fs: push i_mutex and filemap_write_and_wait down into ->fsync() handlers
Btrfs needs to be able to control how filemap_write_and_wait_range() is called
in fsync to make it less of a painful operation, so push down taking i_mutex and
the calling of filemap_write_and_wait() down into the ->fsync() handlers. Some
file systems can drop taking the i_mutex altogether it seems, like ext3 and
ocfs2. For correctness sake I just pushed everything down in all cases to make
sure that we keep the current behavior the same for everybody, and then each
individual fs maintainer can make up their mind about what to do from there.
Thanks,
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
diff --git a/Documentation/filesystems/Locking b/Documentation/filesystems/Locking
index 9b6ed7c..ca7e252 100644
--- a/Documentation/filesystems/Locking
+++ b/Documentation/filesystems/Locking
@@ -412,7 +412,7 @@
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *);
int (*release) (struct inode *, struct file *);
- int (*fsync) (struct file *, int datasync);
+ int (*fsync) (struct file *, loff_t start, loff_t end, int datasync);
int (*aio_fsync) (struct kiocb *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
@@ -438,9 +438,7 @@
locking rules:
All may block except for ->setlease.
- No VFS locks held on entry except for ->fsync and ->setlease.
-
-->fsync() has i_mutex on inode.
+ No VFS locks held on entry except for ->setlease.
->setlease has the file_list_lock held and must not sleep.
diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting
index 6b96773..7f8861d 100644
--- a/Documentation/filesystems/porting
+++ b/Documentation/filesystems/porting
@@ -421,3 +421,10 @@
offset is less than i_size and SEEK_DATA is specified, return the same offset.
If the above is true for the offset and you are given SEEK_HOLE, return the end
of the file. If the offset is i_size or greater return -ENXIO in either case.
+
+[mandatory]
+ If you have your own ->fsync() you must make sure to call
+filemap_write_and_wait_range() so that all dirty pages are synced out properly.
+You must also keep in mind that ->fsync() is not called with i_mutex held
+anymore, so if you require i_mutex locking you must make sure to take it and
+release it yourself.
diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index 6bf85b78..eff6617 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -777,7 +777,7 @@
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *);
int (*release) (struct inode *, struct file *);
- int (*fsync) (struct file *, int datasync);
+ int (*fsync) (struct file *, loff_t, loff_t, int datasync);
int (*aio_fsync) (struct kiocb *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
diff --git a/arch/powerpc/platforms/cell/spufs/file.c b/arch/powerpc/platforms/cell/spufs/file.c
index 3c7c3f8..fb59c46 100644
--- a/arch/powerpc/platforms/cell/spufs/file.c
+++ b/arch/powerpc/platforms/cell/spufs/file.c
@@ -1850,9 +1850,16 @@
return ret;
}
-static int spufs_mfc_fsync(struct file *file, int datasync)
+static int spufs_mfc_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
- return spufs_mfc_flush(file, NULL);
+ struct inode *inode = file->f_path.dentry->d_inode;
+ int err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (!err) {
+ mutex_lock(&inode->i_mutex);
+ err = spufs_mfc_flush(file, NULL);
+ mutex_unlock(&inode->i_mutex);
+ }
+ return err;
}
static int spufs_mfc_fasync(int fd, struct file *file, int on)
diff --git a/drivers/char/ps3flash.c b/drivers/char/ps3flash.c
index 5a06787..d0c57c2 100644
--- a/drivers/char/ps3flash.c
+++ b/drivers/char/ps3flash.c
@@ -309,9 +309,14 @@
return ps3flash_writeback(ps3flash_dev);
}
-static int ps3flash_fsync(struct file *file, int datasync)
+static int ps3flash_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
- return ps3flash_writeback(ps3flash_dev);
+ struct inode *inode = file->f_path.dentry->d_inode;
+ int err;
+ mutex_lock(&inode->i_mutex);
+ err = ps3flash_writeback(ps3flash_dev);
+ mutex_unlock(&inode->i_mutex);
+ return err;
}
static irqreturn_t ps3flash_interrupt(int irq, void *data)
diff --git a/drivers/mtd/ubi/cdev.c b/drivers/mtd/ubi/cdev.c
index 191f3bb..3320a50 100644
--- a/drivers/mtd/ubi/cdev.c
+++ b/drivers/mtd/ubi/cdev.c
@@ -189,12 +189,16 @@
return new_offset;
}
-static int vol_cdev_fsync(struct file *file, int datasync)
+static int vol_cdev_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct ubi_volume_desc *desc = file->private_data;
struct ubi_device *ubi = desc->vol->ubi;
-
- return ubi_sync(ubi->ubi_num);
+ struct inode *inode = file->f_path.dentry->d_inode;
+ int err;
+ mutex_lock(&inode->i_mutex);
+ err = ubi_sync(ubi->ubi_num);
+ mutex_unlock(&inode->i_mutex);
+ return err;
}
diff --git a/drivers/staging/pohmelfs/inode.c b/drivers/staging/pohmelfs/inode.c
index c0f0ac7..f3c6060 100644
--- a/drivers/staging/pohmelfs/inode.c
+++ b/drivers/staging/pohmelfs/inode.c
@@ -887,11 +887,16 @@
/*
* We want fsync() to work on POHMELFS.
*/
-static int pohmelfs_fsync(struct file *file, int datasync)
+static int pohmelfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
-
- return sync_inode_metadata(inode, 1);
+ int err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (!err) {
+ mutex_lock(&inode->i_mutex);
+ err = sync_inode_metadata(inode, 1);
+ mutex_unlock(&inode->i_mutex);
+ }
+ return err;
}
ssize_t pohmelfs_write(struct file *file, const char __user *buf,
diff --git a/drivers/usb/gadget/printer.c b/drivers/usb/gadget/printer.c
index 271ef94..978e6a10 100644
--- a/drivers/usb/gadget/printer.c
+++ b/drivers/usb/gadget/printer.c
@@ -795,12 +795,14 @@
}
static int
-printer_fsync(struct file *fd, int datasync)
+printer_fsync(struct file *fd, loff_t start, loff_t end, int datasync)
{
struct printer_dev *dev = fd->private_data;
+ struct inode *inode = fd->f_path.dentry->d_inode;
unsigned long flags;
int tx_list_empty;
+ mutex_lock(&inode->i_mutex);
spin_lock_irqsave(&dev->lock, flags);
tx_list_empty = (likely(list_empty(&dev->tx_reqs)));
spin_unlock_irqrestore(&dev->lock, flags);
@@ -810,6 +812,7 @@
wait_event_interruptible(dev->tx_flush_wait,
(likely(list_empty(&dev->tx_reqs_active))));
}
+ mutex_unlock(&inode->i_mutex);
return 0;
}
diff --git a/drivers/video/fb_defio.c b/drivers/video/fb_defio.c
index 8040001..32814e8 100644
--- a/drivers/video/fb_defio.c
+++ b/drivers/video/fb_defio.c
@@ -66,19 +66,26 @@
return 0;
}
-int fb_deferred_io_fsync(struct file *file, int datasync)
+int fb_deferred_io_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct fb_info *info = file->private_data;
+ struct inode *inode = file->f_path.dentry->d_inode;
+ int err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
/* Skip if deferred io is compiled-in but disabled on this fbdev */
if (!info->fbdefio)
return 0;
+ mutex_lock(&inode->i_mutex);
/* Kill off the delayed work */
cancel_delayed_work_sync(&info->deferred_work);
/* Run it immediately */
- return schedule_delayed_work(&info->deferred_work, 0);
+ err = schedule_delayed_work(&info->deferred_work, 0);
+ mutex_unlock(&inode->i_mutex);
+ return err;
}
EXPORT_SYMBOL_GPL(fb_deferred_io_fsync);
diff --git a/fs/9p/v9fs_vfs.h b/fs/9p/v9fs_vfs.h
index 4014160..46ce357 100644
--- a/fs/9p/v9fs_vfs.h
+++ b/fs/9p/v9fs_vfs.h
@@ -70,7 +70,8 @@
ssize_t v9fs_fid_readn(struct p9_fid *, char *, char __user *, u32, u64);
void v9fs_blank_wstat(struct p9_wstat *wstat);
int v9fs_vfs_setattr_dotl(struct dentry *, struct iattr *);
-int v9fs_file_fsync_dotl(struct file *filp, int datasync);
+int v9fs_file_fsync_dotl(struct file *filp, loff_t start, loff_t end,
+ int datasync);
ssize_t v9fs_file_write_internal(struct inode *, struct p9_fid *,
const char __user *, size_t, loff_t *, int);
int v9fs_refresh_inode(struct p9_fid *fid, struct inode *inode);
diff --git a/fs/9p/vfs_file.c b/fs/9p/vfs_file.c
index ffed558..3c173fc 100644
--- a/fs/9p/vfs_file.c
+++ b/fs/9p/vfs_file.c
@@ -519,32 +519,50 @@
}
-static int v9fs_file_fsync(struct file *filp, int datasync)
+static int v9fs_file_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct p9_fid *fid;
+ struct inode *inode = filp->f_mapping->host;
struct p9_wstat wstat;
int retval;
+ retval = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (retval)
+ return retval;
+
+ mutex_lock(&inode->i_mutex);
P9_DPRINTK(P9_DEBUG_VFS, "filp %p datasync %x\n", filp, datasync);
fid = filp->private_data;
v9fs_blank_wstat(&wstat);
retval = p9_client_wstat(fid, &wstat);
+ mutex_unlock(&inode->i_mutex);
+
return retval;
}
-int v9fs_file_fsync_dotl(struct file *filp, int datasync)
+int v9fs_file_fsync_dotl(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct p9_fid *fid;
+ struct inode *inode = filp->f_mapping->host;
int retval;
+ retval = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (retval)
+ return retval;
+
+ mutex_lock(&inode->i_mutex);
P9_DPRINTK(P9_DEBUG_VFS, "v9fs_file_fsync_dotl: filp %p datasync %x\n",
filp, datasync);
fid = filp->private_data;
retval = p9_client_fsync(fid, datasync);
+ mutex_unlock(&inode->i_mutex);
+
return retval;
}
diff --git a/fs/affs/affs.h b/fs/affs/affs.h
index 0e95f73..c2b9c79 100644
--- a/fs/affs/affs.h
+++ b/fs/affs/affs.h
@@ -182,7 +182,7 @@
void affs_free_prealloc(struct inode *inode);
extern void affs_truncate(struct inode *);
-int affs_file_fsync(struct file *, int);
+int affs_file_fsync(struct file *, loff_t, loff_t, int);
/* dir.c */
diff --git a/fs/affs/file.c b/fs/affs/file.c
index acf321b..2f4c935c 100644
--- a/fs/affs/file.c
+++ b/fs/affs/file.c
@@ -923,14 +923,20 @@
affs_free_prealloc(inode);
}
-int affs_file_fsync(struct file *filp, int datasync)
+int affs_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
struct inode *inode = filp->f_mapping->host;
int ret, err;
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
+ mutex_lock(&inode->i_mutex);
ret = write_inode_now(inode, 0);
err = sync_blockdev(inode->i_sb->s_bdev);
if (!ret)
ret = err;
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/afs/internal.h b/fs/afs/internal.h
index f396d33..d2b0888 100644
--- a/fs/afs/internal.h
+++ b/fs/afs/internal.h
@@ -750,7 +750,7 @@
extern ssize_t afs_file_write(struct kiocb *, const struct iovec *,
unsigned long, loff_t);
extern int afs_writeback_all(struct afs_vnode *);
-extern int afs_fsync(struct file *, int);
+extern int afs_fsync(struct file *, loff_t, loff_t, int);
/*****************************************************************************/
diff --git a/fs/afs/write.c b/fs/afs/write.c
index b806285..9aa52d9 100644
--- a/fs/afs/write.c
+++ b/fs/afs/write.c
@@ -681,9 +681,10 @@
* - the return status from this call provides a reliable indication of
* whether any write errors occurred for this process.
*/
-int afs_fsync(struct file *file, int datasync)
+int afs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct dentry *dentry = file->f_path.dentry;
+ struct inode *inode = file->f_mapping->host;
struct afs_writeback *wb, *xwb;
struct afs_vnode *vnode = AFS_FS_I(dentry->d_inode);
int ret;
@@ -692,12 +693,19 @@
vnode->fid.vid, vnode->fid.vnode, dentry->d_name.name,
datasync);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
/* use a writeback record as a marker in the queue - when this reaches
* the front of the queue, all the outstanding writes are either
* completed or rejected */
wb = kzalloc(sizeof(*wb), GFP_KERNEL);
- if (!wb)
- return -ENOMEM;
+ if (!wb) {
+ ret = -ENOMEM;
+ goto out;
+ }
wb->vnode = vnode;
wb->first = 0;
wb->last = -1;
@@ -720,7 +728,7 @@
if (ret < 0) {
afs_put_writeback(wb);
_leave(" = %d [wb]", ret);
- return ret;
+ goto out;
}
/* wait for the preceding writes to actually complete */
@@ -729,6 +737,8 @@
vnode->writebacks.next == &wb->link);
afs_put_writeback(wb);
_leave(" = %d", ret);
+out:
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/bad_inode.c b/fs/bad_inode.c
index f024d8a..9205cf2 100644
--- a/fs/bad_inode.c
+++ b/fs/bad_inode.c
@@ -87,7 +87,8 @@
return -EIO;
}
-static int bad_file_fsync(struct file *file, int datasync)
+static int bad_file_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
return -EIO;
}
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 966617a..9fb0b15 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -378,7 +378,7 @@
return retval;
}
-int blkdev_fsync(struct file *filp, int datasync)
+int blkdev_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
struct inode *bd_inode = filp->f_mapping->host;
struct block_device *bdev = I_BDEV(bd_inode);
@@ -389,14 +389,10 @@
* i_mutex and doing so causes performance issues with concurrent
* O_SYNC writers to a block device.
*/
- mutex_unlock(&bd_inode->i_mutex);
-
error = blkdev_issue_flush(bdev, GFP_KERNEL, NULL);
if (error == -EOPNOTSUPP)
error = 0;
- mutex_lock(&bd_inode->i_mutex);
-
return error;
}
EXPORT_SYMBOL(blkdev_fsync);
diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index f1ff62b..82be74ef 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -2605,7 +2605,7 @@
int btrfs_add_inode_defrag(struct btrfs_trans_handle *trans,
struct inode *inode);
int btrfs_run_defrag_inodes(struct btrfs_fs_info *fs_info);
-int btrfs_sync_file(struct file *file, int datasync);
+int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync);
int btrfs_drop_extent_cache(struct inode *inode, u64 start, u64 end,
int skip_pinned);
extern const struct file_operations btrfs_file_operations;
diff --git a/fs/btrfs/file.c b/fs/btrfs/file.c
index bd4d061..59cbdb1 100644
--- a/fs/btrfs/file.c
+++ b/fs/btrfs/file.c
@@ -1452,7 +1452,7 @@
* important optimization for directories because holding the mutex prevents
* new operations on the dir while we write to disk.
*/
-int btrfs_sync_file(struct file *file, int datasync)
+int btrfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
{
struct dentry *dentry = file->f_path.dentry;
struct inode *inode = dentry->d_inode;
@@ -1462,9 +1462,13 @@
trace_btrfs_sync_file(file, datasync);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
/* we wait first, since the writeback may change the inode */
root->log_batch++;
- /* the VFS called filemap_fdatawrite for us */
btrfs_wait_ordered_range(inode, 0, (u64)-1);
root->log_batch++;
@@ -1472,8 +1476,10 @@
* check the transaction that last modified this inode
* and see if its already been committed
*/
- if (!BTRFS_I(inode)->last_trans)
+ if (!BTRFS_I(inode)->last_trans) {
+ mutex_unlock(&inode->i_mutex);
goto out;
+ }
/*
* if the last transaction that changed this file was before
@@ -1484,6 +1490,7 @@
if (BTRFS_I(inode)->last_trans <=
root->fs_info->last_trans_committed) {
BTRFS_I(inode)->last_trans = 0;
+ mutex_unlock(&inode->i_mutex);
goto out;
}
@@ -1496,12 +1503,15 @@
trans = btrfs_start_transaction(root, 0);
if (IS_ERR(trans)) {
ret = PTR_ERR(trans);
+ mutex_unlock(&inode->i_mutex);
goto out;
}
ret = btrfs_log_dentry_safe(trans, root, dentry);
- if (ret < 0)
+ if (ret < 0) {
+ mutex_unlock(&inode->i_mutex);
goto out;
+ }
/* we've logged all the items and now have a consistent
* version of the file in the log. It is possible that
@@ -1513,7 +1523,7 @@
* file again, but that will end up using the synchronization
* inside btrfs_sync_log to keep things safe.
*/
- mutex_unlock(&dentry->d_inode->i_mutex);
+ mutex_unlock(&inode->i_mutex);
if (ret != BTRFS_NO_LOG_SYNC) {
if (ret > 0) {
@@ -1528,7 +1538,6 @@
} else {
ret = btrfs_end_transaction(trans, root);
}
- mutex_lock(&dentry->d_inode->i_mutex);
out:
return ret > 0 ? -EIO : ret;
}
diff --git a/fs/ceph/caps.c b/fs/ceph/caps.c
index f605753..8d74ad7 100644
--- a/fs/ceph/caps.c
+++ b/fs/ceph/caps.c
@@ -1811,7 +1811,7 @@
spin_unlock(&ci->i_unsafe_lock);
}
-int ceph_fsync(struct file *file, int datasync)
+int ceph_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
struct ceph_inode_info *ci = ceph_inode(inode);
@@ -1822,9 +1822,10 @@
dout("fsync %p%s\n", inode, datasync ? " datasync" : "");
sync_write_wait(inode);
- ret = filemap_write_and_wait(inode->i_mapping);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
if (ret < 0)
return ret;
+ mutex_lock(&inode->i_mutex);
dirty = try_flush_caps(inode, NULL, &flush_tid);
dout("fsync dirty caps are %s\n", ceph_cap_string(dirty));
@@ -1841,6 +1842,7 @@
}
dout("fsync %p%s done\n", inode, datasync ? " datasync" : "");
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 0972b45..1065ac7 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -1118,7 +1118,8 @@
* an fsync() on a dir will wait for any uncommitted directory
* operations to commit.
*/
-static int ceph_dir_fsync(struct file *file, int datasync)
+static int ceph_dir_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct ceph_inode_info *ci = ceph_inode(inode);
@@ -1128,6 +1129,11 @@
int ret = 0;
dout("dir_fsync %p\n", inode);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
spin_lock(&ci->i_unsafe_lock);
if (list_empty(head))
goto out;
@@ -1161,6 +1167,8 @@
} while (req->r_tid < last_tid);
out:
spin_unlock(&ci->i_unsafe_lock);
+ mutex_unlock(&inode->i_mutex);
+
return ret;
}
diff --git a/fs/ceph/super.h b/fs/ceph/super.h
index 56c41ef..30446b1 100644
--- a/fs/ceph/super.h
+++ b/fs/ceph/super.h
@@ -728,7 +728,8 @@
extern void ceph_queue_caps_release(struct inode *inode);
extern int ceph_write_inode(struct inode *inode, struct writeback_control *wbc);
-extern int ceph_fsync(struct file *file, int datasync);
+extern int ceph_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync);
extern void ceph_kick_flushing_caps(struct ceph_mds_client *mdsc,
struct ceph_mds_session *session);
extern struct ceph_cap *ceph_get_cap_for_mds(struct ceph_inode_info *ci,
diff --git a/fs/cifs/cifsfs.h b/fs/cifs/cifsfs.h
index 036ca83..fbd050c 100644
--- a/fs/cifs/cifsfs.h
+++ b/fs/cifs/cifsfs.h
@@ -91,8 +91,8 @@
extern ssize_t cifs_strict_writev(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos);
extern int cifs_lock(struct file *, int, struct file_lock *);
-extern int cifs_fsync(struct file *, int);
-extern int cifs_strict_fsync(struct file *, int);
+extern int cifs_fsync(struct file *, loff_t, loff_t, int);
+extern int cifs_strict_fsync(struct file *, loff_t, loff_t, int);
extern int cifs_flush(struct file *, fl_owner_t id);
extern int cifs_file_mmap(struct file * , struct vm_area_struct *);
extern int cifs_file_strict_mmap(struct file * , struct vm_area_struct *);
diff --git a/fs/cifs/file.c b/fs/cifs/file.c
index bb71471..cef58445 100644
--- a/fs/cifs/file.c
+++ b/fs/cifs/file.c
@@ -1401,7 +1401,8 @@
return rc;
}
-int cifs_strict_fsync(struct file *file, int datasync)
+int cifs_strict_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
int xid;
int rc = 0;
@@ -1410,6 +1411,11 @@
struct inode *inode = file->f_path.dentry->d_inode;
struct cifs_sb_info *cifs_sb = CIFS_SB(inode->i_sb);
+ rc = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (rc)
+ return rc;
+ mutex_lock(&inode->i_mutex);
+
xid = GetXid();
cFYI(1, "Sync file - name: %s datasync: 0x%x",
@@ -1428,16 +1434,23 @@
rc = CIFSSMBFlush(xid, tcon, smbfile->netfid);
FreeXid(xid);
+ mutex_unlock(&inode->i_mutex);
return rc;
}
-int cifs_fsync(struct file *file, int datasync)
+int cifs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
int xid;
int rc = 0;
struct cifs_tcon *tcon;
struct cifsFileInfo *smbfile = file->private_data;
struct cifs_sb_info *cifs_sb = CIFS_SB(file->f_path.dentry->d_sb);
+ struct inode *inode = file->f_mapping->host;
+
+ rc = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (rc)
+ return rc;
+ mutex_lock(&inode->i_mutex);
xid = GetXid();
@@ -1449,6 +1462,7 @@
rc = CIFSSMBFlush(xid, tcon, smbfile->netfid);
FreeXid(xid);
+ mutex_unlock(&inode->i_mutex);
return rc;
}
diff --git a/fs/coda/coda_int.h b/fs/coda/coda_int.h
index 6b443ff..b7143cf 100644
--- a/fs/coda/coda_int.h
+++ b/fs/coda/coda_int.h
@@ -11,7 +11,7 @@
void coda_destroy_inodecache(void);
int coda_init_inodecache(void);
-int coda_fsync(struct file *coda_file, int datasync);
+int coda_fsync(struct file *coda_file, loff_t start, loff_t end, int datasync);
void coda_sysctl_init(void);
void coda_sysctl_clean(void);
diff --git a/fs/coda/file.c b/fs/coda/file.c
index 0433057..8edd404 100644
--- a/fs/coda/file.c
+++ b/fs/coda/file.c
@@ -199,7 +199,7 @@
return 0;
}
-int coda_fsync(struct file *coda_file, int datasync)
+int coda_fsync(struct file *coda_file, loff_t start, loff_t end, int datasync)
{
struct file *host_file;
struct inode *coda_inode = coda_file->f_path.dentry->d_inode;
@@ -210,6 +210,11 @@
S_ISLNK(coda_inode->i_mode)))
return -EINVAL;
+ err = filemap_write_and_wait_range(coda_inode->i_mapping, start, end);
+ if (err)
+ return err;
+ mutex_lock(&coda_inode->i_mutex);
+
cfi = CODA_FTOC(coda_file);
BUG_ON(!cfi || cfi->cfi_magic != CODA_MAGIC);
host_file = cfi->cfi_container;
@@ -217,6 +222,7 @@
err = vfs_fsync(host_file, datasync);
if (!err && !datasync)
err = venus_fsync(coda_inode->i_sb, coda_i2f(coda_inode));
+ mutex_unlock(&coda_inode->i_mutex);
return err;
}
diff --git a/fs/ecryptfs/file.c b/fs/ecryptfs/file.c
index 4ec9eb00..c6ac98c 100644
--- a/fs/ecryptfs/file.c
+++ b/fs/ecryptfs/file.c
@@ -270,14 +270,15 @@
}
static int
-ecryptfs_fsync(struct file *file, int datasync)
+ecryptfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
int rc = 0;
- rc = generic_file_fsync(file, datasync);
+ rc = generic_file_fsync(file, start, end, datasync);
if (rc)
goto out;
- rc = vfs_fsync(ecryptfs_file_to_lower(file), datasync);
+ rc = vfs_fsync_range(ecryptfs_file_to_lower(file), start, end,
+ datasync);
out:
return rc;
}
diff --git a/fs/exofs/file.c b/fs/exofs/file.c
index 45ca323..491c6c0 100644
--- a/fs/exofs/file.c
+++ b/fs/exofs/file.c
@@ -42,11 +42,19 @@
* Note, in exofs all metadata is written as part of inode, regardless.
* The writeout is synchronous
*/
-static int exofs_file_fsync(struct file *filp, int datasync)
+static int exofs_file_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
+ struct inode *inode = filp->f_mapping->host;
int ret;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ mutex_lock(&inode->i_mutex);
ret = sync_inode_metadata(filp->f_mapping->host, 1);
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/ext2/ext2.h b/fs/ext2/ext2.h
index 645be9e..af9fc89 100644
--- a/fs/ext2/ext2.h
+++ b/fs/ext2/ext2.h
@@ -150,7 +150,8 @@
extern const struct file_operations ext2_dir_operations;
/* file.c */
-extern int ext2_fsync(struct file *file, int datasync);
+extern int ext2_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync);
extern const struct inode_operations ext2_file_inode_operations;
extern const struct file_operations ext2_file_operations;
extern const struct file_operations ext2_xip_file_operations;
diff --git a/fs/ext2/file.c b/fs/ext2/file.c
index 49eec94..82e0632 100644
--- a/fs/ext2/file.c
+++ b/fs/ext2/file.c
@@ -40,13 +40,13 @@
return 0;
}
-int ext2_fsync(struct file *file, int datasync)
+int ext2_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
int ret;
struct super_block *sb = file->f_mapping->host->i_sb;
struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
- ret = generic_file_fsync(file, datasync);
+ ret = generic_file_fsync(file, start, end, datasync);
if (ret == -EIO || test_and_clear_bit(AS_EIO, &mapping->flags)) {
/* We don't really know where the IO error happened... */
ext2_error(sb, __func__,
diff --git a/fs/ext3/fsync.c b/fs/ext3/fsync.c
index 09b13bb..0bcf63a 100644
--- a/fs/ext3/fsync.c
+++ b/fs/ext3/fsync.c
@@ -43,7 +43,7 @@
* inode to disk.
*/
-int ext3_sync_file(struct file *file, int datasync)
+int ext3_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
struct ext3_inode_info *ei = EXT3_I(inode);
@@ -54,6 +54,17 @@
if (inode->i_sb->s_flags & MS_RDONLY)
return 0;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ /*
+ * Taking the mutex here just to keep consistent with how fsync was
+ * called previously, however it looks like we don't need to take
+ * i_mutex at all.
+ */
+ mutex_lock(&inode->i_mutex);
+
J_ASSERT(ext3_journal_current_handle() == NULL);
/*
@@ -70,8 +81,10 @@
* (they were dirtied by commit). But that's OK - the blocks are
* safe in-journal, which is all fsync() needs to ensure.
*/
- if (ext3_should_journal_data(inode))
+ if (ext3_should_journal_data(inode)) {
+ mutex_unlock(&inode->i_mutex);
return ext3_force_commit(inode->i_sb);
+ }
if (datasync)
commit_tid = atomic_read(&ei->i_datasync_tid);
@@ -91,5 +104,6 @@
*/
if (needs_barrier)
blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 1921392c..fa44df8 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1758,7 +1758,7 @@
extern void ext4_htree_free_dir_info(struct dir_private_info *p);
/* fsync.c */
-extern int ext4_sync_file(struct file *, int);
+extern int ext4_sync_file(struct file *, loff_t, loff_t, int);
extern int ext4_flush_completed_IO(struct inode *);
/* hash.c */
diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c
index ce66d2f..da3bed3 100644
--- a/fs/ext4/fsync.c
+++ b/fs/ext4/fsync.c
@@ -151,6 +151,32 @@
return ret;
}
+/**
+ * __sync_file - generic_file_fsync without the locking and filemap_write
+ * @inode: inode to sync
+ * @datasync: only sync essential metadata if true
+ *
+ * This is just generic_file_fsync without the locking. This is needed for
+ * nojournal mode to make sure this inodes data/metadata makes it to disk
+ * properly. The i_mutex should be held already.
+ */
+static int __sync_inode(struct inode *inode, int datasync)
+{
+ int err;
+ int ret;
+
+ ret = sync_mapping_buffers(inode->i_mapping);
+ if (!(inode->i_state & I_DIRTY))
+ return ret;
+ if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
+ return ret;
+
+ err = sync_inode_metadata(inode, 1);
+ if (ret == 0)
+ ret = err;
+ return ret;
+}
+
/*
* akpm: A new design for ext4_sync_file().
*
@@ -165,7 +191,7 @@
* i_mutex lock is held when entering and exiting this function
*/
-int ext4_sync_file(struct file *file, int datasync)
+int ext4_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
struct ext4_inode_info *ei = EXT4_I(inode);
@@ -178,15 +204,20 @@
trace_ext4_sync_file_enter(file, datasync);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
if (inode->i_sb->s_flags & MS_RDONLY)
- return 0;
+ goto out;
ret = ext4_flush_completed_IO(inode);
if (ret < 0)
goto out;
if (!journal) {
- ret = generic_file_fsync(file, datasync);
+ ret = __sync_inode(inode, datasync);
if (!ret && !list_empty(&inode->i_dentry))
ret = ext4_sync_parent(inode);
goto out;
@@ -220,6 +251,7 @@
if (needs_barrier)
blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
out:
+ mutex_unlock(&inode->i_mutex);
trace_ext4_sync_file_exit(inode, ret);
return ret;
}
diff --git a/fs/fat/fat.h b/fs/fat/fat.h
index a975b41..a5d3853 100644
--- a/fs/fat/fat.h
+++ b/fs/fat/fat.h
@@ -310,7 +310,8 @@
extern void fat_truncate_blocks(struct inode *inode, loff_t offset);
extern int fat_getattr(struct vfsmount *mnt, struct dentry *dentry,
struct kstat *stat);
-extern int fat_file_fsync(struct file *file, int datasync);
+extern int fat_file_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync);
/* fat/inode.c */
extern void fat_attach(struct inode *inode, loff_t i_pos);
diff --git a/fs/fat/file.c b/fs/fat/file.c
index e1587c5..c118acf 100644
--- a/fs/fat/file.c
+++ b/fs/fat/file.c
@@ -149,12 +149,12 @@
return 0;
}
-int fat_file_fsync(struct file *filp, int datasync)
+int fat_file_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
struct inode *inode = filp->f_mapping->host;
int res, err;
- res = generic_file_fsync(filp, datasync);
+ res = generic_file_fsync(filp, start, end, datasync);
err = sync_mapping_buffers(MSDOS_SB(inode->i_sb)->fat_inode->i_mapping);
return res ? res : err;
diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index 02063dd..9f63e49 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -1176,9 +1176,10 @@
return 0;
}
-static int fuse_dir_fsync(struct file *file, int datasync)
+static int fuse_dir_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
- return fuse_fsync_common(file, datasync, 1);
+ return fuse_fsync_common(file, start, end, datasync, 1);
}
static bool update_mtime(unsigned ivalid)
diff --git a/fs/fuse/file.c b/fs/fuse/file.c
index 73b89df..7bb685c 100644
--- a/fs/fuse/file.c
+++ b/fs/fuse/file.c
@@ -400,7 +400,8 @@
fuse_release_nowrite(inode);
}
-int fuse_fsync_common(struct file *file, int datasync, int isdir)
+int fuse_fsync_common(struct file *file, loff_t start, loff_t end,
+ int datasync, int isdir)
{
struct inode *inode = file->f_mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
@@ -412,9 +413,15 @@
if (is_bad_inode(inode))
return -EIO;
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir))
return 0;
+ mutex_lock(&inode->i_mutex);
+
/*
* Start writeback against all dirty pages of the inode, then
* wait for all outstanding writes, before sending the FSYNC
@@ -422,13 +429,15 @@
*/
err = write_inode_now(inode, 0);
if (err)
- return err;
+ goto out;
fuse_sync_writes(inode);
req = fuse_get_req(fc);
- if (IS_ERR(req))
- return PTR_ERR(req);
+ if (IS_ERR(req)) {
+ err = PTR_ERR(req);
+ goto out;
+ }
memset(&inarg, 0, sizeof(inarg));
inarg.fh = ff->fh;
@@ -448,12 +457,15 @@
fc->no_fsync = 1;
err = 0;
}
+out:
+ mutex_unlock(&inode->i_mutex);
return err;
}
-static int fuse_fsync(struct file *file, int datasync)
+static int fuse_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
- return fuse_fsync_common(file, datasync, 0);
+ return fuse_fsync_common(file, start, end, datasync, 0);
}
void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index b788bec..c6aa2d4 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -589,7 +589,8 @@
/**
* Send FSYNC or FSYNCDIR request
*/
-int fuse_fsync_common(struct file *file, int datasync, int isdir);
+int fuse_fsync_common(struct file *file, loff_t start, loff_t end,
+ int datasync, int isdir);
/**
* Notify poll wakeup
diff --git a/fs/gfs2/file.c b/fs/gfs2/file.c
index 89c39e5..f82cb5e 100644
--- a/fs/gfs2/file.c
+++ b/fs/gfs2/file.c
@@ -544,7 +544,9 @@
/**
* gfs2_fsync - sync the dirty data for a file (across the cluster)
- * @file: the file that points to the dentry (we ignore this)
+ * @file: the file that points to the dentry
+ * @start: the start position in the file to sync
+ * @end: the end position in the file to sync
* @datasync: set if we can ignore timestamp changes
*
* The VFS will flush data for us. We only need to worry
@@ -553,23 +555,32 @@
* Returns: errno
*/
-static int gfs2_fsync(struct file *file, int datasync)
+static int gfs2_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = file->f_mapping->host;
int sync_state = inode->i_state & (I_DIRTY_SYNC|I_DIRTY_DATASYNC);
struct gfs2_inode *ip = GFS2_I(inode);
int ret;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
if (datasync)
sync_state &= ~I_DIRTY_SYNC;
if (sync_state) {
ret = sync_inode_metadata(inode, 1);
- if (ret)
+ if (ret) {
+ mutex_unlock(&inode->i_mutex);
return ret;
+ }
gfs2_ail_flush(ip->i_gl);
}
+ mutex_unlock(&inode->i_mutex);
return 0;
}
diff --git a/fs/hfs/inode.c b/fs/hfs/inode.c
index 5e7c3f3..96a1b62 100644
--- a/fs/hfs/inode.c
+++ b/fs/hfs/inode.c
@@ -627,12 +627,18 @@
return 0;
}
-static int hfs_file_fsync(struct file *filp, int datasync)
+static int hfs_file_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = filp->f_mapping->host;
struct super_block * sb;
int ret, err;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
/* sync the inode to buffers */
ret = write_inode_now(inode, 0);
@@ -649,6 +655,7 @@
err = sync_blockdev(sb->s_bdev);
if (!ret)
ret = err;
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/hfsplus/hfsplus_fs.h b/fs/hfsplus/hfsplus_fs.h
index d685752..38184e3 100644
--- a/fs/hfsplus/hfsplus_fs.h
+++ b/fs/hfsplus/hfsplus_fs.h
@@ -392,7 +392,8 @@
int hfsplus_cat_write_inode(struct inode *);
struct inode *hfsplus_new_inode(struct super_block *, int);
void hfsplus_delete_inode(struct inode *);
-int hfsplus_file_fsync(struct file *file, int datasync);
+int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync);
/* ioctl.c */
long hfsplus_ioctl(struct file *filp, unsigned int cmd, unsigned long arg);
diff --git a/fs/hfsplus/inode.c b/fs/hfsplus/inode.c
index 5b1cb98..30486e0 100644
--- a/fs/hfsplus/inode.c
+++ b/fs/hfsplus/inode.c
@@ -308,13 +308,19 @@
return 0;
}
-int hfsplus_file_fsync(struct file *file, int datasync)
+int hfsplus_file_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = file->f_mapping->host;
struct hfsplus_inode_info *hip = HFSPLUS_I(inode);
struct hfsplus_sb_info *sbi = HFSPLUS_SB(inode->i_sb);
int error = 0, error2;
+ error = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (error)
+ return error;
+ mutex_lock(&inode->i_mutex);
+
/*
* Sync inode metadata into the catalog and extent trees.
*/
@@ -342,6 +348,8 @@
if (!test_bit(HFSPLUS_SB_NOBARRIER, &sbi->flags))
blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
+ mutex_unlock(&inode->i_mutex);
+
return error;
}
diff --git a/fs/hostfs/hostfs_kern.c b/fs/hostfs/hostfs_kern.c
index 6e449c5..0d22afd 100644
--- a/fs/hostfs/hostfs_kern.c
+++ b/fs/hostfs/hostfs_kern.c
@@ -362,9 +362,20 @@
return 0;
}
-int hostfs_fsync(struct file *file, int datasync)
+int hostfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
- return fsync_file(HOSTFS_I(file->f_mapping->host)->fd, datasync);
+ struct inode *inode = file->f_mapping->host;
+ int ret;
+
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ mutex_lock(&inode->i_mutex);
+ ret = fsync_file(HOSTFS_I(inode)->fd, datasync);
+ mutex_unlock(&inode->i_mutex);
+
+ return ret;
}
static const struct file_operations hostfs_file_fops = {
diff --git a/fs/hpfs/file.c b/fs/hpfs/file.c
index 89c500e..89d2a58 100644
--- a/fs/hpfs/file.c
+++ b/fs/hpfs/file.c
@@ -18,9 +18,14 @@
return 0;
}
-int hpfs_file_fsync(struct file *file, int datasync)
+int hpfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
+ int ret;
+
+ ret = filemap_write_and_wait_range(file->f_mapping, start, end);
+ if (ret)
+ return ret;
return sync_blockdev(inode->i_sb->s_bdev);
}
diff --git a/fs/hpfs/hpfs_fn.h b/fs/hpfs/hpfs_fn.h
index dd552f8..331b5e2 100644
--- a/fs/hpfs/hpfs_fn.h
+++ b/fs/hpfs/hpfs_fn.h
@@ -258,7 +258,7 @@
/* file.c */
-int hpfs_file_fsync(struct file *, int);
+int hpfs_file_fsync(struct file *, loff_t, loff_t, int);
extern const struct file_operations hpfs_file_ops;
extern const struct inode_operations hpfs_file_iops;
extern const struct address_space_operations hpfs_aops;
diff --git a/fs/hppfs/hppfs.c b/fs/hppfs/hppfs.c
index 85c098a..8635be5 100644
--- a/fs/hppfs/hppfs.c
+++ b/fs/hppfs/hppfs.c
@@ -573,9 +573,10 @@
return err;
}
-static int hppfs_fsync(struct file *file, int datasync)
+static int hppfs_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
- return 0;
+ return filemap_write_and_wait_range(file->f_mapping, start, end);
}
static const struct file_operations hppfs_dir_fops = {
diff --git a/fs/jffs2/file.c b/fs/jffs2/file.c
index 1c0a08d..3989f7e 100644
--- a/fs/jffs2/file.c
+++ b/fs/jffs2/file.c
@@ -27,13 +27,20 @@
struct page **pagep, void **fsdata);
static int jffs2_readpage (struct file *filp, struct page *pg);
-int jffs2_fsync(struct file *filp, int datasync)
+int jffs2_fsync(struct file *filp, loff_t start, loff_t end, int datasync)
{
struct inode *inode = filp->f_mapping->host;
struct jffs2_sb_info *c = JFFS2_SB_INFO(inode->i_sb);
+ int ret;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ mutex_lock(&inode->i_mutex);
/* Trigger GC to flush any pending writes for this inode */
jffs2_flush_wbuf_gc(c, inode->i_ino);
+ mutex_unlock(&inode->i_mutex);
return 0;
}
diff --git a/fs/jffs2/os-linux.h b/fs/jffs2/os-linux.h
index 65c6c43..9c25283 100644
--- a/fs/jffs2/os-linux.h
+++ b/fs/jffs2/os-linux.h
@@ -158,7 +158,7 @@
extern const struct file_operations jffs2_file_operations;
extern const struct inode_operations jffs2_file_inode_operations;
extern const struct address_space_operations jffs2_file_address_operations;
-int jffs2_fsync(struct file *, int);
+int jffs2_fsync(struct file *, loff_t, loff_t, int);
int jffs2_do_readpage_unlock (struct inode *inode, struct page *pg);
/* ioctl.c */
diff --git a/fs/jfs/file.c b/fs/jfs/file.c
index 9f32315..7527855 100644
--- a/fs/jfs/file.c
+++ b/fs/jfs/file.c
@@ -28,19 +28,26 @@
#include "jfs_acl.h"
#include "jfs_debug.h"
-int jfs_fsync(struct file *file, int datasync)
+int jfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
int rc = 0;
+ rc = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (rc)
+ return rc;
+
+ mutex_lock(&inode->i_mutex);
if (!(inode->i_state & I_DIRTY) ||
(datasync && !(inode->i_state & I_DIRTY_DATASYNC))) {
/* Make sure committed changes hit the disk */
jfs_flush_journal(JFS_SBI(inode->i_sb)->log, 1);
+ mutex_unlock(&inode->i_mutex);
return rc;
}
rc |= jfs_commit_inode(inode, 1);
+ mutex_unlock(&inode->i_mutex);
return rc ? -EIO : 0;
}
diff --git a/fs/jfs/jfs_inode.h b/fs/jfs/jfs_inode.h
index ec2fb8b..9271cfe4 100644
--- a/fs/jfs/jfs_inode.h
+++ b/fs/jfs/jfs_inode.h
@@ -21,7 +21,7 @@
struct fid;
extern struct inode *ialloc(struct inode *, umode_t);
-extern int jfs_fsync(struct file *, int);
+extern int jfs_fsync(struct file *, loff_t, loff_t, int);
extern long jfs_ioctl(struct file *, unsigned int, unsigned long);
extern long jfs_compat_ioctl(struct file *, unsigned int, unsigned long);
extern struct inode *jfs_iget(struct super_block *, unsigned long);
diff --git a/fs/libfs.c b/fs/libfs.c
index bd50b11..8f2271a 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -905,21 +905,29 @@
* filesystems which track all non-inode metadata in the buffers list
* hanging off the address_space structure.
*/
-int generic_file_fsync(struct file *file, int datasync)
+int generic_file_fsync(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = file->f_mapping->host;
int err;
int ret;
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
+ mutex_lock(&inode->i_mutex);
ret = sync_mapping_buffers(inode->i_mapping);
if (!(inode->i_state & I_DIRTY))
- return ret;
+ goto out;
if (datasync && !(inode->i_state & I_DIRTY_DATASYNC))
- return ret;
+ goto out;
err = sync_inode_metadata(inode, 1);
if (ret == 0)
ret = err;
+out:
+ mutex_unlock(&inode->i_mutex);
return ret;
}
EXPORT_SYMBOL(generic_file_fsync);
@@ -956,7 +964,7 @@
/*
* No-op implementation of ->fsync for in-memory filesystems.
*/
-int noop_fsync(struct file *file, int datasync)
+int noop_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
return 0;
}
diff --git a/fs/logfs/file.c b/fs/logfs/file.c
index c2ad702..b548c87 100644
--- a/fs/logfs/file.c
+++ b/fs/logfs/file.c
@@ -219,11 +219,20 @@
}
}
-int logfs_fsync(struct file *file, int datasync)
+int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct super_block *sb = file->f_mapping->host->i_sb;
+ struct inode *inode = file->f_mapping->host;
+ int ret;
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+
+ mutex_lock(&inode->i_mutex);
logfs_write_anchor(sb);
+ mutex_unlock(&inode->i_mutex);
+
return 0;
}
diff --git a/fs/logfs/logfs.h b/fs/logfs/logfs.h
index 57afd4a..f22d108 100644
--- a/fs/logfs/logfs.h
+++ b/fs/logfs/logfs.h
@@ -506,7 +506,7 @@
extern const struct address_space_operations logfs_reg_aops;
int logfs_readpage(struct file *file, struct page *page);
long logfs_ioctl(struct file *file, unsigned int cmd, unsigned long arg);
-int logfs_fsync(struct file *file, int datasync);
+int logfs_fsync(struct file *file, loff_t start, loff_t end, int datasync);
/* gc.c */
u32 get_best_cand(struct super_block *sb, struct candidate_list *list, u32 *ec);
diff --git a/fs/ncpfs/file.c b/fs/ncpfs/file.c
index 0ed65e0..64a3264 100644
--- a/fs/ncpfs/file.c
+++ b/fs/ncpfs/file.c
@@ -20,9 +20,9 @@
#include "ncp_fs.h"
-static int ncp_fsync(struct file *file, int datasync)
+static int ncp_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
- return 0;
+ return filemap_write_and_wait_range(file->f_mapping, start, end);
}
/*
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 8a45e6d..57f578e 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -56,7 +56,7 @@
static int nfs_mknod(struct inode *, struct dentry *, int, dev_t);
static int nfs_rename(struct inode *, struct dentry *,
struct inode *, struct dentry *);
-static int nfs_fsync_dir(struct file *, int);
+static int nfs_fsync_dir(struct file *, loff_t, loff_t, int);
static loff_t nfs_llseek_dir(struct file *, loff_t, int);
static void nfs_readdir_clear_array(struct page*);
@@ -945,15 +945,19 @@
* All directory operations under NFS are synchronous, so fsync()
* is a dummy operation.
*/
-static int nfs_fsync_dir(struct file *filp, int datasync)
+static int nfs_fsync_dir(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct dentry *dentry = filp->f_path.dentry;
+ struct inode *inode = dentry->d_inode;
dfprintk(FILE, "NFS: fsync dir(%s/%s) datasync %d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
datasync);
+ mutex_lock(&inode->i_mutex);
nfs_inc_stats(dentry->d_inode, NFSIOS_VFSFSYNC);
+ mutex_unlock(&inode->i_mutex);
return 0;
}
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 2c1705b..28b8c3f 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -55,7 +55,7 @@
static ssize_t nfs_file_write(struct kiocb *, const struct iovec *iov,
unsigned long nr_segs, loff_t pos);
static int nfs_file_flush(struct file *, fl_owner_t id);
-static int nfs_file_fsync(struct file *, int datasync);
+static int nfs_file_fsync(struct file *, loff_t, loff_t, int datasync);
static int nfs_check_flags(int flags);
static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl);
static int nfs_flock(struct file *filp, int cmd, struct file_lock *fl);
@@ -308,7 +308,7 @@
* fall back to doing a synchronous write.
*/
static int
-nfs_file_fsync(struct file *file, int datasync)
+nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct dentry *dentry = file->f_path.dentry;
struct nfs_open_context *ctx = nfs_file_open_context(file);
@@ -316,11 +316,15 @@
int have_error, status;
int ret = 0;
-
dprintk("NFS: fsync file(%s/%s) datasync %d\n",
dentry->d_parent->d_name.name, dentry->d_name.name,
datasync);
+ ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (ret)
+ return ret;
+ mutex_lock(&inode->i_mutex);
+
nfs_inc_stats(inode, NFSIOS_VFSFSYNC);
have_error = test_and_clear_bit(NFS_CONTEXT_ERROR_WRITE, &ctx->flags);
status = nfs_commit_inode(inode, FLUSH_SYNC);
@@ -332,6 +336,7 @@
if (!ret && !datasync)
/* application has asked for meta-data sync */
ret = pnfs_layoutcommit_inode(inode, true);
+ mutex_unlock(&inode->i_mutex);
return ret;
}
diff --git a/fs/nilfs2/file.c b/fs/nilfs2/file.c
index d7eeca6..2660152 100644
--- a/fs/nilfs2/file.c
+++ b/fs/nilfs2/file.c
@@ -27,7 +27,7 @@
#include "nilfs.h"
#include "segment.h"
-int nilfs_sync_file(struct file *file, int datasync)
+int nilfs_sync_file(struct file *file, loff_t start, loff_t end, int datasync)
{
/*
* Called from fsync() system call
@@ -40,8 +40,15 @@
struct inode *inode = file->f_mapping->host;
int err;
- if (!nilfs_inode_dirty(inode))
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+ mutex_lock(&inode->i_mutex);
+
+ if (!nilfs_inode_dirty(inode)) {
+ mutex_unlock(&inode->i_mutex);
return 0;
+ }
if (datasync)
err = nilfs_construct_dsync_segment(inode->i_sb, inode, 0,
@@ -49,6 +56,7 @@
else
err = nilfs_construct_segment(inode->i_sb);
+ mutex_unlock(&inode->i_mutex);
return err;
}
diff --git a/fs/nilfs2/nilfs.h b/fs/nilfs2/nilfs.h
index 6fb7511..255d5e1 100644
--- a/fs/nilfs2/nilfs.h
+++ b/fs/nilfs2/nilfs.h
@@ -235,7 +235,7 @@
struct page *, struct inode *);
/* file.c */
-extern int nilfs_sync_file(struct file *, int);
+extern int nilfs_sync_file(struct file *, loff_t, loff_t, int);
/* ioctl.c */
long nilfs_ioctl(struct file *, unsigned int, unsigned long);
diff --git a/fs/ntfs/dir.c b/fs/ntfs/dir.c
index 0f48e7c..99e3610 100644
--- a/fs/ntfs/dir.c
+++ b/fs/ntfs/dir.c
@@ -1527,13 +1527,20 @@
* this problem for now. We do write the $BITMAP attribute if it is present
* which is the important one for a directory so things are not too bad.
*/
-static int ntfs_dir_fsync(struct file *filp, int datasync)
+static int ntfs_dir_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct inode *bmp_vi, *vi = filp->f_mapping->host;
int err, ret;
ntfs_attr na;
ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
+
+ err = filemap_write_and_wait_range(vi->i_mapping, start, end);
+ if (err)
+ return err;
+ mutex_lock(&vi->i_mutex);
+
BUG_ON(!S_ISDIR(vi->i_mode));
/* If the bitmap attribute inode is in memory sync it, too. */
na.mft_no = vi->i_ino;
@@ -1555,6 +1562,7 @@
else
ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx. Error "
"%u.", datasync ? "data" : "", vi->i_ino, -ret);
+ mutex_unlock(&vi->i_mutex);
return ret;
}
diff --git a/fs/ntfs/file.c b/fs/ntfs/file.c
index b59f5ac..c587e2d 100644
--- a/fs/ntfs/file.c
+++ b/fs/ntfs/file.c
@@ -2152,12 +2152,19 @@
* with this inode but since we have no simple way of getting to them we ignore
* this problem for now.
*/
-static int ntfs_file_fsync(struct file *filp, int datasync)
+static int ntfs_file_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct inode *vi = filp->f_mapping->host;
int err, ret = 0;
ntfs_debug("Entering for inode 0x%lx.", vi->i_ino);
+
+ err = filemap_write_and_wait_range(vi->i_mapping, start, end);
+ if (err)
+ return err;
+ mutex_lock(&vi->i_mutex);
+
BUG_ON(S_ISDIR(vi->i_mode));
if (!datasync || !NInoNonResident(NTFS_I(vi)))
ret = __ntfs_write_inode(vi, 1);
@@ -2175,6 +2182,7 @@
else
ntfs_warning(vi->i_sb, "Failed to f%ssync inode 0x%lx. Error "
"%u.", datasync ? "data" : "", vi->i_ino, -ret);
+ mutex_unlock(&vi->i_mutex);
return ret;
}
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 22d6046..0fc2bd3 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -171,7 +171,8 @@
return 0;
}
-static int ocfs2_sync_file(struct file *file, int datasync)
+static int ocfs2_sync_file(struct file *file, loff_t start, loff_t end,
+ int datasync)
{
int err = 0;
journal_t *journal;
@@ -184,6 +185,16 @@
file->f_path.dentry->d_name.name,
(unsigned long long)datasync);
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
+ /*
+ * Probably don't need the i_mutex at all in here, just putting it here
+ * to be consistent with how fsync used to be called, someone more
+ * familiar with the fs could possibly remove it.
+ */
+ mutex_lock(&inode->i_mutex);
if (datasync && !(inode->i_state & I_DIRTY_DATASYNC)) {
/*
* We still have to flush drive's caches to get data to the
@@ -200,6 +211,7 @@
bail:
if (err)
mlog_errno(err);
+ mutex_unlock(&inode->i_mutex);
return (err < 0) ? -EIO : 0;
}
diff --git a/fs/reiserfs/dir.c b/fs/reiserfs/dir.c
index 198dabf..133e935 100644
--- a/fs/reiserfs/dir.c
+++ b/fs/reiserfs/dir.c
@@ -14,7 +14,8 @@
extern const struct reiserfs_key MIN_KEY;
static int reiserfs_readdir(struct file *, void *, filldir_t);
-static int reiserfs_dir_fsync(struct file *filp, int datasync);
+static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync);
const struct file_operations reiserfs_dir_operations = {
.llseek = generic_file_llseek,
@@ -27,13 +28,21 @@
#endif
};
-static int reiserfs_dir_fsync(struct file *filp, int datasync)
+static int reiserfs_dir_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = filp->f_mapping->host;
int err;
+
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
+ mutex_lock(&inode->i_mutex);
reiserfs_write_lock(inode->i_sb);
err = reiserfs_commit_for_inode(inode);
reiserfs_write_unlock(inode->i_sb);
+ mutex_unlock(&inode->i_mutex);
if (err < 0)
return err;
return 0;
diff --git a/fs/reiserfs/file.c b/fs/reiserfs/file.c
index bbf3100..c7156dc 100644
--- a/fs/reiserfs/file.c
+++ b/fs/reiserfs/file.c
@@ -140,12 +140,18 @@
* be removed...
*/
-static int reiserfs_sync_file(struct file *filp, int datasync)
+static int reiserfs_sync_file(struct file *filp, loff_t start, loff_t end,
+ int datasync)
{
struct inode *inode = filp->f_mapping->host;
int err;
int barrier_done;
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+
+ mutex_lock(&inode->i_mutex);
BUG_ON(!S_ISREG(inode->i_mode));
err = sync_mapping_buffers(inode->i_mapping);
reiserfs_write_lock(inode->i_sb);
@@ -153,6 +159,7 @@
reiserfs_write_unlock(inode->i_sb);
if (barrier_done != 1 && reiserfs_barrier_flush(inode->i_sb))
blkdev_issue_flush(inode->i_sb->s_bdev, GFP_KERNEL, NULL);
+ mutex_unlock(&inode->i_mutex);
if (barrier_done < 0)
return barrier_done;
return (err < 0) ? -EIO : 0;
diff --git a/fs/sync.c b/fs/sync.c
index c38ec16..c98a747 100644
--- a/fs/sync.c
+++ b/fs/sync.c
@@ -165,28 +165,9 @@
*/
int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
{
- struct address_space *mapping = file->f_mapping;
- int err, ret;
-
- if (!file->f_op || !file->f_op->fsync) {
- ret = -EINVAL;
- goto out;
- }
-
- ret = filemap_write_and_wait_range(mapping, start, end);
-
- /*
- * We need to protect against concurrent writers, which could cause
- * livelocks in fsync_buffers_list().
- */
- mutex_lock(&mapping->host->i_mutex);
- err = file->f_op->fsync(file, datasync);
- if (!ret)
- ret = err;
- mutex_unlock(&mapping->host->i_mutex);
-
-out:
- return ret;
+ if (!file->f_op || !file->f_op->fsync)
+ return -EINVAL;
+ return file->f_op->fsync(file, start, end, datasync);
}
EXPORT_SYMBOL(vfs_fsync_range);
diff --git a/fs/ubifs/file.c b/fs/ubifs/file.c
index 5e7fccf..89ef9a2 100644
--- a/fs/ubifs/file.c
+++ b/fs/ubifs/file.c
@@ -1304,7 +1304,7 @@
return NULL;
}
-int ubifs_fsync(struct file *file, int datasync)
+int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct inode *inode = file->f_mapping->host;
struct ubifs_info *c = inode->i_sb->s_fs_info;
@@ -1319,14 +1319,16 @@
*/
return 0;
- /*
- * VFS has already synchronized dirty pages for this inode. Synchronize
- * the inode unless this is a 'datasync()' call.
- */
+ err = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (err)
+ return err;
+ mutex_lock(&inode->i_mutex);
+
+ /* Synchronize the inode unless this is a 'datasync()' call. */
if (!datasync || (inode->i_state & I_DIRTY_DATASYNC)) {
err = inode->i_sb->s_op->write_inode(inode, NULL);
if (err)
- return err;
+ goto out;
}
/*
@@ -1334,10 +1336,9 @@
* them.
*/
err = ubifs_sync_wbufs_by_inode(c, inode);
- if (err)
- return err;
-
- return 0;
+out:
+ mutex_unlock(&inode->i_mutex);
+ return err;
}
/**
diff --git a/fs/ubifs/ubifs.h b/fs/ubifs/ubifs.h
index f79983d..4cd6485 100644
--- a/fs/ubifs/ubifs.h
+++ b/fs/ubifs/ubifs.h
@@ -1720,7 +1720,7 @@
int ubifs_calc_dark(const struct ubifs_info *c, int spc);
/* file.c */
-int ubifs_fsync(struct file *file, int datasync);
+int ubifs_fsync(struct file *file, loff_t start, loff_t end, int datasync);
int ubifs_setattr(struct dentry *dentry, struct iattr *attr);
/* dir.c */
diff --git a/fs/xfs/linux-2.6/xfs_file.c b/fs/xfs/linux-2.6/xfs_file.c
index 7f782af..fbbf657 100644
--- a/fs/xfs/linux-2.6/xfs_file.c
+++ b/fs/xfs/linux-2.6/xfs_file.c
@@ -127,6 +127,8 @@
STATIC int
xfs_file_fsync(
struct file *file,
+ loff_t start,
+ loff_t end,
int datasync)
{
struct inode *inode = file->f_mapping->host;
@@ -138,6 +140,10 @@
trace_xfs_file_fsync(ip);
+ error = filemap_write_and_wait_range(inode->i_mapping, start, end);
+ if (error)
+ return error;
+
if (XFS_FORCED_SHUTDOWN(mp))
return -XFS_ERROR(EIO);
@@ -875,18 +881,11 @@
/* Handle various SYNC-type writes */
if ((file->f_flags & O_DSYNC) || IS_SYNC(inode)) {
loff_t end = pos + ret - 1;
- int error, error2;
xfs_rw_iunlock(ip, iolock);
- error = filemap_write_and_wait_range(mapping, pos, end);
+ ret = -xfs_file_fsync(file, pos, end,
+ (file->f_flags & __O_SYNC) ? 0 : 1);
xfs_rw_ilock(ip, iolock);
-
- error2 = -xfs_file_fsync(file,
- (file->f_flags & __O_SYNC) ? 0 : 1);
- if (error)
- ret = error;
- else if (error2)
- ret = error2;
}
out_unlock:
diff --git a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h
index 5e06acf..0c473fd 100644
--- a/include/linux/ext3_fs.h
+++ b/include/linux/ext3_fs.h
@@ -877,7 +877,7 @@
extern void ext3_htree_free_dir_info(struct dir_private_info *p);
/* fsync.c */
-extern int ext3_sync_file(struct file *, int);
+extern int ext3_sync_file(struct file *, loff_t, loff_t, int);
/* hash.c */
extern int ext3fs_dirhash(const char *name, int len, struct
diff --git a/include/linux/fb.h b/include/linux/fb.h
index 6a82748..1d6836c 100644
--- a/include/linux/fb.h
+++ b/include/linux/fb.h
@@ -1043,7 +1043,8 @@
struct inode *inode,
struct file *file);
extern void fb_deferred_io_cleanup(struct fb_info *info);
-extern int fb_deferred_io_fsync(struct file *file, int datasync);
+extern int fb_deferred_io_fsync(struct file *file, loff_t start,
+ loff_t end, int datasync);
static inline bool fb_be_math(struct fb_info *info)
{
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 4a61f98..9cd2075 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1572,7 +1572,7 @@
int (*open) (struct inode *, struct file *);
int (*flush) (struct file *, fl_owner_t id);
int (*release) (struct inode *, struct file *);
- int (*fsync) (struct file *, int datasync);
+ int (*fsync) (struct file *, loff_t, loff_t, int datasync);
int (*aio_fsync) (struct kiocb *, int datasync);
int (*fasync) (int, struct file *, int);
int (*lock) (struct file *, int, struct file_lock *);
@@ -2360,7 +2360,8 @@
/* fs/block_dev.c */
extern ssize_t blkdev_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos);
-extern int blkdev_fsync(struct file *filp, int datasync);
+extern int blkdev_fsync(struct file *filp, loff_t start, loff_t end,
+ int datasync);
/* fs/splice.c */
extern ssize_t generic_file_splice_read(struct file *, loff_t *,
@@ -2490,7 +2491,7 @@
extern int simple_unlink(struct inode *, struct dentry *);
extern int simple_rmdir(struct inode *, struct dentry *);
extern int simple_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
-extern int noop_fsync(struct file *, int);
+extern int noop_fsync(struct file *, loff_t, loff_t, int);
extern int simple_empty(struct dentry *);
extern int simple_readpage(struct file *file, struct page *page);
extern int simple_write_begin(struct file *file, struct address_space *mapping,
@@ -2515,7 +2516,7 @@
extern ssize_t simple_write_to_buffer(void *to, size_t available, loff_t *ppos,
const void __user *from, size_t count);
-extern int generic_file_fsync(struct file *, int);
+extern int generic_file_fsync(struct file *, loff_t, loff_t, int);
extern int generic_check_addressable(unsigned, u64);
diff --git a/ipc/shm.c b/ipc/shm.c
index ab3385a..27884ad 100644
--- a/ipc/shm.c
+++ b/ipc/shm.c
@@ -277,13 +277,13 @@
return 0;
}
-static int shm_fsync(struct file *file, int datasync)
+static int shm_fsync(struct file *file, loff_t start, loff_t end, int datasync)
{
struct shm_file_data *sfd = shm_file_data(file);
if (!sfd->file->f_op->fsync)
return -EINVAL;
- return sfd->file->f_op->fsync(sfd->file, datasync);
+ return sfd->file->f_op->fsync(sfd->file, start, end, datasync);
}
static unsigned long shm_get_unmapped_area(struct file *file,
