ocfs2: Provide the ocfs2_dlm_lvb_valid() stack API.

The Lock Value Block (LVB) of a DLM lock can be lost when nodes die and
the DLM cannot reconstruct its state.  Clients of the DLM need to know
this.

ocfs2's internal DLM, o2dlm, explicitly zeroes out the LVB when it loses
track of the state.  This is not a standard behavior, but ocfs2 has
always relied on it.  Thus, an o2dlm LVB is always "valid".

ocfs2 now supports both o2dlm and fs/dlm via the stack glue.  When
fs/dlm loses track of an LVBs state, it sets a flag
(DLM_SBF_VALNOTVALID) on the Lock Status Block (LKSB).  The contents of
the LVB may be garbage or merely stale.

ocfs2 doesn't want to try to guess at the validity of the stale LVB.
Instead, it should be checking the VALNOTVALID flag.  As this is the
'standard' way of treating LVBs, we will promote this behavior.

We add a stack glue API ocfs2_dlm_lvb_valid().  It returns non-zero when
the LVB is valid.  o2dlm will always return valid, while fs/dlm will
check VALNOTVALID.

Signed-off-by: Joel Becker <joel.becker@oracle.com>
Acked-by: Mark Fasheh <mfasheh@suse.com>

5 files changed, 38 insertions, 9 deletions

diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 6cdeaa76f27..83d2ddb2718 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -1989,7 +1989,8 @@ static inline int ocfs2_meta_lvb_is_trustable(struct inode *inode,
 {
 	struct ocfs2_meta_lvb *lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
 
-	if (lvb->lvb_version == OCFS2_LVB_VERSION
+	if (ocfs2_dlm_lvb_valid(&lockres->l_lksb)
+	    && lvb->lvb_version == OCFS2_LVB_VERSION
 	    && be32_to_cpu(lvb->lvb_igeneration) == inode->i_generation)
 		return 1;
 	return 0;
@@ -2382,7 +2383,8 @@ int ocfs2_orphan_scan_lock(struct ocfs2_super *osb, u32 *seqno, int ex)
 		return status;
 
 	lvb = ocfs2_dlm_lvb(&lockres->l_lksb);
-	if (lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
+	if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
+	    lvb->lvb_version == OCFS2_ORPHAN_LVB_VERSION)
 		*seqno = be32_to_cpu(lvb->lvb_os_seqno);
 	return status;
 }
@@ -3627,7 +3629,8 @@ static int ocfs2_refresh_qinfo(struct ocfs2_mem_dqinfo *oinfo)
 	struct ocfs2_global_disk_dqinfo *gdinfo;
 	int status = 0;
 
-	if (lvb->lvb_version == OCFS2_QINFO_LVB_VERSION) {
+	if (ocfs2_dlm_lvb_valid(&lockres->l_lksb) &&
+	    lvb->lvb_version == OCFS2_QINFO_LVB_VERSION) {
 		info->dqi_bgrace = be32_to_cpu(lvb->lvb_bgrace);
 		info->dqi_igrace = be32_to_cpu(lvb->lvb_igrace);
 		oinfo->dqi_syncms = be32_to_cpu(lvb->lvb_syncms);
diff --git a/fs/ocfs2/stack_o2cb.c b/fs/ocfs2/stack_o2cb.c
index fcd120f1493..3f661376a2d 100644
--- a/fs/ocfs2/stack_o2cb.c
+++ b/fs/ocfs2/stack_o2cb.c
@@ -236,6 +236,16 @@ static int o2cb_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
 	return dlm_status_to_errno(lksb->lksb_o2dlm.status);
 }
 
+/*
+ * o2dlm aways has a "valid" LVB. If the dlm loses track of the LVB
+ * contents, it will zero out the LVB.  Thus the caller can always trust
+ * the contents.
+ */
+static int o2cb_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+{
+	return 1;
+}
+
 static void *o2cb_dlm_lvb(union ocfs2_dlm_lksb *lksb)
 {
 	return (void *)(lksb->lksb_o2dlm.lvb);
@@ -354,6 +364,7 @@ static struct ocfs2_stack_operations o2cb_stack_ops = {
 	.dlm_lock	= o2cb_dlm_lock,
 	.dlm_unlock	= o2cb_dlm_unlock,
 	.lock_status	= o2cb_dlm_lock_status,
+	.lvb_valid	= o2cb_dlm_lvb_valid,
 	.lock_lvb	= o2cb_dlm_lvb,
 	.dump_lksb	= o2cb_dump_lksb,
 };
diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c
index 9b76d41a8ac..ff4c798a563 100644
--- a/fs/ocfs2/stack_user.c
+++ b/fs/ocfs2/stack_user.c
@@ -738,6 +738,13 @@ static int user_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
 	return lksb->lksb_fsdlm.sb_status;
 }
 
+static int user_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+{
+	int invalid = lksb->lksb_fsdlm.sb_flags & DLM_SBF_VALNOTVALID;
+
+	return !invalid;
+}
+
 static void *user_dlm_lvb(union ocfs2_dlm_lksb *lksb)
 {
 	if (!lksb->lksb_fsdlm.sb_lvbptr)
@@ -873,6 +880,7 @@ static struct ocfs2_stack_operations ocfs2_user_plugin_ops = {
 	.dlm_lock	= user_dlm_lock,
 	.dlm_unlock	= user_dlm_unlock,
 	.lock_status	= user_dlm_lock_status,
+	.lvb_valid	= user_dlm_lvb_valid,
 	.lock_lvb	= user_dlm_lvb,
 	.plock		= user_plock,
 	.dump_lksb	= user_dlm_dump_lksb,
diff --git a/fs/ocfs2/stackglue.c b/fs/ocfs2/stackglue.c
index 68b668b0e60..3f2f1c45b7b 100644
--- a/fs/ocfs2/stackglue.c
+++ b/fs/ocfs2/stackglue.c
@@ -6,7 +6,7 @@
  * Code which implements an OCFS2 specific interface to underlying
  * cluster stacks.
  *
- * Copyright (C) 2007 Oracle.  All rights reserved.
+ * Copyright (C) 2007, 2009 Oracle.  All rights reserved.
  *
  * This program is free software; you can redistribute it and/or
  * modify it under the terms of the GNU General Public
@@ -271,11 +271,12 @@ int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb)
 }
 EXPORT_SYMBOL_GPL(ocfs2_dlm_lock_status);
 
-/*
- * Why don't we cast to ocfs2_meta_lvb?  The "clean" answer is that we
- * don't cast at the glue level.  The real answer is that the header
- * ordering is nigh impossible.
- */
+int ocfs2_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb)
+{
+	return active_stack->sp_ops->lvb_valid(lksb);
+}
+EXPORT_SYMBOL_GPL(ocfs2_dlm_lvb_valid);
+
 void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb)
 {
 	return active_stack->sp_ops->lock_lvb(lksb);
diff --git a/fs/ocfs2/stackglue.h b/fs/ocfs2/stackglue.h
index c571af375ef..03a44d60eac 100644
--- a/fs/ocfs2/stackglue.h
+++ b/fs/ocfs2/stackglue.h
@@ -186,6 +186,11 @@ struct ocfs2_stack_operations {
 	int (*lock_status)(union ocfs2_dlm_lksb *lksb);
 
 	/*
+	 * Return non-zero if the LVB is valid.
+	 */
+	int (*lvb_valid)(union ocfs2_dlm_lksb *lksb);
+
+	/*
 	 * Pull the lvb pointer off of the stack-specific lksb.
 	 */
 	void *(*lock_lvb)(union ocfs2_dlm_lksb *lksb);
@@ -252,6 +257,7 @@ int ocfs2_dlm_unlock(struct ocfs2_cluster_connection *conn,
 		     struct ocfs2_lock_res *astarg);
 
 int ocfs2_dlm_lock_status(union ocfs2_dlm_lksb *lksb);
+int ocfs2_dlm_lvb_valid(union ocfs2_dlm_lksb *lksb);
 void *ocfs2_dlm_lvb(union ocfs2_dlm_lksb *lksb);
 void ocfs2_dlm_dump_lksb(union ocfs2_dlm_lksb *lksb);