[PATCH] ipmi: fix inconsistent spinlock usage

[PATCH] ipmi: fix inconsistent spinlock usage

[Corey Minyard]

Linus,

An obvious bug.  Thanks, Hironobu.

Part of a patch was accidentally reverted, this corrects an
inconsistent spinlock use in the IPMI message handler.

Signed-off-by: Hironobu Ishii <hishii@soft.fujitsu.com>
Signed-off-by: Corey Minyard <minyard@acm.org>

Index: linux-2.6.14-rc4/drivers/char/ipmi/ipmi_msghandler.c
===================================================================
--- linux-2.6.14-rc4.orig/drivers/char/ipmi/ipmi_msghandler.c
+++ linux-2.6.14-rc4/drivers/char/ipmi/ipmi_msghandler.c
@@ -2664,7 +2664,7 @@ void ipmi_smi_msg_received(ipmi_smi_t   
 	spin_lock_irqsave(&intf->waiting_msgs_lock, flags);
 	if (!list_empty(&intf->waiting_msgs)) {
 		list_add_tail(&msg->link, &intf->waiting_msgs);
-		spin_unlock(&intf->waiting_msgs_lock);
+		spin_unlock_irqrestore(&intf->waiting_msgs_lock, flags);
 		goto out;
 	}
 	spin_unlock_irqrestore(&intf->waiting_msgs_lock, flags);
@@ -2673,9 +2673,9 @@ void ipmi_smi_msg_received(ipmi_smi_t   
 	if (rv > 0) {
 		/* Could not handle the message now, just add it to a
                    list to handle later. */
-		spin_lock(&intf->waiting_msgs_lock);
+		spin_lock_irqsave(&intf->waiting_msgs_lock, flags);
 		list_add_tail(&msg->link, &intf->waiting_msgs);
-		spin_unlock(&intf->waiting_msgs_lock);
+		spin_unlock_irqrestore(&intf->waiting_msgs_lock, flags);
 	} else if (rv == 0) {
 		ipmi_free_smi_msg(msg);
 	}