New kind of open files - "location only".

New flag for open(2) - O_PATH.  Semantics:
	* pathname is resolved, but the file itself is _NOT_ opened
as far as filesystem is concerned.
	* almost all operations on the resulting descriptors shall
fail with -EBADF.  Exceptions are:
	1) operations on descriptors themselves (i.e.
		close(), dup(), dup2(), dup3(), fcntl(fd, F_DUPFD),
		fcntl(fd, F_DUPFD_CLOEXEC, ...), fcntl(fd, F_GETFD),
		fcntl(fd, F_SETFD, ...))
	2) fcntl(fd, F_GETFL), for a common non-destructive way to
		check if descriptor is open
	3) "dfd" arguments of ...at(2) syscalls, i.e. the starting
		points of pathname resolution
	* closing such descriptor does *NOT* affect dnotify or
posix locks.
	* permissions are checked as usual along the way to file;
no permission checks are applied to the file itself.  Of course,
giving such thing to syscall will result in permission checks (at
the moment it means checking that starting point of ....at() is
a directory and caller has exec permissions on it).

fget() and fget_light() return NULL on such descriptors; use of
fget_raw() and fget_raw_light() is needed to get them.  That protects
existing code from dealing with those things.

There are two things still missing (they come in the next commits):
one is handling of symlinks (right now we refuse to open them that
way; see the next commit for semantics related to those) and another
is descriptor passing via SCM_RIGHTS datagrams.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>

4 files changed, 110 insertions, 17 deletions

diff --git a/fs/fcntl.c b/fs/fcntl.c
index cb1026181bd..6c82e5bac03 100644
--- a/fs/fcntl.c
+++ b/fs/fcntl.c
@@ -131,7 +131,7 @@ SYSCALL_DEFINE2(dup2, unsigned int, oldfd, unsigned int, newfd)
 SYSCALL_DEFINE1(dup, unsigned int, fildes)
 {
 	int ret = -EBADF;
-	struct file *file = fget(fildes);
+	struct file *file = fget_raw(fildes);
 
 	if (file) {
 		ret = get_unused_fd();
@@ -426,15 +426,35 @@ static long do_fcntl(int fd, unsigned int cmd, unsigned long arg,
 	return err;
 }
 
+static int check_fcntl_cmd(unsigned cmd)
+{
+	switch (cmd) {
+	case F_DUPFD:
+	case F_DUPFD_CLOEXEC:
+	case F_GETFD:
+	case F_SETFD:
+	case F_GETFL:
+		return 1;
+	}
+	return 0;
+}
+
 SYSCALL_DEFINE3(fcntl, unsigned int, fd, unsigned int, cmd, unsigned long, arg)
 {	
 	struct file *filp;
 	long err = -EBADF;
 
-	filp = fget(fd);
+	filp = fget_raw(fd);
 	if (!filp)
 		goto out;
 
+	if (unlikely(filp->f_mode & FMODE_PATH)) {
+		if (!check_fcntl_cmd(cmd)) {
+			fput(filp);
+			goto out;
+		}
+	}
+
 	err = security_file_fcntl(filp, cmd, arg);
 	if (err) {
 		fput(filp);
@@ -456,10 +476,17 @@ SYSCALL_DEFINE3(fcntl64, unsigned int, fd, unsigned int, cmd,
 	long err;
 
 	err = -EBADF;
-	filp = fget(fd);
+	filp = fget_raw(fd);
 	if (!filp)
 		goto out;
 
+	if (unlikely(filp->f_mode & FMODE_PATH)) {
+		if (!check_fcntl_cmd(cmd)) {
+			fput(filp);
+			goto out;
+		}
+	}
+
 	err = security_file_fcntl(filp, cmd, arg);
 	if (err) {
 		fput(filp);
@@ -808,14 +835,14 @@ static int __init fcntl_init(void)
 	 * Exceptions: O_NONBLOCK is a two bit define on parisc; O_NDELAY
 	 * is defined as O_NONBLOCK on some platforms and not on others.
 	 */
-	BUILD_BUG_ON(18 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
+	BUILD_BUG_ON(19 - 1 /* for O_RDONLY being 0 */ != HWEIGHT32(
 		O_RDONLY	| O_WRONLY	| O_RDWR	|
 		O_CREAT		| O_EXCL	| O_NOCTTY	|
 		O_TRUNC		| O_APPEND	| /* O_NONBLOCK	| */
 		__O_SYNC	| O_DSYNC	| FASYNC	|
 		O_DIRECT	| O_LARGEFILE	| O_DIRECTORY	|
 		O_NOFOLLOW	| O_NOATIME	| O_CLOEXEC	|
-		__FMODE_EXEC
+		__FMODE_EXEC	| O_PATH
 		));
 
 	fasync_cache = kmem_cache_create("fasync_cache",
diff --git a/fs/file_table.c b/fs/file_table.c
index eb36b6b17e2..3c16e1ca163 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -276,11 +276,10 @@ struct file *fget(unsigned int fd)
 	rcu_read_lock();
 	file = fcheck_files(files, fd);
 	if (file) {
-		if (!atomic_long_inc_not_zero(&file->f_count)) {
-			/* File object ref couldn't be taken */
-			rcu_read_unlock();
-			return NULL;
-		}
+		/* File object ref couldn't be taken */
+		if (file->f_mode & FMODE_PATH ||
+		    !atomic_long_inc_not_zero(&file->f_count))
+			file = NULL;
 	}
 	rcu_read_unlock();
 
@@ -289,6 +288,23 @@ struct file *fget(unsigned int fd)
 
 EXPORT_SYMBOL(fget);
 
+struct file *fget_raw(unsigned int fd)
+{
+	struct file *file;
+	struct files_struct *files = current->files;
+
+	rcu_read_lock();
+	file = fcheck_files(files, fd);
+	if (file) {
+		/* File object ref couldn't be taken */
+		if (!atomic_long_inc_not_zero(&file->f_count))
+			file = NULL;
+	}
+	rcu_read_unlock();
+
+	return file;
+}
+
 /*
  * Lightweight file lookup - no refcnt increment if fd table isn't shared.
  *
@@ -313,6 +329,33 @@ struct file *fget_light(unsigned int fd, int *fput_needed)
 	*fput_needed = 0;
 	if (atomic_read(&files->count) == 1) {
 		file = fcheck_files(files, fd);
+		if (file && (file->f_mode & FMODE_PATH))
+			file = NULL;
+	} else {
+		rcu_read_lock();
+		file = fcheck_files(files, fd);
+		if (file) {
+			if (!(file->f_mode & FMODE_PATH) &&
+			    atomic_long_inc_not_zero(&file->f_count))
+				*fput_needed = 1;
+			else
+				/* Didn't get the reference, someone's freed */
+				file = NULL;
+		}
+		rcu_read_unlock();
+	}
+
+	return file;
+}
+
+struct file *fget_raw_light(unsigned int fd, int *fput_needed)
+{
+	struct file *file;
+	struct files_struct *files = current->files;
+
+	*fput_needed = 0;
+	if (atomic_read(&files->count) == 1) {
+		file = fcheck_files(files, fd);
 	} else {
 		rcu_read_lock();
 		file = fcheck_files(files, fd);
diff --git a/fs/namei.c b/fs/namei.c
index 33be51a2ddb..e1d9f90d977 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -1544,7 +1544,7 @@ static int path_init(int dfd, const char *name, unsigned int flags,
 	} else {
 		struct dentry *dentry;
 
-		file = fget_light(dfd, &fput_needed);
+		file = fget_raw_light(dfd, &fput_needed);
 		retval = -EBADF;
 		if (!file)
 			goto out_fail;
diff --git a/fs/open.c b/fs/open.c
index 48afc5c139d..14a51de01f5 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -669,11 +669,16 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt,
 					int (*open)(struct inode *, struct file *),
 					const struct cred *cred)
 {
+	static const struct file_operations empty_fops = {};
 	struct inode *inode;
 	int error;
 
 	f->f_mode = OPEN_FMODE(f->f_flags) | FMODE_LSEEK |
 				FMODE_PREAD | FMODE_PWRITE;
+
+	if (unlikely(f->f_flags & O_PATH))
+		f->f_mode = FMODE_PATH;
+
 	inode = dentry->d_inode;
 	if (f->f_mode & FMODE_WRITE) {
 		error = __get_file_write_access(inode, mnt);
@@ -687,9 +692,15 @@ static struct file *__dentry_open(struct dentry *dentry, struct vfsmount *mnt,
 	f->f_path.dentry = dentry;
 	f->f_path.mnt = mnt;
 	f->f_pos = 0;
-	f->f_op = fops_get(inode->i_fop);
 	file_sb_list_add(f, inode->i_sb);
 
+	if (unlikely(f->f_mode & FMODE_PATH)) {
+		f->f_op = &empty_fops;
+		return f;
+	}
+
+	f->f_op = fops_get(inode->i_fop);
+
 	error = security_dentry_open(f, cred);
 	if (error)
 		goto cleanup_all;
@@ -911,9 +922,18 @@ static inline int build_open_flags(int flags, int mode, struct open_flags *op)
 	if (flags & __O_SYNC)
 		flags |= O_DSYNC;
 
-	op->open_flag = flags;
+	/*
+	 * If we have O_PATH in the open flag. Then we
+	 * cannot have anything other than the below set of flags
+	 */
+	if (flags & O_PATH) {
+		flags &= O_DIRECTORY | O_NOFOLLOW | O_PATH;
+		acc_mode = 0;
+	} else {
+		acc_mode = MAY_OPEN | ACC_MODE(flags);
+	}
 
-	acc_mode = MAY_OPEN | ACC_MODE(flags);
+	op->open_flag = flags;
 
 	/* O_TRUNC implies we need access checks for write permissions */
 	if (flags & O_TRUNC)
@@ -926,7 +946,8 @@ static inline int build_open_flags(int flags, int mode, struct open_flags *op)
 
 	op->acc_mode = acc_mode;
 
-	op->intent = LOOKUP_OPEN;
+	op->intent = flags & O_PATH ? 0 : LOOKUP_OPEN;
+
 	if (flags & O_CREAT) {
 		op->intent |= LOOKUP_CREATE;
 		if (flags & O_EXCL)
@@ -1053,8 +1074,10 @@ int filp_close(struct file *filp, fl_owner_t id)
 	if (filp->f_op && filp->f_op->flush)
 		retval = filp->f_op->flush(filp, id);
 
-	dnotify_flush(filp, id);
-	locks_remove_posix(filp, id);
+	if (likely(!(filp->f_mode & FMODE_PATH))) {
+		dnotify_flush(filp, id);
+		locks_remove_posix(filp, id);
+	}
 	fput(filp);
 	return retval;
 }