ext4: Use bitops to read/modify EXT4_I(inode)->i_state
At several places we modify EXT4_I(inode)->i_state without holding
i_mutex (ext4_release_file, ext4_bmap, ext4_journalled_writepage,
ext4_do_update_inode, ...). These modifications are racy and we can
lose updates to i_state. So convert handling of i_state to use bitops
which are atomic.
Cc: Jan Kara <jack@suse.cz>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 307ecd1..ffe1334 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -313,17 +313,6 @@
return flags & EXT4_OTHER_FLMASK;
}
-/*
- * Inode dynamic state flags
- */
-#define EXT4_STATE_JDATA 0x00000001 /* journaled data exists */
-#define EXT4_STATE_NEW 0x00000002 /* inode is newly created */
-#define EXT4_STATE_XATTR 0x00000004 /* has in-inode xattrs */
-#define EXT4_STATE_NO_EXPAND 0x00000008 /* No space for expansion */
-#define EXT4_STATE_DA_ALLOC_CLOSE 0x00000010 /* Alloc DA blks on close */
-#define EXT4_STATE_EXT_MIGRATE 0x00000020 /* Inode is migrating */
-#define EXT4_STATE_DIO_UNWRITTEN 0x00000040 /* need convert on dio done*/
-
/* Used to pass group descriptor data when online resize is done */
struct ext4_new_group_input {
__u32 group; /* Group number for this data */
@@ -631,7 +620,7 @@
* near to their parent directory's inode.
*/
ext4_group_t i_block_group;
- __u32 i_state; /* Dynamic state flags for ext4 */
+ unsigned long i_state_flags; /* Dynamic state flags */
ext4_lblk_t i_dir_start_lookup;
#ifdef CONFIG_EXT4_FS_XATTR
@@ -1051,6 +1040,34 @@
(ino >= EXT4_FIRST_INO(sb) &&
ino <= le32_to_cpu(EXT4_SB(sb)->s_es->s_inodes_count));
}
+
+/*
+ * Inode dynamic state flags
+ */
+enum {
+ EXT4_STATE_JDATA, /* journaled data exists */
+ EXT4_STATE_NEW, /* inode is newly created */
+ EXT4_STATE_XATTR, /* has in-inode xattrs */
+ EXT4_STATE_NO_EXPAND, /* No space for expansion */
+ EXT4_STATE_DA_ALLOC_CLOSE, /* Alloc DA blks on close */
+ EXT4_STATE_EXT_MIGRATE, /* Inode is migrating */
+ EXT4_STATE_DIO_UNWRITTEN, /* need convert on dio done*/
+};
+
+static inline int ext4_test_inode_state(struct inode *inode, int bit)
+{
+ return test_bit(bit, &EXT4_I(inode)->i_state_flags);
+}
+
+static inline void ext4_set_inode_state(struct inode *inode, int bit)
+{
+ set_bit(bit, &EXT4_I(inode)->i_state_flags);
+}
+
+static inline void ext4_clear_inode_state(struct inode *inode, int bit)
+{
+ clear_bit(bit, &EXT4_I(inode)->i_state_flags);
+}
#else
/* Assume that user mode programs are passing in an ext4fs superblock, not
* a kernel struct super_block. This will allow us to call the feature-test
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index c568779..5461615 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -3076,7 +3076,7 @@
if (io)
io->flag = DIO_AIO_UNWRITTEN;
else
- EXT4_I(inode)->i_state |= EXT4_STATE_DIO_UNWRITTEN;
+ ext4_set_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
goto out;
}
/* async DIO end_io complete, convert the filled extent to written */
@@ -3362,8 +3362,8 @@
if (io)
io->flag = DIO_AIO_UNWRITTEN;
else
- EXT4_I(inode)->i_state |=
- EXT4_STATE_DIO_UNWRITTEN;;
+ ext4_set_inode_state(inode,
+ EXT4_STATE_DIO_UNWRITTEN);
}
}
err = ext4_ext_insert_extent(handle, inode, path, &newex, flags);
@@ -3739,7 +3739,7 @@
int error = 0;
/* in-inode? */
- if (EXT4_I(inode)->i_state & EXT4_STATE_XATTR) {
+ if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
struct ext4_iloc iloc;
int offset; /* offset of xattr in inode */
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 9630583..f6071ce 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -35,9 +35,9 @@
*/
static int ext4_release_file(struct inode *inode, struct file *filp)
{
- if (EXT4_I(inode)->i_state & EXT4_STATE_DA_ALLOC_CLOSE) {
+ if (ext4_test_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE)) {
ext4_alloc_da_blocks(inode);
- EXT4_I(inode)->i_state &= ~EXT4_STATE_DA_ALLOC_CLOSE;
+ ext4_clear_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE);
}
/* if we are the last writer on the inode, drop the block reservation */
if ((filp->f_mode & FMODE_WRITE) &&
diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c
index f3624ea..2fab5ad 100644
--- a/fs/ext4/ialloc.c
+++ b/fs/ext4/ialloc.c
@@ -1029,7 +1029,8 @@
inode->i_generation = sbi->s_next_generation++;
spin_unlock(&sbi->s_next_gen_lock);
- ei->i_state = EXT4_STATE_NEW;
+ ei->i_state_flags = 0;
+ ext4_set_inode_state(inode, EXT4_STATE_NEW);
ei->i_extra_isize = EXT4_SB(sb)->s_want_extra_isize;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 1a3d7b2..3e53011 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1307,7 +1307,7 @@
* i_data's format changing. Force the migrate
* to fail by clearing migrate flags
*/
- EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
+ ext4_clear_inode_state(inode, EXT4_STATE_EXT_MIGRATE);
}
/*
@@ -1794,7 +1794,7 @@
new_i_size = pos + copied;
if (new_i_size > inode->i_size)
i_size_write(inode, pos+copied);
- EXT4_I(inode)->i_state |= EXT4_STATE_JDATA;
+ ext4_set_inode_state(inode, EXT4_STATE_JDATA);
if (new_i_size > EXT4_I(inode)->i_disksize) {
ext4_update_i_disksize(inode, new_i_size);
ret2 = ext4_mark_inode_dirty(handle, inode);
@@ -2632,7 +2632,7 @@
ret = err;
walk_page_buffers(handle, page_bufs, 0, len, NULL, bput_one);
- EXT4_I(inode)->i_state |= EXT4_STATE_JDATA;
+ ext4_set_inode_state(inode, EXT4_STATE_JDATA);
out:
return ret;
}
@@ -3303,7 +3303,8 @@
filemap_write_and_wait(mapping);
}
- if (EXT4_JOURNAL(inode) && EXT4_I(inode)->i_state & EXT4_STATE_JDATA) {
+ if (EXT4_JOURNAL(inode) &&
+ ext4_test_inode_state(inode, EXT4_STATE_JDATA)) {
/*
* This is a REALLY heavyweight approach, but the use of
* bmap on dirty files is expected to be extremely rare:
@@ -3322,7 +3323,7 @@
* everything they get.
*/
- EXT4_I(inode)->i_state &= ~EXT4_STATE_JDATA;
+ ext4_clear_inode_state(inode, EXT4_STATE_JDATA);
journal = EXT4_JOURNAL(inode);
jbd2_journal_lock_updates(journal);
err = jbd2_journal_flush(journal);
@@ -3790,8 +3791,8 @@
if (ret != -EIOCBQUEUED && ret <= 0 && iocb->private) {
ext4_free_io_end(iocb->private);
iocb->private = NULL;
- } else if (ret > 0 && (EXT4_I(inode)->i_state &
- EXT4_STATE_DIO_UNWRITTEN)) {
+ } else if (ret > 0 && ext4_test_inode_state(inode,
+ EXT4_STATE_DIO_UNWRITTEN)) {
int err;
/*
* for non AIO case, since the IO is already
@@ -3801,7 +3802,7 @@
offset, ret);
if (err < 0)
ret = err;
- EXT4_I(inode)->i_state &= ~EXT4_STATE_DIO_UNWRITTEN;
+ ext4_clear_inode_state(inode, EXT4_STATE_DIO_UNWRITTEN);
}
return ret;
}
@@ -4457,7 +4458,7 @@
return;
if (inode->i_size == 0 && !test_opt(inode->i_sb, NO_AUTO_DA_ALLOC))
- ei->i_state |= EXT4_STATE_DA_ALLOC_CLOSE;
+ ext4_set_inode_state(inode, EXT4_STATE_DA_ALLOC_CLOSE);
if (EXT4_I(inode)->i_flags & EXT4_EXTENTS_FL) {
ext4_ext_truncate(inode);
@@ -4743,7 +4744,7 @@
{
/* We have all inode data except xattrs in memory here. */
return __ext4_get_inode_loc(inode, iloc,
- !(EXT4_I(inode)->i_state & EXT4_STATE_XATTR));
+ !ext4_test_inode_state(inode, EXT4_STATE_XATTR));
}
void ext4_set_inode_flags(struct inode *inode)
@@ -4837,7 +4838,7 @@
}
inode->i_nlink = le16_to_cpu(raw_inode->i_links_count);
- ei->i_state = 0;
+ ei->i_state_flags = 0;
ei->i_dir_start_lookup = 0;
ei->i_dtime = le32_to_cpu(raw_inode->i_dtime);
/* We now have enough fields to check if the inode was active or not.
@@ -4920,7 +4921,7 @@
EXT4_GOOD_OLD_INODE_SIZE +
ei->i_extra_isize;
if (*magic == cpu_to_le32(EXT4_XATTR_MAGIC))
- ei->i_state |= EXT4_STATE_XATTR;
+ ext4_set_inode_state(inode, EXT4_STATE_XATTR);
}
} else
ei->i_extra_isize = 0;
@@ -5060,7 +5061,7 @@
/* For fields not not tracking in the in-memory inode,
* initialise them to zero for new inodes. */
- if (ei->i_state & EXT4_STATE_NEW)
+ if (ext4_test_inode_state(inode, EXT4_STATE_NEW))
memset(raw_inode, 0, EXT4_SB(inode->i_sb)->s_inode_size);
ext4_get_inode_flags(ei);
@@ -5156,7 +5157,7 @@
rc = ext4_handle_dirty_metadata(handle, inode, bh);
if (!err)
err = rc;
- ei->i_state &= ~EXT4_STATE_NEW;
+ ext4_clear_inode_state(inode, EXT4_STATE_NEW);
ext4_update_inode_fsync_trans(handle, inode, 0);
out_brelse:
@@ -5580,8 +5581,8 @@
entry = IFIRST(header);
/* No extended attributes present */
- if (!(EXT4_I(inode)->i_state & EXT4_STATE_XATTR) ||
- header->h_magic != cpu_to_le32(EXT4_XATTR_MAGIC)) {
+ if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR) ||
+ header->h_magic != cpu_to_le32(EXT4_XATTR_MAGIC)) {
memset((void *)raw_inode + EXT4_GOOD_OLD_INODE_SIZE, 0,
new_extra_isize);
EXT4_I(inode)->i_extra_isize = new_extra_isize;
@@ -5625,7 +5626,7 @@
err = ext4_reserve_inode_write(handle, inode, &iloc);
if (ext4_handle_valid(handle) &&
EXT4_I(inode)->i_extra_isize < sbi->s_want_extra_isize &&
- !(EXT4_I(inode)->i_state & EXT4_STATE_NO_EXPAND)) {
+ !ext4_test_inode_state(inode, EXT4_STATE_NO_EXPAND)) {
/*
* We need extra buffer credits since we may write into EA block
* with this same handle. If journal_extend fails, then it will
@@ -5639,7 +5640,8 @@
sbi->s_want_extra_isize,
iloc, handle);
if (ret) {
- EXT4_I(inode)->i_state |= EXT4_STATE_NO_EXPAND;
+ ext4_set_inode_state(inode,
+ EXT4_STATE_NO_EXPAND);
if (mnt_count !=
le16_to_cpu(sbi->s_es->s_mnt_count)) {
ext4_warning(inode->i_sb, __func__,
diff --git a/fs/ext4/migrate.c b/fs/ext4/migrate.c
index 8141581..46a4101 100644
--- a/fs/ext4/migrate.c
+++ b/fs/ext4/migrate.c
@@ -365,12 +365,12 @@
* happened after we started the migrate. We need to
* fail the migrate
*/
- if (!(EXT4_I(inode)->i_state & EXT4_STATE_EXT_MIGRATE)) {
+ if (!ext4_test_inode_state(inode, EXT4_STATE_EXT_MIGRATE)) {
retval = -EAGAIN;
up_write(&EXT4_I(inode)->i_data_sem);
goto err_out;
} else
- EXT4_I(inode)->i_state &= ~EXT4_STATE_EXT_MIGRATE;
+ ext4_clear_inode_state(inode, EXT4_STATE_EXT_MIGRATE);
/*
* We have the extent map build with the tmp inode.
* Now copy the i_data across
@@ -533,7 +533,7 @@
* allocation.
*/
down_read((&EXT4_I(inode)->i_data_sem));
- EXT4_I(inode)->i_state |= EXT4_STATE_EXT_MIGRATE;
+ ext4_set_inode_state(inode, EXT4_STATE_EXT_MIGRATE);
up_read((&EXT4_I(inode)->i_data_sem));
handle = ext4_journal_start(inode, 1);
diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c
index f3a2f7e..c619a7e 100644
--- a/fs/ext4/xattr.c
+++ b/fs/ext4/xattr.c
@@ -267,7 +267,7 @@
void *end;
int error;
- if (!(EXT4_I(inode)->i_state & EXT4_STATE_XATTR))
+ if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR))
return -ENODATA;
error = ext4_get_inode_loc(inode, &iloc);
if (error)
@@ -396,7 +396,7 @@
void *end;
int error;
- if (!(EXT4_I(inode)->i_state & EXT4_STATE_XATTR))
+ if (!ext4_test_inode_state(inode, EXT4_STATE_XATTR))
return 0;
error = ext4_get_inode_loc(inode, &iloc);
if (error)
@@ -908,7 +908,7 @@
is->s.base = is->s.first = IFIRST(header);
is->s.here = is->s.first;
is->s.end = (void *)raw_inode + EXT4_SB(inode->i_sb)->s_inode_size;
- if (EXT4_I(inode)->i_state & EXT4_STATE_XATTR) {
+ if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {
error = ext4_xattr_check_names(IFIRST(header), is->s.end);
if (error)
return error;
@@ -940,10 +940,10 @@
header = IHDR(inode, ext4_raw_inode(&is->iloc));
if (!IS_LAST_ENTRY(s->first)) {
header->h_magic = cpu_to_le32(EXT4_XATTR_MAGIC);
- EXT4_I(inode)->i_state |= EXT4_STATE_XATTR;
+ ext4_set_inode_state(inode, EXT4_STATE_XATTR);
} else {
header->h_magic = cpu_to_le32(0);
- EXT4_I(inode)->i_state &= ~EXT4_STATE_XATTR;
+ ext4_clear_inode_state(inode, EXT4_STATE_XATTR);
}
return 0;
}
@@ -986,8 +986,8 @@
if (strlen(name) > 255)
return -ERANGE;
down_write(&EXT4_I(inode)->xattr_sem);
- no_expand = EXT4_I(inode)->i_state & EXT4_STATE_NO_EXPAND;
- EXT4_I(inode)->i_state |= EXT4_STATE_NO_EXPAND;
+ no_expand = ext4_test_inode_state(inode, EXT4_STATE_NO_EXPAND);
+ ext4_set_inode_state(inode, EXT4_STATE_NO_EXPAND);
error = ext4_get_inode_loc(inode, &is.iloc);
if (error)
@@ -997,10 +997,10 @@
if (error)
goto cleanup;
- if (EXT4_I(inode)->i_state & EXT4_STATE_NEW) {
+ if (ext4_test_inode_state(inode, EXT4_STATE_NEW)) {
struct ext4_inode *raw_inode = ext4_raw_inode(&is.iloc);
memset(raw_inode, 0, EXT4_SB(inode->i_sb)->s_inode_size);
- EXT4_I(inode)->i_state &= ~EXT4_STATE_NEW;
+ ext4_clear_inode_state(inode, EXT4_STATE_NEW);
}
error = ext4_xattr_ibody_find(inode, &i, &is);
@@ -1052,7 +1052,7 @@
ext4_xattr_update_super_block(handle, inode->i_sb);
inode->i_ctime = ext4_current_time(inode);
if (!value)
- EXT4_I(inode)->i_state &= ~EXT4_STATE_NO_EXPAND;
+ ext4_clear_inode_state(inode, EXT4_STATE_NO_EXPAND);
error = ext4_mark_iloc_dirty(handle, inode, &is.iloc);
/*
* The bh is consumed by ext4_mark_iloc_dirty, even with
@@ -1067,7 +1067,7 @@
brelse(is.iloc.bh);
brelse(bs.bh);
if (no_expand == 0)
- EXT4_I(inode)->i_state &= ~EXT4_STATE_NO_EXPAND;
+ ext4_clear_inode_state(inode, EXT4_STATE_NO_EXPAND);
up_write(&EXT4_I(inode)->xattr_sem);
return error;
}
