vmscan: merge duplicate code in shrink_active_list()
Wu Fengguang [Tue, 16 Jun 2009 22:33:13 +0000 (15:33 -0700)]
The "move pages to active list" and "move pages to inactive list" code
blocks are mostly identical and can be served by a function.

Thanks to Andrew Morton for pointing this out.

Note that buffer_heads_over_limit check will also be carried out for
re-activated pages, which is slightly different from pre-2.6.28 kernels.
Also, Rik's "vmscan: evict use-once pages first" patch could totally stop
scans of active file list when memory pressure is low.  So the net effect
could be, the number of buffer heads is now more likely to grow large.

However that's fine according to Johannes' comments:

  I don't think that this could be harmful.  We just preserve the buffer
  mappings of what we consider the working set and with low memory
  pressure, as you say, this set is not big.

  As to stripping of reactivated pages: the only pages we re-activate
  for now are those VM_EXEC mapped ones.  Since we don't expect IO from
  or to these pages, removing the buffer mappings in case they grow too
  large should be okay, I guess.

Cc: Pekka Enberg <penberg@cs.helsinki.fi>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Reviewed-by: Minchan Kim <minchan.kim@gmail.com>
Reviewed-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/vmscan.c

index 1024979..f3a55d1 100644 (file)
@@ -1204,6 +1204,43 @@ static inline void note_zone_scanning_priority(struct zone *zone, int priority)
  * But we had to alter page->flags anyway.
  */
 
+static void move_active_pages_to_lru(struct zone *zone,
+                                    struct list_head *list,
+                                    enum lru_list lru)
+{
+       unsigned long pgmoved = 0;
+       struct pagevec pvec;
+       struct page *page;
+
+       pagevec_init(&pvec, 1);
+
+       while (!list_empty(list)) {
+               page = lru_to_page(list);
+               prefetchw_prev_lru_page(page, list, flags);
+
+               VM_BUG_ON(PageLRU(page));
+               SetPageLRU(page);
+
+               VM_BUG_ON(!PageActive(page));
+               if (!is_active_lru(lru))
+                       ClearPageActive(page);  /* we are de-activating */
+
+               list_move(&page->lru, &zone->lru[lru].list);
+               mem_cgroup_add_lru_list(page, lru);
+               pgmoved++;
+
+               if (!pagevec_add(&pvec, page) || list_empty(list)) {
+                       spin_unlock_irq(&zone->lru_lock);
+                       if (buffer_heads_over_limit)
+                               pagevec_strip(&pvec);
+                       __pagevec_release(&pvec);
+                       spin_lock_irq(&zone->lru_lock);
+               }
+       }
+       __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
+       if (!is_active_lru(lru))
+               __count_vm_events(PGDEACTIVATE, pgmoved);
+}
 
 static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
                        struct scan_control *sc, int priority, int file)
@@ -1215,8 +1252,6 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
        LIST_HEAD(l_active);
        LIST_HEAD(l_inactive);
        struct page *page;
-       struct pagevec pvec;
-       enum lru_list lru;
        struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc);
 
        lru_add_drain();
@@ -1233,6 +1268,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
        }
        reclaim_stat->recent_scanned[!!file] += pgmoved;
 
+       __count_zone_vm_events(PGREFILL, zone, pgscanned);
        if (file)
                __mod_zone_page_state(zone, NR_ACTIVE_FILE, -pgmoved);
        else
@@ -1275,8 +1311,6 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
        /*
         * Move pages back to the lru list.
         */
-       pagevec_init(&pvec, 1);
-
        spin_lock_irq(&zone->lru_lock);
        /*
         * Count referenced pages from currently used mappings as rotated,
@@ -1286,57 +1320,12 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone,
         */
        reclaim_stat->recent_rotated[!!file] += pgmoved;
 
-       pgmoved = 0;  /* count pages moved to inactive list */
-       lru = LRU_BASE + file * LRU_FILE;
-       while (!list_empty(&l_inactive)) {
-               page = lru_to_page(&l_inactive);
-               prefetchw_prev_lru_page(page, &l_inactive, flags);
-               VM_BUG_ON(PageLRU(page));
-               SetPageLRU(page);
-               VM_BUG_ON(!PageActive(page));
-               ClearPageActive(page);
-
-               list_move(&page->lru, &zone->lru[lru].list);
-               mem_cgroup_add_lru_list(page, lru);
-               pgmoved++;
-               if (!pagevec_add(&pvec, page)) {
-                       spin_unlock_irq(&zone->lru_lock);
-                       if (buffer_heads_over_limit)
-                               pagevec_strip(&pvec);
-                       __pagevec_release(&pvec);
-                       spin_lock_irq(&zone->lru_lock);
-               }
-       }
-       __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
-       __count_zone_vm_events(PGREFILL, zone, pgscanned);
-       __count_vm_events(PGDEACTIVATE, pgmoved);
-
-       pgmoved = 0;  /* count pages moved back to active list */
-       lru = LRU_ACTIVE + file * LRU_FILE;
-       while (!list_empty(&l_active)) {
-               page = lru_to_page(&l_active);
-               prefetchw_prev_lru_page(page, &l_active, flags);
-               VM_BUG_ON(PageLRU(page));
-               SetPageLRU(page);
-               VM_BUG_ON(!PageActive(page));
-
-               list_move(&page->lru, &zone->lru[lru].list);
-               mem_cgroup_add_lru_list(page, lru);
-               pgmoved++;
-               if (!pagevec_add(&pvec, page)) {
-                       spin_unlock_irq(&zone->lru_lock);
-                       if (buffer_heads_over_limit)
-                               pagevec_strip(&pvec);
-                       __pagevec_release(&pvec);
-                       spin_lock_irq(&zone->lru_lock);
-               }
-       }
-       __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved);
+       move_active_pages_to_lru(zone, &l_active,
+                                               LRU_ACTIVE + file * LRU_FILE);
+       move_active_pages_to_lru(zone, &l_inactive,
+                                               LRU_BASE   + file * LRU_FILE);
 
        spin_unlock_irq(&zone->lru_lock);
-       if (buffer_heads_over_limit)
-               pagevec_strip(&pvec);
-       pagevec_release(&pvec);
 }
 
 static int inactive_anon_is_low_global(struct zone *zone)