* [PATCH V2 02/19] ext4: export inline xattr functions.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 03/19] ext4: Add the basic function for inline data support Tao Ma
` (16 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Inline data needs some inline xattr functions, so export them
from fs/ext4/xattr.c so that inline.c can uses them.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/xattr.c | 39 +++++-------------------------------
fs/ext4/xattr.h | 57 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 63 insertions(+), 33 deletions(-)
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index 93a00d8..71ef45e 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -61,11 +61,6 @@
#include "xattr.h"
#include "acl.h"
-#define BHDR(bh) ((struct ext4_xattr_header *)((bh)->b_data))
-#define ENTRY(ptr) ((struct ext4_xattr_entry *)(ptr))
-#define BFIRST(bh) ENTRY(BHDR(bh)+1)
-#define IS_LAST_ENTRY(entry) (*(__u32 *)(entry) == 0)
-
#ifdef EXT4_XATTR_DEBUG
# define ea_idebug(inode, f...) do { \
printk(KERN_DEBUG "inode %s:%lu: ", \
@@ -255,7 +250,7 @@ cleanup:
return error;
}
-static int
+int
ext4_xattr_ibody_get(struct inode *inode, int name_index, const char *name,
void *buffer, size_t buffer_size)
{
@@ -522,21 +517,6 @@ static size_t ext4_xattr_free_space(struct ext4_xattr_entry *last,
return (*min_offs - ((void *)last - base) - sizeof(__u32));
}
-struct ext4_xattr_info {
- int name_index;
- const char *name;
- const void *value;
- size_t value_len;
-};
-
-struct ext4_xattr_search {
- struct ext4_xattr_entry *first;
- void *base;
- void *end;
- struct ext4_xattr_entry *here;
- int not_found;
-};
-
static int
ext4_xattr_set_entry(struct ext4_xattr_info *i, struct ext4_xattr_search *s)
{
@@ -890,14 +870,8 @@ bad_block:
#undef header
}
-struct ext4_xattr_ibody_find {
- struct ext4_xattr_search s;
- struct ext4_iloc iloc;
-};
-
-static int
-ext4_xattr_ibody_find(struct inode *inode, struct ext4_xattr_info *i,
- struct ext4_xattr_ibody_find *is)
+int ext4_xattr_ibody_find(struct inode *inode, struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is)
{
struct ext4_xattr_ibody_header *header;
struct ext4_inode *raw_inode;
@@ -925,10 +899,9 @@ ext4_xattr_ibody_find(struct inode *inode, struct ext4_xattr_info *i,
return 0;
}
-static int
-ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
- struct ext4_xattr_info *i,
- struct ext4_xattr_ibody_find *is)
+int ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
+ struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is)
{
struct ext4_xattr_ibody_header *header;
struct ext4_xattr_search *s = &is->s;
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 25b7387..c4c015d 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -63,6 +63,32 @@ struct ext4_xattr_entry {
EXT4_I(inode)->i_extra_isize))
#define IFIRST(hdr) ((struct ext4_xattr_entry *)((hdr)+1))
+#define BHDR(bh) ((struct ext4_xattr_header *)((bh)->b_data))
+#define ENTRY(ptr) ((struct ext4_xattr_entry *)(ptr))
+#define BFIRST(bh) ENTRY(BHDR(bh)+1)
+#define IS_LAST_ENTRY(entry) (*(__u32 *)(entry) == 0)
+
+
+struct ext4_xattr_info {
+ int name_index;
+ const char *name;
+ const void *value;
+ size_t value_len;
+};
+
+struct ext4_xattr_search {
+ struct ext4_xattr_entry *first;
+ void *base;
+ void *end;
+ struct ext4_xattr_entry *here;
+ int not_found;
+};
+
+struct ext4_xattr_ibody_find {
+ struct ext4_xattr_search s;
+ struct ext4_iloc iloc;
+};
+
# ifdef CONFIG_EXT4_FS_XATTR
extern const struct xattr_handler ext4_xattr_user_handler;
@@ -88,6 +114,15 @@ extern void ext4_exit_xattr(void);
extern const struct xattr_handler *ext4_xattr_handlers[];
+extern int ext4_xattr_ibody_find(struct inode *inode, struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is);
+extern int ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
+ struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is);
+extern int ext4_xattr_ibody_get(struct inode *inode, int name_index,
+ const char *name,
+ void *buffer, size_t buffer_size);
+
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -141,6 +176,28 @@ ext4_expand_extra_isize_ea(struct inode *inode, int new_extra_isize,
#define ext4_xattr_handlers NULL
+static inline int
+ext4_xattr_ibody_find(struct inode *inode, struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is)
+{
+ return -EOPNOTSUPP;
+}
+
+static inline int
+ext4_xattr_ibody_set(handle_t *handle, struct inode *inode,
+ struct ext4_xattr_info *i,
+ struct ext4_xattr_ibody_find *is)
+{
+ return -EOPNOTSUPP;
+}
+
+extern int ext4_xattr_ibody_get(struct inode *inode, int name_index,
+ const char *name,
+ void *buffer, size_t buffer_size)
+{
+ return -EOPNOTSUPP;
+}
+
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 03/19] ext4: Add the basic function for inline data support.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
2011-11-21 16:39 ` [PATCH V2 02/19] ext4: export inline xattr functions Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 04/19] ext4: Add read support for inline data Tao Ma
` (15 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Implement inline data with xattr. This idea is inspired by Andreas.
So now we use "system.data" to store xattr.
For inode_size = 256, currently we uses all the space between i_extra_isize
and inode_size. For inode_size > 256, we use half of that space.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/Makefile | 2 +-
fs/ext4/ext4.h | 5 +
fs/ext4/inline.c | 459 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/inode.c | 4 +-
fs/ext4/xattr.h | 62 ++++++++
5 files changed, 530 insertions(+), 2 deletions(-)
create mode 100644 fs/ext4/inline.c
diff --git a/fs/ext4/Makefile b/fs/ext4/Makefile
index 56fd8f8..e88b7a6 100644
--- a/fs/ext4/Makefile
+++ b/fs/ext4/Makefile
@@ -9,6 +9,6 @@ ext4-y := balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o page-io.o \
ext4_jbd2.o migrate.o mballoc.o block_validity.o move_extent.o \
mmp.o indirect.o
-ext4-$(CONFIG_EXT4_FS_XATTR) += xattr.o xattr_user.o xattr_trusted.o
+ext4-$(CONFIG_EXT4_FS_XATTR) += xattr.o xattr_user.o xattr_trusted.o inline.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 5b0e26a..3e560ba 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -869,6 +869,10 @@ struct ext4_inode_info {
/* on-disk additional length */
__u16 i_extra_isize;
+ /* Indicate the inline data space. */
+ u16 i_inline_off;
+ u16 i_inline_size;
+
#ifdef CONFIG_QUOTA
/* quota space reservation, managed internally by quota code */
qsize_t i_reserved_quota;
@@ -1297,6 +1301,7 @@ enum {
EXT4_STATE_DIO_UNWRITTEN, /* need convert on dio done*/
EXT4_STATE_NEWENTRY, /* File just added to dir */
EXT4_STATE_DELALLOC_RESERVED, /* blks already reserved for delalloc */
+ EXT4_STATE_MAY_INLINE_DATA, /* may have in-inode data */
};
#define EXT4_INODE_BIT_FNS(name, field, offset) \
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
new file mode 100644
index 0000000..d3aa377
--- /dev/null
+++ b/fs/ext4/inline.c
@@ -0,0 +1,459 @@
+/*
+ * Copyright (c) 2011 Taobao.
+ * Written by Tao Ma <boyu.mt@taobao.com>
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of version 2.1 of the GNU Lesser General Public License
+ * as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will 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.
+ */
+#include "ext4_jbd2.h"
+#include "ext4.h"
+#include "xattr.h"
+
+#define EXT4_XATTR_SYSTEM_DATA_NAME "data"
+#define EXT4_MIN_INLINE_DATA_SIZE ((sizeof(__le32) * EXT4_N_BLOCKS))
+
+int ext4_get_inline_size(struct inode *inode)
+{
+ if (EXT4_I(inode)->i_inline_off)
+ return EXT4_I(inode)->i_inline_size;
+
+ return 0;
+}
+
+static int get_max_inline_xattr_value_size(struct inode *inode,
+ struct ext4_iloc *iloc)
+{
+ struct ext4_xattr_ibody_header *header;
+ struct ext4_xattr_entry *entry;
+ struct ext4_inode *raw_inode;
+ int free, min_offs;
+
+ if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR))
+ return EXT4_XATTR_SIZE(EXT4_SB(inode->i_sb)->s_inode_size -
+ EXT4_GOOD_OLD_INODE_SIZE -
+ EXT4_I(inode)->i_extra_isize -
+ sizeof(struct ext4_xattr_ibody_header) -
+ EXT4_XATTR_LEN(strlen(EXT4_XATTR_SYSTEM_DATA_NAME)) -
+ EXT4_XATTR_ROUND - sizeof(__u32));
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ raw_inode = ext4_raw_inode(iloc);
+ header = IHDR(inode, raw_inode);
+ entry = IFIRST(header);
+ min_offs = EXT4_SB(inode->i_sb)->s_inode_size -
+ EXT4_GOOD_OLD_INODE_SIZE -
+ EXT4_I(inode)->i_extra_isize -
+ sizeof(struct ext4_xattr_ibody_header);
+
+ /* Compute min_offs. */
+ for (; !IS_LAST_ENTRY(entry); entry = EXT4_XATTR_NEXT(entry)) {
+ if (!entry->e_value_block && entry->e_value_size) {
+ size_t offs = le16_to_cpu(entry->e_value_offs);
+ if (offs < min_offs)
+ min_offs = offs;
+ }
+ }
+ free = min_offs -
+ ((void *)entry - (void *)IFIRST(header)) - sizeof(__u32);
+
+ if (EXT4_I(inode)->i_inline_off) {
+ entry = (struct ext4_xattr_entry *)
+ ((void *)raw_inode + EXT4_I(inode)->i_inline_off);
+
+ free += le32_to_cpu(entry->e_value_size);
+ goto out;
+ }
+
+ free -= EXT4_XATTR_LEN(strlen(EXT4_XATTR_SYSTEM_DATA_NAME));
+
+ if (free > EXT4_XATTR_ROUND)
+ free = EXT4_XATTR_SIZE(free - EXT4_XATTR_ROUND);
+ else
+ free = 0;
+
+out:
+ up_read(&EXT4_I(inode)->xattr_sem);
+ return free;
+}
+
+/*
+ * Get the maximum size we now can store in an inode.
+ * If we can't find the space for a xattr entry, don't use the space
+ * of the extents since we have no space to indicate the inline data.
+ */
+int ext4_get_max_inline_size(struct inode *inode)
+{
+ int error, max_inline_size;
+ struct ext4_iloc iloc;
+
+ if (EXT4_I(inode)->i_extra_isize == 0)
+ return 0;
+
+ error = ext4_get_inode_loc(inode, &iloc);
+ if (error)
+ return error;
+
+ max_inline_size = get_max_inline_xattr_value_size(inode, &iloc);
+
+ brelse(iloc.bh);
+
+ if (!max_inline_size)
+ return 0;
+
+ return max_inline_size + EXT4_MIN_INLINE_DATA_SIZE;
+}
+
+int ext4_has_inline_data(struct inode *inode)
+{
+ return EXT4_I(inode)->i_inline_off;
+}
+
+int ext4_find_inline_data(struct inode *inode)
+{
+ struct ext4_xattr_ibody_find is = {
+ .s = { .not_found = -ENODATA, },
+ };
+ struct ext4_xattr_info i = {
+ .name_index = EXT4_XATTR_INDEX_SYSTEM_DATA,
+ .name = EXT4_XATTR_SYSTEM_DATA_NAME,
+ };
+ int error;
+
+ if (EXT4_I(inode)->i_extra_isize == 0)
+ return 0;
+
+ error = ext4_get_inode_loc(inode, &is.iloc);
+ if (error)
+ return error;
+
+ error = ext4_xattr_ibody_find(inode, &i, &is);
+ if (error)
+ goto out;
+
+ if (!is.s.not_found) {
+ EXT4_I(inode)->i_inline_off = (u16)((void *)is.s.here -
+ (void *)ext4_raw_inode(&is.iloc));
+ EXT4_I(inode)->i_inline_size = EXT4_MIN_INLINE_DATA_SIZE +
+ le32_to_cpu(is.s.here->e_value_size);
+ ext4_set_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ }
+out:
+ brelse(is.iloc.bh);
+ return error;
+}
+
+int ext4_read_inline_data(struct inode *inode, void *buffer, unsigned int len)
+{
+ struct ext4_xattr_entry *entry;
+ struct ext4_xattr_ibody_header *header;
+ int cp_len = 0, error;
+ struct ext4_inode *raw_inode;
+ struct ext4_iloc iloc;
+
+ if (!len)
+ return 0;
+
+ error = ext4_get_inode_loc(inode, &iloc);
+ if (error)
+ return error;
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ if(!EXT4_I(inode)->i_inline_off) {
+ ext4_warning(inode->i_sb, "inode %lu doesn't have inline data.",
+ inode->i_ino);
+ goto out;
+ }
+ BUG_ON(len > EXT4_I(inode)->i_inline_size);
+
+
+ cp_len = len < EXT4_MIN_INLINE_DATA_SIZE ?
+ len : EXT4_MIN_INLINE_DATA_SIZE;
+
+ raw_inode = ext4_raw_inode(&iloc);
+ memcpy(buffer, (void *)(raw_inode->i_block), cp_len);
+
+ len -= cp_len;
+ buffer += cp_len;
+
+ if (!len)
+ goto out;
+
+ header = IHDR(inode, raw_inode);
+ entry = (struct ext4_xattr_entry *)((void *)raw_inode +
+ EXT4_I(inode)->i_inline_off);
+
+ memcpy(buffer,
+ (void *)IFIRST(header) + le16_to_cpu(entry->e_value_offs), len);
+ cp_len += len;
+
+out:
+ up_read(&EXT4_I(inode)->xattr_sem);
+ brelse(iloc.bh);
+ return cp_len;
+}
+
+/*
+ * write the buffer to the inline inode.
+ * If 'create' is set, we don't need to do the extra copy in the xattr
+ * value since it is already handled by ext4_xattr_ibody_set. That saves
+ * us one memcpy.
+ */
+void ext4_write_inline_data(struct inode *inode, struct ext4_iloc *iloc,
+ void *buffer, loff_t pos, unsigned int len)
+{
+ struct ext4_xattr_entry *entry;
+ struct ext4_xattr_ibody_header *header;
+ struct ext4_inode *raw_inode;
+ int cp_len = 0;
+
+ BUG_ON(!EXT4_I(inode)->i_inline_off);
+ BUG_ON(pos + len > EXT4_I(inode)->i_inline_size);
+
+ raw_inode = ext4_raw_inode(iloc);
+ buffer += pos;
+
+ if (pos < EXT4_MIN_INLINE_DATA_SIZE) {
+ cp_len = pos + len > EXT4_MIN_INLINE_DATA_SIZE ?
+ EXT4_MIN_INLINE_DATA_SIZE - pos : len;
+ memcpy((void *)raw_inode->i_block + pos, buffer, cp_len);
+
+ len -= cp_len;
+ buffer += cp_len;
+ pos += cp_len;
+ }
+
+ if (!len)
+ return;
+
+ pos -= EXT4_MIN_INLINE_DATA_SIZE;
+ header = IHDR(inode, raw_inode);
+ entry = (struct ext4_xattr_entry *)((void *)raw_inode +
+ EXT4_I(inode)->i_inline_off);
+
+ memcpy((void *)IFIRST(header) + le16_to_cpu(entry->e_value_offs) + pos,
+ buffer, len);
+}
+
+static int ext4_create_inline_data(handle_t *handle,
+ struct inode *inode, unsigned len)
+{
+ int error;
+ void *value = NULL;
+ struct ext4_xattr_ibody_find is = {
+ .s = { .not_found = -ENODATA, },
+ };
+ struct ext4_xattr_info i = {
+ .name_index = EXT4_XATTR_INDEX_SYSTEM_DATA,
+ .name = EXT4_XATTR_SYSTEM_DATA_NAME,
+ };
+
+ error = ext4_get_inode_loc(inode, &is.iloc);
+ if (error)
+ return error;
+
+ error = ext4_journal_get_write_access(handle, is.iloc.bh);
+ if (error)
+ goto out;
+
+ if (len > EXT4_MIN_INLINE_DATA_SIZE) {
+ value = (void *)empty_zero_page;
+ len -= EXT4_MIN_INLINE_DATA_SIZE;
+ } else {
+ value = "";
+ len = 0;
+ }
+
+ /* Insert the the xttr entry. */
+ i.value = value;
+ i.value_len = len;
+
+ error = ext4_xattr_ibody_find(inode, &i, &is);
+ if (error)
+ goto out;
+
+ BUG_ON(!is.s.not_found);
+
+ error = ext4_xattr_ibody_set(handle, inode, &i, &is);
+ if (error) {
+ if (error == -ENOSPC)
+ ext4_clear_inode_state(inode,
+ EXT4_STATE_MAY_INLINE_DATA);
+ goto out;
+ }
+
+ memset((void *)ext4_raw_inode(&is.iloc)->i_block,
+ 0, EXT4_MIN_INLINE_DATA_SIZE);
+
+ EXT4_I(inode)->i_inline_off = (u16)((void *)is.s.here -
+ (void *)ext4_raw_inode(&is.iloc));
+ EXT4_I(inode)->i_inline_size = len + EXT4_MIN_INLINE_DATA_SIZE;
+ ext4_set_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ get_bh(is.iloc.bh);
+ error = ext4_mark_iloc_dirty(handle, inode, &is.iloc);
+
+out:
+ brelse(is.iloc.bh);
+ return error;
+}
+
+static int ext4_update_inline_data(handle_t *handle, struct inode *inode,
+ unsigned int len)
+{
+ int error;
+ void *value = NULL;
+ struct ext4_xattr_ibody_find is = {
+ .s = { .not_found = -ENODATA, },
+ };
+ struct ext4_xattr_info i = {
+ .name_index = EXT4_XATTR_INDEX_SYSTEM_DATA,
+ .name = EXT4_XATTR_SYSTEM_DATA_NAME,
+ };
+
+ /* If the old space is ok, write the data directly. */
+ if (len <= EXT4_I(inode)->i_inline_size)
+ return 0;
+
+ error = ext4_get_inode_loc(inode, &is.iloc);
+ if (error)
+ return error;
+
+ error = ext4_xattr_ibody_find(inode, &i, &is);
+ if (error)
+ goto out;
+
+ BUG_ON(is.s.not_found);
+
+ len -= EXT4_MIN_INLINE_DATA_SIZE;
+ value = kzalloc(len, GFP_NOFS);
+ if (!value)
+ goto out;
+
+ error = ext4_xattr_ibody_get(inode, i.name_index, i.name,
+ value, len);
+ if (error == -ENODATA)
+ goto out;
+
+ error = ext4_journal_get_write_access(handle, is.iloc.bh);
+ if (error)
+ goto out;
+
+ /* Update the xttr entry. */
+ i.value = value;
+ i.value_len = len;
+
+ error = ext4_xattr_ibody_set(handle, inode, &i, &is);
+ if (error)
+ goto out;
+
+ EXT4_I(inode)->i_inline_off = (u16)((void *)is.s.here -
+ (void *)ext4_raw_inode(&is.iloc));
+ EXT4_I(inode)->i_inline_size = EXT4_MIN_INLINE_DATA_SIZE +
+ le32_to_cpu(is.s.here->e_value_size);
+ ext4_set_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ get_bh(is.iloc.bh);
+ error = ext4_mark_iloc_dirty(handle, inode, &is.iloc);
+
+out:
+ brelse(is.iloc.bh);
+ return error;
+}
+
+int ext4_prepare_inline_data(handle_t *handle, struct inode *inode,
+ unsigned int len)
+{
+ int ret, size;
+ struct ext4_inode_info *ei = EXT4_I(inode);
+
+ if (!ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA))
+ return -ENOSPC;
+
+ size = ext4_get_max_inline_size(inode);
+ if (size < len)
+ return -ENOSPC;
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+
+ if (ei->i_inline_off)
+ ret = ext4_update_inline_data(handle, inode, len);
+ else
+ ret = ext4_create_inline_data(handle, inode, len);
+
+ up_write(&EXT4_I(inode)->xattr_sem);
+
+ return ret;
+}
+
+static int ext4_destroy_inline_data_nolock(handle_t *handle, struct inode *inode)
+{
+ struct ext4_inode_info *ei = EXT4_I(inode);
+ struct ext4_xattr_ibody_find is = {
+ .s = { .not_found = 0, },
+ };
+ struct ext4_xattr_info i = {
+ .name_index = EXT4_XATTR_INDEX_SYSTEM_DATA,
+ .name = EXT4_XATTR_SYSTEM_DATA_NAME,
+ .value = NULL,
+ .value_len = 0,
+ };
+ int error;
+
+ if (!ei->i_inline_off)
+ return 0;
+
+ error = ext4_get_inode_loc(inode, &is.iloc);
+ if (error)
+ return error;
+
+ error = ext4_xattr_ibody_find(inode, &i, &is);
+ if (error)
+ goto out;
+
+ error = ext4_journal_get_write_access(handle, is.iloc.bh);
+ if (error)
+ goto out;
+
+ error = ext4_xattr_ibody_set(handle, inode, &i, &is);
+ if (error)
+ goto out;
+
+ memset((void *)ext4_raw_inode(&is.iloc)->i_block,
+ 0, EXT4_MIN_INLINE_DATA_SIZE);
+
+ if (EXT4_HAS_INCOMPAT_FEATURE(inode->i_sb,
+ EXT4_FEATURE_INCOMPAT_EXTENTS)) {
+ if (S_ISDIR(inode->i_mode) ||
+ S_ISREG(inode->i_mode) || S_ISLNK(inode->i_mode)) {
+ ext4_set_inode_flag(inode, EXT4_INODE_EXTENTS);
+ ext4_ext_tree_init(handle, inode);
+ }
+ }
+
+ get_bh(is.iloc.bh);
+ error = ext4_mark_iloc_dirty(handle, inode, &is.iloc);
+
+ EXT4_I(inode)->i_inline_off = 0;
+ EXT4_I(inode)->i_inline_size = 0;
+ ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+out:
+ brelse(is.iloc.bh);
+ if (error == -ENODATA)
+ error = 0;
+ return error;
+}
+
+int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
+{
+ int ret;
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+ ret = ext4_destroy_inline_data_nolock(handle, inode);
+ up_write(&EXT4_I(inode)->xattr_sem);
+
+ return ret;
+}
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 15d0460..9b1fbf5 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -3767,8 +3767,10 @@ static inline void ext4_iget_extra_inode(struct inode *inode,
{
__le32 *magic = (void *)raw_inode +
EXT4_GOOD_OLD_INODE_SIZE + ei->i_extra_isize;
- if (*magic == cpu_to_le32(EXT4_XATTR_MAGIC))
+ if (*magic == cpu_to_le32(EXT4_XATTR_MAGIC)) {
ext4_set_inode_state(inode, EXT4_STATE_XATTR);
+ ext4_find_inline_data(inode);
+ }
}
struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index c4c015d..e436d33 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -21,6 +21,7 @@
#define EXT4_XATTR_INDEX_TRUSTED 4
#define EXT4_XATTR_INDEX_LUSTRE 5
#define EXT4_XATTR_INDEX_SECURITY 6
+#define EXT4_XATTR_INDEX_SYSTEM_DATA 7
struct ext4_xattr_header {
__le32 h_magic; /* magic number for identification */
@@ -123,6 +124,21 @@ extern int ext4_xattr_ibody_get(struct inode *inode, int name_index,
const char *name,
void *buffer, size_t buffer_size);
+extern int ext4_has_inline_data(struct inode *inode);
+extern int ext4_get_inline_size(struct inode *inode);
+extern int ext4_get_max_inline_size(struct inode *inode);
+extern int ext4_find_inline_data(struct inode *inode);
+extern int ext4_read_inline_data(struct inode *inode,
+ void *buffer, unsigned int len);
+extern void ext4_write_inline_data(struct inode *inode,
+ struct ext4_iloc *iloc,
+ void *buffer, loff_t pos,
+ unsigned int len);
+extern int ext4_prepare_inline_data(handle_t *handle, struct inode *inode,
+ unsigned int len);
+extern int ext4_init_inline_data(handle_t *handle, struct inode *inode,
+ unsigned int len);
+extern int ext4_destroy_inline_data(handle_t *handle, struct inode *inode);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -198,6 +214,52 @@ extern int ext4_xattr_ibody_get(struct inode *inode, int name_index,
return -EOPNOTSUPP;
}
+static inline int ext4_find_inline_data(struct inode *inode)
+{
+ return 0;
+}
+
+static inline int ext4_has_inline_data(struct inode *inode)
+{
+ return 0;
+}
+
+static inline int ext4_get_inline_size(struct inode *inode)
+{
+ return 0;
+}
+
+static inline int ext4_get_max_inline_size(struct inode *inode)
+{
+ return 0;
+}
+
+static inline int ext4_read_inline_data(struct inode *inode,
+ void *buffer, unsigned int len)
+{
+ return 0;
+}
+
+static inline void ext4_write_inline_data(struct inode *inode,
+ struct ext4_iloc *iloc,
+ void *buffer, loff_t pos,
+ unsigned int len)
+{
+ return;
+}
+
+static inline int ext4_init_inline_data(handle_t *handle,
+ struct inode *inode,
+ unsigned int len)
+{
+ return 0;
+}
+
+static inline int ext4_destroy_inline_data(handle_t *handle,
+ struct inode *inode)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 04/19] ext4: Add read support for inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
2011-11-21 16:39 ` [PATCH V2 02/19] ext4: export inline xattr functions Tao Ma
2011-11-21 16:39 ` [PATCH V2 03/19] ext4: Add the basic function for inline data support Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 05/19] ext4: Add normal write " Tao Ma
` (14 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Let readpage and readpages handle the case when we
want to read an inlined file.
Signed-off-by: Tao Ma <boyu.mt@taobao.comc.
---
fs/ext4/inline.c | 53 ++++++++++++++++++++++++++++++++++++++++++++++-------
fs/ext4/inode.c | 25 +++++++++++++++++++++++++
fs/ext4/xattr.h | 7 +++++++
3 files changed, 78 insertions(+), 7 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index d3aa377..fa8f554 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -163,12 +163,6 @@ int ext4_read_inline_data(struct inode *inode, void *buffer, unsigned int len)
if (error)
return error;
- down_read(&EXT4_I(inode)->xattr_sem);
- if(!EXT4_I(inode)->i_inline_off) {
- ext4_warning(inode->i_sb, "inode %lu doesn't have inline data.",
- inode->i_ino);
- goto out;
- }
BUG_ON(len > EXT4_I(inode)->i_inline_size);
@@ -193,7 +187,6 @@ int ext4_read_inline_data(struct inode *inode, void *buffer, unsigned int len)
cp_len += len;
out:
- up_read(&EXT4_I(inode)->xattr_sem);
brelse(iloc.bh);
return cp_len;
}
@@ -447,6 +440,52 @@ out:
return error;
}
+static int ext4_read_inline_page(struct inode *inode, struct page *page)
+{
+ void *kaddr;
+ loff_t size;
+ int ret = 0;
+
+ BUG_ON(!PageLocked(page));
+ BUG_ON(!ext4_has_inline_data(inode));
+
+ size = i_size_read(inode);
+
+ if (size > PAGE_CACHE_SIZE ||
+ size > ext4_get_inline_size(inode)) {
+ ext4_error_inode(inode, __func__, __LINE__, 0,
+ "bad size %ld for a inline file", (long)size);
+ return -EROFS;
+ }
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ if(!EXT4_I(inode)->i_inline_off) {
+ ext4_warning(inode->i_sb, "inode %lu doesn't have inline data.",
+ inode->i_ino);
+ goto out;
+ }
+
+ kaddr = kmap_atomic(page, KM_USER0);
+ ret = ext4_read_inline_data(inode, kaddr, size);
+ flush_dcache_page(page);
+ kunmap_atomic(kaddr, KM_USER0);
+ SetPageUptodate(page);
+
+out:
+ up_read(&EXT4_I(inode)->xattr_sem);
+ return ret;
+}
+
+int ext4_readpage_inline(struct inode *inode, struct page *page)
+{
+ int ret;
+
+ ret = ext4_read_inline_page(inode, page);
+
+ unlock_page(page);
+ return ret;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 9b1fbf5..4f764b7 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -592,6 +592,9 @@ static int _ext4_get_block(struct inode *inode, sector_t iblock,
int ret = 0, started = 0;
int dio_credits;
+ if (ext4_has_inline_data(inode))
+ return -ERANGE;
+
map.m_lblk = iblock;
map.m_len = bh->b_size >> inode->i_blkbits;
@@ -2633,6 +2636,12 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
journal_t *journal;
int err;
+ /*
+ * XXX: Can we arrive here for a inline file? Maybe not.
+ */
+ if (ext4_has_inline_data(inode))
+ return 0;
+
if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY) &&
test_opt(inode->i_sb, DELALLOC)) {
/*
@@ -2678,7 +2687,13 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
static int ext4_readpage(struct file *file, struct page *page)
{
+ struct inode *inode = file->f_mapping->host;
+
trace_ext4_readpage(page);
+
+ if (ext4_has_inline_data(inode))
+ return ext4_readpage_inline(inode, page);
+
return mpage_readpage(page, ext4_get_block);
}
@@ -2686,6 +2701,12 @@ static int
ext4_readpages(struct file *file, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
+ struct inode *inode = mapping->host;
+
+ /* If the file has inline data, no need to do readpages. */
+ if (ext4_has_inline_data(inode))
+ return 0;
+
return mpage_readpages(mapping, pages, nr_pages, ext4_get_block);
}
@@ -3007,6 +3028,10 @@ static ssize_t ext4_direct_IO(int rw, struct kiocb *iocb,
if (ext4_should_journal_data(inode))
return 0;
+ /* Let buffer I/O handle the inline data case. */
+ if (ext4_has_inline_data(inode))
+ return 0;
+
trace_ext4_direct_IO_enter(inode, offset, iov_length(iov, nr_segs), rw);
if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
ret = ext4_ext_direct_IO(rw, iocb, iov, offset, nr_segs);
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index e436d33..129ead1 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -139,6 +139,8 @@ extern int ext4_prepare_inline_data(handle_t *handle, struct inode *inode,
extern int ext4_init_inline_data(handle_t *handle, struct inode *inode,
unsigned int len);
extern int ext4_destroy_inline_data(handle_t *handle, struct inode *inode);
+
+extern int ext4_readpage_inline(struct inode *inode, struct page *page);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -260,6 +262,11 @@ static inline int ext4_destroy_inline_data(handle_t *handle,
{
return 0;
}
+
+static inline int ext4_readpage_inline(struct inode *inode, struct page *page)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 05/19] ext4: Add normal write support for inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (2 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 04/19] ext4: Add read support for inline data Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 06/19] ext4: Add journalled " Tao Ma
` (13 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
For a normal write case(not journalled write, not delayed allocation),
we write to the inline if the file is small and convert it to an extent
based file when the write is larger than the max inline size.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 11 +++
fs/ext4/extents.c | 9 ++-
fs/ext4/inline.c | 223 +++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/inode.c | 40 +++++++---
fs/ext4/xattr.h | 26 ++++++
5 files changed, 297 insertions(+), 12 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 3e560ba..d9603c1 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1861,8 +1861,19 @@ struct buffer_head *ext4_getblk(handle_t *, struct inode *,
ext4_lblk_t, int, int *);
struct buffer_head *ext4_bread(handle_t *, struct inode *,
ext4_lblk_t, int, int *);
+int ext4_get_block_write(struct inode *inode, sector_t iblock,
+ struct buffer_head *bh_result, int create);
int ext4_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create);
+int walk_page_buffers(handle_t *handle,
+ struct buffer_head *head,
+ unsigned from,
+ unsigned to,
+ int *partial,
+ int (*fn)(handle_t *handle,
+ struct buffer_head *bh));
+int do_journal_get_write_access(handle_t *handle,
+ struct buffer_head *bh);
extern struct inode *ext4_iget(struct super_block *, unsigned long);
extern int ext4_write_inode(struct inode *, struct writeback_control *);
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 61fa9e1..a0063b7 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -42,6 +42,7 @@
#include <asm/uaccess.h>
#include <linux/fiemap.h>
#include "ext4_jbd2.h"
+#include "xattr.h"
#include <trace/events/ext4.h>
@@ -2178,7 +2179,13 @@ int ext4_ext_calc_credits_for_single_extent(struct inode *inode, int nrblocks,
int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks, int chunk)
{
int index;
- int depth = ext_depth(inode);
+ int depth;
+
+ /* If we are converting the inline data, only one is needed here. */
+ if (ext4_has_inline_data(inode))
+ return 1;
+
+ depth = ext_depth(inode);
if (chunk)
index = depth * 2;
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index fa8f554..1d1572b 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -14,6 +14,7 @@
#include "ext4_jbd2.h"
#include "ext4.h"
#include "xattr.h"
+#include "truncate.h"
#define EXT4_XATTR_SYSTEM_DATA_NAME "data"
#define EXT4_MIN_INLINE_DATA_SIZE ((sizeof(__le32) * EXT4_N_BLOCKS))
@@ -486,6 +487,228 @@ int ext4_readpage_inline(struct inode *inode, struct page *page)
return ret;
}
+static int ext4_convert_inline_data_to_extent(struct address_space *mapping,
+ struct inode *inode,
+ unsigned flags)
+{
+ int ret, needed_blocks;
+ handle_t *handle = NULL;
+ int retries = 0;
+ struct page *page = NULL;
+ unsigned from, to;
+ struct ext4_iloc iloc;
+
+ if (!ext4_has_inline_data(inode)) {
+ /*
+ * clear the flag so that no new write
+ * will trap here again.
+ */
+ ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ return 0;
+ }
+
+ needed_blocks = ext4_writepage_trans_blocks(inode);
+ from = 0;
+ to = ext4_get_inline_size(inode);
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret)
+ return ret;
+
+retry:
+ handle = ext4_journal_start(inode, needed_blocks);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ handle = NULL;
+ goto out;
+ }
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+ /* If some one has already done this for us, just exit. */
+ if (!ext4_has_inline_data(inode))
+ goto out;
+
+ /* We cannot recurse into the filesystem as the transaction is already
+ * started */
+ flags |= AOP_FLAG_NOFS;
+
+ page = grab_cache_page_write_begin(mapping, 0, flags);
+ if (!page) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ if (!PageUptodate(page)) {
+ void *kaddr = kmap_atomic(page, KM_USER0);
+ ret = ext4_read_inline_data(inode, kaddr,
+ i_size_read(inode));
+ flush_dcache_page(page);
+ kunmap_atomic(kaddr, KM_USER0);
+ SetPageUptodate(page);
+ if (ret < 0)
+ goto out;
+ }
+
+ ret = ext4_destroy_inline_data_nolock(handle, inode);
+ if (ret)
+ goto out;
+
+ if (ext4_should_dioread_nolock(inode))
+ ret = __block_write_begin(page, from, to, ext4_get_block_write);
+ else
+ ret = __block_write_begin(page, from, to, ext4_get_block);
+
+ if (!ret && ext4_should_journal_data(inode)) {
+ ret = walk_page_buffers(handle, page_buffers(page),
+ from, to, NULL, do_journal_get_write_access);
+ }
+
+ if (ret) {
+ unlock_page(page);
+ page_cache_release(page);
+ ext4_orphan_add(handle, inode);
+ up_write(&EXT4_I(inode)->xattr_sem);
+ ext4_journal_stop(handle);
+ handle = NULL;
+ ext4_truncate_failed_write(inode);
+ /*
+ * If truncate failed early the inode might
+ * still be on the orphan list; we need to
+ * make sure the inode is removed from the
+ * orphan list in that case.
+ */
+ if (inode->i_nlink)
+ ext4_orphan_del(NULL, inode);
+ }
+
+ if (ret == -ENOSPC && ext4_should_retry_alloc(inode->i_sb, &retries))
+ goto retry;
+
+ block_commit_write(page, from, to);
+out:
+ if (page) {
+ unlock_page(page);
+ page_cache_release(page);
+ }
+ up_write(&EXT4_I(inode)->xattr_sem);
+ if (handle)
+ ext4_journal_stop(handle);
+ brelse(iloc.bh);
+ return ret;
+}
+
+/*
+ * Try to write data in the inode.
+ * If the inode has inline data, check whether the new write can be
+ * in the inode also. If not, create the page the handle, move the data
+ * to the page make it update and let the later codes create extent for it.
+ */
+int ext4_try_to_write_inline_data(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep)
+{
+ int ret;
+ handle_t *handle;
+ struct page *page;
+ struct ext4_iloc iloc;
+
+ if (pos + len > ext4_get_max_inline_size(inode))
+ goto convert;
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret)
+ return ret;
+
+ /*
+ * The possible write could happen in the inode,
+ * so try to reserve the space in inode first.
+ */
+ handle = ext4_journal_start(inode, 1);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ handle = NULL;
+ goto out;
+ }
+
+ ret = ext4_prepare_inline_data(handle, inode, pos + len);
+ if (ret && ret != -ENOSPC)
+ goto out;
+
+ /* We don't have space in inline inode, so convert it to extent. */
+ if (ret == -ENOSPC) {
+ ext4_journal_stop(handle);
+ brelse(iloc.bh);
+ goto convert;
+ }
+
+ flags |= AOP_FLAG_NOFS;
+
+ page = grab_cache_page_write_begin(mapping, 0, flags);
+ if (!page) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ *pagep = page;
+
+ if (!PageUptodate(page)) {
+ ret = ext4_read_inline_page(inode, page);
+ if (ret < 0)
+ goto out;
+ }
+
+ ret = 1;
+ handle = NULL;
+out:
+ if (handle)
+ ext4_journal_stop(handle);
+ brelse(iloc.bh);
+ return ret;
+convert:
+ return ext4_convert_inline_data_to_extent(mapping,
+ inode, flags);
+}
+
+int ext4_write_inline_data_end(struct inode *inode, loff_t pos, unsigned len,
+ unsigned copied, struct page *page)
+{
+ int ret;
+ void *kaddr;
+ struct ext4_iloc iloc;
+
+ if (unlikely(copied < len)) {
+ if (!PageUptodate(page)) {
+ copied = 0;
+ goto out;
+ }
+ }
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret) {
+ ext4_std_error(inode->i_sb, ret);
+ copied = 0;
+ goto out;
+ }
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+ BUG_ON(!ext4_has_inline_data(inode));
+
+ kaddr = kmap_atomic(page, KM_USER0);
+ ext4_write_inline_data(inode, &iloc, kaddr, pos, len);
+ kunmap_atomic(kaddr, KM_USER0);
+ SetPageUptodate(page);
+ /* clear page dirty so that writepages wouldn't work for us. */
+ ClearPageDirty(page);
+
+ up_write(&EXT4_I(inode)->xattr_sem);
+ brelse(iloc.bh);
+out:
+ return copied;
+}
+
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 4f764b7..26ca9a7 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -711,7 +711,7 @@ struct buffer_head *ext4_bread(handle_t *handle, struct inode *inode,
return NULL;
}
-static int walk_page_buffers(handle_t *handle,
+int walk_page_buffers(handle_t *handle,
struct buffer_head *head,
unsigned from,
unsigned to,
@@ -767,8 +767,8 @@ static int walk_page_buffers(handle_t *handle,
* is elevated. We'll still have enough credits for the tiny quotafile
* write.
*/
-static int do_journal_get_write_access(handle_t *handle,
- struct buffer_head *bh)
+int do_journal_get_write_access(handle_t *handle,
+ struct buffer_head *bh)
{
int dirty = buffer_dirty(bh);
int ret;
@@ -791,8 +791,6 @@ static int do_journal_get_write_access(handle_t *handle,
return ret;
}
-static int ext4_get_block_write(struct inode *inode, sector_t iblock,
- struct buffer_head *bh_result, int create);
static int ext4_write_begin(struct file *file, struct address_space *mapping,
loff_t pos, unsigned len, unsigned flags,
struct page **pagep, void **fsdata)
@@ -815,6 +813,17 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping,
from = pos & (PAGE_CACHE_SIZE - 1);
to = from + len;
+ if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) {
+ ret = ext4_try_to_write_inline_data(mapping, inode, pos, len,
+ flags, pagep);
+ if (ret < 0)
+ goto out;
+ if (ret == 1) {
+ ret = 0;
+ goto out;
+ }
+ }
+
retry:
handle = ext4_journal_start(inode, needed_blocks);
if (IS_ERR(handle)) {
@@ -832,6 +841,7 @@ retry:
ret = -ENOMEM;
goto out;
}
+
*pagep = page;
if (ext4_should_dioread_nolock(inode))
@@ -896,7 +906,12 @@ static int ext4_generic_write_end(struct file *file,
struct inode *inode = mapping->host;
handle_t *handle = ext4_journal_current_handle();
- copied = block_write_end(file, mapping, pos, len, copied, page, fsdata);
+ if (ext4_has_inline_data(inode))
+ copied = ext4_write_inline_data_end(inode, pos, len,
+ copied, page);
+ else
+ copied = block_write_end(file, mapping, pos,
+ len, copied, page, fsdata);
/*
* No need to use i_size_read() here, the i_size
@@ -2773,7 +2788,7 @@ static int ext4_releasepage(struct page *page, gfp_t wait)
* We allocate an uinitialized extent if blocks haven't been allocated.
* The extent will be converted to initialized after the IO is complete.
*/
-static int ext4_get_block_write(struct inode *inode, sector_t iblock,
+int ext4_get_block_write(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create)
{
ext4_debug("ext4_get_block_write: inode %lu, create flag %d\n",
@@ -3795,7 +3810,8 @@ static inline void ext4_iget_extra_inode(struct inode *inode,
if (*magic == cpu_to_le32(EXT4_XATTR_MAGIC)) {
ext4_set_inode_state(inode, EXT4_STATE_XATTR);
ext4_find_inline_data(inode);
- }
+ } else
+ EXT4_I(inode)->i_inline_off = 0;
}
struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
@@ -3933,9 +3949,10 @@ struct inode *ext4_iget(struct super_block *sb, unsigned long ino)
ret = -EIO;
goto bad_inode;
} else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
- if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
+ if (!EXT4_I(inode)->i_inline_off &&
+ (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
(S_ISLNK(inode->i_mode) &&
- !ext4_inode_is_fast_symlink(inode)))
+ !ext4_inode_is_fast_symlink(inode))))
/* Validate extent which is part of inode */
ret = ext4_ext_check_inode(inode);
} else if (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) ||
@@ -4125,9 +4142,10 @@ static int ext4_do_update_inode(handle_t *handle,
cpu_to_le32(new_encode_dev(inode->i_rdev));
raw_inode->i_block[2] = 0;
}
- } else
+ } else if (!ext4_has_inline_data(inode)) {
for (block = 0; block < EXT4_N_BLOCKS; block++)
raw_inode->i_block[block] = ei->i_data[block];
+ }
raw_inode->i_disk_version = cpu_to_le32(inode->i_version);
if (ei->i_extra_isize) {
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 129ead1..f2d9c53 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -141,6 +141,15 @@ extern int ext4_init_inline_data(handle_t *handle, struct inode *inode,
extern int ext4_destroy_inline_data(handle_t *handle, struct inode *inode);
extern int ext4_readpage_inline(struct inode *inode, struct page *page);
+extern int ext4_try_to_write_inline_data(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep);
+extern int ext4_write_inline_data_end(struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned copied,
+ struct page *page);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -267,6 +276,23 @@ static inline int ext4_readpage_inline(struct inode *inode, struct page *page)
{
return 0;
}
+
+static inline int ext4_try_to_write_inline_data(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep)
+{
+ return 0;
+}
+
+static inline int ext4_write_inline_data_end(struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned copied,
+ struct page *page)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 06/19] ext4: Add journalled write support for inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (3 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 05/19] ext4: Add normal write " Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 07/19] ext4: Add delalloc " Tao Ma
` (12 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/inline.c | 24 ++++++++++++++++++++++
fs/ext4/inode.c | 59 +++++++++++++++++++++++++++++++++++++----------------
fs/ext4/xattr.h | 12 +++++++++++
3 files changed, 77 insertions(+), 18 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 1d1572b..e14d8e8 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -708,6 +708,30 @@ out:
return copied;
}
+struct buffer_head *
+ext4_journalled_write_inline_data(struct inode *inode,
+ unsigned len,
+ struct page *page)
+{
+ int ret;
+ void *kaddr;
+ struct ext4_iloc iloc;
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret) {
+ ext4_std_error(inode->i_sb, ret);
+ return NULL;
+ }
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+ kaddr = kmap_atomic(page, KM_USER0);
+ ext4_write_inline_data(inode, &iloc, kaddr, 0, len);
+ kunmap_atomic(kaddr, KM_USER0);
+ up_write(&EXT4_I(inode)->xattr_sem);
+
+ return iloc.bh;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 26ca9a7..32077cf 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1062,16 +1062,21 @@ static int ext4_journalled_write_end(struct file *file,
BUG_ON(!ext4_handle_valid(handle));
- if (copied < len) {
- if (!PageUptodate(page))
- copied = 0;
- page_zero_new_buffers(page, from+copied, to);
- }
+ if (ext4_has_inline_data(inode))
+ copied = ext4_write_inline_data_end(inode, pos, len,
+ copied, page);
+ else {
+ if (copied < len) {
+ if (!PageUptodate(page))
+ copied = 0;
+ page_zero_new_buffers(page, from+copied, to);
+ }
- ret = walk_page_buffers(handle, page_buffers(page), from,
- to, &partial, write_end_fn);
- if (!partial)
- SetPageUptodate(page);
+ ret = walk_page_buffers(handle, page_buffers(page), from,
+ to, &partial, write_end_fn);
+ if (!partial)
+ SetPageUptodate(page);
+ }
new_i_size = pos + copied;
if (new_i_size > inode->i_size)
i_size_write(inode, pos+copied);
@@ -1830,15 +1835,23 @@ static int __ext4_journalled_writepage(struct page *page,
{
struct address_space *mapping = page->mapping;
struct inode *inode = mapping->host;
- struct buffer_head *page_bufs;
+ struct buffer_head *page_bufs = NULL;
handle_t *handle = NULL;
int ret = 0;
int err;
+ struct buffer_head *inode_bh = NULL;
ClearPageChecked(page);
- page_bufs = page_buffers(page);
- BUG_ON(!page_bufs);
- walk_page_buffers(handle, page_bufs, 0, len, NULL, bget_one);
+
+ if (ext4_has_inline_data(inode)) {
+ BUG_ON(page->index != 0);
+ BUG_ON(len > ext4_get_max_inline_size(inode));
+ inode_bh = ext4_journalled_write_inline_data(inode, len, page);
+ } else {
+ page_bufs = page_buffers(page);
+ BUG_ON(!page_bufs);
+ walk_page_buffers(handle, page_bufs, 0, len, NULL, bget_one);
+ }
/* As soon as we unlock the page, it can go away, but we have
* references to buffers so we are safe */
unlock_page(page);
@@ -1851,11 +1864,19 @@ static int __ext4_journalled_writepage(struct page *page,
BUG_ON(!ext4_handle_valid(handle));
- ret = walk_page_buffers(handle, page_bufs, 0, len, NULL,
- do_journal_get_write_access);
+ if (ext4_has_inline_data(inode) && inode_bh) {
+ ret = ext4_journal_get_write_access(handle, inode_bh);
+
+ err = ext4_handle_dirty_metadata(handle, inode, inode_bh);
+
+ } else {
+ ret = walk_page_buffers(handle, page_bufs, 0, len, NULL,
+ do_journal_get_write_access);
+
+ err = walk_page_buffers(handle, page_bufs, 0, len, NULL,
+ write_end_fn);
+ }
- err = walk_page_buffers(handle, page_bufs, 0, len, NULL,
- write_end_fn);
if (ret == 0)
ret = err;
EXT4_I(inode)->i_datasync_tid = handle->h_transaction->t_tid;
@@ -1863,9 +1884,11 @@ static int __ext4_journalled_writepage(struct page *page,
if (!ret)
ret = err;
- walk_page_buffers(handle, page_bufs, 0, len, NULL, bput_one);
+ if (!ext4_has_inline_data(inode))
+ walk_page_buffers(handle, page_bufs, 0, len, NULL, bput_one);
ext4_set_inode_state(inode, EXT4_STATE_JDATA);
out:
+ brelse(inode_bh);
return ret;
}
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index f2d9c53..85dbcc0 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -150,6 +150,10 @@ extern int ext4_write_inline_data_end(struct inode *inode,
loff_t pos, unsigned len,
unsigned copied,
struct page *page);
+extern struct buffer_head *
+ext4_journalled_write_inline_data(struct inode *inode,
+ unsigned len,
+ struct page *page);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -293,6 +297,14 @@ static inline int ext4_write_inline_data_end(struct inode *inode,
{
return 0;
}
+
+static inline struct buffer_head *
+ext4_journalled_write_inline_data(struct inode *inode,
+ unsigned len,
+ struct page *page)
+{
+ return NULL;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 07/19] ext4: Add delalloc support for inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (4 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 06/19] ext4: Add journalled " Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 08/19] ext4: Create a new function ext4_init_new_dir Tao Ma
` (11 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
For delayed allocation mode, we write to inline data if the file
is small enough. And in case of we write to some offset larger
than the inline size, the 1st page is dirtied, so that
ext4_da_writepages can handle the conversion. When the 1st page
is initialized with blocks, the inline part is removed.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 4 +
fs/ext4/inline.c | 175 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/inode.c | 66 ++++++++++++++++----
fs/ext4/xattr.h | 27 ++++++++
4 files changed, 258 insertions(+), 14 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index d9603c1..3df7d5a 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1865,6 +1865,8 @@ int ext4_get_block_write(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create);
int ext4_get_block(struct inode *inode, sector_t iblock,
struct buffer_head *bh_result, int create);
+int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
+ struct buffer_head *bh, int create);
int walk_page_buffers(handle_t *handle,
struct buffer_head *head,
unsigned from,
@@ -1874,6 +1876,8 @@ int walk_page_buffers(handle_t *handle,
struct buffer_head *bh));
int do_journal_get_write_access(handle_t *handle,
struct buffer_head *bh);
+#define FALL_BACK_TO_NONDELALLOC 1
+#define CONVERT_INLINE_DATA 2
extern struct inode *ext4_iget(struct super_block *, unsigned long);
extern int ext4_write_inode(struct inode *, struct writeback_control *);
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index e14d8e8..f2a212d 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -732,6 +732,181 @@ ext4_journalled_write_inline_data(struct inode *inode,
return iloc.bh;
}
+/*
+ * Try to make the page cache and handle ready for the inline data case.
+ * We can call this function in 2 cases:
+ * 1. The inode is created and the first write exceeds inline size. We can
+ * clear the inode state safely.
+ * 2. The inode has inline data, then we need to read the data, make it
+ * update and dirty so that ext4_da_writepages can handle it. We don't
+ * need to start the journal since the file's metatdata isn't changed now.
+ */
+static int ext4_da_convert_inline_data_to_extent(struct address_space *mapping,
+ struct inode *inode,
+ unsigned flags,
+ void **fsdata)
+{
+ int ret, inline_size;
+ struct page *page;
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ if (!ext4_has_inline_data(inode)) {
+ ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ up_read(&EXT4_I(inode)->xattr_sem);
+ return 0;
+ }
+
+ inline_size = ext4_get_max_inline_size(inode);
+
+ page = grab_cache_page_write_begin(mapping, 0, flags);
+ if (!page) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ if (!PageUptodate(page)) {
+ ret = ext4_read_inline_page(inode, page);
+ if (ret < 0)
+ goto out;
+ }
+
+ ret = __block_write_begin(page, 0, inline_size,
+ ext4_da_get_block_prep);
+ if (ret) {
+ ext4_truncate_failed_write(inode);
+ goto out;
+ }
+
+ SetPageDirty(page);
+ SetPageUptodate(page);
+ ext4_clear_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+ *fsdata = (void *)CONVERT_INLINE_DATA;
+
+out:
+ if (page) {
+ unlock_page(page);
+ page_cache_release(page);
+ }
+ up_read(&EXT4_I(inode)->xattr_sem);
+ return ret;
+}
+
+/*
+ * Prepare the write for the inline data.
+ * If the the data can be written into the inode, we just read
+ * the page and make it uptodate, and start the journal.
+ * Otherwise read the page, makes it dirty so that it can be
+ * handle in writepages(the i_disksize update is left to the
+ * normal ext4_da_write_end).
+ */
+int ext4_da_write_inline_data_begin(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep,
+ void **fsdata)
+{
+ int ret, inline_size;
+ handle_t *handle;
+ struct page *page;
+ struct ext4_iloc iloc;
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret)
+ return ret;
+
+ handle = ext4_journal_start(inode, 1);
+ if (IS_ERR(handle)) {
+ ret = PTR_ERR(handle);
+ handle = NULL;
+ goto out;
+ }
+
+ inline_size = ext4_get_max_inline_size(inode);
+
+ ret = -ENOSPC;
+ if (inline_size >= pos + len) {
+ ret = ext4_prepare_inline_data(handle, inode, pos + len);
+ if (ret && ret != -ENOSPC)
+ goto out;
+ }
+
+ if (ret == -ENOSPC) {
+ ret = ext4_da_convert_inline_data_to_extent(mapping,
+ inode,
+ flags,
+ fsdata);
+ goto out;
+ }
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ /*
+ * We cannot recurse into the filesystem as the transaction
+ * is already started.
+ */
+ flags |= AOP_FLAG_NOFS;
+
+ page = grab_cache_page_write_begin(mapping, 0, flags);
+ if (!page) {
+ ret = -ENOMEM;
+ goto out_up_read;
+ }
+
+ if (!PageUptodate(page)) {
+ ret = ext4_read_inline_page(inode, page);
+ if (ret < 0)
+ goto out_release_page;
+ }
+
+ up_read(&EXT4_I(inode)->xattr_sem);
+ *pagep = page;
+ handle = NULL;
+ return 1;
+out_release_page:
+ unlock_page(page);
+ page_cache_release(page);
+out_up_read:
+ up_read(&EXT4_I(inode)->xattr_sem);
+out:
+ if (handle)
+ ext4_journal_stop(handle);
+ brelse(iloc.bh);
+ return ret;
+}
+
+int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+ unsigned len, unsigned copied,
+ struct page *page)
+{
+ int i_size_changed = 0;
+
+ copied = ext4_write_inline_data_end(inode, pos, len, copied, page);
+
+ /*
+ * No need to use i_size_read() here, the i_size
+ * cannot change under us because we hold i_mutex.
+ *
+ * But it's important to update i_size while still holding page lock:
+ * page writeout could otherwise come in and zero beyond i_size.
+ */
+ if (pos+copied > inode->i_size) {
+ i_size_write(inode, pos+copied);
+ i_size_changed = 1;
+ }
+ unlock_page(page);
+ page_cache_release(page);
+
+ /*
+ * Don't mark the inode dirty under page lock. First, it unnecessarily
+ * makes the holding time of page lock longer. Second, it forces lock
+ * ordering of page lock and transaction start for journaling
+ * filesystems.
+ */
+ if (i_size_changed)
+ mark_inode_dirty(inode);
+
+ return copied;
+}
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 32077cf..9fd344b 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1713,7 +1713,13 @@ static int ext4_da_map_blocks(struct inode *inode, sector_t iblock,
* file system block.
*/
down_read((&EXT4_I(inode)->i_data_sem));
- if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+ if (ext4_has_inline_data(inode)) {
+ /*
+ * We will soon create blocks for this page, and let
+ * us pretend as if the blocks aren't allocated yet.
+ */
+ retval = 0;
+ } else if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
retval = ext4_ext_map_blocks(NULL, inode, map, 0);
else
retval = ext4_ind_map_blocks(NULL, inode, map, 0);
@@ -1760,8 +1766,8 @@ out_unlock:
* We also have b_blocknr = physicalblock mapping unwritten extent and b_bdev
* initialized properly.
*/
-static int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
- struct buffer_head *bh, int create)
+int ext4_da_get_block_prep(struct inode *inode, sector_t iblock,
+ struct buffer_head *bh, int create)
{
struct ext4_map_blocks map;
int ret = 0;
@@ -2035,7 +2041,8 @@ static int ext4_da_writepages_trans_blocks(struct inode *inode)
* mpage_da_map_and_submit to map a single contiguous memory region
* and then write them.
*/
-static int write_cache_pages_da(struct address_space *mapping,
+static int write_cache_pages_da(handle_t *handle,
+ struct address_space *mapping,
struct writeback_control *wbc,
struct mpage_da_data *mpd,
pgoff_t *done_index)
@@ -2114,6 +2121,17 @@ static int write_cache_pages_da(struct address_space *mapping,
wait_on_page_writeback(page);
BUG_ON(PageWriteback(page));
+ /*
+ * If we have inline data and arrive here, it means that
+ * we will soon create the block for the 1st page, so
+ * we'd better clear the inline data here.
+ */
+ if (ext4_has_inline_data(inode)) {
+ BUG_ON(ext4_test_inode_state(inode,
+ EXT4_STATE_MAY_INLINE_DATA));
+ ext4_destroy_inline_data(handle, inode);
+ }
+
if (mpd->next_page != page->index)
mpd->first_page = page->index;
mpd->next_page = page->index + 1;
@@ -2319,7 +2337,8 @@ retry:
* contiguous region of logical blocks that need
* blocks to be allocated by ext4 and submit them.
*/
- ret = write_cache_pages_da(mapping, wbc, &mpd, &done_index);
+ ret = write_cache_pages_da(handle, mapping,
+ wbc, &mpd, &done_index);
/*
* If we have a contiguous extent of pages and we
* haven't done the I/O yet, map the blocks and submit
@@ -2383,7 +2402,6 @@ out_writepages:
return ret;
}
-#define FALL_BACK_TO_NONDELALLOC 1
static int ext4_nonda_switch(struct super_block *sb)
{
s64 free_blocks, dirty_blocks;
@@ -2438,6 +2456,19 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping,
}
*fsdata = (void *)0;
trace_ext4_da_write_begin(inode, pos, len, flags);
+
+ if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) {
+ ret = ext4_da_write_inline_data_begin(mapping, inode,
+ pos, len, flags,
+ pagep, fsdata);
+ if (ret < 0)
+ goto out;
+ if (ret == 1) {
+ ret = 0;
+ goto out;
+ }
+ }
+
retry:
/*
* With delayed allocation, we don't log the i_disksize update
@@ -2546,10 +2577,10 @@ static int ext4_da_write_end(struct file *file,
* changes. So let's piggyback the i_disksize mark_inode_dirty
* into that.
*/
-
new_i_size = pos + copied;
if (new_i_size > EXT4_I(inode)->i_disksize) {
- if (ext4_da_should_update_i_disksize(page, end)) {
+ if (ext4_has_inline_data(inode) ||
+ ext4_da_should_update_i_disksize(page, end)) {
down_write(&EXT4_I(inode)->i_data_sem);
if (new_i_size > EXT4_I(inode)->i_disksize) {
/*
@@ -2570,16 +2601,23 @@ static int ext4_da_write_end(struct file *file,
ext4_mark_inode_dirty(handle, inode);
}
}
- ret2 = generic_write_end(file, mapping, pos, len, copied,
+
+
+ if (write_mode != CONVERT_INLINE_DATA && ext4_has_inline_data(inode))
+ ret2 = ext4_da_write_inline_data_end(inode, pos, len, copied,
+ page);
+ else {
+ ret2 = generic_write_end(file, mapping, pos, len, copied,
page, fsdata);
- page_len = PAGE_CACHE_SIZE -
+ page_len = PAGE_CACHE_SIZE -
((pos + copied - 1) & (PAGE_CACHE_SIZE - 1));
- if (page_len > 0) {
- ret = ext4_discard_partial_page_buffers_no_lock(handle,
- inode, page, pos + copied - 1, page_len,
- EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED);
+ if (page_len > 0) {
+ ret = ext4_discard_partial_page_buffers_no_lock(handle,
+ inode, page, pos + copied - 1, page_len,
+ EXT4_DISCARD_PARTIAL_PG_ZERO_UNMAPPED);
+ }
}
copied = ret2;
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 85dbcc0..c390402 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -154,6 +154,15 @@ extern struct buffer_head *
ext4_journalled_write_inline_data(struct inode *inode,
unsigned len,
struct page *page);
+extern int ext4_da_write_inline_data_begin(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep,
+ void **fsdata);
+extern int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+ unsigned len, unsigned copied,
+ struct page *page);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -305,6 +314,24 @@ ext4_journalled_write_inline_data(struct inode *inode,
{
return NULL;
}
+
+static inline int
+ext4_da_write_inline_data_begin(struct address_space *mapping,
+ struct inode *inode,
+ loff_t pos, unsigned len,
+ unsigned flags,
+ struct page **pagep,
+ void **fsdata)
+{
+ return 0;
+}
+
+static inline int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
+ unsigned len, unsigned copied,
+ struct page *page)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 08/19] ext4: Create a new function ext4_init_new_dir.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (5 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 07/19] ext4: Add delalloc " Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 09/19] ext4: Refactor __ext4_check_dir_entry to accepts start and size Tao Ma
` (10 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Currently, the initialization of dot and dotdot are encapsulated in
ext4_mkdir and also bond with dir_block. So create a new function
named ext4_init_new_dir and the initialization is moved to
ext4_init_dot_dotdot which only accepts a 'de'.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 2 +
fs/ext4/namei.c | 74 +++++++++++++++++++++++++++++++++---------------------
2 files changed, 47 insertions(+), 29 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 3df7d5a..27c5637 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -2254,6 +2254,8 @@ extern loff_t ext4_llseek(struct file *file, loff_t offset, int origin);
extern const struct inode_operations ext4_dir_inode_operations;
extern const struct inode_operations ext4_special_inode_operations;
extern struct dentry *ext4_get_parent(struct dentry *child);
+void ext4_init_dot_dotdot(struct inode *parent, struct inode *inode,
+ struct ext4_dir_entry_2 *de, int blocksize);
/* symlink.c */
extern const struct inode_operations ext4_symlink_inode_operations;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index aa4c782..14ecbd2 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1806,13 +1806,54 @@ retry:
return err;
}
+void ext4_init_dot_dotdot(struct inode *parent, struct inode *inode,
+ struct ext4_dir_entry_2 *de, int blocksize)
+{
+ de->inode = cpu_to_le32(inode->i_ino);
+ de->name_len = 1;
+ de->rec_len = ext4_rec_len_to_disk(EXT4_DIR_REC_LEN(de->name_len),
+ blocksize);
+ strcpy(de->name, ".");
+ ext4_set_de_type(parent->i_sb, de, S_IFDIR);
+ de = ext4_next_entry(de, blocksize);
+ de->inode = cpu_to_le32(parent->i_ino);
+ de->rec_len = ext4_rec_len_to_disk(blocksize - EXT4_DIR_REC_LEN(1),
+ blocksize);
+ de->name_len = 2;
+ strcpy(de->name, "..");
+ ext4_set_de_type(parent->i_sb, de, S_IFDIR);
+ set_nlink(inode, 2);
+}
+
+static int ext4_init_new_dir(handle_t *handle, struct inode *parent,
+ struct inode *inode)
+{
+ struct buffer_head *dir_block = NULL;
+ struct ext4_dir_entry_2 *de;
+ int err;
+ int blocksize = inode->i_sb->s_blocksize;
+
+ inode->i_size = EXT4_I(inode)->i_disksize = blocksize;
+ dir_block = ext4_bread(handle, inode, 0, 1, &err);
+ if (!dir_block)
+ goto out;
+ BUFFER_TRACE(dir_block, "get_write_access");
+ err = ext4_journal_get_write_access(handle, dir_block);
+ if (err)
+ goto out;
+ de = (struct ext4_dir_entry_2 *)dir_block->b_data;
+ ext4_init_dot_dotdot(parent, inode, de, blocksize);
+ BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
+ err = ext4_handle_dirty_metadata(handle, inode, dir_block);
+out:
+ brelse(dir_block);
+ return err;
+}
+
static int ext4_mkdir(struct inode *dir, struct dentry *dentry, int mode)
{
handle_t *handle;
struct inode *inode;
- struct buffer_head *dir_block = NULL;
- struct ext4_dir_entry_2 *de;
- unsigned int blocksize = dir->i_sb->s_blocksize;
int err, retries = 0;
if (EXT4_DIR_LINK_MAX(dir))
@@ -1838,31 +1879,7 @@ retry:
inode->i_op = &ext4_dir_inode_operations;
inode->i_fop = &ext4_dir_operations;
- inode->i_size = EXT4_I(inode)->i_disksize = inode->i_sb->s_blocksize;
- dir_block = ext4_bread(handle, inode, 0, 1, &err);
- if (!dir_block)
- goto out_clear_inode;
- BUFFER_TRACE(dir_block, "get_write_access");
- err = ext4_journal_get_write_access(handle, dir_block);
- if (err)
- goto out_clear_inode;
- de = (struct ext4_dir_entry_2 *) dir_block->b_data;
- de->inode = cpu_to_le32(inode->i_ino);
- de->name_len = 1;
- de->rec_len = ext4_rec_len_to_disk(EXT4_DIR_REC_LEN(de->name_len),
- blocksize);
- strcpy(de->name, ".");
- ext4_set_de_type(dir->i_sb, de, S_IFDIR);
- de = ext4_next_entry(de, blocksize);
- de->inode = cpu_to_le32(dir->i_ino);
- de->rec_len = ext4_rec_len_to_disk(blocksize - EXT4_DIR_REC_LEN(1),
- blocksize);
- de->name_len = 2;
- strcpy(de->name, "..");
- ext4_set_de_type(dir->i_sb, de, S_IFDIR);
- set_nlink(inode, 2);
- BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
- err = ext4_handle_dirty_metadata(handle, inode, dir_block);
+ err = ext4_init_new_dir(handle, dir, inode);
if (err)
goto out_clear_inode;
err = ext4_mark_inode_dirty(handle, inode);
@@ -1884,7 +1901,6 @@ out_clear_inode:
d_instantiate(dentry, inode);
unlock_new_inode(inode);
out_stop:
- brelse(dir_block);
ext4_journal_stop(handle);
if (err == -ENOSPC && ext4_should_retry_alloc(dir->i_sb, &retries))
goto retry;
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 09/19] ext4: Refactor __ext4_check_dir_entry to accepts start and size.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (6 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 08/19] ext4: Create a new function ext4_init_new_dir Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 10/19] ext4: Create __ext4_insert_dentry for dir entry insertion Tao Ma
` (9 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
__ext4_check_dir_entry is used to check whether the de is over
the block boundary. Now with inline data, it could be within the
block boundary while exceeds the inode size. So check this function
to check the overflow more precisely.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/dir.c | 16 ++++++++--------
fs/ext4/ext4.h | 7 ++++---
fs/ext4/namei.c | 13 +++++++++----
3 files changed, 21 insertions(+), 15 deletions(-)
diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index 164c560..6c2199e 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -68,7 +68,7 @@ static unsigned char get_dtype(struct super_block *sb, int filetype)
int __ext4_check_dir_entry(const char *function, unsigned int line,
struct inode *dir, struct file *filp,
struct ext4_dir_entry_2 *de,
- struct buffer_head *bh,
+ struct buffer_head *bh, char *buf, int size,
unsigned int offset)
{
const char *error_msg = NULL;
@@ -81,9 +81,8 @@ int __ext4_check_dir_entry(const char *function, unsigned int line,
error_msg = "rec_len % 4 != 0";
else if (unlikely(rlen < EXT4_DIR_REC_LEN(de->name_len)))
error_msg = "rec_len is too small for name_len";
- else if (unlikely(((char *) de - bh->b_data) + rlen >
- dir->i_sb->s_blocksize))
- error_msg = "directory entry across blocks";
+ else if (unlikely(((char *) de - buf) + rlen > size))
+ error_msg = "directory entry across range";
else if (unlikely(le32_to_cpu(de->inode) >
le32_to_cpu(EXT4_SB(dir->i_sb)->s_es->s_inodes_count)))
error_msg = "inode out of bounds";
@@ -94,14 +93,14 @@ int __ext4_check_dir_entry(const char *function, unsigned int line,
ext4_error_file(filp, function, line, bh ? bh->b_blocknr : 0,
"bad entry in directory: %s - offset=%u(%u), "
"inode=%u, rec_len=%d, name_len=%d",
- error_msg, (unsigned) (offset%bh->b_size),
+ error_msg, (unsigned) (offset%size),
offset, le32_to_cpu(de->inode),
rlen, de->name_len);
else
ext4_error_inode(dir, function, line, bh ? bh->b_blocknr : 0,
"bad entry in directory: %s - offset=%u(%u), "
"inode=%u, rec_len=%d, name_len=%d",
- error_msg, (unsigned) (offset%bh->b_size),
+ error_msg, (unsigned) (offset%size),
offset, le32_to_cpu(de->inode),
rlen, de->name_len);
@@ -210,8 +209,9 @@ revalidate:
while (!error && filp->f_pos < inode->i_size
&& offset < sb->s_blocksize) {
de = (struct ext4_dir_entry_2 *) (bh->b_data + offset);
- if (ext4_check_dir_entry(inode, filp, de,
- bh, offset)) {
+ if (ext4_check_dir_entry(inode, filp, de, bh,
+ bh->b_data, bh->b_size,
+ offset)) {
/*
* On error, skip the f_pos to the next block
*/
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 27c5637..bcc29b1 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1806,10 +1806,11 @@ ext4_fsblk_t ext4_inode_to_goal_block(struct inode *);
extern int __ext4_check_dir_entry(const char *, unsigned int, struct inode *,
struct file *,
struct ext4_dir_entry_2 *,
- struct buffer_head *, unsigned int);
-#define ext4_check_dir_entry(dir, filp, de, bh, offset) \
+ struct buffer_head *, char *, int,
+ unsigned int);
+#define ext4_check_dir_entry(dir, filp, de, bh, buf, size, offset) \
unlikely(__ext4_check_dir_entry(__func__, __LINE__, (dir), (filp), \
- (de), (bh), (offset)))
+ (de), (bh), (buf), (size), (offset)))
extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
__u32 minor_hash,
struct ext4_dir_entry_2 *dirent);
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 14ecbd2..aae0f5b 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -583,6 +583,7 @@ static int htree_dirblock_to_tree(struct file *dir_file,
EXT4_DIR_REC_LEN(0));
for (; de < top; de = ext4_next_entry(de, dir->i_sb->s_blocksize)) {
if (ext4_check_dir_entry(dir, NULL, de, bh,
+ bh->b_data, bh->b_size,
(block<<EXT4_BLOCK_SIZE_BITS(dir->i_sb))
+ ((char *)de - bh->b_data))) {
/* On error, skip the f_pos to the next block. */
@@ -821,7 +822,8 @@ static inline int search_dirblock(struct buffer_head *bh,
if ((char *) de + namelen <= dlimit &&
ext4_match (namelen, name, de)) {
/* found a match - just to be sure, do a full check */
- if (ext4_check_dir_entry(dir, NULL, de, bh, offset))
+ if (ext4_check_dir_entry(dir, NULL, de, bh, bh->b_data,
+ bh->b_size, offset))
return -1;
*res_dir = de;
return 1;
@@ -1267,7 +1269,8 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
de = (struct ext4_dir_entry_2 *)bh->b_data;
top = bh->b_data + blocksize - reclen;
while ((char *) de <= top) {
- if (ext4_check_dir_entry(dir, NULL, de, bh, offset))
+ if (ext4_check_dir_entry(dir, NULL, de, bh, bh->b_data,
+ bh->b_size, offset))
return -EIO;
if (ext4_match(namelen, name, de))
return -EEXIST;
@@ -1650,7 +1653,8 @@ static int ext4_delete_entry(handle_t *handle,
pde = NULL;
de = (struct ext4_dir_entry_2 *) bh->b_data;
while (i < bh->b_size) {
- if (ext4_check_dir_entry(dir, NULL, de, bh, i))
+ if (ext4_check_dir_entry(dir, NULL, de, bh,
+ bh->b_data, bh->b_size, i))
return -EIO;
if (de == de_del) {
BUFFER_TRACE(bh, "get_write_access");
@@ -1963,7 +1967,8 @@ static int empty_dir(struct inode *inode)
}
de = (struct ext4_dir_entry_2 *) bh->b_data;
}
- if (ext4_check_dir_entry(inode, NULL, de, bh, offset)) {
+ if (ext4_check_dir_entry(inode, NULL, de, bh,
+ bh->b_data, bh->b_size, offset)) {
de = (struct ext4_dir_entry_2 *)(bh->b_data +
sb->s_blocksize);
offset = (offset | (sb->s_blocksize - 1)) + 1;
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 10/19] ext4: Create __ext4_insert_dentry for dir entry insertion.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (7 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 09/19] ext4: Refactor __ext4_check_dir_entry to accepts start and size Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 11/19] ext4: let add_dir_entry handle inline data properly Tao Ma
` (8 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
The old add_dirent_to_buf handles all the work related to the
work of adding dir entry to a dir block. Now we have inline data,
so create 2 new function __ext4_find_dest_de and __ext4_insert_dentry
that do the real work and let add_dirent_to_buf call them.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 15 +++++++
fs/ext4/namei.c | 111 +++++++++++++++++++++++++++++++++---------------------
2 files changed, 83 insertions(+), 43 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index bcc29b1..620c168 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1815,6 +1815,21 @@ extern int ext4_htree_store_dirent(struct file *dir_file, __u32 hash,
__u32 minor_hash,
struct ext4_dir_entry_2 *dirent);
extern void ext4_htree_free_dir_info(struct dir_private_info *p);
+extern int ext4_find_dest_de(struct inode *dir, struct inode *inode,
+ struct buffer_head *bh,
+ void *buf, int buf_size,
+ const char *name, int namelen,
+ struct ext4_dir_entry_2 **dest_de);
+void ext4_insert_dentry(struct inode *inode,
+ struct ext4_dir_entry_2 *de,
+ int buf_size,
+ const char *name, int namelen);
+static inline void ext4_update_dx_flag(struct inode *inode)
+{
+ if (!EXT4_HAS_COMPAT_FEATURE(inode->i_sb,
+ EXT4_FEATURE_COMPAT_DIR_INDEX))
+ ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
+}
/* fsync.c */
extern int ext4_sync_file(struct file *, loff_t, loff_t, int);
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index aae0f5b..21acbc9 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -775,13 +775,6 @@ static void dx_insert_block(struct dx_frame *frame, u32 hash, ext4_lblk_t block)
dx_set_count(entries, count + 1);
}
-static void ext4_update_dx_flag(struct inode *inode)
-{
- if (!EXT4_HAS_COMPAT_FEATURE(inode->i_sb,
- EXT4_FEATURE_COMPAT_DIR_INDEX))
- ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
-}
-
/*
* NOTE! unlike strncmp, ext4_match returns 1 for success, 0 for failure.
*
@@ -1243,6 +1236,66 @@ errout:
return NULL;
}
+int ext4_find_dest_de(struct inode *dir, struct inode *inode,
+ struct buffer_head *bh,
+ void *buf, int buf_size,
+ const char *name, int namelen,
+ struct ext4_dir_entry_2 **dest_de)
+{
+ struct ext4_dir_entry_2 *de;
+ unsigned short reclen = EXT4_DIR_REC_LEN(namelen);
+ int nlen, rlen;
+ unsigned int offset = 0;
+ char *top;
+
+ de = (struct ext4_dir_entry_2 *)buf;
+ top = buf + buf_size - reclen;
+ while ((char *) de <= top) {
+ if (ext4_check_dir_entry(dir, NULL, de, bh,
+ buf, buf_size, offset))
+ return -EIO;
+ if (ext4_match(namelen, name, de))
+ return -EEXIST;
+ nlen = EXT4_DIR_REC_LEN(de->name_len);
+ rlen = ext4_rec_len_from_disk(de->rec_len, buf_size);
+ if ((de->inode ? rlen - nlen : rlen) >= reclen)
+ break;
+ de = (struct ext4_dir_entry_2 *)((char *)de + rlen);
+ offset += rlen;
+ }
+ if ((char *) de > top)
+ return -ENOSPC;
+
+ *dest_de = de;
+ return 0;
+}
+
+void ext4_insert_dentry(struct inode *inode,
+ struct ext4_dir_entry_2 *de,
+ int buf_size,
+ const char *name, int namelen)
+{
+
+ int nlen, rlen;
+
+ nlen = EXT4_DIR_REC_LEN(de->name_len);
+ rlen = ext4_rec_len_from_disk(de->rec_len, buf_size);
+ if (de->inode) {
+ struct ext4_dir_entry_2 *de1 =
+ (struct ext4_dir_entry_2 *)((char *)de + nlen);
+ de1->rec_len = ext4_rec_len_to_disk(rlen - nlen, buf_size);
+ de->rec_len = ext4_rec_len_to_disk(nlen, buf_size);
+ de = de1;
+ }
+ de->file_type = EXT4_FT_UNKNOWN;
+ if (inode) {
+ de->inode = cpu_to_le32(inode->i_ino);
+ ext4_set_de_type(inode->i_sb, de, inode->i_mode);
+ } else
+ de->inode = 0;
+ de->name_len = namelen;
+ memcpy(de->name, name, namelen);
+}
/*
* Add a new entry into a directory (leaf) block. If de is non-NULL,
* it points to a directory entry which is guaranteed to be large
@@ -1258,31 +1311,17 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
struct inode *dir = dentry->d_parent->d_inode;
const char *name = dentry->d_name.name;
int namelen = dentry->d_name.len;
- unsigned int offset = 0;
unsigned int blocksize = dir->i_sb->s_blocksize;
unsigned short reclen;
- int nlen, rlen, err;
- char *top;
+ int err;
reclen = EXT4_DIR_REC_LEN(namelen);
if (!de) {
- de = (struct ext4_dir_entry_2 *)bh->b_data;
- top = bh->b_data + blocksize - reclen;
- while ((char *) de <= top) {
- if (ext4_check_dir_entry(dir, NULL, de, bh, bh->b_data,
- bh->b_size, offset))
- return -EIO;
- if (ext4_match(namelen, name, de))
- return -EEXIST;
- nlen = EXT4_DIR_REC_LEN(de->name_len);
- rlen = ext4_rec_len_from_disk(de->rec_len, blocksize);
- if ((de->inode? rlen - nlen: rlen) >= reclen)
- break;
- de = (struct ext4_dir_entry_2 *)((char *)de + rlen);
- offset += rlen;
- }
- if ((char *) de > top)
- return -ENOSPC;
+ err = ext4_find_dest_de(dir, inode,
+ bh, bh->b_data, blocksize,
+ name, namelen, &de);
+ if (err)
+ return err;
}
BUFFER_TRACE(bh, "get_write_access");
err = ext4_journal_get_write_access(handle, bh);
@@ -1292,22 +1331,8 @@ static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry,
}
/* By now the buffer is marked for journaling */
- nlen = EXT4_DIR_REC_LEN(de->name_len);
- rlen = ext4_rec_len_from_disk(de->rec_len, blocksize);
- if (de->inode) {
- struct ext4_dir_entry_2 *de1 = (struct ext4_dir_entry_2 *)((char *)de + nlen);
- de1->rec_len = ext4_rec_len_to_disk(rlen - nlen, blocksize);
- de->rec_len = ext4_rec_len_to_disk(nlen, blocksize);
- de = de1;
- }
- de->file_type = EXT4_FT_UNKNOWN;
- if (inode) {
- de->inode = cpu_to_le32(inode->i_ino);
- ext4_set_de_type(dir->i_sb, de, inode->i_mode);
- } else
- de->inode = 0;
- de->name_len = namelen;
- memcpy(de->name, name, namelen);
+ ext4_insert_dentry(inode, de, blocksize, name, namelen);
+
/*
* XXX shouldn't update any times until successful
* completion of syscall, but too many callers depend
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 11/19] ext4: let add_dir_entry handle inline data properly.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (8 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 10/19] ext4: Create __ext4_insert_dentry for dir entry insertion Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 12/19] ext4: Let ext4_readdir handle inline data Tao Ma
` (7 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
This patch let add_dir_entry handle the inline data case. So the
dir is initialized as inline dir first and then we can try to add
some files to it, when the inline space can't hold all the entries,
a dir block will be created and the dir entry will be moved to it.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 2 +
fs/ext4/inline.c | 276 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/namei.c | 19 ++++
fs/ext4/xattr.h | 19 ++++
4 files changed, 316 insertions(+), 0 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 620c168..7edfbd4 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -2260,6 +2260,8 @@ static inline void ext4_mark_super_dirty(struct super_block *sb)
/* dir.c */
extern const struct file_operations ext4_dir_operations;
+extern void ext4_show_inline_dir(struct inode *dir, struct buffer_head *bh,
+ void *inline_start, int inline_size);
/* file.c */
extern const struct inode_operations ext4_file_inode_operations;
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index f2a212d..848bcd3 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -908,6 +908,282 @@ int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
return copied;
}
+#ifdef INLINE_DIR_DEBUG
+void ext4_show_inline_dir(struct inode *dir, struct buffer_head *bh,
+ void *inline_start, int inline_size)
+{
+ int offset;
+ unsigned short de_len;
+ struct ext4_dir_entry_2 *de = inline_start;
+ void *dlimit = inline_start + inline_size;
+
+ trace_printk("inode %lu\n", dir->i_ino);
+ offset = 0;
+ while ((void *)de < dlimit) {
+ de_len = ext4_rec_len_from_disk(de->rec_len, inline_size);
+ trace_printk("de: offset %u reclen %u name %*.s "
+ "namelen %u ino %u\n",
+ offset, de_len, de->name_len, de->name,
+ de->name_len, le32_to_cpu(de->inode));
+ if (ext4_check_dir_entry(dir, NULL, de, bh,
+ inline_start, inline_size, offset))
+ BUG();
+
+ offset += de_len;
+ de = (struct ext4_dir_entry_2 *) ((char *) de + de_len);
+ }
+}
+#else
+#define ext4_show_inline_dir(dir, bh, inline_start, inline_size)
+#endif
+
+/*
+ * Add a new entry into a inline dir.
+ * It will return -ENOSPC if no space is available, and -EIO
+ * and -EEXIST if directory entry already exists.
+ */
+static int ext4_add_dirent_to_inline(handle_t *handle,
+ struct dentry *dentry,
+ struct inode *inode,
+ struct ext4_iloc *iloc,
+ void *inline_start, int inline_size)
+{
+ struct inode *dir = dentry->d_parent->d_inode;
+ const char *name = dentry->d_name.name;
+ int namelen = dentry->d_name.len;
+ unsigned short reclen;
+ int err;
+ struct ext4_dir_entry_2 *de;
+
+ reclen = EXT4_DIR_REC_LEN(namelen);
+ err = ext4_find_dest_de(dir, inode, iloc->bh,
+ inline_start, inline_size,
+ name, namelen, &de);
+ if (err)
+ return err;
+
+ err = ext4_journal_get_write_access(handle, iloc->bh);
+ if (err)
+ return err;
+ ext4_insert_dentry(inode, de, inline_size, name, namelen);
+
+ ext4_show_inline_dir(dir, iloc->bh, inline_start, inline_size);
+
+ /*
+ * XXX shouldn't update any times until successful
+ * completion of syscall, but too many callers depend
+ * on this.
+ *
+ * XXX similarly, too many callers depend on
+ * ext4_new_inode() setting the times, but error
+ * recovery deletes the inode, so the worst that can
+ * happen is that the times are slightly out of date
+ * and/or different from the directory change time.
+ */
+ dir->i_mtime = dir->i_ctime = ext4_current_time(dir);
+ ext4_update_dx_flag(dir);
+ dir->i_version++;
+ ext4_mark_inode_dirty(handle, dir);
+ return 1;
+}
+
+void* ext4_get_inline_xattr_pos(struct inode *inode, struct ext4_iloc *iloc)
+{
+ struct ext4_xattr_entry *entry;
+ struct ext4_xattr_ibody_header *header;
+
+ BUG_ON(!EXT4_I(inode)->i_inline_off);
+
+ header = IHDR(inode, ext4_raw_inode(iloc));
+ entry = (struct ext4_xattr_entry *)((void *)ext4_raw_inode(iloc) +
+ EXT4_I(inode)->i_inline_off);
+
+ return (void *)IFIRST(header) + le16_to_cpu(entry->e_value_offs);
+}
+
+/* Set the final de to cover the whole block. */
+static void ext4_update_final_de(void *de_buf, int old_size, int new_size)
+{
+ struct ext4_dir_entry_2 *de, *prev_de;
+ void *limit;
+ int de_len;
+
+ de = (struct ext4_dir_entry_2 *)de_buf;
+ if (old_size) {
+ limit = de_buf + old_size;
+ do {
+ prev_de = de;
+ de_len = ext4_rec_len_from_disk(de->rec_len, old_size);
+ de_buf += de_len;
+ de = (struct ext4_dir_entry_2 *)de_buf;
+ } while (de_buf < limit);
+
+ prev_de->rec_len = ext4_rec_len_to_disk(de_len + new_size -
+ old_size, new_size);
+ } else {
+ /* this is just created, so create an empty entry. */
+ de->inode = 0;
+ de->rec_len = ext4_rec_len_to_disk(new_size, new_size);
+ }
+}
+
+static int ext4_update_inline_dir(handle_t *handle, struct dentry *dentry,
+ struct inode *dir,
+ struct ext4_iloc *iloc)
+{
+ int ret, reclen = EXT4_DIR_REC_LEN(dentry->d_name.len);
+ int old_size = EXT4_I(dir)->i_inline_size - EXT4_MIN_INLINE_DATA_SIZE;
+
+ ret = ext4_update_inline_data(handle, dir,
+ EXT4_I(dir)->i_inline_size + reclen);
+ if (ret)
+ return ret;
+
+ ext4_update_final_de(ext4_get_inline_xattr_pos(dir, iloc), old_size,
+ EXT4_I(dir)->i_inline_size -
+ EXT4_MIN_INLINE_DATA_SIZE);
+ dir->i_size = EXT4_I(dir)->i_disksize = EXT4_I(dir)->i_inline_size;
+ return 0;
+}
+
+/*
+ * Try to add the new entry to the inline data.
+ * If succeeds, return 0. If not, extended the inline dir and copied data to
+ * the new created block.
+ */
+int ext4_try_add_inline_entry(handle_t *handle, struct dentry *dentry,
+ struct inode *inode)
+{
+ int ret, inline_size;
+ void *inline_start, *backup_buf = NULL;
+ struct buffer_head *dir_block = NULL;
+ struct ext4_iloc iloc;
+ int blocksize = inode->i_sb->s_blocksize;
+ struct inode *dir = dentry->d_parent->d_inode;
+
+ ret = ext4_get_inode_loc(dir, &iloc);
+ if (ret)
+ return ret;
+
+ down_write(&EXT4_I(inode)->xattr_sem);
+ if (!ext4_has_inline_data(inode))
+ goto out;
+
+ inline_start = ext4_raw_inode(&iloc)->i_block;
+ inline_size = EXT4_MIN_INLINE_DATA_SIZE;
+
+ ret = ext4_add_dirent_to_inline(handle, dentry, inode, &iloc,
+ inline_start, inline_size);
+ if (ret != -ENOSPC)
+ goto out;
+
+ /* check whether it can be inserted to inline xattr space. */
+ inline_size = EXT4_I(dir)->i_inline_size -
+ EXT4_MIN_INLINE_DATA_SIZE;
+ if (inline_size > 0) {
+ inline_start = ext4_get_inline_xattr_pos(dir, &iloc);
+
+ ret = ext4_add_dirent_to_inline(handle, dentry, inode, &iloc,
+ inline_start, inline_size);
+ if (ret != -ENOSPC)
+ goto out;
+ }
+
+ /* Try to add more xattr space.*/
+ ret = ext4_update_inline_dir(handle, dentry, dir, &iloc);
+ if (ret && ret != -ENOSPC)
+ goto out;
+ else if (!ret) {
+ inline_size = EXT4_I(dir)->i_inline_size -
+ EXT4_MIN_INLINE_DATA_SIZE;
+ inline_start = ext4_get_inline_xattr_pos(dir, &iloc);
+
+ ret = ext4_add_dirent_to_inline(handle, dentry, inode, &iloc,
+ inline_start, inline_size);
+ if (ret != -ENOSPC)
+ goto out;
+ }
+
+ /*
+ * The inline space is filled up, so create a new block for it.
+ * As the extent tree will be created, we have to save the inline
+ * dir first.
+ */
+ inline_size = EXT4_I(dir)->i_inline_size;
+ backup_buf = kmalloc(inline_size, GFP_NOFS);
+ if (!backup_buf) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ memcpy(backup_buf, (void *)ext4_raw_inode(&iloc)->i_block,
+ EXT4_MIN_INLINE_DATA_SIZE);
+ if (inline_size > EXT4_MIN_INLINE_DATA_SIZE)
+ memcpy(backup_buf + EXT4_MIN_INLINE_DATA_SIZE,
+ ext4_get_inline_xattr_pos(dir, &iloc),
+ inline_size - EXT4_MIN_INLINE_DATA_SIZE);
+
+ /* clear the entry and the flag in dir now. */
+ ret = ext4_destroy_inline_data_nolock(handle, dir);
+ if (ret)
+ goto out;
+
+ dir->i_size = EXT4_I(dir)->i_disksize = blocksize;
+ dir_block = ext4_bread(handle, dir, 0, 1, &ret);
+ if (!dir_block)
+ goto out;
+
+ BUFFER_TRACE(dir_block, "get_write_access");
+ ret = ext4_journal_get_write_access(handle, dir_block);
+ if (ret)
+ goto out;
+ memcpy(dir_block->b_data, backup_buf, inline_size);
+
+ /* Set the final de to cover the whole block. */
+ ext4_update_final_de(dir_block->b_data, inline_size,
+ blocksize);
+
+ BUFFER_TRACE(dir_block, "call ext4_handle_dirty_metadata");
+ ret = ext4_handle_dirty_metadata(handle, dir, dir_block);
+
+out:
+ kfree(backup_buf);
+ brelse(dir_block);
+ if (!ret || ret == 1)
+ ext4_mark_inode_dirty(handle, dir);
+ up_write(&EXT4_I(inode)->xattr_sem);
+ brelse(iloc.bh);
+ return ret;
+}
+
+/*
+ * Try to create the inline data for the new dir.
+ * If it succeeds, return 0, otherwise return the error.
+ * In case of ENOSPC, the caller should create the normal disk layout dir.
+ */
+int ext4_try_create_inline_dir(handle_t *handle, struct inode *parent,
+ struct inode *inode)
+{
+ int ret, inline_size = EXT4_MIN_INLINE_DATA_SIZE;
+ struct ext4_iloc iloc;
+ struct ext4_dir_entry_2 *de;
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret)
+ return ret;
+
+ ret = ext4_prepare_inline_data(handle, inode, inline_size);
+ if (ret)
+ goto out;
+
+ de = (struct ext4_dir_entry_2 *)ext4_raw_inode(&iloc)->i_block;
+ ext4_init_dot_dotdot(parent, inode, de, inline_size);
+ inode->i_size = EXT4_I(inode)->i_disksize = inline_size;
+out:
+ brelse(iloc.bh);
+ return ret;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 21acbc9..77312dc 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1486,6 +1486,17 @@ static int ext4_add_entry(handle_t *handle, struct dentry *dentry,
blocksize = sb->s_blocksize;
if (!dentry->d_name.len)
return -EINVAL;
+
+ if (ext4_has_inline_data(dir)) {
+ retval = ext4_try_add_inline_entry(handle, dentry, inode);
+ if (retval < 0)
+ return retval;
+ if (retval == 1) {
+ retval = 0;
+ return retval;
+ }
+ }
+
if (is_dx(dir)) {
retval = ext4_dx_add_entry(handle, dentry, inode);
if (!retval || (retval != ERR_BAD_DX_DIR))
@@ -1862,6 +1873,14 @@ static int ext4_init_new_dir(handle_t *handle, struct inode *parent,
int err;
int blocksize = inode->i_sb->s_blocksize;
+ if (ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) {
+ err = ext4_try_create_inline_dir(handle, parent, inode);
+ if (err < 0 && err != ENOSPC)
+ goto out;
+ if (!err)
+ goto out;
+ }
+
inode->i_size = EXT4_I(inode)->i_disksize = blocksize;
dir_block = ext4_bread(handle, inode, 0, 1, &err);
if (!dir_block)
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index c390402..5343aa1 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -163,6 +163,11 @@ extern int ext4_da_write_inline_data_begin(struct address_space *mapping,
extern int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
unsigned len, unsigned copied,
struct page *page);
+extern int ext4_try_add_inline_entry(handle_t *handle, struct dentry *dentry,
+ struct inode *inode);
+extern int ext4_try_create_inline_dir(handle_t *handle,
+ struct inode *parent,
+ struct inode *inode);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -332,6 +337,20 @@ static inline int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos,
{
return 0;
}
+
+static inline int ext4_try_add_inline_entry(handle_t *handle,
+ struct dentry *dentry,
+ struct inode *inode)
+{
+ return 0;
+}
+
+static inline int ext4_try_create_inline_dir(handle_t *handle,
+ struct inode *parent,
+ struct inode *inode)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 12/19] ext4: Let ext4_readdir handle inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (9 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 11/19] ext4: let add_dir_entry handle inline data properly Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 13/19] ext4: Create a new function search_dir Tao Ma
` (6 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/dir.c | 23 ++++------
fs/ext4/ext4.h | 12 +++++
fs/ext4/inline.c | 124 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/xattr.h | 7 +++
4 files changed, 153 insertions(+), 13 deletions(-)
diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c
index 6c2199e..691548c 100644
--- a/fs/ext4/dir.c
+++ b/fs/ext4/dir.c
@@ -27,10 +27,7 @@
#include <linux/slab.h>
#include <linux/rbtree.h>
#include "ext4.h"
-
-static unsigned char ext4_filetype_table[] = {
- DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
-};
+#include "xattr.h"
static int ext4_readdir(struct file *, void *, filldir_t);
static int ext4_dx_readdir(struct file *filp,
@@ -51,19 +48,13 @@ const struct file_operations ext4_dir_operations = {
};
-static unsigned char get_dtype(struct super_block *sb, int filetype)
-{
- if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FILETYPE) ||
- (filetype >= EXT4_FT_MAX))
- return DT_UNKNOWN;
-
- return (ext4_filetype_table[filetype]);
-}
-
/*
* Return 0 if the directory entry is OK, and 1 if there is a problem
*
* Note: this is the opposite of what ext2 and ext3 historically returned...
+ *
+ * bh passed here can be an inode block or a dir data block, depending
+ * on the inode inline data flag.
*/
int __ext4_check_dir_entry(const char *function, unsigned int line,
struct inode *dir, struct file *filp,
@@ -122,6 +113,12 @@ static int ext4_readdir(struct file *filp,
sb = inode->i_sb;
+ if (ext4_has_inline_data(inode)) {
+ ret = ext4_read_inline_dir(filp, dirent, filldir);
+ if (ret > 0 || ret < 0)
+ return ret;
+ }
+
if (EXT4_HAS_COMPAT_FEATURE(inode->i_sb,
EXT4_FEATURE_COMPAT_DIR_INDEX) &&
((ext4_test_inode_flag(inode, EXT4_INODE_INDEX)) ||
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 7edfbd4..cb4dfbd 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1830,6 +1830,18 @@ static inline void ext4_update_dx_flag(struct inode *inode)
EXT4_FEATURE_COMPAT_DIR_INDEX))
ext4_clear_inode_flag(inode, EXT4_INODE_INDEX);
}
+static unsigned char ext4_filetype_table[] = {
+ DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK
+};
+
+static inline unsigned char get_dtype(struct super_block *sb, int filetype)
+{
+ if (!EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_FILETYPE) ||
+ (filetype >= EXT4_FT_MAX))
+ return DT_UNKNOWN;
+
+ return ext4_filetype_table[filetype];
+}
/* fsync.c */
extern int ext4_sync_file(struct file *, loff_t, loff_t, int);
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 848bcd3..a6b6593 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1184,6 +1184,130 @@ out:
return ret;
}
+/*
+ * Get the inline dentry at offset.
+ */
+static inline struct ext4_dir_entry_2 *
+ext4_get_inline_entry(struct inode *inode,
+ struct ext4_iloc *iloc,
+ unsigned int offset,
+ void **inline_start,
+ int *inline_size)
+{
+ void *inline_pos;
+
+ BUG_ON(offset > ext4_get_inline_size(inode));
+
+ if (offset < EXT4_MIN_INLINE_DATA_SIZE) {
+ inline_pos = ext4_raw_inode(iloc)->i_block;
+ *inline_size = EXT4_MIN_INLINE_DATA_SIZE;
+ } else {
+ inline_pos = ext4_get_inline_xattr_pos(inode, iloc);
+ offset -= EXT4_MIN_INLINE_DATA_SIZE;
+ *inline_size = ext4_get_inline_size(inode) -
+ EXT4_MIN_INLINE_DATA_SIZE;
+ }
+
+ if (inline_start)
+ *inline_start = inline_pos;
+ return (struct ext4_dir_entry_2 *)(inline_pos + offset);
+}
+
+int ext4_read_inline_dir(struct file *filp,
+ void *dirent, filldir_t filldir)
+{
+ int error = 0;
+ unsigned int offset;
+ int i, stored;
+ void *inline_pos;
+ struct ext4_dir_entry_2 *de;
+ struct super_block *sb;
+ struct inode *inode = filp->f_path.dentry->d_inode;
+ int ret, inline_size = 0;
+ struct ext4_iloc iloc;
+
+ ret = ext4_get_inode_loc(inode, &iloc);
+ if (ret)
+ return ret;
+
+ down_read(&EXT4_I(inode)->xattr_sem);
+ if (!ext4_has_inline_data(inode))
+ goto out;
+
+ sb = inode->i_sb;
+ stored = 0;
+ offset = filp->f_pos & (sb->s_blocksize - 1);
+
+ while (!error && !stored && filp->f_pos < inode->i_size) {
+revalidate:
+ /* If the version has changed since the last call to
+ * readdir(2), then we might be pointing to an invalid
+ * dirent right now. Scan from the start of the block
+ * to make sure. */
+ if (filp->f_version != inode->i_version) {
+ for (i = 0; i < inode->i_size && i < offset; ) {
+ de = ext4_get_inline_entry(inode, &iloc, i,
+ NULL, &inline_size);
+ /* It's too expensive to do a full
+ * dirent test each time round this
+ * loop, but we do have to test at
+ * least that it is non-zero. A
+ * failure will be detected in the
+ * dirent test below. */
+ if (ext4_rec_len_from_disk(de->rec_len,
+ inline_size) < EXT4_DIR_REC_LEN(1))
+ break;
+ i += ext4_rec_len_from_disk(de->rec_len,
+ inline_size);
+ }
+ offset = i;
+ filp->f_pos = offset;
+ filp->f_version = inode->i_version;
+ }
+
+ while (!error && filp->f_pos < inode->i_size) {
+ de = ext4_get_inline_entry(inode, &iloc, offset,
+ &inline_pos, &inline_size);
+ if (ext4_check_dir_entry(inode, filp, de,
+ iloc.bh, inline_pos,
+ inline_size, offset)) {
+ ret = stored;
+ goto out;
+ }
+ offset += ext4_rec_len_from_disk(de->rec_len,
+ inline_size);
+ if (le32_to_cpu(de->inode)) {
+ /* We might block in the next section
+ * if the data destination is
+ * currently swapped out. So, use a
+ * version stamp to detect whether or
+ * not the directory has been modified
+ * during the copy operation.
+ */
+ u64 version = filp->f_version;
+
+ error = filldir(dirent, de->name,
+ de->name_len,
+ filp->f_pos,
+ le32_to_cpu(de->inode),
+ get_dtype(sb, de->file_type));
+ if (error)
+ break;
+ if (version != filp->f_version)
+ goto revalidate;
+ stored++;
+ }
+ filp->f_pos += ext4_rec_len_from_disk(de->rec_len,
+ inline_size);
+ }
+ offset = 0;
+ }
+out:
+ up_read(&EXT4_I(inode)->xattr_sem);
+ brelse(iloc.bh);
+ return ret;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 5343aa1..1e042b4 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -168,6 +168,8 @@ extern int ext4_try_add_inline_entry(handle_t *handle, struct dentry *dentry,
extern int ext4_try_create_inline_dir(handle_t *handle,
struct inode *parent,
struct inode *inode);
+extern int ext4_read_inline_dir(struct file *filp,
+ void *dirent, filldir_t filldir);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -351,6 +353,11 @@ static inline int ext4_try_create_inline_dir(handle_t *handle,
{
return 0;
}
+static inline int ext4_read_inline_dir(struct file *filp,
+ void *dirent, filldir_t filldir)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 13/19] ext4: Create a new function search_dir.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (10 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 12/19] ext4: Let ext4_readdir handle inline data Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 14/19] ext4: let ext4_find_entry handle inline data Tao Ma
` (5 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
search_dirblock is used to search a dir block, but the code is
almost the same for searching an inline dir.
So create a new fuction search_dir and let search_dirblock
call it.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 7 +++++++
fs/ext4/namei.c | 26 +++++++++++++++++++-------
2 files changed, 26 insertions(+), 7 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index cb4dfbd..9776c06 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1965,6 +1965,13 @@ extern int ext4_orphan_add(handle_t *, struct inode *);
extern int ext4_orphan_del(handle_t *, struct inode *);
extern int ext4_htree_fill_tree(struct file *dir_file, __u32 start_hash,
__u32 start_minor_hash, __u32 *next_hash);
+extern int search_dir(struct buffer_head *bh,
+ char *search_buf,
+ int buf_size,
+ struct inode *dir,
+ const struct qstr *d_name,
+ unsigned int offset,
+ struct ext4_dir_entry_2 **res_dir);
/* resize.c */
extern int ext4_group_add(struct super_block *sb,
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 77312dc..6e97dd2 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -794,11 +794,13 @@ static inline int ext4_match (int len, const char * const name,
/*
* Returns 0 if not found, -1 on failure, and 1 on success
*/
-static inline int search_dirblock(struct buffer_head *bh,
- struct inode *dir,
- const struct qstr *d_name,
- unsigned int offset,
- struct ext4_dir_entry_2 ** res_dir)
+int search_dir(struct buffer_head *bh,
+ char *search_buf,
+ int buf_size,
+ struct inode *dir,
+ const struct qstr *d_name,
+ unsigned int offset,
+ struct ext4_dir_entry_2 **res_dir)
{
struct ext4_dir_entry_2 * de;
char * dlimit;
@@ -806,8 +808,8 @@ static inline int search_dirblock(struct buffer_head *bh,
const char *name = d_name->name;
int namelen = d_name->len;
- de = (struct ext4_dir_entry_2 *) bh->b_data;
- dlimit = bh->b_data + dir->i_sb->s_blocksize;
+ de = (struct ext4_dir_entry_2 *)search_buf;
+ dlimit = search_buf + buf_size;
while ((char *) de < dlimit) {
/* this code is executed quadratically often */
/* do minimal checking `by hand' */
@@ -832,6 +834,16 @@ static inline int search_dirblock(struct buffer_head *bh,
return 0;
}
+static inline int search_dirblock(struct buffer_head *bh,
+ struct inode *dir,
+ const struct qstr *d_name,
+ unsigned int offset,
+ struct ext4_dir_entry_2 **res_dir)
+{
+ return search_dir(bh, bh->b_data, dir->i_sb->s_blocksize, dir,
+ d_name, offset, res_dir);
+}
+
/*
* ext4_find_entry()
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 14/19] ext4: let ext4_find_entry handle inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (11 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 13/19] ext4: Create a new function search_dir Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 15/19] ext4: make ext4_delete_entry generic Tao Ma
` (4 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Create a new ext4_find_inline_entry to handle the case
of inline data.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/inline.c | 47 +++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/namei.c | 10 +++++++++-
fs/ext4/xattr.h | 12 ++++++++++++
3 files changed, 68 insertions(+), 1 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index a6b6593..c76ae3b 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1308,6 +1308,53 @@ out:
return ret;
}
+struct buffer_head *ext4_find_inline_entry(struct inode *dir,
+ const struct qstr *d_name,
+ struct ext4_dir_entry_2 **res_dir,
+ int *has_inline_data)
+{
+ int ret;
+ struct ext4_iloc iloc;
+ void *inline_start;
+ int inline_size;
+
+ if (ext4_get_inode_loc(dir, &iloc))
+ return NULL;
+
+ down_read(&EXT4_I(dir)->xattr_sem);
+ if (!ext4_has_inline_data(dir)) {
+ *has_inline_data = 0;
+ goto out;
+ }
+
+ inline_start = ext4_raw_inode(&iloc)->i_block;
+ inline_size = EXT4_MIN_INLINE_DATA_SIZE;
+ ret = search_dir(iloc.bh, inline_start, inline_size,
+ dir, d_name, 0, res_dir);
+ if (ret == 1)
+ goto out_find;
+ if (ret < 0)
+ goto out;
+
+ if (ext4_get_inline_size(dir) == EXT4_MIN_INLINE_DATA_SIZE)
+ goto out;
+
+ inline_start = ext4_get_inline_xattr_pos(dir, &iloc);
+ inline_size = ext4_get_max_inline_size(dir);
+
+ ret = search_dir(iloc.bh, inline_start, inline_size,
+ dir, d_name, 0, res_dir);
+ if (ret == 1)
+ goto out_find;
+
+out:
+ brelse(iloc.bh);
+ iloc.bh = NULL;
+out_find:
+ up_read(&EXT4_I(dir)->xattr_sem);
+ return iloc.bh;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 6e97dd2..92f7b67 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -844,7 +844,6 @@ static inline int search_dirblock(struct buffer_head *bh,
d_name, offset, res_dir);
}
-
/*
* ext4_find_entry()
*
@@ -879,6 +878,15 @@ static struct buffer_head * ext4_find_entry (struct inode *dir,
namelen = d_name->len;
if (namelen > EXT4_NAME_LEN)
return NULL;
+
+ if (ext4_has_inline_data(dir)) {
+ int has_inline_data = 1;
+ ret = ext4_find_inline_entry(dir, d_name, res_dir,
+ &has_inline_data);
+ if (has_inline_data)
+ return ret;
+ }
+
if ((namelen <= 2) && (name[0] == '.') &&
(name[1] == '.' || name[1] == '\0')) {
/*
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index 1e042b4..f4d4cb1 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -170,6 +170,10 @@ extern int ext4_try_create_inline_dir(handle_t *handle,
struct inode *inode);
extern int ext4_read_inline_dir(struct file *filp,
void *dirent, filldir_t filldir);
+extern struct buffer_head *ext4_find_inline_entry(struct inode *dir,
+ const struct qstr *d_name,
+ struct ext4_dir_entry_2 **res_dir,
+ int *has_inline_data);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -358,6 +362,14 @@ static inline int ext4_read_inline_dir(struct file *filp,
{
return 0;
}
+
+struct buffer_head *ext4_find_inline_entry(struct inode *dir,
+ const struct qstr *d_name,
+ struct ext4_dir_entry_2 **res_dir,
+ int *has_inline_data)
+{
+ return NULL;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 15/19] ext4: make ext4_delete_entry generic.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (12 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 14/19] ext4: let ext4_find_entry handle inline data Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 16/19] ext4: let ext4_delete_entry handle inline data Tao Ma
` (3 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Currently ext4_delete_entry is used only for dir entry
removing from a dir block. So let us create a new function
ext4_generic_delete_entry and this function takes a entry_buf
and a buf_size so that it can be used for inline data.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 6 +++++
fs/ext4/namei.c | 62 ++++++++++++++++++++++++++++++++++++------------------
2 files changed, 47 insertions(+), 21 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 9776c06..a88bc00 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1972,6 +1972,12 @@ extern int search_dir(struct buffer_head *bh,
const struct qstr *d_name,
unsigned int offset,
struct ext4_dir_entry_2 **res_dir);
+extern int ext4_generic_delete_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh,
+ void *entry_buf,
+ int buf_size);
/* resize.c */
extern int ext4_group_add(struct super_block *sb,
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 92f7b67..f706b0f 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1693,32 +1693,28 @@ cleanup:
}
/*
- * ext4_delete_entry deletes a directory entry by merging it with the
- * previous entry
+ * ext4_generic_delete_entry deletes a directory entry by merging it
+ * with the previous entry
*/
-static int ext4_delete_entry(handle_t *handle,
- struct inode *dir,
- struct ext4_dir_entry_2 *de_del,
- struct buffer_head *bh)
+int ext4_generic_delete_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh,
+ void *entry_buf,
+ int buf_size)
{
struct ext4_dir_entry_2 *de, *pde;
unsigned int blocksize = dir->i_sb->s_blocksize;
- int i, err;
+ int i;
i = 0;
pde = NULL;
- de = (struct ext4_dir_entry_2 *) bh->b_data;
- while (i < bh->b_size) {
+ de = (struct ext4_dir_entry_2 *)entry_buf;
+ while (i < buf_size) {
if (ext4_check_dir_entry(dir, NULL, de, bh,
bh->b_data, bh->b_size, i))
return -EIO;
if (de == de_del) {
- BUFFER_TRACE(bh, "get_write_access");
- err = ext4_journal_get_write_access(handle, bh);
- if (unlikely(err)) {
- ext4_std_error(dir->i_sb, err);
- return err;
- }
if (pde)
pde->rec_len = ext4_rec_len_to_disk(
ext4_rec_len_from_disk(pde->rec_len,
@@ -1729,12 +1725,6 @@ static int ext4_delete_entry(handle_t *handle,
else
de->inode = 0;
dir->i_version++;
- BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
- err = ext4_handle_dirty_metadata(handle, dir, bh);
- if (unlikely(err)) {
- ext4_std_error(dir->i_sb, err);
- return err;
- }
return 0;
}
i += ext4_rec_len_from_disk(de->rec_len, blocksize);
@@ -1744,6 +1734,36 @@ static int ext4_delete_entry(handle_t *handle,
return -ENOENT;
}
+static int ext4_delete_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh)
+{
+ int err;
+
+ BUFFER_TRACE(bh, "get_write_access");
+ err = ext4_journal_get_write_access(handle, bh);
+ if (unlikely(err))
+ goto out;
+
+ err = ext4_generic_delete_entry(handle, dir, de_del,
+ bh, bh->b_data,
+ dir->i_sb->s_blocksize);
+ if (err)
+ goto out;
+
+ BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+ err = ext4_handle_dirty_metadata(handle, dir, bh);
+ if (unlikely(err))
+ goto out;
+
+ return 0;
+out:
+ if (err != -ENOENT)
+ ext4_std_error(dir->i_sb, err);
+ return err;
+}
+
/*
* DIR_NLINK feature is set if 1) nlinks > EXT4_LINK_MAX or 2) nlinks == 2,
* since this indicates that nlinks count was previously 1.
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 16/19] ext4: let ext4_delete_entry handle inline data.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (13 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 15/19] ext4: make ext4_delete_entry generic Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 17/19] ext4: let empty_dir handle inline dir Tao Ma
` (2 subsequent siblings)
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/inline.c | 52 ++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/namei.c | 9 +++++++++
fs/ext4/xattr.h | 11 +++++++++++
3 files changed, 72 insertions(+), 0 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index c76ae3b..302925e 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1355,6 +1355,58 @@ out_find:
return iloc.bh;
}
+int ext4_delete_inline_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh)
+{
+ int err, inline_size;
+ struct ext4_iloc iloc;
+ void *inline_start;
+
+ err = ext4_get_inode_loc(dir, &iloc);
+ if (err)
+ return err;
+
+ down_write(&EXT4_I(dir)->xattr_sem);
+ if (!ext4_has_inline_data(dir))
+ goto out;
+
+ if ((void *)de_del - ((void *)ext4_raw_inode(&iloc)->i_block) <
+ EXT4_MIN_INLINE_DATA_SIZE) {
+ inline_start = ext4_raw_inode(&iloc)->i_block;
+ inline_size = EXT4_MIN_INLINE_DATA_SIZE;
+ } else {
+ inline_start = ext4_get_inline_xattr_pos(dir, &iloc);
+ inline_size = ext4_get_inline_size(dir) -
+ EXT4_MIN_INLINE_DATA_SIZE;
+ }
+
+ err = ext4_journal_get_write_access(handle, bh);
+ if (err)
+ goto out;
+
+ err = ext4_generic_delete_entry(handle, dir, de_del, bh,
+ inline_start, inline_size);
+ if (err)
+ goto out;
+
+ BUFFER_TRACE(bh, "call ext4_handle_dirty_metadata");
+ get_bh(iloc.bh);
+ err = ext4_mark_inode_dirty(handle, dir);
+ if (unlikely(err))
+ goto out;
+
+ ext4_show_inline_dir(dir, iloc.bh, inline_start, inline_size);
+ err = 1;
+out:
+ up_write(&EXT4_I(dir)->xattr_sem);
+ brelse(iloc.bh);
+ if (err != 1 && err != -ENOENT)
+ ext4_std_error(dir->i_sb, err);
+ return err;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index f706b0f..cefcec6 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -1741,6 +1741,15 @@ static int ext4_delete_entry(handle_t *handle,
{
int err;
+ if (ext4_has_inline_data(dir)) {
+ err = ext4_delete_inline_entry(handle, dir, de_del, bh);
+ if (err) {
+ if (err == 1)
+ err = 0;
+ return err;
+ }
+ }
+
BUFFER_TRACE(bh, "get_write_access");
err = ext4_journal_get_write_access(handle, bh);
if (unlikely(err))
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index f4d4cb1..ccd994e 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -174,6 +174,10 @@ extern struct buffer_head *ext4_find_inline_entry(struct inode *dir,
const struct qstr *d_name,
struct ext4_dir_entry_2 **res_dir,
int *has_inline_data);
+extern int ext4_delete_inline_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -370,6 +374,13 @@ struct buffer_head *ext4_find_inline_entry(struct inode *dir,
{
return NULL;
}
+static inline int ext4_delete_inline_entry(handle_t *handle,
+ struct inode *dir,
+ struct ext4_dir_entry_2 *de_del,
+ struct buffer_head *bh)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 17/19] ext4: let empty_dir handle inline dir.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (14 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 16/19] ext4: let ext4_delete_entry handle inline data Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 18/19] ext4: let ext4_rename " Tao Ma
2011-11-21 16:39 ` [PATCH V2 19/19] ext4: Enable ext4 inline support Tao Ma
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
empty_dir is used when deleting a dir. So it should handle
inline dir properly.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/inline.c | 69 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
fs/ext4/namei.c | 8 ++++++
fs/ext4/xattr.h | 6 ++++
3 files changed, 83 insertions(+), 0 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 302925e..4930e2e 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1407,6 +1407,75 @@ out:
return err;
}
+int empty_inline_dir(struct inode *dir, int *has_inline_data)
+{
+ int err, inline_size;
+ struct ext4_iloc iloc;
+ void *inline_pos;
+ unsigned int offset;
+ struct ext4_dir_entry_2 *de, *de1;
+ int ret = 1;
+
+ err = ext4_get_inode_loc(dir, &iloc);
+ if (err) {
+ EXT4_ERROR_INODE(dir, "error %d getting inode %lu block",
+ err, dir->i_ino);
+ return 1;
+ }
+
+ down_read(&EXT4_I(dir)->xattr_sem);
+ if (!ext4_has_inline_data(dir)) {
+ *has_inline_data = 0;
+ goto out;
+ }
+
+ de = ext4_get_inline_entry(dir, &iloc, 0,
+ &inline_pos, &inline_size);
+ de1 = ext4_get_inline_entry(dir, &iloc,
+ ext4_rec_len_from_disk(de->rec_len, inline_size),
+ &inline_pos, &inline_size);
+ if (le32_to_cpu(de->inode) != dir->i_ino ||
+ !le32_to_cpu(de1->inode) ||
+ strcmp(".", de->name) ||
+ strcmp("..", de1->name)) {
+ ext4_warning(dir->i_sb,
+ "bad directory (dir #%lu) - no `.' or `..'",
+ dir->i_ino);
+ ret = 1;
+ goto out;
+ }
+
+ offset = ext4_rec_len_from_disk(de->rec_len, inline_size) +
+ ext4_rec_len_from_disk(de1->rec_len, inline_size);
+ while (offset < dir->i_size) {
+ de = ext4_get_inline_entry(dir, &iloc, offset,
+ &inline_pos, &inline_size);
+ if (ext4_check_dir_entry(dir, NULL, de,
+ iloc.bh, inline_pos,
+ inline_size, offset)) {
+ ext4_warning(dir->i_sb,
+ "bad inline directory (dir #%lu) - "
+ "inode %u, rec_len %u, name_len %d"
+ "inline size %d\n",
+ dir->i_ino, le32_to_cpu(de->inode),
+ le16_to_cpu(de->rec_len), de->name_len,
+ inline_size);
+ ret = 1;
+ goto out;
+ }
+ if (le32_to_cpu(de->inode)) {
+ ret = 0;
+ goto out;
+ }
+ offset += ext4_rec_len_from_disk(de->rec_len, inline_size);
+ }
+
+out:
+ up_read(&EXT4_I(dir)->xattr_sem);
+ brelse(iloc.bh);
+ return ret;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index cefcec6..24cdb61 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -2015,6 +2015,14 @@ static int empty_dir(struct inode *inode)
struct super_block *sb;
int err = 0;
+ if (ext4_has_inline_data(inode)) {
+ int has_inline_data = 1;
+
+ err = empty_inline_dir(inode, &has_inline_data);
+ if (has_inline_data)
+ return err;
+ }
+
sb = inode->i_sb;
if (inode->i_size < EXT4_DIR_REC_LEN(1) + EXT4_DIR_REC_LEN(2) ||
!(bh = ext4_bread(NULL, inode, 0, 0, &err))) {
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index ccd994e..d889582 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -178,6 +178,7 @@ extern int ext4_delete_inline_entry(handle_t *handle,
struct inode *dir,
struct ext4_dir_entry_2 *de_del,
struct buffer_head *bh);
+extern int empty_inline_dir(struct inode *dir, int *has_inline_data);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -381,6 +382,11 @@ static inline int ext4_delete_inline_entry(handle_t *handle,
{
return 0;
}
+
+static inline int empty_inline_dir(struct inode *dir, int *has_inline_data)
+{
+ return 0;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 18/19] ext4: let ext4_rename handle inline dir.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (15 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 17/19] ext4: let empty_dir handle inline dir Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
2011-11-21 16:39 ` [PATCH V2 19/19] ext4: Enable ext4 inline support Tao Ma
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
In case of we rename a dir, ext4_rename has to read the dir block
and change its dotdot's information. The old ext4_rename encapsulated
the dir_block read into itself. So this patch try to add a new function
ext4_get_dir_block which get the dir buffer information so the
ext4_rename can handle it properly.
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/inline.c | 15 +++++++++++++++
fs/ext4/namei.c | 38 ++++++++++++++++++++++++++++++++------
fs/ext4/xattr.h | 12 ++++++++++++
3 files changed, 59 insertions(+), 6 deletions(-)
diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c
index 4930e2e..d73585e 100644
--- a/fs/ext4/inline.c
+++ b/fs/ext4/inline.c
@@ -1476,6 +1476,21 @@ out:
return ret;
}
+struct buffer_head *ext4_get_first_inline_block(struct inode *inode,
+ void **buf, int *buf_size, int *retval)
+{
+ struct ext4_iloc iloc;
+
+ *retval = ext4_get_inode_loc(inode, &iloc);
+ if (*retval)
+ return NULL;
+
+ *buf = ext4_raw_inode(&iloc)->i_block;
+ *buf_size = EXT4_MIN_INLINE_DATA_SIZE;
+
+ return iloc.bh;
+}
+
int ext4_destroy_inline_data(handle_t *handle, struct inode *inode)
{
int ret;
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c
index 24cdb61..cc4db68 100644
--- a/fs/ext4/namei.c
+++ b/fs/ext4/namei.c
@@ -2507,6 +2507,31 @@ retry:
(ext4_next_entry((struct ext4_dir_entry_2 *)(buffer), size)->inode)
/*
+ * Try to find buffer head where contains the parent block.
+ * It should be the inode block if it is inlined or the 1st block
+ * if it is a normal dir.
+ */
+static struct buffer_head *ext4_get_first_dir_block(handle_t *handle,
+ struct inode *inode,
+ void **buf,
+ int *buf_size,
+ int *retval)
+{
+ struct buffer_head *bh;
+
+ if (!ext4_has_inline_data(inode)) {
+ bh = ext4_bread(handle, inode, 0, 0, retval);
+ if (!bh)
+ return NULL;
+ *buf = bh->b_data;
+ *buf_size = inode->i_sb->s_blocksize;
+ return bh;
+ }
+
+ return ext4_get_first_inline_block(inode, buf, buf_size, retval);
+}
+
+/*
* Anybody can rename anything with this: the permission checks are left to the
* higher-level routines.
*/
@@ -2517,7 +2542,8 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
struct inode *old_inode, *new_inode;
struct buffer_head *old_bh, *new_bh, *dir_bh;
struct ext4_dir_entry_2 *old_de, *new_de;
- int retval, force_da_alloc = 0;
+ int buf_size, retval, force_da_alloc = 0;
+ void *dir_buf = NULL;
dquot_initialize(old_dir);
dquot_initialize(new_dir);
@@ -2564,11 +2590,12 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
goto end_rename;
}
retval = -EIO;
- dir_bh = ext4_bread(handle, old_inode, 0, 0, &retval);
+ dir_bh = ext4_get_first_dir_block(handle, old_inode,
+ &dir_buf, &buf_size, &retval);
if (!dir_bh)
goto end_rename;
- if (le32_to_cpu(PARENT_INO(dir_bh->b_data,
- old_dir->i_sb->s_blocksize)) != old_dir->i_ino)
+ if (le32_to_cpu(PARENT_INO(dir_buf,
+ buf_size)) != old_dir->i_ino)
goto end_rename;
retval = -EMLINK;
if (!new_inode && new_dir != old_dir &&
@@ -2648,8 +2675,7 @@ static int ext4_rename(struct inode *old_dir, struct dentry *old_dentry,
old_dir->i_ctime = old_dir->i_mtime = ext4_current_time(old_dir);
ext4_update_dx_flag(old_dir);
if (dir_bh) {
- PARENT_INO(dir_bh->b_data, new_dir->i_sb->s_blocksize) =
- cpu_to_le32(new_dir->i_ino);
+ PARENT_INO(dir_buf, buf_size) = cpu_to_le32(new_dir->i_ino);
BUFFER_TRACE(dir_bh, "call ext4_handle_dirty_metadata");
retval = ext4_handle_dirty_metadata(handle, old_inode, dir_bh);
if (retval) {
diff --git a/fs/ext4/xattr.h b/fs/ext4/xattr.h
index d889582..5300206 100644
--- a/fs/ext4/xattr.h
+++ b/fs/ext4/xattr.h
@@ -179,6 +179,10 @@ extern int ext4_delete_inline_entry(handle_t *handle,
struct ext4_dir_entry_2 *de_del,
struct buffer_head *bh);
extern int empty_inline_dir(struct inode *dir, int *has_inline_data);
+extern struct buffer_head *ext4_get_first_inline_block(struct inode *inode,
+ void **buf,
+ int *buf_size,
+ int *retval);
# else /* CONFIG_EXT4_FS_XATTR */
static inline int
@@ -387,6 +391,14 @@ static inline int empty_inline_dir(struct inode *dir, int *has_inline_data)
{
return 0;
}
+
+static inline struct buffer_head *
+ext4_get_first_inline_block(struct inode *inode,
+ void **buf, int *buf_size,
+ int *retval)
+{
+ return NULL;
+}
# endif /* CONFIG_EXT4_FS_XATTR */
#ifdef CONFIG_EXT4_FS_SECURITY
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread* [PATCH V2 19/19] ext4: Enable ext4 inline support.
2011-11-21 16:39 ` [PATCH V2 01/19] ext4: Move extra inode read to a new function Tao Ma
` (16 preceding siblings ...)
2011-11-21 16:39 ` [PATCH V2 18/19] ext4: let ext4_rename " Tao Ma
@ 2011-11-21 16:39 ` Tao Ma
17 siblings, 0 replies; 20+ messages in thread
From: Tao Ma @ 2011-11-21 16:39 UTC (permalink / raw)
To: linux-ext4; +Cc: tytso, linux-fsdevel, linux-kernel, adilger
From: Tao Ma <boyu.mt@taobao.com>
Signed-off-by: Tao Ma <boyu.mt@taobao.com>
---
fs/ext4/ext4.h | 1 +
fs/ext4/ialloc.c | 4 ++++
2 files changed, 5 insertions(+), 0 deletions(-)
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index a88bc00..784893a 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1414,6 +1414,7 @@ static inline void ext4_clear_state_flags(struct ext4_inode_info *ei)
#define EXT4_FEATURE_INCOMPAT_FLEX_BG 0x0200
#define EXT4_FEATURE_INCOMPAT_EA_INODE 0x0400 /* EA in inode */
#define EXT4_FEATURE_INCOMPAT_DIRDATA 0x1000 /* data in dirent */
+#define EXT4_FEATURE_INCOMPAT_INLINEDATA 0x2000 /* data in inode */
#define EXT2_FEATURE_COMPAT_SUPP EXT4_FEATURE_COMPAT_EXT_ATTR
#define EXT2_FEATURE_INCOMPAT_SUPP (EXT4_FEATURE_INCOMPAT_FILETYPE| \
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index 00beb4f..4f7fe21 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -897,6 +897,10 @@ got:
ei->i_extra_isize = EXT4_SB(sb)->s_want_extra_isize;
+ ei->i_inline_off = 0;
+ if (EXT4_HAS_INCOMPAT_FEATURE(sb, EXT4_FEATURE_INCOMPAT_INLINEDATA))
+ ext4_set_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA);
+
ret = inode;
dquot_initialize(inode);
err = dquot_alloc_inode(inode);
--
1.7.0.4
^ permalink raw reply related [flat|nested] 20+ messages in thread