We recently fixed the writeback of pages prior to performing
direct i/o, however the initial fix was perhaps a bit heavy
handed. There is no need to invalidate pages if the direct i/o
is only a read, since they will be identical to what has been
flushed to disk anyway.
Signed-off-by: Steven Whitehouse <swhit...@redhat.com>
diff --git a/fs/gfs2/aops.c b/fs/gfs2/aops.c
index cf858da..49436fa 100644
--- a/fs/gfs2/aops.c
+++ b/fs/gfs2/aops.c
@@ -1032,8 +1032,9 @@ static ssize_t gfs2_direct_IO(int rw, struct kiocb *iocb,
unmap_shared_mapping_range(ip->i_inode.i_mapping,
offset, len);
rv = filemap_write_and_wait_range(mapping, lstart, end);
if (rv)
- return rv;
- truncate_inode_pages_range(mapping, lstart, end);
+ goto out;
+ if (rw == WRITE)
+ truncate_inode_pages_range(mapping, lstart, end);
}
rv = __blockdev_direct_IO(rw, iocb, inode, inode->i_sb->s_bdev, iov,
--
1.8.3.1
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
