Don't attempt to clean device_pager backed objects at terminate time.

There is similar bogusness in the pageout daemon that will be fixed soon.
This fixes a panic pointed out to me by Bruce Evans that occurs when
/dev/mem is used to map managed memory.
This commit is contained in:
David Greenman 1995-01-25 20:36:29 +00:00
parent 673f8dc3be
commit a465acdaf8
Notes: svn2git 2020-12-20 02:59:44 +00:00
svn path=/head/; revision=5903

View File

@ -61,7 +61,7 @@
* any improvements or extensions that they make and grant Carnegie the
* rights to redistribute these changes.
*
* $Id: vm_object.c,v 1.18 1995/01/13 13:30:24 davidg Exp $
* $Id: vm_object.c,v 1.19 1995/01/24 10:13:14 davidg Exp $
*/
/*
@ -384,7 +384,8 @@ vm_object_terminate(object)
* object are gone, so we don't need to lock it.
*/
if ((object->flags & OBJ_INTERNAL) == 0) {
if (((object->flags & OBJ_INTERNAL) == 0) &&
object->pager && (object->pager->pg_type != PG_DEVICE)) {
(void) vm_object_page_clean(object, 0, 0, TRUE, TRUE);
}
/*