NOTE: libkvm, w, ps, 'top', and any other utility which depends on struct

      proc or any VM system structure will have to be rebuilt!!!

Much needed overhaul of the VM system. Included in this first round of
changes:

1) Improved pager interfaces: init, alloc, dealloc, getpages, putpages,
   haspage, and sync operations are supported. The haspage interface now
   provides information about clusterability. All pager routines now take
   struct vm_object's instead of "pagers".

2) Improved data structures. In the previous paradigm, there is constant
   confusion caused by pagers being both a data structure ("allocate a
   pager") and a collection of routines. The idea of a pager structure has
   escentially been eliminated. Objects now have types, and this type is
   used to index the appropriate pager. In most cases, items in the pager
   structure were duplicated in the object data structure and thus were
   unnecessary. In the few cases that remained, a un_pager structure union
   was created in the object to contain these items.

3) Because of the cleanup of #1 & #2, a lot of unnecessary layering can now
   be removed. For instance, vm_object_enter(), vm_object_lookup(),
   vm_object_remove(), and the associated object hash list were some of the
   things that were removed.

4) simple_lock's removed. Discussion with several people reveals that the
   SMP locking primitives used in the VM system aren't likely the mechanism
   that we'll be adopting. Even if it were, the locking that was in the code
   was very inadequate and would have to be mostly re-done anyway. The
   locking in a uni-processor kernel was a no-op but went a long way toward
   making the code difficult to read and debug.

5) Places that attempted to kludge-up the fact that we don't have kernel
   thread support have been fixed to reflect the reality that we are really
   dealing with processes, not threads. The VM system didn't have complete
   thread support, so the comments and mis-named routines were just wrong.
   We now use tsleep and wakeup directly in the lock routines, for instance.

6) Where appropriate, the pagers have been improved, especially in the
   pager_alloc routines. Most of the pager_allocs have been rewritten and
   are now faster and easier to maintain.

7) The pagedaemon pageout clustering algorithm has been rewritten and
   now tries harder to output an even number of pages before and after
   the requested page. This is sort of the reverse of the ideal pagein
   algorithm and should provide better overall performance.

8) Unnecessary (incorrect) casts to caddr_t in calls to tsleep & wakeup
   have been removed. Some other unnecessary casts have also been removed.

9) Some almost useless debugging code removed.

10) Terminology of shadow objects vs. backing objects straightened out.
    The fact that the vm_object data structure escentially had this
    backwards really confused things. The use of "shadow" and "backing
    object" throughout the code is now internally consistent and correct
    in the Mach terminology.

11) Several minor bug fixes, including one in the vm daemon that caused
    0 RSS objects to not get purged as intended.

12) A "default pager" has now been created which cleans up the transition
    of objects to the "swap" type. The previous checks throughout the code
    for swp->pg_data != NULL were really ugly. This change also provides
    the rudiments for future backing of "anonymous" memory by something
    other than the swap pager (via the vnode pager, for example), and it
    allows the decision about which of these pagers to use to be made
    dynamically (although will need some additional decision code to do
    this, of course).

13) (dyson) MAP_COPY has been deprecated and the corresponding "copy
    object" code has been removed. MAP_COPY was undocumented and non-
    standard. It was furthermore broken in several ways which caused its
    behavior to degrade to MAP_PRIVATE. Binaries that use MAP_COPY will
    continue to work correctly, but via the slightly different semantics
    of MAP_PRIVATE.

14) (dyson) Sharing maps have been removed. It's marginal usefulness in a
    threads design can be worked around in other ways. Both #12 and #13
    were done to simplify the code and improve readability and maintain-
    ability. (As were most all of these changes)

TODO:

1) Rewrite most of the vnode pager to use VOP_GETPAGES/PUTPAGES. Doing
   this will reduce the vnode pager to a mere fraction of its current size.

2) Rewrite vm_fault and the swap/vnode pagers to use the clustering
   information provided by the new haspage pager interface. This will
   substantially reduce the overhead by eliminating a large number of
   VOP_BMAP() calls. The VOP_BMAP() filesystem interface should be
   improved to provide both a "behind" and "ahead" indication of
   contiguousness.

3) Implement the extended features of pager_haspage in swap_pager_haspage().
   It currently just says 0 pages ahead/behind.

4) Re-implement the swap device (swstrategy) in a more elegant way, perhaps
   via a much more general mechanism that could also be used for disk
   striping of regular filesystems.

5) Do something to improve the architecture of vm_object_collapse(). The
   fact that it makes calls into the swap pager and knows too much about
   how the swap pager operates really bothers me. It also doesn't allow
   for collapsing of non-swap pager objects ("unnamed" objects backed by
   other pagers).

1 files changed, 20 insertions, 393 deletions

diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c
index c87fe7725861..73e5611c21c1 100644
--- a/sys/vm/vm_map.c
+++ b/sys/vm/vm_map.c
@@ -61,7 +61,7 @@
  * any improvements or extensions that they make and grant Carnegie the
  * rights to redistribute these changes.
  *
- * $Id: vm_map.c,v 1.21 1995/04/16 12:56:17 davidg Exp $
+ * $Id: vm_map.c,v 1.22 1995/05/30 08:16:07 rgrimes Exp $
  */
 
 /*
@@ -77,6 +77,7 @@
 #include <vm/vm_page.h>
 #include <vm/vm_object.h>
 #include <vm/vm_kern.h>
+#include <vm/vm_pager.h>
 
 /*
  *	Virtual memory maps provide for the mapping, protection,
@@ -290,8 +291,6 @@ vm_map_init(map, min, max, pageable)
 	map->hint = &map->header;
 	map->timestamp = 0;
 	lock_init(&map->lock, TRUE);
-	simple_lock_init(&map->ref_lock);
-	simple_lock_init(&map->hint_lock);
 }
 
 /*
@@ -436,9 +435,7 @@ vm_map_reference(map)
 	if (map == NULL)
 		return;
 
-	simple_lock(&map->ref_lock);
 	map->ref_count++;
-	simple_unlock(&map->ref_lock);
 }
 
 /*
@@ -457,16 +454,14 @@ vm_map_deallocate(map)
 	if (map == NULL)
 		return;
 
-	simple_lock(&map->ref_lock);
 	c = map->ref_count;
-	simple_unlock(&map->ref_lock);
 
 	if (c == 0)
 		panic("vm_map_deallocate: deallocating already freed map");
 
 	if (c != 1) {
 		--map->ref_count;
-		wakeup((caddr_t) &map->ref_count);
+		wakeup(&map->ref_count);
 		return;
 	}
 	/*
@@ -609,12 +604,10 @@ vm_map_insert(map, object, offset, start, end)
  *	SAVE_HINT:
  *
  *	Saves the specified entry as the hint for
- *	future lookups.  Performs necessary interlocks.
+ *	future lookups.
  */
 #define	SAVE_HINT(map,value) \
-		simple_lock(&(map)->hint_lock); \
-		(map)->hint = (value); \
-		simple_unlock(&(map)->hint_lock);
+		(map)->hint = (value);
 
 /*
  *	vm_map_lookup_entry:	[ internal use only ]
@@ -639,9 +632,7 @@ vm_map_lookup_entry(map, address, entry)
 	 * Start looking either from the head of the list, or from the hint.
 	 */
 
-	simple_lock(&map->hint_lock);
 	cur = map->hint;
-	simple_unlock(&map->hint_lock);
 
 	if (cur == &map->header)
 		cur = cur->next;
@@ -828,9 +819,7 @@ vm_map_simplify_entry(map, entry)
 		int count;
 
 		my_share_map = entry->object.share_map;
-		simple_lock(&my_share_map->ref_lock);
 		count = my_share_map->ref_count;
-		simple_unlock(&my_share_map->ref_lock);
 
 		if (count == 1) {
 			/*
@@ -1291,7 +1280,7 @@ vm_map_pageable(map, start, end, new_pageable)
 		 * 1).
 		 *
 		 * Downgrading to a read lock for vm_fault_wire avoids a possible
-		 * deadlock with another thread that may have faulted on one
+		 * deadlock with another process that may have faulted on one
 		 * of the pages to be wired (it would mark the page busy,
 		 * blocking us, then in turn block on the map lock that we
 		 * hold).  Because of problems in the recursive lock package,
@@ -1329,7 +1318,7 @@ vm_map_pageable(map, start, end, new_pageable)
 						entry->needs_copy = FALSE;
 					} else if (entry->object.vm_object == NULL) {
 						entry->object.vm_object =
-						    vm_object_allocate((vm_size_t) (entry->end
+						    vm_object_allocate(OBJT_DEFAULT, (vm_size_t) (entry->end
 							- entry->start));
 						entry->offset = (vm_offset_t) 0;
 					}
@@ -1367,12 +1356,12 @@ vm_map_pageable(map, start, end, new_pageable)
 		/*
 		 * HACK HACK HACK HACK
 		 *
-		 * If we are wiring in the kernel map or a submap of it, unlock
-		 * the map to avoid deadlocks.  We trust that the kernel
-		 * threads are well-behaved, and therefore will not do
-		 * anything destructive to this region of the map while we
-		 * have it unlocked.  We cannot trust user threads to do the
-		 * same.
+		 * If we are wiring in the kernel map or a submap of it,
+		 * unlock the map to avoid deadlocks.  We trust that the
+		 * kernel is well-behaved, and therefore will not do
+		 * anything destructive to this region of the map while
+		 * we have it unlocked.  We cannot trust user processes
+		 * to do the same.
 		 *
 		 * HACK HACK HACK HACK
 		 */
@@ -1493,9 +1482,7 @@ vm_map_clean(map, start, end, syncio, invalidate)
 		} else {
 			object = current->object.vm_object;
 		}
-		if (object && (object->pager != NULL) &&
-		    (object->pager->pg_type == PG_VNODE)) {
-			vm_object_lock(object);
+		if (object && (object->type == OBJT_VNODE)) {
 			/*
 			 * Flush pages if writing is allowed. XXX should we continue
 			 * on an error?
@@ -1505,10 +1492,9 @@ vm_map_clean(map, start, end, syncio, invalidate)
 			 *     idea.
 			 */
 			if (current->protection & VM_PROT_WRITE)
-		   	    	vm_object_page_clean(object, offset, offset + size, syncio);
+		   	    	vm_object_page_clean(object, offset, offset + size, syncio, TRUE);
 			if (invalidate)
 				vm_object_page_remove(object, offset, offset + size, FALSE);
-			vm_object_unlock(object);
 		}
 		start += size;
 	}
@@ -1746,9 +1732,8 @@ vm_map_copy_entry(src_map, dst_map, src_entry, dst_entry)
 	if (src_entry->is_sub_map || dst_entry->is_sub_map)
 		return;
 
-	if (dst_entry->object.vm_object != NULL &&
-	    (dst_entry->object.vm_object->flags & OBJ_INTERNAL) == 0)
-		printf("vm_map_copy_entry: copying over permanent data!\n");
+	if (dst_entry->object.vm_object != NULL)
+		printf("vm_map_copy_entry: dst_entry object not NULL!\n");
 
 	/*
 	 * If our destination map was wired down, unwire it now.
@@ -1788,9 +1773,7 @@ vm_map_copy_entry(src_map, dst_map, src_entry, dst_entry)
 			 * just protect the virtual address range.
 			 */
 			if (!(su = src_map->is_main_map)) {
-				simple_lock(&src_map->ref_lock);
 				su = (src_map->ref_count == 1);
-				simple_unlock(&src_map->ref_lock);
 			}
 			if (su) {
 				pmap_protect(src_map->pmap,
@@ -1807,7 +1790,6 @@ vm_map_copy_entry(src_map, dst_map, src_entry, dst_entry)
 		/*
 		 * Make a copy of the object.
 		 */
-		temp_object = dst_entry->object.vm_object;
 		vm_object_copy(src_entry->object.vm_object,
 		    src_entry->offset,
 		    (vm_size_t) (src_entry->end -
@@ -1834,10 +1816,6 @@ vm_map_copy_entry(src_map, dst_map, src_entry, dst_entry)
 		 */
 		src_entry->copy_on_write = TRUE;
 		dst_entry->copy_on_write = TRUE;
-		/*
-		 * Get rid of the old object.
-		 */
-		vm_object_deallocate(temp_object);
 
 		pmap_copy(dst_map->pmap, src_map->pmap, dst_entry->start,
 		    dst_entry->end - dst_entry->start, src_entry->start);
@@ -1852,292 +1830,6 @@ vm_map_copy_entry(src_map, dst_map, src_entry, dst_entry)
 }
 
 /*
- *	vm_map_copy:
- *
- *	Perform a virtual memory copy from the source
- *	address map/range to the destination map/range.
- *
- *	If src_destroy or dst_alloc is requested,
- *	the source and destination regions should be
- *	disjoint, not only in the top-level map, but
- *	in the sharing maps as well.  [The best way
- *	to guarantee this is to use a new intermediate
- *	map to make copies.  This also reduces map
- *	fragmentation.]
- */
-int
-vm_map_copy(dst_map, src_map,
-    dst_addr, len, src_addr,
-    dst_alloc, src_destroy)
-	vm_map_t dst_map;
-	vm_map_t src_map;
-	vm_offset_t dst_addr;
-	vm_size_t len;
-	vm_offset_t src_addr;
-	boolean_t dst_alloc;
-	boolean_t src_destroy;
-{
-	register
-	vm_map_entry_t src_entry;
-	register
-	vm_map_entry_t dst_entry;
-	vm_map_entry_t tmp_entry;
-	vm_offset_t src_start;
-	vm_offset_t src_end;
-	vm_offset_t dst_start;
-	vm_offset_t dst_end;
-	vm_offset_t src_clip;
-	vm_offset_t dst_clip;
-	int result;
-	boolean_t old_src_destroy;
-
-	/*
-	 * XXX While we figure out why src_destroy screws up, we'll do it by
-	 * explicitly vm_map_delete'ing at the end.
-	 */
-
-	old_src_destroy = src_destroy;
-	src_destroy = FALSE;
-
-	/*
-	 * Compute start and end of region in both maps
-	 */
-
-	src_start = src_addr;
-	src_end = src_start + len;
-	dst_start = dst_addr;
-	dst_end = dst_start + len;
-
-	/*
-	 * Check that the region can exist in both source and destination.
-	 */
-
-	if ((dst_end < dst_start) || (src_end < src_start))
-		return (KERN_NO_SPACE);
-
-	/*
-	 * Lock the maps in question -- we avoid deadlock by ordering lock
-	 * acquisition by map value
-	 */
-
-	if (src_map == dst_map) {
-		vm_map_lock(src_map);
-	} else if ((int) src_map < (int) dst_map) {
-		vm_map_lock(src_map);
-		vm_map_lock(dst_map);
-	} else {
-		vm_map_lock(dst_map);
-		vm_map_lock(src_map);
-	}
-
-	result = KERN_SUCCESS;
-
-	/*
-	 * Check protections... source must be completely readable and
-	 * destination must be completely writable.  [Note that if we're
-	 * allocating the destination region, we don't have to worry about
-	 * protection, but instead about whether the region exists.]
-	 */
-
-	if (src_map->is_main_map && dst_map->is_main_map) {
-		if (!vm_map_check_protection(src_map, src_start, src_end,
-			VM_PROT_READ)) {
-			result = KERN_PROTECTION_FAILURE;
-			goto Return;
-		}
-		if (dst_alloc) {
-			/* XXX Consider making this a vm_map_find instead */
-			if ((result = vm_map_insert(dst_map, NULL,
-				    (vm_offset_t) 0, dst_start, dst_end)) != KERN_SUCCESS)
-				goto Return;
-		} else if (!vm_map_check_protection(dst_map, dst_start, dst_end,
-			VM_PROT_WRITE)) {
-			result = KERN_PROTECTION_FAILURE;
-			goto Return;
-		}
-	}
-	/*
-	 * Find the start entries and clip.
-	 *
-	 * Note that checking protection asserts that the lookup cannot fail.
-	 *
-	 * Also note that we wait to do the second lookup until we have done the
-	 * first clip, as the clip may affect which entry we get!
-	 */
-
-	(void) vm_map_lookup_entry(src_map, src_addr, &tmp_entry);
-	src_entry = tmp_entry;
-	vm_map_clip_start(src_map, src_entry, src_start);
-
-	(void) vm_map_lookup_entry(dst_map, dst_addr, &tmp_entry);
-	dst_entry = tmp_entry;
-	vm_map_clip_start(dst_map, dst_entry, dst_start);
-
-	/*
-	 * If both source and destination entries are the same, retry the
-	 * first lookup, as it may have changed.
-	 */
-
-	if (src_entry == dst_entry) {
-		(void) vm_map_lookup_entry(src_map, src_addr, &tmp_entry);
-		src_entry = tmp_entry;
-	}
-	/*
-	 * If source and destination entries are still the same, a null copy
-	 * is being performed.
-	 */
-
-	if (src_entry == dst_entry)
-		goto Return;
-
-	/*
-	 * Go through entries until we get to the end of the region.
-	 */
-
-	while (src_start < src_end) {
-		/*
-		 * Clip the entries to the endpoint of the entire region.
-		 */
-
-		vm_map_clip_end(src_map, src_entry, src_end);
-		vm_map_clip_end(dst_map, dst_entry, dst_end);
-
-		/*
-		 * Clip each entry to the endpoint of the other entry.
-		 */
-
-		src_clip = src_entry->start + (dst_entry->end - dst_entry->start);
-		vm_map_clip_end(src_map, src_entry, src_clip);
-
-		dst_clip = dst_entry->start + (src_entry->end - src_entry->start);
-		vm_map_clip_end(dst_map, dst_entry, dst_clip);
-
-		/*
-		 * Both entries now match in size and relative endpoints.
-		 *
-		 * If both entries refer to a VM object, we can deal with them
-		 * now.
-		 */
-
-		if (!src_entry->is_a_map && !dst_entry->is_a_map) {
-			vm_map_copy_entry(src_map, dst_map, src_entry,
-			    dst_entry);
-		} else {
-			register vm_map_t new_dst_map;
-			vm_offset_t new_dst_start;
-			vm_size_t new_size;
-			vm_map_t new_src_map;
-			vm_offset_t new_src_start;
-
-			/*
-			 * We have to follow at least one sharing map.
-			 */
-
-			new_size = (dst_entry->end - dst_entry->start);
-
-			if (src_entry->is_a_map) {
-				new_src_map = src_entry->object.share_map;
-				new_src_start = src_entry->offset;
-			} else {
-				new_src_map = src_map;
-				new_src_start = src_entry->start;
-				lock_set_recursive(&src_map->lock);
-			}
-
-			if (dst_entry->is_a_map) {
-				vm_offset_t new_dst_end;
-
-				new_dst_map = dst_entry->object.share_map;
-				new_dst_start = dst_entry->offset;
-
-				/*
-				 * Since the destination sharing entries will
-				 * be merely deallocated, we can do that now,
-				 * and replace the region with a null object.
-				 * [This prevents splitting the source map to
-				 * match the form of the destination map.]
-				 * Note that we can only do so if the source
-				 * and destination do not overlap.
-				 */
-
-				new_dst_end = new_dst_start + new_size;
-
-				if (new_dst_map != new_src_map) {
-					vm_map_lock(new_dst_map);
-					(void) vm_map_delete(new_dst_map,
-					    new_dst_start,
-					    new_dst_end);
-					(void) vm_map_insert(new_dst_map,
-					    NULL,
-					    (vm_offset_t) 0,
-					    new_dst_start,
-					    new_dst_end);
-					vm_map_unlock(new_dst_map);
-				}
-			} else {
-				new_dst_map = dst_map;
-				new_dst_start = dst_entry->start;
-				lock_set_recursive(&dst_map->lock);
-			}
-
-			/*
-			 * Recursively copy the sharing map.
-			 */
-
-			(void) vm_map_copy(new_dst_map, new_src_map,
-			    new_dst_start, new_size, new_src_start,
-			    FALSE, FALSE);
-
-			if (dst_map == new_dst_map)
-				lock_clear_recursive(&dst_map->lock);
-			if (src_map == new_src_map)
-				lock_clear_recursive(&src_map->lock);
-		}
-
-		/*
-		 * Update variables for next pass through the loop.
-		 */
-
-		src_start = src_entry->end;
-		src_entry = src_entry->next;
-		dst_start = dst_entry->end;
-		dst_entry = dst_entry->next;
-
-		/*
-		 * If the source is to be destroyed, here is the place to do
-		 * it.
-		 */
-
-		if (src_destroy && src_map->is_main_map &&
-		    dst_map->is_main_map)
-			vm_map_entry_delete(src_map, src_entry->prev);
-	}
-
-	/*
-	 * Update the physical maps as appropriate
-	 */
-
-	if (src_map->is_main_map && dst_map->is_main_map) {
-		if (src_destroy)
-			pmap_remove(src_map->pmap, src_addr, src_addr + len);
-	}
-	/*
-	 * Unlock the maps
-	 */
-
-Return:;
-
-	if (old_src_destroy)
-		vm_map_delete(src_map, src_addr, src_addr + len);
-
-	vm_map_unlock(src_map);
-	if (src_map != dst_map)
-		vm_map_unlock(dst_map);
-
-	return (result);
-}
-
-/*
  * vmspace_fork:
  * Create a new process vmspace structure and vm_map
  * based on those of an existing process.  The new map
@@ -2178,58 +1870,12 @@ vmspace_fork(vm1)
 
 		case VM_INHERIT_SHARE:
 			/*
-			 * If we don't already have a sharing map:
-			 */
-
-			if (!old_entry->is_a_map) {
-				vm_map_t new_share_map;
-				vm_map_entry_t new_share_entry;
-
-				/*
-				 * Create a new sharing map
-				 */
-
-				new_share_map = vm_map_create(NULL,
-				    old_entry->start,
-				    old_entry->end,
-				    TRUE);
-				new_share_map->is_main_map = FALSE;
-
-				/*
-				 * Create the only sharing entry from the old
-				 * task map entry.
-				 */
-
-				new_share_entry =
-				    vm_map_entry_create(new_share_map);
-				*new_share_entry = *old_entry;
-				new_share_entry->wired_count = 0;
-
-				/*
-				 * Insert the entry into the new sharing map
-				 */
-
-				vm_map_entry_link(new_share_map,
-				    new_share_map->header.prev,
-				    new_share_entry);
-
-				/*
-				 * Fix up the task map entry to refer to the
-				 * sharing map now.
-				 */
-
-				old_entry->is_a_map = TRUE;
-				old_entry->object.share_map = new_share_map;
-				old_entry->offset = old_entry->start;
-			}
-			/*
 			 * Clone the entry, referencing the sharing map.
 			 */
-
 			new_entry = vm_map_entry_create(new_map);
 			*new_entry = *old_entry;
 			new_entry->wired_count = 0;
-			vm_map_reference(new_entry->object.share_map);
+			++new_entry->object.vm_object->ref_count;
 
 			/*
 			 * Insert the entry into the new map -- we know we're
@@ -2261,22 +1907,7 @@ vmspace_fork(vm1)
 			new_entry->is_a_map = FALSE;
 			vm_map_entry_link(new_map, new_map->header.prev,
 			    new_entry);
-			if (old_entry->is_a_map) {
-				int check;
-
-				check = vm_map_copy(new_map,
-				    old_entry->object.share_map,
-				    new_entry->start,
-				    (vm_size_t) (new_entry->end -
-					new_entry->start),
-				    old_entry->offset,
-				    FALSE, FALSE);
-				if (check != KERN_SUCCESS)
-					printf("vm_map_fork: copy in share_map region failed\n");
-			} else {
-				vm_map_copy_entry(old_map, new_map, old_entry,
-				    new_entry);
-			}
+			vm_map_copy_entry(old_map, new_map, old_entry, new_entry);
 			break;
 		}
 		old_entry = old_entry->next;
@@ -2350,9 +1981,7 @@ RetryLookup:;
 	 * blown lookup routine.
 	 */
 
-	simple_lock(&map->hint_lock);
 	entry = map->hint;
-	simple_unlock(&map->hint_lock);
 
 	*out_entry = entry;
 
@@ -2483,7 +2112,7 @@ RetryLookup:;
 				vm_map_unlock_read(map);
 			goto RetryLookup;
 		}
-		entry->object.vm_object = vm_object_allocate(
+		entry->object.vm_object = vm_object_allocate(OBJT_DEFAULT,
 		    (vm_size_t) (entry->end - entry->start));
 		entry->offset = 0;
 		lock_write_to_read(&share_map->lock);
@@ -2501,9 +2130,7 @@ RetryLookup:;
 	 */
 
 	if (!su) {
-		simple_lock(&share_map->ref_lock);
 		su = (share_map->ref_count == 1);
-		simple_unlock(&share_map->ref_lock);
 	}
 	*out_prot = prot;
 	*single_use = su;