[PATCH] temporarily disable swap token on memory pressure

Some users (hi Zwane) have seen a problem when running a workload that
eats nearly all of physical memory - th system does an OOM kill, even
when there is still a lot of swap free.

The problem appears to be a very big task that is holding the swap
token, and the VM has a very hard time finding any other page in the
system that is swappable.

Instead of ignoring the swap token when sc->priority reaches 0, we could
simply take the swap token away from the memory hog and make sure we
don't give it back to the memory hog for a few seconds.

This patch resolves the problem Zwane ran into.

Signed-off-by: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/include/linux/rmap.h b/include/linux/rmap.h
index 35b30e6c8cf8c7b0b4fb4bd09c11b154a45c6ba6..33261f1d22395fba7d9a54a45efcaff12493336d 100644 (file)
--- a/include/linux/rmap.h
+++ b/include/linux/rmap.h
@@ -89,7 +89,7 @@ static inline void page_dup_rmap(struct page *page)
 /*
  * Called from mm/vmscan.c to handle paging out
  */
-int page_referenced(struct page *, int is_locked, int ignore_token);
+int page_referenced(struct page *, int is_locked);
 int try_to_unmap(struct page *);
 
 /*
@@ -109,7 +109,7 @@ unsigned long page_address_in_vma(struct page *, struct vm_area_struct *);
 #define anon_vma_prepare(vma)  (0)
 #define anon_vma_link(vma)     do {} while (0)
 
-#define page_referenced(page,l,i) TestClearPageReferenced(page)
+#define page_referenced(page,l) TestClearPageReferenced(page)
 #define try_to_unmap(page)     SWAP_FAIL
 
 #endif /* CONFIG_MMU */

diff --git a/include/linux/swap.h b/include/linux/swap.h
index 20c975642cab4c7f17e1cd9c8b2a6b9a2938aa03..508668f840b6a072998ce931bfc492a68c83f473 100644 (file)
--- a/include/linux/swap.h
+++ b/include/linux/swap.h
@@ -239,6 +239,11 @@ static inline void put_swap_token(struct mm_struct *mm)
                __put_swap_token(mm);
 }
 
+static inline void disable_swap_token(void)
+{
+       put_swap_token(swap_token_mm);
+}
+
 #else /* CONFIG_SWAP */
 
 #define total_swap_pages                       0
@@ -283,6 +288,7 @@ static inline swp_entry_t get_swap_page(void)
 #define put_swap_token(x) do { } while(0)
 #define grab_swap_token()  do { } while(0)
 #define has_swap_token(x) 0
+#define disable_swap_token() do { } while(0)
 
 #endif /* CONFIG_SWAP */
 #endif /* __KERNEL__*/

diff --git a/mm/rmap.c b/mm/rmap.c
index 6389cda02a20e50d3edfa836afa21895941b6e90..491ac350048f2546c8aa5c09a2f7356a46e22993 100644 (file)
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -290,7 +290,7 @@ pte_t *page_check_address(struct page *page, struct mm_struct *mm,
  * repeatedly from either page_referenced_anon or page_referenced_file.
  */
 static int page_referenced_one(struct page *page,
-       struct vm_area_struct *vma, unsigned int *mapcount, int ignore_token)
+       struct vm_area_struct *vma, unsigned int *mapcount)
 {
        struct mm_struct *mm = vma->vm_mm;
        unsigned long address;
@@ -311,7 +311,7 @@ static int page_referenced_one(struct page *page,
 
        /* Pretend the page is referenced if the task has the
           swap token and is in the middle of a page fault. */
-       if (mm != current->mm && !ignore_token && has_swap_token(mm) &&
+       if (mm != current->mm && has_swap_token(mm) &&
                        rwsem_is_locked(&mm->mmap_sem))
                referenced++;
 
@@ -321,7 +321,7 @@ out:
        return referenced;
 }
 
-static int page_referenced_anon(struct page *page, int ignore_token)
+static int page_referenced_anon(struct page *page)
 {
        unsigned int mapcount;
        struct anon_vma *anon_vma;
@@ -334,8 +334,7 @@ static int page_referenced_anon(struct page *page, int ignore_token)
 
        mapcount = page_mapcount(page);
        list_for_each_entry(vma, &anon_vma->head, anon_vma_node) {
-               referenced += page_referenced_one(page, vma, &mapcount,
-                                                       ignore_token);
+               referenced += page_referenced_one(page, vma, &mapcount);
                if (!mapcount)
                        break;
        }
@@ -354,7 +353,7 @@ static int page_referenced_anon(struct page *page, int ignore_token)
  *
  * This function is only called from page_referenced for object-based pages.
  */
-static int page_referenced_file(struct page *page, int ignore_token)
+static int page_referenced_file(struct page *page)
 {
        unsigned int mapcount;
        struct address_space *mapping = page->mapping;
@@ -392,8 +391,7 @@ static int page_referenced_file(struct page *page, int ignore_token)
                        referenced++;
                        break;
                }
-               referenced += page_referenced_one(page, vma, &mapcount,
-                                                       ignore_token);
+               referenced += page_referenced_one(page, vma, &mapcount);
                if (!mapcount)
                        break;
        }
@@ -410,13 +408,10 @@ static int page_referenced_file(struct page *page, int ignore_token)
  * Quick test_and_clear_referenced for all mappings to a page,
  * returns the number of ptes which referenced the page.
  */
-int page_referenced(struct page *page, int is_locked, int ignore_token)
+int page_referenced(struct page *page, int is_locked)
 {
        int referenced = 0;
 
-       if (!swap_token_default_timeout)
-               ignore_token = 1;
-
        if (page_test_and_clear_young(page))
                referenced++;
 
@@ -425,15 +420,14 @@ int page_referenced(struct page *page, int is_locked, int ignore_token)
 
        if (page_mapped(page) && page->mapping) {
                if (PageAnon(page))
-                       referenced += page_referenced_anon(page, ignore_token);
+                       referenced += page_referenced_anon(page);
                else if (is_locked)
-                       referenced += page_referenced_file(page, ignore_token);
+                       referenced += page_referenced_file(page);
                else if (TestSetPageLocked(page))
                        referenced++;
                else {
                        if (page->mapping)
-                               referenced += page_referenced_file(page,
-                                                               ignore_token);
+                               referenced += page_referenced_file(page);
                        unlock_page(page);
                }
        }

diff --git a/mm/thrash.c b/mm/thrash.c
index eff3c18c33a10ba802b9743c514a83ccd04aaa09..f4c560b4a2b79862340fe7457af001f8e6141715 100644 (file)
--- a/mm/thrash.c
+++ b/mm/thrash.c
@@ -57,14 +57,17 @@ void grab_swap_token(void)
        /* We have the token. Let others know we still need it. */
        if (has_swap_token(current->mm)) {
                current->mm->recent_pagein = 1;
+               if (unlikely(!swap_token_default_timeout))
+                       disable_swap_token();
                return;
        }
 
        if (time_after(jiffies, swap_token_check)) {
 
-               /* Can't get swapout protection if we exceed our RSS limit. */
-               // if (current->mm->rss > current->mm->rlimit_rss)
-               //      return;
+               if (!swap_token_default_timeout) {
+                       swap_token_check = jiffies + SWAP_TOKEN_CHECK_INTERVAL;
+                       return;
+               }
 
                /* ... or if we recently held the token. */
                if (time_before(jiffies, current->mm->swap_token_time))
@@ -95,6 +98,7 @@ void __put_swap_token(struct mm_struct *mm)
 {
        spin_lock(&swap_token_lock);
        if (likely(mm == swap_token_mm)) {
+               mm->swap_token_time = jiffies + SWAP_TOKEN_CHECK_INTERVAL;
                swap_token_mm = &init_mm;
                swap_token_check = jiffies;
        }

diff --git a/mm/vmscan.c b/mm/vmscan.c
index 28130541270f5ad35b3e18ac878b349c61c6e01f..078cf920208af05e1e5d250dc7941ee134d02ada 100644 (file)
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -407,7 +407,7 @@ static int shrink_list(struct list_head *page_list, struct scan_control *sc)
                if (PageWriteback(page))
                        goto keep_locked;
 
-               referenced = page_referenced(page, 1, sc->priority <= 0);
+               referenced = page_referenced(page, 1);
                /* In active use or really unfreeable?  Activate it. */
                if (referenced && page_mapping_inuse(page))
                        goto activate_locked;
@@ -756,7 +756,7 @@ refill_inactive_zone(struct zone *zone, struct scan_control *sc)
                if (page_mapped(page)) {
                        if (!reclaim_mapped ||
                            (total_swap_pages == 0 && PageAnon(page)) ||
-                           page_referenced(page, 0, sc->priority <= 0)) {
+                           page_referenced(page, 0)) {
                                list_add(&page->lru, &l_active);
                                continue;
                        }
@@ -960,6 +960,8 @@ int try_to_free_pages(struct zone **zones, gfp_t gfp_mask)
                sc.nr_reclaimed = 0;
                sc.priority = priority;
                sc.swap_cluster_max = SWAP_CLUSTER_MAX;
+               if (!priority)
+                       disable_swap_token();
                shrink_caches(zones, &sc);
                shrink_slab(sc.nr_scanned, gfp_mask, lru_pages);
                if (reclaim_state) {
@@ -1056,6 +1058,10 @@ loop_again:
                int end_zone = 0;       /* Inclusive.  0 = ZONE_DMA */
                unsigned long lru_pages = 0;
 
+               /* The swap token gets in the way of swapout... */
+               if (!priority)
+                       disable_swap_token();
+
                all_zones_ok = 1;
 
                if (nr_pages == 0) {
@@ -1360,6 +1366,7 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
        sc.nr_reclaimed = 0;
        /* scan at the highest priority */
        sc.priority = 0;
+       disable_swap_token();
 
        if (nr_pages > SWAP_CLUSTER_MAX)
                sc.swap_cluster_max = nr_pages;