Changes to make the OOM killer much more effective:

- Allow the OOM killer to target processes currently locked in
  memory.  These very often are the ones doing the memory hogging.
- Drop the wakeup priority of processes currently sleeping while
  waiting for their page fault to complete.  In order for the OOM
  killer to work well, the killed process and other system processes
  waiting on memory must be allowed to wakeup first.

Reviewed by:	dillon
MFC after:	1 week

4 files changed, 28 insertions, 4 deletions

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 79e1224..77e1615 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -347,7 +347,7 @@ RetryFault:;
 			if ((queue - fs.m->pc) == PQ_CACHE && vm_page_count_severe()) {
 				vm_page_activate(fs.m);
 				unlock_and_deallocate(&fs);
-				VM_WAIT;
+				VM_WAITPFAULT;
 				goto RetryFault;
 			}
 
@@ -388,7 +388,7 @@ RetryFault:;
 			}
 			if (fs.m == NULL) {
 				unlock_and_deallocate(&fs);
-				VM_WAIT;
+				VM_WAITPFAULT;
 				goto RetryFault;
 			}
 		}
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index abc4194..2c0052b 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -967,6 +967,7 @@ loop:
  *	vm_wait:	(also see VM_WAIT macro)
  *
  *	Block until free pages are available for allocation
+ *	- Called in various places before memory allocations.
  */
 
 void
@@ -989,6 +990,28 @@ vm_wait(void)
 }
 
 /*
+ *	vm_waitpfault:	(also see VM_WAITPFAULT macro)
+ *
+ *	Block until free pages are available for allocation
+ *	- Called only in vm_fault so that processes page faulting
+ *	  can be easily tracked.
+ */
+
+void
+vm_waitpfault(void)
+{
+	int s;
+
+	s = splvm();
+	if (!vm_pages_needed) {
+		vm_pages_needed = 1;
+		wakeup(&vm_pages_needed);
+	}
+	tsleep(&cnt.v_free_count, PUSER, "pfault", 0);
+	splx(s);
+}
+
+/*
  *	vm_page_activate:
  *
  *	Put the specified page on the active list (if appropriate).
diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 57f22b2..657f5e7 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -1141,8 +1141,7 @@ rescan0:
 			/*
 			 * if this is a system process, skip it
 			 */
-			if ((p->p_flag & P_SYSTEM) || (p->p_lock > 0) ||
-			    (p->p_pid == 1) ||
+			if ((p->p_flag & P_SYSTEM) || (p->p_pid == 1) ||
 			    ((p->p_pid < 48) && (vm_swap_size != 0))) {
 				PROC_UNLOCK(p);
 				continue;
diff --git a/sys/vm/vm_pageout.h b/sys/vm/vm_pageout.h
index bfcfbb1..fc88589 100644
--- a/sys/vm/vm_pageout.h
+++ b/sys/vm/vm_pageout.h
@@ -100,7 +100,9 @@ extern int vm_pageout_deficit;
 
 extern void pagedaemon_wakeup __P((void));
 #define VM_WAIT vm_wait()
+#define VM_WAITPFAULT vm_waitpfault()
 extern void vm_wait __P((void));
+extern void vm_waitpfault __P((void));
 
 #ifdef _KERNEL
 void vm_pageout_page __P((vm_page_t, vm_object_t));