Revert change in revision 1.53 and add a small comment to protect

the revived code.

vm pages newly allocated are marked busy (PG_BUSY), thus calling
vm_page_delete before the pages has been freed or unbusied will
cause a deadlock since vm_page_object_page_remove will wait for the
busy flag to be cleared.  This can be triggered by calling malloc
with size > PAGE_SIZE and the M_NOWAIT flag on systems low on
physical free memory.

A kernel module that reproduces the problem, written by Logan Gabriel
<logan@mail.2cactus.com>, can be found in the freebsd-hackers mail
archive (12 Apr 2001).  The problem was recently noticed again by
Archie Cobbs <archie@dellroad.org>.

Reviewed by:	dillon

1 files changed, 12 insertions, 0 deletions

diff --git a/sys/vm/vm_kern.c b/sys/vm/vm_kern.c
index 2e5ea30..5a0dfda 100644
--- a/sys/vm/vm_kern.c
+++ b/sys/vm/vm_kern.c
@@ -378,6 +378,18 @@ retry:
 				vm_map_lock(map);
 				goto retry;
 			}
+			/* 
+			 * Free the pages before removing the map entry.
+			 * They are already marked busy.  Calling
+			 * vm_map_delete before the pages has been freed or
+			 * unbusied will cause a deadlock.
+			 */
+			while (i != 0) {
+				i -= PAGE_SIZE;
+				m = vm_page_lookup(kmem_object,
+						   OFF_TO_IDX(offset + i));
+				vm_page_free(m);
+			}
 			vm_map_delete(map, addr, addr + size);
 			vm_map_unlock(map);
 			goto bad;