HWPOISON, hugetlb: add free check to dequeue_hwpoison_huge_page()

This check is necessary to avoid race between dequeue and allocation,
which can cause a free hugepage to be dequeued twice and get kernel unstable.

Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Acked-by: Mel Gorman <mel@csn.ul.ie>
Reviewed-by: Christoph Lameter <cl@linux.com>
Signed-off-by: Andi Kleen <ak@linux.intel.com>

include/linux/hugetlb.h

@@ -43,7 +43,7 @@ int hugetlb_reserve_pages(struct inode *inode, long from, long to,
 						struct vm_area_struct *vma,
 						int acctflags);
 void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed);
-void __isolate_hwpoisoned_huge_page(struct page *page);
+int dequeue_hwpoisoned_huge_page(struct page *page);
 void copy_huge_page(struct page *dst, struct page *src);
 
 extern unsigned long hugepages_treat_as_movable;
@@ -102,7 +102,7 @@ static inline void hugetlb_report_meminfo(struct seq_file *m)
 #define hugetlb_free_pgd_range(tlb, addr, end, floor, ceiling) ({BUG(); 0; })
 #define hugetlb_fault(mm, vma, addr, flags)	({ BUG(); 0; })
 #define huge_pte_offset(mm, address)	0
-#define __isolate_hwpoisoned_huge_page(page)	0
+#define dequeue_hwpoisoned_huge_page(page)	0
 static inline void copy_huge_page(struct page *dst, struct page *src)
 {
 }

mm/hugetlb.c

@@ -2955,18 +2955,39 @@ void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed)
 	hugetlb_acct_memory(h, -(chg - freed));
 }
 
+/* Should be called in hugetlb_lock */
+static int is_hugepage_on_freelist(struct page *hpage)
+{
+	struct page *page;
+	struct page *tmp;
+	struct hstate *h = page_hstate(hpage);
+	int nid = page_to_nid(hpage);
+
+	list_for_each_entry_safe(page, tmp, &h->hugepage_freelists[nid], lru)
+		if (page == hpage)
+			return 1;
+	return 0;
+}
+
+#ifdef CONFIG_MEMORY_FAILURE
 /*
  * This function is called from memory failure code.
  * Assume the caller holds page lock of the head page.
  */
-void __isolate_hwpoisoned_huge_page(struct page *hpage)
+int dequeue_hwpoisoned_huge_page(struct page *hpage)
 {
 	struct hstate *h = page_hstate(hpage);
 	int nid = page_to_nid(hpage);
+	int ret = -EBUSY;
 
 	spin_lock(&hugetlb_lock);
-	list_del(&hpage->lru);
-	h->free_huge_pages--;
-	h->free_huge_pages_node[nid]--;
+	if (is_hugepage_on_freelist(hpage)) {
+		list_del(&hpage->lru);
+		h->free_huge_pages--;
+		h->free_huge_pages_node[nid]--;
+		ret = 0;
+	}
 	spin_unlock(&hugetlb_lock);
+	return ret;
 }
+#endif

mm/memory-failure.c

@@ -698,6 +698,7 @@ static int me_swapcache_clean(struct page *p, unsigned long pfn)
  */
 static int me_huge_page(struct page *p, unsigned long pfn)
 {
+	int res = 0;
 	struct page *hpage = compound_head(p);
 	/*
 	 * We can safely recover from error on free or reserved (i.e.
@@ -710,8 +711,9 @@ static int me_huge_page(struct page *p, unsigned long pfn)
 	 * so there is no race between isolation and mapping/unmapping.
 	 */
 	if (!(page_mapping(hpage) || PageAnon(hpage))) {
-		__isolate_hwpoisoned_huge_page(hpage);
-		return RECOVERED;
+		res = dequeue_hwpoisoned_huge_page(hpage);
+		if (!res)
+			return RECOVERED;
 	}
 	return DELAYED;
 }