From: Ian Kent <ikent@redhat.com>
Date: Wed, 18 Mar 2009 16:05:14 +0900
Subject: [fs] autofs4: add miscelaneous device for ioctls
Message-id: 20090318070514.9081.59293.stgit@zeus.themaw.net
O-Subject: [RHEL 5.4 PATCH 3/5] autofs4 - add miscelaneous device for ioctls
Bugzilla: 452120
RH-Acked-by: Jeff Layton <jlayton@redhat.com>

From: Ian Kent <raven@themaw.net>

This patch is the third of four patches needed to addresses bug 452120.

Patch to add a miscellaneous device to the autofs4 module for routing
ioctls. This provides the ability to obtain an ioctl file handle for
an autofs mount point that is possibly covered by another mount.

The actual problem with autofs is that it can't reconnect to existing
mounts. Immediately one things of just adding the ability to remount
autofs file systems would solve it, but alas, that can't work. This is
because autofs direct mounts and the implementation of "on demand mount
and expire" of nested mount trees have the file system mounted on top of
the mount trigger dentry.

To resolve this a miscellaneous device node for routing ioctl commands
to these mount points has been implemented in the autofs4 kernel module
and a library added to autofs. This provides the ability to open a file
descriptor for these over mounted autofs mount points.

Please refer to Documentation/filesystems/autofs4-mount-control.txt for
a discussion of the problem, implementation alternatives considered and
a description of the interface.

Signed-off-by: Ian Kent <raven@themaw.net>

diff --git a/fs/autofs4/Makefile b/fs/autofs4/Makefile
index f2c3b79..a811c1f 100644
--- a/fs/autofs4/Makefile
+++ b/fs/autofs4/Makefile
@@ -4,4 +4,4 @@
 
 obj-$(CONFIG_AUTOFS4_FS) += autofs4.o
 
-autofs4-objs := init.o inode.o root.o symlink.o waitq.o expire.o
+autofs4-objs := init.o inode.o root.o symlink.o waitq.o expire.o dev-ioctl.o
diff --git a/fs/autofs4/autofs_i.h b/fs/autofs4/autofs_i.h
index d39c197..18c27f9 100644
--- a/fs/autofs4/autofs_i.h
+++ b/fs/autofs4/autofs_i.h
@@ -14,6 +14,7 @@
 /* Internal header file for autofs */
 
 #include <linux/auto_fs4.h>
+#include <linux/auto_dev-ioctl.h>
 #include <linux/mutex.h>
 #include <linux/list.h>
 
@@ -21,6 +22,9 @@
 #define AUTOFS_IOC_FIRST     AUTOFS_IOC_READY
 #define AUTOFS_IOC_COUNT     32
 
+#define AUTOFS_DEV_IOCTL_IOC_FIRST	(AUTOFS_DEV_IOCTL_VERSION)
+#define AUTOFS_DEV_IOCTL_IOC_COUNT	(AUTOFS_IOC_COUNT - 11)
+
 #define AUTOFS_TYPE_TRIGGER	(AUTOFS_TYPE_DIRECT|AUTOFS_TYPE_OFFSET)
 
 #include <linux/kernel.h>
@@ -37,11 +41,27 @@
 /* #define DEBUG */
 
 #ifdef DEBUG
-#define DPRINTK(fmt,args...) do { printk(KERN_DEBUG "pid %d: %s: " fmt "\n" , current->pid , __FUNCTION__ , ##args); } while(0)
+#define DPRINTK(fmt,args...) 				\
+do {							\
+	printk(KERN_DEBUG "pid %d: %s: " fmt "\n",i	\
+		current->pid , __FUNCTION__ , ##args);	\
+} while(0)
 #else
-#define DPRINTK(fmt,args...) do {} while(0)
+#define DPRINTK(fmt, args...) do {} while(0)
 #endif
 
+#define AUTOFS_WARN(fmt, args...)			\
+do {							\
+	printk(KERN_WARNING "pid %d: %s: " fmt "\n",	\
+		current->pid, __FUNCTION__, ##args);	\
+} while (0)
+
+#define AUTOFS_ERROR(fmt, args...)			\
+do {							\
+	printk(KERN_ERR "pid %d: %s: " fmt "\n",	\
+		current->pid, __FUNCTION__, ##args);	\
+} while (0)
+
 #define AUTOFS_SUPER_MAGIC 0x0187
 
 /* Unified info structure.  This is pointed to by both the dentry and
@@ -169,8 +189,21 @@ int autofs4_expire_wait(struct dentry *dentry);
 int autofs4_expire_run(struct super_block *, struct vfsmount *,
 			struct autofs_sb_info *,
 			struct autofs_packet_expire __user *);
+int autofs4_do_expire_multi(struct super_block *sb, struct vfsmount *mnt,
+			    struct autofs_sb_info *sbi, int when);
 int autofs4_expire_multi(struct super_block *, struct vfsmount *,
 			struct autofs_sb_info *, int __user *);
+struct dentry *autofs4_expire_direct(struct super_block *sb,
+				     struct vfsmount *mnt,
+				     struct autofs_sb_info *sbi, int how);
+struct dentry *autofs4_expire_indirect(struct super_block *sb,
+				       struct vfsmount *mnt,
+				       struct autofs_sb_info *sbi, int how);
+
+/* Device node initialization */
+
+int autofs_dev_ioctl_init(void);
+void autofs_dev_ioctl_exit(void);
 
 /* Operations structures */
 
diff --git a/fs/autofs4/dev-ioctl.c b/fs/autofs4/dev-ioctl.c
new file mode 100644
index 0000000..13dce64
--- /dev/null
+++ b/fs/autofs4/dev-ioctl.c
@@ -0,0 +1,840 @@
+/*
+ * Copyright 2008 Red Hat, Inc. All rights reserved.
+ * Copyright 2008 Ian Kent <raven@themaw.net>
+ *
+ * This file is part of the Linux kernel and is made available under
+ * the terms of the GNU General Public License, version 2, or at your
+ * option, any later version, incorporated herein by reference.
+ */
+
+#include <linux/module.h>
+#include <linux/vmalloc.h>
+#include <linux/miscdevice.h>
+#include <linux/init.h>
+#include <linux/wait.h>
+#include <linux/dcache.h>
+#include <linux/namei.h>
+#include <linux/fcntl.h>
+#include <linux/file.h>
+#include <linux/sched.h>
+#include <linux/compat.h>
+#include <linux/syscalls.h>
+#include <linux/smp_lock.h>
+#include <linux/uaccess.h>
+
+#include "autofs_i.h"
+
+/*
+ * This module implements an interface for routing autofs ioctl control
+ * commands via a miscellaneous device file.
+ *
+ * The alternate interface is needed because we need to be able open
+ * an ioctl file descriptor on an autofs mount that may be covered by
+ * another mount. This situation arises when starting automount(8)
+ * or other user space daemon which uses direct mounts or offset
+ * mounts (used for autofs lazy mount/umount of nested mount trees),
+ * which have been left busy at at service shutdown.
+ */
+
+#define AUTOFS_DEV_IOCTL_SIZE	sizeof(struct autofs_dev_ioctl)
+
+typedef int (*ioctl_fn)(struct file *, struct autofs_sb_info *,
+			struct autofs_dev_ioctl *);
+
+static int check_name(const char *name)
+{
+	if (!strchr(name, '/'))
+		return -EINVAL;
+	return 0;
+}
+
+/*
+ * Check a string doesn't overrun the chunk of
+ * memory we copied from user land.
+ */
+static int invalid_str(char *str, void *end)
+{
+	while ((void *) str <= end)
+		if (!*str++)
+			return 0;
+	return -EINVAL;
+}
+
+/*
+ * Check that the user compiled against correct version of autofs
+ * misc device code.
+ *
+ * As well as checking the version compatibility this always copies
+ * the kernel interface version out.
+ */
+static int check_dev_ioctl_version(int cmd, struct autofs_dev_ioctl *param)
+{
+	int err = 0;
+
+	if ((AUTOFS_DEV_IOCTL_VERSION_MAJOR != param->ver_major) ||
+	    (AUTOFS_DEV_IOCTL_VERSION_MINOR < param->ver_minor)) {
+		AUTOFS_WARN("ioctl control interface version mismatch: "
+		     "kernel(%u.%u), user(%u.%u), cmd(%d)",
+		     AUTOFS_DEV_IOCTL_VERSION_MAJOR,
+		     AUTOFS_DEV_IOCTL_VERSION_MINOR,
+		     param->ver_major, param->ver_minor, cmd);
+		err = -EINVAL;
+	}
+
+	/* Fill in the kernel version. */
+	param->ver_major = AUTOFS_DEV_IOCTL_VERSION_MAJOR;
+	param->ver_minor = AUTOFS_DEV_IOCTL_VERSION_MINOR;
+
+	return err;
+}
+
+/*
+ * Copy parameter control struct, including a possible path allocated
+ * at the end of the struct.
+ */
+static struct autofs_dev_ioctl *copy_dev_ioctl(struct autofs_dev_ioctl __user *in)
+{
+	struct autofs_dev_ioctl tmp, *ads;
+
+	if (copy_from_user(&tmp, in, sizeof(tmp)))
+		return ERR_PTR(-EFAULT);
+
+	if (tmp.size < sizeof(tmp))
+		return ERR_PTR(-EINVAL);
+
+	ads = kmalloc(tmp.size, GFP_KERNEL);
+	if (!ads)
+		return ERR_PTR(-ENOMEM);
+
+	if (copy_from_user(ads, in, tmp.size)) {
+		kfree(ads);
+		return ERR_PTR(-EFAULT);
+	}
+
+	return ads;
+}
+
+static inline void free_dev_ioctl(struct autofs_dev_ioctl *param)
+{
+	kfree(param);
+	return;
+}
+
+/*
+ * Check sanity of parameter control fields and if a path is present
+ * check that it is terminated and contains at least one "/".
+ */
+static int validate_dev_ioctl(int cmd, struct autofs_dev_ioctl *param)
+{
+	int err;
+
+	err = check_dev_ioctl_version(cmd, param);
+	if (err) {
+		AUTOFS_WARN("invalid device control module version "
+		     "supplied for cmd(0x%08x)", cmd);
+		goto out;
+	}
+
+	if (param->size > sizeof(*param)) {
+		err = invalid_str(param->path,
+				 (void *) ((size_t) param + param->size));
+		if (err) {
+			AUTOFS_WARN(
+			  "path string terminator missing for cmd(0x%08x)",
+			  cmd);
+			goto out;
+		}
+
+		err = check_name(param->path);
+		if (err) {
+			AUTOFS_WARN("invalid path supplied for cmd(0x%08x)",
+				    cmd);
+			goto out;
+		}
+	}
+
+	err = 0;
+out:
+	return err;
+}
+
+/*
+ * Get the autofs super block info struct from the file opened on
+ * the autofs mount point.
+ */
+static struct autofs_sb_info *autofs_dev_ioctl_sbi(struct file *f)
+{
+	struct autofs_sb_info *sbi = NULL;
+	struct inode *inode;
+
+	if (f) {
+		inode = f->f_dentry->d_inode;
+		sbi = autofs4_sbi(inode->i_sb);
+	}
+	return sbi;
+}
+
+/* Return autofs module protocol version */
+static int autofs_dev_ioctl_protover(struct file *fp,
+				     struct autofs_sb_info *sbi,
+				     struct autofs_dev_ioctl *param)
+{
+	param->protover.version = sbi->version;
+	return 0;
+}
+
+/* Return autofs module protocol sub version */
+static int autofs_dev_ioctl_protosubver(struct file *fp,
+					struct autofs_sb_info *sbi,
+					struct autofs_dev_ioctl *param)
+{
+	param->protosubver.sub_version = sbi->sub_version;
+	return 0;
+}
+
+/*
+ * Walk down the mount stack looking for an autofs mount that
+ * has the requested device number (aka. new_encode_dev(sb->s_dev).
+ */
+static int autofs_dev_ioctl_find_super(struct nameidata *nd, dev_t devno)
+{
+	struct dentry *dentry;
+	struct inode *inode;
+	struct super_block *sb;
+	dev_t s_dev;
+	unsigned int err;
+
+	err = -ENOENT;
+
+	/* Lookup the dentry name at the base of our mount point */
+	dentry = d_lookup(nd->dentry, &nd->last);
+	if (!dentry)
+		goto out;
+
+	dput(nd->dentry);
+	nd->dentry = dentry;
+
+	/* And follow the mount stack looking for our autofs mount */
+	while (follow_down(&nd->mnt, &nd->dentry)) {
+		inode = nd->dentry->d_inode;
+		if (!inode)
+			break;
+
+		sb = inode->i_sb;
+		s_dev = new_encode_dev(sb->s_dev);
+		if (devno == s_dev) {
+			if (sb->s_magic == AUTOFS_SUPER_MAGIC) {
+				err = 0;
+				break;
+			}
+		}
+	}
+out:
+	return err;
+}
+
+/*
+ * Walk down the mount stack looking for an autofs mount that
+ * has the requested mount type (ie. indirect, direct or offset).
+ */
+static int autofs_dev_ioctl_find_sbi_type(struct nameidata *nd, unsigned int type)
+{
+	struct dentry *dentry;
+	struct autofs_info *ino;
+	unsigned int err;
+
+	err = -ENOENT;
+
+	/* Lookup the dentry name at the base of our mount point */
+	dentry = d_lookup(nd->dentry, &nd->last);
+	if (!dentry)
+		goto out;
+
+	dput(nd->dentry);
+	nd->dentry = dentry;
+
+	/* And follow the mount stack looking for our autofs mount */
+	while (follow_down(&nd->mnt, &nd->dentry)) {
+		ino = autofs4_dentry_ino(nd->dentry);
+		if (ino && ino->sbi->type & type) {
+			err = 0;
+			break;
+		}
+	}
+out:
+	return err;
+}
+
+static void autofs_dev_ioctl_fd_install(unsigned int fd, struct file *file)
+{
+	struct files_struct *files = current->files;
+	struct fdtable *fdt;
+
+	spin_lock(&files->file_lock);
+	fdt = files_fdtable(files);
+	BUG_ON(fdt->fd[fd] != NULL);
+	rcu_assign_pointer(fdt->fd[fd], file);
+	FD_SET(fd, fdt->close_on_exec);
+	spin_unlock(&files->file_lock);
+}
+
+
+/*
+ * Open a file descriptor on the autofs mount point corresponding
+ * to the given path and device number (aka. new_encode_dev(sb->s_dev)).
+ */
+static int autofs_dev_ioctl_open_mountpoint(const char *path, dev_t devid)
+{
+	struct file *filp;
+	struct nameidata nd;
+	int err, fd;
+
+	fd = get_unused_fd();
+	if (likely(fd >= 0)) {
+		/* Get nameidata of the parent directory */
+		err = path_lookup(path, LOOKUP_PARENT, &nd);
+		if (err)
+			goto out;
+
+		/*
+		 * Search down, within the parent, looking for an
+		 * autofs super block that has the device number
+		 * corresponding to the autofs fs we want to open.
+		 */
+		err = autofs_dev_ioctl_find_super(&nd, devid);
+		if (err) {
+			path_release(&nd);
+			goto out;
+		}
+
+		filp = dentry_open(nd.dentry, nd.mnt, O_RDONLY);
+		if (IS_ERR(filp)) {
+			err = PTR_ERR(filp);
+			goto out;
+		}
+
+		autofs_dev_ioctl_fd_install(fd, filp);
+	}
+
+	return fd;
+
+out:
+	put_unused_fd(fd);
+	return err;
+}
+
+/* Open a file descriptor on an autofs mount point */
+static int autofs_dev_ioctl_openmount(struct file *fp,
+				      struct autofs_sb_info *sbi,
+				      struct autofs_dev_ioctl *param)
+{
+	const char *path;
+	dev_t devid;
+	int err, fd;
+
+	/* param->path has already been checked */
+	if (!param->openmount.devid)
+		return -EINVAL;
+
+	param->ioctlfd = -1;
+
+	path = param->path;
+	devid = param->openmount.devid;
+
+	err = 0;
+	fd = autofs_dev_ioctl_open_mountpoint(path, devid);
+	if (unlikely(fd < 0)) {
+		err = fd;
+		goto out;
+	}
+
+	param->ioctlfd = fd;
+out:
+	return err;
+}
+
+/* Close file descriptor allocated above (user can also use close(2)). */
+static int autofs_dev_ioctl_closemount(struct file *fp,
+				       struct autofs_sb_info *sbi,
+				       struct autofs_dev_ioctl *param)
+{
+	return sys_close(param->ioctlfd);
+}
+
+/*
+ * Send "ready" status for an existing wait (either a mount or an expire
+ * request).
+ */
+static int autofs_dev_ioctl_ready(struct file *fp,
+				  struct autofs_sb_info *sbi,
+				  struct autofs_dev_ioctl *param)
+{
+	autofs_wqt_t token;
+
+	token = (autofs_wqt_t) param->ready.token;
+	return autofs4_wait_release(sbi, token, 0);
+}
+
+/*
+ * Send "fail" status for an existing wait (either a mount or an expire
+ * request).
+ */
+static int autofs_dev_ioctl_fail(struct file *fp,
+				 struct autofs_sb_info *sbi,
+				 struct autofs_dev_ioctl *param)
+{
+	autofs_wqt_t token;
+	int status;
+
+	token = (autofs_wqt_t) param->fail.token;
+	status = param->fail.status ? param->fail.status : -ENOENT;
+	return autofs4_wait_release(sbi, token, status);
+}
+
+/*
+ * Set the pipe fd for kernel communication to the daemon.
+ *
+ * Normally this is set at mount using an option but if we
+ * are reconnecting to a busy mount then we need to use this
+ * to tell the autofs mount about the new kernel pipe fd. In
+ * order to protect mounts against incorrectly setting the
+ * pipefd we also require that the autofs mount be catatonic.
+ *
+ * This also sets the process group id used to identify the
+ * controlling process (eg. the owning automount(8) daemon).
+ */
+static int autofs_dev_ioctl_setpipefd(struct file *fp,
+				      struct autofs_sb_info *sbi,
+				      struct autofs_dev_ioctl *param)
+{
+	int pipefd;
+	int err = 0;
+
+	if (param->setpipefd.pipefd == -1)
+		return -EINVAL;
+
+	pipefd = param->setpipefd.pipefd;
+
+	mutex_lock(&sbi->wq_mutex);
+	if (!sbi->catatonic) {
+		mutex_unlock(&sbi->wq_mutex);
+		return -EBUSY;
+	} else {
+		struct file *pipe = fget(pipefd);
+		if (!pipe->f_op || !pipe->f_op->write) {
+			err = -EPIPE;
+			fput(pipe);
+			goto out;
+		}
+		sbi->oz_pgrp = process_group(current);
+		sbi->pipefd = pipefd;
+		sbi->pipe = pipe;
+		sbi->catatonic = 0;
+	}
+out:
+	mutex_unlock(&sbi->wq_mutex);
+	return err;
+}
+
+/*
+ * Make the autofs mount point catatonic, no longer responsive to
+ * mount requests. Also closes the kernel pipe file descriptor.
+ */
+static int autofs_dev_ioctl_catatonic(struct file *fp,
+				      struct autofs_sb_info *sbi,
+				      struct autofs_dev_ioctl *param)
+{
+	autofs4_catatonic_mode(sbi);
+	return 0;
+}
+
+/* Set the autofs mount timeout */
+static int autofs_dev_ioctl_timeout(struct file *fp,
+				    struct autofs_sb_info *sbi,
+				    struct autofs_dev_ioctl *param)
+{
+	unsigned long timeout;
+
+	timeout = param->timeout.timeout;
+	param->timeout.timeout = sbi->exp_timeout / HZ;
+	sbi->exp_timeout = timeout * HZ;
+	return 0;
+}
+
+/*
+ * Return the uid and gid of the last request for the mount
+ *
+ * When reconstructing an autofs mount tree with active mounts
+ * we need to re-connect to mounts that may have used the original
+ * process uid and gid (or string variations of them) for mount
+ * lookups within the map entry.
+ */
+static int autofs_dev_ioctl_requester(struct file *fp,
+				      struct autofs_sb_info *sbi,
+				      struct autofs_dev_ioctl *param)
+{
+	struct autofs_info *ino;
+	struct nameidata nd;
+	const char *path;
+	dev_t devid;
+	int err = -ENOENT;
+
+	if (param->size <= sizeof(*param)) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	path = param->path;
+	devid = sbi->sb->s_dev;
+
+	param->requester.uid = param->requester.gid = -1;
+
+	/* Get nameidata of the parent directory */
+	err = path_lookup(path, LOOKUP_PARENT, &nd);
+	if (err)
+		goto out;
+
+	err = autofs_dev_ioctl_find_super(&nd, devid);
+	if (err)
+		goto out_release;
+
+	ino = autofs4_dentry_ino(nd.dentry);
+	if (ino) {
+		err = 0;
+		autofs4_expire_wait(nd.dentry);
+		spin_lock(&sbi->fs_lock);
+		param->requester.uid = ino->uid;
+		param->requester.gid = ino->gid;
+		spin_unlock(&sbi->fs_lock);
+	}
+
+out_release:
+	path_release(&nd);
+out:
+	return err;
+}
+
+/*
+ * Call repeatedly until it returns -EAGAIN, meaning there's nothing
+ * more that can be done.
+ */
+static int autofs_dev_ioctl_expire(struct file *fp,
+				   struct autofs_sb_info *sbi,
+				   struct autofs_dev_ioctl *param)
+{
+	struct vfsmount *mnt;
+	int how;
+
+	how = param->expire.how;
+	mnt = fp->f_vfsmnt;
+
+	return autofs4_do_expire_multi(sbi->sb, mnt, sbi, how);
+}
+
+/* Check if autofs mount point is in use */
+static int autofs_dev_ioctl_askumount(struct file *fp,
+				      struct autofs_sb_info *sbi,
+				      struct autofs_dev_ioctl *param)
+{
+	param->askumount.may_umount = 0;
+	if (may_umount(fp->f_vfsmnt))
+		param->askumount.may_umount = 1;
+	return 0;
+}
+
+/*
+ * Check if the given path is a mountpoint.
+ *
+ * If we are supplied with the file descriptor of an autofs
+ * mount we're looking for a specific mount. In this case
+ * the path is considered a mountpoint if it is itself a
+ * mountpoint or contains a mount, such as a multi-mount
+ * without a root mount. In this case we return 1 if the
+ * path is a mount point and the super magic of the covering
+ * mount if there is one or 0 if it isn't a mountpoint.
+ *
+ * If we aren't supplied with a file descriptor then we
+ * lookup the nameidata of the path and check if it is the
+ * root of a mount. If a type is given we are looking for
+ * a particular autofs mount and if we don't find a match
+ * we return fail. If the located nameidata path is the
+ * root of a mount we return 1 along with the super magic
+ * of the mount or 0 otherwise.
+ *
+ * In both cases the the device number (as returned by
+ * new_encode_dev()) is also returned.
+ */
+static int autofs_dev_ioctl_ismountpoint(struct file *fp,
+					 struct autofs_sb_info *sbi,
+					 struct autofs_dev_ioctl *param)
+{
+	struct nameidata nd;
+	const char *path;
+	unsigned int type;
+	unsigned int devid, magic;
+	int err = -ENOENT;
+
+	if (param->size <= sizeof(*param)) {
+		err = -EINVAL;
+		goto out;
+	}
+
+	path = param->path;
+	type = param->ismountpoint.in.type;
+
+	param->ismountpoint.out.devid = devid = 0;
+	param->ismountpoint.out.magic = magic = 0;
+
+	if (!fp || param->ioctlfd == -1) {
+		if (type == AUTOFS_TYPE_ANY) {
+			struct super_block *sb;
+
+			err = path_lookup(path, LOOKUP_FOLLOW, &nd);
+			if (err)
+				goto out;
+
+			sb = nd.dentry->d_sb;
+			devid = new_encode_dev(sb->s_dev);
+		} else {
+			struct autofs_info *ino;
+
+			err = path_lookup(path, LOOKUP_PARENT, &nd);
+			if (err)
+				goto out;
+
+			err = autofs_dev_ioctl_find_sbi_type(&nd, type);
+			if (err)
+				goto out_release;
+
+			ino = autofs4_dentry_ino(nd.dentry);
+			devid = autofs4_get_dev(ino->sbi);
+		}
+
+		err = 0;
+		if (nd.dentry->d_inode &&
+		    nd.mnt->mnt_root == nd.dentry) {
+			err = 1;
+			magic = nd.dentry->d_inode->i_sb->s_magic;
+		}
+	} else {
+		dev_t dev = autofs4_get_dev(sbi);
+
+		err = path_lookup(path, LOOKUP_PARENT, &nd);
+		if (err)
+			goto out;
+
+		err = autofs_dev_ioctl_find_super(&nd, dev);
+		if (err)
+			goto out_release;
+
+		devid = dev;
+
+		err = have_submounts(nd.dentry);
+
+		if (nd.mnt->mnt_mountpoint != nd.mnt->mnt_root) {
+			if (follow_down(&nd.mnt, &nd.dentry)) {
+				struct inode *inode = nd.dentry->d_inode;
+				magic = inode->i_sb->s_magic;
+			}
+		}
+	}
+
+	param->ismountpoint.out.devid = devid;
+	param->ismountpoint.out.magic = magic;
+
+out_release:
+	path_release(&nd);
+out:
+	return err;
+}
+
+/*
+ * Our range of ioctl numbers isn't 0 based so we need to shift
+ * the array index by _IOC_NR(AUTOFS_CTL_IOC_FIRST) for the table
+ * lookup.
+ */
+#define cmd_idx(cmd)	(cmd - _IOC_NR(AUTOFS_DEV_IOCTL_IOC_FIRST))
+
+static ioctl_fn lookup_dev_ioctl(unsigned int cmd)
+{
+	static struct {
+		int cmd;
+		ioctl_fn fn;
+	} _ioctls[] = {
+		{cmd_idx(AUTOFS_DEV_IOCTL_VERSION_CMD), NULL},
+		{cmd_idx(AUTOFS_DEV_IOCTL_PROTOVER_CMD),
+			 autofs_dev_ioctl_protover},
+		{cmd_idx(AUTOFS_DEV_IOCTL_PROTOSUBVER_CMD),
+			 autofs_dev_ioctl_protosubver},
+		{cmd_idx(AUTOFS_DEV_IOCTL_OPENMOUNT_CMD),
+			 autofs_dev_ioctl_openmount},
+		{cmd_idx(AUTOFS_DEV_IOCTL_CLOSEMOUNT_CMD),
+			 autofs_dev_ioctl_closemount},
+		{cmd_idx(AUTOFS_DEV_IOCTL_READY_CMD),
+			 autofs_dev_ioctl_ready},
+		{cmd_idx(AUTOFS_DEV_IOCTL_FAIL_CMD),
+			 autofs_dev_ioctl_fail},
+		{cmd_idx(AUTOFS_DEV_IOCTL_SETPIPEFD_CMD),
+			 autofs_dev_ioctl_setpipefd},
+		{cmd_idx(AUTOFS_DEV_IOCTL_CATATONIC_CMD),
+			 autofs_dev_ioctl_catatonic},
+		{cmd_idx(AUTOFS_DEV_IOCTL_TIMEOUT_CMD),
+			 autofs_dev_ioctl_timeout},
+		{cmd_idx(AUTOFS_DEV_IOCTL_REQUESTER_CMD),
+			 autofs_dev_ioctl_requester},
+		{cmd_idx(AUTOFS_DEV_IOCTL_EXPIRE_CMD),
+			 autofs_dev_ioctl_expire},
+		{cmd_idx(AUTOFS_DEV_IOCTL_ASKUMOUNT_CMD),
+			 autofs_dev_ioctl_askumount},
+		{cmd_idx(AUTOFS_DEV_IOCTL_ISMOUNTPOINT_CMD),
+			 autofs_dev_ioctl_ismountpoint}
+	};
+	unsigned int idx = cmd_idx(cmd);
+
+	return (idx >= ARRAY_SIZE(_ioctls)) ? NULL : _ioctls[idx].fn;
+}
+
+/* ioctl dispatcher */
+static int _autofs_dev_ioctl(unsigned int command, struct autofs_dev_ioctl __user *user)
+{
+	struct autofs_dev_ioctl *param;
+	struct file *fp;
+	struct autofs_sb_info *sbi;
+	unsigned int cmd_first, cmd;
+	ioctl_fn fn = NULL;
+	int err = 0;
+
+	/* only root can play with this */
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	cmd_first = _IOC_NR(AUTOFS_DEV_IOCTL_IOC_FIRST);
+	cmd = _IOC_NR(command);
+
+	if (_IOC_TYPE(command) != _IOC_TYPE(AUTOFS_DEV_IOCTL_IOC_FIRST) ||
+	    cmd - cmd_first >= AUTOFS_DEV_IOCTL_IOC_COUNT) {
+		return -ENOTTY;
+	}
+
+	/* Copy the parameters into kernel space. */
+	param = copy_dev_ioctl(user);
+	if (IS_ERR(param))
+		return PTR_ERR(param);
+
+	err = validate_dev_ioctl(command, param);
+	if (err)
+		goto out;
+
+	/* The validate routine above always sets the version */
+	if (cmd == AUTOFS_DEV_IOCTL_VERSION_CMD)
+		goto done;
+
+	fn = lookup_dev_ioctl(cmd);
+	if (!fn) {
+		AUTOFS_WARN("unknown command 0x%08x", command);
+		return -ENOTTY;
+	}
+
+	fp = NULL;
+	sbi = NULL;
+
+	/*
+	 * For obvious reasons the openmount can't have a file
+	 * descriptor yet. We don't take a reference to the
+	 * file during close to allow for immediate release.
+	 */
+	if (cmd != AUTOFS_DEV_IOCTL_OPENMOUNT_CMD &&
+	    cmd != AUTOFS_DEV_IOCTL_CLOSEMOUNT_CMD) {
+		fp = fget(param->ioctlfd);
+		if (!fp) {
+			if (cmd == AUTOFS_DEV_IOCTL_ISMOUNTPOINT_CMD)
+				goto cont;
+			err = -EBADF;
+			goto out;
+		}
+
+		if (!fp->f_op) {
+			err = -ENOTTY;
+			fput(fp);
+			goto out;
+		}
+
+		sbi = autofs_dev_ioctl_sbi(fp);
+		if (!sbi || sbi->magic != AUTOFS_SBI_MAGIC) {
+			err = -EINVAL;
+			fput(fp);
+			goto out;
+		}
+
+		/*
+		 * Admin needs to be able to set the mount catatonic in
+		 * order to be able to perform the re-open.
+		 */
+		if (!autofs4_oz_mode(sbi) &&
+		    cmd != AUTOFS_DEV_IOCTL_CATATONIC_CMD) {
+			err = -EACCES;
+			fput(fp);
+			goto out;
+		}
+	}
+cont:
+	err = fn(fp, sbi, param);
+
+	if (fp)
+		fput(fp);
+done:
+	if (err >= 0 && copy_to_user(user, param, AUTOFS_DEV_IOCTL_SIZE))
+		err = -EFAULT;
+out:
+	free_dev_ioctl(param);
+	return err;
+}
+
+static long autofs_dev_ioctl(struct file *file, uint command, ulong u)
+{
+	int err;
+	err = _autofs_dev_ioctl(command, (struct autofs_dev_ioctl __user *) u);
+	return (long) err;
+}
+
+#ifdef CONFIG_COMPAT
+static long autofs_dev_ioctl_compat(struct file *file, uint command, ulong u)
+{
+	return (long) autofs_dev_ioctl(file, command, (ulong) compat_ptr(u));
+}
+#else
+#define autofs_dev_ioctl_compat NULL
+#endif
+
+static const struct file_operations _dev_ioctl_fops = {
+	.unlocked_ioctl	 = autofs_dev_ioctl,
+	.compat_ioctl = autofs_dev_ioctl_compat,
+	.owner	 = THIS_MODULE,
+};
+
+static struct miscdevice _autofs_dev_ioctl_misc = {
+	.minor 		= MISC_DYNAMIC_MINOR,
+	.name  		= AUTOFS_DEVICE_NAME,
+	.fops  		= &_dev_ioctl_fops
+};
+
+/* Register/deregister misc character device */
+int autofs_dev_ioctl_init(void)
+{
+	int r;
+
+	r = misc_register(&_autofs_dev_ioctl_misc);
+	if (r) {
+		AUTOFS_ERROR("misc_register failed for control device");
+		return r;
+	}
+
+	return 0;
+}
+
+void autofs_dev_ioctl_exit(void)
+{
+	misc_deregister(&_autofs_dev_ioctl_misc);
+	return;
+}
+
diff --git a/fs/autofs4/expire.c b/fs/autofs4/expire.c
index e96e81b..69d7451 100644
--- a/fs/autofs4/expire.c
+++ b/fs/autofs4/expire.c
@@ -255,10 +255,10 @@ cont:
 }
 
 /* Check if we can expire a direct mount (possibly a tree) */
-static struct dentry *autofs4_expire_direct(struct super_block *sb,
-					    struct vfsmount *mnt,
-					    struct autofs_sb_info *sbi,
-					    int how)
+struct dentry *autofs4_expire_direct(struct super_block *sb,
+				     struct vfsmount *mnt,
+				     struct autofs_sb_info *sbi,
+				     int how)
 {
 	unsigned long timeout;
 	struct dentry *root = dget(sb->s_root);
@@ -294,10 +294,10 @@ static struct dentry *autofs4_expire_direct(struct super_block *sb,
  *  - it is unused by any user process
  *  - it has been unused for exp_timeout time
  */
-static struct dentry *autofs4_expire_indirect(struct super_block *sb,
-					      struct vfsmount *mnt,
-					      struct autofs_sb_info *sbi,
-					      int how)
+struct dentry *autofs4_expire_indirect(struct super_block *sb,
+				       struct vfsmount *mnt,
+				       struct autofs_sb_info *sbi,
+				       int how)
 {
 	unsigned long timeout;
 	struct dentry *root = sb->s_root;
@@ -478,22 +478,16 @@ int autofs4_expire_run(struct super_block *sb,
 	return ret;
 }
 
-/* Call repeatedly until it returns -EAGAIN, meaning there's nothing
-   more to be done */
-int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt,
-			struct autofs_sb_info *sbi, int __user *arg)
+int autofs4_do_expire_multi(struct super_block *sb, struct vfsmount *mnt,
+			    struct autofs_sb_info *sbi, int when)
 {
 	struct dentry *dentry;
 	int ret = -EAGAIN;
-	int do_now = 0;
-
-	if (arg && get_user(do_now, arg))
-		return -EFAULT;
 
 	if (sbi->type & AUTOFS_TYPE_TRIGGER)
-		dentry = autofs4_expire_direct(sb, mnt, sbi, do_now);
+		dentry = autofs4_expire_direct(sb, mnt, sbi, when);
 	else
-		dentry = autofs4_expire_indirect(sb, mnt, sbi, do_now);
+		dentry = autofs4_expire_indirect(sb, mnt, sbi, when);
 
 	if (dentry) {
 		struct autofs_info *ino = autofs4_dentry_ino(dentry);
@@ -516,3 +510,16 @@ int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt,
 	return ret;
 }
 
+/* Call repeatedly until it returns -EAGAIN, meaning there's nothing
+   more to be done */
+int autofs4_expire_multi(struct super_block *sb, struct vfsmount *mnt,
+			struct autofs_sb_info *sbi, int __user *arg)
+{
+	int do_now = 0;
+
+	if (arg && get_user(do_now, arg))
+		return -EFAULT;
+
+	return autofs4_do_expire_multi(sb, mnt, sbi, do_now);
+}
+
diff --git a/fs/autofs4/init.c b/fs/autofs4/init.c
index 723a1c5..9722e4b 100644
--- a/fs/autofs4/init.c
+++ b/fs/autofs4/init.c
@@ -29,11 +29,20 @@ static struct file_system_type autofs_fs_type = {
 
 static int __init init_autofs4_fs(void)
 {
-	return register_filesystem(&autofs_fs_type);
+	int err;
+
+	err = register_filesystem(&autofs_fs_type);
+	if (err)
+		return err;
+
+	autofs_dev_ioctl_init();
+
+	return err;
 }
 
 static void __exit exit_autofs4_fs(void)
 {
+	autofs_dev_ioctl_exit();
 	unregister_filesystem(&autofs_fs_type);
 }
 
diff --git a/include/linux/auto_dev-ioctl.h b/include/linux/auto_dev-ioctl.h
new file mode 100644
index 0000000..91a7739
--- /dev/null
+++ b/include/linux/auto_dev-ioctl.h
@@ -0,0 +1,224 @@
+/*
+ * Copyright 2008 Red Hat, Inc. All rights reserved.
+ * Copyright 2008 Ian Kent <raven@themaw.net>
+ *
+ * This file is part of the Linux kernel and is made available under
+ * the terms of the GNU General Public License, version 2, or at your
+ * option, any later version, incorporated herein by reference.
+ */
+
+#ifndef _LINUX_AUTO_DEV_IOCTL_H
+#define _LINUX_AUTO_DEV_IOCTL_H
+
+#include <linux/string.h>
+#include <linux/types.h>
+
+#define AUTOFS_DEVICE_NAME		"autofs"
+
+#define AUTOFS_DEV_IOCTL_VERSION_MAJOR	1
+#define AUTOFS_DEV_IOCTL_VERSION_MINOR	0
+
+#define AUTOFS_DEVID_LEN		16
+
+#define AUTOFS_DEV_IOCTL_SIZE		sizeof(struct autofs_dev_ioctl)
+
+/*
+ * An ioctl interface for autofs mount point control.
+ */
+
+struct args_protover {
+	__u32	version;
+};
+
+struct args_protosubver {
+	__u32	sub_version;
+};
+
+struct args_openmount {
+	__u32	devid;
+};
+
+struct args_ready {
+	__u32	token;
+};
+
+struct args_fail {
+	__u32	token;
+	__s32	status;
+};
+
+struct args_setpipefd {
+	__s32	pipefd;
+};
+
+struct args_timeout {
+	__u64	timeout;
+};
+
+struct args_requester {
+	__u32	uid;
+	__u32	gid;
+};
+
+struct args_expire {
+	__u32	how;
+};
+
+struct args_askumount {
+	__u32	may_umount;
+};
+
+struct args_ismountpoint {
+	union {
+		struct args_in {
+			__u32	type;
+		} in;
+		struct args_out {
+			__u32	devid;
+			__u32	magic;
+		} out;
+	};
+};
+
+/*
+ * All the ioctls use this structure.
+ * When sending a path size must account for the total length
+ * of the chunk of memory otherwise is is the size of the
+ * structure.
+ */
+
+struct autofs_dev_ioctl {
+	__u32 ver_major;
+	__u32 ver_minor;
+	__u32 size;		/* total size of data passed in
+				 * including this struct */
+	__s32 ioctlfd;		/* automount command fd */
+
+	/* Command parameters */
+
+	union {
+		struct args_protover		protover;
+		struct args_protosubver		protosubver;
+		struct args_openmount		openmount;
+		struct args_ready		ready;
+		struct args_fail		fail;
+		struct args_setpipefd		setpipefd;
+		struct args_timeout		timeout;
+		struct args_requester		requester;
+		struct args_expire		expire;
+		struct args_askumount		askumount;
+		struct args_ismountpoint	ismountpoint;
+	};
+
+	char path[0];
+};
+
+static inline void init_autofs_dev_ioctl(struct autofs_dev_ioctl *in)
+{
+	memset(in, 0, sizeof(struct autofs_dev_ioctl));
+	in->ver_major = AUTOFS_DEV_IOCTL_VERSION_MAJOR;
+	in->ver_minor = AUTOFS_DEV_IOCTL_VERSION_MINOR;
+	in->size = sizeof(struct autofs_dev_ioctl);
+	in->ioctlfd = -1;
+	return;
+}
+
+/*
+ * If you change this make sure you make the corresponding change
+ * to autofs-dev-ioctl.c:lookup_ioctl()
+ */
+enum {
+	/* Get various version info */
+	AUTOFS_DEV_IOCTL_VERSION_CMD = 0x71,
+	AUTOFS_DEV_IOCTL_PROTOVER_CMD,
+	AUTOFS_DEV_IOCTL_PROTOSUBVER_CMD,
+
+	/* Open mount ioctl fd */
+	AUTOFS_DEV_IOCTL_OPENMOUNT_CMD,
+
+	/* Close mount ioctl fd */
+	AUTOFS_DEV_IOCTL_CLOSEMOUNT_CMD,
+
+	/* Mount/expire status returns */
+	AUTOFS_DEV_IOCTL_READY_CMD,
+	AUTOFS_DEV_IOCTL_FAIL_CMD,
+
+	/* Activate/deactivate autofs mount */
+	AUTOFS_DEV_IOCTL_SETPIPEFD_CMD,
+	AUTOFS_DEV_IOCTL_CATATONIC_CMD,
+
+	/* Expiry timeout */
+	AUTOFS_DEV_IOCTL_TIMEOUT_CMD,
+
+	/* Get mount last requesting uid and gid */
+	AUTOFS_DEV_IOCTL_REQUESTER_CMD,
+
+	/* Check for eligible expire candidates */
+	AUTOFS_DEV_IOCTL_EXPIRE_CMD,
+
+	/* Request busy status */
+	AUTOFS_DEV_IOCTL_ASKUMOUNT_CMD,
+
+	/* Check if path is a mountpoint */
+	AUTOFS_DEV_IOCTL_ISMOUNTPOINT_CMD,
+};
+
+#define AUTOFS_IOCTL 0x93
+
+#define AUTOFS_DEV_IOCTL_VERSION \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_VERSION_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_PROTOVER \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_PROTOVER_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_PROTOSUBVER \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_PROTOSUBVER_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_OPENMOUNT \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_OPENMOUNT_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_CLOSEMOUNT \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_CLOSEMOUNT_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_READY \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_READY_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_FAIL \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_FAIL_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_SETPIPEFD \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_SETPIPEFD_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_CATATONIC \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_CATATONIC_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_TIMEOUT \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_TIMEOUT_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_REQUESTER \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_REQUESTER_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_EXPIRE \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_EXPIRE_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_ASKUMOUNT \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_ASKUMOUNT_CMD, struct autofs_dev_ioctl)
+
+#define AUTOFS_DEV_IOCTL_ISMOUNTPOINT \
+	_IOWR(AUTOFS_IOCTL, \
+	      AUTOFS_DEV_IOCTL_ISMOUNTPOINT_CMD, struct autofs_dev_ioctl)
+
+#endif	/* _LINUX_AUTO_DEV_IOCTL_H */
diff --git a/include/linux/auto_fs4.h b/include/linux/auto_fs4.h
index 48dd618..d0d0d12 100644
--- a/include/linux/auto_fs4.h
+++ b/include/linux/auto_fs4.h
@@ -23,7 +23,7 @@
 #define AUTOFS_MIN_PROTO_VERSION	3
 #define AUTOFS_MAX_PROTO_VERSION	5
 
-#define AUTOFS_PROTO_SUBVERSION		0
+#define AUTOFS_PROTO_SUBVERSION		1
 
 /* Mask for expire behaviour */
 #define AUTOFS_EXP_IMMEDIATE		1