- Introduce and use a new tunable "debug.mpsafevm". At present, setting

   "debug.mpsafevm" results in (almost) Giant-free execution of zero-fill
   page faults.  (Giant is held only briefly, just long enough to determine
   if there is a vnode backing the faulting address.)

   Also, condition the acquisition and release of Giant around calls to
   pmap_remove() on "debug.mpsafevm".

   The effect on performance is significant.  On my dual Opteron, I see a
   3.6% reduction in "buildworld" time.

 - Use atomic operations to update several counters in vm_fault().

1 files changed, 4 insertions, 4 deletions

diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c
index 765126c..2a486e8 100644
--- a/sys/vm/vm_map.c
+++ b/sys/vm/vm_map.c
@@ -2114,9 +2114,9 @@ vm_map_sync(
 	}
 
 	if (invalidate) {
-		mtx_lock(&Giant);
+		VM_LOCK_GIANT();
 		pmap_remove(map->pmap, start, end);
-		mtx_unlock(&Giant);
+		VM_UNLOCK_GIANT();
 	}
 	/*
 	 * Make a second pass, cleaning/uncaching pages from the indicated
@@ -2282,10 +2282,10 @@ vm_map_delete(vm_map_t map, vm_offset_t start, vm_offset_t end)
 		}
 
 		if (!map->system_map)
-			mtx_lock(&Giant);
+			VM_LOCK_GIANT();
 		pmap_remove(map->pmap, entry->start, entry->end);
 		if (!map->system_map)
-			mtx_unlock(&Giant);
+			VM_UNLOCK_GIANT();
 
 		/*
 		 * Delete the entry (which may delete the object) only after