Merge tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4

Pull ext4 updates from Ted Ts'o:

 - add GETFSMAP support

 - some performance improvements for very large file systems and for
   random write workloads into a preallocated file

 - bug fixes and cleanups.

* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  jbd2: cleanup write flags handling from jbd2_write_superblock()
  ext4: mark superblock writes synchronous for nobarrier mounts
  ext4: inherit encryption xattr before other xattrs
  ext4: replace BUG_ON with WARN_ONCE in ext4_end_bio()
  ext4: avoid unnecessary transaction stalls during writeback
  ext4: preload block group descriptors
  ext4: make ext4_shutdown() static
  ext4: support GETFSMAP ioctls
  vfs: add common GETFSMAP ioctl definitions
  ext4: evict inline data when writing to memory map
  ext4: remove ext4_xattr_check_entry()
  ext4: rename ext4_xattr_check_names() to ext4_xattr_check_entries()
  ext4: merge ext4_xattr_list() into ext4_listxattr()
  ext4: constify static data that is never modified
  ext4: trim return value and 'dir' argument from ext4_insert_dentry()
  jbd2: fix dbench4 performance regression for 'nobarrier' mounts
  jbd2: Fix lockdep splat with generic/270 test
  mm: retry writepages() on ENOMEM when doing an data integrity writeback

15 files changed, 1060 insertions, 86 deletions

diff --git a/fs/ext4/Makefile b/fs/ext4/Makefile
index 354103f3490c..d9beca1653c5 100644
--- a/fs/ext4/Makefile
+++ b/fs/ext4/Makefile
@@ -4,11 +4,11 @@
 
 obj-$(CONFIG_EXT4_FS) += ext4.o
 
-ext4-y	:= balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o page-io.o \
-		ioctl.o namei.o super.o symlink.o hash.o resize.o extents.o \
-		ext4_jbd2.o migrate.o mballoc.o block_validity.o move_extent.o \
-		mmp.o indirect.o extents_status.o xattr.o xattr_user.o \
-		xattr_trusted.o inline.o readpage.o sysfs.o
+ext4-y	:= balloc.o bitmap.o block_validity.o dir.o ext4_jbd2.o extents.o \
+		extents_status.o file.o fsmap.o fsync.o hash.o ialloc.o \
+		indirect.o inline.o inode.o ioctl.o mballoc.o migrate.o \
+		mmp.o move_extent.o namei.o page-io.o readpage.o resize.o \
+		super.o symlink.o sysfs.o xattr.o xattr_trusted.o xattr_user.o
 
 ext4-$(CONFIG_EXT4_FS_POSIX_ACL)	+= acl.o
 ext4-$(CONFIG_EXT4_FS_SECURITY)		+= xattr_security.o
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index f7b465b4fb69..8e8046104f4d 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -2356,17 +2356,16 @@ extern int ext4_find_dest_de(struct inode *dir, struct inode *inode,
 			     void *buf, int buf_size,
 			     struct ext4_filename *fname,
 			     struct ext4_dir_entry_2 **dest_de);
-int ext4_insert_dentry(struct inode *dir,
-		       struct inode *inode,
-		       struct ext4_dir_entry_2 *de,
-		       int buf_size,
-		       struct ext4_filename *fname);
+void ext4_insert_dentry(struct inode *inode,
+			struct ext4_dir_entry_2 *de,
+			int buf_size,
+			struct ext4_filename *fname);
 static inline void ext4_update_dx_flag(struct inode *inode)
 {
 	if (!ext4_has_feature_dir_index(inode->i_sb))
 		ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
 }
-static unsigned char ext4_filetype_table[] = {
+static const unsigned char ext4_filetype_table[] = {
 	DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
 };
 
@@ -3050,7 +3049,7 @@ extern int ext4_handle_dirty_dirent_node(handle_t *handle,
 					 struct inode *inode,
 					 struct buffer_head *bh);
 #define S_SHIFT 12
-static unsigned char ext4_type_by_mode[S_IFMT >> S_SHIFT] = {
+static const unsigned char ext4_type_by_mode[S_IFMT >> S_SHIFT] = {
 	[S_IFREG >> S_SHIFT]	= EXT4_FT_REG_FILE,
 	[S_IFDIR >> S_SHIFT]	= EXT4_FT_DIR,
 	[S_IFCHR >> S_SHIFT]	= EXT4_FT_CHRDEV,
diff --git a/fs/ext4/fsmap.c b/fs/ext4/fsmap.c
new file mode 100644
index 000000000000..b19436098837
--- /dev/null
+++ b/fs/ext4/fsmap.c
@@ -0,0 +1,722 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "ext4.h"
+#include <linux/fsmap.h>
+#include "fsmap.h"
+#include "mballoc.h"
+#include <linux/sort.h>
+#include <linux/list_sort.h>
+#include <trace/events/ext4.h>
+
+/* Convert an ext4_fsmap to an fsmap. */
+void ext4_fsmap_from_internal(struct super_block *sb, struct fsmap *dest,
+			      struct ext4_fsmap *src)
+{
+	dest->fmr_device = src->fmr_device;
+	dest->fmr_flags = src->fmr_flags;
+	dest->fmr_physical = src->fmr_physical << sb->s_blocksize_bits;
+	dest->fmr_owner = src->fmr_owner;
+	dest->fmr_offset = 0;
+	dest->fmr_length = src->fmr_length << sb->s_blocksize_bits;
+	dest->fmr_reserved[0] = 0;
+	dest->fmr_reserved[1] = 0;
+	dest->fmr_reserved[2] = 0;
+}
+
+/* Convert an fsmap to an ext4_fsmap. */
+void ext4_fsmap_to_internal(struct super_block *sb, struct ext4_fsmap *dest,
+			    struct fsmap *src)
+{
+	dest->fmr_device = src->fmr_device;
+	dest->fmr_flags = src->fmr_flags;
+	dest->fmr_physical = src->fmr_physical >> sb->s_blocksize_bits;
+	dest->fmr_owner = src->fmr_owner;
+	dest->fmr_length = src->fmr_length >> sb->s_blocksize_bits;
+}
+
+/* getfsmap query state */
+struct ext4_getfsmap_info {
+	struct ext4_fsmap_head	*gfi_head;
+	ext4_fsmap_format_t	gfi_formatter;	/* formatting fn */
+	void			*gfi_format_arg;/* format buffer */
+	ext4_fsblk_t		gfi_next_fsblk;	/* next fsblock we expect */
+	u32			gfi_dev;	/* device id */
+	ext4_group_t		gfi_agno;	/* bg number, if applicable */
+	struct ext4_fsmap	gfi_low;	/* low rmap key */
+	struct ext4_fsmap	gfi_high;	/* high rmap key */
+	struct ext4_fsmap	gfi_lastfree;	/* free ext at end of last bg */
+	struct list_head	gfi_meta_list;	/* fixed metadata list */
+	bool			gfi_last;	/* last extent? */
+};
+
+/* Associate a device with a getfsmap handler. */
+struct ext4_getfsmap_dev {
+	int			(*gfd_fn)(struct super_block *sb,
+				      struct ext4_fsmap *keys,
+				      struct ext4_getfsmap_info *info);
+	u32			gfd_dev;
+};
+
+/* Compare two getfsmap device handlers. */
+static int ext4_getfsmap_dev_compare(const void *p1, const void *p2)
+{
+	const struct ext4_getfsmap_dev *d1 = p1;
+	const struct ext4_getfsmap_dev *d2 = p2;
+
+	return d1->gfd_dev - d2->gfd_dev;
+}
+
+/* Compare a record against our starting point */
+static bool ext4_getfsmap_rec_before_low_key(struct ext4_getfsmap_info *info,
+					     struct ext4_fsmap *rec)
+{
+	return rec->fmr_physical < info->gfi_low.fmr_physical;
+}
+
+/*
+ * Format a reverse mapping for getfsmap, having translated rm_startblock
+ * into the appropriate daddr units.
+ */
+static int ext4_getfsmap_helper(struct super_block *sb,
+				struct ext4_getfsmap_info *info,
+				struct ext4_fsmap *rec)
+{
+	struct ext4_fsmap fmr;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	ext4_fsblk_t rec_fsblk = rec->fmr_physical;
+	ext4_group_t agno;
+	ext4_grpblk_t cno;
+	int error;
+
+	if (fatal_signal_pending(current))
+		return -EINTR;
+
+	/*
+	 * Filter out records that start before our startpoint, if the
+	 * caller requested that.
+	 */
+	if (ext4_getfsmap_rec_before_low_key(info, rec)) {
+		rec_fsblk += rec->fmr_length;
+		if (info->gfi_next_fsblk < rec_fsblk)
+			info->gfi_next_fsblk = rec_fsblk;
+		return EXT4_QUERY_RANGE_CONTINUE;
+	}
+
+	/* Are we just counting mappings? */
+	if (info->gfi_head->fmh_count == 0) {
+		if (rec_fsblk > info->gfi_next_fsblk)
+			info->gfi_head->fmh_entries++;
+
+		if (info->gfi_last)
+			return EXT4_QUERY_RANGE_CONTINUE;
+
+		info->gfi_head->fmh_entries++;
+
+		rec_fsblk += rec->fmr_length;
+		if (info->gfi_next_fsblk < rec_fsblk)
+			info->gfi_next_fsblk = rec_fsblk;
+		return EXT4_QUERY_RANGE_CONTINUE;
+	}
+
+	/*
+	 * If the record starts past the last physical block we saw,
+	 * then we've found a gap.  Report the gap as being owned by
+	 * whatever the caller specified is the missing owner.
+	 */
+	if (rec_fsblk > info->gfi_next_fsblk) {
+		if (info->gfi_head->fmh_entries >= info->gfi_head->fmh_count)
+			return EXT4_QUERY_RANGE_ABORT;
+
+		ext4_get_group_no_and_offset(sb, info->gfi_next_fsblk,
+				&agno, &cno);
+		trace_ext4_fsmap_mapping(sb, info->gfi_dev, agno,
+				EXT4_C2B(sbi, cno),
+				rec_fsblk - info->gfi_next_fsblk,
+				EXT4_FMR_OWN_UNKNOWN);
+
+		fmr.fmr_device = info->gfi_dev;
+		fmr.fmr_physical = info->gfi_next_fsblk;
+		fmr.fmr_owner = EXT4_FMR_OWN_UNKNOWN;
+		fmr.fmr_length = rec_fsblk - info->gfi_next_fsblk;
+		fmr.fmr_flags = FMR_OF_SPECIAL_OWNER;
+		error = info->gfi_formatter(&fmr, info->gfi_format_arg);
+		if (error)
+			return error;
+		info->gfi_head->fmh_entries++;
+	}
+
+	if (info->gfi_last)
+		goto out;
+
+	/* Fill out the extent we found */
+	if (info->gfi_head->fmh_entries >= info->gfi_head->fmh_count)
+		return EXT4_QUERY_RANGE_ABORT;
+
+	ext4_get_group_no_and_offset(sb, rec_fsblk, &agno, &cno);
+	trace_ext4_fsmap_mapping(sb, info->gfi_dev, agno, EXT4_C2B(sbi, cno),
+			rec->fmr_length, rec->fmr_owner);
+
+	fmr.fmr_device = info->gfi_dev;
+	fmr.fmr_physical = rec_fsblk;
+	fmr.fmr_owner = rec->fmr_owner;
+	fmr.fmr_flags = FMR_OF_SPECIAL_OWNER;
+	fmr.fmr_length = rec->fmr_length;
+	error = info->gfi_formatter(&fmr, info->gfi_format_arg);
+	if (error)
+		return error;
+	info->gfi_head->fmh_entries++;
+
+out:
+	rec_fsblk += rec->fmr_length;
+	if (info->gfi_next_fsblk < rec_fsblk)
+		info->gfi_next_fsblk = rec_fsblk;
+	return EXT4_QUERY_RANGE_CONTINUE;
+}
+
+static inline ext4_fsblk_t ext4_fsmap_next_pblk(struct ext4_fsmap *fmr)
+{
+	return fmr->fmr_physical + fmr->fmr_length;
+}
+
+/* Transform a blockgroup's free record into a fsmap */
+static int ext4_getfsmap_datadev_helper(struct super_block *sb,
+					ext4_group_t agno, ext4_grpblk_t start,
+					ext4_grpblk_t len, void *priv)
+{
+	struct ext4_fsmap irec;
+	struct ext4_getfsmap_info *info = priv;
+	struct ext4_fsmap *p;
+	struct ext4_fsmap *tmp;
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	ext4_fsblk_t fsb;
+	ext4_fsblk_t fslen;
+	int error;
+
+	fsb = (EXT4_C2B(sbi, start) + ext4_group_first_block_no(sb, agno));
+	fslen = EXT4_C2B(sbi, len);
+
+	/* If the retained free extent record is set... */
+	if (info->gfi_lastfree.fmr_owner) {
+		/* ...and abuts this one, lengthen it and return. */
+		if (ext4_fsmap_next_pblk(&info->gfi_lastfree) == fsb) {
+			info->gfi_lastfree.fmr_length += fslen;
+			return 0;
+		}
+
+		/*
+		 * There's a gap between the two free extents; emit the
+		 * retained extent prior to merging the meta_list.
+		 */
+		error = ext4_getfsmap_helper(sb, info, &info->gfi_lastfree);
+		if (error)
+			return error;
+		info->gfi_lastfree.fmr_owner = 0;
+	}
+
+	/* Merge in any relevant extents from the meta_list */
+	list_for_each_entry_safe(p, tmp, &info->gfi_meta_list, fmr_list) {
+		if (p->fmr_physical + p->fmr_length <= info->gfi_next_fsblk) {
+			list_del(&p->fmr_list);
+			kfree(p);
+		} else if (p->fmr_physical < fsb) {
+			error = ext4_getfsmap_helper(sb, info, p);
+			if (error)
+				return error;
+
+			list_del(&p->fmr_list);
+			kfree(p);
+		}
+	}
+
+	irec.fmr_device = 0;
+	irec.fmr_physical = fsb;
+	irec.fmr_length = fslen;
+	irec.fmr_owner = EXT4_FMR_OWN_FREE;
+	irec.fmr_flags = 0;
+
+	/* If this is a free extent at the end of a bg, buffer it. */
+	if (ext4_fsmap_next_pblk(&irec) ==
+			ext4_group_first_block_no(sb, agno + 1)) {
+		info->gfi_lastfree = irec;
+		return 0;
+	}
+
+	/* Otherwise, emit it */
+	return ext4_getfsmap_helper(sb, info, &irec);
+}
+
+/* Execute a getfsmap query against the log device. */
+static int ext4_getfsmap_logdev(struct super_block *sb, struct ext4_fsmap *keys,
+				struct ext4_getfsmap_info *info)
+{
+	journal_t *journal = EXT4_SB(sb)->s_journal;
+	struct ext4_fsmap irec;
+
+	/* Set up search keys */
+	info->gfi_low = keys[0];
+	info->gfi_low.fmr_length = 0;
+
+	memset(&info->gfi_high, 0xFF, sizeof(info->gfi_high));
+
+	trace_ext4_fsmap_low_key(sb, info->gfi_dev, 0,
+			info->gfi_low.fmr_physical,
+			info->gfi_low.fmr_length,
+			info->gfi_low.fmr_owner);
+
+	trace_ext4_fsmap_high_key(sb, info->gfi_dev, 0,
+			info->gfi_high.fmr_physical,
+			info->gfi_high.fmr_length,
+			info->gfi_high.fmr_owner);
+
+	if (keys[0].fmr_physical > 0)
+		return 0;
+
+	/* Fabricate an rmap entry for the external log device. */
+	irec.fmr_physical = journal->j_blk_offset;
+	irec.fmr_length = journal->j_maxlen;
+	irec.fmr_owner = EXT4_FMR_OWN_LOG;
+	irec.fmr_flags = 0;
+
+	return ext4_getfsmap_helper(sb, info, &irec);
+}
+
+/* Helper to fill out an ext4_fsmap. */
+static inline int ext4_getfsmap_fill(struct list_head *meta_list,
+				     ext4_fsblk_t fsb, ext4_fsblk_t len,
+				     uint64_t owner)
+{
+	struct ext4_fsmap *fsm;
+
+	fsm = kmalloc(sizeof(*fsm), GFP_NOFS);
+	if (!fsm)
+		return -ENOMEM;
+	fsm->fmr_device = 0;
+	fsm->fmr_flags = 0;
+	fsm->fmr_physical = fsb;
+	fsm->fmr_owner = owner;
+	fsm->fmr_length = len;
+	list_add_tail(&fsm->fmr_list, meta_list);
+
+	return 0;
+}
+
+/*
+ * This function returns the number of file system metadata blocks at
+ * the beginning of a block group, including the reserved gdt blocks.
+ */
+static unsigned int ext4_getfsmap_find_sb(struct super_block *sb,
+					  ext4_group_t agno,
+					  struct list_head *meta_list)
+{
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	ext4_fsblk_t fsb = ext4_group_first_block_no(sb, agno);
+	ext4_fsblk_t len;
+	unsigned long first_meta_bg = le32_to_cpu(sbi->s_es->s_first_meta_bg);
+	unsigned long metagroup = agno / EXT4_DESC_PER_BLOCK(sb);
+	int error;
+
+	/* Record the superblock. */
+	if (ext4_bg_has_super(sb, agno)) {
+		error = ext4_getfsmap_fill(meta_list, fsb, 1, EXT4_FMR_OWN_FS);
+		if (error)
+			return error;
+		fsb++;
+	}
+
+	/* Record the group descriptors. */
+	len = ext4_bg_num_gdb(sb, agno);
+	if (!len)
+		return 0;
+	error = ext4_getfsmap_fill(meta_list, fsb, len,
+				   EXT4_FMR_OWN_GDT);
+	if (error)
+		return error;
+	fsb += len;
+
+	/* Reserved GDT blocks */
+	if (!ext4_has_feature_meta_bg(sb) || metagroup < first_meta_bg) {
+		len = le16_to_cpu(sbi->s_es->s_reserved_gdt_blocks);
+		error = ext4_getfsmap_fill(meta_list, fsb, len,
+					   EXT4_FMR_OWN_RESV_GDT);
+		if (error)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Compare two fsmap items. */
+static int ext4_getfsmap_compare(void *priv,
+				 struct list_head *a,
+				 struct list_head *b)
+{
+	struct ext4_fsmap *fa;
+	struct ext4_fsmap *fb;
+
+	fa = container_of(a, struct ext4_fsmap, fmr_list);
+	fb = container_of(b, struct ext4_fsmap, fmr_list);
+	if (fa->fmr_physical < fb->fmr_physical)
+		return -1;
+	else if (fa->fmr_physical > fb->fmr_physical)
+		return 1;
+	return 0;
+}
+
+/* Merge adjacent extents of fixed metadata. */
+static void ext4_getfsmap_merge_fixed_metadata(struct list_head *meta_list)
+{
+	struct ext4_fsmap *p;
+	struct ext4_fsmap *prev = NULL;
+	struct ext4_fsmap *tmp;
+
+	list_for_each_entry_safe(p, tmp, meta_list, fmr_list) {
+		if (!prev) {
+			prev = p;
+			continue;
+		}
+
+		if (prev->fmr_owner == p->fmr_owner &&
+		    prev->fmr_physical + prev->fmr_length == p->fmr_physical) {
+			prev->fmr_length += p->fmr_length;
+			list_del(&p->fmr_list);
+			kfree(p);
+		} else
+			prev = p;
+	}
+}
+
+/* Free a list of fixed metadata. */
+static void ext4_getfsmap_free_fixed_metadata(struct list_head *meta_list)
+{
+	struct ext4_fsmap *p;
+	struct ext4_fsmap *tmp;
+
+	list_for_each_entry_safe(p, tmp, meta_list, fmr_list) {
+		list_del(&p->fmr_list);
+		kfree(p);
+	}
+}
+
+/* Find all the fixed metadata in the filesystem. */
+int ext4_getfsmap_find_fixed_metadata(struct super_block *sb,
+				      struct list_head *meta_list)
+{
+	struct ext4_group_desc *gdp;
+	ext4_group_t agno;
+	int error;
+
+	INIT_LIST_HEAD(meta_list);
+
+	/* Collect everything. */
+	for (agno = 0; agno < EXT4_SB(sb)->s_groups_count; agno++) {
+		gdp = ext4_get_group_desc(sb, agno, NULL);
+		if (!gdp) {
+			error = -EFSCORRUPTED;
+			goto err;
+		}
+
+		/* Superblock & GDT */
+		error = ext4_getfsmap_find_sb(sb, agno, meta_list);
+		if (error)
+			goto err;
+
+		/* Block bitmap */
+		error = ext4_getfsmap_fill(meta_list,
+					   ext4_block_bitmap(sb, gdp), 1,
+					   EXT4_FMR_OWN_BLKBM);
+		if (error)
+			goto err;
+
+		/* Inode bitmap */
+		error = ext4_getfsmap_fill(meta_list,
+					   ext4_inode_bitmap(sb, gdp), 1,
+					   EXT4_FMR_OWN_INOBM);
+		if (error)
+			goto err;
+
+		/* Inodes */
+		error = ext4_getfsmap_fill(meta_list,
+					   ext4_inode_table(sb, gdp),
+					   EXT4_SB(sb)->s_itb_per_group,
+					   EXT4_FMR_OWN_INODES);
+		if (error)
+			goto err;
+	}
+
+	/* Sort the list */
+	list_sort(NULL, meta_list, ext4_getfsmap_compare);
+
+	/* Merge adjacent extents */
+	ext4_getfsmap_merge_fixed_metadata(meta_list);
+
+	return 0;
+err:
+	ext4_getfsmap_free_fixed_metadata(meta_list);
+	return error;
+}
+
+/* Execute a getfsmap query against the buddy bitmaps */
+static int ext4_getfsmap_datadev(struct super_block *sb,
+				 struct ext4_fsmap *keys,
+				 struct ext4_getfsmap_info *info)
+{
+	struct ext4_sb_info *sbi = EXT4_SB(sb);
+	ext4_fsblk_t start_fsb;
+	ext4_fsblk_t end_fsb;
+	ext4_fsblk_t eofs;
+	ext4_group_t start_ag;
+	ext4_group_t end_ag;
+	ext4_grpblk_t first_cluster;
+	ext4_grpblk_t last_cluster;
+	int error = 0;
+
+	eofs = ext4_blocks_count(sbi->s_es);
+	if (keys[0].fmr_physical >= eofs)
+		return 0;
+	if (keys[1].fmr_physical >= eofs)
+		keys[1].fmr_physical = eofs - 1;
+	start_fsb = keys[0].fmr_physical;
+	end_fsb = keys[1].fmr_physical;
+
+	/* Determine first and last group to examine based on start and end */
+	ext4_get_group_no_and_offset(sb, start_fsb, &start_ag, &first_cluster);
+	ext4_get_group_no_and_offset(sb, end_fsb, &end_ag, &last_cluster);
+
+	/*
+	 * Convert the fsmap low/high keys to bg based keys.  Initialize
+	 * low to the fsmap low key and max out the high key to the end
+	 * of the bg.
+	 */
+	info->gfi_low = keys[0];
+	info->gfi_low.fmr_physical = EXT4_C2B(sbi, first_cluster);
+	info->gfi_low.fmr_length = 0;
+
+	memset(&info->gfi_high, 0xFF, sizeof(info->gfi_high));
+
+	/* Assemble a list of all the fixed-location metadata. */
+	error = ext4_getfsmap_find_fixed_metadata(sb, &info->gfi_meta_list);
+	if (error)
+		goto err;
+
+	/* Query each bg */
+	for (info->gfi_agno = start_ag;
+	     info->gfi_agno <= end_ag;
+	     info->gfi_agno++) {
+		/*
+		 * Set the bg high key from the fsmap high key if this
+		 * is the last bg that we're querying.
+		 */
+		if (info->gfi_agno == end_ag) {
+			info->gfi_high = keys[1];
+			info->gfi_high.fmr_physical = EXT4_C2B(sbi,
+					last_cluster);
+			info->gfi_high.fmr_length = 0;
+		}
+
+		trace_ext4_fsmap_low_key(sb, info->gfi_dev, info->gfi_agno,
+				info->gfi_low.fmr_physical,
+				info->gfi_low.fmr_length,
+				info->gfi_low.fmr_owner);
+
+		trace_ext4_fsmap_high_key(sb, info->gfi_dev, info->gfi_agno,
+				info->gfi_high.fmr_physical,
+				info->gfi_high.fmr_length,
+				info->gfi_high.fmr_owner);
+
+		error = ext4_mballoc_query_range(sb, info->gfi_agno,
+				EXT4_B2C(sbi, info->gfi_low.fmr_physical),
+				EXT4_B2C(sbi, info->gfi_high.fmr_physical),
+				ext4_getfsmap_datadev_helper, info);
+		if (error)
+			goto err;
+
+		/*
+		 * Set the bg low key to the start of the bg prior to
+		 * moving on to the next bg.
+		 */
+		if (info->gfi_agno == start_ag)
+			memset(&info->gfi_low, 0, sizeof(info->gfi_low));
+	}
+
+	/* Do we have a retained free extent? */
+	if (info->gfi_lastfree.fmr_owner) {
+		error = ext4_getfsmap_helper(sb, info, &info->gfi_lastfree);
+		if (error)
+			goto err;
+	}
+
+	/* Report any gaps at the end of the bg */
+	info->gfi_last = true;
+	error = ext4_getfsmap_datadev_helper(sb, end_ag, last_cluster, 0, info);
+	if (error)
+		goto err;
+
+err:
+	ext4_getfsmap_free_fixed_metadata(&info->gfi_meta_list);
+	return error;
+}
+
+/* Do we recognize the device? */
+static bool ext4_getfsmap_is_valid_device(struct super_block *sb,
+					  struct ext4_fsmap *fm)
+{
+	if (fm->fmr_device == 0 || fm->fmr_device == UINT_MAX ||
+	    fm->fmr_device == new_encode_dev(sb->s_bdev->bd_dev))
+		return true;
+	if (EXT4_SB(sb)->journal_bdev &&
+	    fm->fmr_device == new_encode_dev(EXT4_SB(sb)->journal_bdev->bd_dev))
+		return true;
+	return false;
+}
+
+/* Ensure that the low key is less than the high key. */
+static bool ext4_getfsmap_check_keys(struct ext4_fsmap *low_key,
+				     struct ext4_fsmap *high_key)
+{
+	if (low_key->fmr_device > high_key->fmr_device)
+		return false;
+	if (low_key->fmr_device < high_key->fmr_device)
+		return true;
+
+	if (low_key->fmr_physical > high_key->fmr_physical)
+		return false;
+	if (low_key->fmr_physical < high_key->fmr_physical)
+		return true;
+
+	if (low_key->fmr_owner > high_key->fmr_owner)
+		return false;
+	if (low_key->fmr_owner < high_key->fmr_owner)
+		return true;
+
+	return false;
+}
+
+#define EXT4_GETFSMAP_DEVS	2
+/*
+ * Get filesystem's extents as described in head, and format for
+ * output.  Calls formatter to fill the user's buffer until all
+ * extents are mapped, until the passed-in head->fmh_count slots have
+ * been filled, or until the formatter short-circuits the loop, if it
+ * is tracking filled-in extents on its own.
+ *
+ * Key to Confusion
+ * ----------------
+ * There are multiple levels of keys and counters at work here:
+ * _fsmap_head.fmh_keys		-- low and high fsmap keys passed in;
+ * 				   these reflect fs-wide block addrs.
+ * dkeys			-- fmh_keys used to query each device;
+ * 				   these are fmh_keys but w/ the low key
+ * 				   bumped up by fmr_length.
+ * _getfsmap_info.gfi_next_fsblk-- next fs block we expect to see; this
+ *				   is how we detect gaps in the fsmap
+ *				   records and report them.
+ * _getfsmap_info.gfi_low/high	-- per-bg low/high keys computed from
+ * 				   dkeys; used to query the free space.
+ */
+int ext4_getfsmap(struct super_block *sb, struct ext4_fsmap_head *head,
+		  ext4_fsmap_format_t formatter, void *arg)
+{
+	struct ext4_fsmap dkeys[2];	/* per-dev keys */
+	struct ext4_getfsmap_dev handlers[EXT4_GETFSMAP_DEVS];
+	struct ext4_getfsmap_info info = {0};
+	int i;
+	int error = 0;
+
+	if (head->fmh_iflags & ~FMH_IF_VALID)
+		return -EINVAL;
+	if (!ext4_getfsmap_is_valid_device(sb, &head->fmh_keys[0]) ||
+	    !ext4_getfsmap_is_valid_device(sb, &head->fmh_keys[1]))
+		return -EINVAL;
+
+	head->fmh_entries = 0;
+
+	/* Set up our device handlers. */
+	memset(handlers, 0, sizeof(handlers));
+	handlers[0].gfd_dev = new_encode_dev(sb->s_bdev->bd_dev);
+	handlers[0].gfd_fn = ext4_getfsmap_datadev;
+	if (EXT4_SB(sb)->journal_bdev) {
+		handlers[1].gfd_dev = new_encode_dev(
+				EXT4_SB(sb)->journal_bdev->bd_dev);
+		handlers[1].gfd_fn = ext4_getfsmap_logdev;
+	}
+
+	sort(handlers, EXT4_GETFSMAP_DEVS, sizeof(struct ext4_getfsmap_dev),
+			ext4_getfsmap_dev_compare, NULL);
+
+	/*
+	 * To continue where we left off, we allow userspace to use the
+	 * last mapping from a previous call as the low key of the next.
+	 * This is identified by a non-zero length in the low key. We
+	 * have to increment the low key in this scenario to ensure we
+	 * don't return the same mapping again, and instead return the
+	 * very next mapping.
+	 *
+	 * Bump the physical offset as there can be no other mapping for
+	 * the same physical block range.
+	 */
+	dkeys[0] = head->fmh_keys[0];
+	dkeys[0].fmr_physical += dkeys[0].fmr_length;
+	dkeys[0].fmr_owner = 0;
+	dkeys[0].fmr_length = 0;
+	memset(&dkeys[1], 0xFF, sizeof(struct ext4_fsmap));
+
+	if (!ext4_getfsmap_check_keys(dkeys, &head->fmh_keys[1]))
+		return -EINVAL;
+
+	info.gfi_next_fsblk = head->fmh_keys[0].fmr_physical +
+			  head->fmh_keys[0].fmr_length;
+	info.gfi_formatter = formatter;
+	info.gfi_format_arg = arg;
+	info.gfi_head = head;
+
+	/* For each device we support... */
+	for (i = 0; i < EXT4_GETFSMAP_DEVS; i++) {
+		/* Is this device within the range the user asked for? */
+		if (!handlers[i].gfd_fn)
+			continue;
+		if (head->fmh_keys[0].fmr_device > handlers[i].gfd_dev)
+			continue;
+		if (head->fmh_keys[1].fmr_device < handlers[i].gfd_dev)
+			break;
+
+		/*
+		 * If this device number matches the high key, we have
+		 * to pass the high key to the handler to limit the
+		 * query results.  If the device number exceeds the
+		 * low key, zero out the low key so that we get
+		 * everything from the beginning.
+		 */
+		if (handlers[i].gfd_dev == head->fmh_keys[1].fmr_device)
+			dkeys[1] = head->fmh_keys[1];
+		if (handlers[i].gfd_dev > head->fmh_keys[0].fmr_device)
+			memset(&dkeys[0], 0, sizeof(struct ext4_fsmap));
+
+		info.gfi_dev = handlers[i].gfd_dev;
+		info.gfi_last = false;
+		info.gfi_agno = -1;
+		error = handlers[i].gfd_fn(sb, dkeys, &info);
+		if (error)
+			break;
+		info.gfi_next_fsblk = 0;
+	}
+
+	head->fmh_oflags = FMH_OF_DEV_T;
+	return error;
+}
diff --git a/fs/ext4/fsmap.h b/fs/ext4/fsmap.h
new file mode 100644
index 000000000000..9a2cd367cc66
--- /dev/null
+++ b/fs/ext4/fsmap.h
@@ -0,0 +1,69 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __EXT4_FSMAP_H__
+#define	__EXT4_FSMAP_H__
+
+struct fsmap;
+
+/* internal fsmap representation */
+struct ext4_fsmap {
+	struct list_head	fmr_list;
+	dev_t		fmr_device;	/* device id */
+	uint32_t	fmr_flags;	/* mapping flags */
+	uint64_t	fmr_physical;	/* device offset of segment */
+	uint64_t	fmr_owner;	/* owner id */
+	uint64_t	fmr_length;	/* length of segment, blocks */
+};
+
+struct ext4_fsmap_head {
+	uint32_t	fmh_iflags;	/* control flags */
+	uint32_t	fmh_oflags;	/* output flags */
+	unsigned int	fmh_count;	/* # of entries in array incl. input */
+	unsigned int	fmh_entries;	/* # of entries filled in (output). */
+
+	struct ext4_fsmap fmh_keys[2];	/* low and high keys */
+};
+
+void ext4_fsmap_from_internal(struct super_block *sb, struct fsmap *dest,
+		struct ext4_fsmap *src);
+void ext4_fsmap_to_internal(struct super_block *sb, struct ext4_fsmap *dest,
+		struct fsmap *src);
+
+/* fsmap to userspace formatter - copy to user & advance pointer */
+typedef int (*ext4_fsmap_format_t)(struct ext4_fsmap *, void *);
+
+int ext4_getfsmap(struct super_block *sb, struct ext4_fsmap_head *head,
+		ext4_fsmap_format_t formatter, void *arg);
+
+#define EXT4_QUERY_RANGE_ABORT		1
+#define EXT4_QUERY_RANGE_CONTINUE	0
+
+/*	fmr_owner special values for FS_IOC_GETFSMAP; some share w/ XFS */
+#define EXT4_FMR_OWN_FREE	FMR_OWN_FREE      /* free space */
+#define EXT4_FMR_OWN_UNKNOWN	FMR_OWN_UNKNOWN   /* unknown owner */
+#define EXT4_FMR_OWN_FS		FMR_OWNER('X', 1) /* static fs metadata */
+#define EXT4_FMR_OWN_LOG	FMR_OWNER('X', 2) /* journalling log */
+#define EXT4_FMR_OWN_INODES	FMR_OWNER('X', 5) /* inodes */
+#define EXT4_FMR_OWN_GDT	FMR_OWNER('f', 1) /* group descriptors */
+#define EXT4_FMR_OWN_RESV_GDT	FMR_OWNER('f', 2) /* reserved gdt blocks */
+#define EXT4_FMR_OWN_BLKBM	FMR_OWNER('f', 3) /* inode bitmap */
+#define EXT4_FMR_OWN_INOBM	FMR_OWNER('f', 4) /* block bitmap */
+
+#endif /* __EXT4_FSMAP_H__ */
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 17bc043308f3..98ac2f1f23b3 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -1098,6 +1098,17 @@ got:
 	if (err)
 		goto fail_drop;
 
+	/*
+	 * Since the encryption xattr will always be unique, create it first so
+	 * that it's less likely to end up in an external xattr block and
+	 * prevent its deduplication.
+	 */
+	if (encrypt) {
+		err = fscrypt_inherit_context(dir, inode, handle, true);
+		if (err)
+			goto fail_free_drop;
+	}
+
 	err = ext4_init_acl(handle, inode, dir);
 	if (err)
 		goto fail_free_drop;
@@ -1119,12 +1130,6 @@ got:
 		ei->i_datasync_tid = handle->h_transaction->t_tid;
 	}
 
-	if (encrypt) {
-		err = fscrypt_inherit_context(dir, inode, handle, true);
-		if (err)
-			goto fail_free_drop;
-	}
-
 	err = ext4_mark_inode_dirty(handle, inode);
 	if (err) {
 		ext4_std_error(sb, err);
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 375fb1c05d49..d5dea4c293ef 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1034,7 +1034,7 @@ static int ext4_add_dirent_to_inline(handle_t *handle,
 	err = ext4_journal_get_write_access(handle, iloc->bh);
 	if (err)
 		return err;
-	ext4_insert_dentry(dir, inode, de, inline_size, fname);
+	ext4_insert_dentry(inode, de, inline_size, fname);
 
 	ext4_show_inline_dir(dir, iloc->bh, inline_start, inline_size);
 
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index f0729b0705c7..5834c4d76be8 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1643,6 +1643,7 @@ struct mpage_da_data {
 	 */
 	struct ext4_map_blocks map;
 	struct ext4_io_submit io_submit;	/* IO submission data */
+	unsigned int do_map:1;
 };
 
 static void mpage_release_unused_pages(struct mpage_da_data *mpd,
@@ -2179,6 +2180,9 @@ static bool mpage_add_bh_to_extent(struct mpage_da_data *mpd, ext4_lblk_t lblk,
 
 	/* First block in the extent? */
 	if (map->m_len == 0) {
+		/* We cannot map unless handle is started... */
+		if (!mpd->do_map)
+			return false;
 		map->m_lblk = lblk;
 		map->m_len = 1;
 		map->m_flags = bh->b_state & BH_FLAGS;
@@ -2231,6 +2235,9 @@ static int mpage_process_page_bufs(struct mpage_da_data *mpd,
 			/* Found extent to map? */
 			if (mpd->map.m_len)
 				return 0;
+			/* Buffer needs mapping and handle is not started? */
+			if (!mpd->do_map)
+				return 0;
 			/* Everything mapped so far and we hit EOF */
 			break;
 		}
@@ -2747,6 +2754,29 @@ retry:
 		tag_pages_for_writeback(mapping, mpd.first_page, mpd.last_page);
 	done = false;
 	blk_start_plug(&plug);
+
+	/*
+	 * First writeback pages that don't need mapping - we can avoid
+	 * starting a transaction unnecessarily and also avoid being blocked
+	 * in the block layer on device congestion while having transaction
+	 * started.
+	 */
+	mpd.do_map = 0;
+	mpd