[PATCH] mm: vma_adjust insert file earlier

For those arches (arm and parisc) which use the i_mmap tree to implement
flush_dcache_page, during split_vma there's a small window in vma_adjust when
flush_dcache_mmap_lock is dropped, and pages in the split-off part of the vma
might for an instant be invisible to __flush_dcache_page.

Though we're more solid there than ever before, I guess it's a bad idea to
leave that window: so (with regret, it was structurally nicer before) take
__vma_link_file (and vma_prio_tree_init) out of __vma_link.

vma_prio_tree_init (which NULLs a few fields) is actually only needed when
copying a vma, not when a new one has just been memset to 0.

__insert_vm_struct is used by nothing but vma_adjust's split_vma case:
co\10mment it accordingly, remove its mark_mm_hugetlb (it can never create
a new kind of vma) and its validate_mm (another follows immediately).

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/mm/mmap.c b/mm/mmap.c
index c77fec7b00c4e877faa13ea49104ad20c67cc1b0..d6fd2fe133197d98fcd2a3d23ea791237e86a90c 100644 (file)
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -293,10 +293,8 @@ __vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
        struct vm_area_struct *prev, struct rb_node **rb_link,
        struct rb_node *rb_parent)
 {
-       vma_prio_tree_init(vma);
        __vma_link_list(mm, vma, prev, rb_parent);
        __vma_link_rb(mm, vma, rb_link, rb_parent);
-       __vma_link_file(vma);
        __anon_vma_link(vma);
 }
 
@@ -312,7 +310,10 @@ static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
        if (mapping)
                spin_lock(&mapping->i_mmap_lock);
        anon_vma_lock(vma);
+
        __vma_link(mm, vma, prev, rb_link, rb_parent);
+       __vma_link_file(vma);
+
        anon_vma_unlock(vma);
        if (mapping)
                spin_unlock(&mapping->i_mmap_lock);
@@ -323,9 +324,9 @@ static void vma_link(struct mm_struct *mm, struct vm_area_struct *vma,
 }
 
 /*
- * Insert vm structure into process list sorted by address and into the
- * inode's i_mmap tree. The caller should hold mm->mmap_sem and
- * ->f_mappping->i_mmap_lock if vm_file is non-NULL.
+ * Helper for vma_adjust in the split_vma insert case:
+ * insert vm structure into list and rbtree and anon_vma,
+ * but it has already been inserted into prio_tree earlier.
  */
 static void
 __insert_vm_struct(struct mm_struct * mm, struct vm_area_struct * vma)
@@ -337,9 +338,7 @@ __insert_vm_struct(struct mm_struct * mm, struct vm_area_struct * vma)
        if (__vma && __vma->vm_start < vma->vm_end)
                BUG();
        __vma_link(mm, vma, prev, rb_link, rb_parent);
-       mark_mm_hugetlb(mm, vma);
        mm->map_count++;
-       validate_mm(mm);
 }
 
 static inline void
@@ -403,6 +402,15 @@ again:                     remove_next = 1 + (end > next->vm_end);
                if (!(vma->vm_flags & VM_NONLINEAR))
                        root = &mapping->i_mmap;
                spin_lock(&mapping->i_mmap_lock);
+               if (insert) {
+                       /*
+                        * Put into prio_tree now, so instantiated pages
+                        * are visible to arm/parisc __flush_dcache_page
+                        * throughout; but we cannot insert into address
+                        * space until vma start or end is updated.
+                        */
+                       __vma_link_file(insert);
+               }
        }
 
        /*
@@ -1463,6 +1471,7 @@ int split_vma(struct mm_struct * mm, struct vm_area_struct * vma,
 
        /* most fields are the same, copy all, and then fixup */
        *new = *vma;
+       vma_prio_tree_init(new);
 
        if (new_below)
                new->vm_end = addr;
@@ -1775,6 +1784,7 @@ struct vm_area_struct *copy_vma(struct vm_area_struct **vmap,
                new_vma = kmem_cache_alloc(vm_area_cachep, SLAB_KERNEL);
                if (new_vma) {
                        *new_vma = *vma;
+                       vma_prio_tree_init(new_vma);
                        pol = mpol_copy(vma_policy(vma));
                        if (IS_ERR(pol)) {
                                kmem_cache_free(vm_area_cachep, new_vma);