Use vm_page_replace_checked() instead of vm_page_rename() for implementing

optimized copy-on-write faults.  This has two advantages: (1) one less radix
tree operation is performed and (2) vm_page_replace_checked() cannot fail,
making the code simpler.

Submitted by:	Ryan Libby
Reviewed by:	kib
Sponsored by:	EMC / Isilon Storage Division
Differential Revision:	https://reviews.freebsd.org/D4478

1 files changed, 10 insertions, 17 deletions

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 66506f3..50bf725 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -799,26 +799,15 @@ vnode_locked:
 				 * We don't chase down the shadow chain
 				 */
 			    fs.object == fs.first_object->backing_object) {
-				/*
-				 * get rid of the unnecessary page
-				 */
-				vm_page_lock(fs.first_m);
-				vm_page_remove(fs.first_m);
-				vm_page_unlock(fs.first_m);
-				/*
-				 * grab the page and put it into the 
-				 * process'es object.  The page is 
-				 * automatically made dirty.
-				 */
-				if (vm_page_rename(fs.m, fs.first_object,
-				    fs.first_pindex)) {
-					VM_OBJECT_WUNLOCK(fs.first_object);
-					unlock_and_deallocate(&fs);
-					goto RetryFault;
-				}
+				vm_page_lock(fs.m);
+				vm_page_remove(fs.m);
+				vm_page_unlock(fs.m);
 				vm_page_lock(fs.first_m);
+				vm_page_replace_checked(fs.m, fs.first_object,
+				    fs.first_pindex, fs.first_m);
 				vm_page_free(fs.first_m);
 				vm_page_unlock(fs.first_m);
+				vm_page_dirty(fs.m);
 #if VM_NRESERVLEVEL > 0
 				/*
 				 * Rename the reservation.
@@ -827,6 +816,10 @@ vnode_locked:
 				    fs.object, OFF_TO_IDX(
 				    fs.first_object->backing_object_offset));
 #endif
+				/*
+				 * Removing the page from the backing object
+				 * unbusied it.
+				 */
 				vm_page_xbusy(fs.m);
 				fs.first_m = fs.m;
 				fs.m = NULL;