This patch contains:
- pdirops support for ldiskfs
- - N-level htree directory
- integrate with osd-ldiskfs
-Index: linux-3.10.0-123.13.2.el7.x86_64/include/linux/htree_lock.h
+Index: linux-3.10.0-229.1.2.fc21.x86_64/include/linux/htree_lock.h
===================================================================
--- /dev/null
-+++ linux-3.10.0-123.13.2.el7.x86_64/include/linux/htree_lock.h
++++ linux-3.10.0-229.1.2.fc21.x86_64/include/linux/htree_lock.h
@@ -0,0 +1,187 @@
+/*
+ * include/linux/htree_lock.h
+ ((lck)->lk_nodes[dep].ln_mode == HTREE_LOCK_NL)
+
+#endif
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/htree_lock.c
+Index: linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/htree_lock.c
===================================================================
--- /dev/null
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/htree_lock.c
++++ linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/htree_lock.c
@@ -0,0 +1,880 @@
+/*
+ * fs/ext4/htree_lock.c
+ htree_lock_mode_t mode, u32 key, unsigned dep,
+ int wait, void *event)
+{
-+ LIST_HEAD (list);
++ LIST_HEAD(list);
+ struct htree_lock *tmp;
+ struct htree_lock *tmp2;
+ u16 major;
+ kfree(lck);
+}
+EXPORT_SYMBOL(htree_lock_free);
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/ext4.h
+Index: linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/Makefile
===================================================================
---- linux-3.10.0-123.13.2.el7.x86_64.orig/fs/ext4/ext4.h
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/ext4.h
+--- linux-3.10.0-229.1.2.fc21.x86_64.orig/fs/ext4/Makefile
++++ linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/Makefile
+@@ -6,6 +6,7 @@ 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 \
++ htree_lock.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
+Index: linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/ext4.h
+===================================================================
+--- linux-3.10.0-229.1.2.fc21.x86_64.orig/fs/ext4/ext4.h
++++ linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/ext4.h
@@ -27,6 +27,7 @@
#include <linux/mutex.h>
#include <linux/timer.h>
#include <linux/blockgroup_lock.h>
#include <linux/percpu_counter.h>
#include <linux/ratelimit.h>
-@@ -810,6 +811,9 @@ struct ext4_inode_info {
+@@ -821,6 +822,9 @@ struct ext4_inode_info {
__u32 i_dtime;
ext4_fsblk_t i_file_acl;
/*
* i_block_group is the number of the block group which contains
* this file's inode. Constant across the lifetime of the inode,
-@@ -1536,6 +1540,7 @@ static inline void ext4_clear_state_flag
- EXT4_FEATURE_INCOMPAT_META_BG| \
- EXT4_FEATURE_INCOMPAT_EXTENTS| \
- EXT4_FEATURE_INCOMPAT_64BIT| \
-+ EXT4_FEATURE_INCOMPAT_LARGEDIR|\
- EXT4_FEATURE_INCOMPAT_FLEX_BG| \
- EXT4_FEATURE_INCOMPAT_EA_INODE| \
- EXT4_FEATURE_INCOMPAT_MMP | \
-@@ -1954,6 +1959,76 @@ struct mmpd_data {
- # define NORET_TYPE /**/
- # define ATTRIB_NORET __attribute__((noreturn))
- # define NORET_AND noreturn,
-+/* htree levels for ext4 */
-+#define EXT4_HTREE_LEVEL_COMPAT 2
-+#define EXT4_HTREE_LEVEL 3
-+
-+static inline int
-+ext4_dir_htree_level(struct super_block *sb)
-+{
-+ return EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_LARGEDIR) ?
-+ EXT4_HTREE_LEVEL : EXT4_HTREE_LEVEL_COMPAT;
-+}
-+
+@@ -1846,6 +1850,71 @@ struct dx_hash_info
+ */
+ #define HASH_NB_ALWAYS 1
+
+/* assume name-hash is protected by upper layer */
+#define EXT4_HTREE_LOCK_HASH 0
+
+ struct inode *dir, unsigned flags);
+#define ext4_htree_unlock(lck) htree_unlock(lck)
+
++extern struct buffer_head *__ext4_find_entry(struct inode *dir,
++ const struct qstr *d_name,
++ struct ext4_dir_entry_2 **res_dir,
++ int *inlined, struct htree_lock *lck);
++extern int __ext4_add_entry(handle_t *handle, struct dentry *dentry,
++ struct inode *inode, struct htree_lock *lck);
- struct ext4_xattr_ino_array {
- unsigned int xia_count; /* # of used item in the array */
-@@ -2050,9 +2125,17 @@ void ext4_insert_dentry(struct inode *in
+ /*
+ * Describe an inode's exact location on disk and in memory
+@@ -2088,9 +2157,17 @@ void ext4_insert_dentry(struct inode *in
const char *name, int namelen, void *data);
static inline void ext4_update_dx_flag(struct inode *inode)
{
}
static unsigned char ext4_filetype_table[] = {
DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
-@@ -2212,14 +2295,14 @@ extern int ext4_htree_fill_tree(struct f
- extern struct inode *ext4_create_inode(handle_t *handle,
- struct inode * dir, int mode);
- extern int ext4_add_entry(handle_t *handle, struct dentry *dentry,
-- struct inode *inode);
-+ struct inode *inode, struct htree_lock *lck);
- extern int ext4_delete_entry(handle_t *handle, struct inode * dir,
- struct ext4_dir_entry_2 * de_del,
- struct buffer_head * bh);
- extern struct buffer_head * ext4_find_entry(struct inode *dir,
- const struct qstr *d_name,
- struct ext4_dir_entry_2 ** res_dir,
-- int *inlined);
-+ int *inlined, struct htree_lock *lck);
- extern int ext4_add_dot_dotdot(handle_t *handle, struct inode *dir,
- struct inode *inode, const void *, const void *);
- extern int search_dir(struct buffer_head *bh,
-@@ -2382,13 +2465,15 @@ static inline void ext4_r_blocks_count_s
- es->s_r_blocks_count_hi = cpu_to_le32(blk >> 32);
- }
-
--static inline loff_t ext4_isize(struct ext4_inode *raw_inode)
-+static inline loff_t ext4_isize(struct super_block *sb,
-+ struct ext4_inode *raw_inode)
- {
-- if (S_ISREG(le16_to_cpu(raw_inode->i_mode)))
-+ if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_LARGEDIR) ||
-+ S_ISREG(le16_to_cpu(raw_inode->i_mode)))
- return ((loff_t)le32_to_cpu(raw_inode->i_size_high) << 32) |
- le32_to_cpu(raw_inode->i_size_lo);
-- else
-- return (loff_t) le32_to_cpu(raw_inode->i_size_lo);
-+
-+ return (loff_t) le32_to_cpu(raw_inode->i_size_lo);
- }
-
- static inline void ext4_isize_set(struct ext4_inode *raw_inode, loff_t i_size)
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/namei.c
+Index: linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/namei.c
===================================================================
---- linux-3.10.0-123.13.2.el7.x86_64.orig/fs/ext4/namei.c
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/namei.c
+--- linux-3.10.0-229.1.2.fc21.x86_64.orig/fs/ext4/namei.c
++++ linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/namei.c
@@ -53,6 +53,7 @@ struct buffer_head *ext4_append(handle_t
ext4_lblk_t *block)
{
+ }
inode->i_size += inode->i_sb->s_blocksize;
EXT4_I(inode)->i_disksize = inode->i_size;
- BUFFER_TRACE(bh, "get_write_access");
+ BUFFER_TRACE(bh, "get_write_access");
err = ext4_journal_get_write_access(handle, bh);
+ up(&ei->i_append_sem);
if (err) {
/* checksumming functions */
void initialize_dirent_tail(struct ext4_dir_entry_tail *t,
-@@ -517,7 +525,7 @@ struct dx_root_info * dx_get_dx_info(str
-
- static inline ext4_lblk_t dx_get_block(struct dx_entry *entry)
- {
-- return le32_to_cpu(entry->block) & 0x00ffffff;
-+ return le32_to_cpu(entry->block) & 0x0fffffff;
- }
-
- static inline void dx_set_block(struct dx_entry *entry, ext4_lblk_t value)
-@@ -667,6 +675,223 @@ struct stats dx_show_entries(struct dx_h
+@@ -668,6 +676,227 @@ struct stats dx_show_entries(struct dx_h
}
#endif /* DX_DEBUG */
+};
+
+#define ext4_htree_lock_data(l) ((struct ext4_dir_lock_data *)(l)->lk_private)
++#define ext4_find_entry(dir, name, dirent, inline) \
++ __ext4_find_entry(dir, name, dirent, inline, NULL)
++#define ext4_add_entry(handle, dentry, inode) \
++ __ext4_add_entry(handle, dentry, inode, NULL)
+
+/* NB: ext4_lblk_t is 32 bits so we use high bits to identify invalid blk */
+#define EXT4_HTREE_NODE_CHANGED (0xcafeULL << 32)
/*
* Probe for a directory leaf block to search.
*
-@@ -678,16 +903,17 @@ struct stats dx_show_entries(struct dx_h
+@@ -679,10 +908,11 @@ struct stats dx_show_entries(struct dx_h
*/
static struct dx_frame *
dx_probe(const struct qstr *d_name, struct inode *dir,
unsigned count, indirect;
- struct dx_entry *at, *entries, *p, *q, *m;
+ struct dx_entry *at, *entries, *p, *q, *m, *dx = NULL;
- struct dx_root_info * info;
+ struct dx_root_info *info;
struct buffer_head *bh;
struct dx_frame *frame = frame_in;
- u32 hash;
-
-- frame->bh = NULL;
-+ memset(frame_in, 0, EXT4_HTREE_LEVEL * sizeof(frame_in[0]));
- bh = ext4_read_dirblock(dir, 0, INDEX);
- if (IS_ERR(bh)) {
- *err = PTR_ERR(bh);
-@@ -720,9 +946,16 @@ dx_probe(const struct qstr *d_name, stru
- goto fail;
- }
-
-- if ((indirect = info->indirect_levels) > 1) {
-- ext4_warning(dir->i_sb, "Unimplemented inode hash depth: %#06x",
-- info->indirect_levels);
-+ indirect = info->indirect_levels;
-+ if (indirect >= ext4_dir_htree_level(dir->i_sb)) {
-+ ext4_warning(dir->i_sb,
-+ "Directory (ino: %lu) htree depth %#06x exceed "
-+ "supported value", dir->i_ino,
-+ ext4_dir_htree_level(dir->i_sb));
-+ if (ext4_dir_htree_level(dir->i_sb) < EXT4_HTREE_LEVEL) {
-+ ext4_warning(dir->i_sb, "Enable large directory "
-+ "feature to access it");
-+ }
- brelse(bh);
- *err = ERR_BAD_DX_DIR;
- goto fail;
-@@ -742,8 +975,15 @@ dx_probe(const struct qstr *d_name, stru
+@@ -750,8 +980,15 @@ dx_probe(const struct qstr *d_name, stru
dxtrace(printk("Look up %x", hash));
while (1)
{
ext4_warning(dir->i_sb,
"dx entry: no count or count > limit");
brelse(bh);
-@@ -784,7 +1024,70 @@ dx_probe(const struct qstr *d_name, stru
+@@ -792,7 +1029,70 @@ dx_probe(const struct qstr *d_name, stru
frame->bh = bh;
frame->entries = entries;
frame->at = at;
bh = ext4_read_dirblock(dir, dx_get_block(at), INDEX);
if (IS_ERR(bh)) {
*err = PTR_ERR(bh);
-@@ -818,13 +1121,18 @@ fail:
- static void dx_release (struct dx_frame *frames)
- {
- struct dx_root_info *info;
-+ int i;
-+
- if (frames[0].bh == NULL)
- return;
-
- info = dx_get_dx_info((struct ext4_dir_entry_2*)frames[0].bh->b_data);
-- if (info->indirect_levels)
-- brelse(frames[1].bh);
-- brelse(frames[0].bh);
-+ for (i = 0; i <= info->indirect_levels; i++) {
-+ if (frames[i].bh == NULL)
-+ break;
-+ brelse(frames[i].bh);
-+ frames[i].bh = NULL;
-+ }
- }
-
- /*
-@@ -847,7 +1155,7 @@ static void dx_release (struct dx_frame
+@@ -860,7 +1160,7 @@ static void dx_release (struct dx_frame
static int ext4_htree_next_block(struct inode *dir, __u32 hash,
struct dx_frame *frame,
struct dx_frame *frames,
{
struct dx_frame *p;
struct buffer_head *bh;
-@@ -862,12 +1170,22 @@ static int ext4_htree_next_block(struct
+@@ -875,12 +1175,22 @@ static int ext4_htree_next_block(struct
* this loop, num_frames indicates the number of interior
* nodes need to be read.
*/
p--;
}
-@@ -890,6 +1208,13 @@ static int ext4_htree_next_block(struct
+@@ -903,6 +1213,13 @@ static int ext4_htree_next_block(struct
* block so no check is necessary
*/
while (num_frames--) {
bh = ext4_read_dirblock(dir, dx_get_block(p->at), INDEX);
if (IS_ERR(bh))
return PTR_ERR(bh);
-@@ -898,6 +1223,7 @@ static int ext4_htree_next_block(struct
+@@ -911,6 +1228,7 @@ static int ext4_htree_next_block(struct
p->bh = bh;
p->at = p->entries = ((struct dx_node *) bh->b_data)->entries;
}
return 1;
}
-@@ -966,7 +1292,7 @@ int ext4_htree_fill_tree(struct file *di
- {
- struct dx_hash_info hinfo;
- struct ext4_dir_entry_2 *de;
-- struct dx_frame frames[2], *frame;
-+ struct dx_frame frames[EXT4_HTREE_LEVEL], *frame;
- struct inode *dir;
- ext4_lblk_t block;
- int count = 0;
-@@ -1000,10 +1326,10 @@ int ext4_htree_fill_tree(struct file *di
+@@ -1013,10 +1331,10 @@ int ext4_htree_fill_tree(struct file *di
}
hinfo.hash = start_hash;
hinfo.minor_hash = 0;
/* Add '.' and '..' from the htree header */
if (!start_hash && !start_minor_hash) {
de = (struct ext4_dir_entry_2 *) frames[0].bh->b_data;
-@@ -1030,7 +1356,7 @@ int ext4_htree_fill_tree(struct file *di
+@@ -1043,7 +1361,7 @@ int ext4_htree_fill_tree(struct file *di
count += ret;
hashval = ~0;
ret = ext4_htree_next_block(dir, HASH_NB_ALWAYS,
*next_hash = hashval;
if (ret < 0) {
err = ret;
-@@ -1226,7 +1552,7 @@ static int is_dx_internal_node(struct in
- struct buffer_head * ext4_find_entry(struct inode *dir,
+@@ -1236,10 +1554,10 @@ static int is_dx_internal_node(struct in
+ * The returned buffer_head has ->b_count elevated. The caller is expected
+ * to brelse() it when appropriate.
+ */
+-static struct buffer_head * ext4_find_entry (struct inode *dir,
++struct buffer_head *__ext4_find_entry(struct inode *dir,
const struct qstr *d_name,
struct ext4_dir_entry_2 **res_dir,
- int *inlined)
{
struct super_block *sb;
struct buffer_head *bh_use[NAMEI_RA_SIZE];
-@@ -1270,7 +1596,7 @@ struct buffer_head * ext4_find_entry(str
+@@ -1283,7 +1601,7 @@ static struct buffer_head * ext4_find_en
goto restart;
}
if (is_dx(dir)) {
/*
* On success, or if the error was file not found,
* return. Otherwise, fall back to doing a search the
-@@ -1280,6 +1606,7 @@ struct buffer_head * ext4_find_entry(str
+@@ -1297,6 +1615,7 @@ static struct buffer_head * ext4_find_en
return bh;
dxtrace(printk(KERN_DEBUG "ext4_find_entry: dx failed, "
"falling back\n"));
}
nblocks = dir->i_size >> EXT4_BLOCK_SIZE_BITS(sb);
start = EXT4_I(dir)->i_dir_start_lookup;
-@@ -1369,17 +1696,19 @@ cleanup_and_exit:
+@@ -1389,9 +1708,12 @@ cleanup_and_exit:
+ brelse(bh_use[ra_ptr]);
+ return ret;
}
- EXPORT_SYMBOL(ext4_find_entry);
++EXPORT_SYMBOL(__ext4_find_entry);
-static struct buffer_head * ext4_dx_find_entry(struct inode *dir, const struct qstr *d_name,
- struct ext4_dir_entry_2 **res_dir, int *err)
-+static struct buffer_head * ext4_dx_find_entry(struct inode *dir,
++static struct buffer_head *ext4_dx_find_entry(struct inode *dir,
+ const struct qstr *d_name,
+ struct ext4_dir_entry_2 **res_dir,
+ struct htree_lock *lck, int *err)
{
struct super_block * sb = dir->i_sb;
struct dx_hash_info hinfo;
-- struct dx_frame frames[2], *frame;
-+ struct dx_frame frames[EXT4_HTREE_LEVEL], *frame;
- struct buffer_head *bh;
+@@ -1400,7 +1722,7 @@ static struct buffer_head * ext4_dx_find
ext4_lblk_t block;
int retval;
return NULL;
do {
block = dx_get_block(frame->at);
-@@ -1403,7 +1732,7 @@ static struct buffer_head * ext4_dx_find
+@@ -1424,7 +1746,7 @@ static struct buffer_head * ext4_dx_find
/* Check to see if we should continue to search */
retval = ext4_htree_next_block(dir, hinfo.hash, frame,
if (retval < 0) {
ext4_warning(sb,
"error reading index page in directory #%lu",
-@@ -1429,7 +1758,7 @@ static struct dentry *ext4_lookup(struct
- if (dentry->d_name.len > EXT4_NAME_LEN)
- return ERR_PTR(-ENAMETOOLONG);
-
-- bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL);
-+ bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL, NULL);
- if (IS_ERR(bh))
- return (struct dentry *) bh;
- inode = NULL;
-@@ -1489,7 +1818,7 @@ struct dentry *ext4_get_parent(struct de
- struct ext4_dir_entry_2 * de;
- struct buffer_head *bh;
-
-- bh = ext4_find_entry(child->d_inode, &dotdot, &de, NULL);
-+ bh = ext4_find_entry(child->d_inode, &dotdot, &de, NULL, NULL);
- if (IS_ERR(bh))
- return (struct dentry *) bh;
- if (!bh)
-@@ -1559,8 +1888,9 @@ static struct ext4_dir_entry_2* dx_pack_
+@@ -1583,8 +1905,9 @@ static struct ext4_dir_entry_2* dx_pack_
* Returns pointer to de in block into which the new entry will be inserted.
*/
static struct ext4_dir_entry_2 *do_split(handle_t *handle, struct inode *dir,
{
unsigned blocksize = dir->i_sb->s_blocksize;
unsigned count, continued;
-@@ -1624,7 +1954,14 @@ static struct ext4_dir_entry_2 *do_split
+@@ -1647,7 +1970,14 @@ static struct ext4_dir_entry_2 *do_split
hash2, split, count-split));
/* Fancy dance to stay within two buffers */
de = dx_pack_dirents(data1, blocksize);
de->rec_len = ext4_rec_len_to_disk(data1 + (blocksize - csum_size) -
(char *) de,
-@@ -1643,13 +1980,21 @@ static struct ext4_dir_entry_2 *do_split
+@@ -1666,13 +1996,21 @@ static struct ext4_dir_entry_2 *do_split
dxtrace(dx_show_leaf (hinfo, (struct ext4_dir_entry_2 *) data1, blocksize, 1));
dxtrace(dx_show_leaf (hinfo, (struct ext4_dir_entry_2 *) data2, blocksize, 1));
err = ext4_handle_dirty_dirent_node(handle, dir, bh2);
if (err)
goto journal_error;
-@@ -1809,7 +2154,7 @@ static int add_dirent_to_buf(handle_t *h
- */
- dir->i_mtime = dir->i_ctime = ext4_current_time(dir);
- ext4_update_dx_flag(dir);
-- dir->i_version++;
-+ inode_inc_iversion(dir);
- ext4_mark_inode_dirty(handle, dir);
- BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
- err = ext4_handle_dirty_dirent_node(handle, dir, bh);
-@@ -1829,7 +2174,7 @@ static int make_indexed_dir(handle_t *ha
- const char *name = dentry->d_name.name;
- int namelen = dentry->d_name.len;
- struct buffer_head *bh2;
-- struct dx_frame frames[2], *frame;
-+ struct dx_frame frames[EXT4_HTREE_LEVEL], *frame;
- struct dx_entry *entries;
- struct ext4_dir_entry_2 *de, *de2, *dot_de, *dotdot_de;
- struct ext4_dir_entry_tail *t;
-@@ -1923,7 +2268,7 @@ static int make_indexed_dir(handle_t *ha
+@@ -1945,7 +2283,7 @@ static int make_indexed_dir(handle_t *ha
ext4_handle_dirty_dx_node(handle, dir, frame->bh);
ext4_handle_dirty_dirent_node(handle, dir, bh);
- de = do_split(handle,dir, &bh, frame, &hinfo, &retval);
-+ de = do_split(handle,dir, &bh, frames, frame, &hinfo, NULL, &retval);
++ de = do_split(handle, dir, &bh, frames, frame, &hinfo, NULL, &retval);
if (!de) {
/*
* Even if the block split failed, we have to properly write
-@@ -2030,7 +2375,7 @@ out:
+@@ -2051,8 +2389,8 @@ out:
+ * may not sleep between calling this and putting something into
* the entry, as someone else might have used it while you slept.
*/
- int ext4_add_entry(handle_t *handle, struct dentry *dentry,
-- struct inode *inode)
-+ struct inode *inode, struct htree_lock *lck)
+-static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
+- struct inode *inode)
++int __ext4_add_entry(handle_t *handle, struct dentry *dentry,
++ struct inode *inode, struct htree_lock *lck)
{
struct inode *dir = dentry->d_parent->d_inode;
struct buffer_head *bh;
-@@ -2066,9 +2411,10 @@ int ext4_add_entry(handle_t *handle, str
+@@ -2087,9 +2425,10 @@ static int ext4_add_entry(handle_t *hand
if (dentry->d_name.len == 2 &&
memcmp(dentry->d_name.name, "..", 2) == 0)
return ext4_update_dotdot(handle, dentry, inode);
ext4_clear_inode_flag(dir, EXT4_INODE_INDEX);
dx_fallback++;
ext4_mark_inode_dirty(handle, dir);
-@@ -2114,18 +2460,21 @@ EXPORT_SYMBOL(ext4_add_entry);
+@@ -2129,12 +2468,13 @@ static int ext4_add_entry(handle_t *hand
+ ext4_set_inode_state(inode, EXT4_STATE_NEWENTRY);
+ return retval;
+ }
++EXPORT_SYMBOL(__ext4_add_entry);
+
+ /*
* Returns 0 for success, or a negative error value
*/
static int ext4_dx_add_entry(handle_t *handle, struct dentry *dentry,
- struct inode *inode)
+ struct inode *inode, struct htree_lock *lck)
{
-- struct dx_frame frames[2], *frame;
-+ struct dx_frame frames[EXT4_HTREE_LEVEL], *frame;
+ struct dx_frame frames[EXT4_HTREE_LEVEL], *frame;
struct dx_entry *entries, *at;
- struct dx_hash_info hinfo;
- struct buffer_head *bh;
- struct inode *dir = dentry->d_parent->d_inode;
- struct super_block *sb = dir->i_sb;
- struct ext4_dir_entry_2 *de;
-+ int restart;
- int err;
+@@ -2148,7 +2488,7 @@ static int ext4_dx_add_entry(handle_t *h
+ again:
+ restart = 0;
- frame = dx_probe(&dentry->d_name, dir, &hinfo, frames, &err);
-+again:
-+ restart = 0;
+ frame = dx_probe(&dentry->d_name, dir, &hinfo, frames, lck, &err);
if (!frame)
return err;
entries = frame->entries;
-@@ -2137,33 +2486,53 @@ static int ext4_dx_add_entry(handle_t *h
- goto cleanup;
- }
-
-- BUFFER_TRACE(bh, "get_write_access");
-- err = ext4_journal_get_write_access(handle, bh);
-- if (err)
-- goto journal_error;
--
- err = add_dirent_to_buf(handle, dentry, inode, NULL, bh);
- if (err != -ENOSPC)
- goto cleanup;
-
-+ err = 0;
- /* Block full, should compress but for now just split */
- dxtrace(printk(KERN_DEBUG "using %u of %u node entries\n",
- dx_get_count(entries), dx_get_limit(entries)));
- /* Need to split index? */
- if (dx_get_count(entries) == dx_get_limit(entries)) {
- ext4_lblk_t newblock;
-- unsigned icount = dx_get_count(entries);
-- int levels = frame - frames;
-+ int levels = frame - frames + 1;
-+ unsigned icount;
-+ int add_level = 1;
- struct dx_entry *entries2;
+@@ -2178,6 +2518,11 @@ again:
struct dx_node *node2;
struct buffer_head *bh2;
-- if (levels && (dx_get_count(frames->entries) ==
-- dx_get_limit(frames->entries))) {
-- ext4_warning(sb, "Directory index full!");
+ if (!ext4_htree_safe_locked(lck)) { /* retry with EX lock */
+ ext4_htree_safe_relock(lck);
+ restart = 1;
+ goto cleanup;
+ }
-+ while (frame > frames) {
-+ if (dx_get_count((frame - 1)->entries) <
-+ dx_get_limit((frame - 1)->entries)) {
-+ add_level = 0;
-+ break;
-+ }
-+ frame--; /* split higher index block */
-+ at = frame->at;
-+ entries = frame->entries;
-+ restart = 1;
-+ }
-+ if (add_level && levels == ext4_dir_htree_level(sb)) {
-+ ext4_warning(sb, "Directory (ino: %lu) index full, "
-+ "reach max htree level :%d",
-+ dir->i_ino, levels);
-+ if (ext4_dir_htree_level(sb) < EXT4_HTREE_LEVEL) {
-+ ext4_warning(sb, "Large directory feature is"
-+ "not enabled on this "
-+ "filesystem");
-+ }
- err = -ENOSPC;
+ while (frame > frames) {
+ if (dx_get_count((frame - 1)->entries) <
+ dx_get_limit((frame - 1)->entries)) {
+@@ -2277,16 +2622,43 @@ again:
+ restart = 1;
goto cleanup;
}
-+ icount = dx_get_count(entries);
- bh2 = ext4_append(handle, dir, &newblock);
- if (IS_ERR(bh2)) {
- err = PTR_ERR(bh2);
-@@ -2178,7 +2547,7 @@ static int ext4_dx_add_entry(handle_t *h
- err = ext4_journal_get_write_access(handle, frame->bh);
- if (err)
- goto journal_error;
-- if (levels) {
-+ if (!add_level) {
- unsigned icount1 = icount/2, icount2 = icount - icount1;
- unsigned hash2 = dx_get_hash(entries + icount1);
- dxtrace(printk(KERN_DEBUG "Split index %i/%i\n",
-@@ -2186,7 +2555,7 @@ static int ext4_dx_add_entry(handle_t *h
-
- BUFFER_TRACE(frame->bh, "get_write_access"); /* index root */
- err = ext4_journal_get_write_access(handle,
-- frames[0].bh);
-+ (frame - 1)->bh);
- if (err)
- goto journal_error;
-
-@@ -2202,18 +2571,24 @@ static int ext4_dx_add_entry(handle_t *h
- frame->entries = entries = entries2;
- swap(frame->bh, bh2);
- }
-- dx_insert_block(frames + 0, hash2, newblock);
-- dxtrace(dx_show_index("node", frames[1].entries));
-+ dx_insert_block((frame - 1), hash2, newblock);
-+ dxtrace(dx_show_index("node", frame->entries));
- dxtrace(dx_show_index("node",
- ((struct dx_node *) bh2->b_data)->entries));
- err = ext4_handle_dirty_dx_node(handle, dir, bh2);
- if (err)
- goto journal_error;
- brelse (bh2);
-+ ext4_handle_dirty_metadata(handle, inode,
-+ (frame - 1)->bh);
-+ if (restart) {
-+ ext4_handle_dirty_metadata(handle, inode,
-+ frame->bh);
-+ goto cleanup;
-+ }
- } else {
- struct dx_root_info * info;
-- dxtrace(printk(KERN_DEBUG
-- "Creating second level index...\n"));
-+
- memcpy((char *) entries2, (char *) entries,
- icount * sizeof(struct dx_entry));
- dx_set_limit(entries2, dx_node_limit(dir));
-@@ -2223,35 +2598,63 @@ static int ext4_dx_add_entry(handle_t *h
- dx_set_block(entries + 0, newblock);
- info = dx_get_dx_info((struct ext4_dir_entry_2*)
- frames[0].bh->b_data);
-- info->indirect_levels = 1;
-+ info->indirect_levels += 1;
-+ dxtrace(printk(KERN_DEBUG
-+ "Creating %d level index...\n",
-+ info->indirect_levels));
-+ ext4_handle_dirty_metadata(handle, inode, frame->bh);
-+ ext4_handle_dirty_metadata(handle, inode, bh2);
-+ brelse(bh2);
-+ restart = 1;
-+ goto cleanup;
-+ }
+ } else if (!ext4_htree_dx_locked(lck)) {
+ struct ext4_dir_lock_data *ld = ext4_htree_lock_data(lck);
-
-- /* Add new access path frame */
-- frame = frames + 1;
-- frame->at = at = at - entries + entries2;
-- frame->entries = entries = entries2;
-- frame->bh = bh2;
-- err = ext4_journal_get_write_access(handle,
-- frame->bh);
-- if (err)
-- goto journal_error;
++
+ /* not well protected, require DX lock */
+ ext4_htree_dx_need_lock(lck);
+ at = frame > frames ? (frame - 1)->at : NULL;
+ (ld->ld_count != dx_get_count(entries))) {
+ restart = 1;
+ goto cleanup;
- }
-- err = ext4_handle_dirty_dx_node(handle, dir, frames[0].bh);
++ }
+ /* OK, I've got DX lock and nothing changed */
+ frame->at = ld->ld_at;
- if (err) {
- ext4_std_error(inode->i_sb, err);
- goto cleanup;
- }
}
- de = do_split(handle, dir, &bh, frame, &hinfo, &err);
+ de = do_split(handle, dir, &bh, frames, frame, &hinfo, lck, &err);
+ ext4_htree_de_unlock(lck);
brelse(bh);
dx_release(frames);
-+ /* @restart is true means htree-path has been changed, we need to
-+ * repeat dx_probe() to find out valid htree-path */
-+ if (restart && err == 0)
-+ goto again;
- return err;
- }
-
-@@ -2288,7 +2691,7 @@ int ext4_generic_delete_entry(handle_t *
- blocksize);
- else
- de->inode = 0;
-- dir->i_version++;
-+ inode_inc_iversion(dir);
- return 0;
- }
- i += ext4_rec_len_from_disk(de->rec_len, blocksize);
-@@ -2373,7 +2776,7 @@ EXPORT_SYMBOL(ext4_dec_count);
- static int ext4_add_nondir(handle_t *handle,
- struct dentry *dentry, struct inode *inode)
- {
-- int err = ext4_add_entry(handle, dentry, inode);
-+ int err = ext4_add_entry(handle, dentry, inode, NULL);
- if (!err) {
- ext4_mark_inode_dirty(handle, inode);
- unlock_new_inode(inode);
-@@ -2641,7 +3044,7 @@ retry:
- goto out_clear_inode;
- err = ext4_mark_inode_dirty(handle, inode);
- if (!err)
-- err = ext4_add_entry(handle, dentry, inode);
-+ err = ext4_add_entry(handle, dentry, inode, NULL);
- if (err) {
- out_clear_inode:
- clear_nlink(inode);
-@@ -2907,7 +3310,7 @@ static int ext4_rmdir(struct inode *dir,
- dquot_initialize(dentry->d_inode);
-
- retval = -ENOENT;
-- bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL);
-+ bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL, NULL);
- if (IS_ERR(bh))
- return PTR_ERR(bh);
- if (!bh)
-@@ -2974,7 +3377,7 @@ static int ext4_unlink(struct inode *dir
- dquot_initialize(dentry->d_inode);
-
- retval = -ENOENT;
-- bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL);
-+ bh = ext4_find_entry(dir, &dentry->d_name, &de, NULL, NULL);
- if (IS_ERR(bh))
- return PTR_ERR(bh);
- if (!bh)
-@@ -3153,7 +3556,7 @@ retry:
- ext4_inc_count(handle, inode);
- ihold(inode);
-
-- err = ext4_add_entry(handle, dentry, inode);
-+ err = ext4_add_entry(handle, dentry, inode, NULL);
- if (!err) {
- ext4_mark_inode_dirty(handle, inode);
- d_instantiate(dentry, inode);
-@@ -3183,7 +3556,7 @@ retry:
- struct buffer_head *bh;
- struct ext4_dir_entry_2 *de;
-
-- bh = ext4_find_entry(dir, d_name, &de, NULL);
-+ bh = ext4_find_entry(dir, d_name, &de, NULL, NULL);
- if (IS_ERR(bh))
- return PTR_ERR(bh);
- if (bh) {
-@@ -3230,7 +3633,7 @@ static int ext4_rename(struct inode *old
- if (IS_DIRSYNC(old_dir) || IS_DIRSYNC(new_dir))
- ext4_handle_sync(handle);
-
-- old.bh = ext4_find_entry(old.dir, &old.dentry->d_name, &old.de, NULL);
-+ old.bh = ext4_find_entry(old.dir, &old.dentry->d_name, &old.de, NULL, NULL);
- if (IS_ERR(old.bh))
- return PTR_ERR(old.bh);
- /*
-@@ -3244,7 +3647,7 @@ static int ext4_rename(struct inode *old
-
- new_inode = new_dentry->d_inode;
- new.bh = ext4_find_entry(new.dir, &new.dentry->d_name,
-- &new.de, &new.inlined);
-+ &new.de, &new.inlined, NULL);
- if (IS_ERR(new.bh)) {
- if (!new_inode) {
- brelse(new_bh);
-@@ -3275,7 +3678,7 @@ static int ext4_rename(struct inode *old
- goto end_rename;
- }
- if (!new.bh) {
-- retval = ext4_add_entry(handle, new.dentry, old.inode);
-+ retval = ext4_add_entry(handle, new.dentry, old.inode, NULL);
- if (retval)
- goto end_rename;
- } else {
-@@ -3375,7 +3678,7 @@ static int ext4_rename(struct inode *old
- dquot_initialize(new.dir);
-
- old.bh = ext4_find_entry(old.dir, &old.dentry->d_name,
-- &old.de, &old.inlined);
-+ &old.de, &old.inlined, NULL);
- /*
- * Check for inode number is _not_ due to possible IO errors.
- * We might rmdir the source, keep it as pwd of some process
-@@ -3475,7 +3678,7 @@ static int ext4_rename(struct inode *old
- goto end_rename;
-
- new.bh = ext4_find_entry(new.dir, &new.dentry->d_name,
-- &new.de, &new.inlined);
-+ &new.de, &new.inlined, NULL);
-
- /* RENAME_EXCHANGE case: old *and* new must both exist */
- if (!new.bh || le32_to_cpu(new.de->inode) != new.inode->i_ino)
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/inode.c
-===================================================================
---- linux-3.10.0-123.13.2.el7.x86_64.orig/fs/ext4/inode.c
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/inode.c
-@@ -4264,7 +4264,7 @@ struct inode *ext4_iget(struct super_blo
- if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_64BIT))
- ei->i_file_acl |=
- ((__u64)le16_to_cpu(raw_inode->i_file_acl_high)) << 32;
-- inode->i_size = ext4_isize(raw_inode);
-+ inode->i_size = ext4_isize(sb, raw_inode);
- ei->i_disksize = inode->i_size;
- #ifdef CONFIG_QUOTA
- ei->i_reserved_quota = 0;
-@@ -4499,7 +4499,7 @@ static int ext4_do_update_inode(handle_t
- raw_inode->i_file_acl_high =
- cpu_to_le16(ei->i_file_acl >> 32);
- raw_inode->i_file_acl_lo = cpu_to_le32(ei->i_file_acl);
-- if (ei->i_disksize != ext4_isize(raw_inode)) {
-+ if (ei->i_disksize != ext4_isize(inode->i_sb, raw_inode)) {
- ext4_isize_set(raw_inode, ei->i_disksize);
- need_datasync = 1;
- }
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/Makefile
-===================================================================
---- linux-3.10.0-123.13.2.el7.x86_64.orig/fs/ext4/Makefile
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/Makefile
-@@ -8,7 +8,7 @@ ext4-y := balloc.o bitmap.o dir.o file.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
-+ xattr_trusted.o inline.o htree_lock.o
-
- ext4-$(CONFIG_EXT4_FS_POSIX_ACL) += acl.o
- ext4-$(CONFIG_EXT4_FS_SECURITY) += xattr_security.o
-Index: linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/super.c
+ /* @restart is true means htree-path has been changed, we need to
+Index: linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/super.c
===================================================================
---- linux-3.10.0-123.13.2.el7.x86_64.orig/fs/ext4/super.c
-+++ linux-3.10.0-123.13.2.el7.x86_64/fs/ext4/super.c
-@@ -872,6 +872,7 @@ static struct inode *ext4_alloc_inode(st
+--- linux-3.10.0-229.1.2.fc21.x86_64.orig/fs/ext4/super.c
++++ linux-3.10.0-229.1.2.fc21.x86_64/fs/ext4/super.c
+@@ -875,6 +875,7 @@ static struct inode *ext4_alloc_inode(st
ei->vfs_inode.i_version = 1;
- spin_lock_init(&ei->i_raw_lock);
+ spin_lock_init(&ei->i_raw_lock);
+ sema_init(&ei->i_append_sem, 1);
INIT_LIST_HEAD(&ei->i_prealloc_list);
spin_lock_init(&ei->i_prealloc_lock);
git://git.whamcloud.com / fs / lustre-release.git / blobdiff