ext4: online defrag -- Add the EXT4_IOC_FREE_BLOCKS_INFO ioctl.
From: Akira Fujita <a-fujita@xxxxxxxxxxxxx>
The EXT4_IOC_FREE_BLOCKS_INFO ioctl gets free extents
information of the target block group.
This ioctl is used only in the force defrag (-f).
Defragger calculates the victim extents (which will be moved
into other block group in order to make contiguous free space for the target file)
based on used/unused extents information.
Signed-off-by: Akira Fujita <a-fujita@xxxxxxxxxxxxx>
Signed-off-by: Takashi Sato <t-sato@xxxxxxxxxxxxx>
---
fs/ext4/balloc.c | 2 +-
fs/ext4/defrag.c | 112 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/ext4.h | 27 +++++++++++++
fs/ext4/ioctl.c | 3 +-
4 files changed, 142 insertions(+), 2 deletions(-)
diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 02068fa..64ec04c 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -919,7 +919,7 @@ static int ext4_test_allocatable(ext4_grpblk_t nr, struct buffer_head *bh)
* bitmap on disk and the last-committed copy in journal, until we find a
* bit free in both bitmaps.
*/
-static ext4_grpblk_t
+ext4_grpblk_t
bitmap_search_next_usable_block(ext4_grpblk_t start, struct buffer_head *bh,
ext4_grpblk_t maxblocks)
{
diff --git a/fs/ext4/defrag.c b/fs/ext4/defrag.c
index cb7d237..941414b 100644
--- a/fs/ext4/defrag.c
+++ b/fs/ext4/defrag.c
@@ -90,6 +90,102 @@ err:
return -EIO;
}
+/**
+ * ext4_defrag_fblocks_distribution - Search free blocks distribution
+ *
+ * @org_inode: original inode
+ * @ext_info: ext4_extents_info
+ *
+ * This function returns 0 if succeed, otherwise returns error value.
+ */
+static int
+ext4_defrag_fblocks_distribution(struct inode *org_inode,
+ struct ext4_extents_info *ext_info)
+{
+ struct buffer_head *bitmap_bh = NULL;
+ struct super_block *sb = org_inode->i_sb;
+ handle_t *handle;
+ ext4_group_t group_no;
+ ext4_grpblk_t start, end;
+ ext4_fsblk_t start_block = 0;
+ int i, err;
+ int num = 0;
+ int len = 0;
+ int block_set = 0;
+ int extra_block = 0;
+
+ if (!sb) {
+ printk(KERN_ERR "ext4 defrag: Non-existent device\n");
+ return -ENOSPC;
+ }
+
+ group_no = (org_inode->i_ino - 1) / EXT4_INODES_PER_GROUP(sb);
+ start = ext_info->g_offset;
+ end = EXT4_BLOCKS_PER_GROUP(sb) - 1;
+
+ /* We consider about the boot block if bs = 1k */
+ if (sb->s_blocksize == 1024)
+ extra_block = 1;
+
+ handle = ext4_journal_start(org_inode, 1);
+ if (IS_ERR(handle)) {
+ err = PTR_ERR(handle);
+ return err;
+ }
+
+ bitmap_bh = ext4_read_block_bitmap(sb, group_no);
+ if (!bitmap_bh) {
+ err = -EIO;
+ goto out;
+ }
+
+ BUFFER_TRACE(bitmap_bh, "get undo access for new block");
+ err = ext4_journal_get_undo_access(handle, bitmap_bh);
+ if (err)
+ goto out;
+
+ for (i = start; i <= end ; i++) {
+ if (bitmap_search_next_usable_block(i, bitmap_bh, i + 1) >= 0) {
+ len++;
+ /*
+ * Reset start_block if the free block is
+ * the head of region.
+ */
+ if (!block_set) {
+ start_block =
+ i + group_no * EXT4_BLOCKS_PER_GROUP(sb) +
+ extra_block;
+ block_set = 1;
+ }
+ } else if (len) {
+ ext_info->ext[num].start = start_block;
+ ext_info->ext[num].len = len;
+ num++;
+ len = 0;
+ block_set = 0;
+ if (num == ext_info->max_entries) {
+ ext_info->g_offset = i + 1;
+ break;
+ }
+ }
+ if (i == end && len) {
+ ext_info->ext[num].start = start_block;
+ ext_info->ext[num].len = len;
+ num++;
+ }
+ }
+
+ ext_info->entries = num;
+out:
+ ext4_journal_release_buffer(handle, bitmap_bh);
+ brelse(bitmap_bh);
+
+ if (handle)
+ ext4_journal_stop(handle);
+
+ return err;
+}
+
int ext4_defrag_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
unsigned long arg)
{
@@ -125,6 +221,22 @@ int ext4_defrag_ioctl(struct inode *inode, struct file *filp, unsigned int cmd,
if (copy_to_user((struct ext4_group_data_info __user *)arg,
&grp_data, sizeof(grp_data)))
return -EFAULT;
+ } else if (cmd == EXT4_IOC_FREE_BLOCKS_INFO) {
+ struct ext4_extents_info ext_info;
+
+ if (copy_from_user(&ext_info,
+ (struct ext4_extents_info __user *)arg,
+ sizeof(ext_info)))
+ return -EFAULT;
+
+ BUG_ON(ext_info.ino != inode->i_ino);
+
+ err = ext4_defrag_fblocks_distribution(inode, &ext_info);
+
+ if (!err)
+ err = copy_to_user(
+ (struct ext4_extents_info __user *)arg,
+ &ext_info, sizeof(ext_info));
} else if (cmd == EXT4_IOC_DEFRAG) {
struct ext4_ext_defrag_data defrag;
struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es;
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index a635a76..ffe687b 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -305,6 +305,7 @@ struct ext4_new_group_data {
#define EXT4_IOC_DEFRAG _IOW('f', 15, struct ext4_ext_defrag_data)
#define EXT4_IOC_FIBMAP _IOW('f', 16, ext4_fsblk_t)
#define EXT4_IOC_GROUP_INFO _IOW('f', 17, struct ext4_group_data_info)
+#define EXT4_IOC_FREE_BLOCKS_INFO _IOW('f', 18, struct ext4_extents_info)
/*
* ioctl commands in 32 bit emulation
@@ -322,6 +323,20 @@ struct ext4_new_group_data {
#define EXT4_IOC32_GETVERSION_OLD FS_IOC32_GETVERSION
#define EXT4_IOC32_SETVERSION_OLD FS_IOC32_SETVERSION
+/*
+ * The following four macros are used for the defrag force mode.
+ *
+ * DEFRAG_MAX_ENT: the maximum number of extents for exchanging between
+ * kernel-space and user-space per an ioctl
+ */
+#define DEFRAG_MAX_ENT 32
+
+struct ext4_extent_data {
+ ext4_lblk_t block; /* start logical block number */
+ ext4_fsblk_t start; /* start physical block number */
+ int len; /* blocks count */
+};
+
struct ext4_ext_defrag_data {
ext4_lblk_t start_offset; /* start offset to defrag in blocks */
ext4_lblk_t defrag_size; /* size of defrag in blocks */
@@ -333,6 +348,16 @@ struct ext4_group_data_info {
int s_inodes_per_group; /* inodes per group */
};
+struct ext4_extents_info {
+ unsigned long long ino; /* inode number */
+ int max_entries; /* maximum extents count */
+ int entries; /* extent number/count */
+ ext4_lblk_t f_offset; /* file offset */
+ ext4_grpblk_t g_offset; /* group offset */
+ ext4_fsblk_t goal; /* block offset for allocation */
+ struct ext4_extent_data ext[DEFRAG_MAX_ENT];
+};
+
#define EXT4_TRANS_META_BLOCKS 4 /* bitmap + group desc + sb + inode */
/*
@@ -1015,6 +1040,8 @@ extern struct ext4_group_desc * ext4_get_group_desc(struct super_block * sb,
extern int ext4_should_retry_alloc(struct super_block *sb, int *retries);
extern void ext4_init_block_alloc_info(struct inode *);
extern void ext4_rsv_window_add(struct super_block *sb, struct ext4_reserve_window_node *rsv);
+extern ext4_grpblk_t bitmap_search_next_usable_block(ext4_grpblk_t,
+ struct buffer_head *, ext4_grpblk_t);
/* dir.c */
extern int ext4_check_dir_entry(const char *, struct inode *,
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index c05502f..40bc2e1 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -258,7 +258,8 @@ setversion_out:
}
case EXT4_IOC_FIBMAP:
case EXT4_IOC_DEFRAG:
- case EXT4_IOC_GROUP_INFO: {
+ case EXT4_IOC_GROUP_INFO:
+ case EXT4_IOC_FREE_BLOCKS_INFO: {
return ext4_defrag_ioctl(inode, filp, cmd, arg);
}
case EXT4_IOC_GROUP_ADD: {
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
