Re: [PATCH 4/6] VM: Annotate pagealloc

[Pekka Enberg]

Hi Steve,

On Sat, Aug 30, 2008 at 2:16 AM, Steve VanDeBogart
<vandebo-lkml@xxxxxxxxxxx> wrote:
> Valgrind annotations for valgrind: memory is addressable once it's been
> alloced, and unaddressable when it is freed again.  Can't use malloc-like
> and free-like because valgrind considers a malloc-like chunk indivisible.
>
> Signed-off-by: Steve VanDeBogart <vandebo-lkml@xxxxxxxxxxx>
> ---
>
> Index: linux-2.6.27-rc5/mm/page_alloc.c
> ===================================================================
> --- linux-2.6.27-rc5.orig/mm/page_alloc.c       2008-08-29
> 14:24:27.000000000 -0700
> +++ linux-2.6.27-rc5/mm/page_alloc.c    2008-08-29 14:24:37.000000000 -0700
> @@ -46,6 +46,7 @@
>  #include <linux/page-isolation.h>
>  #include <linux/memcontrol.h>
>  #include <linux/debugobjects.h>
> +#include <linux/memcheck.h>
>
>  #include <asm/tlbflush.h>
>  #include <asm/div64.h>
> @@ -1080,6 +1081,7 @@
>                if (!page)
>                        goto failed;
>        }
> +       VALGRIND_MAKE_MEM_UNDEFINED(page_address(page), PAGE_SIZE << order);

No SHOUTING macros please, and I think you should just pass page and
order here and do the calculation elsewhere.

>
>        __count_zone_vm_events(PGALLOC, zone, 1 << order);
>        zone_statistics(preferred_zone, zone);
> @@ -1679,6 +1681,8 @@
>  void __free_pages(struct page *page, unsigned int order)
>  {
>        if (put_page_testzero(page)) {
> +               VALGRIND_MAKE_MEM_NOACCESS(page_address(page),
> +                                               PAGE_SIZE << order);

Same here.

>                if (order == 0)
>                        free_hot_page(page);
>                else

Other than that, looks good to me.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/