commit 9da224f4cba000a2470c926d4c27bc20a45c3952
Author: Yiannis Pericleous <[EMAIL PROTECTED]>
Date: Mon May 28 11:50:18 2007 -0400
comment on why we cannot create entries in the odf (assign inode nums)
when creating the odf cached dir
diff --git a/fs/unionfs/dirhelper.c b/fs/unionfs/dirhelper.c
index d4a89fa..83b006d 100644
--- a/fs/unionfs/dirhelper.c
+++ b/fs/unionfs/dirhelper.c
@@ -183,6 +183,13 @@ find:
if (err || buf->mode != RD_CACHE_ODF || !buf->filp || whiteout)
goto out;
+ /* Ideally we would want all entries we find here to be in the
+ * odf and have a persisten inode. However it is possible for the
+ * entries not to be in the odf yet, and thus not have a valid odf
+ * inode number. We could have created an entry in the odf at this
+ * point, but since we do not have a unionfs dentry here we wouldn't
+ * be able to check for hardlinks
+ */
if (odf_dentry)
ino = odf_dentry->d_inode->i_ino;
err = odf_write_dirent(buf->filp, name, namelen, ino, d_type);
_______________________________________________
unionfs-cvs mailing list: http://unionfs.filesystems.org/
[email protected]
http://www.fsl.cs.sunysb.edu/mailman/listinfo/unionfs-cvs
