Merge branch 'master'

13 files changed, 599 insertions, 214 deletions

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 67f29516662a..508707704d2c 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -85,7 +85,7 @@ void free_huge_page(struct page *page)
         BUG_ON(page_count(page));
 
         INIT_LIST_HEAD(&page->lru);
-        page[1].mapping = NULL;
+        page[1].lru.next = NULL;                        /* reset dtor */
 
         spin_lock(&hugetlb_lock);
         enqueue_huge_page(page);
@@ -105,7 +105,7 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, unsigned long addr)
         }
         spin_unlock(&hugetlb_lock);
         set_page_count(page, 1);
-        page[1].mapping = (void *)free_huge_page;
+        page[1].lru.next = (void *)free_huge_page;      /* set dtor */
         for (i = 0; i < (HPAGE_SIZE/PAGE_SIZE); ++i)
                 clear_user_highpage(&page[i], addr);
         return page;
diff --git a/mm/madvise.c b/mm/madvise.c
index ae0ae3ea299a..af3d573b0141 100644
--- a/mm/madvise.c
+++ b/mm/madvise.c
@@ -22,16 +22,23 @@ static long madvise_behavior(struct vm_area_struct * vma,
         struct mm_struct * mm = vma->vm_mm;
         int error = 0;
         pgoff_t pgoff;
-        int new_flags = vma->vm_flags & ~VM_READHINTMASK;
+        int new_flags = vma->vm_flags;
 
         switch (behavior) {
+        case MADV_NORMAL:
+                new_flags = new_flags & ~VM_RAND_READ & ~VM_SEQ_READ;
+                break;
         case MADV_SEQUENTIAL:
-                new_flags |= VM_SEQ_READ;
+                new_flags = (new_flags & ~VM_RAND_READ) | VM_SEQ_READ;
                 break;
         case MADV_RANDOM:
-                new_flags |= VM_RAND_READ;
+                new_flags = (new_flags & ~VM_SEQ_READ) | VM_RAND_READ;
                 break;
-        default:
+        case MADV_DONTFORK:
+                new_flags |= VM_DONTCOPY;
+                break;
+        case MADV_DOFORK:
+                new_flags &= ~VM_DONTCOPY;
                 break;
         }
 
@@ -177,6 +184,12 @@ madvise_vma(struct vm_area_struct *vma, struct vm_area_struct **prev,
         long error;
 
         switch (behavior) {
+        case MADV_DOFORK:
+                if (vma->vm_flags & VM_IO) {
+                        error = -EINVAL;
+                        break;
+                }
+        case MADV_DONTFORK:
         case MADV_NORMAL:
         case MADV_SEQUENTIAL:
         case MADV_RANDOM:
diff --git a/mm/memory.c b/mm/memory.c
index 2bee1f21aa8a..9abc6008544b 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -82,6 +82,16 @@ EXPORT_SYMBOL(num_physpages);
 EXPORT_SYMBOL(high_memory);
 EXPORT_SYMBOL(vmalloc_earlyreserve);
 
+int randomize_va_space __read_mostly = 1;
+
+static int __init disable_randmaps(char *s)
+{
+        randomize_va_space = 0;
+        return 0;
+}
+__setup("norandmaps", disable_randmaps);
+
+
 /*
  * If a p?d_bad entry is found while walking page tables, report
  * the error, before resetting entry to p?d_none.  Usually (but
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index a918f77f02f3..1fe76d963ac2 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -130,6 +130,7 @@ int online_pages(unsigned long pfn, unsigned long nr_pages)
                 onlined_pages++;
         }
         zone->present_pages += onlined_pages;
+        zone->zone_pgdat->node_present_pages += onlined_pages;
 
         setup_per_zone_pages_min();
 
diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 3bd7fb7e4b75..2a8206009422 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -132,19 +132,29 @@ static int mpol_check_policy(int mode, nodemask_t *nodes)
         }
         return nodes_subset(*nodes, node_online_map) ? 0 : -EINVAL;
 }
+
 /* Generate a custom zonelist for the BIND policy. */
 static struct zonelist *bind_zonelist(nodemask_t *nodes)
 {
         struct zonelist *zl;
-        int num, max, nd;
+        int num, max, nd, k;
 
         max = 1 + MAX_NR_ZONES * nodes_weight(*nodes);
-        zl = kmalloc(sizeof(void *) * max, GFP_KERNEL);
+        zl = kmalloc(sizeof(struct zone *) * max, GFP_KERNEL);
         if (!zl)
                 return NULL;
         num = 0;
-        for_each_node_mask(nd, *nodes)
-                zl->zones[num++] = &NODE_DATA(nd)->node_zones[policy_zone];
+        /* First put in the highest zones from all nodes, then all the next 
+           lower zones etc. Avoid empty zones because the memory allocator
+           doesn't like them. If you implement node hot removal you
+           have to fix that. */
+        for (k = policy_zone; k >= 0; k--) { 
+                for_each_node_mask(nd, *nodes) { 
+                        struct zone *z = &NODE_DATA(nd)->node_zones[k];
+                        if (z->present_pages > 0) 
+                                zl->zones[num++] = z;
+                }
+        }
         zl->zones[num] = NULL;
         return zl;
 }
@@ -187,7 +197,7 @@ static struct mempolicy *mpol_new(int mode, nodemask_t *nodes)
         return policy;
 }
 
-static void gather_stats(struct page *, void *);
+static void gather_stats(struct page *, void *, int pte_dirty);
 static void migrate_page_add(struct page *page, struct list_head *pagelist,
                                 unsigned long flags);
 
@@ -229,7 +239,7 @@ static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
                         continue;
 
                 if (flags & MPOL_MF_STATS)
-                        gather_stats(page, private);
+                        gather_stats(page, private, pte_dirty(*pte));
                 else if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
                         migrate_page_add(page, private, flags);
                 else
@@ -542,7 +552,7 @@ static void migrate_page_add(struct page *page, struct list_head *pagelist,
          */
         if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1) {
                 if (isolate_lru_page(page))
-                        list_add(&page->lru, pagelist);
+                        list_add_tail(&page->lru, pagelist);
         }
 }
 
@@ -559,6 +569,7 @@ static int migrate_pages_to(struct list_head *pagelist,
         LIST_HEAD(moved);
         LIST_HEAD(failed);
         int err = 0;
+        unsigned long offset = 0;
         int nr_pages;
         struct page *page;
         struct list_head *p;
@@ -566,8 +577,21 @@ static int migrate_pages_to(struct list_head *pagelist,
 redo:
         nr_pages = 0;
         list_for_each(p, pagelist) {
-                if (vma)
-                        page = alloc_page_vma(GFP_HIGHUSER, vma, vma->vm_start);
+                if (vma) {
+                        /*
+                         * The address passed to alloc_page_vma is used to
+                         * generate the proper interleave behavior. We fake
+                         * the address here by an increasing offset in order
+                         * to get the proper distribution of pages.
+                         *
+                         * No decision has been made as to which page
+                         * a certain old page is moved to so we cannot
+                         * specify the correct address.
+                         */
+                        page = alloc_page_vma(GFP_HIGHUSER, vma,
+                                        offset + vma->vm_start);
+                        offset += PAGE_SIZE;
+                }
                 else
                         page = alloc_pages_node(dest, GFP_HIGHUSER, 0);
 
@@ -575,9 +599,9 @@ redo:
                         err = -ENOMEM;
                         goto out;
                 }
-                list_add(&page->lru, &newlist);
+                list_add_tail(&page->lru, &newlist);
                 nr_pages++;
-                if (nr_pages > MIGRATE_CHUNK_SIZE);
+                if (nr_pages > MIGRATE_CHUNK_SIZE)
                         break;
         }
         err = migrate_pages(pagelist, &newlist, &moved, &failed);
@@ -724,7 +748,7 @@ long do_mbind(unsigned long start, unsigned long len,
                                       MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
             || mode > MPOL_MAX)
                 return -EINVAL;
-        if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_RESOURCE))
+        if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE))
                 return -EPERM;
 
         if (start & ~PAGE_MASK)
@@ -798,6 +822,8 @@ static int get_nodes(nodemask_t *nodes, const unsigned long __user *nmask,
         nodes_clear(*nodes);
         if (maxnode == 0 || !nmask)
                 return 0;
+        if (maxnode > PAGE_SIZE*BITS_PER_BYTE)
+                return -EINVAL;
 
         nlongs = BITS_TO_LONGS(maxnode);
         if ((maxnode % BITS_PER_LONG) == 0)
@@ -916,19 +942,20 @@ asmlinkage long sys_migrate_pages(pid_t pid, unsigned long maxnode,
          */
         if ((current->euid != task->suid) && (current->euid != task->uid) &&
             (current->uid != task->suid) && (current->uid != task->uid) &&
-            !capable(CAP_SYS_ADMIN)) {
+            !capable(CAP_SYS_NICE)) {
                 err = -EPERM;
                 goto out;
         }
 
         task_nodes = cpuset_mems_allowed(task);
         /* Is the user allowed to access the target nodes? */
-        if (!nodes_subset(new, task_nodes) && !capable(CAP_SYS_ADMIN)) {
+        if (!nodes_subset(new, task_nodes) && !capable(CAP_SYS_NICE)) {
                 err = -EPERM;
                 goto out;
         }
 
-        err = do_migrate_pages(mm, &old, &new, MPOL_MF_MOVE);
+        err = do_migrate_pages(mm, &old, &new,
+                capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
 out:
         mmput(mm);
         return err;
@@ -1726,66 +1753,145 @@ static inline int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol)
 struct numa_maps {
         unsigned long pages;
         unsigned long anon;
-        unsigned long mapped;
+        unsigned long active;
+        unsigned long writeback;
         unsigned long mapcount_max;
+        unsigned long dirty;
+        unsigned long swapcache;
         unsigned long node[MAX_NUMNODES];
 };
 
-static void gather_stats(struct page *page, void *private)
+static void gather_stats(struct page *page, void *private, int pte_dirty)
 {
         struct numa_maps *md = private;
         int count = page_mapcount(page);
 
-        if (count)
-                md->mapped++;
+        md->pages++;
+        if (pte_dirty || PageDirty(page))
+                md->dirty++;
 
-        if (count > md->mapcount_max)
-                md->mapcount_max = count;
+        if (PageSwapCache(page))
+                md->swapcache++;
 
-        md->pages++;
+        if (PageActive(page))
+                md->active++;
+
+        if (PageWriteback(page))
+                md->writeback++;
 
         if (PageAnon(page))
                 md->anon++;
 
+        if (count > md->mapcount_max)
+                md->mapcount_max = count;
+
         md->node[page_to_nid(page)]++;
         cond_resched();
 }
 
+#ifdef CONFIG_HUGETLB_PAGE
+static void check_huge_range(struct vm_area_struct *vma,
+                unsigned long start, unsigned long end,
+                struct numa_maps *md)
+{
+        unsigned long addr;
+        struct page *page;
+
+        for (addr = start; addr < end; addr += HPAGE_SIZE) {
+                pte_t *ptep = huge_pte_offset(vma->vm_mm, addr & HPAGE_MASK);
+                pte_t pte;
+
+                if (!ptep)
+                        continue;
+
+                pte = *ptep;
+                if (pte_none(pte))
+                        continue;
+
+                page = pte_page(pte);
+                if (!page)
+                        continue;
+
+                gather_stats(page, md, pte_dirty(*ptep));
+        }
+}
+#else
+static inline void check_huge_range(struct vm_area_struct *vma,
+                unsigned long start, unsigned long end,
+                struct numa_maps *md)
+{
+}
+#endif
+
 int show_numa_map(struct seq_file *m, void *v)
 {
         struct task_struct *task = m->private;
         struct vm_area_struct *vma = v;
         struct numa_maps *md;
+        struct file *file = vma->vm_file;
+        struct mm_struct *mm = vma->vm_mm;
         int n;
         char buffer[50];
 
-        if (!vma->vm_mm)
+        if (!mm)
                 return 0;
 
         md = kzalloc(sizeof(struct numa_maps), GFP_KERNEL);
         if (!md)
                 return 0;
 
-        check_pgd_range(vma, vma->vm_start, vma->vm_end,
-                    &node_online_map, MPOL_MF_STATS, md);
+        mpol_to_str(buffer, sizeof(buffer),
+                        get_vma_policy(task, vma, vma->vm_start));
 
-        if (md->pages) {
-                mpol_to_str(buffer, sizeof(buffer),
-                            get_vma_policy(task, vma, vma->vm_start));
+        seq_printf(m, "%08lx %s", vma->vm_start, buffer);
 
-                seq_printf(m, "%08lx %s pages=%lu mapped=%lu maxref=%lu",
-                           vma->vm_start, buffer, md->pages,
-                           md->mapped, md->mapcount_max);
+        if (file) {
+                seq_printf(m, " file=");
+                seq_path(m, file->f_vfsmnt, file->f_dentry, "\n\t= ");
+        } else if (vma->vm_start <= mm->brk && vma->vm_end >= mm->start_brk) {
+                seq_printf(m, " heap");
+        } else if (vma->vm_start <= mm->start_stack &&
+                        vma->vm_end >= mm->start_stack) {
+                seq_printf(m, " stack");
+        }
 
-                if (md->anon)
-                        seq_printf(m," anon=%lu",md->anon);
+        if (is_vm_hugetlb_page(vma)) {
+                check_huge_range(vma, vma->vm_start, vma->vm_end, md);
+                seq_printf(m, " huge");
+        } else {
+                check_pgd_range(vma, vma->vm_start, vma->vm_end,
+                                &node_online_map, MPOL_MF_STATS, md);
+        }
 
-                for_each_online_node(n)
-                        if (md->node[n])
-                                seq_printf(m, " N%d=%lu", n, md->node[n]);
+        if (!md->pages)
+                goto out;
 
-                seq_putc(m, '\n');
-        }
+        if (md->anon)
+                seq_printf(m," anon=%lu",md->anon);
+
+        if (md->dirty)
+                seq_printf(m," dirty=%lu",md->dirty);
+
+        if (md->pages != md->anon && md->pages != md->dirty)
+                seq_printf(m, " mapped=%lu", md->pages);
+
+        if (md->mapcount_max > 1)
+                seq_printf(m, " mapmax=%lu", md->mapcount_max);
+
+        if (md->swapcache)
+                seq_printf(m," swapcache=%lu", md->swapcache);
+
+        if (md->active < md->pages && !is_vm_hugetlb_page(vma))
+                seq_printf(m," active=%lu", md->active);
+
+        if (md->writeback)
+                seq_printf(m," writeback=%lu", md->writeback);
+
+        for_each_online_node(n)
+                if (md->node[n])
+                        seq_printf(m, " N%d=%lu", n, md->node[n]);
+out:
+        seq_putc(m, '\n');
         kfree(md);
 
         if (m->count < m->size)
diff --git a/mm/nommu.c b/mm/nommu.c
index c10262d68232..4951f4786f28 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -53,10 +53,11 @@ DECLARE_RWSEM(nommu_vma_sem);
 struct vm_operations_struct generic_file_vm_ops = {
 };
 
-EXPORT_SYMBOL(vmalloc);
 EXPORT_SYMBOL(vfree);
 EXPORT_SYMBOL(vmalloc_to_page);
 EXPORT_SYMBOL(vmalloc_32);
+EXPORT_SYMBOL(vmap);
+EXPORT_SYMBOL(vunmap);
 
 /*
  * Handle all mappings that got truncated by a "truncate()"
@@ -203,6 +204,13 @@ void *vmalloc(unsigned long size)
 {
        return __vmalloc(size, GFP_KERNEL | __GFP_HIGHMEM, PAGE_KERNEL);
 }
+EXPORT_SYMBOL(vmalloc);
+
+void *vmalloc_node(unsigned long size, int node)
+{
+        return vmalloc(size);
+}
+EXPORT_SYMBOL(vmalloc_node);
 
 /*
  *      vmalloc_32  -  allocate virtually continguos memory (32bit addressable)
diff --git a/mm/oom_kill.c b/mm/oom_kill.c
index b05ab8f2a562..78747afad6b0 100644
--- a/mm/oom_kill.c
+++ b/mm/oom_kill.c
@@ -58,15 +58,17 @@ unsigned long badness(struct task_struct *p, unsigned long uptime)
 
         /*
          * Processes which fork a lot of child processes are likely
-         * a good choice. We add the vmsize of the children if they
+         * a good choice. We add half the vmsize of the children if they
          * have an own mm. This prevents forking servers to flood the
-         * machine with an endless amount of children
+         * machine with an endless amount of children. In case a single
+         * child is eating the vast majority of memory, adding only half
+         * to the parents will make the child our kill candidate of choice.
          */
         list_for_each(tsk, &p->children) {
                 struct task_struct *chld;
                 chld = list_entry(tsk, struct task_struct, sibling);
                 if (chld->mm != p->mm && chld->mm)
-                        points += chld->mm->total_vm;
+                        points += chld->mm->total_vm/2 + 1;
         }
 
         /*
@@ -131,17 +133,47 @@ unsigned long badness(struct task_struct *p, unsigned long uptime)
 }
 
 /*
+ * Types of limitations to the nodes from which allocations may occur
+ */
+#define CONSTRAINT_NONE 1
+#define CONSTRAINT_MEMORY_POLICY 2
+#define CONSTRAINT_CPUSET 3
+
+/*
+ * Determine the type of allocation constraint.
+ */
+static inline int constrained_alloc(struct zonelist *zonelist, gfp_t gfp_mask)
+{
+#ifdef CONFIG_NUMA
+        struct zone **z;
+        nodemask_t nodes = node_online_map;
+
+        for (z = zonelist->zones; *z; z++)
+                if (cpuset_zone_allowed(*z, gfp_mask))
+                        node_clear((*z)->zone_pgdat->node_id,
+                                        nodes);
+                else
+                        return CONSTRAINT_CPUSET;
+
+        if (!nodes_empty(nodes))
+                return CONSTRAINT_MEMORY_POLICY;
+#endif
+
+        return CONSTRAINT_NONE;
+}
+
+/*
  * Simple selection loop. We chose the process with the highest
  * number of 'points'. We expect the caller will lock the tasklist.
  *
  * (not docbooked, we don't want this one cluttering up the manual)
  */
-static struct task_struct * select_bad_process(void)
+static struct task_struct *select_bad_process(unsigned long *ppoints)
 {
-        unsigned long maxpoints = 0;
         struct task_struct *g, *p;
         struct task_struct *chosen = NULL;
         struct timespec uptime;
+        *ppoints = 0;
 
         do_posix_clock_monotonic_gettime(&uptime);
         do_each_thread(g, p) {
@@ -169,9 +201,9 @@ static struct task_struct * select_bad_process(void)
                         return p;
 
                 points = badness(p, uptime.tv_sec);
-                if (points > maxpoints || !chosen) {
+                if (points > *ppoints || !chosen) {
                         chosen = p;
-                        maxpoints = points;
+                        *ppoints = points;
                 }
         } while_each_thread(g, p);
         return chosen;
@@ -182,7 +214,7 @@ static struct task_struct * select_bad_process(void)
  * CAP_SYS_RAW_IO set, send SIGTERM instead (but it's unlikely that
  * we select a process with CAP_SYS_RAW_IO set).
  */
-static void __oom_kill_task(task_t *p)
+static void __oom_kill_task(task_t *p, const char *message)
 {
         if (p->pid == 1) {
                 WARN_ON(1);
@@ -198,8 +230,8 @@ static void __oom_kill_task(task_t *p)
                 return;
         }
         task_unlock(p);
-        printk(KERN_ERR "Out of Memory: Killed process %d (%s).\n",
-                                                        p->pid, p->comm);
+        printk(KERN_ERR "%s: Killed process %d (%s).\n",
+                                message, p->pid, p->comm);
 
         /*
          * We give our sacrificial lamb high priority and access to
@@ -212,7 +244,7 @@ static void __oom_kill_task(task_t *p)
         force_sig(SIGKILL, p);
 }
 
-static struct mm_struct *oom_kill_task(task_t *p)
+static struct mm_struct *oom_kill_task(task_t *p, const char *message)
 {
         struct mm_struct *mm = get_task_mm(p);
         task_t * g, * q;
@@ -224,35 +256,38 @@ static struct mm_struct *oom_kill_task(task_t *p)
                 return NULL;
         }
 
-        __oom_kill_task(p);
+        __oom_kill_task(p, message);
         /*
          * kill all processes that share the ->mm (i.e. all threads),
          * but are in a different thread group
          */
         do_each_thread(g, q)
                 if (q->mm == mm && q->tgid != p->tgid)
-                        __oom_kill_task(q);
+                        __oom_kill_task(q, message);
         while_each_thread(g, q);
 
         return mm;
 }
 
-static struct mm_struct *oom_kill_process(struct task_struct *p)
+static struct mm_struct *oom_kill_process(struct task_struct *p,
+                                unsigned long points, const char *message)
 {
         struct mm_struct *mm;
         struct task_struct *c;
         struct list_head *tsk;
 
+        printk(KERN_ERR "Out of Memory: Kill process %d (%s) score %li and "
+                "children.\n", p->pid, p->comm, points);
         /* Try to kill a child first */
         list_for_each(tsk, &p->children) {
                 c = list_entry(tsk, struct task_struct, sibling);
                 if (c->mm == p->mm)
                         continue;
-                mm = oom_kill_task(c);
+                mm = oom_kill_task(c, message);
                 if (mm)
                         return mm;
         }
-        return oom_kill_task(p);
+        return oom_kill_task(p, message);
 }
 
 /**
@@ -263,10 +298,11 @@ static struct mm_struct *oom_kill_process(struct task_struct *p)
  * OR try to be smart about which process to kill. Note that we
  * don't have to be perfect here, we just have to be good.
  */
-void out_of_memory(gfp_t gfp_mask, int order)
+void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, int order)
 {
         struct mm_struct *mm = NULL;
-        task_t * p;
+        task_t *p;
+        unsigned long points = 0;
 
         if (printk_ratelimit()) {
                 printk("oom-killer: gfp_mask=0x%x, order=%d\n",
@@ -277,24 +313,48 @@ void out_of_memory(gfp_t gfp_mask, int order)
 
         cpuset_lock();
         read_lock(&tasklist_lock);
+
+        /*
+         * Check if there were limitations on the allocation (only relevant for
+         * NUMA) that may require different handling.
+         */
+        switch (constrained_alloc(zonelist, gfp_mask)) {
+        case CONSTRAINT_MEMORY_POLICY:
+                mm = oom_kill_process(current, points,
+                                "No available memory (MPOL_BIND)");
+                break;
+
+        case CONSTRAINT_CPUSET:
+                mm = oom_kill_process(current, points,
+                                "No available memory in cpuset");
+                break;
+
+        case CONSTRAINT_NONE:
 retry:
-        p = select_bad_process();
+                /*
+                 * Rambo mode: Shoot down a process and hope it solves whatever
+                 * issues we may have.
+                 */
+                p = select_bad_process(&points);
 
-        if (PTR_ERR(p) == -1UL)
-                goto out;
+                if (PTR_ERR(p) == -1UL)
+                        goto out;
 
-        /* Found nothing?!?! Either we hang forever, or we panic. */
-        if (!p) {
-                read_unlock(&tasklist_lock);
-                cpuset_unlock();
-                panic("Out of memory and no killable processes...\n");
-        }
+                /* Found nothing?!?! Either we hang forever, or we panic. */
+                if (!p) {
+                        read_unlock(&tasklist_lock);
+                        cpuset_unlock();
+                        panic("Out of memory and no killable processes...\n");
+                }
 
-        mm = oom_kill_process(p);
-        if (!mm)
-                goto retry;
+                mm = oom_kill_process(p, points, "Out of memory");
+                if (!mm)
+                        goto retry;
+
+                break;
+        }
 
- out:
+out:
         read_unlock(&tasklist_lock);
         cpuset_unlock();
         if (mm)
@@ -305,5 +365,5 @@ retry:
          * retry to allocate memory unless "p" is current
          */
         if (!test_thread_flag(TIF_MEMDIE))
-                schedule_timeout_interruptible(1);
+                schedule_timeout_uninterruptible(1);
 }
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index dde04ff4be31..234bd4895d14 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -56,6 +56,7 @@ long nr_swap_pages;
 int percpu_pagelist_fraction;
 
 static void fastcall free_hot_cold_page(struct page *page, int cold);
+static void __free_pages_ok(struct page *page, unsigned int order);
 
 /*
  * results with 256, 32 in the lowmem_reserve sysctl:
@@ -169,20 +170,23 @@ static void bad_page(struct page *page)
  * All pages have PG_compound set.  All pages have their ->private pointing at
  * the head page (even the head page has this).
  *
- * The first tail page's ->mapping, if non-zero, holds the address of the
- * compound page's put_page() function.
- *
- * The order of the allocation is stored in the first tail page's ->index
- * This is only for debug at present.  This usage means that zero-order pages
- * may not be compound.
+ * The first tail page's ->lru.next holds the address of the compound page's
+ * put_page() function.  Its ->lru.prev holds the order of allocation.
+ * This usage means that zero-order pages may not be compound.
  */
+
+static void free_compound_page(struct page *page)
+{
+        __free_pages_ok(page, (unsigned long)page[1].lru.prev);
+}
+
 static void prep_compound_page(struct page *page, unsigned long order)
 {
         int i;
         int nr_pages = 1 << order;
 
-        page[1].mapping = NULL;
-        page[1].index = order;
+        page[1].lru.next = (void *)free_compound_page;  /* set dtor */
+        page[1].lru.prev = (void *)order;
         for (i = 0; i < nr_pages; i++) {
                 struct page *p = page + i;
 
@@ -196,7 +200,7 @@ static void destroy_compound_page(struct page *page, unsigned long order)
         int i;
         int nr_pages = 1 << order;
 
-        if (unlikely(page[1].index != order))
+        if (unlikely((unsigned long)page[1].lru.prev != order))
                 bad_page(page);
 
         for (i = 0; i < nr_pages; i++) {
@@ -586,21 +590,20 @@ static int rmqueue_bulk(struct zone *zone, unsigned int order,
 }
 
 #ifdef CONFIG_NUMA
-/* Called from the slab reaper to drain remote pagesets */
-void drain_remote_pages(void)
+/*
+ * Called from the slab reaper to drain pagesets on a particular node that
+ * belong to the currently executing processor.
+ */
+void drain_node_pages(int nodeid)
 {
-        struct zone *zone;
-        int i;
+        int i, z;
         unsigned long flags;
 
         local_irq_save(flags);
-        for_each_zone(zone) {
+        for (z = 0; z < MAX_NR_ZONES; z++) {
+                struct zone *zone = NODE_DATA(nodeid)->node_zones + z;
                 struct per_cpu_pageset *pset;
 
-                /* Do not drain local pagesets */
-                if (zone->zone_pgdat->node_id == numa_node_id())
-                        continue;
-
                 pset = zone_pcp(zone, smp_processor_id());
                 for (i = 0; i < ARRAY_SIZE(pset->pcp); i++) {
                         struct per_cpu_pages *pcp;
@@ -1011,7 +1014,7 @@ rebalance:
                 if (page)
                         goto got_pg;
 
-                out_of_memory(gfp_mask, order);
+                out_of_memory(zonelist, gfp_mask, order);
                 goto restart;
         }
 
@@ -1537,29 +1540,29 @@ static int __initdata node_load[MAX_NUMNODES];
  */
 static int __init find_next_best_node(int node, nodemask_t *used_node_mask)
 {
-        int i, n, val;
+        int n, val;
         int min_val = INT_MAX;
         int best_node = -1;
 
-        for_each_online_node(i) {
-                cpumask_t tmp;
+        /* Use the local node if we haven't already */
+        if (!node_isset(node, *used_node_mask)) {
+                node_set(node, *used_node_mask);
+                return node;
+        }
 
-                /* Start from local node */
-                n = (node+i) % num_online_nodes();
+        for_each_online_node(n) {
+                cpumask_t tmp;
 
                 /* Don't want a node to appear more than once */
                 if (node_isset(n, *used_node_mask))
                         continue;
 
-                /* Use the local node if we haven't already */
-                if (!node_isset(node, *used_node_mask)) {
-                        best_node = node;
-                        break;
-                }
-
                 /* Use the distance array to find the distance */
                 val = node_distance(node, n);
 
+                /* Penalize nodes under us ("prefer the next node") */
+                val += (n < node);
+
                 /* Give preference to headless and unused nodes */
                 tmp = node_to_cpumask(n);
                 if (!cpus_empty(tmp))
diff --git a/mm/rmap.c b/mm/rmap.c
index df2c41c2a9a2..67f0e20b101f 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -212,25 +212,33 @@ out:
  * through real pte's pointing to valid pages and then releasing
  * the page from the swap cache.
  *
- * Must hold page lock on page.
+ * Must hold page lock on page and mmap_sem of one vma that contains
+ * the page.
  */
 void remove_from_swap(struct page *page)
 {
         struct anon_vma *anon_vma;
         struct vm_area_struct *vma;
+        unsigned long mapping;
 
-        if (!PageAnon(page) || !PageSwapCache(page))
+        if (!PageSwapCache(page))
                 return;
 
-        anon_vma = page_lock_anon_vma(page);
-        if (!anon_vma)
+        mapping = (unsigned long)page->mapping;
+
+        if (!mapping || (mapping & PAGE_MAPPING_ANON) == 0)
                 return;
 
+        /*
+         * We hold the mmap_sem lock. So no need to call page_lock_anon_vma.
+         */
+        anon_vma = (struct anon_vma *) (mapping - PAGE_MAPPING_ANON);
+        spin_lock(&anon_vma->lock);
+
         list_for_each_entry(vma, &anon_vma->head, anon_vma_node)
                 remove_vma_swap(vma, page);
 
         spin_unlock(&anon_vma->lock);
-
         delete_from_swap_cache(page);
 }
 EXPORT_SYMBOL(remove_from_swap);
@@ -529,9 +537,6 @@ void page_add_new_anon_rmap(struct page *page,
  */
 void page_add_file_rmap(struct page *page)
 {
-        BUG_ON(PageAnon(page));
-        BUG_ON(!pfn_valid(page_to_pfn(page)));
-
         if (atomic_inc_and_test(&page->_mapcount))
                 __inc_page_state(nr_mapped);
 }
diff --git a/mm/shmem.c b/mm/shmem.c
index f7ac7b812f92..7c455fbaff7b 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -45,6 +45,7 @@
 #include <linux/swapops.h>
 #include <linux/mempolicy.h>
 #include <linux/namei.h>
+#include <linux/ctype.h>
 #include <asm/uaccess.h>
 #include <asm/div64.h>
 #include <asm/pgtable.h>
@@ -874,6 +875,51 @@ redirty:
 }
 
 #ifdef CONFIG_NUMA
+static int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
+{
+        char *nodelist = strchr(value, ':');
+        int err = 1;
+
+        if (nodelist) {
+                /* NUL-terminate policy string */
+                *nodelist++ = '\0';
+                if (nodelist_parse(nodelist, *policy_nodes))
+                        goto out;
+        }
+        if (!strcmp(value, "default")) {
+                *policy = MPOL_DEFAULT;
+                /* Don't allow a nodelist */
+                if (!nodelist)
+                        err = 0;
+        } else if (!strcmp(value, "prefer")) {
+                *policy = MPOL_PREFERRED;
+                /* Insist on a nodelist of one node only */
+                if (nodelist) {
+                        char *rest = nodelist;
+                        while (isdigit(*rest))
+                                rest++;
+                        if (!*rest)
+                                err = 0;
+                }
+        } else if (!strcmp(value, "bind")) {
+                *policy = MPOL_BIND;
+                /* Insist on a nodelist */
+                if (nodelist)
+                        err = 0;
+        } else if (!strcmp(value, "interleave")) {
+                *policy = MPOL_INTERLEAVE;
+                /* Default to nodes online if no nodelist */
+                if (!nodelist)
+                        *policy_nodes = node_online_map;
+                err = 0;
+        }
+out:
+        /* Restore string for error message */
+        if (nodelist)
+                *--nodelist = ':';
+        return err;
+}
+
 static struct page *shmem_swapin_async(struct shared_policy *p,
                                        swp_entry_t entry, unsigned long idx)
 {
@@ -926,6 +972,11 @@ shmem_alloc_page(gfp_t gfp, struct shmem_inode_info *info,
         return page;
 }
 #else
+static inline int shmem_parse_mpol(char *value, int *policy, nodemask_t *policy_nodes)
+{
+        return 1;
+}
+
 static inline struct page *
 shmem_swapin(struct shmem_inode_info *info,swp_entry_t entry,unsigned long idx)
 {
@@ -1859,7 +1910,23 @@ static int shmem_parse_options(char *options, int *mode, uid_t *uid,
 {
         char *this_char, *value, *rest;
 
-        while ((this_char = strsep(&options, ",")) != NULL) {
+        while (options != NULL) {
+                this_char = options;
+                for (;;) {
+                        /*
+                         * NUL-terminate this option: unfortunately,
+                         * mount options form a comma-separated list,
+                         * but mpol's nodelist may also contain commas.
+                         */
+                        options = strchr(options, ',');
+                        if (options == NULL)
+                                break;
+                        options++;
+                        if (!isdigit(*options)) {
+                                options[-1] = '\0';
+                                break;
+                        }
+                }
                 if (!*this_char)
                         continue;
                 if ((value = strchr(this_char,'=')) != NULL) {
@@ -1910,18 +1977,8 @@ static int shmem_parse_options(char *options, int *mode, uid_t *uid,
                         if (*rest)
                                 goto bad_val;
                 } else if (!strcmp(this_char,"mpol")) {
-                        if (!strcmp(value,"default"))
-                                *policy = MPOL_DEFAULT;
-                        else if (!strcmp(value,"preferred"))
-                                *policy = MPOL_PREFERRED;
-                        else if (!strcmp(value,"bind"))
-                                *policy = MPOL_BIND;
-                        else if (!strcmp(value,"interleave"))
-                                *policy = MPOL_INTERLEAVE;
-                        else
+                        if (shmem_parse_mpol(value,policy,policy_nodes))
                                 goto bad_val;
-                } else if (!strcmp(this_char,"mpol_nodelist")) {
-                        nodelist_parse(value, *policy_nodes);
                 } else {
                         printk(KERN_ERR "tmpfs: Bad mount option %s\n",
                                this_char);
diff --git a/mm/slab.c b/mm/slab.c
index d66c2b0d9715..d0bd7f07ab04 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -789,6 +789,47 @@ static void __slab_error(const char *function, struct kmem_cache *cachep, char *
         dump_stack();
 }
 
+#ifdef CONFIG_NUMA
+/*
+ * Special reaping functions for NUMA systems called from cache_reap().
+ * These take care of doing round robin flushing of alien caches (containing
+ * objects freed on different nodes from which they were allocated) and the
+ * flushing of remote pcps by calling drain_node_pages.
+ */
+static DEFINE_PER_CPU(unsigned long, reap_node);
+
+static void init_reap_node(int cpu)
+{
+        int node;
+
+        node = next_node(cpu_to_node(cpu), node_online_map);
+        if (node == MAX_NUMNODES)
+                node = 0;
+
+        __get_cpu_var(reap_node) = node;
+}
+
+static void next_reap_node(void)
+{
+        int node = __get_cpu_var(reap_node);
+
+        /*
+         * Also drain per cpu pages on remote zones
+         */
+        if (node != numa_node_id())
+                drain_node_pages(node);
+
+        node = next_node(node, node_online_map);
+        if (unlikely(node >= MAX_NUMNODES))
+                node = first_node(node_online_map);
+        __get_cpu_var(reap_node) = node;
+}
+
+#else
+#define init_reap_node(cpu) do { } while (0)
+#define next_reap_node(void) do { } while (0)
+#endif
+
 /*
  * Initiate the reap timer running on the target CPU.  We run at around 1 to 2Hz
  * via the workqueue/eventd.
@@ -806,6 +847,7 @@ static void __devinit start_cpu_timer(int cpu)
          * at that time.
          */
         if (keventd_up() && reap_work->func == NULL) {
+                init_reap_node(cpu);
                 INIT_WORK(reap_work, cache_reap, NULL);
                 schedule_delayed_work_on(cpu, reap_work, HZ + 3 * cpu);
         }
@@ -884,6 +926,23 @@ static void __drain_alien_cache(struct kmem_cache *cachep,
         }
 }
 
+/*
+ * Called from cache_reap() to regularly drain alien caches round robin.
+ */
+static void reap_alien(struct kmem_cache *cachep, struct kmem_list3 *l3)
+{
+        int node = __get_cpu_var(reap_node);
+
+        if (l3->alien) {
+                struct array_cache *ac = l3->alien[node];
+                if (ac && ac->avail) {
+                        spin_lock_irq(&ac->lock);
+                        __drain_alien_cache(cachep, ac, node);
+                        spin_unlock_irq(&ac->lock);
+                }
+        }
+}
+
 static void drain_alien_cache(struct kmem_cache *cachep, struct array_cache **alien)
 {
         int i = 0;
@@ -902,6 +961,7 @@ static void drain_alien_cache(struct kmem_cache *cachep, struct array_cache **al
 #else
 
 #define drain_alien_cache(cachep, alien) do { } while (0)
+#define reap_alien(cachep, l3) do { } while (0)
 
 static inline struct array_cache **alloc_alien_cache(int node, int limit)
 {
@@ -1124,6 +1184,7 @@ void __init kmem_cache_init(void)
         struct cache_sizes *sizes;
         struct cache_names *names;
         int i;
+        int order;
 
         for (i = 0; i < NUM_INIT_LISTS; i++) {
                 kmem_list3_init(&initkmem_list3[i]);
@@ -1167,11 +1228,15 @@ void __init kmem_cache_init(void)
 
         cache_cache.buffer_size = ALIGN(cache_cache.buffer_size, cache_line_size());
 
-        cache_estimate(0, cache_cache.buffer_size, cache_line_size(), 0,
-                       &left_over, &cache_cache.num);
+        for (order = 0; order < MAX_ORDER; order++) {
+                cache_estimate(order, cache_cache.buffer_size,
+                        cache_line_size(), 0, &left_over, &cache_cache.num);
+                if (cache_cache.num)
+                        break;
+        }
         if (!cache_cache.num)
                 BUG();
-
+        cache_cache.gfporder = order;
         cache_cache.colour = left_over / cache_cache.colour_off;
         cache_cache.slab_size = ALIGN(cache_cache.num * sizeof(kmem_bufctl_t) +
                                       sizeof(struct slab), cache_line_size());
@@ -1628,36 +1693,44 @@ static inline size_t calculate_slab_order(struct kmem_cache *cachep,
                         size_t size, size_t align, unsigned long flags)
 {
         size_t left_over = 0;
+        int gfporder;
 
-        for (;; cachep->gfporder++) {
+        for (gfporder = 0 ; gfporder <= MAX_GFP_ORDER; gfporder++) {
                 unsigned int num;
                 size_t remainder;
 
-                if (cachep->gfporder > MAX_GFP_ORDER) {
-                        cachep->num = 0;
-                        break;
-                }
-
-                cache_estimate(cachep->gfporder, size, align, flags,
-                               &remainder, &num);
+                cache_estimate(gfporder, size, align, flags, &remainder, &num);
                 if (!num)
                         continue;
+
                 /* More than offslab_limit objects will cause problems */
-                if (flags & CFLGS_OFF_SLAB && cachep->num > offslab_limit)
+                if ((flags & CFLGS_OFF_SLAB) && num > offslab_limit)
                         break;
 
+                /* Found something acceptable - save it away */
                 cachep->num = num;
+                cachep->gfporder = gfporder;
                 left_over = remainder;
 
                 /*
+                 * A VFS-reclaimable slab tends to have most allocations
+                 * as GFP_NOFS and we really don't want to have to be allocating
+                 * higher-order pages when we are unable to shrink dcache.
+                 */
+                if (flags & SLAB_RECLAIM_ACCOUNT)
+                        break;
+
+                /*
                  * Large number of objects is good, but very large slabs are
                  * currently bad for the gfp()s.
                  */
-                if (cachep->gfporder >= slab_break_gfp_order)
+                if (gfporder >= slab_break_gfp_order)
                         break;
 
-                if ((left_over * 8) <= (PAGE_SIZE << cachep->gfporder))
-                        /* Acceptable internal fragmentation */
+                /*
+                 * Acceptable internal fragmentation?
+                 */
+                if ((left_over * 8) <= (PAGE_SIZE << gfporder))
                         break;
         }
         return left_over;
@@ -1717,6 +1790,12 @@ kmem_cache_create (const char *name, size_t size, size_t align,
                 BUG();
         }
 
+        /*
+         * Prevent CPUs from coming and going.
+         * lock_cpu_hotplug() nests outside cache_chain_mutex
+         */
+        lock_cpu_hotplug();
+
         mutex_lock(&cache_chain_mutex);
 
         list_for_each(p, &cache_chain) {
@@ -1863,17 +1942,7 @@ kmem_cache_create (const char *name, size_t size, size_t align,
 
         size = ALIGN(size, align);
 
-        if ((flags & SLAB_RECLAIM_ACCOUNT) && size <= PAGE_SIZE) {
-                /*
-                 * A VFS-reclaimable slab tends to have most allocations
-                 * as GFP_NOFS and we really don't want to have to be allocating
-                 * higher-order pages when we are unable to shrink dcache.
-                 */
-                cachep->gfporder = 0;
-                cache_estimate(cachep->gfporder, size, align, flags,
-                               &left_over, &cachep->num);
-        } else
-                left_over = calculate_slab_order(cachep, size, align, flags);
+        left_over = calculate_slab_order(cachep, size, align, flags);
 
         if (!cachep->num) {
                 printk("kmem_cache_create: couldn't create cache %s.\n", name);
@@ -1918,8 +1987,6 @@ kmem_cache_create (const char *name, size_t size, size_t align,
         cachep->dtor = dtor;
         cachep->name = name;
 
-        /* Don't let CPUs to come and go */
-        lock_cpu_hotplug();
 
         if (g_cpucache_up == FULL) {
                 enable_cpucache(cachep);
@@ -1978,12 +2045,12 @@ kmem_cache_create (const char *name, size_t size, size_t align,
 
         /* cache setup completed, link it into the list */
         list_add(&cachep->next, &cache_chain);
-        unlock_cpu_hotplug();
       oops:
         if (!cachep && (flags & SLAB_PANIC))
                 panic("kmem_cache_create(): failed to create slab `%s'\n",
                       name);
         mutex_unlock(&cache_chain_mutex);
+        unlock_cpu_hotplug();
         return cachep;
 }
 EXPORT_SYMBOL(kmem_cache_create);
@@ -2550,7 +2617,7 @@ static void check_slabp(struct kmem_cache *cachep, struct slab *slabp)
                        "slab: Internal list corruption detected in cache '%s'(%d), slabp %p(%d). Hexdump:\n",
                        cachep->name, cachep->num, slabp, slabp->inuse);
                 for (i = 0;
-                     i < sizeof(slabp) + cachep->num * sizeof(kmem_bufctl_t);
+                     i < sizeof(*slabp) + cachep->num * sizeof(kmem_bufctl_t);
                      i++) {
                         if ((i % 16) == 0)
                                 printk("\n%03x:", i);
@@ -3490,8 +3557,7 @@ static void cache_reap(void *unused)
                 check_irq_on();
 
                 l3 = searchp->nodelists[numa_node_id()];
-                if (l3->alien)
-                        drain_alien_cache(searchp, l3->alien);
+                reap_alien(searchp, l3);
                 spin_lock_irq(&l3->list_lock);
 
                 drain_array_locked(searchp, cpu_cache_get(searchp), 0,
@@ -3541,7 +3607,7 @@ static void cache_reap(void *unused)
         }
         check_irq_on();
         mutex_unlock(&cache_chain_mutex);
-        drain_remote_pages();
+        next_reap_node();
         /* Setup the next iteration */
         schedule_delayed_work(&__get_cpu_var(reap_work), REAPTIMEOUT_CPUC);
 }
diff --git a/mm/swap.c b/mm/swap.c
index 76247424dea1..e9ec06d845e8 100644
--- a/mm/swap.c
+++ b/mm/swap.c
@@ -40,7 +40,7 @@ static void put_compound_page(struct page *page)
         if (put_page_testzero(page)) {
                 void (*dtor)(struct page *page);
 
-                dtor = (void (*)(struct page *))page[1].mapping;
+                dtor = (void (*)(struct page *))page[1].lru.next;
                 (*dtor)(page);
         }
 }
@@ -489,13 +489,34 @@ void percpu_counter_mod(struct percpu_counter *fbc, long amount)
         if (count >= FBC_BATCH || count <= -FBC_BATCH) {
                 spin_lock(&fbc->lock);
                 fbc->count += count;
+                *pcount = 0;
                 spin_unlock(&fbc->lock);
-                count = 0;
+        } else {
+                *pcount = count;
         }
-        *pcount = count;
         put_cpu();
 }
 EXPORT_SYMBOL(percpu_counter_mod);
+
+/*
+ * Add up all the per-cpu counts, return the result.  This is a more accurate
+ * but much slower version of percpu_counter_read_positive()
+ */
+long percpu_counter_sum(struct percpu_counter *fbc)
+{
+        long ret;
+        int cpu;
+
+        spin_lock(&fbc->lock);
+        ret = fbc->count;
+        for_each_cpu(cpu) {
+                long *pcount = per_cpu_ptr(fbc->counters, cpu);
+                ret += *pcount;
+        }
+        spin_unlock(&fbc->lock);
+        return ret < 0 ? 0 : ret;
+}
+EXPORT_SYMBOL(percpu_counter_sum);
 #endif
 
 /*
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 5a610804cd06..4fe7e3aa02e2 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -443,6 +443,10 @@ static int shrink_list(struct list_head *page_list, struct scan_control *sc)
                 BUG_ON(PageActive(page));
 
                 sc->nr_scanned++;
+
+                if (!sc->may_swap && page_mapped(page))
+                        goto keep_locked;
+
                 /* Double the slab pressure for mapped and swapcache pages */
                 if (page_mapped(page) || PageSwapCache(page))
                         sc->nr_scanned++;
@@ -632,7 +636,7 @@ static int swap_page(struct page *page)
         struct address_space *mapping = page_mapping(page);
 
         if (page_mapped(page) && mapping)
-                if (try_to_unmap(page, 0) != SWAP_SUCCESS)
+                if (try_to_unmap(page, 1) != SWAP_SUCCESS)
                         goto unlock_retry;
 
         if (PageDirty(page)) {
@@ -696,7 +700,7 @@ int migrate_page_remove_references(struct page *newpage,
          * the page.
          */
         if (!mapping || page_mapcount(page) + nr_refs != page_count(page))
-                return 1;
+                return -EAGAIN;
 
         /*
          * Establish swap ptes for anonymous pages or destroy pte
@@ -717,13 +721,15 @@ int migrate_page_remove_references(struct page *newpage,
          * If the page was not migrated then the PageSwapCache bit
          * is still set and the operation may continue.
          */
-        try_to_unmap(page, 1);
+        if (try_to_unmap(page, 1) == SWAP_FAIL)
+                /* A vma has VM_LOCKED set -> Permanent failure */
+                return -EPERM;
 
         /*
          * Give up if we were unable to remove all mappings.
          */
         if (page_mapcount(page))
-                return 1;
+                return -EAGAIN;
 
         write_lock_irq(&mapping->tree_lock);
 
@@ -734,7 +740,7 @@ int migrate_page_remove_references(struct page *newpage,
         if (!page_mapping(page) || page_count(page) != nr_refs ||
                         *radix_pointer != page) {
                 write_unlock_irq(&mapping->tree_lock);
-                return 1;
+                return -EAGAIN;
         }
 
         /*
@@ -809,10 +815,14 @@ EXPORT_SYMBOL(migrate_page_copy);
  */
 int migrate_page(struct page *newpage, struct page *page)
 {
+        int rc;
+
         BUG_ON(PageWriteback(page));    /* Writeback must be complete */
 
-        if (migrate_page_remove_references(newpage, page, 2))
-                return -EAGAIN;
+        rc = migrate_page_remove_references(newpage, page, 2);
+
+        if (rc)
+                return rc;
 
         migrate_page_copy(newpage, page);
 
@@ -839,7 +849,7 @@ EXPORT_SYMBOL(migrate_page);
  * pages are swapped out.
  *
  * The function returns after 10 attempts or if no pages
- * are movable anymore because t has become empty
+ * are movable anymore because to has become empty
  * or no retryable pages exist anymore.
  *
  * Return: Number of pages not migrated when "to" ran empty.
@@ -928,12 +938,21 @@ redo:
                         goto unlock_both;
 
                 if (mapping->a_ops->migratepage) {
+                        /*
+                         * Most pages have a mapping and most filesystems
+                         * should provide a migration function. Anonymous
+                         * pages are part of swap space which also has its
+                         * own migration function. This is the most common
+                         * path for page migration.
+                         */
                         rc = mapping->a_ops->migratepage(newpage, page);
                         goto unlock_both;
                 }
 
                 /*
-                 * Trigger writeout if page is dirty
+                 * Default handling if a filesystem does not provide
+                 * a migration function. We can only migrate clean
+                 * pages so try to write out any dirty pages first.
                  */
                 if (PageDirty(page)) {
                         switch (pageout(page, mapping)) {
@@ -949,9 +968,10 @@ redo:
                                 ; /* try to migrate the page below */
                         }
                 }
+
                 /*
-                 * If we have no buffer or can release the buffer
-                 * then do a simple migration.
+                 * Buffers are managed in a filesystem specific way.
+                 * We must have no buffers or drop them.
                  */
                 if (!page_has_buffers(page) ||
                     try_to_release_page(page, GFP_KERNEL)) {
@@ -966,6 +986,11 @@ redo:
                  * swap them out.
                  */
                 if (pass > 4) {
+                        /*
+                         * Persistently unable to drop buffers..... As a
+                         * measure of last resort we fall back to
+                         * swap_page().
+                         */
                         unlock_page(newpage);
                         newpage = NULL;
                         rc = swap_page(page);
@@ -1176,9 +1201,47 @@ refill_inactive_zone(struct zone *zone, struct scan_control *sc)
         struct page *page;
         struct pagevec pvec;
         int reclaim_mapped = 0;
-        long mapped_ratio;
-        long distress;
-        long swap_tendency;
+
+        if (unlikely(sc->may_swap)) {
+                long mapped_ratio;
+                long distress;
+                long swap_tendency;
+
+                /*
+                 * `distress' is a measure of how much trouble we're having
+                 * reclaiming pages.  0 -> no problems.  100 -> great trouble.
+                 */
+                distress = 100 >> zone->prev_priority;
+
+                /*
+                 * The point of this algorithm is to decide when to start
+                 * reclaiming mapped memory instead of just pagecache.  Work out
+                 * how much memory
+                 * is mapped.
+                 */
+                mapped_ratio = (sc->nr_mapped * 100) / total_memory;
+
+                /*
+                 * Now decide how much we really want to unmap some pages.  The
+                 * mapped ratio is downgraded - just because there's a lot of
+                 * mapped memory doesn't necessarily mean that page reclaim
+                 * isn't succeeding.
+                 *
+                 * The distress ratio is important - we don't want to start
+                 * going oom.
+                 *
+                 * A 100% value of vm_swappiness overrides this algorithm
+                 * altogether.
+                 */
+                swap_tendency = mapped_ratio / 2 + distress + vm_swappiness;
+
+                /*
+                 * Now use this metric to decide whether to start moving mapped
+                 * memory onto the inactive list.
+                 */
+                if (swap_tendency >= 100)
+                        reclaim_mapped = 1;
+        }
 
         lru_add_drain();
         spin_lock_irq(&zone->lru_lock);
@@ -1188,37 +1251,6 @@ refill_inactive_zone(struct zone *zone, struct scan_control *sc)
         zone->nr_active -= pgmoved;
         spin_unlock_irq(&zone->lru_lock);
 
-        /*
-         * `distress' is a measure of how much trouble we're having reclaiming
-         * pages.  0 -> no problems.  100 -> great trouble.
-         */
-        distress = 100 >> zone->prev_priority;
-
-        /*
-         * The point of this algorithm is to decide when to start reclaiming
-         * mapped memory instead of just pagecache.  Work out how much memory
-         * is mapped.
-         */
-        mapped_ratio = (sc->nr_mapped * 100) / total_memory;
-
-        /*
-         * Now decide how much we really want to unmap some pages.  The mapped
-         * ratio is downgraded - just because there's a lot of mapped memory
-         * doesn't necessarily mean that page reclaim isn't succeeding.
-         *
-         * The distress ratio is important - we don't want to start going oom.
-         *
-         * A 100% value of vm_swappiness overrides this algorithm altogether.
-         */
-        swap_tendency = mapped_ratio / 2 + distress + vm_swappiness;
-
-        /*
-         * Now use this metric to decide whether to start moving mapped memory
-         * onto the inactive list.
-         */
-        if (swap_tendency >= 100)
-                reclaim_mapped = 1;
-
         while (!list_empty(&l_hold)) {
                 cond_resched();
                 page = lru_to_page(&l_hold);
@@ -1595,9 +1627,7 @@ scan:
                         sc.nr_reclaimed = 0;
                         sc.priority = priority;
                         sc.swap_cluster_max = nr_pages? nr_pages : SWAP_CLUSTER_MAX;
-                        atomic_inc(&zone->reclaim_in_progress);
                         shrink_zone(zone, &sc);
-                        atomic_dec(&zone->reclaim_in_progress);
                         reclaim_state->reclaimed_slab = 0;
                         nr_slab = shrink_slab(sc.nr_scanned, GFP_KERNEL,
                                                 lru_pages);
@@ -1859,7 +1889,8 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
 
         if (!(gfp_mask & __GFP_WAIT) ||
                 zone->all_unreclaimable ||
-                atomic_read(&zone->reclaim_in_progress) > 0)
+                atomic_read(&zone->reclaim_in_progress) > 0 ||
+                (p->flags & PF_MEMALLOC))
                         return 0;
 
         node_id = zone->zone_pgdat->node_id;
@@ -1884,7 +1915,12 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
                 sc.swap_cluster_max = SWAP_CLUSTER_MAX;
 
         cond_resched();
-        p->flags |= PF_MEMALLOC;
+        /*
+         * We need to be able to allocate from the reserves for RECLAIM_SWAP
+         * and we also need to be able to write out pages for RECLAIM_WRITE
+         * and RECLAIM_SWAP.
+         */
+        p->flags |= PF_MEMALLOC | PF_SWAPWRITE;
         reclaim_state.reclaimed_slab = 0;
         p->reclaim_state = &reclaim_state;
 
@@ -1908,11 +1944,10 @@ int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order)
                  * a long time.
                  */
                 shrink_slab(sc.nr_scanned, gfp_mask, order);
-                sc.nr_reclaimed = 1;    /* Avoid getting the off node timeout */
         }
 
         p->reclaim_state = NULL;
-        current->flags &= ~PF_MEMALLOC;
+        current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE);
 
         if (sc.nr_reclaimed == 0)
                 zone->last_unsuccessful_zone_reclaim = jiffies;