Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1416035

[PATCHv9-rebased 15/32] thp, mlock: do not mlock PTE-mapped file huge pages

From "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Newsgroups linux.kernel
Subject [PATCHv9-rebased 15/32] thp, mlock: do not mlock PTE-mapped file huge pages
Date 2016-06-07 13:20 +0200
Message-ID <rHnd8-1YN-19@gated-at.bofh.it> (permalink)
References <rH3o5-5Yv-3@gated-at.bofh.it> <rHn3r-1Vt-11@gated-at.bofh.it>
Organization linux.* mail to news gateway

Show all headers | View raw


As with anon THP, we only mlock file huge pages if we can prove that the
page is not mapped with PTE. This way we can avoid mlock leak into
non-mlocked vma on split.

We rely on PageDoubleMap() under lock_page() to check if the the page
may be PTE mapped. PG_double_map is set by page_add_file_rmap() when the
page mapped with PTEs.

Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
---
 include/linux/page-flags.h | 13 ++++++++++++-
 mm/huge_memory.c           | 27 ++++++++++++++++++++-------
 mm/mmap.c                  |  6 ++++++
 mm/page_alloc.c            |  2 ++
 mm/rmap.c                  | 16 ++++++++++++++--
 5 files changed, 54 insertions(+), 10 deletions(-)

diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index 7c8e82ac2eb7..8cf09639185a 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -581,6 +581,17 @@ static inline int PageDoubleMap(struct page *page)
 	return PageHead(page) && test_bit(PG_double_map, &page[1].flags);
 }
 
+static inline void SetPageDoubleMap(struct page *page)
+{
+	VM_BUG_ON_PAGE(!PageHead(page), page);
+	set_bit(PG_double_map, &page[1].flags);
+}
+
+static inline void ClearPageDoubleMap(struct page *page)
+{
+	VM_BUG_ON_PAGE(!PageHead(page), page);
+	clear_bit(PG_double_map, &page[1].flags);
+}
 static inline int TestSetPageDoubleMap(struct page *page)
 {
 	VM_BUG_ON_PAGE(!PageHead(page), page);
@@ -598,7 +609,7 @@ TESTPAGEFLAG_FALSE(TransHuge)
 TESTPAGEFLAG_FALSE(TransCompound)
 TESTPAGEFLAG_FALSE(TransCompoundMap)
 TESTPAGEFLAG_FALSE(TransTail)
-TESTPAGEFLAG_FALSE(DoubleMap)
+PAGEFLAG_FALSE(DoubleMap)
 	TESTSETFLAG_FALSE(DoubleMap)
 	TESTCLEARFLAG_FALSE(DoubleMap)
 #endif
diff --git a/mm/huge_memory.c b/mm/huge_memory.c
index b3b2bf3da167..28e7d963cc1f 100644
--- a/mm/huge_memory.c
+++ b/mm/huge_memory.c
@@ -1438,6 +1438,8 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma,
 		 * We don't mlock() pte-mapped THPs. This way we can avoid
 		 * leaking mlocked pages into non-VM_LOCKED VMAs.
 		 *
+		 * For anon THP:
+		 *
 		 * In most cases the pmd is the only mapping of the page as we
 		 * break COW for the mlock() -- see gup_flags |= FOLL_WRITE for
 		 * writable private mappings in populate_vma_page_range().
@@ -1445,15 +1447,26 @@ struct page *follow_trans_huge_pmd(struct vm_area_struct *vma,
 		 * The only scenario when we have the page shared here is if we
 		 * mlocking read-only mapping shared over fork(). We skip
 		 * mlocking such pages.
+		 *
+		 * For file THP:
+		 *
+		 * We can expect PageDoubleMap() to be stable under page lock:
+		 * for file pages we set it in page_add_file_rmap(), which
+		 * requires page to be locked.
 		 */
-		if (compound_mapcount(page) == 1 && !PageDoubleMap(page) &&
-				page->mapping && trylock_page(page)) {
-			lru_add_drain();
-			if (page->mapping)
-				mlock_vma_page(page);
-			unlock_page(page);
-		}
+
+		if (PageAnon(page) && compound_mapcount(page) != 1)
+			goto skip_mlock;
+		if (PageDoubleMap(page) || !page->mapping)
+			goto skip_mlock;
+		if (!trylock_page(page))
+			goto skip_mlock;
+		lru_add_drain();
+		if (page->mapping && !PageDoubleMap(page))
+			mlock_vma_page(page);
+		unlock_page(page);
 	}
+skip_mlock:
 	page += (addr & ~HPAGE_PMD_MASK) >> PAGE_SHIFT;
 	VM_BUG_ON_PAGE(!PageCompound(page), page);
 	if (flags & FOLL_GET)
diff --git a/mm/mmap.c b/mm/mmap.c
index 02990e7dd70e..daabef097c78 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -2591,6 +2591,12 @@ SYSCALL_DEFINE5(remap_file_pages, unsigned long, start, unsigned long, size,
 		/* drop PG_Mlocked flag for over-mapped range */
 		for (tmp = vma; tmp->vm_start >= start + size;
 				tmp = tmp->vm_next) {
+			/*
+			 * Split pmd and munlock page on the border
+			 * of the range.
+			 */
+			vma_adjust_trans_huge(tmp, start, start + size, 0);
+
 			munlock_vma_pages_range(tmp,
 					max(tmp->vm_start, start),
 					min(tmp->vm_end, start + size));
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index a46547389e53..e32ff3abe9da 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1005,6 +1005,8 @@ static __always_inline bool free_pages_prepare(struct page *page,
 
 		VM_BUG_ON_PAGE(compound && compound_order(page) != order, page);
 
+		if (compound)
+			ClearPageDoubleMap(page);
 		for (i = 1; i < (1 << order); i++) {
 			if (compound)
 				bad += free_tail_pages_check(page, page + i);
diff --git a/mm/rmap.c b/mm/rmap.c
index b78374519bac..26e3e784ad75 100644
--- a/mm/rmap.c
+++ b/mm/rmap.c
@@ -1287,6 +1287,12 @@ void page_add_file_rmap(struct page *page, bool compound)
 		if (!atomic_inc_and_test(compound_mapcount_ptr(page)))
 			goto out;
 	} else {
+		if (PageTransCompound(page)) {
+			VM_BUG_ON_PAGE(!PageLocked(page), page);
+			SetPageDoubleMap(compound_head(page));
+			if (PageMlocked(page))
+				clear_page_mlock(compound_head(page));
+		}
 		if (!atomic_inc_and_test(&page->_mapcount))
 			goto out;
 	}
@@ -1460,8 +1466,14 @@ static int try_to_unmap_one(struct page *page, struct vm_area_struct *vma,
 	 */
 	if (!(flags & TTU_IGNORE_MLOCK)) {
 		if (vma->vm_flags & VM_LOCKED) {
-			/* Holding pte lock, we do *not* need mmap_sem here */
-			mlock_vma_page(page);
+			/* PTE-mapped THP are never mlocked */
+			if (!PageTransCompound(page)) {
+				/*
+				 * Holding pte lock, we do *not* need
+				 * mmap_sem here
+				 */
+				mlock_vma_page(page);
+			}
 			ret = SWAP_MLOCK;
 			goto out_unmap;
 		}
-- 
2.8.1

Back to linux.kernel | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

[PATCHv9 00/32] THP-enabled tmpfs/shmem using compound pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:20 +0200
  [PATCHv9 11/32] thp: skip file huge pmd on copy_huge_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:20 +0200
  [PATCHv9 09/32] thp: handle file pages in split_huge_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:30 +0200
  [PATCHv9 06/32] mm: introduce do_set_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:30 +0200
  [PATCHv9 13/32] thp: run vma_adjust_trans_huge() outside i_mmap_rwsem "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:30 +0200
  [PATCHv9 04/32] mm: postpone page table allocation until we have page to map "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-06 16:30 +0200
  [PATCHv9-rebased 24/32] shmem: add huge pages support "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 25/32] shmem, thp: respect MADV_{NO,}HUGEPAGE for file mappings "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 06/32] mm: introduce do_set_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 16/32] vmscan: split file huge pages before paging them out "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 07/32] thp, vmstats: add counters for huge file pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 23/32] shmem: get_unmapped_area align huge page "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 19/32] filemap: prepare find and delete operations for huge pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 30/32] thp: introduce CONFIG_TRANSPARENT_HUGE_PAGECACHE "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 01/32] thp, mlock: update unevictable-lru.txt "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 13/32] thp: run vma_adjust_trans_huge() outside i_mmap_rwsem "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 04/32] mm: postpone page table allocation until we have page to map "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 28/32] shmem: make shmem_inode_info::lock irq-safe "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 32/32] thp: update Documentation/{vm/transhuge,filesystems/proc}.txt "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 27/32] khugepaged: move up_read(mmap_sem) out of khugepaged_alloc_page() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 10/32] thp: handle file COW faults "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 29/32] khugepaged: add support of collapse for tmpfs/shmem pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 31/32] shmem: split huge pages beyond i_size under memory pressure "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:10 +0200
  [PATCHv9-rebased 02/32] mm: do not pass mm_struct into handle_mm_fault "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 21/32] mm, rmap: account shmem thp pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 12/32] thp: prepare change_huge_pmd() for file thp "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 08/32] thp: support file pages in zap_huge_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 05/32] rmap: support file thp "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 11/32] thp: skip file huge pmd on copy_huge_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
  [PATCHv9-rebased 00/32] THP-enabled tmpfs/shmem using compound pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 22/32] shmem: prepare huge= mount option and sysfs knob "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 15/32] thp, mlock: do not mlock PTE-mapped file huge pages "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 17/32] page-flags: relax policy for PG_mappedtodisk and PG_reclaim "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 20/32] truncate: handle file thp "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 18/32] radix-tree: implement radix_tree_maybe_preload_order() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 09/32] thp: handle file pages in split_huge_pmd() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200
    [PATCHv9-rebased 14/32] thp: file pages support for split_huge_page() "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> - 2016-06-07 13:20 +0200

csiph-web