[Ocfs2-commits] mfasheh commits r2329 - trunk/fs/ocfs2

svn-commits at oss.oracle.com svn-commits at oss.oracle.com
Thu May 26 17:31:01 CDT 2005


Author: mfasheh
Signed-off-by: zab
Date: 2005-05-26 17:30:59 -0500 (Thu, 26 May 2005)
New Revision: 2329

Modified:
   trunk/fs/ocfs2/inode.c
Log:
* We never unlocked the inode at the end of delete inode which was fine
  because we would be dropping the locks in clear inode anyway. Now however,
  it trips an assert in ocfs2_lock_res_free so fix that.

Signed-off-by: zab



Modified: trunk/fs/ocfs2/inode.c
===================================================================
--- trunk/fs/ocfs2/inode.c	2005-05-26 22:22:52 UTC (rev 2328)
+++ trunk/fs/ocfs2/inode.c	2005-05-26 22:30:59 UTC (rev 2329)
@@ -445,6 +445,7 @@
 void ocfs2_delete_inode(struct inode *inode)
 {
 	int status = 0;
+	int unlock = 0;
 	int orphaned_slot;
 	struct inode *orphan_dir_inode = NULL;
 	struct inode *inode_alloc_inode = NULL;
@@ -509,6 +510,7 @@
 		mlog_errno(status);
 		goto bail_unblock;
 	}
+	unlock = 1;
 
 	/* While we were waiting for the lock, another node might have
 	 * asked to delete the inode. Recheck our flags to catch this
@@ -671,6 +673,8 @@
 bail:
 	if (handle)
 		ocfs_commit_trans(handle);
+	if (unlock)
+		ocfs2_meta_unlock(inode, 1);
 	if (orphan_dir_bh)
 		brelse(orphan_dir_bh);
 	if (inode_alloc_bh)



More information about the Ocfs2-commits mailing list