[RFC] mm:add zero_page _mapcount when mapped into user space

Commit Message

Wang, Yalin Dec. 2, 2014, 9:17 a.m. UTC

This patch add/dec zero_page's _mapcount to make sure
the mapcount is correct for zero_page,
so that when read from /proc/kpagecount, zero_page's
mapcount is also correct, userspace process like procrank can
calculate PSS correctly.

Signed-off-by: Yalin Wang <[e-mail]>
---
 mm/memory.c | 2 ++
 1 file changed, 2 insertions(+)

Patch

diff --git a/mm/memory.c b/mm/memory.c
index 3e50383..a6309c1 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -2198,6 +2198,7 @@  gotten:
 		new_page = alloc_zeroed_user_highpage_movable(vma, address);
 		if (!new_page)
 			goto oom;
+		atomic_dec(&pfn_to_page(pte_pfn(orig_pte))->_mapcount);
 	} else {
 		new_page = alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
 		if (!new_page)
@@ -2647,6 +2648,7 @@  static int do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma,
 		page_table = pte_offset_map_lock(mm, pmd, address, &ptl);
 		if (!pte_none(*page_table))
 			goto unlock;
+		atomic_inc(&(pfn_to_page(my_zero_pfn(address)))->_mapcount);
 		goto setpte;
 	}