Merge branch 'master' into for-2.6.34

12 files changed, 287 insertions, 131 deletions

diff --git a/mm/filemap.c b/mm/filemap.c
index 96ac6b0eb6c..698ea80f210 100644
--- a/mm/filemap.c
+++ b/mm/filemap.c
@@ -1634,14 +1634,15 @@ EXPORT_SYMBOL(generic_file_readonly_mmap);
 static struct page *__read_cache_page(struct address_space *mapping,
 				pgoff_t index,
 				int (*filler)(void *,struct page*),
-				void *data)
+				void *data,
+				gfp_t gfp)
 {
 	struct page *page;
 	int err;
 repeat:
 	page = find_get_page(mapping, index);
 	if (!page) {
-		page = page_cache_alloc_cold(mapping);
+		page = __page_cache_alloc(gfp | __GFP_COLD);
 		if (!page)
 			return ERR_PTR(-ENOMEM);
 		err = add_to_page_cache_lru(page, mapping, index, GFP_KERNEL);
@@ -1661,31 +1662,18 @@ repeat:
 	return page;
 }
 
-/**
- * read_cache_page_async - read into page cache, fill it if needed
- * @mapping:	the page's address_space
- * @index:	the page index
- * @filler:	function to perform the read
- * @data:	destination for read data
- *
- * Same as read_cache_page, but don't wait for page to become unlocked
- * after submitting it to the filler.
- *
- * Read into the page cache. If a page already exists, and PageUptodate() is
- * not set, try to fill the page but don't wait for it to become unlocked.
- *
- * If the page does not get brought uptodate, return -EIO.
- */
-struct page *read_cache_page_async(struct address_space *mapping,
+static struct page *do_read_cache_page(struct address_space *mapping,
 				pgoff_t index,
 				int (*filler)(void *,struct page*),
-				void *data)
+				void *data,
+				gfp_t gfp)
+
 {
 	struct page *page;
 	int err;
 
 retry:
-	page = __read_cache_page(mapping, index, filler, data);
+	page = __read_cache_page(mapping, index, filler, data, gfp);
 	if (IS_ERR(page))
 		return page;
 	if (PageUptodate(page))
@@ -1710,8 +1698,67 @@ out:
 	mark_page_accessed(page);
 	return page;
 }
+
+/**
+ * read_cache_page_async - read into page cache, fill it if needed
+ * @mapping:	the page's address_space
+ * @index:	the page index
+ * @filler:	function to perform the read
+ * @data:	destination for read data
+ *
+ * Same as read_cache_page, but don't wait for page to become unlocked
+ * after submitting it to the filler.
+ *
+ * Read into the page cache. If a page already exists, and PageUptodate() is
+ * not set, try to fill the page but don't wait for it to become unlocked.
+ *
+ * If the page does not get brought uptodate, return -EIO.
+ */
+struct page *read_cache_page_async(struct address_space *mapping,
+				pgoff_t index,
+				int (*filler)(void *,struct page*),
+				void *data)
+{
+	return do_read_cache_page(mapping, index, filler, data, mapping_gfp_mask(mapping));
+}
 EXPORT_SYMBOL(read_cache_page_async);
 
+static struct page *wait_on_page_read(struct page *page)
+{
+	if (!IS_ERR(page)) {
+		wait_on_page_locked(page);
+		if (!PageUptodate(page)) {
+			page_cache_release(page);
+			page = ERR_PTR(-EIO);
+		}
+	}
+	return page;
+}
+
+/**
+ * read_cache_page_gfp - read into page cache, using specified page allocation flags.
+ * @mapping:	the page's address_space
+ * @index:	the page index
+ * @gfp:	the page allocator flags to use if allocating
+ *
+ * This is the same as "read_mapping_page(mapping, index, NULL)", but with
+ * any new page allocations done using the specified allocation flags. Note
+ * that the Radix tree operations will still use GFP_KERNEL, so you can't
+ * expect to do this atomically or anything like that - but you can pass in
+ * other page requirements.
+ *
+ * If the page does not get brought uptodate, return -EIO.
+ */
+struct page *read_cache_page_gfp(struct address_space *mapping,
+				pgoff_t index,
+				gfp_t gfp)
+{
+	filler_t *filler = (filler_t *)mapping->a_ops->readpage;
+
+	return wait_on_page_read(do_read_cache_page(mapping, index, filler, NULL, gfp));
+}
+EXPORT_SYMBOL(read_cache_page_gfp);
+
 /**
  * read_cache_page - read into page cache, fill it if needed
  * @mapping:	the page's address_space
@@ -1729,18 +1776,7 @@ struct page *read_cache_page(struct address_space *mapping,
 				int (*filler)(void *,struct page*),
 				void *data)
 {
-	struct page *page;
-
-	page = read_cache_page_async(mapping, index, filler, data);
-	if (IS_ERR(page))
-		goto out;
-	wait_on_page_locked(page);
-	if (!PageUptodate(page)) {
-		page_cache_release(page);
-		page = ERR_PTR(-EIO);
-	}
- out:
-	return page;
+	return wait_on_page_read(read_cache_page_async(mapping, index, filler, data));
 }
 EXPORT_SYMBOL(read_cache_page);
 
@@ -2196,6 +2232,9 @@ again:
 		if (unlikely(status))
 			break;
 
+		if (mapping_writably_mapped(mapping))
+			flush_dcache_page(page);
+
 		pagefault_disable();
 		copied = iov_iter_copy_from_user_atomic(page, i, offset, bytes);
 		pagefault_enable();
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index 65f38c21820..2d16fa6b8c2 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -402,7 +402,7 @@ static void clear_huge_page(struct page *page,
 {
 	int i;
 
-	if (unlikely(sz > MAX_ORDER_NR_PAGES)) {
+	if (unlikely(sz/PAGE_SIZE > MAX_ORDER_NR_PAGES)) {
 		clear_gigantic_page(page, addr, sz);
 		return;
 	}
@@ -1515,10 +1515,9 @@ static struct attribute_group hstate_attr_group = {
 	.attrs = hstate_attrs,
 };
 
-static int __init hugetlb_sysfs_add_hstate(struct hstate *h,
-				struct kobject *parent,
-				struct kobject **hstate_kobjs,
-				struct attribute_group *hstate_attr_group)
+static int hugetlb_sysfs_add_hstate(struct hstate *h, struct kobject *parent,
+				    struct kobject **hstate_kobjs,
+				    struct attribute_group *hstate_attr_group)
 {
 	int retval;
 	int hi = h - hstates;
diff --git a/mm/maccess.c b/mm/maccess.c
index 9073695ff25..4e348dbaecd 100644
--- a/mm/maccess.c
+++ b/mm/maccess.c
@@ -14,7 +14,11 @@
  * Safely read from address @src to the buffer at @dst.  If a kernel fault
  * happens, handle that and return -EFAULT.
  */
-long probe_kernel_read(void *dst, void *src, size_t size)
+
+long __weak probe_kernel_read(void *dst, void *src, size_t size)
+    __attribute__((alias("__probe_kernel_read")));
+
+long __probe_kernel_read(void *dst, void *src, size_t size)
 {
 	long ret;
 	mm_segment_t old_fs = get_fs();
@@ -39,7 +43,10 @@ EXPORT_SYMBOL_GPL(probe_kernel_read);
  * Safely write to address @dst from the buffer at @src.  If a kernel fault
  * happens, handle that and return -EFAULT.
  */
-long notrace __weak probe_kernel_write(void *dst, void *src, size_t size)
+long __weak probe_kernel_write(void *dst, void *src, size_t size)
+    __attribute__((alias("__probe_kernel_write")));
+
+long __probe_kernel_write(void *dst, void *src, size_t size)
 {
 	long ret;
 	mm_segment_t old_fs = get_fs();
diff --git a/mm/memcontrol.c b/mm/memcontrol.c
index 488b644e0e8..954032b80be 100644
--- a/mm/memcontrol.c
+++ b/mm/memcontrol.c
@@ -2586,7 +2586,7 @@ static int mem_cgroup_force_empty(struct mem_cgroup *mem, bool free_all)
 	if (free_all)
 		goto try_to_free;
 move_account:
-	while (mem->res.usage > 0) {
+	do {
 		ret = -EBUSY;
 		if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children))
 			goto out;
@@ -2614,8 +2614,8 @@ move_account:
 		if (ret == -ENOMEM)
 			goto try_to_free;
 		cond_resched();
-	}
-	ret = 0;
+	/* "ret" should also be checked to ensure all lists are empty. */
+	} while (mem->res.usage > 0 || ret);
 out:
 	css_put(&mem->css);
 	return ret;
@@ -2648,10 +2648,7 @@ try_to_free:
 	}
 	lru_add_drain();
 	/* try move_account...there may be some *locked* pages. */
-	if (mem->res.usage)
-		goto move_account;
-	ret = 0;
-	goto out;
+	goto move_account;
 }
 
 int mem_cgroup_force_empty_write(struct cgroup *cont, unsigned int event)
diff --git a/mm/migrate.c b/mm/migrate.c
index efddbf0926b..9a0db5bbabe 100644
--- a/mm/migrate.c
+++ b/mm/migrate.c
@@ -912,6 +912,9 @@ static int do_pages_move(struct mm_struct *mm, struct task_struct *task,
 				goto out_pm;
 
 			err = -ENODEV;
+			if (node < 0 || node >= MAX_NUMNODES)
+				goto out_pm;
+
 			if (!node_state(node, N_HIGH_MEMORY))
 				goto out_pm;
 
diff --git a/mm/nommu.c b/mm/nommu.c
index 6f9248f89bd..48a2ecfaf05 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -432,6 +432,7 @@ SYSCALL_DEFINE1(brk, unsigned long, brk)
 	/*
 	 * Ok, looks good - let it rip.
 	 */
+	flush_icache_range(mm->brk, brk);
 	return mm->brk = brk;
 }
 
@@ -551,11 +552,11 @@ static void free_page_series(unsigned long from, unsigned long to)
 static void __put_nommu_region(struct vm_region *region)
 	__releases(nommu_region_sem)
 {
-	kenter("%p{%d}", region, atomic_read(&region->vm_usage));
+	kenter("%p{%d}", region, region->vm_usage);
 
 	BUG_ON(!nommu_region_tree.rb_node);
 
-	if (atomic_dec_and_test(&region->vm_usage)) {
+	if (--region->vm_usage == 0) {
 		if (region->vm_top > region->vm_start)
 			delete_nommu_region(region);
 		up_write(&nommu_region_sem);
@@ -1204,7 +1205,7 @@ unsigned long do_mmap_pgoff(struct file *file,
 	if (!vma)
 		goto error_getting_vma;
 
-	atomic_set(&region->vm_usage, 1);
+	region->vm_usage = 1;
 	region->vm_flags = vm_flags;
 	region->vm_pgoff = pgoff;
 
@@ -1271,7 +1272,7 @@ unsigned long do_mmap_pgoff(struct file *file,
 			}
 
 			/* we've found a region we can share */
-			atomic_inc(&pregion->vm_usage);
+			pregion->vm_usage++;
 			vma->vm_region = pregion;
 			start = pregion->vm_start;
 			start += (pgoff - pregion->vm_pgoff) << PAGE_SHIFT;
@@ -1288,7 +1289,7 @@ unsigned long do_mmap_pgoff(struct file *file,
 					vma->vm_region = NULL;
 					vma->vm_start = 0;
 					vma->vm_end = 0;
-					atomic_dec(&pregion->vm_usage);
+					pregion->vm_usage--;
 					pregion = NULL;
 					goto error_just_free;
 				}
@@ -1353,10 +1354,14 @@ unsigned long do_mmap_pgoff(struct file *file,
 share:
 	add_vma_to_mm(current->mm, vma);
 
-	up_write(&nommu_region_sem);
+	/* we flush the region from the icache only when the first executable
+	 * mapping of it is made  */
+	if (vma->vm_flags & VM_EXEC && !region->vm_icache_flushed) {
+		flush_icache_range(region->vm_start, region->vm_end);
+		region->vm_icache_flushed = true;
+	}
 
-	if (prot & PROT_EXEC)
-		flush_icache_range(result, result + len);
+	up_write(&nommu_region_sem);
 
 	kleave(" = %lx", result);
 	return result;
@@ -1436,10 +1441,9 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
 
 	kenter("");
 
-	/* we're only permitted to split anonymous regions that have a single
-	 * owner */
-	if (vma->vm_file ||
-	    atomic_read(&vma->vm_region->vm_usage) != 1)
+	/* we're only permitted to split anonymous regions (these should have
+	 * only a single usage on the region) */
+	if (vma->vm_file)
 		return -ENOMEM;
 
 	if (mm->map_count >= sysctl_max_map_count)
@@ -1513,7 +1517,7 @@ static int shrink_vma(struct mm_struct *mm,
 
 	/* cut the backing region down to size */
 	region = vma->vm_region;
-	BUG_ON(atomic_read(&region->vm_usage) != 1);
+	BUG_ON(region->vm_usage != 1);
 
 	down_write(&nommu_region_sem);
 	delete_nommu_region(region);
@@ -1757,27 +1761,6 @@ void unmap_mapping_range(struct address_space *mapping,
 EXPORT_SYMBOL(unmap_mapping_range);
 
 /*
- * ask for an unmapped area at which to create a mapping on a file
- */
-unsigned long get_unmapped_area(struct file *file, unsigned long addr,
-				unsigned long len, unsigned long pgoff,
-				unsigned long flags)
-{
-	unsigned long (*get_area)(struct file *, unsigned long, unsigned long,
-				  unsigned long, unsigned long);
-
-	get_area = current->mm->get_unmapped_area;
-	if (file && file->f_op && file->f_op->get_unmapped_area)
-		get_area = file->f_op->get_unmapped_area;
-
-	if (!get_area)
-		return -ENOSYS;
-
-	return get_area(file, addr, len, pgoff, flags);
-}
-EXPORT_SYMBOL(get_unmapped_area);
-
-/*
  * Check that a process has enough memory to allocate a new virtual
  * mapping. 0 means there is enough memory for the allocation to
  * succeed and -ENOMEM implies there is not.
@@ -1916,9 +1899,11 @@ int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, in
 
 		/* only read or write mappings where it is permitted */
 		if (write && vma->vm_flags & VM_MAYWRITE)
-			len -= copy_to_user((void *) addr, buf, len);
+			copy_to_user_page(vma, NULL, addr,
+					 (void *) addr, buf, len);
 		else if (!write && vma->vm_flags & VM_MAYREAD)
-			len -= copy_from_user(buf, (void *) addr, len);
+			copy_from_user_page(vma, NULL, addr,
+					    buf, (void *) addr, len);
 		else
 			len = 0;
 	} else {
@@ -1929,3 +1914,65 @@ int access_process_vm(struct task_struct *tsk, unsigned long addr, void *buf, in
 	mmput(mm);
 	return len;
 }
+
+/**
+ * nommu_shrink_inode_mappings - Shrink the shared mappings on an inode
+ * @inode: The inode to check
+ * @size: The current filesize of the inode
+ * @newsize: The proposed filesize of the inode
+ *
+ * Check the shared mappings on an inode on behalf of a shrinking truncate to
+ * make sure that that any outstanding VMAs aren't broken and then shrink the
+ * vm_regions that extend that beyond so that do_mmap_pgoff() doesn't
+ * automatically grant mappings that are too large.
+ */
+int nommu_shrink_inode_mappings(struct inode *inode, size_t size,
+				size_t newsize)
+{
+	struct vm_area_struct *vma;
+	struct prio_tree_iter iter;
+	struct vm_region *region;
+	pgoff_t low, high;
+	size_t r_size, r_top;
+
+	low = newsize >> PAGE_SHIFT;
+	high = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
+
+	down_write(&nommu_region_sem);
+
+	/* search for VMAs that fall within the dead zone */
+	vma_prio_tree_foreach(vma, &iter, &inode->i_mapping->i_mmap,
+			      low, high) {
+		/* found one - only interested if it's shared out of the page
+		 * cache */
+		if (vma->vm_flags & VM_SHARED) {
+			up_write(&nommu_region_sem);
+			return -ETXTBSY; /* not quite true, but near enough */
+		}
+	}
+
+	/* reduce any regions that overlap the dead zone - if in existence,
+	 * these will be pointed to by VMAs that don't overlap the dead zone
+	 *
+	 * we don't check for any regions that start beyond the EOF as there
+	 * shouldn't be any
+	 */
+	vma_prio_tree_foreach(vma, &iter, &inode->i_mapping->i_mmap,
+			      0, ULONG_MAX) {
+		if (!(vma->vm_flags & VM_SHARED))
+			continue;
+
+		region = vma->vm_region;
+		r_size = region->vm_top - region->vm_start;
+		r_top = (region->vm_pgoff << PAGE_SHIFT) + r_size;
+
+		if (r_top > newsize) {
+			region->vm_top -= r_top - newsize;
+			if (region->vm_end > region->vm_top)
+				region->vm_end = region->vm_top;
+		}
+	}
+
+	up_write(&nommu_region_sem);
+	return 0;
+}
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 4e9f5cc5fb5..8deb9d0fd5b 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -556,8 +556,9 @@ static void free_pcppages_bulk(struct zone *zone, int count,
 			page = list_entry(list->prev, struct page, lru);
 			/* must delete as __free_one_page list manipulates */
 			list_del(&page->lru);
-			__free_one_page(page, zone, 0, migratetype);
-			trace_mm_page_pcpu_drain(page, 0, migratetype);
+			/* MIGRATE_MOVABLE list may include MIGRATE_RESERVEs */
+			__free_one_page(page, zone, 0, page_private(page));
+			trace_mm_page_pcpu_drain(page, 0, page_private(page));
 		} while (--count && --batch_free && !list_empty(list));
 	}
 	spin_unlock(&zone->lock);
@@ -1222,10 +1223,10 @@ again:
 		}
 		spin_lock_irqsave(&zone->lock, flags);
 		page = __rmqueue(zone, order, migratetype);
-		__mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order));
 		spin_unlock(&zone->lock);
 		if (!page)
 			goto failed;
+		__mod_zone_page_state(zone, NR_FREE_PAGES, -(1 << order));
 	}
 
 	__count_zone_vm_events(PGALLOC, zone, 1 << order);
@@ -3998,7 +3999,7 @@ void __init add_active_range(unsigned int nid, unsigned long start_pfn,
 		}
 
 		/* Merge backward if suitable */
-		if (start_pfn < early_node_map[i].end_pfn &&
+		if (start_pfn < early_node_map[i].start_pfn &&
 				end_pfn >= early_node_map[i].start_pfn) {
 			early_node_map[i].start_pfn = start_pfn;
 			return;
diff --git a/mm/percpu.c b/mm/percpu.c
index 442010cc91c..083e7c91e5f 100644
--- a/mm/percpu.c
+++ b/mm/percpu.c
@@ -1271,7 +1271,7 @@ static void pcpu_reclaim(struct work_struct *work)
  */
 void free_percpu(void *ptr)
 {
-	void *addr = __pcpu_ptr_to_addr(ptr);
+	void *addr;
 	struct pcpu_chunk *chunk;
 	unsigned long flags;
 	int off;
@@ -1279,6 +1279,8 @@ void free_percpu(void *ptr)
 	if (!ptr)
 		return;
 
+	addr = __pcpu_ptr_to_addr(ptr);
+
 	spin_lock_irqsave(&pcpu_lock, flags);
 
 	chunk = pcpu_chunk_addr_search(addr);
diff --git a/mm/truncate.c b/mm/truncate.c
index 342deee2268..e87e3724482 100644
--- a/mm/truncate.c
+++ b/mm/truncate.c
@@ -522,22 +522,20 @@ EXPORT_SYMBOL_GPL(invalidate_inode_pages2);
  */
 void truncate_pagecache(struct inode *inode, loff_t old, loff_t new)
 {
-	if (new < old) {
-		struct address_space *mapping = inode->i_mapping;
-
-		/*
-		 * unmap_mapping_range is called twice, first simply for
-		 * efficiency so that truncate_inode_pages does fewer
-		 * single-page unmaps.  However after this first call, and
-		 * before truncate_inode_pages finishes, it is possible for
-		 * private pages to be COWed, which remain after
-		 * truncate_inode_pages finishes, hence the second
-		 * unmap_mapping_range call must be made for correctness.
-		 */
-		unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
-		truncate_inode_pages(mapping, new);
-		unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
-	}
+	struct address_space *mapping = inode->i_mapping;
+
+	/*
+	 * unmap_mapping_range is called twice, first simply for
+	 * efficiency so that truncate_inode_pages does fewer
+	 * single-page unmaps.  However after this first call, and
+	 * before truncate_inode_pages finishes, it is possible for
+	 * private pages to be COWed, which remain after
+	 * truncate_inode_pages finishes, hence the second
+	 * unmap_mapping_range call must be made for correctness.
+	 */
+	unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
+	truncate_inode_pages(mapping, new);
+	unmap_mapping_range(mapping, new + PAGE_SIZE - 1, 0, 1);
 }
 EXPORT_SYMBOL(truncate_pagecache);
 
diff --git a/mm/util.c b/mm/util.c
index 7c35ad95f92..834db7be240 100644
--- a/mm/util.c
+++ b/mm/util.c
@@ -220,7 +220,7 @@ char *strndup_user(const char __user *s, long n)
 }
 EXPORT_SYMBOL(strndup_user);
 
-#ifndef HAVE_ARCH_PICK_MMAP_LAYOUT
+#if defined(CONFIG_MMU) && !defined(HAVE_ARCH_PICK_MMAP_LAYOUT)
 void arch_pick_mmap_layout(struct mm_struct *mm)
 {
 	mm->mmap_base = TASK_UNMAPPED_BASE;
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 37e69295f25..ae007462b7f 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -509,6 +509,9 @@ static unsigned long lazy_max_pages(void)
 
 static atomic_t vmap_lazy_nr = ATOMIC_INIT(0);
 
+/* for per-CPU blocks */
+static void purge_fragmented_blocks_allcpus(void);
+
 /*
  * Purges all lazily-freed vmap areas.
  *
@@ -539,6 +542,9 @@ static void __purge_vmap_area_lazy(unsigned long *start, unsigned long *end,
 	} else
 		spin_lock(&purge_lock);
 
+	if (sync)
+		purge_fragmented_blocks_allcpus();
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(va, &vmap_area_list, list) {
 		if (va->flags & VM_LAZY_FREE) {
@@ -555,10 +561,8 @@ static void __purge_vmap_area_lazy(unsigned long *start, unsigned long *end,
 	}
 	rcu_read_unlock();
 
-	if (nr) {
-		BUG_ON(nr > atomic_read(&vmap_lazy_nr));
+	if (nr)
 		atomic_sub(nr, &vmap_lazy_nr);
-	}
 
 	if (nr || force_flush)
 		flush_tlb_kernel_range(*start, *end);
@@ -669,8 +673,6 @@ static bool vmap_initialized __read_mostly = false;
 struct vmap_block_queue {
 	spinlock_t lock;
 	struct list_head free;
-	struct list_head dirty;
-	unsigned int nr_dirty;
 };
 
 struct vmap_block {
@@ -680,10 +682,9 @@ struct vmap_block {
 	unsigned long free, dirty;
 	DECLARE_BITMAP(alloc_map, VMAP_BBMAP_BITS);
 	DECLARE_BITMAP(dirty_map, VMAP_BBMAP_BITS);
-	union {
-		struct list_head free_list;
-		struct rcu_head rcu_head;
-	};
+	struct list_head free_list;
+	struct rcu_head rcu_head;
+	struct list_head purge;
 };
 
 /* Queue of free and dirty vmap blocks, for allocation and flushing purposes */
@@ -759,7 +760,7 @@ static struct vmap_block *new_vmap_block(gfp_t gfp_mask)
 	vbq = &get_cpu_var(vmap_block_queue);
 	vb->vbq = vbq;
 	spin_lock(&vbq->lock);
-	list_add(&vb->free_list, &vbq->free);
+	list_add_rcu(&vb->free_list, &vbq->free);
 	spin_unlock(&vbq->lock);
 	put_cpu_var(vmap_block_queue);
 
@@ -778,8 +779,6 @@ static void free_vmap_block(struct vmap_block *vb)
 	struct vmap_block *tmp;
 	unsigned long vb_idx;
 
-	BUG_ON(!list_empty(&vb->free_list));
-
 	vb_idx = addr_to_vb_idx(vb->va->va_start);
 	spin_lock(&vmap_block_tree_lock);
 	tmp = radix_tree_delete(&vmap_block_tree, vb_idx);
@@ -790,12 +789,61 @@ static void free_vmap_block(struct vmap_block *vb)
 	call_rcu(&vb->rcu_head, rcu_free_vb);
 }
 
+static void purge_fragmented_blocks(int cpu)
+{
+	LIST_HEAD(purge);
+	struct vmap_block *vb;
+	struct vmap_block *n_vb;
+	struct vmap_block_queue *vbq = &per_cpu(vmap_block_queue, cpu);
+
+	rcu_read_lock();
+	list_for_each_entry_rcu(vb, &vbq->free, free_list) {
+
+		if (!(vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS))
+			continue;
+
+		spin_lock(&vb->lock);
+		if (vb->free + vb->dirty == VMAP_BBMAP_BITS && vb->dirty != VMAP_BBMAP_BITS) {
+			vb->free = 0; /* prevent further allocs after releasing lock */
+			vb->dirty = VMAP_BBMAP_BITS; /* prevent purging it again */
+			bitmap_fill(vb->alloc_map, VMAP_BBMAP_BITS);
+			bitmap_fill(vb->dirty_map, VMAP_BBMAP_BITS);
+			spin_lock(&vbq->lock);
+			list_del_rcu(&vb->free_list);
+			spin_unlock(&vbq->lock);
+			spin_unlock(&vb->lock);
+			list_add_tail(&vb->purge, &purge);
+		} else
+			spin_unlock(&vb->lock);
+	}
+	rcu_read_unlock();
+
+	list_for_each_entry_safe(vb, n_vb, &purge, purge) {
+		list_del(&vb->purge);
+		free_vmap_block(vb);
+	}
+}
+
+static void purge_fragmented_blocks_thiscpu(void)
+{
+	purge_fragmented_blocks(smp_processor_id());
+}
+
+static void purge_fragmented_blocks_allcpus(void)
+{
+	int cpu;
+
+	for_each_possible_cpu(cpu)
+		purge_fragmented_blocks(cpu);
+}
+
 static void *vb_alloc(unsigned long size, gfp_t gfp_mask)
 {
 	struct vmap_block_queue *vbq;
 	struct vmap_block *vb;
 	unsigned long addr = 0;
 	unsigned int order;
+	int purge = 0;
 
 	BUG_ON(size & ~PAGE_MASK);
 	BUG_ON(size > PAGE_SIZE*VMAP_MAX_ALLOC);
@@ -808,24 +856,38 @@ again:
 		int i;
 
 		spin_lock(&vb->lock);
+		if (vb->free < 1UL << order)
+			goto next;
+
 		i = bitmap_find_free_region(vb->alloc_map,
 						VMAP_BBMAP_BITS, order);
 
-		if (i >= 0) {
-			addr = vb->va->va_start + (i << PAGE_SHIFT);
-			BUG_ON(addr_to_vb_idx(addr) !=
-					addr_to_vb_idx(vb->va->va_start));
-			vb->free -= 1UL << order;
-			if (vb->free == 0) {
-				spin_lock(&vbq->lock);
-				list_del_init(&vb->free_list);
-				spin_unlock(&vbq->lock);
+		if (i < 0) {
+			if (vb->free + vb->dirty == VMAP_BBMAP_BITS) {
+				/* fragmented and no outstanding allocations */
+				BUG_ON(vb->dirty != VMAP_BBMAP_BITS);
+				purge = 1;
 			}
-			spin_unlock(&vb->lock);
-			break;
+			goto next;
+		}
+		addr = vb->va->va_start + (i << PAGE_SHIFT);
+		BUG_ON(addr_to_vb_idx(addr) !=
+				addr_to_vb_idx(vb->va->va_start));
+		vb->free -= 1UL << order;
+		if (vb->free == 0) {
+			spin_lock(&vbq->lock);
+			list_del_rcu(&vb->free_list);
+			spin_unlock(&vbq->lock);
 		}
 		spin_unlock(&vb->lock);
+		break;
+next:
+		spin_unlock(&vb->lock);
 	}
+
+	if (purge)
+		purge_fragmented_blocks_thiscpu();
+
 	put_cpu_var(vmap_block_queue);
 	rcu_read_unlock();
 
@@ -862,11 +924,11 @@ static void vb_free(const void *addr, unsigned long size)
 	BUG_ON(!vb);
 
 	spin_lock(&vb->lock);
-	bitmap_allocate_region(vb->dirty_map, offset >> PAGE_SHIFT, order);
+	BUG_ON(bitmap_allocate_region(vb->dirty_map, offset >> PAGE_SHIFT, order));
 
 	vb->dirty += 1UL << order;
 	if (vb->dirty == VMAP_BBMAP_BITS) {
-		BUG_ON(vb->free || !list_empty(&vb->free_list));
+		BUG_ON(vb->free);
 		spin_unlock(&vb->lock);
 		free_vmap_block(vb);
 	} else
@@ -1035,8 +1097,6 @@ void __init vmalloc_init(void)
 		vbq = &per_cpu(vmap_block_queue, i);
 		spin_lock_init(&vbq->lock);
 		INIT_LIST_HEAD(&vbq->free);
-		INIT_LIST_HEAD(&vbq->dirty);
-		vbq->nr_dirty = 0;
 	}
 
 	/* Import existing vmlist entries. */
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 885207a6b6b..c26986c85ce 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -1922,6 +1922,9 @@ static int sleeping_prematurely(pg_data_t *pgdat, int order, long remaining)
 		if (!populated_zone(zone))
 			continue;
 
+		if (zone_is_all_unreclaimable(zone))
+			continue;
+
 		if (!zone_watermark_ok(zone, order, high_wmark_pages(zone),
 								0, 0))
 			return 1;