26 files changed, 16978 insertions, 0 deletions

diff --git a/fs/ext4/Makefile b/fs/ext4/Makefile
new file mode 100644
index 000000000000..704cd44a40c2
--- /dev/null
+++ b/fs/ext4/Makefile
@@ -0,0 +1,12 @@
+#
+# Makefile for the linux ext3-filesystem routines.
+#
+
+obj-$(CONFIG_EXT3_FS) += ext3.o
+
+ext3-y	:= balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o \
+	   ioctl.o namei.o super.o symlink.o hash.o resize.o
+
+ext3-$(CONFIG_EXT3_FS_XATTR)	 += xattr.o xattr_user.o xattr_trusted.o
+ext3-$(CONFIG_EXT3_FS_POSIX_ACL) += acl.o
+ext3-$(CONFIG_EXT3_FS_SECURITY)	 += xattr_security.o
diff --git a/fs/ext4/acl.c b/fs/ext4/acl.c
new file mode 100644
index 000000000000..1e5038d9a01b
--- /dev/null
+++ b/fs/ext4/acl.c
@@ -0,0 +1,551 @@
+/*
+ * linux/fs/ext3/acl.c
+ *
+ * Copyright (C) 2001-2003 Andreas Gruenbacher, <agruen@suse.de>
+ */
+
+#include <linux/init.h>
+#include <linux/sched.h>
+#include <linux/slab.h>
+#include <linux/capability.h>
+#include <linux/fs.h>
+#include <linux/ext3_jbd.h>
+#include <linux/ext3_fs.h>
+#include "xattr.h"
+#include "acl.h"
+
+/*
+ * Convert from filesystem to in-memory representation.
+ */
+static struct posix_acl *
+ext3_acl_from_disk(const void *value, size_t size)
+{
+	const char *end = (char *)value + size;
+	int n, count;
+	struct posix_acl *acl;
+
+	if (!value)
+		return NULL;
+	if (size < sizeof(ext3_acl_header))
+		 return ERR_PTR(-EINVAL);
+	if (((ext3_acl_header *)value)->a_version !=
+	    cpu_to_le32(EXT3_ACL_VERSION))
+		return ERR_PTR(-EINVAL);
+	value = (char *)value + sizeof(ext3_acl_header);
+	count = ext3_acl_count(size);
+	if (count < 0)
+		return ERR_PTR(-EINVAL);
+	if (count == 0)
+		return NULL;
+	acl = posix_acl_alloc(count, GFP_KERNEL);
+	if (!acl)
+		return ERR_PTR(-ENOMEM);
+	for (n=0; n < count; n++) {
+		ext3_acl_entry *entry =
+			(ext3_acl_entry *)value;
+		if ((char *)value + sizeof(ext3_acl_entry_short) > end)
+			goto fail;
+		acl->a_entries[n].e_tag  = le16_to_cpu(entry->e_tag);
+		acl->a_entries[n].e_perm = le16_to_cpu(entry->e_perm);
+		switch(acl->a_entries[n].e_tag) {
+			case ACL_USER_OBJ:
+			case ACL_GROUP_OBJ:
+			case ACL_MASK:
+			case ACL_OTHER:
+				value = (char *)value +
+					sizeof(ext3_acl_entry_short);
+				acl->a_entries[n].e_id = ACL_UNDEFINED_ID;
+				break;
+
+			case ACL_USER:
+			case ACL_GROUP:
+				value = (char *)value + sizeof(ext3_acl_entry);
+				if ((char *)value > end)
+					goto fail;
+				acl->a_entries[n].e_id =
+					le32_to_cpu(entry->e_id);
+				break;
+
+			default:
+				goto fail;
+		}
+	}
+	if (value != end)
+		goto fail;
+	return acl;
+
+fail:
+	posix_acl_release(acl);
+	return ERR_PTR(-EINVAL);
+}
+
+/*
+ * Convert from in-memory to filesystem representation.
+ */
+static void *
+ext3_acl_to_disk(const struct posix_acl *acl, size_t *size)
+{
+	ext3_acl_header *ext_acl;
+	char *e;
+	size_t n;
+
+	*size = ext3_acl_size(acl->a_count);
+	ext_acl = kmalloc(sizeof(ext3_acl_header) + acl->a_count *
+			sizeof(ext3_acl_entry), GFP_KERNEL);
+	if (!ext_acl)
+		return ERR_PTR(-ENOMEM);
+	ext_acl->a_version = cpu_to_le32(EXT3_ACL_VERSION);
+	e = (char *)ext_acl + sizeof(ext3_acl_header);
+	for (n=0; n < acl->a_count; n++) {
+		ext3_acl_entry *entry = (ext3_acl_entry *)e;
+		entry->e_tag  = cpu_to_le16(acl->a_entries[n].e_tag);
+		entry->e_perm = cpu_to_le16(acl->a_entries[n].e_perm);
+		switch(acl->a_entries[n].e_tag) {
+			case ACL_USER:
+			case ACL_GROUP:
+				entry->e_id =
+					cpu_to_le32(acl->a_entries[n].e_id);
+				e += sizeof(ext3_acl_entry);
+				break;
+
+			case ACL_USER_OBJ:
+			case ACL_GROUP_OBJ:
+			case ACL_MASK:
+			case ACL_OTHER:
+				e += sizeof(ext3_acl_entry_short);
+				break;
+
+			default:
+				goto fail;
+		}
+	}
+	return (char *)ext_acl;
+
+fail:
+	kfree(ext_acl);
+	return ERR_PTR(-EINVAL);
+}
+
+static inline struct posix_acl *
+ext3_iget_acl(struct inode *inode, struct posix_acl **i_acl)
+{
+	struct posix_acl *acl = EXT3_ACL_NOT_CACHED;
+
+	spin_lock(&inode->i_lock);
+	if (*i_acl != EXT3_ACL_NOT_CACHED)
+		acl = posix_acl_dup(*i_acl);
+	spin_unlock(&inode->i_lock);
+
+	return acl;
+}
+
+static inline void
+ext3_iset_acl(struct inode *inode, struct posix_acl **i_acl,
+                  struct posix_acl *acl)
+{
+	spin_lock(&inode->i_lock);
+	if (*i_acl != EXT3_ACL_NOT_CACHED)
+		posix_acl_release(*i_acl);
+	*i_acl = posix_acl_dup(acl);
+	spin_unlock(&inode->i_lock);
+}
+
+/*
+ * Inode operation get_posix_acl().
+ *
+ * inode->i_mutex: don't care
+ */
+static struct posix_acl *
+ext3_get_acl(struct inode *inode, int type)
+{
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	int name_index;
+	char *value = NULL;
+	struct posix_acl *acl;
+	int retval;
+
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return NULL;
+
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			acl = ext3_iget_acl(inode, &ei->i_acl);
+			if (acl != EXT3_ACL_NOT_CACHED)
+				return acl;
+			name_index = EXT3_XATTR_INDEX_POSIX_ACL_ACCESS;
+			break;
+
+		case ACL_TYPE_DEFAULT:
+			acl = ext3_iget_acl(inode, &ei->i_default_acl);
+			if (acl != EXT3_ACL_NOT_CACHED)
+				return acl;
+			name_index = EXT3_XATTR_INDEX_POSIX_ACL_DEFAULT;
+			break;
+
+		default:
+			return ERR_PTR(-EINVAL);
+	}
+	retval = ext3_xattr_get(inode, name_index, "", NULL, 0);
+	if (retval > 0) {
+		value = kmalloc(retval, GFP_KERNEL);
+		if (!value)
+			return ERR_PTR(-ENOMEM);
+		retval = ext3_xattr_get(inode, name_index, "", value, retval);
+	}
+	if (retval > 0)
+		acl = ext3_acl_from_disk(value, retval);
+	else if (retval == -ENODATA || retval == -ENOSYS)
+		acl = NULL;
+	else
+		acl = ERR_PTR(retval);
+	kfree(value);
+
+	if (!IS_ERR(acl)) {
+		switch(type) {
+			case ACL_TYPE_ACCESS:
+				ext3_iset_acl(inode, &ei->i_acl, acl);
+				break;
+
+			case ACL_TYPE_DEFAULT:
+				ext3_iset_acl(inode, &ei->i_default_acl, acl);
+				break;
+		}
+	}
+	return acl;
+}
+
+/*
+ * Set the access or default ACL of an inode.
+ *
+ * inode->i_mutex: down unless called from ext3_new_inode
+ */
+static int
+ext3_set_acl(handle_t *handle, struct inode *inode, int type,
+	     struct posix_acl *acl)
+{
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	int name_index;
+	void *value = NULL;
+	size_t size = 0;
+	int error;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+
+	switch(type) {
+		case ACL_TYPE_ACCESS:
+			name_index = EXT3_XATTR_INDEX_POSIX_ACL_ACCESS;
+			if (acl) {
+				mode_t mode = inode->i_mode;
+				error = posix_acl_equiv_mode(acl, &mode);
+				if (error < 0)
+					return error;
+				else {
+					inode->i_mode = mode;
+					ext3_mark_inode_dirty(handle, inode);
+					if (error == 0)
+						acl = NULL;
+				}
+			}
+			break;
+
+		case ACL_TYPE_DEFAULT:
+			name_index = EXT3_XATTR_INDEX_POSIX_ACL_DEFAULT;
+			if (!S_ISDIR(inode->i_mode))
+				return acl ? -EACCES : 0;
+			break;
+
+		default:
+			return -EINVAL;
+	}
+	if (acl) {
+		value = ext3_acl_to_disk(acl, &size);
+		if (IS_ERR(value))
+			return (int)PTR_ERR(value);
+	}
+
+	error = ext3_xattr_set_handle(handle, inode, name_index, "",
+				      value, size, 0);
+
+	kfree(value);
+	if (!error) {
+		switch(type) {
+			case ACL_TYPE_ACCESS:
+				ext3_iset_acl(inode, &ei->i_acl, acl);
+				break;
+
+			case ACL_TYPE_DEFAULT:
+				ext3_iset_acl(inode, &ei->i_default_acl, acl);
+				break;
+		}
+	}
+	return error;
+}
+
+static int
+ext3_check_acl(struct inode *inode, int mask)
+{
+	struct posix_acl *acl = ext3_get_acl(inode, ACL_TYPE_ACCESS);
+
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl) {
+		int error = posix_acl_permission(inode, acl, mask);
+		posix_acl_release(acl);
+		return error;
+	}
+
+	return -EAGAIN;
+}
+
+int
+ext3_permission(struct inode *inode, int mask, struct nameidata *nd)
+{
+	return generic_permission(inode, mask, ext3_check_acl);
+}
+
+/*
+ * Initialize the ACLs of a new inode. Called from ext3_new_inode.
+ *
+ * dir->i_mutex: down
+ * inode->i_mutex: up (access to inode is still exclusive)
+ */
+int
+ext3_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
+{
+	struct posix_acl *acl = NULL;
+	int error = 0;
+
+	if (!S_ISLNK(inode->i_mode)) {
+		if (test_opt(dir->i_sb, POSIX_ACL)) {
+			acl = ext3_get_acl(dir, ACL_TYPE_DEFAULT);
+			if (IS_ERR(acl))
+				return PTR_ERR(acl);
+		}
+		if (!acl)
+			inode->i_mode &= ~current->fs->umask;
+	}
+	if (test_opt(inode->i_sb, POSIX_ACL) && acl) {
+		struct posix_acl *clone;
+		mode_t mode;
+
+		if (S_ISDIR(inode->i_mode)) {
+			error = ext3_set_acl(handle, inode,
+					     ACL_TYPE_DEFAULT, acl);
+			if (error)
+				goto cleanup;
+		}
+		clone = posix_acl_clone(acl, GFP_KERNEL);
+		error = -ENOMEM;
+		if (!clone)
+			goto cleanup;
+
+		mode = inode->i_mode;
+		error = posix_acl_create_masq(clone, &mode);
+		if (error >= 0) {
+			inode->i_mode = mode;
+			if (error > 0) {
+				/* This is an extended ACL */
+				error = ext3_set_acl(handle, inode,
+						     ACL_TYPE_ACCESS, clone);
+			}
+		}
+		posix_acl_release(clone);
+	}
+cleanup:
+	posix_acl_release(acl);
+	return error;
+}
+
+/*
+ * Does chmod for an inode that may have an Access Control List. The
+ * inode->i_mode field must be updated to the desired value by the caller
+ * before calling this function.
+ * Returns 0 on success, or a negative error number.
+ *
+ * We change the ACL rather than storing some ACL entries in the file
+ * mode permission bits (which would be more efficient), because that
+ * would break once additional permissions (like  ACL_APPEND, ACL_DELETE
+ * for directories) are added. There are no more bits available in the
+ * file mode.
+ *
+ * inode->i_mutex: down
+ */
+int
+ext3_acl_chmod(struct inode *inode)
+{
+	struct posix_acl *acl, *clone;
+        int error;
+
+	if (S_ISLNK(inode->i_mode))
+		return -EOPNOTSUPP;
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return 0;
+	acl = ext3_get_acl(inode, ACL_TYPE_ACCESS);
+	if (IS_ERR(acl) || !acl)
+		return PTR_ERR(acl);
+	clone = posix_acl_clone(acl, GFP_KERNEL);
+	posix_acl_release(acl);
+	if (!clone)
+		return -ENOMEM;
+	error = posix_acl_chmod_masq(clone, inode->i_mode);
+	if (!error) {
+		handle_t *handle;
+		int retries = 0;
+
+	retry:
+		handle = ext3_journal_start(inode,
+				EXT3_DATA_TRANS_BLOCKS(inode->i_sb));
+		if (IS_ERR(handle)) {
+			error = PTR_ERR(handle);
+			ext3_std_error(inode->i_sb, error);
+			goto out;
+		}
+		error = ext3_set_acl(handle, inode, ACL_TYPE_ACCESS, clone);
+		ext3_journal_stop(handle);
+		if (error == -ENOSPC &&
+		    ext3_should_retry_alloc(inode->i_sb, &retries))
+			goto retry;
+	}
+out:
+	posix_acl_release(clone);
+	return error;
+}
+
+/*
+ * Extended attribute handlers
+ */
+static size_t
+ext3_xattr_list_acl_access(struct inode *inode, char *list, size_t list_len,
+			   const char *name, size_t name_len)
+{
+	const size_t size = sizeof(POSIX_ACL_XATTR_ACCESS);
+
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return 0;
+	if (list && size <= list_len)
+		memcpy(list, POSIX_ACL_XATTR_ACCESS, size);
+	return size;
+}
+
+static size_t
+ext3_xattr_list_acl_default(struct inode *inode, char *list, size_t list_len,
+			    const char *name, size_t name_len)
+{
+	const size_t size = sizeof(POSIX_ACL_XATTR_DEFAULT);
+
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return 0;
+	if (list && size <= list_len)
+		memcpy(list, POSIX_ACL_XATTR_DEFAULT, size);
+	return size;
+}
+
+static int
+ext3_xattr_get_acl(struct inode *inode, int type, void *buffer, size_t size)
+{
+	struct posix_acl *acl;
+	int error;
+
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return -EOPNOTSUPP;
+
+	acl = ext3_get_acl(inode, type);
+	if (IS_ERR(acl))
+		return PTR_ERR(acl);
+	if (acl == NULL)
+		return -ENODATA;
+	error = posix_acl_to_xattr(acl, buffer, size);
+	posix_acl_release(acl);
+
+	return error;
+}
+
+static int
+ext3_xattr_get_acl_access(struct inode *inode, const char *name,
+			  void *buffer, size_t size)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ext3_xattr_get_acl(inode, ACL_TYPE_ACCESS, buffer, size);
+}
+
+static int
+ext3_xattr_get_acl_default(struct inode *inode, const char *name,
+			   void *buffer, size_t size)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ext3_xattr_get_acl(inode, ACL_TYPE_DEFAULT, buffer, size);
+}
+
+static int
+ext3_xattr_set_acl(struct inode *inode, int type, const void *value,
+		   size_t size)
+{
+	handle_t *handle;
+	struct posix_acl *acl;
+	int error, retries = 0;
+
+	if (!test_opt(inode->i_sb, POSIX_ACL))
+		return -EOPNOTSUPP;
+	if ((current->fsuid != inode->i_uid) && !capable(CAP_FOWNER))
+		return -EPERM;
+
+	if (value) {
+		acl = posix_acl_from_xattr(value, size);
+		if (IS_ERR(acl))
+			return PTR_ERR(acl);
+		else if (acl) {
+			error = posix_acl_valid(acl);
+			if (error)
+				goto release_and_out;
+		}
+	} else
+		acl = NULL;
+
+retry:
+	handle = ext3_journal_start(inode, EXT3_DATA_TRANS_BLOCKS(inode->i_sb));
+	if (IS_ERR(handle))
+		return PTR_ERR(handle);
+	error = ext3_set_acl(handle, inode, type, acl);
+	ext3_journal_stop(handle);
+	if (error == -ENOSPC && ext3_should_retry_alloc(inode->i_sb, &retries))
+		goto retry;
+
+release_and_out:
+	posix_acl_release(acl);
+	return error;
+}
+
+static int
+ext3_xattr_set_acl_access(struct inode *inode, const char *name,
+			  const void *value, size_t size, int flags)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ext3_xattr_set_acl(inode, ACL_TYPE_ACCESS, value, size);
+}
+
+static int
+ext3_xattr_set_acl_default(struct inode *inode, const char *name,
+			   const void *value, size_t size, int flags)
+{
+	if (strcmp(name, "") != 0)
+		return -EINVAL;
+	return ext3_xattr_set_acl(inode, ACL_TYPE_DEFAULT, value, size);
+}
+
+struct xattr_handler ext3_xattr_acl_access_handler = {
+	.prefix	= POSIX_ACL_XATTR_ACCESS,
+	.list	= ext3_xattr_list_acl_access,
+	.get	= ext3_xattr_get_acl_access,
+	.set	= ext3_xattr_set_acl_access,
+};
+
+struct xattr_handler ext3_xattr_acl_default_handler = {
+	.prefix	= POSIX_ACL_XATTR_DEFAULT,
+	.list	= ext3_xattr_list_acl_default,
+	.get	= ext3_xattr_get_acl_default,
+	.set	= ext3_xattr_set_acl_default,
+};
diff --git a/fs/ext4/acl.h b/fs/ext4/acl.h
new file mode 100644
index 000000000000..0d1e6279cbfd
--- /dev/null
+++ b/fs/ext4/acl.h
@@ -0,0 +1,81 @@
+/*
+  File: fs/ext3/acl.h
+
+  (C) 2001 Andreas Gruenbacher, <a.gruenbacher@computer.org>
+*/
+
+#include <linux/posix_acl_xattr.h>
+
+#define EXT3_ACL_VERSION	0x0001
+
+typedef struct {
+	__le16		e_tag;
+	__le16		e_perm;
+	__le32		e_id;
+} ext3_acl_entry;
+
+typedef struct {
+	__le16		e_tag;
+	__le16		e_perm;
+} ext3_acl_entry_short;
+
+typedef struct {
+	__le32		a_version;
+} ext3_acl_header;
+
+static inline size_t ext3_acl_size(int count)
+{
+	if (count <= 4) {
+		return sizeof(ext3_acl_header) +
+		       count * sizeof(ext3_acl_entry_short);
+	} else {
+		return sizeof(ext3_acl_header) +
+		       4 * sizeof(ext3_acl_entry_short) +
+		       (count - 4) * sizeof(ext3_acl_entry);
+	}
+}
+
+static inline int ext3_acl_count(size_t size)
+{
+	ssize_t s;
+	size -= sizeof(ext3_acl_header);
+	s = size - 4 * sizeof(ext3_acl_entry_short);
+	if (s < 0) {
+		if (size % sizeof(ext3_acl_entry_short))
+			return -1;
+		return size / sizeof(ext3_acl_entry_short);
+	} else {
+		if (s % sizeof(ext3_acl_entry))
+			return -1;
+		return s / sizeof(ext3_acl_entry) + 4;
+	}
+}
+
+#ifdef CONFIG_EXT3_FS_POSIX_ACL
+
+/* Value for inode->u.ext3_i.i_acl and inode->u.ext3_i.i_default_acl
+   if the ACL has not been cached */
+#define EXT3_ACL_NOT_CACHED ((void *)-1)
+
+/* acl.c */
+extern int ext3_permission (struct inode *, int, struct nameidata *);
+extern int ext3_acl_chmod (struct inode *);
+extern int ext3_init_acl (handle_t *, struct inode *, struct inode *);
+
+#else  /* CONFIG_EXT3_FS_POSIX_ACL */
+#include <linux/sched.h>
+#define ext3_permission NULL
+
+static inline int
+ext3_acl_chmod(struct inode *inode)
+{
+	return 0;
+}
+
+static inline int
+ext3_init_acl(handle_t *handle, struct inode *inode, struct inode *dir)
+{
+	return 0;
+}
+#endif  /* CONFIG_EXT3_FS_POSIX_ACL */
+
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
new file mode 100644
index 000000000000..b41a7d7e20f0
--- /dev/null
+++ b/fs/ext4/balloc.c
@@ -0,0 +1,1818 @@
+/*
+ *  linux/fs/ext3/balloc.c
+ *
+ * Copyright (C) 1992, 1993, 1994, 1995
+ * Remy Card (card@masi.ibp.fr)
+ * Laboratoire MASI - Institut Blaise Pascal
+ * Universite Pierre et Marie Curie (Paris VI)
+ *
+ *  Enhanced block allocation by Stephen Tweedie (sct@redhat.com), 1993
+ *  Big-endian to little-endian byte-swapping/bitmaps by
+ *        David S. Miller (davem@caip.rutgers.edu), 1995
+ */
+
+#include <linux/time.h>
+#include <linux/capability.h>
+#include <linux/fs.h>
+#include <linux/jbd.h>
+#include <linux/ext3_fs.h>
+#include <linux/ext3_jbd.h>
+#include <linux/quotaops.h>
+#include <linux/buffer_head.h>
+
+/*
+ * balloc.c contains the blocks allocation and deallocation routines
+ */
+
+/*
+ * The free blocks are managed by bitmaps.  A file system contains several
+ * blocks groups.  Each group contains 1 bitmap block for blocks, 1 bitmap
+ * block for inodes, N blocks for the inode table and data blocks.
+ *
+ * The file system contains group descriptors which are located after the
+ * super block.  Each descriptor contains the number of the bitmap block and
+ * the free blocks count in the block.  The descriptors are loaded in memory
+ * when a file system is mounted (see ext3_read_super).
+ */
+
+
+#define in_range(b, first, len)	((b) >= (first) && (b) <= (first) + (len) - 1)
+
+/**
+ * ext3_get_group_desc() -- load group descriptor from disk
+ * @sb:			super block
+ * @block_group:	given block group
+ * @bh:			pointer to the buffer head to store the block
+ *			group descriptor
+ */
+struct ext3_group_desc * ext3_get_group_desc(struct super_block * sb,
+					     unsigned int block_group,
+					     struct buffer_head ** bh)
+{
+	unsigned long group_desc;
+	unsigned long offset;
+	struct ext3_group_desc * desc;
+	struct ext3_sb_info *sbi = EXT3_SB(sb);
+
+	if (block_group >= sbi->s_groups_count) {
+		ext3_error (sb, "ext3_get_group_desc",
+			    "block_group >= groups_count - "
+			    "block_group = %d, groups_count = %lu",
+			    block_group, sbi->s_groups_count);
+
+		return NULL;
+	}
+	smp_rmb();
+
+	group_desc = block_group >> EXT3_DESC_PER_BLOCK_BITS(sb);
+	offset = block_group & (EXT3_DESC_PER_BLOCK(sb) - 1);
+	if (!sbi->s_group_desc[group_desc]) {
+		ext3_error (sb, "ext3_get_group_desc",
+			    "Group descriptor not loaded - "
+			    "block_group = %d, group_desc = %lu, desc = %lu",
+			     block_group, group_desc, offset);
+		return NULL;
+	}
+
+	desc = (struct ext3_group_desc *) sbi->s_group_desc[group_desc]->b_data;
+	if (bh)
+		*bh = sbi->s_group_desc[group_desc];
+	return desc + offset;
+}
+
+/**
+ * read_block_bitmap()
+ * @sb:			super block
+ * @block_group:	given block group
+ *
+ * Read the bitmap for a given block_group, reading into the specified
+ * slot in the superblock's bitmap cache.
+ *
+ * Return buffer_head on success or NULL in case of failure.
+ */
+static struct buffer_head *
+read_block_bitmap(struct super_block *sb, unsigned int block_group)
+{
+	struct ext3_group_desc * desc;
+	struct buffer_head * bh = NULL;
+
+	desc = ext3_get_group_desc (sb, block_group, NULL);
+	if (!desc)
+		goto error_out;
+	bh = sb_bread(sb, le32_to_cpu(desc->bg_block_bitmap));
+	if (!bh)
+		ext3_error (sb, "read_block_bitmap",
+			    "Cannot read block bitmap - "
+			    "block_group = %d, block_bitmap = %u",
+			    block_group, le32_to_cpu(desc->bg_block_bitmap));
+error_out:
+	return bh;
+}
+/*
+ * The reservation window structure operations
+ * --------------------------------------------
+ * Operations include:
+ * dump, find, add, remove, is_empty, find_next_reservable_window, etc.
+ *
+ * We use a red-black tree to represent per-filesystem reservation
+ * windows.
+ *
+ */
+
+/**
+ * __rsv_window_dump() -- Dump the filesystem block allocation reservation map
+ * @rb_root:		root of per-filesystem reservation rb tree
+ * @verbose:		verbose mode
+ * @fn:			function which wishes to dump the reservation map
+ *
+ * If verbose is turned on, it will print the whole block reservation
+ * windows(start, end).	Otherwise, it will only print out the "bad" windows,
+ * those windows that overlap with their immediate neighbors.
+ */
+#if 1
+static void __rsv_window_dump(struct rb_root *root, int verbose,
+			      const char *fn)
+{
+	struct rb_node *n;
+	struct ext3_reserve_window_node *rsv, *prev;
+	int bad;
+
+restart:
+	n = rb_first(root);
+	bad = 0;
+	prev = NULL;
+
+	printk("Block Allocation Reservation Windows Map (%s):\n", fn);
+	while (n) {
+		rsv = list_entry(n, struct ext3_reserve_window_node, rsv_node);
+		if (verbose)
+			printk("reservation window 0x%p "
+			       "start:  %lu, end:  %lu\n",
+			       rsv, rsv->rsv_start, rsv->rsv_end);
+		if (rsv->rsv_start && rsv->rsv_start >= rsv->rsv_end) {
+			printk("Bad reservation %p (start >= end)\n",
+			       rsv);
+			bad = 1;
+		}
+		if (prev && prev->rsv_end >= rsv->rsv_start) {
+			printk("Bad reservation %p (prev->end >= start)\n",
+			       rsv);
+			bad = 1;
+		}
+		if (bad) {
+			if (!verbose) {
+				printk("Restarting reservation walk in verbose mode\n");
+				verbose = 1;
+				goto restart;
+			}
+		}
+		n = rb_next(n);
+		prev = rsv;
+	}
+	printk("Window map complete.\n");
+	if (bad)
+		BUG();
+}
+#define rsv_window_dump(root, verbose) \
+	__rsv_window_dump((root), (verbose), __FUNCTION__)
+#else
+#define rsv_window_dump(root, verbose) do {} while (0)
+#endif
+
+/**
+ * goal_in_my_reservation()
+ * @rsv:		inode's reservation window
+ * @grp_goal:		given goal block relative to the allocation block group
+ * @group:		the current allocation block group
+ * @sb:			filesystem super block
+ *
+ * Test if the given goal block (group relative) is within the file's
+ * own block reservation window range.
+ *
+ * If the reservation window is outside the goal allocation group, return 0;
+ * grp_goal (given goal block) could be -1, which means no specific
+ * goal block. In this case, always return 1.
+ * If the goal block is within the reservation window, return 1;
+ * otherwise, return 0;
+ */
+static int
+goal_in_my_reservation(struct ext3_reserve_window *rsv, ext3_grpblk_t grp_goal,
+			unsigned int group, struct super_block * sb)
+{
+	ext3_fsblk_t group_first_block, group_last_block;
+
+	group_first_block = ext3_group_first_block_no(sb, group);
+	group_last_block = group_first_block + (EXT3_BLOCKS_PER_GROUP(sb) - 1);
+
+	if ((rsv->_rsv_start > group_last_block) ||
+	    (rsv->_rsv_end < group_first_block))
+		return 0;
+	if ((grp_goal >= 0) && ((grp_goal + group_first_block < rsv->_rsv_start)
+		|| (grp_goal + group_first_block > rsv->_rsv_end)))
+		return 0;
+	return 1;
+}
+
+/**
+ * search_reserve_window()
+ * @rb_root:		root of reservation tree
+ * @goal:		target allocation block
+ *
+ * Find the reserved window which includes the goal, or the previous one
+ * if the goal is not in any window.
+ * Returns NULL if there are no windows or if all windows start after the goal.
+ */
+static struct ext3_reserve_window_node *
+search_reserve_window(struct rb_root *root, ext3_fsblk_t goal)
+{
+	struct rb_node *n = root->rb_node;
+	struct ext3_reserve_window_node *rsv;
+
+	if (!n)
+		return NULL;
+
+	do {
+		rsv = rb_entry(n, struct ext3_reserve_window_node, rsv_node);
+
+		if (goal < rsv->rsv_start)
+			n = n->rb_left;
+		else if (goal > rsv->rsv_end)
+			n = n->rb_right;
+		else
+			return rsv;
+	} while (n);
+	/*
+	 * We've fallen off the end of the tree: the goal wasn't inside
+	 * any particular node.  OK, the previous node must be to one
+	 * side of the interval containing the goal.  If it's the RHS,
+	 * we need to back up one.
+	 */
+	if (rsv->rsv_start > goal) {
+		n = rb_prev(&rsv->rsv_node);
+		rsv = rb_entry(n, struct ext3_reserve_window_node, rsv_node);
+	}
+	return rsv;
+}
+
+/**
+ * ext3_rsv_window_add() -- Insert a window to the block reservation rb tree.
+ * @sb:			super block
+ * @rsv:		reservation window to add
+ *
+ * Must be called with rsv_lock hold.
+ */
+void ext3_rsv_window_add(struct super_block *sb,
+		    struct ext3_reserve_window_node *rsv)
+{
+	struct rb_root *root = &EXT3_SB(sb)->s_rsv_window_root;
+	struct rb_node *node = &rsv->rsv_node;
+	ext3_fsblk_t start = rsv->rsv_start;
+
+	struct rb_node ** p = &root->rb_node;
+	struct rb_node * parent = NULL;
+	struct ext3_reserve_window_node *this;
+
+	while (*p)
+	{
+		parent = *p;
+		this = rb_entry(parent, struct ext3_reserve_window_node, rsv_node);
+
+		if (start < this->rsv_start)
+			p = &(*p)->rb_left;
+		else if (start > this->rsv_end)
+			p = &(*p)->rb_right;
+		else {
+			rsv_window_dump(root, 1);
+			BUG();
+		}
+	}
+
+	rb_link_node(node, parent, p);
+	rb_insert_color(node, root);
+}
+
+/**
+ * ext3_rsv_window_remove() -- unlink a window from the reservation rb tree
+ * @sb:			super block
+ * @rsv:		reservation window to remove
+ *
+ * Mark the block reservation window as not allocated, and unlink it
+ * from the filesystem reservation window rb tree. Must be called with
+ * rsv_lock hold.
+ */
+static void rsv_window_remove(struct super_block *sb,
+			      struct ext3_reserve_window_node *rsv)
+{
+	rsv->rsv_start = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+	rsv->rsv_end = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+	rsv->rsv_alloc_hit = 0;
+	rb_erase(&rsv->rsv_node, &EXT3_SB(sb)->s_rsv_window_root);
+}
+
+/*
+ * rsv_is_empty() -- Check if the reservation window is allocated.
+ * @rsv:		given reservation window to check
+ *
+ * returns 1 if the end block is EXT3_RESERVE_WINDOW_NOT_ALLOCATED.
+ */
+static inline int rsv_is_empty(struct ext3_reserve_window *rsv)
+{
+	/* a valid reservation end block could not be 0 */
+	return rsv->_rsv_end == EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+}
+
+/**
+ * ext3_init_block_alloc_info()
+ * @inode:		file inode structure
+ *
+ * Allocate and initialize the	reservation window structure, and
+ * link the window to the ext3 inode structure at last
+ *
+ * The reservation window structure is only dynamically allocated
+ * and linked to ext3 inode the first time the open file
+ * needs a new block. So, before every ext3_new_block(s) call, for
+ * regular files, we should check whether the reservation window
+ * structure exists or not. In the latter case, this function is called.
+ * Fail to do so will result in block reservation being turned off for that
+ * open file.
+ *
+ * This function is called from ext3_get_blocks_handle(), also called
+ * when setting the reservation window size through ioctl before the file
+ * is open for write (needs block allocation).
+ *
+ * Needs truncate_mutex protection prior to call this function.
+ */
+void ext3_init_block_alloc_info(struct inode *inode)
+{
+	struct ext3_inode_info *ei = EXT3_I(inode);
+	struct ext3_block_alloc_info *block_i = ei->i_block_alloc_info;
+	struct super_block *sb = inode->i_sb;
+
+	block_i = kmalloc(sizeof(*block_i), GFP_NOFS);
+	if (block_i) {
+		struct ext3_reserve_window_node *rsv = &block_i->rsv_window_node;
+
+		rsv->rsv_start = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+		rsv->rsv_end = EXT3_RESERVE_WINDOW_NOT_ALLOCATED;
+
+		/*
+		 * if filesystem is mounted with NORESERVATION, the goal
+		 * reservation window size is set to zero to indicate
+		 * block reservation is off
+		 */
+		if (!test_opt(sb, RESERVATION))
+			rsv->rsv_goal_size = 0;
+		else
+			rsv->rsv_goal_size = EXT3_DEFAULT_RESERVE_BLOCKS;
+		rsv->rsv_alloc_hit = 0;
+		block_i->last_alloc_logical_block = 0;
+		block_i->last_alloc_physical_block = 0;
+	}
+	ei->i_block_