include "pnode.h"
 
 /* return the next shared peer mount of @p */
@@ -211,8 +212,7 @@ int propagate_mnt(struct vfsmount *dest_mnt, struct dentry *dest_dentry,
 out:
 	spin_lock(&vfsmount_lock);
 	while (!list_empty(&tmp_list)) {
-		child = list_entry(tmp_list.next, struct vfsmount, mnt_hash);
-		list_del_init(&child->mnt_hash);
+		child = list_first_entry(&tmp_list, struct vfsmount, mnt_hash);
 		umount_tree(child, 0, &umount_list);
 	}
 	spin_unlock(&vfsmount_lock);
diff --git a/fs/pnode.h b/fs/pnode.h
index f249be2..973c3f8 100644
--- a/fs/pnode.h
+++ b/fs/pnode.h
@@ -35,4 +35,5 @@ int propagate_mnt(struct vfsmount *, struct dentry *, struct vfsmount *,
 		struct list_head *);
 int propagate_umount(struct list_head *);
 int propagate_mount_busy(struct vfsmount *, int);
+void mnt_release_group_id(struct vfsmount *);
 #endif /* _LINUX_PNODE_H */
diff --git a/fs/proc/base.c b/fs/proc/base.c
index 81d7d14..7313c62 100644
--- a/fs/proc/base.c
+++ b/fs/proc/base.c
@@ -1626,7 +1626,6 @@ static int proc_readfd_common(struct file * filp, void * dirent,
 	unsigned int fd, ino;
 	int retval;
 	struct files_struct * files;
-	struct fdtable *fdt;
 
 	retval = -ENOENT;
 	if (!p)
@@ -1649,9 +1648,8 @@ static int proc_readfd_common(struct file * filp, void * dirent,
 			if (!files)
 				goto out;
 			rcu_read_lock();
-			fdt = files_fdtable(files);
 			for (fd = filp->f_pos-2;
-			     fd < fdt->max_fds;
+			     fd < files_fdtable(files)->max_fds;
 			     fd++, filp->f_pos++) {
 				char name[PROC_NUMBUF];
 				int len;
diff --git a/fs/seq_file.c b/fs/seq_file.c
index 9943408..cdfd996 100644
--- a/fs/seq_file.c
+++ b/fs/seq_file.c
@@ -25,6 +25,7 @@
  *	into the buffer.  In case of error ->start() and ->next() return
  *	ERR_PTR(error).  In the end of sequence they return %NULL. ->show()
  *	returns 0 in case of success and negative number in case of error.
+ *	Returning SEQ_SKIP means "discard this element and move on".
  */
 int seq_open(struct file *file, const struct seq_operations *op)
 {
@@ -114,8 +115,10 @@ ssize_t seq_read(struct file *file, char __user *buf, size_t size, loff_t *ppos)
 		if (!p || IS_ERR(p))
 			break;
 		err = m->op->show(m, p);
-		if (err)
+		if (err < 0)
 			break;
+		if (unlikely(err))
+			m->count = 0;
 		if (m->count < m->size)
 			goto Fill;
 		m->op->stop(m, p);
@@ -140,9 +143,10 @@ Fill:
 			break;
 		}
 		err = m->op->show(m, p);
-		if (err || m->count == m->size) {
+		if (m->count == m->size || err) {
 			m->count = offs;
-			break;
+			if (likely(err <= 0))
+				break;
 		}
 		pos = next;
 	}
@@ -199,8 +203,12 @@ static int traverse(struct seq_file *m, loff_t offset)
 		if (IS_ERR(p))
 			break;
 		error = m->op->show(m, p);
-		if (error)
+		if (error < 0)
 			break;
+		if (unlikely(error)) {
+			error = 0;
+			m->count = 0;
+		}
 		if (m->count == m->size)
 			goto Eoverflow;
 		if (pos + m->count > offset) {
diff --git a/fs/super.c b/fs/super.c
index 1f8f05e..4798350 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -39,6 +39,7 @@
 #include <linux/mutex.h>
 #include <linux/file.h>
 #include <asm/uaccess.h>
+#include "internal.h"
 
 
 LIST_HEAD(super_blocks);