summaryrefslogtreecommitdiff
path: root/fs/ext4
diff options
context:
space:
mode:
Diffstat (limited to 'fs/ext4')
-rw-r--r--fs/ext4/Makefile10
-rw-r--r--fs/ext4/fsmap.c722
-rw-r--r--fs/ext4/fsmap.h69
-rw-r--r--fs/ext4/ioctl.c90
-rw-r--r--fs/ext4/mballoc.c49
-rw-r--r--fs/ext4/mballoc.h17
-rw-r--r--fs/ext4/super.c1
7 files changed, 953 insertions, 5 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/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/ioctl.c b/fs/ext4/ioctl.c
index a4273ddb9922..dfcb815275f1 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -19,6 +19,9 @@
#include <linux/delay.h>
#include "ext4_jbd2.h"
#include "ext4.h"
+#include <linux/fsmap.h>
+#include "fsmap.h"
+#include <trace/events/ext4.h>
/**
* Swap memory between @a and @b for @len bytes.
@@ -489,6 +492,90 @@ int ext4_shutdown(struct super_block *sb, unsigned long arg)
return 0;
}
+struct getfsmap_info {
+ struct super_block *gi_sb;
+ struct fsmap_head __user *gi_data;
+ unsigned int gi_idx;
+ __u32 gi_last_flags;
+};
+
+static int ext4_getfsmap_format(struct ext4_fsmap *xfm, void *priv)
+{
+ struct getfsmap_info *info = priv;
+ struct fsmap fm;
+
+ trace_ext4_getfsmap_mapping(info->gi_sb, xfm);
+
+ info->gi_last_flags = xfm->fmr_flags;
+ ext4_fsmap_from_internal(info->gi_sb, &fm, xfm);
+ if (copy_to_user(&info->gi_data->fmh_recs[info->gi_idx++], &fm,
+ sizeof(struct fsmap)))
+ return -EFAULT;
+
+ return 0;
+}
+
+static int ext4_ioc_getfsmap(struct super_block *sb,
+ struct fsmap_head __user *arg)
+{
+ struct getfsmap_info info = {0};
+ struct ext4_fsmap_head xhead = {0};
+ struct fsmap_head head;
+ bool aborted = false;
+ int error;
+
+ if (copy_from_user(&head, arg, sizeof(struct fsmap_head)))
+ return -EFAULT;
+ if (memchr_inv(head.fmh_reserved, 0, sizeof(head.fmh_reserved)) ||
+ memchr_inv(head.fmh_keys[0].fmr_reserved, 0,
+ sizeof(head.fmh_keys[0].fmr_reserved)) ||
+ memchr_inv(head.fmh_keys[1].fmr_reserved, 0,
+ sizeof(head.fmh_keys[1].fmr_reserved)))
+ return -EINVAL;
+ /*
+ * ext4 doesn't report file extents at all, so the only valid
+ * file offsets are the magic ones (all zeroes or all ones).
+ */
+ if (head.fmh_keys[0].fmr_offset ||
+ (head.fmh_keys[1].fmr_offset != 0 &&
+ head.fmh_keys[1].fmr_offset != -1ULL))
+ return -EINVAL;
+
+ xhead.fmh_iflags = head.fmh_iflags;
+ xhead.fmh_count = head.fmh_count;
+ ext4_fsmap_to_internal(sb, &xhead.fmh_keys[0], &head.fmh_keys[0]);
+ ext4_fsmap_to_internal(sb, &xhead.fmh_keys[1], &head.fmh_keys[1]);
+
+ trace_ext4_getfsmap_low_key(sb, &xhead.fmh_keys[0]);
+ trace_ext4_getfsmap_high_key(sb, &xhead.fmh_keys[1]);
+
+ info.gi_sb = sb;
+ info.gi_data = arg;
+ error = ext4_getfsmap(sb, &xhead, ext4_getfsmap_format, &info);
+ if (error == EXT4_QUERY_RANGE_ABORT) {
+ error = 0;
+ aborted = true;
+ } else if (error)
+ return error;
+
+ /* If we didn't abort, set the "last" flag in the last fmx */
+ if (!aborted && info.gi_idx) {
+ info.gi_last_flags |= FMR_OF_LAST;
+ if (copy_to_user(&info.gi_data->fmh_recs[info.gi_idx - 1].fmr_flags,
+ &info.gi_last_flags,
+ sizeof(info.gi_last_flags)))
+ return -EFAULT;
+ }
+
+ /* copy back header */
+ head.fmh_entries = xhead.fmh_entries;
+ head.fmh_oflags = xhead.fmh_oflags;
+ if (copy_to_user(arg, &head, sizeof(struct fsmap_head)))
+ return -EFAULT;
+
+ return 0;
+}
+
long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
{
struct inode *inode = file_inode(filp);
@@ -499,6 +586,8 @@ long ext4_ioctl(struct file *filp, unsigned int cmd, unsigned long arg)
ext4_debug("cmd = %u, arg = %lu\n", cmd, arg);
switch (cmd) {
+ case FS_IOC_GETFSMAP:
+ return ext4_ioc_getfsmap(sb, (void __user *)arg);
case EXT4_IOC_GETFLAGS:
ext4_get_inode_flags(ei);
flags = ei->i_flags & EXT4_FL_USER_VISIBLE;
@@ -1009,6 +1098,7 @@ long ext4_compat_ioctl(struct file *file, unsigned int cmd, unsigned long arg)
case EXT4_IOC_GET_ENCRYPTION_PWSALT:
case EXT4_IOC_GET_ENCRYPTION_POLICY:
case EXT4_IOC_SHUTDOWN:
+ case FS_IOC_GETFSMAP:
break;
default:
return -ENOIOCTLCMD;
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index dbe51301c2bb..36de58a37653 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -5277,3 +5277,52 @@ out:
range->len = EXT4_C2B(EXT4_SB(sb), trimmed) << sb->s_blocksize_bits;
return ret;
}
+
+/* Iterate all the free extents in the group. */
+int
+ext4_mballoc_query_range(
+ struct super_block *sb,
+ ext4_group_t group,
+ ext4_grpblk_t start,
+ ext4_grpblk_t end,
+ ext4_mballoc_query_range_fn formatter,
+ void *priv)
+{
+ void *bitmap;
+ ext4_grpblk_t next;
+ struct ext4_buddy e4b;
+ int error;
+
+ error = ext4_mb_load_buddy(sb, group, &e4b);
+ if (error)
+ return error;
+ bitmap = e4b.bd_bitmap;
+
+ ext4_lock_group(sb, group);
+
+ start = (e4b.bd_info->bb_first_free > start) ?
+ e4b.bd_info->bb_first_free : start;
+ if (end >= EXT4_CLUSTERS_PER_GROUP(sb))
+ end = EXT4_CLUSTERS_PER_GROUP(sb) - 1;
+
+ while (start <= end) {
+ start = mb_find_next_zero_bit(bitmap, end + 1, start);
+ if (start > end)
+ break;
+ next = mb_find_next_bit(bitmap, end + 1, start);
+
+ ext4_unlock_group(sb, group);
+ error = formatter(sb, group, start, next - start, priv);
+ if (error)
+ goto out_unload;
+ ext4_lock_group(sb, group);
+
+ start = next + 1;
+ }
+
+ ext4_unlock_group(sb, group);
+out_unload:
+ ext4_mb_unload_buddy(&e4b);
+
+ return error;
+}
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index 1aba469f8220..2bed62084a8c 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -199,4 +199,21 @@ static inline ext4_fsblk_t ext4_grp_offs_to_block(struct super_block *sb,
return ext4_group_first_block_no(sb, fex->fe_group) +
(fex->fe_start << EXT4_SB(sb)->s_cluster_bits);
}
+
+typedef int (*ext4_mballoc_query_range_fn)(
+ struct super_block *sb,
+ ext4_group_t agno,
+ ext4_grpblk_t start,
+ ext4_grpblk_t len,
+ void *priv);
+
+int
+ext4_mballoc_query_range(
+ struct super_block *sb,
+ ext4_group_t agno,
+ ext4_grpblk_t start,
+ ext4_grpblk_t end,
+ ext4_mballoc_query_range_fn formatter,
+ void *priv);
+
#endif
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 7b82487401d2..9ec8963ba604 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -49,6 +49,7 @@
#include "xattr.h"
#include "acl.h"
#include "mballoc.h"
+#include "fsmap.h"
#define CREATE_TRACE_POINTS
#include <trace/events/ext4.h>