Re: [RFC][PATCH 2/3] mm/page_poison.c: Enable PAGE_POISONING as a separate option

[Laura Abbott]

On 01/25/2016 10:39 PM, Jianyu Zhan wrote:

On Tue, Jan 26, 2016 at 12:55 AM, Laura Abbott
<labb...@fedoraproject.org> wrote:

--- a/mm/debug-pagealloc.c
+++ b/mm/debug-pagealloc.c
@@ -8,11 +8,5 @@

  void __kernel_map_pages(struct page *page, int numpages, int enable)
  {
-       if (!page_poisoning_enabled())
-               return;
-
-       if (enable)
-               unpoison_pages(page, numpages);
-       else
-               poison_pages(page, numpages);
+       kernel_poison_pages(page, numpages, enable);
  }
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 63358d9..c733421 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1002,6 +1002,7 @@ static bool free_pages_prepare(struct page *page, 
unsigned int order)
                                            PAGE_SIZE << order);
         }
         arch_free_page(page, order);
+       kernel_poison_pages(page, 1 << order, 0);
         kernel_map_pages(page, 1 << order, 0);

         return true;
@@ -1396,6 +1397,7 @@ static int prep_new_page(struct page *page, unsigned int 
order, gfp_t gfp_flags,
         set_page_refcounted(page);

         arch_alloc_page(page, order);
+       kernel_poison_pages(page, 1 << order, 1);
         kernel_map_pages(page, 1 << order, 1);
         kasan_alloc_pages(page, order);

kernel_map_pages() will fall back to page poisoning scheme for
!ARCH_SUPPORTS_DEBUG_PAGEALLOC.

IIUC,  calling kernel_poison_pages() before kernel_map_pages() will be
equivalent to call kernel_poison_pages()
twice?!

Yes, you are absolutely right. In the !ARCH_SUPPORTS_DEBUG_PAGEALLOC
case we shouldn't need to do anything in kernel_map_pages.

Thanks,
Jianyu Zhan

Thanks,
Laura