[Ocfs2-commits] smushran commits r2973 - branches/ocfs2-1.2/fs/ocfs2/dlm

[svn-commits at oss.oracle.com]

Author: smushran
Date: 2007-01-29 14:26:24 -0800 (Mon, 29 Jan 2007)
New Revision: 2973

Modified:
   branches/ocfs2-1.2/fs/ocfs2/dlm/dlmdomain.c
Log:
ocfs2_dlm: Add timeout to dlm join domain

Currently dlm has no timeout during dlm join domain. While this is
not a problem in normal operation, this does become an issue if, say,
the other node is refusing to let the node join the domain because
of a stuck recovery. This patch adds a 90 sec timeout.

Signed-off-by: mfasheh

Modified: branches/ocfs2-1.2/fs/ocfs2/dlm/dlmdomain.c
===================================================================
--- branches/ocfs2-1.2/fs/ocfs2/dlm/dlmdomain.c	2007-01-29 22:24:16 UTC (rev 2972)
+++ branches/ocfs2-1.2/fs/ocfs2/dlm/dlmdomain.c	2007-01-29 22:26:24 UTC (rev 2973)
@@ -1246,6 +1246,8 @@
 static int dlm_join_domain(struct dlm_ctxt *dlm)
 {
 	int status;
+	unsigned int backoff;
+	unsigned int tot_backoff = 0;
 
 	BUG_ON(!dlm);
 
@@ -1277,18 +1279,27 @@
 	}
 
 	do {
-		unsigned int backoff;
 		status = dlm_try_to_join_domain(dlm);
 
 		/* If we're racing another node to the join, then we
 		 * need to back off temporarily and let them
 		 * complete. */
+#define	DLM_JOIN_TIMEOUT_MSECS	90000
 		if (status == -EAGAIN) {
 			if (signal_pending(current)) {
 				status = -ERESTARTSYS;
 				goto bail;
 			}
 
+			if (tot_backoff >
+			    msecs_to_jiffies(DLM_JOIN_TIMEOUT_MSECS)) {
+				status = -ERESTARTSYS;
+				mlog(ML_NOTICE, "Timed out joining dlm domain "
+				     "%s after %u msecs\n", dlm->name,
+				     jiffies_to_msecs(tot_backoff));
+				goto bail;
+			}
+
 			/*
 			 * <chip> After you!
 			 * <dale> No, after you!
@@ -1298,6 +1309,7 @@
 			 */
 			backoff = (unsigned int)(jiffies & 0x3);
 			backoff *= DLM_DOMAIN_BACKOFF_MS;
+			tot_backoff += backoff;
 			mlog(0, "backoff %d\n", backoff);
 			msleep(backoff);
 		}