korg/linux
0
0
Fork 0
mirror of https://mirrors.bfsu.edu.cn/git/linux.git synced 2024-11-11 12:28:41 +08:00
Code Issues Packages Projects Releases Wiki Activity

f2fs: enhance on-disk inode structure scalability

Browse Source 
This patch add new flag F2FS_EXTRA_ATTR storing in inode.i_inline
to indicate that on-disk structure of current inode is extended.

In order to extend, we changed the inode structure a bit:

Original one:

struct f2fs_inode {
	...
	struct f2fs_extent i_ext;
	__le32 i_addr[DEF_ADDRS_PER_INODE];
	__le32 i_nid[DEF_NIDS_PER_INODE];
}

Extended one:

struct f2fs_inode {
        ...
        struct f2fs_extent i_ext;
	union {
		struct {
			__le16 i_extra_isize;
			__le16 i_padding;
			__le32 i_extra_end[0];
		};
		__le32 i_addr[DEF_ADDRS_PER_INODE];
	};
        __le32 i_nid[DEF_NIDS_PER_INODE];
}

Once F2FS_EXTRA_ATTR is set, we will steal four bytes in the head of
i_addr field for storing i_extra_isize and i_padding. with i_extra_isize,
we can calculate actual size of reserved space in i_addr, available
attribute fields included in total extra attribute fields for current
inode can be described as below:

  +--------------------+
  | .i_mode            |
  | ...                |
  | .i_ext             |
  +--------------------+
  | .i_extra_isize     |-----+
  | .i_padding         |     |
  | .i_prjid           |     |
  | .i_atime_extra     |     |
  | .i_ctime_extra     |     |
  | .i_mtime_extra     |<----+
  | .i_inode_cs        |<----- store blkaddr/inline from here
  | .i_xattr_cs        |
  | ...                |
  +--------------------+
  |                    |
  |    block address   |
  |                    |
  +--------------------+
  | .i_nid             |
  +--------------------+
  |   node_footer      |
  | (nid, ino, offset) |
  +--------------------+

Hence, with this patch, we would enhance scalability of f2fs inode for
storing more newly added attribute.

Signed-off-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
This commit is contained in:
Chao Yu 2017-07-19 00:19:06 +08:00 committed by Jaegeuk Kim
parent f247037120
commit 7a2af766af
10 changed files with 135 additions and 52 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

15
fs/f2fs/data.c
View File

@ -460,10 +460,14 @@ static void __set_data_blkaddr(struct dnode_of_data *dn)
{
struct f2fs_node *rn = F2FS_NODE(dn->node_page);
__le32 *addr_array;
int base = 0;
if (IS_INODE(dn->node_page) && f2fs_has_extra_attr(dn->inode))
base = get_extra_isize(dn->inode);
/* Get physical address of data block */
addr_array = blkaddr_in_node(rn);
addr_array[dn->ofs_in_node] = cpu_to_le32(dn->data_blkaddr);
addr_array[base + dn->ofs_in_node] = cpu_to_le32(dn->data_blkaddr);
}
/*
@ -507,8 +511,8 @@ int reserve_new_blocks(struct dnode_of_data *dn, blkcnt_t count)
f2fs_wait_on_page_writeback(dn->node_page, NODE, true);
for (; count > 0; dn->ofs_in_node++) {
block_t blkaddr =
datablock_addr(dn->node_page, dn->ofs_in_node);
block_t blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
if (blkaddr == NULL_ADDR) {
dn->data_blkaddr = NEW_ADDR;
__set_data_blkaddr(dn);
@ -755,7 +759,8 @@ static int __allocate_data_block(struct dnode_of_data *dn)
if (unlikely(is_inode_flag_set(dn->inode, FI_NO_ALLOC)))
return -EPERM;
dn->data_blkaddr = datablock_addr(dn->node_page, dn->ofs_in_node);
dn->data_blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
if (dn->data_blkaddr == NEW_ADDR)
goto alloc;
@ -902,7 +907,7 @@ next_dnode:
end_offset = ADDRS_PER_PAGE(dn.node_page, inode);
next_block:
blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
blkaddr = datablock_addr(dn.inode, dn.node_page, dn.ofs_in_node);
if (blkaddr == NEW_ADDR || blkaddr == NULL_ADDR) {
if (create) {

72
fs/f2fs/f2fs.h
View File

@ -110,9 +110,10 @@ struct f2fs_mount_info {
unsigned int opt;
};
#define F2FS_FEATURE_ENCRYPT 0x0001
#define F2FS_FEATURE_BLKZONED 0x0002
#define F2FS_FEATURE_ATOMIC_WRITE 0x0004
#define F2FS_FEATURE_ENCRYPT 0x0001
#define F2FS_FEATURE_BLKZONED 0x0002
#define F2FS_FEATURE_ATOMIC_WRITE 0x0004
#define F2FS_FEATURE_EXTRA_ATTR 0x0008
#define F2FS_HAS_FEATURE(sb, mask) \
((F2FS_SB(sb)->raw_super->feature & cpu_to_le32(mask)) != 0)
@ -359,10 +360,10 @@ struct f2fs_flush_device {
/* for inline stuff */
#define DEF_INLINE_RESERVED_SIZE 1
static inline int get_inline_reserved_size(struct inode *inode);
#define MAX_INLINE_DATA(inode) (sizeof(__le32) * (DEF_ADDRS_PER_INODE -\
get_inline_reserved_size(inode) -\
static inline int get_extra_isize(struct inode *inode);
#define MAX_INLINE_DATA(inode) (sizeof(__le32) * \
(CUR_ADDRS_PER_INODE(inode) - \
DEF_INLINE_RESERVED_SIZE - \
F2FS_INLINE_XATTR_ADDRS))
/* for inline dir */
@ -568,7 +569,7 @@ struct f2fs_inode_info {
struct rw_semaphore dio_rwsem[2];/* avoid racing between dio and gc */
struct rw_semaphore i_mmap_sem;
int i_inline_reserved; /* reserved size in inline data */
int i_extra_isize; /* size of extra space located in i_addr */
};
static inline void get_extent_info(struct extent_info *ext,
@ -1792,20 +1793,38 @@ static inline bool IS_INODE(struct page *page)
return RAW_IS_INODE(p);
}
static inline int offset_in_addr(struct f2fs_inode *i)
{
return (i->i_inline & F2FS_EXTRA_ATTR) ?
(le16_to_cpu(i->i_extra_isize) / sizeof(__le32)) : 0;
}
static inline __le32 *blkaddr_in_node(struct f2fs_node *node)
{
return RAW_IS_INODE(node) ? node->i.i_addr : node->dn.addr;
}
static inline block_t datablock_addr(struct page *node_page,
unsigned int offset)
static inline int f2fs_has_extra_attr(struct inode *inode);
static inline block_t datablock_addr(struct inode *inode,
struct page *node_page, unsigned int offset)
{
struct f2fs_node *raw_node;
__le32 *addr_array;
int base = 0;
bool is_inode = IS_INODE(node_page);
raw_node = F2FS_NODE(node_page);
/* from GC path only */
if (!inode) {
if (is_inode)
base = offset_in_addr(&raw_node->i);
} else if (f2fs_has_extra_attr(inode) && is_inode) {
base = get_extra_isize(inode);
}
addr_array = blkaddr_in_node(raw_node);
return le32_to_cpu(addr_array[offset]);
return le32_to_cpu(addr_array[base + offset]);
}
static inline int f2fs_test_bit(unsigned int nr, char *addr)
@ -1896,6 +1915,7 @@ enum {
FI_DIRTY_FILE, /* indicate regular/symlink has dirty pages */
FI_NO_PREALLOC, /* indicate skipped preallocated blocks */
FI_HOT_DATA, /* indicate file is hot */
FI_EXTRA_ATTR, /* indicate file has extra attribute */
};
static inline void __mark_inode_dirty_flag(struct inode *inode,
@ -2015,6 +2035,8 @@ static inline void get_inline_info(struct inode *inode, struct f2fs_inode *ri)
set_bit(FI_DATA_EXIST, &fi->flags);
if (ri->i_inline & F2FS_INLINE_DOTS)
set_bit(FI_INLINE_DOTS, &fi->flags);
if (ri->i_inline & F2FS_EXTRA_ATTR)
set_bit(FI_EXTRA_ATTR, &fi->flags);
}
static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri)
@ -2031,6 +2053,13 @@ static inline void set_raw_inline(struct inode *inode, struct f2fs_inode *ri)
ri->i_inline |= F2FS_DATA_EXIST;
if (is_inode_flag_set(inode, FI_INLINE_DOTS))
ri->i_inline |= F2FS_INLINE_DOTS;
if (is_inode_flag_set(inode, FI_EXTRA_ATTR))
ri->i_inline |= F2FS_EXTRA_ATTR;
}
static inline int f2fs_has_extra_attr(struct inode *inode)
{
return is_inode_flag_set(inode, FI_EXTRA_ATTR);
}
static inline int f2fs_has_inline_xattr(struct inode *inode)
@ -2041,8 +2070,8 @@ static inline int f2fs_has_inline_xattr(struct inode *inode)
static inline unsigned int addrs_per_inode(struct inode *inode)
{
if (f2fs_has_inline_xattr(inode))
return DEF_ADDRS_PER_INODE - F2FS_INLINE_XATTR_ADDRS;
return DEF_ADDRS_PER_INODE;
return CUR_ADDRS_PER_INODE(inode) - F2FS_INLINE_XATTR_ADDRS;
return CUR_ADDRS_PER_INODE(inode);
}
static inline void *inline_xattr_addr(struct page *page)
@ -2104,9 +2133,9 @@ static inline bool f2fs_is_drop_cache(struct inode *inode)
static inline void *inline_data_addr(struct inode *inode, struct page *page)
{
struct f2fs_inode *ri = F2FS_INODE(page);
int reserved_size = get_inline_reserved_size(inode);
int extra_size = get_extra_isize(inode);
return (void *)&(ri->i_addr[reserved_size]);
return (void *)&(ri->i_addr[extra_size + DEF_INLINE_RESERVED_SIZE]);
}
static inline int f2fs_has_inline_dentry(struct inode *inode)
@ -2197,15 +2226,19 @@ static inline void *f2fs_kmalloc(struct f2fs_sb_info *sbi,
return kmalloc(size, flags);
}
static inline int get_inline_reserved_size(struct inode *inode)
static inline int get_extra_isize(struct inode *inode)
{
return F2FS_I(inode)->i_inline_reserved;
return F2FS_I(inode)->i_extra_isize / sizeof(__le32);
}
#define get_inode_mode(i) \
((is_inode_flag_set(i, FI_ACL_MODE)) ? \
(F2FS_I(i)->i_acl_mode) : ((i)->i_mode))
#define F2FS_TOTAL_EXTRA_ATTR_SIZE \
(offsetof(struct f2fs_inode, i_extra_end) - \
offsetof(struct f2fs_inode, i_extra_isize)) \
/*
* file.c
*/
@ -2798,6 +2831,11 @@ static inline int f2fs_sb_mounted_blkzoned(struct super_block *sb)
return F2FS_HAS_FEATURE(sb, F2FS_FEATURE_BLKZONED);
}
static inline int f2fs_sb_has_extra_attr(struct super_block *sb)
{
return F2FS_HAS_FEATURE(sb, F2FS_FEATURE_EXTRA_ATTR);
}
#ifdef CONFIG_BLK_DEV_ZONED
static inline int get_blkz_type(struct f2fs_sb_info *sbi,
struct block_device *bdev, block_t blkaddr)

23
fs/f2fs/file.c
View File

@ -382,7 +382,8 @@ static loff_t f2fs_seek_block(struct file *file, loff_t offset, int whence)
dn.ofs_in_node++, pgofs++,
data_ofs = (loff_t)pgofs << PAGE_SHIFT) {
block_t blkaddr;
blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
blkaddr = datablock_addr(dn.inode,
dn.node_page, dn.ofs_in_node);
if (__found_offset(blkaddr, dirty, pgofs, whence)) {
f2fs_put_dnode(&dn);
@ -467,9 +468,13 @@ int truncate_data_blocks_range(struct dnode_of_data *dn, int count)
struct f2fs_node *raw_node;
int nr_free = 0, ofs = dn->ofs_in_node, len = count;
__le32 *addr;
int base = 0;
if (IS_INODE(dn->node_page) && f2fs_has_extra_attr(dn->inode))
base = get_extra_isize(dn->inode);
raw_node = F2FS_NODE(dn->node_page);
addr = blkaddr_in_node(raw_node) + ofs;
addr = blkaddr_in_node(raw_node) + base + ofs;
for (; count > 0; count--, addr++, dn->ofs_in_node++) {
block_t blkaddr = le32_to_cpu(*addr);
@ -927,7 +932,8 @@ next_dnode:
done = min((pgoff_t)ADDRS_PER_PAGE(dn.node_page, inode) -
dn.ofs_in_node, len);
for (i = 0; i < done; i++, blkaddr++, do_replace++, dn.ofs_in_node++) {
*blkaddr = datablock_addr(dn.node_page, dn.ofs_in_node);
*blkaddr = datablock_addr(dn.inode,
dn.node_page, dn.ofs_in_node);
if (!is_checkpointed_data(sbi, *blkaddr)) {
if (test_opt(sbi, LFS)) {
@ -1003,8 +1009,8 @@ static int __clone_blkaddrs(struct inode *src_inode, struct inode *dst_inode,
ADDRS_PER_PAGE(dn.node_page, dst_inode) -
dn.ofs_in_node, len - i);
do {
dn.data_blkaddr = datablock_addr(dn.node_page,
dn.ofs_in_node);
dn.data_blkaddr = datablock_addr(dn.inode,
dn.node_page, dn.ofs_in_node);
truncate_data_blocks_range(&dn, 1);
if (do_replace[i]) {
@ -1173,7 +1179,8 @@ static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start,
int ret;
for (; index < end; index++, dn->ofs_in_node++) {
if (datablock_addr(dn->node_page, dn->ofs_in_node) == NULL_ADDR)
if (datablock_addr(dn->inode, dn->node_page,
dn->ofs_in_node) == NULL_ADDR)
count++;
}
@ -1184,8 +1191,8 @@ static int f2fs_do_zero_range(struct dnode_of_data *dn, pgoff_t start,
dn->ofs_in_node = ofs_in_node;
for (index = start; index < end; index++, dn->ofs_in_node++) {
dn->data_blkaddr =
datablock_addr(dn->node_page, dn->ofs_in_node);
dn->data_blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
/*
* reserve_new_blocks will not guarantee entire block
* allocation.

2
fs/f2fs/gc.c
View File

@ -587,7 +587,7 @@ static bool is_alive(struct f2fs_sb_info *sbi, struct f2fs_summary *sum,
}
*nofs = ofs_of_node(node_page);
source_blkaddr = datablock_addr(node_page, ofs_in_node);
source_blkaddr = datablock_addr(NULL, node_page, ofs_in_node);
f2fs_put_page(node_page, 1);
if (source_blkaddr != blkaddr)

32
fs/f2fs/inode.c
View File

@ -49,20 +49,22 @@ void f2fs_set_inode_flags(struct inode *inode)
static void __get_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
{
int extra_size = get_extra_isize(inode);
if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode) ||
S_ISFIFO(inode->i_mode) || S_ISSOCK(inode->i_mode)) {
if (ri->i_addr[0])
inode->i_rdev =
old_decode_dev(le32_to_cpu(ri->i_addr[0]));
if (ri->i_addr[extra_size])
inode->i_rdev = old_decode_dev(
le32_to_cpu(ri->i_addr[extra_size]));
else
inode->i_rdev =
new_decode_dev(le32_to_cpu(ri->i_addr[1]));
inode->i_rdev = new_decode_dev(
le32_to_cpu(ri->i_addr[extra_size + 1]));
}
}
static bool __written_first_block(struct f2fs_inode *ri)
{
block_t addr = le32_to_cpu(ri->i_addr[0]);
block_t addr = le32_to_cpu(ri->i_addr[offset_in_addr(ri)]);
if (addr != NEW_ADDR && addr != NULL_ADDR)
return true;
@ -71,16 +73,18 @@ static bool __written_first_block(struct f2fs_inode *ri)
static void __set_inode_rdev(struct inode *inode, struct f2fs_inode *ri)
{
int extra_size = get_extra_isize(inode);
if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) {
if (old_valid_dev(inode->i_rdev)) {
ri->i_addr[0] =
ri->i_addr[extra_size] =
cpu_to_le32(old_encode_dev(inode->i_rdev));
ri->i_addr[1] = 0;
ri->i_addr[extra_size + 1] = 0;
} else {
ri->i_addr[0] = 0;
ri->i_addr[1] =
ri->i_addr[extra_size] = 0;
ri->i_addr[extra_size + 1] =
cpu_to_le32(new_encode_dev(inode->i_rdev));
ri->i_addr[2] = 0;
ri->i_addr[extra_size + 2] = 0;
}
}
}
@ -153,6 +157,9 @@ static int do_read_inode(struct inode *inode)
get_inline_info(inode, ri);
fi->i_extra_isize = f2fs_has_extra_attr(inode) ?
le16_to_cpu(ri->i_extra_isize) : 0;
/* check data exist */
if (f2fs_has_inline_data(inode) && !f2fs_exist_data(inode))
__recover_inline_status(inode, node_page);
@ -292,6 +299,9 @@ int update_inode(struct inode *inode, struct page *node_page)
ri->i_generation = cpu_to_le32(inode->i_generation);
ri->i_dir_level = F2FS_I(inode)->i_dir_level;
if (f2fs_has_extra_attr(inode))
ri->i_extra_isize = cpu_to_le16(F2FS_I(inode)->i_extra_isize);
__set_inode_rdev(inode, ri);
set_cold_node(inode, node_page);

5
fs/f2fs/namei.c
View File

@ -72,6 +72,11 @@ static struct inode *f2fs_new_inode(struct inode *dir, umode_t mode)
set_inode_flag(inode, FI_NEW_INODE);
if (f2fs_sb_has_extra_attr(sbi->sb)) {
set_inode_flag(inode, FI_EXTRA_ATTR);
F2FS_I(inode)->i_extra_isize = F2FS_TOTAL_EXTRA_ATTR_SIZE;
}
if (test_opt(sbi, INLINE_XATTR))
set_inode_flag(inode, FI_INLINE_XATTR);
if (test_opt(sbi, INLINE_DATA) && f2fs_may_inline_data(inode))

7
fs/f2fs/node.c
View File

@ -655,7 +655,8 @@ int get_dnode_of_data(struct dnode_of_data *dn, pgoff_t index, int mode)
dn->nid = nids[level];
dn->ofs_in_node = offset[level];
dn->node_page = npage[level];
dn->data_blkaddr = datablock_addr(dn->node_page, dn->ofs_in_node);
dn->data_blkaddr = datablock_addr(dn->inode,
dn->node_page, dn->ofs_in_node);
return 0;
release_pages:
@ -2263,7 +2264,9 @@ retry:
dst->i_blocks = cpu_to_le64(1);
dst->i_links = cpu_to_le32(1);
dst->i_xattr_nid = 0;
dst->i_inline = src->i_inline & F2FS_INLINE_XATTR;
dst->i_inline = src->i_inline & (F2FS_INLINE_XATTR | F2FS_EXTRA_ATTR);
if (dst->i_inline & F2FS_EXTRA_ATTR)
dst->i_extra_isize = src->i_extra_isize;
new_ni = old_ni;
new_ni.ino = ino;

7
fs/f2fs/recovery.c
View File

@ -361,7 +361,8 @@ out:
return 0;
truncate_out:
if (datablock_addr(tdn.node_page, tdn.ofs_in_node) == blkaddr)
if (datablock_addr(tdn.inode, tdn.node_page,
tdn.ofs_in_node) == blkaddr)
truncate_data_blocks_range(&tdn, 1);
if (dn->inode->i_ino == nid && !dn->inode_page_locked)
unlock_page(dn->inode_page);
@ -414,8 +415,8 @@ retry_dn:
for (; start < end; start++, dn.ofs_in_node++) {
block_t src, dest;
src = datablock_addr(dn.node_page, dn.ofs_in_node);
dest = datablock_addr(page, dn.ofs_in_node);
src = datablock_addr(dn.inode, dn.node_page, dn.ofs_in_node);
dest = datablock_addr(dn.inode, page, dn.ofs_in_node);
/* skip recovering if dest is the same as src */
if (src == dest)

11
fs/f2fs/super.c
View File

@ -447,8 +447,6 @@ static struct inode *f2fs_alloc_inode(struct super_block *sb)
/* Will be used by directory only */
fi->i_dir_level = F2FS_SB(sb)->dir_level;
fi->i_inline_reserved = DEF_INLINE_RESERVED_SIZE;
return &fi->vfs_inode;
}
@ -1306,9 +1304,16 @@ static const struct export_operations f2fs_export_ops = {
static loff_t max_file_blocks(void)
{
loff_t result = (DEF_ADDRS_PER_INODE - F2FS_INLINE_XATTR_ADDRS);
loff_t result = 0;
loff_t leaf_count = ADDRS_PER_BLOCK;
/*
* note: previously, result is equal to (DEF_ADDRS_PER_INODE -
* F2FS_INLINE_XATTR_ADDRS), but now f2fs try to reserve more
* space in inode.i_addr, it will be more safe to reassign
* result as zero.
*/
/* two direct node blocks */
result += (leaf_count * 2);

13
include/linux/f2fs_fs.h
View File

@ -186,6 +186,8 @@ struct f2fs_extent {
#define F2FS_NAME_LEN 255
#define F2FS_INLINE_XATTR_ADDRS 50 /* 200 bytes for inline xattrs */
#define DEF_ADDRS_PER_INODE 923 /* Address Pointers in an Inode */
#define CUR_ADDRS_PER_INODE(inode) (DEF_ADDRS_PER_INODE - \
get_extra_isize(inode))
#define DEF_NIDS_PER_INODE 5 /* Node IDs in an Inode */
#define ADDRS_PER_INODE(inode) addrs_per_inode(inode)
#define ADDRS_PER_BLOCK 1018 /* Address Pointers in a Direct Block */
@ -205,6 +207,7 @@ struct f2fs_extent {
#define F2FS_INLINE_DENTRY 0x04 /* file inline dentry flag */
#define F2FS_DATA_EXIST 0x08 /* file inline data exist flag */
#define F2FS_INLINE_DOTS 0x10 /* file having implicit dot dentries */
#define F2FS_EXTRA_ATTR 0x20 /* file having extra attribute */
struct f2fs_inode {
__le16 i_mode; /* file mode */
@ -232,8 +235,14 @@ struct f2fs_inode {
struct f2fs_extent i_ext; /* caching a largest extent */
__le32 i_addr[DEF_ADDRS_PER_INODE]; /* Pointers to data blocks */
union {
struct {
__le16 i_extra_isize; /* extra inode attribute size */
__le16 i_padding; /* padding */
__le32 i_extra_end[0]; /* for attribute size calculation */
};
__le32 i_addr[DEF_ADDRS_PER_INODE]; /* Pointers to data blocks */
};
__le32 i_nid[DEF_NIDS_PER_INODE]; /* direct(2), indirect(2),
double_indirect(1) node id */
} __packed;