get mount write in __dentry_open()

diff -puN fs/open.c~get-write-in-__dentry_open fs/open.c
@@ -778,9 +778,15 @@ static struct file *__dentry_open(struct
 				FMODE_PREAD | FMODE_PWRITE;
 	inode = dentry->d_inode;
 	if (f->f_mode & FMODE_WRITE) {
-		error = get_write_access(inode);
+		error = mnt_want_write(mnt);
 		if (error)
 			goto cleanup_file;
+
+		error = get_write_access(inode);
+		if (error) {
+			mnt_drop_write(mnt);
+			goto cleanup_file;
+		}
 	}
 
 	f->f_mapping = inode->i_mapping;

@@ -820,8 +826,10 @@ static struct file *__dentry_open(struct
 
 cleanup_all:
 	fops_put(f->f_op);
-	if (f->f_mode & FMODE_WRITE)
+	if (f->f_mode & FMODE_WRITE) {
 		put_write_access(inode);
+		mnt_drop_write(mnt);
+	}
 	file_kill(f);
 	f->f_path.dentry = NULL;
 	f->f_path.mnt = NULL;

diff -puN fs/namei.c~move-mnt_want_write-into-open_namei_create fs/namei.c
@@ -1686,6 +1686,19 @@ static int open_namei_create(struct name
 	int error;
 	struct dentry *dir = nd->dentry;
 
+	/*
+	 * This ensures that the mnt stays writable
+	 * over the vfs_create() call to may_open(),
+	 * which takes a more persistent
+	 * mnt_want_write().
+	 */
+	error = mnt_want_write(nd->mnt);
+	if (error) {
+		mutex_unlock(&dir->d_inode->i_mutex);
+		dput(nd->dentry);
+		return error;
+	}
+
 	if (!IS_POSIXACL(dir->d_inode))
 		mode &= ~current->fs->umask;
 	error = vfs_create(dir->d_inode, path->dentry, mode, nd);

@@ -1693,9 +1706,16 @@ static int open_namei_create(struct name
 	dput(nd->dentry);
 	nd->dentry = path->dentry;
 	if (error)
-		return error;
+		goto out;
 	/* Don't check for write permission, don't truncate */
-	return may_open(nd, 0, flag & ~O_TRUNC);
+	error = may_open(nd, 0, flag & ~O_TRUNC);
+out:
+	/*
+	 * if needed, may_open() has taken a write
+	 * on the mnt for us, so we can release ours.
+	 */
+	mnt_drop_write(nd->mnt);
+	return error;
 }
 
 /*

@@ -1778,11 +1798,7 @@ do_last:
 
 	/* Negative dentry, just create the file */
 	if (!path.dentry->d_inode) {
-		error = mnt_want_write(nd->mnt);
-		if (error)
-			goto exit_mutex_unlock;
 		error = open_namei_create(nd, &path, flag, mode);
-		mnt_drop_write(nd->mnt);
 		if (error)
 			goto exit;
 		return 0;

diff -puN fs/namei.c~move-mnt_want_write-out-of-may_open fs/namei.c
@@ -1621,14 +1621,6 @@ int may_open(struct nameidata *nd, int a
 			return -EACCES;
 
 		flag &= ~O_TRUNC;
-	} else if (flag & FMODE_WRITE) {
-		/*
-		 * effectively: !special_file()
-		 * balanced by __fput()
-		 */
-		error = mnt_want_write(nd->mnt);
-		if (error)
-			return error;
 	}
 
 	error = vfs_permission(nd, acc_mode);

@@ -1687,10 +1679,8 @@ static int open_namei_create(struct name
 	struct dentry *dir = nd->dentry;
 
 	/*
-	 * This ensures that the mnt stays writable
-	 * over the vfs_create() call to may_open(),
-	 * which takes a more persistent
-	 * mnt_want_write().
+	 * This mnt_want_write() is potentially persistent,
+	 * and balanced in __fput()
 	 */
 	error = mnt_want_write(nd->mnt);
 	if (error) {

@@ -1710,14 +1700,18 @@ static int open_namei_create(struct name
 	/* Don't check for write permission, don't truncate */
 	error = may_open(nd, 0, flag & ~O_TRUNC);
 out:
-	/*
-	 * if needed, may_open() has taken a write
-	 * on the mnt for us, so we can release ours.
-	 */
-	mnt_drop_write(nd->mnt);
+	if (error)
+		mnt_drop_write(nd->mnt);
 	return error;
 }
 
+static inline int write_may_occur_to_file(struct inode *inode, int acc_mode)
+{
+	if (special_file(inode->i_mode))
+		return 0;
+	return (acc_mode & FMODE_WRITE);
+}
+
 /*
  *	open_namei()
  *

@@ -1831,9 +1825,20 @@ do_last:
 	if (path.dentry->d_inode && S_ISDIR(path.dentry->d_inode->i_mode))
 		goto exit;
 ok:
-	error = may_open(nd, acc_mode, flag);
+	/*
+	 * This mnt_want_write() is potentially persistent,
+	 * and balanced in __fput()
+	 */
+	if (write_may_occur_to_file(nd->dentry->d_inode, acc_mode))
+		error = mnt_want_write(nd->mnt);
 	if (error)
 		goto exit;
+	error = may_open(nd, acc_mode, flag);
+	if (error) {
+		if (write_may_occur_to_file(nd->dentry->d_inode, acc_mode))
+			mnt_drop_write(nd->mnt);
+		goto exit;
+	}
 	return 0;
 
 exit_mutex_unlock:

diff -puN fs/nfsctl.c~move-mnt_want_write-out-of-may_open fs/nfsctl.c
@@ -22,6 +22,7 @@
 
 static struct file *do_open(char *name, int flags)
 {
+	struct file *filp;
 	struct nameidata nd;
 	struct vfsmount *mnt;
 	int error;

@@ -35,14 +36,29 @@ static struct file *do_open(char *name, 
 	if (error)
 		return ERR_PTR(error);
 
+	/*
+	 * This is balanced when you __fput() the 'struct file'
+	 * returned by dentry_open() below.  If the may_open()
+	 * ever stops containing FMODE_WRITE, this needs to
+	 * be changed.
+	 */
+	error = mnt_want_write(mnt);
+	if (error)
+		return ERR_PTR(error);
 	if (flags == O_RDWR)
 		error = may_open(&nd,MAY_READ|MAY_WRITE,FMODE_READ|FMODE_WRITE);
 	else
 		error = may_open(&nd, MAY_WRITE, FMODE_WRITE);
 
-	if (!error)
-		return dentry_open(nd.dentry, nd.mnt, flags);
+	if (error)
+		goto out;
 
+	filp = dentry_open(nd.dentry, nd.mnt, flags);
+	if (!IS_ERR(filp))
+		return filp;
+	error = PTR_ERR(error);
+out:
+	mnt_drop_write(mnt);
 	path_release(&nd);
 	return ERR_PTR(error);
 }

diff -puN fs/namei.c~do-namei_flags-calculation-inside-open_namei fs/namei.c
@@ -1672,7 +1672,12 @@ int may_open(struct nameidata *nd, int a
 	return 0;
 }
 
-static int open_namei_create(struct nameidata *nd, struct path *path,
+/*
+ * Be careful about ever adding any more callers of this
+ * function.  Its flags must be in the namei format, not
+ * what get passed to sys_open().
+ */
+static int __open_namei_create(struct nameidata *nd, struct path *path,
 				int flag, int mode)
 {
 	int error;

@@ -1713,26 +1718,46 @@ static inline int write_may_occur_to_fil
 }
 
 /*
+ * Note that while the flag value (low two bits) for sys_open means:
+ *	00 - read-only
+ *	01 - write-only
+ *	10 - read-write
+ *	11 - special
+ * it is changed into
+ *	00 - no permissions needed
+ *	01 - read-permission
+ *	10 - write-permission
+ *	11 - read-write
+ * for the internal routines (ie open_namei()/follow_link() etc)
+ * This is more logical, and also allows the 00 "no perm needed"
+ * to be used for symlinks (where the permissions are checked
+ * later).
+ *
+*/
+static inline int sys_open_flags_to_namei_flags(int flag)
+{
+	if ((flag+1) & O_ACCMODE)
+		flag++;
+	return flag;
+}
+
+/*
  *	open_namei()
  *
  * namei for open - this is in fact almost the whole open-routine.
  *
  * Note that the low bits of "flag" aren't the same as in the open
- * system call - they are 00 - no permissions needed
- *			  01 - read permission needed
- *			  10 - write permission needed
- *			  11 - read/write permissions needed
- * which is a lot more logical, and also allows the "no perm" needed
- * for symlinks (where the permissions are checked later).
+ * system call.  See sys_open_flags_to_namei_flags().
  * SMP-safe
  */
-int open_namei(int dfd, const char *pathname, int flag,
+int open_namei(int dfd, const char *pathname, int sys_open_flag,
 		int mode, struct nameidata *nd)
 {
 	int acc_mode, error;
 	struct path path;
 	struct dentry *dir;
 	int count = 0;
+	int flag = sys_open_flags_to_namei_flags(sys_open_flag);
 
 	acc_mode = ACC_MODE(flag);
 

@@ -1792,7 +1817,7 @@ do_last:
 
 	/* Negative dentry, just create the file */
 	if (!path.dentry->d_inode) {
-		error = open_namei_create(nd, &path, flag, mode);
+		error = __open_namei_create(nd, &path, flag, mode);
 		if (error)
 			goto exit;
 		return 0;

diff -puN fs/open.c~do-namei_flags-calculation-inside-open_namei fs/open.c
@@ -840,31 +840,13 @@ cleanup_file:
 	return ERR_PTR(error);
 }
 
-/*
- * Note that while the flag value (low two bits) for sys_open means:
- *	00 - read-only
- *	01 - write-only
- *	10 - read-write
- *	11 - special
- * it is changed into
- *	00 - no permissions needed
- *	01 - read-permission
- *	10 - write-permission
- *	11 - read-write
- * for the internal routines (ie open_namei()/follow_link() etc). 00 is
- * used by symlinks.
- */
 static struct file *do_filp_open(int dfd, const char *filename, int flags,
 				 int mode)
 {
-	int namei_flags, error;
+	int error;
 	struct nameidata nd;
 
-	namei_flags = flags;
-	if ((namei_flags+1) & O_ACCMODE)
-		namei_flags++;
-
-	error = open_namei(dfd, filename, namei_flags, mode, &nd);
+	error = open_namei(dfd, filename, flags, mode, &nd);
 	if (!error)
 		return nameidata_to_filp(&nd, flags);
 

diff -puN fs/namei.c~make-open_namei-return-a-filp fs/namei.c
@@ -1750,8 +1750,8 @@ static inline int sys_open_flags_to_name
  * system call.  See sys_open_flags_to_namei_flags().
  * SMP-safe
  */
-int open_namei(int dfd, const char *pathname, int sys_open_flag,
-		int mode, struct nameidata *nd)
+struct file *open_namei(int dfd, const char *pathname, int sys_open_flag,
+			int mode, struct nameidata *nd)
 {
 	int acc_mode, error;
 	struct path path;

@@ -1777,7 +1777,7 @@ int open_namei(int dfd, const char *path
 		error = path_lookup_open(dfd, pathname, lookup_flags(flag),
 					 nd, flag);
 		if (error)
-			return error;
+			return ERR_PTR(error);
 		goto ok;
 	}
 

@@ -1786,7 +1786,7 @@ int open_namei(int dfd, const char *path
 	 */
 	error = path_lookup_create(dfd,pathname,LOOKUP_PARENT,nd,flag,mode);
 	if (error)
-		return error;
+		return ERR_PTR(error);
 
 	/*
 	 * We have the parent and last component. First of all, check

@@ -1820,7 +1820,7 @@ do_last:
 		error = __open_namei_create(nd, &path, flag, mode);
 		if (error)
 			goto exit;
-		return 0;
+		return nameidata_to_filp(nd, sys_open_flag);
 	}
 
 	/*

@@ -1864,7 +1864,7 @@ ok:
 			mnt_drop_write(nd->mnt);
 		goto exit;
 	}
-	return 0;
+	return nameidata_to_filp(nd, sys_open_flag);
 
 exit_mutex_unlock:
 	mutex_unlock(&dir->d_inode->i_mutex);

@@ -1874,7 +1874,7 @@ exit:
 	if (!IS_ERR(nd->intent.open.file))
 		release_open_intent(nd);
 	path_release(nd);
-	return error;
+	return ERR_PTR(error);
 
 do_link:
 	error = -ELOOP;

@@ -1901,7 +1901,7 @@ do_link:
 		 * with "intent.open".
 		 */
 		release_open_intent(nd);
-		return error;
+		return ERR_PTR(error);
 	}
 	nd->flags &= ~LOOKUP_PARENT;
 	if (nd->last_type == LAST_BIND)

diff -puN fs/open.c~make-open_namei-return-a-filp fs/open.c
@@ -846,11 +846,7 @@ static struct file *do_filp_open(int dfd
 	int error;
 	struct nameidata nd;
 
-	error = open_namei(dfd, filename, flags, mode, &nd);
-	if (!error)
-		return nameidata_to_filp(&nd, flags);
-
-	return ERR_PTR(error);
+	return open_namei(dfd, filename, flags, mode, &nd);
 }
 
 struct file *filp_open(const char *filename, int flags, int mode)

diff -puN include/linux/fs.h~make-open_namei-return-a-filp include/linux/fs.h
@@ -1716,7 +1716,7 @@ extern struct file *create_read_pipe(str
 extern struct file *create_write_pipe(void);
 extern void free_write_pipe(struct file *);
 
-extern int open_namei(int dfd, const char *, int, int, struct nameidata *);
+extern struct file *open_namei(int dfd, const char *, int, int, struct nameidata *);
 extern int may_open(struct nameidata *, int, int);
 
 extern int kernel_read(struct file *, unsigned long, char *, unsigned long);