hugetlb/cgroup: assign the page hugetlb cgroup when we move the page to active list.

A page's hugetlb cgroup assignment and movement to the active list should
occur with hugetlb_lock held.  Otherwise when we remove the hugetlb cgroup
we will iterate the active list and find pages with NULL hugetlb cgroup
values.

Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Reviewed-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

2 files changed, 13 insertions, 14 deletions

diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ec7b86ebf9d9..c39e4beeb63a 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -928,14 +928,8 @@ struct page *alloc_huge_page_node(struct hstate *h, int nid)
         page = dequeue_huge_page_node(h, nid);
         spin_unlock(&hugetlb_lock);
 
-        if (!page) {
+        if (!page)
                 page = alloc_buddy_huge_page(h, nid);
-                if (page) {
-                        spin_lock(&hugetlb_lock);
-                        list_move(&page->lru, &h->hugepage_activelist);
-                        spin_unlock(&hugetlb_lock);
-                }
-        }
 
         return page;
 }
@@ -1150,9 +1144,13 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma,
         }
         spin_lock(&hugetlb_lock);
         page = dequeue_huge_page_vma(h, vma, addr, avoid_reserve);
-        spin_unlock(&hugetlb_lock);
-
-        if (!page) {
+        if (page) {
+                /* update page cgroup details */
+                hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h),
+                                             h_cg, page);
+                spin_unlock(&hugetlb_lock);
+        } else {
+                spin_unlock(&hugetlb_lock);
                 page = alloc_buddy_huge_page(h, NUMA_NO_NODE);
                 if (!page) {
                         hugetlb_cgroup_uncharge_cgroup(idx,
@@ -1162,6 +1160,8 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma,
                         return ERR_PTR(-ENOSPC);
                 }
                 spin_lock(&hugetlb_lock);
+                hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h),
+                                             h_cg, page);
                 list_move(&page->lru, &h->hugepage_activelist);
                 spin_unlock(&hugetlb_lock);
         }
@@ -1169,8 +1169,6 @@ static struct page *alloc_huge_page(struct vm_area_struct *vma,
         set_page_private(page, (unsigned long)spool);
 
         vma_commit_reservation(h, vma, addr);
-        /* update page cgroup details */
-        hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h), h_cg, page);
         return page;
 }
 
diff --git a/mm/hugetlb_cgroup.c b/mm/hugetlb_cgroup.c
index 680e4819e077..9834a01c79dc 100644
--- a/mm/hugetlb_cgroup.c
+++ b/mm/hugetlb_cgroup.c
@@ -213,6 +213,7 @@ done:
         return ret;
 }
 
+/* Should be called with hugetlb_lock held */
 void hugetlb_cgroup_commit_charge(int idx, unsigned long nr_pages,
                                   struct hugetlb_cgroup *h_cg,
                                   struct page *page)
@@ -220,9 +221,7 @@ void hugetlb_cgroup_commit_charge(int idx, unsigned long nr_pages,
         if (hugetlb_cgroup_disabled() || !h_cg)
                 return;
 
-        spin_lock(&hugetlb_lock);
         set_hugetlb_cgroup(page, h_cg);
-        spin_unlock(&hugetlb_lock);
         return;
 }
 
@@ -389,6 +388,7 @@ int __init hugetlb_cgroup_file_init(int idx)
 void hugetlb_cgroup_migrate(struct page *oldhpage, struct page *newhpage)
 {
         struct hugetlb_cgroup *h_cg;
+        struct hstate *h = page_hstate(oldhpage);
 
         if (hugetlb_cgroup_disabled())
                 return;
@@ -401,6 +401,7 @@ void hugetlb_cgroup_migrate(struct page *oldhpage, struct page *newhpage)
 
         /* move the h_cg details to new cgroup */
         set_hugetlb_cgroup(newhpage, h_cg);
+        list_move(&newhpage->lru, &h->hugepage_activelist);
         spin_unlock(&hugetlb_lock);
         cgroup_release_and_wakeup_rmdir(&h_cg->css);
         return;