We have to recompute pgoff if the given page is huge, since result based on

HPAGE_SIZE is inappropriate for scanning the vma interval tree, as shown

by commit 36e4f20af833(hugetlb: do not use vma_hugecache_offset() for

vma_prio_tree_foreach)

Signed-off-by: Hillf Danton <dhillf@gmail.com>

---

--- a/mm/rmap.c Mon Mar  4 20:00:00 2013

+++ b/mm/rmap.c Mon Mar  4 20:02:16 2013

@@ -1513,6 +1513,9 @@ static int try_to_unmap_file(struct page

  unsigned long max_nl_size = 0;

  unsigned int mapcount;

 

+ if (PageHuge(page))

+ pgoff = page->index << compound_order(page);

+

  mutex_lock(&mapping->i_mmap_mutex);

  vma_interval_tree_foreach(vma, &mapping->i_mmap, pgoff, pgoff) {

  unsigned long address = vma_address(page, vma);

--