mm: migration: do not lose soft dirty bit if page is in migration state
authorCyrill Gorcunov <gorcunov@gmail.com>
Wed, 16 Oct 2013 20:46:51 +0000 (13:46 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 17 Oct 2013 04:35:52 +0000 (21:35 -0700)
If page migration is turned on in config and the page is migrating, we
may lose the soft dirty bit.  If fork and mprotect are called on
migrating pages (once migration is complete) pages do not obtain the
soft dirty bit in the correspond pte entries.  Fix it adding an
appropriate test on swap entries.

Signed-off-by: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Pavel Emelyanov <xemul@parallels.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Matt Mackall <mpm@selenic.com>
Cc: Xiao Guangrong <xiaoguangrong@linux.vnet.ibm.com>
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Aneesh Kumar K.V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Mel Gorman <mel@csn.ul.ie>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memory.c
mm/migrate.c
mm/mprotect.c

index ca00039..f7b7692 100644 (file)
@@ -837,6 +837,8 @@ copy_one_pte(struct mm_struct *dst_mm, struct mm_struct *src_mm,
                                         */
                                        make_migration_entry_read(&entry);
                                        pte = swp_entry_to_pte(entry);
+                                       if (pte_swp_soft_dirty(*src_pte))
+                                               pte = pte_swp_mksoft_dirty(pte);
                                        set_pte_at(src_mm, addr, src_pte, pte);
                                }
                        }
index a26bccd..7a7325e 100644 (file)
@@ -161,6 +161,8 @@ static int remove_migration_pte(struct page *new, struct vm_area_struct *vma,
 
        get_page(new);
        pte = pte_mkold(mk_pte(new, vma->vm_page_prot));
+       if (pte_swp_soft_dirty(*ptep))
+               pte = pte_mksoft_dirty(pte);
        if (is_write_migration_entry(entry))
                pte = pte_mkwrite(pte);
 #ifdef CONFIG_HUGETLB_PAGE
index 94722a4..a3af058 100644 (file)
@@ -94,13 +94,16 @@ static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
                        swp_entry_t entry = pte_to_swp_entry(oldpte);
 
                        if (is_write_migration_entry(entry)) {
+                               pte_t newpte;
                                /*
                                 * A protection check is difficult so
                                 * just be safe and disable write
                                 */
                                make_migration_entry_read(&entry);
-                               set_pte_at(mm, addr, pte,
-                                       swp_entry_to_pte(entry));
+                               newpte = swp_entry_to_pte(entry);
+                               if (pte_swp_soft_dirty(oldpte))
+                                       newpte = pte_swp_mksoft_dirty(newpte);
+                               set_pte_at(mm, addr, pte, newpte);
                        }
                        pages++;
                }