mm: fix crashes from mbind() merging vmas

In v2.6.34 commit 9d8cebd4bcd7 ("mm: fix mbind vma merge problem")
introduced vma merging to mbind(), but it should have also changed the
convention of passing start vma from queue_pages_range() (formerly
check_range()) to new_vma_page(): vma merging may have already freed
that structure, resulting in BUG at mm/mempolicy.c:1738 and probably
worse crashes.

Fixes: 9d8cebd4bcd7 ("mm: fix mbind vma merge problem")
Reported-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Tested-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Christoph Lameter <cl@linux.com>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Minchan Kim <minchan.kim@gmail.com>
Cc: <stable@vger.kernel.org>	[2.6.34+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

1 files changed, 20 insertions, 26 deletions

diff --git a/mm/mempolicy.c b/mm/mempolicy.c
index 284974230459..eb58de19f815 100644
--- a/mm/mempolicy.c
+++ b/mm/mempolicy.c
@@ -656,19 +656,18 @@ static unsigned long change_prot_numa(struct vm_area_struct *vma,
  * @nodes and @flags,) it's isolated and queued to the pagelist which is
  * passed via @private.)
  */
-static struct vm_area_struct *
+static int
 queue_pages_range(struct mm_struct *mm, unsigned long start, unsigned long end,
                 const nodemask_t *nodes, unsigned long flags, void *private)
 {
-        int err;
-        struct vm_area_struct *first, *vma, *prev;
-
+        int err = 0;
+        struct vm_area_struct *vma, *prev;
 
-        first = find_vma(mm, start);
-        if (!first)
-                return ERR_PTR(-EFAULT);
+        vma = find_vma(mm, start);
+        if (!vma)
+                return -EFAULT;
         prev = NULL;
-        for (vma = first; vma && vma->vm_start < end; vma = vma->vm_next) {
+        for (; vma && vma->vm_start < end; vma = vma->vm_next) {
                 unsigned long endvma = vma->vm_end;
 
                 if (endvma > end)
@@ -678,9 +677,9 @@ queue_pages_range(struct mm_struct *mm, unsigned long start, unsigned long end,
 
                 if (!(flags & MPOL_MF_DISCONTIG_OK)) {
                         if (!vma->vm_next && vma->vm_end < end)
-                                return ERR_PTR(-EFAULT);
+                                return -EFAULT;
                         if (prev && prev->vm_end < vma->vm_start)
-                                return ERR_PTR(-EFAULT);
+                                return -EFAULT;
                 }
 
                 if (flags & MPOL_MF_LAZY) {
@@ -694,15 +693,13 @@ queue_pages_range(struct mm_struct *mm, unsigned long start, unsigned long end,
 
                         err = queue_pages_pgd_range(vma, start, endvma, nodes,
                                                 flags, private);
-                        if (err) {
-                                first = ERR_PTR(err);
+                        if (err)
                                 break;
-                        }
                 }
 next:
                 prev = vma;
         }
-        return first;
+        return err;
 }
 
 /*
@@ -1156,16 +1153,17 @@ out:
 
 /*
  * Allocate a new page for page migration based on vma policy.
- * Start assuming that page is mapped by vma pointed to by @private.
+ * Start by assuming the page is mapped by the same vma as contains @start.
  * Search forward from there, if not.  N.B., this assumes that the
  * list of pages handed to migrate_pages()--which is how we get here--
  * is in virtual address order.
  */
-static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
+static struct page *new_page(struct page *page, unsigned long start, int **x)
 {
-        struct vm_area_struct *vma = (struct vm_area_struct *)private;
+        struct vm_area_struct *vma;
         unsigned long uninitialized_var(address);
 
+        vma = find_vma(current->mm, start);
         while (vma) {
                 address = page_address_in_vma(page, vma);
                 if (address != -EFAULT)
@@ -1195,7 +1193,7 @@ int do_migrate_pages(struct mm_struct *mm, const nodemask_t *from,
         return -ENOSYS;
 }
 
-static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
+static struct page *new_page(struct page *page, unsigned long start, int **x)
 {
         return NULL;
 }
@@ -1205,7 +1203,6 @@ static long do_mbind(unsigned long start, unsigned long len,
                      unsigned short mode, unsigned short mode_flags,
                      nodemask_t *nmask, unsigned long flags)
 {
-        struct vm_area_struct *vma;
         struct mm_struct *mm = current->mm;
         struct mempolicy *new;
         unsigned long end;
@@ -1271,11 +1268,9 @@ static long do_mbind(unsigned long start, unsigned long len,
         if (err)
                 goto mpol_out;
 
-        vma = queue_pages_range(mm, start, end, nmask,
+        err = queue_pages_range(mm, start, end, nmask,
                           flags | MPOL_MF_INVERT, &pagelist);
-
-        err = PTR_ERR(vma);     /* maybe ... */
-        if (!IS_ERR(vma))
+        if (!err)
                 err = mbind_range(mm, start, end, new);
 
         if (!err) {
@@ -1283,9 +1278,8 @@ static long do_mbind(unsigned long start, unsigned long len,
 
                 if (!list_empty(&pagelist)) {
                         WARN_ON_ONCE(flags & MPOL_MF_LAZY);
-                        nr_failed = migrate_pages(&pagelist, new_vma_page,
-                                        NULL, (unsigned long)vma,
-                                        MIGRATE_SYNC, MR_MEMPOLICY_MBIND);
+                        nr_failed = migrate_pages(&pagelist, new_page, NULL,
+                                start, MIGRATE_SYNC, MR_MEMPOLICY_MBIND);
                         if (nr_failed)
                                 putback_movable_pages(&pagelist);
                 }