We need to wait for all pending direct I/O requests before taking care of
metadata in fsync and write_inode.
Note that the write_inode hunk will causes conflicts when merging with
mainline, I'm not sure how to best deal with that.
Signed-off-by: Christoph Hellwig <hch@xxxxxx>
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c 2010-03-06 14:10:12.247004588
+0100
+++ xfs/fs/xfs/linux-2.6/xfs_file.c 2010-03-06 14:16:48.460003752 +0100
@@ -115,6 +115,8 @@ xfs_file_fsync(
xfs_iflags_clear(ip, XFS_ITRUNCATED);
+ xfs_ioend_wait(ip);
+
/*
* We always need to make sure that the required inode state is safe on
* disk. The inode might be clean but we still might need to force the
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c 2010-03-06 14:10:12.259026379
+0100
+++ xfs/fs/xfs/linux-2.6/xfs_super.c 2010-03-06 14:16:48.461003193 +0100
@@ -1079,6 +1079,8 @@ xfs_fs_write_inode(
if (error)
goto out;
+ xfs_ioend_wait(ip);
+
/*
* Make sure the inode has hit stable storage. By using the
* log and the fsync transactions we reduce the IOs we have
|