diff --git a/sys/sys/vmmeter.h b/sys/sys/vmmeter.h
--- a/sys/sys/vmmeter.h
+++ b/sys/sys/vmmeter.h
@@ -120,6 +120,7 @@
 	counter_u64_t v_rforkpages;	/* (p) pages affected by rfork() */
 	counter_u64_t v_kthreadpages;	/* (p) ... and by kernel fork() */
 	counter_u64_t v_wire_count;	/* (p) pages wired down */
+	counter_u64_t v_nofree_count;	/* (p) permanently allocated pages */
 #define	VM_METER_NCOUNTERS	\
 	(offsetof(struct vmmeter, v_page_size) / sizeof(counter_u64_t))
 	/*
@@ -174,6 +175,13 @@
 	return (VM_CNT_FETCH(v_wire_count));
 }
 
+static inline u_int
+vm_nofree_count(void)
+{
+
+	return (VM_CNT_FETCH(v_nofree_count));
+}
+
 /*
  * Return TRUE if we are under our severe low-free-pages threshold
  *
diff --git a/sys/vm/vm_meter.c b/sys/vm/vm_meter.c
--- a/sys/vm/vm_meter.c
+++ b/sys/vm/vm_meter.c
@@ -90,6 +90,7 @@
 	.v_rforkpages = EARLY_COUNTER,
 	.v_kthreadpages = EARLY_COUNTER,
 	.v_wire_count = EARLY_COUNTER,
+	.v_nofree_count = EARLY_COUNTER,
 };
 
 u_long __exclusive_cache_line vm_user_wire_count;
@@ -386,6 +387,7 @@
 VM_STATS_UINT(v_free_min, "Minimum low-free-pages threshold");
 VM_STATS_PROC(v_free_count, "Free pages", vm_free_count);
 VM_STATS_PROC(v_wire_count, "Wired pages", vm_wire_count);
+VM_STATS_PROC(v_nofree_count, "Permanently allocated pages", vm_nofree_count);
 VM_STATS_PROC(v_active_count, "Active pages", vm_active_count);
 VM_STATS_UINT(v_inactive_target, "Desired inactive pages");
 VM_STATS_PROC(v_inactive_count, "Inactive pages", vm_inactive_count);