[PATCH] vmtrunc: truncate_count not atomic

Why is mapping->truncate_count atomic?  It's incremented inside i_mmap_lock
(and i_sem), and the reads don't need it to be atomic.

And why smp_rmb() before call to ->nopage?  The compiler cannot reorder the
initial assignment of sequence after the call to ->nopage, and no cpu (yet!)
can read from the future, which is all that matters there.

And delete totally bogus reset of truncate_count from blkmtd add_device.
truncate_count is all about detecting i_size changes: i_size does not change
there; and if it did, the count should be incremented not reset.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

4 files changed, 6 insertions, 10 deletions

diff --git a/drivers/mtd/devices/blkmtd.c b/drivers/mtd/devices/blkmtd.c
index 80d3393ee817..199824a6c3de 100644
--- a/drivers/mtd/devices/blkmtd.c
+++ b/drivers/mtd/devices/blkmtd.c
@@ -661,7 +661,6 @@ static struct blkmtd_dev *add_device(char *devname, int readonly, int erase_size
 
 	memset(dev, 0, sizeof(struct blkmtd_dev));
 	dev->blkdev = bdev;
-	atomic_set(&(dev->blkdev->bd_inode->i_mapping->truncate_count), 0);
 	if(!readonly) {
 		init_MUTEX(&dev->wrbuf_mutex);
 	}
diff --git a/fs/inode.c b/fs/inode.c
index b6066773ad82..0beebc00a645 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -198,7 +198,6 @@ void inode_init_once(struct inode *inode)
 	INIT_RADIX_TREE(&inode->i_data.page_tree, GFP_ATOMIC);
 	spin_lock_init(&inode->i_data.tree_lock);
 	spin_lock_init(&inode->i_data.i_mmap_lock);
-	atomic_set(&inode->i_data.truncate_count, 0);
 	INIT_LIST_HEAD(&inode->i_data.private_list);
 	spin_lock_init(&inode->i_data.private_lock);
 	INIT_RAW_PRIO_TREE_ROOT(&inode->i_data.i_mmap);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 2d9885aa2626..cd0b31b4afba 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -341,7 +341,7 @@ struct address_space {
 	struct prio_tree_root	i_mmap;		/* tree of private and shared mappings */
 	struct list_head	i_mmap_nonlinear;/*list VM_NONLINEAR mappings */
 	spinlock_t		i_mmap_lock;	/* protect tree, count, list */
-	atomic_t		truncate_count;	/* Cover race condition with truncate */
+	unsigned int		truncate_count;	/* Cover race condition with truncate */
 	unsigned long		nrpages;	/* number of total pages */
 	pgoff_t			writeback_index;/* writeback starts here */
 	struct address_space_operations *a_ops;	/* methods */
diff --git a/mm/memory.c b/mm/memory.c
index 6224cb1933c1..5c6a4db990ac 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -1424,7 +1424,7 @@ void unmap_mapping_range(struct address_space *mapping,
 
 	spin_lock(&mapping->i_mmap_lock);
 	/* Protect against page fault */
-	atomic_inc(&mapping->truncate_count);
+	mapping->truncate_count++;
 
 	if (unlikely(!prio_tree_empty(&mapping->i_mmap)))
 		unmap_mapping_range_list(&mapping->i_mmap, &details);
@@ -1726,7 +1726,7 @@ do_no_page(struct mm_struct *mm, struct vm_area_struct *vma,
 	struct page * new_page;
 	struct address_space *mapping = NULL;
 	pte_t entry;
-	int sequence = 0;
+	unsigned int sequence = 0;
 	int ret = VM_FAULT_MINOR;
 	int anon = 0;
 
@@ -1738,9 +1738,8 @@ do_no_page(struct mm_struct *mm, struct vm_area_struct *vma,
 
 	if (vma->vm_file) {
 		mapping = vma->vm_file->f_mapping;
-		sequence = atomic_read(&mapping->truncate_count);
+		sequence = mapping->truncate_count;
 	}
-	smp_rmb();  /* Prevent CPU from reordering lock-free ->nopage() */
 retry:
 	cond_resched();
 	new_page = vma->vm_ops->nopage(vma, address & PAGE_MASK, &ret);
@@ -1774,9 +1773,8 @@ retry:
 	 * invalidated this page.  If unmap_mapping_range got called,
 	 * retry getting the page.
 	 */
-	if (mapping &&
-	      (unlikely(sequence != atomic_read(&mapping->truncate_count)))) {
-		sequence = atomic_read(&mapping->truncate_count);
+	if (mapping && unlikely(sequence != mapping->truncate_count)) {
+		sequence = mapping->truncate_count;
 		spin_unlock(&mm->page_table_lock);
 		page_cache_release(new_page);
 		goto retry;