diff mbox

[5/7] Record this node's information

Message ID 20130927170855.GA11732@shrek.lan (mailing list archive)
State New, archived
Headers show

Commit Message

Goldwyn Rodrigues Sept. 27, 2013, 5:09 p.m. UTC
---
 fs/ocfs2/stack_user.c | 44 ++++++++++++++++++++++++++++++++++----------
 1 file changed, 34 insertions(+), 10 deletions(-)

Comments

Joel Becker Sept. 27, 2013, 7:03 p.m. UTC | #1
See, this patch needs to be applied before we ever try to call the new
lockspace code.

Joel

On Fri, Sep 27, 2013 at 12:09:00PM -0500, Goldwyn Rodrigues wrote:
> ---
>  fs/ocfs2/stack_user.c | 44 ++++++++++++++++++++++++++++++++++----------
>  1 file changed, 34 insertions(+), 10 deletions(-)
> 
> diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c
> index ffefbb5..cccc1e1 100644
> --- a/fs/ocfs2/stack_user.c
> +++ b/fs/ocfs2/stack_user.c
> @@ -23,6 +23,7 @@
>  #include <linux/mutex.h>
>  #include <linux/slab.h>
>  #include <linux/reboot.h>
> +#include <linux/sched.h>
>  #include <asm/uaccess.h>
>  
>  #include "stackglue.h"
> @@ -116,6 +117,9 @@ struct ocfs2_live_connection {
>  	struct list_head		oc_list;
>  	struct ocfs2_cluster_connection	*oc_conn;
>  	enum ocfs2_connection_type	oc_type;
> +	atomic_t			oc_this_node;
> +	int				oc_our_slot;
> +	wait_queue_head_t		oc_wait;
>  };
>  
>  struct ocfs2_control_private {
> @@ -205,15 +209,13 @@ static struct ocfs2_live_connection *ocfs2_connection_find(const char *name)
>   * fill_super(), we can't get dupes here.
>   */
>  static int ocfs2_live_connection_new(struct ocfs2_cluster_connection *conn,
> -				     struct ocfs2_live_connection *c,
> -				     enum ocfs2_connection_type type)
> +				     struct ocfs2_live_connection *c)
>  {
>  	int rc = 0;
>  	mutex_lock(&ocfs2_control_lock);
>  	c->oc_conn = conn;
> -	c->oc_type = type;
>  
> -	if ((type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened))
> +	if ((c->oc_type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened))
>  		list_add(&c->oc_list, &ocfs2_live_connection_list);
>  	else {
>  		printk(KERN_ERR
> @@ -802,12 +804,30 @@ static void user_recover_prep(void *arg)
>  
>  static void user_recover_slot(void *arg, struct dlm_slot *slot)
>  {
> +	struct ocfs2_cluster_connection *conn =
> +		(struct ocfs2_cluster_connection  *) arg;
> +	printk(KERN_INFO "ocfs2: Node %d/%d down. Initiating recovery.\n",
> +			slot->nodeid, slot->slot);
> +	conn->cc_recovery_handler(slot->nodeid, conn->cc_recovery_data);
>  }
>  
>  static void user_recover_done(void *arg, struct dlm_slot *slots,
>  			      int num_slots, int our_slot,
>  			      uint32_t generation)
>  {
> +	struct ocfs2_cluster_connection *conn =
> +		(struct ocfs2_cluster_connection *)arg;
> +	struct ocfs2_live_connection *lc = conn->cc_private;
> +	int i;
> +
> +	for (i = 0; i < num_slots; i++)
> +		if (slots[i].slot == our_slot) {
> +			atomic_set(&lc->oc_this_node, slots[i].nodeid);
> +			break;
> +		}
> +
> +	lc->oc_our_slot = our_slot;
> +	wake_up(&lc->oc_wait);
>  }
>  
>  const struct dlm_lockspace_ops ocfs2_ls_ops = {
> @@ -830,7 +850,6 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
>  	dlm_lockspace_t *fsdlm;
>  	struct ocfs2_live_connection *lc = NULL;
>  	int rc = 0, ops_rv;
> -	enum ocfs2_connection_type type = NO_CONTROLD;
>  
>  	BUG_ON(conn == NULL);
>  
> @@ -840,6 +859,11 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
>  		goto out;
>  	}
>  
> +	init_waitqueue_head(&lc->oc_wait);
> +	atomic_set(&lc->oc_this_node, 0);
> +	conn->cc_private = lc;
> +	lc->oc_type = NO_CONTROLD;
> +
>  	rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name,
>  			       DLM_LSFL_FS, DLM_LVB_LEN,
>  				&ocfs2_ls_ops, conn, &ops_rv, &fsdlm);
> @@ -848,7 +872,7 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
>  		goto out;
>  
>  	if (ops_rv == -EOPNOTSUPP) {
> -		type = WITH_CONTROLD;
> +		lc->oc_type = WITH_CONTROLD;
>  		printk(KERN_NOTICE "ocfs2: You seem to be using an older "
>  				"version of dlm_controld and/or ocfs2-tools."
>  				" Please consider upgrading.\n");
> @@ -858,11 +882,11 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
>  	}
>  	conn->cc_lockspace = fsdlm;
>  
> -	rc = ocfs2_live_connection_new(conn, lc, type);
> +	rc = ocfs2_live_connection_new(conn, lc);
>  	if (rc)
>  		goto out;
>  
> -	if (type == WITH_CONTROLD) {
> +	if (lc->oc_type == WITH_CONTROLD) {
>  		/*
>  		 * running_proto must have been set before we allowed any mounts
>  		 * to proceed.
> @@ -890,9 +914,9 @@ static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
>  			lc = NULL;
>  			goto out;
>  		}
> -	}
> +	} else if (lc->oc_type == NO_CONTROLD)
> +		wait_event(lc->oc_wait, (atomic_read(&lc->oc_this_node) > 0));
>  
> -	conn->cc_private = lc;
>  out:
>  	if (rc && lc)
>  		kfree(lc);
> -- 
> 1.8.1.4
> 
> 
> -- 
> Goldwyn
> 
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel@oss.oracle.com
> https://oss.oracle.com/mailman/listinfo/ocfs2-devel
diff mbox

Patch

diff --git a/fs/ocfs2/stack_user.c b/fs/ocfs2/stack_user.c
index ffefbb5..cccc1e1 100644
--- a/fs/ocfs2/stack_user.c
+++ b/fs/ocfs2/stack_user.c
@@ -23,6 +23,7 @@ 
 #include <linux/mutex.h>
 #include <linux/slab.h>
 #include <linux/reboot.h>
+#include <linux/sched.h>
 #include <asm/uaccess.h>
 
 #include "stackglue.h"
@@ -116,6 +117,9 @@  struct ocfs2_live_connection {
 	struct list_head		oc_list;
 	struct ocfs2_cluster_connection	*oc_conn;
 	enum ocfs2_connection_type	oc_type;
+	atomic_t			oc_this_node;
+	int				oc_our_slot;
+	wait_queue_head_t		oc_wait;
 };
 
 struct ocfs2_control_private {
@@ -205,15 +209,13 @@  static struct ocfs2_live_connection *ocfs2_connection_find(const char *name)
  * fill_super(), we can't get dupes here.
  */
 static int ocfs2_live_connection_new(struct ocfs2_cluster_connection *conn,
-				     struct ocfs2_live_connection *c,
-				     enum ocfs2_connection_type type)
+				     struct ocfs2_live_connection *c)
 {
 	int rc = 0;
 	mutex_lock(&ocfs2_control_lock);
 	c->oc_conn = conn;
-	c->oc_type = type;
 
-	if ((type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened))
+	if ((c->oc_type == NO_CONTROLD) || atomic_read(&ocfs2_control_opened))
 		list_add(&c->oc_list, &ocfs2_live_connection_list);
 	else {
 		printk(KERN_ERR
@@ -802,12 +804,30 @@  static void user_recover_prep(void *arg)
 
 static void user_recover_slot(void *arg, struct dlm_slot *slot)
 {
+	struct ocfs2_cluster_connection *conn =
+		(struct ocfs2_cluster_connection  *) arg;
+	printk(KERN_INFO "ocfs2: Node %d/%d down. Initiating recovery.\n",
+			slot->nodeid, slot->slot);
+	conn->cc_recovery_handler(slot->nodeid, conn->cc_recovery_data);
 }
 
 static void user_recover_done(void *arg, struct dlm_slot *slots,
 			      int num_slots, int our_slot,
 			      uint32_t generation)
 {
+	struct ocfs2_cluster_connection *conn =
+		(struct ocfs2_cluster_connection *)arg;
+	struct ocfs2_live_connection *lc = conn->cc_private;
+	int i;
+
+	for (i = 0; i < num_slots; i++)
+		if (slots[i].slot == our_slot) {
+			atomic_set(&lc->oc_this_node, slots[i].nodeid);
+			break;
+		}
+
+	lc->oc_our_slot = our_slot;
+	wake_up(&lc->oc_wait);
 }
 
 const struct dlm_lockspace_ops ocfs2_ls_ops = {
@@ -830,7 +850,6 @@  static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 	dlm_lockspace_t *fsdlm;
 	struct ocfs2_live_connection *lc = NULL;
 	int rc = 0, ops_rv;
-	enum ocfs2_connection_type type = NO_CONTROLD;
 
 	BUG_ON(conn == NULL);
 
@@ -840,6 +859,11 @@  static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 		goto out;
 	}
 
+	init_waitqueue_head(&lc->oc_wait);
+	atomic_set(&lc->oc_this_node, 0);
+	conn->cc_private = lc;
+	lc->oc_type = NO_CONTROLD;
+
 	rc = dlm_new_lockspace(conn->cc_name, conn->cc_cluster_name,
 			       DLM_LSFL_FS, DLM_LVB_LEN,
 				&ocfs2_ls_ops, conn, &ops_rv, &fsdlm);
@@ -848,7 +872,7 @@  static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 		goto out;
 
 	if (ops_rv == -EOPNOTSUPP) {
-		type = WITH_CONTROLD;
+		lc->oc_type = WITH_CONTROLD;
 		printk(KERN_NOTICE "ocfs2: You seem to be using an older "
 				"version of dlm_controld and/or ocfs2-tools."
 				" Please consider upgrading.\n");
@@ -858,11 +882,11 @@  static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 	}
 	conn->cc_lockspace = fsdlm;
 
-	rc = ocfs2_live_connection_new(conn, lc, type);
+	rc = ocfs2_live_connection_new(conn, lc);
 	if (rc)
 		goto out;
 
-	if (type == WITH_CONTROLD) {
+	if (lc->oc_type == WITH_CONTROLD) {
 		/*
 		 * running_proto must have been set before we allowed any mounts
 		 * to proceed.
@@ -890,9 +914,9 @@  static int user_cluster_connect(struct ocfs2_cluster_connection *conn)
 			lc = NULL;
 			goto out;
 		}
-	}
+	} else if (lc->oc_type == NO_CONTROLD)
+		wait_event(lc->oc_wait, (atomic_read(&lc->oc_this_node) > 0));
 
-	conn->cc_private = lc;
 out:
 	if (rc && lc)
 		kfree(lc);